Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L.

ICU Book, 2nd Editio n Chapter 1 CIRCULATORY BLOOD FLOW When is a piece of matter said to be alive? When it goes on doing something, movin g, exchanging material with its environment. --Erwin Schrodinger The adult human being has an estimated 100 trillion cells that must go on exchanging material w ith the external environment to stay alive. To accomplish this, the circulatory system includes a vascular network that stretches more than 60,000 miles (more t han twice the circumference of the Earth), and an average of 8000 liters of bloo d is pumped through this vascular network every day (1). This chapter describes the flow of blood through the circulatory system, and includes a description of flow through the heart (cardiac output) and flow through distant regions of the vascular circuit (peripheral blood flow). Most of these concepts are old friends from the physiology classroom, but this chapter applies them to actual practice at the bedside. CARDIAC OUTPUT Circulatory flow originates in the muscular contractions of the heart. Because b lood is an incompressible fluid that flows through a closed hydraulic loop, the volume of blood ejected by the left side of the heart (in a given time period) m ust equal the volume of blood returning to the right side of the heart (over the same period of time). This reflection of the conservation of mass (volume) in a closed hydraulic system is known as the principle of continuity (2). It predict s that the volume flow of blood (volumetric flow rate), which is determined by t he stroke output of the heart, will be the same at all points along the circulat ory system. Therefore, the forces that determine cardiac stroke output also dete rmine volumetric blood flow. The determinants of cardiac stroke output that can be measured or derived in a clinical setting are shown P.4 in Table 1.1. Each of these is described briefly in the paragraphs that follow. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (1 of 18)03-May-05 17:25:32

Ovid: ICU Book TABLE 1.1. REFERENCE RANGES FOR HEMODYNAMIC PARAMETERS IN ADULTS PRELOAD When a weight is attached to one end of a resting muscle, the muscle stretches t o a new length. The weight in this situation represents a force called the prelo ad, that is, the load imposed on a muscle before the onset of contraction. The p reload force acts indirectly to augment the force of muscle contraction. That is , the preload force stretches the muscle to a new resting length and (according to the lengthtension relationship of muscle) the increase in muscle length then l eads to a more forceful muscle contraction. PressureVolume Curves In the intact heart, the stretch imposed on the cardiac muscle at rest is a func tion of the volume in the ventricles at the end of diastole. Therefore, ventricu lar end-diastolic volume (EDV) is used as a reflection of the preload force for the intact heart (3). The pressurevolume curves in Figure 1.1 describe the influe nce of preload on the mechanical performance of the left ventricle during diasto le (lower curves) and systole (upper curves). The solid curves represent the nor mal pressurevolume relationships for diastole and systole. Note that the uppermos t curve in the figure has a rapid ascent, indicating that small changes in diast olic volume are associated with large changes in systolic pressure. The normal r elationship between diastolic volume (preload) and the strength of ventricular c ontraction was described independently by Otto Frank and Ernest Starling, and is commonly known as the FrankStarling phenomenon (3). This relationship can be restated as follows: http://gateway.ut.ovid.com/gw1/ovidweb.cgi (2 of 18)03-May-05 17:25:32

Ovid: ICU Book Figure 1.1. Pressurevolume curves for the intact ventricle. Solid lines represent normal pressurevolume relationships. In the normal heart, the diastolic volume (preload) is the principal force that governs the strength of ventricular contraction. P.5 This indicates that the stroke output of the normal heart is primarily a ref lection of the diastolic volume. Therefore, the most effective measure for prese rving the cardiac output is to maintain an adequate diastolic volume. This empha sizes the value of avoiding hypovolemia and correcting volume deficits when they exist. Ventricular Function Curves Ventricular end-diastolic volume is not easily measured at the bedside, and the ventricular enddiastolic pressure (EDP) is more commonly used as a reflection of ventricular preload in clinical practice (see Chapter 11 for more information o n end-diastolic pressure). The relationship between end-diastolic pressure (prel oad) and stroke volume (systolic performance) is used to monitor the FrankStarlin g relationship in the clinical setting. The curves that define this relationship , known as ventricular function curves (4), are shown in Figure 1.2. Unfortunate ly, the interpretation of ventricular function curves can be misleading, as is d emonstrated in the sections http://gateway.ut.ovid.com/gw1/ovidweb.cgi (3 of 18)03-May-05 17:25:32

Ovid: ICU Book that follow. Figure 1.2. Ventricular function curves. Ventricular Compliance The stretch imposed on cardiac muscle is determined not only by the volume of bl ood in the ventricular chambers, but also by the tendency of the ventricular wal l to distend or stretch at any given chamber P.6 volume. This latter property is described as the compliance (distensibility) of the ventricle. Compliance is de fined by the following relationship between changes in EDP and volume (EDV) (5): The lower curves in Figure 1.1 show the end-diastolic pressurevolume relationship s for the normal ventricle and a noncompliant (stiff) ventricle. As the ventricl e becomes less compliant (e. g., when the ventricle hypertrophies), there is les s of a change in diastolic volume relative to the change in diastolic pressure. Early in this process, the EDV remains normal, but the EDP increases above norma l. As the compliance decreases further, the increase in EDP eventually reduces v enous inflow into the heart, thereby causing a reduction in EDV. The decrease in EDV then leads to a decrease in the force of ventricular contraction. This illu strates how changes in http://gateway.ut.ovid.com/gw1/ovidweb.cgi (4 of 18)03-May-05 17:25:32

Ovid: ICU Book ventricular compliance can lead to changes in cardiac stroke output, and how cha nges in cardiac stroke output can be independent of changes in systolic function . The decrease in stroke output that accompanies a decrease in ventricular compl iance is known as diastolic heart failure (6). The difference between heart fail ure caused by systolic and diastolic dysfunction is presented in Chapter 16. The Preload Measurement Changes in ventricular compliance also influence the reliability of EDP as a ref lection of EDV. For example, a decrease in ventricular P.7 compliance results in a higher-than-expected EDP at any given EDV. Therefore, the EDP overestimates t he actual preload (EDV) when the ventricle is noncompliant. The following points are important to remember when EDP is used as an index of ventricular preload: EDP provides an accurate reflection of preload only when ventricular compliance is normal. Changes in EDP provide an accurate reflection of changes in preload o nly when ventricular compliance is constant. The influence of ventricular compliance on the assessment of preload surfaces ag ain in Chapter 11. Chapter 16 describes the importance of an accurate preload me asurement in distinguishing systolic from diastolic forms of heart failure. AFTERLOAD When a weight is attached to one end of a contracting muscle, the force of muscl e contraction must overcome the opposing force of the weight before the muscle b egins to shorten its length. The weight in this situation represents a force cal led the afterload, the load imposed on the muscle after the onset of contraction . The afterload is an opposing force that determines the force of muscle contrac tion needed to initiate muscle shortening (i.e., isotonic muscle contraction). I n the intact heart, the afterload force is equivalent to the tension developed a cross the wall of the ventricles during systole (3). The determinants of ventric ular wall tension are derived from observations on soap bubbles made by the Marq uis de Laplace in 1820. These observations were the basis for the law of Laplace , which states that the tension across a thin-walled sphere is directly related to the internal pressure and radius of the sphere: T = Pr. Because the ventricle s are not thin-walled spheres, the Laplace relationship for the intact heart inc orporates a factor that reflects the http://gateway.ut.ovid.com/gw1/ovidweb.cgi (5 of 18)03-May-05 17:25:32

Ovid: ICU Book average thickness of the ventricular wall (5). The law of Laplace applied to the intact heart is then expressed as T = Pr/t, where T represents the tension acro ss the wall of the ventricle during systole, P represents the transmural pressur e across the ventricle at the end of systole, r represents the chamber radius at the end of diastole, and t represents the average thickness of the ventricular wall. The forces that contribute to ventricular wall tension are shown in Figure 1.3. Figure 1.3. The forces that contribute to ventricular afterload. Forces enclosed in rectangles are readily defined or measured. Pleural Pressures Because afterload is a transmural force, it is influenced by the pleural pressur es at the outer surface of heart. Negative pleural pressures increase transmural pressure and increase ventricular afterload, whereas positive pleural pressures have the opposite effect. Negative pressures surrounding the heart can impede v entricular emptying by opposing the inward http://gateway.ut.ovid.com/gw1/ovidweb.cgi (6 of 18)03-May-05 17:25:32

Ovid: ICU Book displacement of the ventricular wall during systole (7). P.8 This action is resp onsible for the decrease in systolic blood pressure (reflecting a decrease in st roke volume) that occurs during the inspiratory phase of spontaneous breathing. When this inspiratory-related drop in pressure is greater than 15 mm Hg, the con dition is called pulsus paradoxus (which is a misnomer, because the response is not paradoxical, but is an exaggerated version of the normal response). Positive pleural pressures can promote ventricular emptying by facilitating the inward d isplacement of the ventricular wall during systole (7). Rapid and forceful rises in positive pressure surrounding the heart might also produce a massagelike act ion to expel blood from the heart and great vessels in the thorax. This is the p roposed explanation for the success of cough CPR, which uses forceful coughing t o P.9 maintain circulatory flow in patients with ventricular tachycardia (8). In fact, positive pleural pressure swings may be responsible for the hemodynamic e ffects of closed chest cardiac massage, as discussed in Chapter 18 (8). Impedance versus Resistance A major component of afterload is the resistance to ventricular outflow in the a orta and large, proximal arteries. The total hydraulic force that opposes pulsat ile flow is known as impedance (9). This force is a combination of two forces: ( a) a force that opposes the rate of change in flow, known as compliance, and (b) a force that opposes mean or volumetric flow, known as resistance. Vascular com pliance is not easily measured at the bedside (10). On the other hand, vascular resistance is derived by assuming that hydraulic resistance is analogous to elec trical resistance. That is, Ohm's law predicts that resistance to flow of an ele ctric current (R) is directly proportional to the voltage drop across a circuit (E) and inversely proportional to the flow of current (I); R = E/I. The hydrauli c analogy then states that resistance to the flow of fluid through a tube is dir ectly proportional to the pressure drop along a tube (Pin Pout), and inversely p roportional to the flow of volume (Q): This relationship is applied to the systemic and pulmonary circulations, creatin g the following derivations: where MABP is mean arterial blood pressure, CVP is central venous pressure, MPAP is mean pulmonary artery pressure, LAP is left-atrial pressure, and CO is the c ardiac output. Vascular http://gateway.ut.ovid.com/gw1/ovidweb.cgi (7 of 18)03-May-05 17:25:32

Ovid: ICU Book resistance is expressed in units of pressure and flow. Because the pressures are measured in mm Hg, the units would be mm Hg per mL/second. However, the dislike for expressing pressures in mm Hg has led to the common practice of expressing vascular resistance in CGS (centimetergram-second) units, or dynes second/cm5. T he conversion is dynes second/cm5 = 1333 mm Hg/mL/second. Clinical Monitoring Although afterload is a combination of several forces that oppose ventricular em ptying, most of the component forces of afterload cannot be measured easily or r eliably at the bedside. As a result, the vascular resistance, derived as shown a bove, is often used as the sole measure of ventricular afterload. However, as mi ght be expected, vascular resistance is not an accurate measure of total ventric ular afterload (11). P.10 A shift in the height and slope of the ventricular fun ction curve could be an indirect marker of changes in afterload, as shown in Fig ure 1.2. However, shifts in the ventricular function curve can also be caused by changes in the contractile state of the myocardium, and because it is not possi ble to determine whether myocardial contractility is constant using bedside meas urements, a shift in the position of ventricular function curves cannot be used as a marker of changes in afterload. CONTRACTILITY The contraction of striated muscle is attributed to interactions between contrac tile proteins arranged in parallel rows in the sarcomere. The number of bridges formed between adjacent rows of contractile elements determines the contractile state or contractility of the muscle fiber. The contractile state of a muscle is reflected by the force and velocity of muscle contraction (3). The contractile state of cardiac muscle in the intact heart is reflected in the systolic perform ance of the ventricles. This is demonstrated in the upper curves in Figure 1.1. The systolic pressures in this figure reflect isovolumetric contraction (i.e., t he pressures are generated before the aortic valve opens), which eliminates the influence of outflow impedance (afterload) on systolic pressure. Therefore, the changes in systolic pressure at any given diastolic volume (preload constant) re flect changes in the contractile state of the myocardium. Clinical Monitoring Changes in myocardial contractility alter the height and slope of the ventricula r function curve, as demonstrated in Figure 1.2. However, as just mentioned, cha nges in the position of ventricular function curves can also be the result of ch anges in ventricular afterload. Therefore, because it is http://gateway.ut.ovid.com/gw1/ovidweb.cgi (8 of 18)03-May-05 17:25:32

Ovid: ICU Book not possible to monitor afterload to determine whether it is constant, a shift i n the ventricular functions curve is not a reliable method for detecting changes in myocardial contractility (4). PERIPHERAL BLOOD FLOW The cardiac stroke output travels through a vast array of vascular channels that can differ markedly in size. The focus of the remainder of the chapter is the f actors that govern flow through these vascular channels. Caution: The determinants of flow through vascular conduits are derived from idealized hy draulic models that differ considerably from the conditions that exist in the in tact circulatory system. For example, the flow in small tubes usually is steady or nonpulsatile flow, and does not represent the continually changing pulsatile pattern of flow that occurs in many regions of the native circulation. Because o f discrepancies P.11 like this, the description of blood flow that follows shoul d be used more as a qualitative than quantitative description of the hydraulics of vascular flow. FLOW IN RIGID TUBES The hydraulic analogy of Ohm's law, as mentioned previously, states that steady volumetric flow (Q) through a rigid tube is proportional to the pressure gradien t between the inlet and outlet of the tube (Pin Pout), and the constant of propo rtionality is the hydraulic resistance to flow (R): The flow of fluids through small tubes was described independently by a German e ngineer (G. Hagen) and a French physician (J. Poiseuille), and their observation s are combined in the equation shown below, called the HagenPoiseuille equation ( 12,13). This equation identifies the components of hydraulic resistance as the inner rad ius (r) and length of the tube (L), and the viscosity of the fluid ( ). Because th e final term in the HagenPoiseuille equation is the reciprocal of resistance (i.e ., 1/R), the hydraulic resistance to steady, volumetric flow is expressed as The components of the HagenPoisseuille equation are shown in the diagram in Figur e 1.4. Note that flow varies according to the fourth power of the inner radius o f the tube. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (9 of 18)03-May-05 17:25:32

Ovid: ICU Book Figure 1.4. The forces that govern steady flow in rigid tubes. Pin = inlet press ure, Pout = outlet pressure, Q = flow, = viscosity, r = radius, L = length. P.12 Thus, a twofold increase in the inner radius of the tube will result in a s ixteenfold increase in flow: (2r)4 = 16 r. Flow varies much less with the other determinants of resistance; that is, a twofold increase in the length of the tub e or the viscosity of the fluid results in a 50% decrease in flow rate. The infl uence of tube dimensions on flow rate has more practical applications as determi nants of flow through vascular catheters, as presented in Chapter 4. FLOW IN TUBES OF VARYING DIAMETER The HagenPoisseuille equation predicts that as blood moves away from the heart an d encounters vessels of decreasing diameter, the resistance to flow should incre ase and flow rate should decrease. However, the principle of continuity describe d earlier predicts that blood flow will be the same at all points along the vasc ular circuit. This apparent dilemma can be resolved by considering the relations hip between flow velocity and cross-sectional area of a tube. For a rigid tube o f varying diameter, the velocity of flow (v) at any point along the tube is dire ctly proportional to the volumetric flow rate (Q) and inversely proportional to the cross-sectional area of the tube (A). These relationships are described belo w (2). If flow is constant, a decrease in the cross-sectional area of a tube results in an increase in the velocity of flow. This is how the nozzle on a garden hose wo rks, and is the rationale for jet ventilation. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (10 of 18)03-May-05 17:25:32

Ovid: ICU Book Equation 1.8 can be rearranged to yield the relationship Q = v A. This relations hip indicates that proportional changes in velocity and cross-sectional area in opposite directions result in a constant volume flow rate. This means that blood flow can remain unchanged in blood vessels of diminishing diameter if there are equal and opposite changes in the velocity of flow and the crosssectional area of the vessels. The trick here is to use the total cross-sectional area of the v essels in a region rather than the cross-sectional area of individual vessels. T his resolves the discrepancy between the principle of continuity and the HagenPoi seuille relationships. Circulatory Design The graph in Figure 1.5 shows the changes in flow velocity and cross-sectional a rea in different regions of the circulation (13). As expected, when blood moves toward the periphery, there are proportionate and reciprocal changes in cross-se ctional area and velocity of flow. The high velocity of flow in the proximal art eries seems well suited for delivering blood quickly to the microcirculation, to allow more time for diffusional exchange with the tissues. The low velocity and large cross-sectional area in the capillaries are also well-suited for diffusio nal P.13 exchange. These features show a rational design in the circulatory syst em. Figure 1.5. Regional differences in flow velocity and cross-sectional area in th e human circulatory system. (Adapted from Little RC, Little WC. Physiology of th e heart and circulation. 4th ed. Chicago: Year Book, 1989.) http://gateway.ut.ovid.com/gw1/ovidweb.cgi (11 of 18)03-May-05 17:25:32

Ovid: ICU Book FLOW IN COLLAPSIBLE TUBES The hydraulic relationships described above apply to flow through rigid tubes, b ut blood vessels are not rigid tubes. The determinants of flow in collapsible tu bes are explained with the aid of the apparatus shown in Figure 1.6 (14). The ap paratus shows a tube with collapsible walls passing through a fluid reservoir. T he height of the fluid in the reservoir can be adjusted to vary the external pre ssure on the tube. As mentioned earlier, flow in a rigid tube is proportional to the pressure difference between the inlet and outlet of the tube (Pin Pout). Th is is also the case in collapsible tubes as long as the external pressure is not high enough to compress the tube. However, as shown in Figure 1.6, when the ext ernal pressure exceeds the outlet pressure (Pext Pout) and compresses the tube, the driving force for flow is pressure difference between the inlet pressure and the external pressure (Pin Pext). In this situation, the driving pressure for f low is independent of the pressure gradient along the tube. Figure 1.6. The influence of external compression on flow through collapsible tu bes. Pext = extravascular pressure, Pout = pressure at the tube outlet. Pin = in let pressure, Pout = outlet pressure, Pext = external pressure. The Pulmonary Circulation Vascular compression has been demonstrated in the cerebral, pulmonary, and syste mic circulations. Extravascular compression is a particular P.14 concern in pati ents who require positive-pressure mechanical ventilation (14). In this situatio n, http://gateway.ut.ovid.com/gw1/ovidweb.cgi (12 of 18)03-May-05 17:25:32

Ovid: ICU Book pressures in the alveoli can exceed pressures in the underlying pulmonary capill aries, and the resultant capillary compression changes the driving force for flo w across the lungs, as illustrated in Figure 1.6. Thus, whereas the normal drivi ng pressure for flow across the lungs is the difference between the mean pulmona ry artery pressure and the left-atrial pressure (PAP LAP), the driving pressure for flow when pulmonary capillaries are compressed is the difference between the pulmonary artery pressure and the alveolar pressure (PAP Palv). The pulmonary v ascular resistance (PVR) will then differ as follows: The problems created by vascular compression in the lungs are discussed in Chapt er 11 and Chapter 26. VISCOSITY Solids resist being deformed (changing shape), whereas fluids deform continuousl y (i.e., flow) but resist changes in the rate of deformation (i.e., the flow rat e). The inherent resistance of a fluid to changes in flow rate is expressed as t he viscosity of the fluid (12,15). When a force is applied that changes flow rat e (i.e., a shear force), the change in flow rate varies inversely with the visco sity of the fluid. Thus, as the viscosity of a fluid increases, the fluid flows less rapidly in response to a shear force. The influence of viscosity is easily demonstrated by comparing the flow of molasses (high viscosity) and the flow of water P.15 (low viscosity) when the force of gravity is applied (i.e., when both are spilled). Blood Viscosity The viscosity of whole blood is determined by the number and strength of interac tions between plasma fibrinogen and the circulating erythrocytes (15,16). The co ncentration of circulating erythrocytes (i.e., the hematocrit) is the principal determinant of whole blood viscosity. The relationship between hematocrit and bl ood viscosity is shown in Table 1.2. Note that viscosity is expressed in absolut e units (centipoise) and also as a relative value (the ratio of blood viscosity to the viscosity of water). Whole blood with a normal hematocrit (i.e., 40%) has a viscosity that is three to four times higher than that of water. Thus, to mov e whole blood with a normal hematocrit, the circulatory system must generate a p ressure that is three to four times higher than the pressure needed to move wate r the same distance. The acellular blood (zero hematocrit) in Table 1.2 is equiv alent to plasma, and has a viscosity that more closely approximates the viscosit y of water. Thus, moving plasma does not require nearly the work involved in mov ing whole blood. This difference in work load can have significant implications in the patient with coronary disease or limited cardiac reserve. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (13 of 18)03-May-05 17:25:32

Ovid: ICU Book TABLE 1.2. BLOOD VISCOSITY AS A FUNCTION OF HEMATOCRIT Other factors that influence viscosity are body temperature and the flow rate (1 6). Viscosity rises in response to decreases in temperature and flow rate. The i ncrease in blood viscosity in low flow states might represent an adaptive respon se aimed at promoting coagulation at sites of hemorrhage (15). However, the rise in viscosity can also serve to further reduce blood flow and thereby provoke is chemic injury. The tendency of viscosity to increase with decreases in blood flo w is a potential problem in the ICU patient population, and deserves further stu dy. P.16 Hemodynamic Effects The HagenPoisseuille equation predicts that (all other variables constant) blood flow will change in the same proportion as the change in blood viscosity; that i s, if viscosity is reduced by onehalf, blood flow will double (15). The graph in Figure 1.7 demonstrates the hemodynamic effects of a progressive decrease in bl ood viscosity. In this case, the subject was an elderly male with secondary poly cythemia, and the reduction in viscosity was achieved by progressive (isovolemic ) hemodilution. As shown in the graph, the progressive reduction in hematocrit w as associated with a progressive rise in cardiac output. The disproportionate im provement in cardiac output may be caused by the fact that low flow rates can in crease viscosity, and thus an increase in flow could itself produce a further in crease in flow. The ability to modulate blood flow by manipulating the hematocri t is presented in more detail in Chapter 44. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (14 of 18)03-May-05 17:25:32

Ovid: ICU Book Figure 1.7. The influence of progressive hemodilution on cardiac output in a pat ient with polycythemia. CO = cardiac output. (From LeVeen HH et al. Lowering blo od viscosity to overcome vascular resistance. Surg Gynecol Obstet 1980;150:139149 .) Clinical Monitoring Viscosity is measured by placing a fluid sample between two parallel plates that are sliding past each other, and recording the resistance or stickiness in the mo vement of the plates. The instrument that performs this task is called a viscome ter. The units of measurement for viscosity are the poise (or dyne second/cm2) i n the CGS system, P.17 and the pascal second (Pa s) in the SI system. To convert units, use the relationship 1 poise = 0.1 Pa s. Viscosity is also expressed in relative terms (relative to the viscosity of water), a method that may be prefer red for its simplicity. The major drawback in monitoring viscosity is the tenden cy of viscosity to vary with changes in temperature, hematocrit, and flow rate. As a result, local conditions in the microcirculation can produce changes in blo od viscosity that will go undetected in the in vitro (viscometer) measurement of viscosity. In states of adequate blood flow, the measurement is considered to b e reasonably accurate. However, for the critically ill patient with suspected lo w flow who might benefit from measurements of blood viscosity, the reliability o f the measurement is likely to be uncertain. A more feasible application of the viscosity measurement would be to monitor the effects of packed cell transfusion s on blood viscosity to determine the point at which http://gateway.ut.ovid.com/gw1/ovidweb.cgi (15 of 18)03-May-05 17:25:32

Ovid: ICU Book hemoconcentration can be counterproductive in individual patients. REFERENCES GENERAL TEXTS Guyton AC, Jones CE, Coleman TG. Circulatory physiology: cardiac output and its regulation. 2nd ed. Philadelphia: WB Saunders, 1973. Nichols WW, O'Rourke M. McDonald's blood flow in arteries. 3rd ed. Baltimore: Wi lliams & Wilkins, 1990. Berne R, Levy M. Cardiovascular physiology. 6th ed. St. Louis: CV Mosby, 1992. Warltier DC. Ventricular function. Baltimore: Williams & Wilkins, 1995. CARDIAC OUTPUT 1. Vogel S. Vital circuits. New York: Oxford University Press, 1992:117. 2. Vogel S. Life in moving fluids. Princeton, NJ: Princeton University Press, 19 81:2528. 3. Braunwald E, Sonnenblick EH, Ross J Jr. Mechanisms of cardiac contraction and relaxation. In: Braunwald E, ed. Heart disease: a textbook of cardiovascular me dicine. 4th ed. Philadelphia: WB Saunders, 1992;351392. 4. Parmley WM, Talbot L. The heart as a pump. In: Berne RM, ed. Handbook of phys iology. The cardiovascular system. Bethesda, MD: American Physiological Society, 1979;429460. 5. Gilbert JC, Glantz SA. Determinants of left ventricular filling and of the di astolic pressurevolume relation. Circ Res 1989;64:827852. 6. Grossman W. Diastolic dysfunction in congestive heart failure. N Engl J Med http://gateway.ut.ovid.com/gw1/ovidweb.cgi (16 of 18)03-May-05 17:25:32

Ovid: ICU Book 1991;325:15571564. 7. Pinsky MR. Cardiopulmonary interactions: the effects of negative and positive changes in pleural pressures on cardiac output. In: Dantzger DR, ed. Cardiopulm onary critical care. 2nd ed. Philadelphia: WB Saunders, 1991;87120. 8. Weil MH, Gazmuri RJ, Rackow EC. The clinical rationale of cardiac resuscitati on. Dis Mon 1990;36:423468. 9. Finkelstein SM, Collins R. Vascular impedance measurement. Prog Cardiovasc Di s 1982;24:401418. P.18 10. Laskey WK, Parker G, Ferrari VA, et al. Estimation of arterial compliance in humans. J Appl Physiol 1990;69:112119. 11. Lang RM, Borrow KM, Neumann A, et al. Systemic vascular resistance: an unrel iable index of left ventricular afterload. Circulation 1986;74:11141123. PERIPHERAL BLOOD FLOW 12. Chien S, Usami S, Skalak R. Blood flow in small tubes. In: Renkin EM, Michel CC, eds. Handbook of physiology. Section 2: the cardiovascular system. Vol. IV. The microcirculation. Bethesda: American Physiological Society, 1984;217249. 13. Little RC, Little WC. Physiology of the heart and circulation. 4th ed. Chica go: Year Book, 1989;219236. 14. Gorback MS. Problems associated with the determination of pulmonary vascular resistance. J Clin Monit 1990;6:118127. 15. Merrill EW. Rheology of blood. Physiol Rev 1969;49:863888. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (17 of 18)03-May-05 17:25:32

Ovid: ICU Book 16. Lowe GOD. Blood rheology in vitro and in vivo. Bailleres Clin Hematol 1987;1 :597. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (18 of 18)03-May-05 17:25:32

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 2 RESPIRATORY GAS TRANSPORT Respiration is thus a process of combustion, in truth very slow, but otherwise e xactly like that of charcoal. --Antoine Lavoisier One of the basic elements of a erobic life is the combustion reaction, in which oxygen releases the energy stor ed in organic fuels and carbon dioxide is generated as a chemical byproduct. The business of aerobic metabolism is the combustion of nutrient fuels, and the cir culatory system plays a dual role in supporting this process by delivering oxyge n to drive the reaction and removing the carbon dioxide that is generated. Becau se both processes have a common purpose, the transport of oxygen and carbon diox ide in blood has been designated the respiratory function of blood. This chapter describes the basic features of each transport system, and demonstrates the cen tral role of hemoglobin in the transport of both oxygen and carbon dioxide. OXYGEN TRANSPORT Oxygen is the most abundant element on the surface of the earth (1), yet because it does not dissolve readily in water, it is unavailable to cells in the interi or of the body. Thus, we depend on a steady supply of this element to survive, y et we function as a natural barrier to its movement. A possible explanation for this is the role of oxygen in promoting oxidative damage, discussed in Chapter 3 . For now, we will assume that oxygen is a wonderful element. The transport syst em for oxygen is separated into four components: blood oxygen content, oxygen de livery in arterial blood, oxygen uptake from the microcirculation, and oxygen ex traction ratio. WHOLE BLOOD O2 CONTENT The concentration of oxygen in arterial blood (CaO2), often called the oxygen co ntent, is described by Equation 2.1. P.20 http://gateway.ut.ovid.com/gw1/ovidweb.cgi (1 of 15)03-May-05 17:27:15

Ovid: ICU Book The contribution of hemoglobin is described in the first term of the equation: 1 .34 Hb SaO2. This relationship states that each gram of hemoglobin (Hb) will bin d 1.34 mL O2 when it is fully saturated with oxygen. The arterial O2 saturation (SaO2) is expressed as a fraction, not a percentage (i.e., 1.0 instead of 100%). One gram of hemoglobin can actually bind 1.39 mL O2 at full saturation (2). How ever, a small fraction of the circulating hemoglobin is represented by forms tha t do not readily bind oxygen (i.e., methemoglobin and carboxyhemoglobin), so the lower binding capacity of 1.34 mL/ g more accurately describes the behavior of the total pool of circulating hemoglobin. The oxygen bound by hemoglobin at a co ncentration of 15 g/dL and an O2 saturation of 98% will then be Note that because hemoglobin concentration is expressed in g/dL (g/100 mL), the concentration of hemoglobin-bound oxygen is expressed in mL/100 mL (volume %). T he contribution of the oxygen dissolved in plasma is defined by the solubility c oefficients shown in Table 2.1. At a normal body temperature of 37 C, the solubil ity of oxygen in plasma is .028 mL/L/mm Hg. To express the concentration in mL/1 00 mL, the solubility coefficient is divided by 10, creating the second term sho wn in Equation 2.1: 0.003 PaO2. Thus, at a PaO2 of 100 mm Hg, the expected conce ntration of dissolved oxygen is TABLE 2.1. SOLUBILITY OF O2 AND CO2 IN PLASMA The total concentration of oxygen in arterial blood is then 19.7 + 0.3 = 20 mL/1 00 mL, or 200 mL/L. Repeating this calculation using an O2 saturation of 75% der ives the O2 content in mixed venous (pulmonary artery) blood, as shown in Table 2.2. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (2 of 15)03-May-05 17:27:15

Ovid: ICU Book TABLE 2.2. CONCENTRATION OF O2 AND CO2 IN WHOLE BLOOD* A comparison of the total and dissolved O2 concentrations in Table 2.2 shows tha t hemoglobin carries 98.5% of the oxygen in arterial blood and 99.5% in mixed ve nous blood. If we were forced to rely solely on the 3 mL/L of dissolved oxygen i n arterial blood, a cardiac output of P.21 89 L/minute would be necessary to sus tain a normal whole-body O2 consumption of 250 mL/minute. Hemoglobin versus PaO2 To illustrate the relative strength of hemoglobin and PaO2 in determining the O2 content of blood, Table 2.3 shows the relative influence of anemia and hypoxemi a on arterial oxygenation. A 50% reduction in hemoglobin (from 15 to 7.5 mg/dL) is fully expressed as a 50% reduction in CaO2. However, a 50% reduction in PaO2 (from 90 to 45 mm Hg) results in only a 20% decrease in CaO2 (which is similar t o the 18% decrease in SaO2). These examples emphasize the following points: http://gateway.ut.ovid.com/gw1/ovidweb.cgi (3 of 15)03-May-05 17:27:15

Ovid: ICU Book TABLE 2.3. RELATIVE INFLUENCE OF ANEMIA AND HYPOXEMIA ON ARTERIAL OXYGENATION Changes in hemoglobin concentration have a larger impact on arterial oxygenation than changes in PaO2. Hypoxemia (a decrease in PaO2) has a relatively minor imp act on arterial oxygenation if the accompanying change in SaO2 is small. PO2 inf luences blood oxygenation only to the extent that it influences the saturation o f hemoglobin with oxygen. Therefore, SaO2 is a more reliable index of arterial o xygenation than PaO2. P.22 The Hemoglobin Mass The hemoglobin concentration is traditionally expressed in grams per deciliter r ather than in grams per liter, and this tends to create an underappreciation for the size of the hemoglobin pool in blood. For example, a hemoglobin concentrati on of 15 g/dL means that there are 150 grams of hemoglobin per liter of whole bl ood. Thus, a normal blood volume of 5.5 liters contains 0.825 kg, or 1.85 lb, of hemoglobin! To place this in perspective, consider that the normal weight of th e heart is 0.3 kg, or about one-third the mass of circulating hemoglobin. Theref ore, the heart must push three times its weight to move hemoglobin through the c irculatory system. This represents a substantial work load for the heart. The re ason for the size of hemoglobin pool is not clear because there is much more tha n needed for oxygen transport. The answer may be that hemoglobin has other funct ions in addition to oxygen transport, as discussed later in the chapter. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (4 of 15)03-May-05 17:27:15

Ovid: ICU Book OXYGEN DELIVERY (DO2) The oxygen transport parameters are shown in Table 2.4. Transport in arterial bl ood is described by oxygen delivery (DO2), which is defined as the product of th e cardiac output (Q) and the arterial oxygen content (CaO2). TABLE 2.4. PARAMETERS OF OXYGEN AND CARBON DIOXIDE TRANSPORT Note that the dissolved oxygen component is eliminated. The factor 10 converts t he final units to mL/ minute. If the cardiac index (cardiac output divided by th e body surface area) is used to derive the DO2, the units are expressed as mL/mi nute/m2. As shown in Table 2.4, the normal range for DO2 is 520 to 570 mL/minute /m2. OXYGEN UPTAKE (VO2) The oxygen uptake from the microcirculation is a function of the cardiac output and the difference in oxygen content between arterial P.23 and venous blood; tha t is, VO2 = Q (CaO2 CvO2). Because CaO2 and CvO2 share the same term for hemoglo bin binding (1.34 Hb), this term can be isolated to create Equation 2.5. The factor 1.34 has been multiplied by 10 to convert units. This relationship is depicted schematically in Figure 2.1. As indicated in Table 2.4, the normal ran ge for the VO2 is 110 160 mL/minute/m2. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (5 of 15)03-May-05 17:27:15

Ovid: ICU Book Figure 2.1. Schematic view of oxygen uptake (VO2) from the microcirculation. SaO 2 = arterial oxygen saturation, SvO2 = venous oxygen saturation, Hb = hemoglobin , PO2 = oxygen pressure. OXYGEN EXTRACTION RATIO (O2ER) The oxygen extraction ratio (O2ER) is the ratio of O2 uptake to O2 delivery (VO2 /DO2), and is the fraction of the oxygen delivered to the microcirculation that is taken up into the tissues. The ratio can be multiplied by 100 to generate a p ercentage. The normal O2ER is 0.2 to 0.3 (20 to 30%), indicating that 20 to 30% of the oxyg en delivered to the capillaries is taken up into the tissues. Thus, only a small fraction of the available oxygen in capillary blood is used to support aerobic metabolism. Oxygen extraction is adjustable, and in conditions where O2 delivery is impaired, the O2ER can increase to 0.5 to 0.6. In trained athletes, the O2ER can be as high as 0.8 during maximal exercise (3). Adjustments in O2 extraction play an important role in maintaining oxygen uptake when oxygen delivery is var iable, as described in the next section. P.24 CONTROL OF OXYGEN UPTAKE The oxygen transport system normally operates to maintain a constant rate of oxy gen uptake (VO2) in conditions where O2 delivery (DO2) can vary widely (3). This is accomplished by compensatory adjustments in O2ER in response to changes in D O2. The O2ER describes the relationship between http://gateway.ut.ovid.com/gw1/ovidweb.cgi (6 of 15)03-May-05 17:27:15

Ovid: ICU Book O2 and DO2; that is, O2ER = VO2/DO2, which can be rearranged as shown in Equatio n 2.7. According to this relationship, when a decrease in DO2 is accompanied by a propo rtional increase in O2ER, the VO2 remains constant. However, when the O2ER is fi xed, a decrease in DO2 is accompanied by a proportional decrease in VO2. The adj ustability of the O2 extraction therefore defines the tendency for VO2 to change in response to variations in O2 delivery. The normal relationship between DO2 a nd VO2 is described in the next section. THE DO2VO2 CURVE The relationship between O2 delivery and O2 uptake is described by the curve in Figure 2.2, where O2 delivery is the independent variable. As the O2 delivery de creases below normal, the O2ER increases P.25 proportionally and the VO2 remains constant. When the O2ER reaches its maximum level (0.5 to 0.6), further decreas es in DO2 are accompanied by proportional decreases in VO2. In the linear portio n of the curve, the VO2 is supply-dependent and the production of adenosine trip hosphate (ATP) is limited by the supply of oxygen. This condition of oxygen-limi ted energy production is called dysoxia (4). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (7 of 15)03-May-05 17:27:15

Ovid: ICU Book Figure 2.2. Graph describing the normal relationship between O2 delivery (DO2) a nd O2 uptake (VO2). O2ER = oxygen extraction ratio. Critical O2 Delivery The DO2 at which the VO2 becomes supply-dependent is called the critical oxygen delivery, and is the point at which energy production in cells becomes oxygen-li mited (dysoxia). The critical DO2 in anesthetized subjects is in the vicinity of 300 mL/minute/m2, but in critically ill patients, the critical DO2 varies widel y, from 150 to 1000 mL/minute/m2 (3). Regardless of the source of this variabili ty, it means that the critical DO2 must be determined for each individual patien t in the ICU. Supply-Dependent VO2 In critically ill patients, the DO2VO2 relationship can be linear over a wide ran ge, and the supplydependent VO2 in these patients can be the result of three pos sible conditions (3,4,5 and 6). One condition is pathologic supply dependency, w here ATP production is limited by the supply of oxygen (dysoxia). This condition produces the supply dependency at very low levels of DO2 in Figure 2.2. Another condition is physiologic supply dependency, where VO2 is the independent variab le and DO2 changes in response to a primary change in the metabolic rate (6). Th is condition is responsible for the linear DO2VO2 relationship seen during exerci se, and may be responsible for supply dependency in critically ill patients. Mos t importantly, it indicates that a linear DO2VO2 relationship may not be the resu lt of a pathologic process. Finally, supply dependency may be an artifact produc ed when VO2 is calculated and not directly measured. This latter possibility is discussed in more detail in Chapter 13. The relationship between DO2 and VO2 is an important component of oxygen transport monitoring in the ICU, and can be use d to identify tissue ischemia (e.g., pathologic supply dependency) or to create a therapeutic strategy (e.g., increasing DO2 when the O2 extraction is maximal). The applications of oxygen transport monitoring in the care of critically ill p atients is described in Chapter 13. CARBON DIOXIDE TRANSPORT Carbon dioxide is the major end-product of oxidative metabolism (7) and, because it is capable of transforming into carbonic acid when hydrated, it can cause pr oblems if allowed to accumulate. The value of CO2 elimination is apparent in the operation of the ventilatory control system; that is, the ventilatory control s ystem is designed to regulate P.26 http://gateway.ut.ovid.com/gw1/ovidweb.cgi (8 of 15)03-May-05 17:27:15

Ovid: ICU Book carbon dioxide and to promote its elimination through the lungs. An increase in PCO2 of 5 mm Hg can result in a twofold increase in minute ventilation. To produ ce an equivalent increment in ventilation, the arterial PO2 must drop to 55 mm H g (8). Thus, the ventilatory control system keeps a close eye on CO2, but pays l ittle attention to oxygen (whereas clinicians keep a close eye on oxygen and pay little attention to CO2). TOTAL BODY CO2 Carbon dioxide is more soluble in water than oxygen, and thus moves more freely in the body fluids. However, the total body content of CO2 is reported as 130 L (9), which seems to be quite a trick, considering that the average adult has no more than 40 to 50 L of water to spare. The explanation for this is that the CO2 enters into a chemical reaction with water. This allows large volumes of CO2 to enter a solution because the reaction with water dissociates CO2 and maintains the gradient that drives the gas into solution. Opening a bottle of warm champag ne or a warm beer demonstrates how much CO2 can be present in a solution. WHOLE BLOOD CO2 CONTENT Unlike oxygen, the CO2 content of whole blood cannot be derived using simple equ ations. The reason for this will become apparent as the CO2 transport process is revealed. However, the fraction of CO2 dissolved in plasma can be defined using the solubility coefficients for CO2 shown in Table 2.1. At a normal body temper ature of 37 C, the concentration of dissolved CO2 is 0.686 mL/L/mm Hg. At a PCO2 of 40 mm Hg, the dissolved CO2 in arterial blood is (40 .68) 26 mL/L, as shown i n Table 2.2. When compared with the total CO2 content, it is apparent that only a small fraction of the CO2 is present in the dissolved form. TRANSPORT SCHEME The centerpiece of CO2 transport is the hydration reaction, and Figure 2.3 shows how the reaction participates in the transport process. The first step in the h ydration reaction is the formation of carbonic acid (H2CO3). This is normally a slow reaction, and takes about 40 seconds to complete (10). The reaction speeds up considerably in the presence of the enzyme carbonic anhydrase, and takes less than 10 milliseconds to complete (10). Carbonic anhydrase is confined to the re d cell, and is not present in plasma. Thus, CO2 is rapidly hydrated only in the red cell, so CO2 is drawn into the red cell. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (9 of 15)03-May-05 17:27:15

Ovid: ICU Book Figure 2.3. The chemical reactions involved in CO2 transport. Values represent a mounts in 1 L whole blood (venous). The double arrows indicate favored pathways. Hb = hemoglobin. P.27 Carbonic acid dissociates to generate hydrogen and bicarbonate ions. A larg e fraction of the bicarbonate generated in the red cell is pumped back into the plasma in exchange for chloride. The hydrogen ion generated in the red cell is b uffered by the hemoglobin. In this way, the CO2 that enters the red cell is dism antled, and the parts stored (hemoglobin) or discarded (bicarbonate) to create r oom for more CO2 to enter the red cell. These processes, along with the carbonic anhydrase facilitated hydration reaction, create a sink for large volumes of CO2 in the red cell. A small fraction of CO2 in the red cell reacts with free amino groups on hemoglobin. This produces carbamic acids, which dissociate into hydro gen ions and carbamino residues (HbNHCOO-. This reaction is not a major componen t of CO2 transport. Unit Conversions If the values in Figure 2.3 are added up, the total CO2 content is 23 mEq/L in w hole blood, with 17 mEq/L in plasma and 6 mEq/L in the red cell. Thus, most of t he CO2 appears to be in the plasma, but this is deceiving because much of the pl asma component is in the form of bicarbonate that has been expelled from the red blood cell. Because CO2 is a ready source of ions (hydrogen and bicarbonate), t he concentration of CO2 is often expressed in mEq/L. This is the case in Figure 2.3. The conversion is based on the following: 1 mole of CO2 has a volume of 22. 3 L (STPD), so 1 mM CO2 is approximately 22.3 P.28 http://gateway.ut.ovid.com/gw1/ovidweb.cgi (10 of 15)03-May-05 17:27:15

Ovid: ICU Book mL and 1 mM/L CO2 is approximately 22.3 mL/L or 2.23 mL/100 mL (vol%). Therefore , HEMOGLOBIN AS A BUFFER As mentioned earlier, the mass of hemoglobin in circulating blood is far greater than what is needed to transport oxygen, and moving this excess hemoglobin repr esents a considerable work load for the heart. This would make the circulatory s ystem an energy-inefficient system unless the excess hemoglobin is required for some other vital function. As shown in Figure 2.3, hemoglobin has an important r ole in the transport of carbon dioxide. Considering the large volume of CO2 in t he blood, the large size of the hemoglobin pool becomes more understandable. The function of hemoglobin in CO2 transport is to act as a buffer for the hydrogen ions that are produced by the hydration of CO2 in the red blood cell. The abilit y of hemoglobin to act as a buffer has been recognized since the 1930s, but this property of hemoglobin receives little attention. The buffer capacity of hemogl obin is shown in Table 2.5 (11). Note that the total buffering capacity of hemog lobin is six times as great as the buffering capacity of all the plasma proteins combined. A small part of this difference is due to the enhanced buffer capacit y of the hemoglobin molecule, but most of the difference is due to the enormous size of the circulating hemoglobin pool. TABLE 2.5. BUFFERING CAPACITY OF BLOOD PROTEINS The buffering actions of hemoglobin are caused by the histidine residues on the molecule. The imidazole group in histidine is responsible for the buffering acti ons, and is most effective at a pH of 7.0 (the dissociation constant of imidazol e has a pK = 7.0, and buffers are most effective within 1 pH unit on either side of the pK). This means that hemoglobin is an effective buffer in the usual pH r ange of blood. In fact, hemoglobin should be more effective as a buffer than bic arbonate because the pK http://gateway.ut.ovid.com/gw1/ovidweb.cgi (11 of 15)03-May-05 17:27:15

Ovid: ICU Book of carbonic acid is 6.1, which is outside the usual pH range of blood. Thus, hem oglobin is the focal point of CO2 transport because it can bind the acid equival ence of CO2. The binding of hydrogen ions by hemoglobin creates a sink that main tains CO2 flux into the red cell. Only because of its large mass can hemoglobin accomplish this. P.29 Haldane Effect Hemoglobin has a greater buffer capacity when it is in the desaturated form, and blood that is fully desaturated can bind an additional 60 mL/L CO2. The increas e in CO2 content that occurs when blood is desaturated is known as the Haldane e ffect. The graph in Figure 2.4 shows that the Haldane effect is responsible for a significant portion of the change in CO2 content between arterial and venous b lood. This reaffirms the important role played by hemoglobin in CO2. Figure 2.4. Graph showing the factors that contribute to the increase in CO2 con tent from arterial to venous blood. Sat = saturation, PCO2 = carbon dioxide pres sure. CO2 ELIMINATION (VCO2) Although CO2 is dismantled for transport in the peripheral venous blood, it is r econstituted when the blood reaches the lungs. The next step is elimination thro ugh the lungs, and this process is http://gateway.ut.ovid.com/gw1/ovidweb.cgi (12 of 15)03-May-05 17:27:15

Ovid: ICU Book described in Figure 2.5. The elimination of CO2 (VCO2) is a Fick relationship, l ike VO2, but with the arterial and venous components reversed. Because there are no simple derivative equations for CO2 content in blood, VCO2 is usually measur ed directly. When VCO2 is expressed as volume/time, the normal value is about 80 % of the VO2 (Table 2.4). The ratio VCO2/VO2 P.30 is thus 0.8 normally. This rat io is known as the respiratory quotient (RQ), and it varies according to the typ e of nutrient being metabolized. (See Chapter 46 for more information on the RQ. ) Figure 2.5. Schematic view of CO2 elimination (VCO2) via the lungs. The dotted l ine represents the alveolarcapillary interface. CvCO2 = venous carbon dioxide con centration, CaCO2 = arterial carbon dioxide concentration, Q = cardiac output. Acid Excretion When VCO2 is expressed in mEq/L, it describes the rate of volatile acid excretio n. As shown in Figure 2.5, this rate is normally 10 mEq/minute, or 14,400 mEq/24 hours. During exercise, the excretion of volatile acids via the lungs can incre ase to 40,000 mEq/24 hours. Considering that the kidneys excrete only 40 to 80 m Eq acid/24 hours, the major organ of acid excretion in the human body is the lun gs, not the kidneys. This method of describing CO2 elimination emphasizes the ac id burden of metabolism. This emphasis on the production side of metabolism is a n important addition to the current supply-dominant approach to metabolic suppor t. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (13 of 15)03-May-05 17:27:15

Ovid: ICU Book REFERENCES GENERAL WORKS Edwards JD, Shoemaker WC, Vincent J-L. Oxygen transport. Principles and practice . Philadelphia: WB Saunders, 1993. Zander R, Mertzlufft F, eds. The oxygen status of arterial blood. Basel: S. Karg er, 1991. P.31 OXYGEN TRANSPORT 1. Pauling L. General chemistry. 3rd ed. Mineola NY: Dover Publications, 1988:21 5. 2. Zander R. Calculation of oxygen concentration. In: Zander R, Mertzlufft F, ed s. The oxygen status of arterial blood. Basel: S. Karger, 1991;203209. 3. Leach RM, Treacher DF. The relationship between oxygen delivery and consumpti on. Dis Mon 1994;30:301368. 4. Connett RJ, Honig CR, Gayeski TEJ, Brooks GA. Defining hypoxia: a systems vie w of VO2, glycolysis, energetics, and intracellular PO2. J Appl Physiol 1990;68: 833842. 5. Russel JA, Phang PT. The oxygen delivery/consumption controversy. Approaches to management in the critically ill. Am Rev Respir Crit Care Med 1994;149:533553. 6. Weissman C, Kemper M. Stressing the critically ill patient: the cardiopulmona ry and metabolic responses to an acute increase in oxygen consumption. J Crit Ca re 1993;7:100109. CARBON DIOXIDE TRANSPORT 7. Nunn JF. Nunn's applied respiratory physiology. 4th ed. Stoneham, MA: Butterw orth, 1993:219246. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (14 of 15)03-May-05 17:27:15

Ovid: ICU Book 8. Lambertson CJ. Carbon dioxide and respiration in acidbase homeostasis. Anesthe siology 1960;21:642651. 9. Henneberg S, Soderberg D, Groth T, et al. Carbon dioxide production during me chanical ventilation. Crit Care Med 1987;15:813. 10. Brahm J. The red cell anion-transport system: kinetics and physiologic impli cations. In: Gunn RB, Parker C, eds. Cell physiology of blood. New York: Rockefe ller Press 1988;142150. 11. Comroe JH Jr. Physiology of respiration. 2nd ed. Chicago: Year Book, 1974:20 1210. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (15 of 15)03-May-05 17:27:15

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 3 THE THREAT OF OXIDANT INJURY All human things are subject to decay. --John Dryden The treatment of critically ill patients is dominated by the notion that promoting the supply of oxygen to the vital organs is a necessary and life-sustaining measure. Oxygen is provided in a liberal and unregulated fashion, while the tendency for oxygen to degrade a nd decompose organic (carbonbased) matter is either overlooked or underestimated . In contrast to the notion that oxygen protects cells from injury in the critic ally ill patient, the accumulated evidence over the past 15 years suggests that oxygen is responsible for the cell injury that accompanies critical illness. Oxy gen's ability to act as a lethal toxin has monumental implications for the way w e treat critically ill patients. THE OXIDATION REACTION An oxidation reaction is a chemical reaction between oxygen and another chemical species. Because oxygen removes electrons from other atoms and molecules, oxida tion is also described as the loss of electrons by an atom or molecule. The chem ical species that removes the electrons is called an oxidizing agent or oxidant. The companion process (i.e., the gain of electrons by an atom or molecule) is c alled a reduction reaction, and the chemical species that donates the electrons is called a reducing agent. Because oxidation of one atom or molecule must be ac companied by reduction of another atom or molecule, the overall reaction is ofte n called a redox reaction. When an organic molecule (a molecule with a carbon sk eleton) reacts with oxygen, electrons are removed from carbon atoms in the molec ule. This disrupts one or more covalent bonds, and as each bond ruptures, energy is released in the form of heat and light (and sometimes P.33 sound). The organ ic molecule then breaks into smaller fragments. When oxidation is complete, the parent molecule is broken down into the smallest molecules capable of independen t existence. Because organic matter is composed mainly of carbon and hydrogen, t he end-products of oxidation are simple combinations of oxygen with carbon and h ydrogen: carbon dioxide and water. OXYGEN METABOLISM http://gateway.ut.ovid.com/gw1/ovidweb.cgi (1 of 21)03-May-05 17:28:54

Ovid: ICU Book Oxygen is a weak oxidizing agent, but some of its metabolites are potent oxidant s capable of producing widespread and lethal cell injury (1). The mechanism wher eby oxygen metabolism can produce more powerful oxidants than the parent molecul e is related to the atomic structure of the oxygen molecule, which is described below. THE OXYGEN MOLECULE Oxygen in its natural state is a diatomic molecule, as shown by the familiar O2 symbol at the top of Figure 3.1. The orbital diagram to the right of the O2 symb ol shows how the outer electrons of the oxygen molecule are arranged. The circle s in the diagram represent orbitals. (An orbital is an energy field that can be occupied by electrons. It is distinct from an orbit, which is a path that repres ents a specific point in space and time.) The arrows in the diagram represent el ectrons that are spinning in the same or opposite directions (indicated by the d irection of the arrows). Note that one of oxygen's orbitals contains two electro ns with opposing spins, and the other two orbitals each contain a single electro n spinning in the same direction. The orbital with the paired electrons is obeyi ng one of the basic rules of the quantum atom: An electron orbital can be occupi ed by two electrons if they have opposing spins. Thus, the two outermost orbital s that contain single electrons are only half full, and their electrons are unpa ired. An atom or molecule that has one or more unpaired electrons in its outer o rbitals is called a free radical (2). (The term free indicates that the atom or molecule is capable of independent existenceit is free-living.) http://gateway.ut.ovid.com/gw1/ovidweb.cgi (2 of 21)03-May-05 17:28:54

Ovid: ICU Book Figure 3.1. The metabolism of molecular oxygen to water. Orbital diagrams on the right side of the figure show the electron configuration (arrows) in the outer orbitals (circles) of each reactant. The highest orbitals in each diagram are fu rthest from the nucleus. Single electron reductions indicated by e- . Free radicals tend to be highly reactive chemical species by virtue of their unp aired electrons. However, not all free radicals are highly reactive. This is the case with oxygen, which is not a highly reactive molecule despite having two un paired electrons. The reason for oxygen's sluggish reactivity is the directional spin of its two unpaired electrons. No two electrons can occupy the same orbita l if they have the same directional spin. Thus, an electron pair cannot be added to oxygen because one orbital would have two electrons with the same directiona l spin, which is a quantum impossibility. This spin restriction limits oxygen to single electron additions, and this not only increases the number of reactions needed to reduce molecular P.34 oxygen to water, but it also produces more highl y reactive intermediates. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (3 of 21)03-May-05 17:28:54

Ovid: ICU Book THE METABOLIC PATHWAY Oxygen is metabolized at the very end of the electron transport chain, where the electrons and protons that have completed the transport P.35 process are left t o accumulate. The complete reduction of molecular oxygen to water requires the a ddition of four electrons and four protons, as shown in the reaction sequence in Figure 3.1. Each metabolite in this sequence is accompanied by an orbital diagr am to demonstrate the changes occurring at each point in the pathway. Superoxide Radical The first reaction adds one electron to oxygen, and produces the superoxide radi cal. Note the superscript dot on the superoxide symbol. This signifies an unpaired el ectron, and is the conventional symbol for a free radical. The superoxide radica l has one unpaired electron, and thus is less of a free radical than oxygen. Sup eroxide is neither a highly reactive radical nor a potent oxidant (3). Neverthel ess, it has been implicated in conditions associated with widespread tissue dama ge, such as the reperfusion injury that follows a period of ischemia (2). The to xicity of superoxide may be caused by the large daily production, which is estim ated at 1 billion molecules per cell, or 1.75 kg (4 lb) for a 70-kg adult (4). Hydrogen Peroxide The addition of one electron to superoxide creates hydrogen peroxide, a strong o xidizing agent (and the source of acid rain in the atmosphere) (5). Hydrogen peroxide is very mobile, and crosses cell membranes easily. It is a pow erful cytotoxin and is well known for its ability to damage endothelial cells. I t is not a free radical, but it may have to generate a free radical (a hydroxyl radical) to express its toxicity. Hydrogen peroxide is loosely held together by a weak oxygenoxygen bond (this bond is represented by the lower orbital in the or bital diagram for hydrogen peroxide). This bond ruptures easily, producing two h ydroxyl radicals, each with one unpaired electron. An electron is donated to one of the hydroxyl radicals, creating one hydroxyl ion (OH-) and one hydroxyl radi cal (OH). The electron is donated by iron in its reduced form, Fe(II), which serv es as a catalyst for the reaction. Iron is involved in many free radical reactio ns, and is considered a powerful pro-oxidant. The role of transition metals in f ree radical reactions is discussed again later in the chapter. Hydroxyl Radical http://gateway.ut.ovid.com/gw1/ovidweb.cgi (4 of 21)03-May-05 17:28:54

Ovid: ICU Book The iron-catalyzed dissociation of hydrogen peroxide proceeds as follows: P.36 (Note that Roman numerals are used instead of plus signs to designate the o xidation state of iron, as recommended by the International Union of Chemistry.) The hydroxyl radical is the ace of free radicals. It is one of the most reactiv e molecules in biochemistry and often reacts with another chemical species withi n five molecular diameters from its point of origin (2). This high degree of rea ctivity limits the mobility of the hydroxyl radical, and this may serve as a pro tective device to limit the toxicity of the hydroxyl radical. However, the hydro xyl radical is always dangerous because it can oxidize any molecule in the human body. Hypochlorous Acid The metabolism of oxygen in neutrophils has an additional pathway (not shown in Figure 3.1) that uses a myeloperoxidase enzyme to chlorinate hydrogen peroxide, creating hypochlorous acid (hypochlorite). When neutrophils are activated, the conversion of oxygen to superoxide increases twentyfold. This is called the respiratory burst, which is an unfortunate term because the increased O2 consumption has nothing to do with energy metabolism. W hen the increased metabolic traffic reaches hydrogen peroxide, about 40% is dive rted to hypochlorite production and the remainder forms hydroxyl radicals (6). H ypochlorite is the active ingredient in household bleach. It is a powerful germi cidal agent and requires only milliseconds to produce lethal damage in bacteria (7). Water The final reaction in oxygen metabolism adds an electron to the hydroxyl radical and produces two molecules of water. Therefore, the metabolism of one molecule of oxygen requires four chemical react ions, each involving the addition of a single electron. This process, then, requ ires four reducing equivalents (electrons and protons). Under normal conditions, about 98% of the oxygen metabolism is completed, and less than 2% of the metabo lites escape into the cytoplasm (3). This is a tribute to cytochrome oxidase, wh ich carries on the reactions in a deep recess that effectively blocks any radica l escape. This degree of suppression is necessary because of the ability of free radicals to start chain reactions (see next section). Proposed Scheme http://gateway.ut.ovid.com/gw1/ovidweb.cgi (5 of 21)03-May-05 17:28:54

Ovid: ICU Book The superoxide radical is mobile but not toxic, whereas the hydroxyl radical is toxic but not mobile. Combining the advantages of each P.37 oxidant yields a sch eme that has the mobile oxidant serving as a transport vehicle that can reach di stant places. Once at the desired location, this metabolite could then generate hydroxyl radicals to produce local damage (3). This scheme is intuitively satisf ying, regardless of its validity. FREE RADICAL REACTIONS The damaging effects of oxidation are largely the result of free radical reactio ns. This section describes the two basic types of free radical reactions: those involving free radicals and nonradicals and those involving two free radicals. RADICAL AND NONRADICAL When a free radical reacts with a nonradical, the nonradical loses an electron a nd is transformed into a free radical. Therefore, the union of a radical and a n onradical begets another radical (thus illustrating the survival value of the fr ee radical). Because free radicals are often highly reactive in nature. This typ e of radical-regenerating reaction tends to become repetitive, creating a series of selfsustaining reactions known collectively as a chain reaction (3). The ten dency to produce chain reactions is one of the most characteristic features of f ree radical reactions. A fire is one example of a chain reaction involving free radicals, and fires illustrate a very important feature of chain reactions: the tendency to produce widespread damage. A chain reaction that is capable of produ cing widespread organ damage is described below. Lipid Peroxidation The rancidity that develops in decaying food is the result of oxidative changes in polyunsaturated fatty acids (8). This same process, called lipid peroxidation , is also responsible for the oxidative damage of membrane lipids. The lipophili c interior of cell membranes is rich in polyunsaturated fatty acids (e.g., arach idonic acid) and the characteristic low melting point of these fatty acids may b e responsible for the fluidity of cell membranes. Oxidation increases the meltin g point of membrane fatty acids and reduces membrane fluidity. The membranes eve ntually lose their selective permeability and become leaky, predisposing cells t o osmotic disruption (8). The peroxidation of membrane lipids proceeds as shown in Figure 3.2. The reaction sequence is initiated by a strong oxidant such as th e hydroxyl radical, which removes an entire hydrogen atom (proton and electron) from one of the carbon atoms in a polyunsaturated fatty acid. This creates a car bon-centered radical (C), which is then transformed into an oxygen-centered perox y radical (COO) that can remove a hydrogen atom from an adjacent fatty acid and i nitiate a new series of reactions. The final propagation reaction creates a self -sustaining chain P.38 http://gateway.ut.ovid.com/gw1/ovidweb.cgi (6 of 21)03-May-05 17:28:54

Ovid: ICU Book reaction that will continue until the substrate (i.e., fatty acid) is exhausted, or until something interferes with the propagation reaction. (The latter mechan ism is the basis for the antioxidant action of vitamin E, which is described lat er.) Figure 3.2. The reaction sequence for the peroxidation of polyunsaturated fatty acids (PUFAs) in cell membranes. Implications Free radical reactions have been implicated in the diseases (9), but it is not clear whether oxidant uence of disease (9,10). However, a chain reaction e., independent of the initiating process), and if comes an independent pathologic process (a primary pathogenesis of more than 100 injury is a cause or a conseq is an independent process (i. it causes tissue injury it be illness).

RADICAL AND RADICAL Two radicals can react by sharing electrons to form a covalent bond. This elimin ates the free radicals but does not necessarily eliminate the P.39 risk of toxic ity. In the example below, the product of a radicalradical reaction is much more destructive than both radicals combined. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (7 of 21)03-May-05 17:28:54

Ovid: ICU Book Nitric Oxide Transformation Nitric oxide has been placed in a category of its own as a free radical because of its beneficial actions as a vasodilator, neurotransmitter, and bactericidal a gent (11). The regard for nitric oxide has been so favorable that it was named Mo lecule of the Year by Science Magazine in 1992. However, despite its favorable pr ofile, nitric oxide can become a toxin in the presence of superoxide. The reacti on of superoxide with nitric oxide generates a powerful oxidant called peroxynit rite, which is 2000 times more potent than hydrogen peroxide as an oxidizing age nt (12). Peroxynitrite can either cause direct tissue damage or can decompose an d produce hydroxyl radicals and nitrogen dioxide. The transformation of nitric oxide into a source of oxidant injury demonstrates how free radicals can promote oxidant damage indirectly. ANTIOXIDANT PROTECTION Evidence for endogenous antioxidant protection is provided by the simple observa tion that accelerated decay begins at the moment of death. This section presents the substances that are believed to play a major role in this protection. An an tioxidant is defined as any chemical species that can reduce or delay the oxidat ion of an oxidizable substrate (2). The nonenzyme antioxidants are included in T able 3.1. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (8 of 21)03-May-05 17:28:54

Ovid: ICU Book http://gateway.ut.ovid.com/gw1/ovidweb.cgi (9 of 21)03-May-05 17:28:54

Ovid: ICU Book TABLE 3.1. ENDOGENOUS AND EXOGENOUS ANTIOXIDANTS ENZYME ANTIOXIDANTS There are three enzymes that function as antioxidants, shown in Figure 3.3. Note that the reaction sequence in this figure is the same as in Figure 3.1. Figure 3.3. The actions of three antioxidant enzymes and a free radical scavenge r. The reaction sequence depicts oxygen metabolism, as shown in Figure 3.1. Cofa ctors for superoxide dismutase are iron (Fe), zinc (Zn), and copper (Cu), but ar e never present as a triad on the same enzyme. Cofactor for glutathione peroxida se is selenium (Se). GSH = reduced glutathione, GSSG = oxidized glutathione as a dipeptide. Superoxide Dismutase The discovery of superoxide dismutase (SOD) enzyme in 1969 was the first indicat ion of free radical activity in humans, and this began the frenzy of interest in free radicals. The role of SOD as an antioxidant is not clear. In fact, SOD pro motes the formation of hydrogen peroxide, which is an oxidant. How can an enzyme that promotes the formation of an oxidant be defined as an antioxidant? In fact , if SOD increases http://gateway.ut.ovid.com/gw1/ovidweb.cgi (10 of 21)03-May-05 17:28:54

Ovid: ICU Book P.40 P.41 the metabolic traffic flowing through hydrogen peroxide, and the catal ase and peroxidase reactions are unable to increase their activity proportionall y, the hydrogen peroxide levels may rise, and in this situation SOD functions as a pro-oxidant (13). Thus, SOD is not an antioxidant at least some of the time. Catalase Catalase is an iron-containing heme protein that reduces hydrogen peroxide to wa ter. It is present in most cells, but is lowest in cardiac cells and neurones. I nhibition of the catalase enzyme does not enhance the toxicity of hydrogen perox ide for endothelial cells (14), so the role of this enzyme as an antioxidant is unclear. Glutathione Peroxidase The peroxidase enzyme reduces hydrogen peroxide to water by removing electrons f rom glutathione in its reduced form and then donating the electrons to hydrogen peroxide. Glutathione is returned to P.42 its reduced state by a reductase enzym e that transfers the reducing equivalents from NADPH. The total reaction can be written as follows: where GSSG and GSH are oxidized and reduced glutathione, respectively. Selenium The activity of the glutathione peroxidase enzyme in humans depends on the trace element selenium. Selenium is an essential nutrient with a recommended dietary allowance of 70 g daily for men and 55 g daily for women (15). Despite this recomm endation, selenium is not included in most total parenteral nutrition regimens. Because the absence of dietary selenium produces measurable differences in gluta thione peroxidase activity after just 1 week (16), the routine administration of selenium seems justified. However, selenium, has no clear-cut deficiency syndro me in humans, so there is little impetus to provide selenium on a routine basis. Selenium status can be monitored with whole blood selenium levels. The normal r ange is 0.5 to 2.5 mg/L. Selenium can be provided intravenously as sodium seleni te (17). The highest daily dose that is considered safe is 200 g, given in divide d doses (50 g intravenously every 6 hours). NONENZYME ANTIOXIDANTS http://gateway.ut.ovid.com/gw1/ovidweb.cgi (11 of 21)03-May-05 17:28:54

Ovid: ICU Book Glutathione One of the major antioxidants in the human body is a sulfur-containing tripeptid e glutathione (glycine cysteineglutamine), which is present in molar concentration s (0.5 to 10 mM/L) in most cells (18,19). Glutathione is a reducing agent by vir tue of a sulfhydryl group in the cysteine residue of the molecule. It is normall y in the reduced state (GSH), and the ratio of reduced to oxidized forms is 10:1 . The major antioxidant action of glutathione is to reduce hydrogen peroxide dir ectly to water, which diverts hydrogen peroxide from producing hydroxyl radicals . Glutathione is found in all organs, but is particularly prevalent in the lung, liver, endothelium, and intestinal mucosa. It is primarily an intracellular ant ioxidant, and plasma levels of glutathione are three orders of magnitude lower t han intracellular levels. Glutathione does not cross cell membranes directly, bu t is broken down first into its constituent amino acids and then reconstituted o n the other side of the membrane. It is synthesized in every cell of the body, a nd largely remains sequestered within cells. Exogenous glutathione P.43 has litt le effect on intracellular levels (20), which limits the therapeutic value of th is agent. N-Acetylcysteine N-Acetylcysteine, a popular mucolytic agent (Mucomyst), is a sulfhydryl-containi ng glutathione analog capable of passing readily across cell membranes. N-Acetyl cysteine is effective as a glutathione analog in acetaminophen toxicity, which i s the result of an overwhelmed glutathione detoxification pathway (see Chapter 5 3). Therefore, N-acetylcysteine has a proven track record as an exogenous glutat hione analog. N-Acetylcysteine may prove to be a valuable antioxidant for therap eutic use. It protects the myocardium from ischemic injury, and has been success ful in reducing the incidence of reperfusion injury during cardiac catheterizati on (21). It has also been used with some success in treating critically ill pati ents with acute respiratory distress syndrome and inflammatory shock syndromes ( 22,23). Vitamin E Vitamin E (-tocopherol) is lipid-soluble vit min th t functions prim rily s n ntioxid nt th t nt gonizes the peroxid tive injury of membr ne lipids. It is the only ntioxid nt c p ble of h lting the prop g tion of lipid peroxid tion. T he mech nism for this ction is shown in Figure 3.4. Vit min E inh bits the lipo philic interior of cell membr nes, where the polyuns tur ted f tty cids re ls o loc ted. When prop g ting w ve of lipid peroxid tion re ches vit min E, it i s oxidized to free r dic l, thereby sp ring ny dj cent polyuns tur ted f tty cids from oxid tion. The vit min E r dic l is poorly re ctive, nd this h lts the prop g tion of the peroxid tion re ctions. This ction h s e rned vit min E the title of ch in-bre king ntioxid nt. P.44 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 21)03-M y-05 17:28:54

Ovid: ICU Book

Vit min E deficiency m y be common in critic lly ill p tients (24). The norm l v it min E level in pl sm is 1 mg/dL, nd level below 0.5 mg/dL is evidence of deficiency (25). Considering the import nt role of vit min E s n ntioxid nt, it seems wise to check the vit min E st tus in p tients who re t risk for oxid nt injury (see T ble 3.2). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 21)03-M y-05 17:28:54

Figure 3.4. The ch in-bre king tion in cell membr nes.

ction of vit min E th t termin tes lipid peroxid

The vit min E r dic l is tr nsformed b ck to vit min E, the electron donor in this re ction.

nd vit min C c n

ct s

Ovid: ICU Book TABLE 3.2. CLINICAL CONDITIONS THAT ARE ACCOMPANIED BY OXIDANT STRESS* Vit min C Vit min C ( scorbic cid) is reducing gent th t c n don te electrons to free r dic ls nd fill their electron orbit ls. It is w ter-soluble P.45 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 21)03-M y-05 17:28:54

Ovid: ICU Book ntioxid nt, nd oper tes prim rily in the extr cellul r sp ce. Vit min C is fou nd in bund nce in the lung, where it m y pl y protective role in in ctiv ting pollut nts th t enter the irw ys. The problem with vit min C is its tendency t o promote (r ther th n ret rd) the form tion of oxid nts in the presence of iron nd copper (26,27 nd 28). Vit min C reduces iron to the Fe(II) st te, nd this norm lly ids in the bsorption of iron from the intestin l tr ct. However, Fe( II) c n promote the production of hydroxyl r dic ls, s described e rlier. Thus, vit min C c n function s pro-oxid nt by m int ining iron in its reduced or F e(II) st te. The re ctions involved re s follows: Sever l conditions th t re common in ICU p tients c n promote n incre se in fr ee iron. Among these re infl mm tion, blood tr nsfusions, nd reductions in bin ding proteins. The prev lence of these conditions r ises serious concerns bout the use of vit min C s n exogenous ntioxid nt in the ICU p tient popul tion. Pl sm Antioxid nts The pl sm components with ntioxid nt ctivity re listed t the very bottom of T ble 3.2. Most of the ntioxid nt ctivity in pl sm c n be tr ced to two prot eins th t m ke up only 4% of tot l pl sm protein pool (27): cerulopl smin (the copper tr nsport or stor ge protein) nd tr nsferrin. Tr nsferrin binds iron in the Fe(III) st te, nd cerulopl smin oxidizes iron from the Fe(II) to Fe(III) st te. Therefore, cerulopl smin helps tr nsferrin to bind iron, nd both proteins then ct to limit free iron in the pl sm . For this re son, iron sequestr tion h s been proposed s the m jor ntioxid nt ctivity in pl sm (24). This is consi stent with the ctions of Fe(II) to promote free r dic l production, s shown in Figure 3.1. OXIDANT STRESS The risk nd severity of oxid tion-induced tissue injury re determined by the b l nce between oxid nt nd ntioxid nt ctivities. When oxid nt ctivity exceeds the neutr lizing c p city of the ntioxid nts, the excess or unopposed oxid nt ctivity c n promote tissue injury. This condition of unopposed biologic l oxid tion is known s oxid nt stress (29). P.46 PREDISPOSING CONDITIONS Any imb l nce in the ctivities of oxid nts nd ntioxid nts c n result in unopp osed oxid tion. The box plots in Figure 3.5 show the effects of two conditions t h t promote oxid nt ntioxid nt imb l nce on the level of oxid nt stress in hum ns . The index of oxid nt stress in this study is the ctivity of lipid hydroperoxi des in urine, me sured s spont neous chemiluminescence nd reported in counts p er minute (CPM). He lthy, nonsmoking dults (the control group) show the lowest level of oxid nt ctivity. The effects of n incre se in oxid nt burden is shown in group of he vy smokers (one puff of cig rette cont ins roughly one billi on free r dic ls). The effects of deficiency in ntioxid nt http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 21)03-M y-05 17:28:54

Ovid: ICU Book protection re shown for group of p tients with hum n immunodeficiency virus ( HIV) infection (glut thione deficiency is common in HIV infections) (30). E ch o f the predisposing conditions is ccomp nied by signific nt incre se in oxid n t ctivity in comp rison to the ctivity in the he lthy, control subjects. Note lso th t the HIV p tients h ve ongoing oxid nt stress when they re symptomfree . P.47 This supports the notion th t oxid nt stress is c use nd not consequ ence of p thologic org n injury.

Figure 3.5. Box di gr ms showing the influence of predisposing conditions on oxi d nt stress. The vertic l xis shows the urin ry ctivity of lipid hydroperoxide s, me sured s spont neous chemiluminescence nd recorded in counts per minute ( CPM). The three groups represent gem tched dults, nd N indic tes the number o f subjects in e ch group. The control group is represented on the left. HIV = hu m n immunodeficiency virus. (From Millili J, M rino PL, Nusb um M. The p ttern o f spont neous chemiluminescence in hum ns nd its use in me suring oxid nt stres s. Proceedings of the First Intern tion l Conference on Clinic l Chemiluminescen ce. Berlin: Humboldt University Press, 1994.) CLINICAL DISEASE http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 21)03-M y-05 17:28:54

Ovid: ICU Book As mentioned e rlier, oxid nts h ve been implic ted in the p thogenesis of more th n 100 clinic l dise ses (9); the ones most likely to be encountered in the IC U re listed in T ble 3.2. Unopposed biologic l oxid tion h s been documented in e ch of these clinic l conditions (9,10,30,31,32,33,34 nd 35). This does not e st blish c us l role for oxid tion (this will require evidence th t ntioxid n t ther py c n improve outcome), but the tendency for oxid tion to c use independ ent nd progressive tissue d m ge (e.g., in ch in re ctions) is re son enough to suspect th t oxid nt-induced injury pl ys role in these illnesses. Infl mm tion Most of the clinic l conditions in T ble 3.2 re ccomp nied by infl mm tion, n d the conditions with multiorg n involvement re often ssoci ted with progres sive, systemic infl mm tory response. As result, infl mm tion h s been propose d s princip l offender in p thologic forms of oxid nt injury. The rele se of free r dic ls from ctiv ted neutrophils nd m croph ges cre tes n oxid nt-inte nse environment, nd the bility of host cells to withst nd this oxid tive ss u lt m y be the import nt f ctor in determining the clinic l course of infl mm tor y conditions. In the desir ble world, leukocytederived oxid nts would nnihil te ll inv ding microbes, but would not ffect the host cells. In the undesir ble world, the infl mm tory oxid nts would destroy both the inv der nd the host. Th is proposed scheme is intuitively ppe ling, nd emph sizes the v lue of providi ng ntioxid nts routinely in infl mm tory illnesses. Antioxid nt Ther py The v lue of m int ining ntioxid nt protection is not deb t ble issue bec use loss of ntioxid nt defenses is known c use of tissue destruction; the best e x mple of this is the cceler ted dec y th t occurs fter de th. Therefore, nti oxid nt supplements should be considered s routine me sure in p tients who sp end more th n few d ys in the ICU. In p tients with ny of the clinic l condit ions in T ble 3.2, it would be wise to monitor some of the endogenous ntioxid n ts (vit min E, vit min C, nd selenium). A reduced ntioxid nt level in blood (o r ny body fluid) m y not indic te deficiency st te (it m y indic te th t the ntioxid nt is being used), but would cert inly be n indic tion to provide supp lements. The re l v lue of ntioxid nt ther py will be determined by clinic l st udies, some of which re currently in progress. P.48

METABOLIC SUPPORT The tendency for erobic met bolism to gener te toxins h s signific nt implic ti ons for the ppro ch to met bolic support in critic lly ill p tients. When met b olic lly-gener ted oxid nts overwhelm the body's ntioxid nt defenses, the commo n pr ctice of supporting met bolism by promoting the v il bility of oxygen nd nutrients serves only to gener te more toxic met bolites. The proper m neuver he re is to support the ntioxid nt defenses. This ppro ch dds nother dimension to the concept of met bolic support by considering the output side of met bolism . Remember th t http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (17 of 21)03-M y-05 17:28:54

Ovid: ICU Book met bolism is n engine (i.e., n energy converter) nd like ll engines, it h s n exh ust th t cont ins noxious byproducts of combustion. The exh ust from n utomobile engine dds pollut nts to the tmosphere; likewise, the exh ust from met bolic engine dds pollut nts to the biosphere. REFERENCES GENERAL WORKS D vies KJA, Ursini F, eds. The oxygen p r dox. P dov , It ly: CLEUP University P ress, 1995.

H lliwell B, Gutteridge JM. Free r dic ls in biology nd medicine. 2nd ed. Oxfor d: Cl rendon Press, 1989.

OXIDANTS 1. Ch nce B, Sies H, Boveris A. Hydroperoxide met bolism in m mm li n org ns. Ph ysiol Rev 1979;59:527605. 2. H lliwell B, Gutteridge JM. Free r dic ls in biology nd medicine. 2nd ed. Cl rendon: Oxford University Press, 1989:280.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (18 of 21)03-M y-05 17:28:54

3. Liochev SI, Fridovich I. The role of O2 in the production of HO in vitro n vivo. Free R dic Biol Med 1994;16:2933.

Y gi K, ed. Active oxygens, lipid peroxides, nd RC Press, 1993.

ntioxid nts. Boc

R ton, FL: C

Sies H, ed. Oxid tive stress II: oxid nts nd ress, 1991.

ntioxid nts. New York: Ac demic P

Moslen MT, Smith CV. Free r dic l mech nisms of tissue injury. Boc RC Press, 1992.

Grish m MB. Re ctive met bolites of oxygen nd nitrogen in biology Austin, TX: RG L ndes, 1992.

nd medicine.

R ton, FL: C

nd i

Ovid: ICU Book

5. Thompson AM. The oxidizing c p city of the e rth's tmosphere: prob ble p st nd future ch nges. Science 1992;256:11571165. 6. Anderson BO, Brown JM, H rken A. Mech nisms of neutrophil-medi ted tissue inj ury. J Surg Res 1991;51:170179.

7. Bernovsky C. Nucleotide chlor mines nd neutrophil-medi ted cytotoxicity. FAS EB J 1991;5:295300. P.49 8. H lliwell B, Gutteridge JM. Free r dic ls in biology nd medicine. 2nd ed. Cl rendon: Oxford University Press, 1989:188204.

12. Freem n B. Free r dic l chemistry of nitric oxide. Looking t the d rk side. Chest 1994;105 (Suppl):7984. ANTIOXIDANTS 13. Michiels C, R es M, Touss nt O, Rem cle J. Import nce of Se-glut thione, per oxid se, c t l se, nd CU/ZN-SOD for cell surviv l g inst oxid tive stress. Fre e R dic Biol Med 1994;17:235248. 14. Suttorp N, Toepfer W, Rok L. Antioxid nt defense mech nisms of endotheli l cells: glut thione redox cycle versus c t l se. Am J Physiol 1986;251:671680. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (19 of 21)03-M y-05 17:28:54

11. Angg rd E. Nitric oxide: medi tor, murderer, 1991206.

nd medicine. L ncet 1994;343:1

10. H lliwell B. Free r dic l, ntioxid nts, , or consequence. L ncet 1994;344:721724.

nd hum n dise se: curiosity, c use

9. Gutteridge JMC. Free r dic ls in dise se processes: d consequence. Free R dic Res Commun 1993;19:141158.

compil tion of c use n

4. Frei B. Re ctive oxygen species nd n. Am J Med 1994;97(Suppl 3A):523.

ntioxid nt vit mins: mech nisms of ctio

Ovid: ICU Book

16. S ndo K, Hoki M, Nezu R, et l. Pl telet glut thione peroxid se ctivity in long-term tot l p renter l nutrition with nd without selenium supplement tion. J P rent Ent Nutr 1992;16:54 58.

18. Meister A. On the ntioxid nt effects of scorbic cid nd glut thione. Bioc hem Ph rm col 1992;44:19051915. 19. C ntin AM, Begin R. Glut thione nd infl mm tory disorders of the lung. Lung 1991;169:123138. 20. Robinson M, Ahn MS, Rounds JD, et l. P renter l glut thione monoester enh n ces tissue ntioxid nt stores. J P rent Ent Nutr 1992;16:413418. 21. Ferr ri R, Ceconi C, Curello S, et l. Oxygen free r dic ls nd myoc rdi l d m ge: protective role of thiol cont ining gents. Am J Med 1991;91(Suppl 3C):951 12. 22. Henderson A, H yes P. Acetylcysteine s cytoprotective ntioxid nt in p ti ents with severe sepsis: potenti l new use for n old drug. Ann Ph rm cother 199 4;28:10861088. 23. Suter PM, Domenighetti G, Sch ller MD, et l. N-Acetylcysteine enh nces reco very from cute lung injury in m n: r ndomized, double-blind pl cebo-controlle d clinic l study. Chest 1994;105:190194. 24. Pincem il J, Bertr nd Y, H nique G, et l. Ev lu tion of vit min E deficienc y in p tients with dult respir tory distress syndrome. Ann N Y Ac d Sci 1989;57 0:498500. 25. Meyd ni M. Vit min E. L ncet 1995;345:170176. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (20 of 21)03-M y-05 17:28:54

17. World He lth Org niz tion. Selenium. Environment l He lth Criteri , Switzerl nd, 1987.

58. Genev

15. N tion l Rese rch Council. Subcommittee on the tenth edition of the RDAs. W shington, DC: N tion l Ac demic Press, 1989:220.

Ovid: ICU Book

27. H lliwell B, Gutteridge JMC. Role of free r dic ls nd c t lytic met l ions in hum n dise se. Methods Enzymol 1990;186:185. 28. Herbert V, Sh w S, J y tilleke E, Stopler-K sd n T. Most free-r dic l injury is iron-rel ted: it is promoted by iron, hemin, h loferritin nd vit min C, nd inhibited by desferriox mine nd poferritin. Stem Cells 1994;12:289303. P.50 OXIDANT STRESS 29. Smith CV. Correl tions nd pp rent contr dictions in ssessment of oxid nt stress in vivo. Free R dic Biol Med 1991;10:217224. 30. St l FJ, El SW, Roederer M, et l. Glut thione deficiency nd hum n immuno deficiency virus infection. L ncet 1992;339:909912. 31. Weitz ZW, Birnb um AI, Sobotk PA, et l. High bre th pent ne concentr tions during cute myoc rdi l inf rction. L ncet 1991;337:933936. 32. Cross CE, v n der Vilet A, O'Neill CA, Eiserich JP. Re ctive oxygen species nd the lung. L ncet 1994;344:930935.

33. Gr ce PA. Ischemi -reperfusion injury. Br J Surg 1994;81:637647. 34. Deitch EA. Multiple org n f ilure. Ann Surg 1992;216:117134. 35. N t nson C, Hoffm n WD, Suffredini AF, et l. Selected tre tment str tegies for septic shock b sed on proposed mech nisms of p thogenesis. Ann Intern Med 19 94;120:771783. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (21 of 21)03-M y-05 17:28:54

26. Herbert V, Sh w S, J y tilleke E. Vit min C supplements re h rmful to leth l for over 10% of Americ ns with high iron stores. FASEB J 1994;8:A678.

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 4 VASCULAR ACCESS He who works with his h nds is l borer. He who works with his he d nd his h n ds is cr ftsm n. --St. Fr ncis of Assisi The c re of critic lly ill p tients r equires one or more pipelines to the v scul r system, for both monitoring nd in terventions. This ch pter presents some guidelines for the insertion of v scul r c theters, including brief description of the common percut neous ccess rout es (1,2 nd 3). The emph sis here is on the cr ft of est blishing v scul r cces s. The l bor of v scul r c nnul tion is skill le rned t the bedside. PREPARING FOR VASCULAR CANNULATION HANDWASHING H ndw shing is m nd tory ( nd often overlooked) before the insertion of v scul r devices. Scrubbing with ntimicrobi l cle nsing solutions does not reduce the i ncidence of c theter-rel ted sepsis (4), so simple so p- nd-w ter scrub is suf ficient. UNIVERSAL PRECAUTIONS In 1985, the Centers for Dise se Control introduced str tegy for blood nd bod y fluid prec utions known s univers l prec utions (5). This str tegy is b sed o n the ssumption th t ll p tients re potenti l sources of hum n immunodeficien cy virus (HIV) nd other blood-borne p thogens (e.g., hep titis viruses) until p roven otherwise. The following recommend tions pply to the insertion of v scul r c theters. P.54 Use protective gloves for ll v scul r c nnul tions. Use steri le gloves for ll c nnul tions except those involving the introduction of shor t c theter into peripher l vein. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 26)03-M y-05 17:31:06

Ovid: ICU Book C ps, gowns, m sks, nd protective eyewe r re not necess ry unless spl shes of blood re nticip ted (e.g., in tr um victim). These me sures do not reduce t he incidence of c theter-rel ted sepsis (6). Avoid needlestick injuries. Do not rec p needles or m nu lly remove needles from syringes. Pl ce ll sh rp instrume nts in puncture-resist nt cont iners immedi tely fter use. If needlestick inj ury is sust ined during the procedure, follow the recommend tions in T ble 4.1. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 26)03-M y-05 17:31:06

Ovid: ICU Book TABLE 4.1. RECOMMENDATIONS FOR NEEDLESTICK INJURIES Needlestick injuries re reported in up to 80% of medic l students nd interns ( 9). Therefore, in p tients who re known risks for tr nsmitting HIV or vir l hep titis, v scul r c nnul tion should be performed only by n experienced senior-l evel resident or fellow.

LATEX ALLERGY The incre sed use of rubber gloves (m de of l tex or vinyl) s protection g ins t HIV infections h s resulted in n incre sed recognition of llergic re ctions to l tex (10). These re ctions c n be m nifest s cont ct derm titis (urtic ri l lesions of the h nds nd f ce), or s conjunctivitis, rhinitis, or sthm . The l tter three m nifest tions re re ctions to irborne l tex p rticles, nd t hey do not require direct physic l cont ct with gloves. They often ppe r when t he ffected individu l enters n re where l tex gloves re being used. Therefo re, l tex llergy should be suspected in ny ICU te m member who develops top ic symptoms when in the ICU. When this occurs, switch to vinyl gloves will eli min te the problem. L tex llergy c n be m nifest s n phyl xis (10), so the tr nsition to vinyl gloves for suspected l tex llergy should not be del yed. CLEANSING THE SKIN Agents th t reduce skin microflor re c lled ntiseptics, where s gents th t r educe the microflor on in nim te objects re c lled disinfect nts. Common ntis eptic gents re listed in T ble 4.2 (11,12). The most widely used ntiseptic g ents re lcohol nd iodine, both of which h ve bro d spectrum of ntimicrobi l ctivity. Alcohol (commonly used s 70% solution) m y not work well on dirty skin (th t is, it does not h ve detergent ction), so it is often used in com bin tion with nother ntiseptic gent. The most popul r ntiseptic solution cur rently in use is povidoneiodine prep r tion (e.g., Bet dine), lso known s n iodophor, w ter-soluble complex of iodine nd c rrier molecule. The iodine i s rele sed slowly from the c rrier molecule, nd this reduces the irrit ting eff ects of iodine on the skin. This prep r tion P.55 should be left in cont ct with the skin for t le st 2 minutes to llow sufficient time for iodine to be rele sed from the c rrier molecule. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 26)03-M y-05 17:31:06

Ovid: ICU Book TABLE 4.2. COMMON ANTISEPTIC AGENTS HAIR REMOVAL Sh ving is not recommended for h ir remov l bec use it br des the skin nd prom otes b cteri l coloniz tion. If h ir remov l is necess ry, the h ir c n be clipp ed or depil tory c n be pplied (6). P.56

CATHETER-OVER-NEEDLE A c theter-over-needle device is shown in Figure 4.1. The c theter fits snugly o ver the insertion needle, nd h s t pered end to minimize d m ge to the c thet er tip nd soft tissues during insertion. This device c n be held like pencil (i.e., between the thumb nd forefinger) s it is inserted through the skin nd directed to the t rget vessel. When the tip of the needle enters the lumen of p tent blood vessel, blood moves up the needle by c pill ry ction nd enters th e fl shb ck ch mber. When this occurs, the c theter is thre ded over the needle nd into the lumen of the blood vessel s the needle is withdr wn. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 26)03-M y-05 17:31:06

CATHETER INSERTION DEVICES V scul r c nnul tion c n be performed by dv ncing the c theter over guidewire th t is in cont ct with the lumen of blood vessel.

needle or

Ovid: ICU Book Figure 4.1. A c theter-over-needle device. The dv nt ge of c theter-over-needle device is the bility to c nnul te P.57 vessels in simple one-step procedure. The dis dv nt ge is the tendency for the c theter tip to become fr yed s it p sses through the skin nd soft tissues, nd to subsequently d m ge the v scul r endothelium nd promote phlebitis nd thr ombosis. To minimize this risk, the c theter-overneedle device is usu lly reserv ed for c nnul tion of superfici l vessels. CATHETER-OVER-GUIDEWIRE Guidewire- ssisted v scul r c nnul tion w s introduced in the e rly 1950s nd is often c lled the Seldinger technique, fter its inventor. This technique is ill ustr ted in Figure 4.2. A sm ll-bore needle (usu lly 20 g uge) is used to probe for the t rget vessel. When the tip of the needle enters the vessel, thin wire with flexible tip (c lled J-tip bec use of its sh pe) is p ssed through the needle nd into the vessel lumen. The needle is then removed, le ving the wire in pl ce to serve s guide for c nnul tion of the vessel. When c nnul ting dee p vessels, rigid dil tor c theter is first thre ded over the guidewire nd rem oved; this cre tes tr ct th t f cilit tes insertion of the v scul r c theter. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 26)03-M y-05 17:31:06

Ovid: ICU Book

The guidewire technique h s the presumed dv nt ge of minimizing d m ge to soft tissues nd blood vessels by using sm ll-bore probe needle. However, the use o f rigid dil tor c theter ( s expl ined bove) seems to nullify this dv nt ge. Nevertheless, the guidewire technique is currently the preferred method for cen tr l venous nd rteri l c nnul tion (1,2). THE CATHETERS V scul r c theters re composed of pl stic polymers impregn ted with b rium or t ungsten s lts to enh nce r diop city. C theters intended for short-term c nnul t ion (d ys) re usu lly m de of polyureth ne; c theters used for long-term venous ccess (weeks to months) re composed of more flexible nd less thrombogenic deriv tive of silicone. The silicone c theters (e.g., Hickm n nd Brovi c c thet ers) re too flexible for routine percut neous insertion, nd re not ppropri t e for use in the ICU. HEPARIN BONDING http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 26)03-M y-05 17:31:06

Figure 4.2. V scul r c nnul tion with

guidewire (the Seldinger technique).

Ovid: ICU Book Some v scul r c theters re impregn ted or co ted with hep rin to reduce thrombo genicity. However, this me sure h s not proven effective in reducing the inciden ce of c theter- ssoci ted thrombosis (13). Bec use hep rin-co ted c theters c n be source of hep rin-induced thrombocytopeni (see Ch pter 45), c theters used in the ICU should not be impregn ted or co ted with hep rin. P.58 CATHETER SIZE The size of v scul r c theters is commonly expressed in terms of the outside di meter, nd the units of me surement re shown in T ble 4.3. The French size is metric deriv tive equiv lent to the outside di meter in millimeters multiplied by 3; th t is, French size = outside di meter (in mm) 3. The g uge system w s de veloped for wires nd needles, nd h s been dopted for c theters. There is no s imple m them tic l rel tionship between g uge size nd other units of me suremen t, nd t ble of reference v lues such s T ble 4.3 is needed P.59 to m ke the ppropri te conversions (14). G uge sizes usu lly r nge from 14 (l rgest di mete r) to 27 (sm llest di meter). TABLE 4.3. CATHETER SIZES AND COMPARATIVE FLOW RATES As described in Ch pter 1, the ste dy or l min r flow through rigid tube is in fluenced most by the r dius of the tube (see the H genPoisseuille equ tion in the section on peripher l blood flow in Ch pter 1). The influence of tube di meter on flow r te is demonstr ted in T ble 4.3 for gr vity flow of one unit of p cked red blood cells diluted with 250 mL norm l s line flowing through c theters of equ l length (15). Note th t little more th n doubling of tube di meter (fro m 0.7 mm to 1.65 mm) is ssoci ted with lmost qu drupling of flow r te (from 24.7 to 96.3 mL/minute). Thus, c theter size (di meter) is n import nt consider tion if r pid flow r tes re desired. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 26)03-M y-05 17:31:06

Ovid: ICU Book MULTILUMEN CATHETERS Multilumen c theters were introduced for clinic l use in the e rly 1980s, nd r e now used routinely for centr l venous c nnul tion. The design of triple-lume n c theter is shown in Figure 4.3. These c theters h ve n outside di meter of 2 .3 mm (6.9 French) nd m y h ve three ch nnels of equ l di meter (usu lly 18 g u ge) or m y h ve one l rger ch nnel (16 g uge) nd two sm ller ch nnels of equ l di meter (18 g uge). The dist l opening of e ch ch nnel is sep r ted from the ot her by t le st 1 cm to help minimize mixing of infus te solutions.

Multilumen c theters h ve proven to be v lu ble ids bec use they minimize the n umber of venipunctures needed for monitoring nd infusion ther py, yet they do n ot incre se the risk of thrombosis or infection when comp red with single-lumen c theters (13). INTRODUCER CATHETERS Another v lu ble ddition to the f mily of v scul r c theters is the introducer c theter, shown in Figure 4.3. These l rge-bore c theters (8 P.60 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 26)03-M y-05 17:31:06

Figure 4.3. A multilumen c theter (top) ttom).

nd

l rge-bore introducer c theter (bo

Ovid: ICU Book to 9 French) c n be used s conduits for insertion nd remov l of sm ller v scul r c theters (including multilumen c theters nd pulmon ry rtery c theters) thr ough single venipuncture. The side- rm infusion port on the c theter provides n ddition l infusion line, nd llows continuous flush to prevent thrombus f orm tion round sm ller c theters th t sit in the lumen of the introducer c thet er. This side- rm infusion port lso llows the introducer c theter to be used s st nd- lone infusion device ( rubber membr ne on the hub of the c theter pr ovides n effective se l when fluids re infusing through the side- rm port of t he c theter). The l rge di meter of introducer c theters m kes them p rticul rly v lu ble s infusion devices when r pid infusion r tes re necess ry (e.g., in the resuscit tion of m ssive hemorrh ge).

ACCESS ROUTES The following is brief description of common v scul r ccess routes in the rm ( ntecubit l veins nd r di l rtery), the thor cic inlet (subcl vi n nd jugul r veins), nd the groin (femor l rtery nd vein). P.61 ANTECUBITAL VEINS The veins in the ntecubit l foss provide r pid nd s fe v scul r ccess for c ute resuscit tive ther py. Although long c theters c n be inserted into the nte cubit l veins nd dv nced into the superior ven c v , such peripher lly insert ed centr l venous c theters (PICC devices) re more ppropri te for home infusio n ther py th n for tre ting critic lly ill p tients (16). Short c theters (5 to 7 cm) re preferred for cute resuscit tion vi the ntecubit l veins bec use th ey re more e sily inserted nd llow more r pid infusion r tes th n the longer PICC c theters. An tomy The surf ce n tomy of the ntecubit l veins is shown in Figure 4.4. The b silic vein runs long the medi l spect of the ntecubit l foss , nd the ceph lic ve in is situ ted on the opposite side. The b silic vein is preferred for c nnul ti on bec use it runs str ighter nd less v ri ble course up the rm th n the cep h lic vein. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 26)03-M y-05 17:31:06

Ovid: ICU Book

Blind Insertion If the ntecubit l veins re neither visible nor p lp ble, p lp te the br chi l rtery pulse t point 1 inch bove the ntecubit l cre se. The b silic vein (o r br chi l vein) should lie just medi l to the p lp ted pulse t this point, nd c n be entered by inserting the c theter-over-needle device through the skin t 35 to 45 ngle nd dv ncing the needle until blood return is noted. This ppro ch h s reported success r te of 80% (17). Injury to the medi n nerve (which i s lso medi l to the rtery, but deep to the veins) c n occur with excessive mov ement of the probe needle. Comment C nnul tion of the ntecubit l veins is recommended (18,19). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 26)03-M y-05 17:31:06

Insertion Technique The p tient need not be supine, but the rm should be str ight nd ntecubit l veins c n be distended by tourniquet or by infl ting re cuff to just bove the di stolic pressure (this llows rteri l impeding venous outflow). Once the veins re visible or p lp ble, r-needle device is used to insert short 16- or 18-g uge c theter lic or ceph lic vein.

Figure 4.4. The v scul r topology of the ntecubit l foss

nd wrist. bducted. The blood pressu inflow while c theter-ove into the b si

Ovid: ICU Book For r pid venous ccess (e.g., c rdiopulmon ry resuscit tion) For thrombolytic t her py in cute myoc rdi l inf rction For tr um victims who require thor cotomy P.62 Remember th t the shorter the c theter, the more r pid the flow r te throu gh the c theter (see Ch pter 1). Thus, insertion of short c theters into the nt ecubit l veins permits more r pid volume resuscit tion th n insertion of the lon ger centr l venous c theters. RADIAL ARTERY The r di l rtery is f vored site for rteri l c nnul tion bec use the vessel is superfici l nd ccessible nd the insertion site is e sy to keep cle n. The m jor dis dv nt ge of the r di l rtery is its sm ll size, which limits the succ ess r te of c nnul tion nd promotes v scul r occlusion. An tomy The surf ce n tomy of the r di l nd uln r rteries is shown in Figure 4.4. The r di l rtery is usu lly p lp ble t point just medi l P.63 to the styloid pr ocess of the r dius. The uln r rtery is on the opposite (medi l) side of the wr ist, just l ter l to the pisiform bone. Although the r di l rtery is preferred, the uln r rtery is the l rger of the two rteries nd should be e sier to c nn ul te (20). The Allen Test The Allen test ev lu tes the c p city of the uln r rtery to supply blood to the digits when the r di l rtery is occluded. The test is performed by first occlu ding the r di l nd uln r rteries with the thumb nd index finger. The p tient is then instructed to r ise the wrist bove the he d nd to m ke fist repe ted ly until the fingers turn white. The uln r rtery is then rele sed, nd the time required for return of the norm l color to the fingers is recorded. A norm l re sponse time is 7 seconds or less, nd del y of 14 seconds or gre ter is eviden ce of insufficient flow in the uln r rtery. Although positive Allen test (i.e ., 14 seconds or longer for return of color to the digits) is often st ted s contr indic tion to r di l rtery c nnul tion, in numerous inst nces the Allen t est h s indic ted in dequ te flow in the uln r rtery, yet subsequent r di l rt ery c nnul tion h s been uneventful (2,21). Thus, positive Allen test is not contr indic tion to r di l rtery c nnul tion. Another limit tion is the need f or p tient cooper tion to perform the test. Therefore, this test is not worth th e time it t kes. Insertion Technique http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 26)03-M y-05 17:31:06

Ovid: ICU Book The wrist should be hyperextended to bring the rtery closer to the surf ce. A s hort 20-g uge c theter is ppropri te, nd c n be inserted by c theter-over-ne edle device or by the guidewire technique. When using c theter-over-needle dev ice, the following through- nd-through technique is recommended: When the needle tip first punctures the rtery ( nd blood ppe rs in the fl shb ck ch mber), th e tip of the c theter is just outside the vessel. To position the c theter tip i n the lumen of the vessel, the needle is p ssed completely through the rtery n d then withdr wn until blood returns g in through the needle. At this point, th e c theter tip should be in the lumen of the rtery, nd the c theter c n be dv nced while the needle is retr cted. If two ttempts t c nnul tion re unsucces sful, switch to n ltern tive site (to reduce tr um to the vessel). m ny s 25% of r di l rtery c nnul tions, but d Despite being well toler ted in most p tients, c (or ny rtery) should be reserved for monitorin be used s P.64 convenience me sure for monito components (23).

THE SUBCLAVIAN VEIN More th n 3 million centr l venous c nnul tions re performed ye rly in the Unit ed St tes (24), nd m jority of these procedures re performed vi the subcl v i n vein (25). The subcl vi n vein is well suited for c nnul tion bec use it is l rge vessel ( bout 20 mm in di meter) nd is prevented from coll psing by its surrounding structures. The immedi te risks of subcl vi n vein c nnul tion incl ude pneumothor x (1% to 2%) nd hemothor x ( less th n 1%) (25). The incidence o f bleeding is no different in the presence or bsence of co gulop thy (26); th t is, the presence of co gul tion disorder is not contr indic tion to subcl vi n vein c nnul tion.

An tomy The subcl vi n vein is continu tion of the xill ry vein s it p sses over the first rib, nd the pic l pleur lies bout 5 mm deep to the vein t its point of origin. As shown in Figure 4.5, the subcl vi n vein runs most of its course long the underside of the cl vicle. The vein runs long the outer surf ce of the nterior sc lene muscle, which sep r tes the vein from its comp nion rtery on the underbelly of the muscle. At the thor cic inlet, the subcl vi n vein meets t he intern l jugul r vein to form the br chioceph lic vein. The convergence of th e right nd left br chioceph lic veins forms the superior ven c v . http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 26)03-M y-05 17:31:06

Comment Arteri l occlusion occurs in s igit l ischemi is r re (2,22). nnul tion of the r di l rtery g blood pressure, nd is not to ring blood g ses or other blood

Ovid: ICU Book Figure 4.5. A surf ce view of the l rge centr l veins s they converge t the th or cic inlet. The circul r m rkers indic te skin insertion sites for c nnul tion of the subcl vi n vein (1 nd 2) nd intern l jugul r vein (3 nd 4). An tomic Dist nces The lengths of the v scul r segments involved in subcl vi n ( nd intern l jugul r) vein c nnul tion re shown in T ble 4.4. The ver ge dist nce from venipunctu re site to the right trium is 14.5 cm nd P.65 18.5 cm for right-sided nd left -sided c nnul tions, respectively. These dist nces re f r shorter th n c theter lengths recommended for right-sided (20 cm) nd left-sided (30 cm) centr l veno us c nnul tions, nd re more consistent with recent report showing th t the ver ge dist nce to the right trium is 16.5 cm in centr l venous c nnul tion fro m either side in dults (27). Therefore, to void pl cing c theter tips in the r ight trium (which c n le d to c rdi c perfor tion nd f t l c rdi c t mpon de), ll centr l venous c theters should be no longer th n 15 or 16 cm in length (27 ). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 26)03-M y-05 17:31:06

Ovid: ICU Book TABLE 4.4. ANATOMIC DISTANCES IN CENTRAL VENOUS CANNULATION INSERTION TECHNIQUE The p tient is pl ced supine, with rms t the sides nd he d f ced w y from th e insertion site. A towel roll c n be pl ced between the shoulder bl des, but th is is uncomfort ble nd unnecess ry. Identify the cl vicul r insertion of the st ernocleidom stoid muscle. The subcl vi n vein lies just underne th the cl vicle where the muscle inserts P.66 onto the cl vicle. The vein c n be entered from ei ther side of the cl vicle. Infr cl vicul r Appro ch (Insertion Site 1 in Figure 4.5). Identify the l ter l m rgin of the sternocleidom stoid muscle s it inserts on t he cl vicle. The c theter is inserted in line with this m rgin, but below the cl vicle. Insert the probing needle (18 or 20 g uge) with the bevel pointing upw r d (tow rd the ceiling) nd dv nce the needle long the underside of the cl vicl e nd tow rd the supr stern l notch. The p th of the needle should be p r llel t o the p tient's b ck. When the vein is entered, turn the bevel of the needle to 3 o'clock so the guidewire thre ds in the direction of the superior ven c v . Supr cl vicul r Appro ch (Insertion Site 2 in Figure 4.5). Identify the ngle formed by the l ter l m rgin of the sternocleidom stoid muscl e nd the cl vicle. The probe needle is inserted so th t it bisects this ngle. Keep the bevel of the needle f cing upw rd nd direct the needle under the cl vi cle in the direction of the opposite nipple. The vein should be entered t dis t nce of 1 to 2 cm from the skin surf ce (the subcl vi n vein is more superfici l in the supr cl vicul r ppro ch). When the vein is entered, turn the bevel of the needle to 9 o'clock so the http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 26)03-M y-05 17:31:06

Ovid: ICU Book

Comment P tient comfort nd e se of insertion re the most compelling re sons to select the subcl vi n vein for centr l venous ccess. Selection of the infr cl vicul r versus supr cl vicul r ppro ch is l rgely m tter of person l preference. Some recommend voiding the subcl vi n vein in ventil tor-dependent p tients bec use of the risk of pneumothor x. However, the risk of pneumothor x is too sm ll to justify b ndoning the subcl vi n vein in p tients with respir tory f ilure. THE INTERNAL JUGULAR VEIN C nnul tion of the intern l jugul r vein reduces (but does not elimin te) the ri sk of pneumothor x, but introduces new risks (e.g., c rotid rtery puncture nd thor cic duct injury). An tomy The intern l jugul r vein is loc ted under the sternocleidom stoid muscle in the neck nd, s shown in Figure 4.5, the vein follows n oblique course s it runs down the neck. When the he d is turned to the opposite side, the vein forms s tr ight line from the pinn of the e r to the sternocl vicul r joint. Ne r the b se of the neck, the intern l jugul r vein becomes the most l ter l structure in the c rotid she th (which cont ins the c rotid rtery s ndwiched between the ve in l ter lly nd the v gus nerve medi lly). P.67

Insertion Technique The right side is preferred bec use the vessels run str ighter course to the r ight trium. The p tient is pl ced in supine or Trendelenburg position, with t he he d turned to the opposite side. The intern l jugul r vein c n be entered fr om n nterior or posterior ppro ch. The Anterior Appro ch (Insertion Site 4 in Figure 4.5). The nterior ppro ch is through tri ngul r region cre ted by two he ds of the sternocleidom stoid muscle. The c rotid rtery is p lp ted in the tri ngle nd retr cted medi lly. The probe needle is inserted t the pex of the tri ngle wit h the bevel f cing up, nd the needle is dv nced tow rd the ipsil ter l nipple, t 45 ngle with the skin surf ce. If the vein is not encountered by depth o f 5 cm, the needle is withdr wn 4 cm nd dv nced g in in more l ter l direct ion. When vessel is entered, look for puls tions. If the blood is red nd puls ting, you h ve entered the c rotid rtery. In this situ tion, remove the needle nd t mpon de the re for 5 to 10 minutes. When the c rotid rtery h s been pu nctured, no further ttempts should be m de on either side bec use puncture of b oth rteries c n h ve serious consequences.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 26)03-M y-05 17:31:06

guidewire thre ds in the direction of the superior ven

c v .

Ovid: ICU Book The Posterior Appro ch (Insertion Site 3 in Figure 4.5). The insertion site for this ppro ch is 1 centimeter superior to the point where the extern l jugul r vein crosses over the l ter l edge of the sternocleidom st oid muscle. The probe needle is inserted with the bevel positioned t 3 o'clock. The needle is dv nced long the underbelly of the muscle in direction pointi ng to the supr stern l notch. The intern l jugul r vein should be encountered 5 to 6 cm from the skin surf ce with this ppro ch (28). C rotid Artery Puncture. If the c rotid rtery h s been punctured with probing needle, the needle shoul d be removed nd pressure should be pplied to the site for t le st 5 minutes ( 10 minutes is recommended for p tients with co gulop thy). No further ttempts should be m de to c nnul te the intern l jugul r vein on either side, to void puncture of both c rotid rteries. If the c rotid rtery h s been in dvertently c nnul ted, the c theter should not be removed, s this could provoke serious he morrh ge. In this situ tion, v scul r surgeon should be consulted immedi tely. Comment As with the subcl vi n vein, c nnul tion of the intern l jugul r vein is s fe n d effective when performed by skilled oper tors. However, sever l dis dv nt ges of intern l jugul r c nnul tion deserve mention. ( ) Accident l puncture of the c rotid rtery is reported in 2 to 10% of ttempted c nnul tions (28). (b) Aw ke p tients often compl in of the limited neck mobility when the intern l jugul r vein is c nnul ted. (c) In git ted p tients, in ppropri te neck flexion c n res ult in thrombotic occlusion of the c theter nd vein. (d) In p tients with tr ch eostomies, the insertion site c n be exposed to infected secretions th t dr in f rom the tr che l stom . P.68 THE EXTERNAL JUGULAR VEIN C nnul tion of the extern l jugul r vein h s two dv nt ges: ( ) There is no ris k of pneumothor x, nd (b) hemorrh ge is e sily controlled. The m jor dr wb ck i s difficulty dv ncing the c theter.

Insertion Technique http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 26)03-M y-05 17:31:06

An tomy The extern l jugul r vein runs long line extending from the ngle of the j to point midw y long the cl vicle. The vein runs obliquely cross the surf of the sternocleidom stoid muscle nd joins the subcl vi n vein t n cute le. This cute ngle is the m jor impediment to dv ncing c theters th t h ve en inserted into the extern l jugul r vein.

w ce ng be

Ovid: ICU Book The p tient is pl ced in the supine or Trendelenburg position, with the he d tur ned w y from the insertion site. If necess ry, the vein c n be occluded just b ove the cl vicle (with the forefinger of the nondomin nt h nd) to engorge the en try site. As m ny s 15% of p tients so not h ve n identifi ble extern l jugul r vein, even under optim l conditions of vein engorgement (28). The extern l jug ul r vein h s little support from surrounding structures, so the vein should be nchored between the thumb nd forefinger when the needle is inserted. The bevel of the needle should be pointing upw rd when it enters the vein. The recommende d insertion point is midw y between the ngle of the j w nd the cl vicle (see F ig. 4.5). Use 16-g uge single-lumen c theter th t is 10 to 15 cm in length. If the c theter does not dv nce e sily, do not force it, s this m y result in v scul r perfor tion t the junction between the extern l jugul r nd subcl vi n v eins. Comment This ppro ch is best reserved for tempor ry ccess in p tients with severe co gulop thy, p rticul rly when the oper tor is inexperienced nd does not feel co mfort ble c nnul ting the subcl vi n or intern l jugul r veins. Contr ry to popu l r belief, c nnul tion of the extern l jugul r is not lw ys e sier to ccompli sh th n centr l venous c nnul tion bec use of the difficulty in dv ncing c thet ers p st the cute ngle t the junction of the subcl vi n vein.

THE FEMORAL VEIN The femor l vein is the e siest of the l rge veins to c nnul te nd lso does no t c rry risk of pneumothor x. The dis dv nt ges ssoci ted with this route re venous thrombosis (10%), femor l rtery puncture P.69 (5%), nd limited bility to flex the hip (which c n be bothersome for w ke p tients). Contr ry to popul r belief, the infection r te with femor l vein c theters is no different from t h t of subcl vi n or intern l jugul r vein c theters (28). An tomy The n tomy of the femor l she th is shown in Figure 4.6. The femor l vein is th e most medi l structure in the femor l she th nd is situ ted just medi l to the femor l rtery. At the inguin l lig ment, the femor l vessels re just few ce ntimeters below the skin surf ce. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (17 of 26)03-M y-05 17:31:06

Ovid: ICU Book Figure 4.6. The n tomy of the femor l she th. Insertion Technique P lp te the femor l rtery just below the inguin l cre se nd insert the needle (bevel up) 1 to 2 cm medi l to the p lp ted pulse. Adv nce P.70 the needle t 45 ngle to the skin surf ce, entering the vein t depth of 2 to 4 cm. Once in the vessel, if the c theter or guidewire will not p ss beyond the tip of the nee dle, tilt the needle so th t it is more p r llel to the skin surf ce (this m y m ove the needle tip w y from the f r side of the vessel w ll nd into more direc t cont ct with the lumen of the vessel). Femor l vein c theters should be t le st 15 cm long. Blind Insertion If the femor l rtery pulse is not p lp ble, dr w n im gin ry line from the nt erior superior ili c crest to http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (18 of 26)03-M y-05 17:31:06

Ovid: ICU Book the pubic tubercle, nd divide the line into three equ l segments. The femor l rtery lies t the junction between the middle nd most medi l segment, nd the f emor l vein is 1 to 2 cm medi l to this point. This method of loc ting the femor l vein h s reported success r te of over 90% (29).

Comment Femor l vein c nnul tion is usu lly reserved for p tients who re p r lyzed or c om tose nd immobile. This ppro ch is not recommended for c rdiopulmon ry resus cit tion (bec use of the del yed tr nsit times for bolus drug injections) (18) o r in p tients with bleeding disorders (28). THE FEMORAL ARTERY C nnul tion of the femor l rtery is usu lly reserved for situ tions where r di l rtery c nnul tion is unsuccessful or contr indic ted. Despite its reserve st tus, the femor l rtery is l rger th n the r di l rtery, nd is e sier to c nnu l te. The complic tions of femor l rtery c nnul tion re the s me s for r di l rtery c nnul tion (thrombosis, bleeding, nd infection). The incidence of infe ction is the s me with r di l nd femor l rtery c theters, nd the incidence of thrombosis is lower with femor l rtery c nnul tion (2). Thrombosis of the femo r l rtery, like th t in the r di l rtery, only r rely results in troublesome i schemi in the dist l extremity (2). Loc liz tion nd c nnul tion of the femor l rtery proceeds s described in the section on femor l vein c nnul tion. The Se ldinger technique is preferred for c theter insertion, nd c theters should be 1 8 g uge in di meter nd 15 to 20 cm long. Comment Femor l rtery c nnul tion is vi ble ltern tive nd m y be prefer ble to r di l rtery c nnul tion in p tients who re p r lyzed or otherwise immobile, unles s they h ve signific nt co gulop thy (in P.71 which c se the r di l rtery is preferred). The incidence of thrombotic complic tions is lower in femor l rtery c nnul tions, nd the pressure in the femor l rtery more closely pproxim tes the pressure in the ort th n does the pressure in the r di l rtery (see Ch pt er 8). IMMEDIATE CONCERNS VENOUS AIR EMBOLISM In dvertent ir entry is one of the most fe red complic tions of centr l venous c nnul tion. The import nce of m int ining closed system during insertion is h ighlighted by the following st tement:

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (19 of 26)03-M y-05 17:31:06

A pressure gr dient of 4 mm Hg long

14-g uge c theter c n entr in

ir

Ovid: ICU Book

Preventive Me sures Prevention is the h llm rk of reducing the morbidity nd mort lity of venous ir embolism. The most effective method of preventing ir entry is to keep the veno us pressure more positive th n tmospheric pressure. This is f cilit ted by pl c ing the p tient in the Trendelenburg position with the he d 15 below the horizont l pl ne. Remember th t the Trendelenburg position does not prevent venous ir e ntry bec use p tients still gener te neg tive intr thor cic pressures while in t he Trendelenburg position. When ch nging connections in centr l venous line, tempor ry positive pressure c n be cre ted by h ving the p tient hum udibly. T his not only produces positive intr thor cic pressure, but llows clinici ns t o he r when the intr thor cic pressure is positive. In ventil tor-dependent p ti ents, the nurse or respir tory ther pist should initi te mech nic l lung infl tion when ch nging connections. the procedur cross p t n cute isch the right he rt,

Ther peutic M neuvers If venous ir embolism is suspected, immedi tely pl ce the p tient with the le ft side down, nd ttempt to spir te ir directly from the venous line. In dire circumst nces, needle should be inserted through the chest w ll nd into the right ventricle to spir te the ir. Unfortun tely, P.72 the mort lity in severe c ses of venous ir embolism rem ins high despite these m neuvers. PNEUMOTHORAX Pneumothor x is concern prim rily with subcl vi n vein c nnul tion but c n ls o complic te jugul r vein c nnul tion (2,30). This is one re son th t postinsert ion chest films re recommended fter ll centr l venous c nnul tions (or ttemp ts). If possible, postinsertion films should be obt ined in the upright position nd during expir tion. Expir tory films f cilit te the detection of sm ll pneum othor xes bec use expir tion decre ses the volume of ir in the lungs, but not t he volume of ir in the pleur l sp ce. Thus, during expir tion, the volume of i r in the pleur l sp ce is l rger fr ction of the tot l volume of the hemithor x, thereby m gnifying the r diogr phic ppe r nce of the pneumothor x (31). Upri ght films re not lw ys possible in ICU p tients. When supine films re necess ry, remember th t http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (20 of 26)03-M y-05 17:31:06

Clinic l Present tion The usu l present tion is cute onset of dyspne th t occurs during e. Hypotension nd c rdi c rrest c n develop r pidly. Air c n p ss ent for men ov le nd obstruct the cerebr l circul tion, producing emic stroke. A ch r cteristic mill wheel murmur c n be he rd over but this murmur m y be fleeting.

r te of 90 mL/second

nd c n produce

f t l ir embolus in 1 second (30).

Ovid: ICU Book

Del yed Pneumothor x Pneumothor xes m y not be r diogr phic lly evident until 24 to 48 hours fter ce ntr l venous c nnul tion (31,33). Therefore, the bsence of pneumothor x on n immedi te postinsertion chest film does not bsolutely exclude the possibility of c theter-induced pneumothor x. This is n import nt consider tion in p tien ts who develop dyspne or other signs of pneumothor x in the first few d ys fte r centr l venous c nnul tion. In the bsence of signs nd symptoms, there is lit tle justific tion for seri l chest films following centr l venous c theter pl ce ment. CATHETER TIP POSITION The properly pl ced centr l venous c theter should run p r llel to the superior ven c v , nd the tip of the c theter should be positioned bove the junction o f the superior ven c v nd right trium. The following conditions w rr nt corr ective me sures. Tip Ag inst the W ll of the Ven C v C theters inserted from the left side must m ke n cute turn downw rd when they re ch the superior ven c v . If they f il to m ke this turn, the c theters c n end up in position like the one shown in Figure 4.7. P.73 The tip of the c th eter is up g inst the l ter l w ll of the ven c v , nd in this position, the c theter tip c n st b the vessel w ll nd perfor te the ven c v . Therefore, c theters th t but the w ll of the ven c v should be repositioned s soon s po ssible. (The problem of v scul r perfor tion is discussed in more det il in Ch p ter 5.) http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (21 of 26)03-M y-05 17:31:06

pleur n the n the ee Ch

l ir does not often collect t the pex of the lung when the p tient is i supine position (32,33). In this situ tion, pleur l ir tends to collect i subpulmonic recess nd long the nteromedi l border of the medi stinum (s pter 28).

Ovid: ICU Book

Tip in the Right Atrium The Food nd Drug Administr tion h s issued strong w rning bout the risk of c rdi c perfor tion from c theter tips th t re dv nced into the he rt (24). How ever, c rdi c perfor tion is r re complic tion of centr l venous c nnul tion ( 27), even though over h lf of centr l venous c theters m y be mispl ced in the r ight trium (27). Nevertheless, t mpon de is often f t l, so c rdi c pl cement o f c theters should be voided. A few me sures help to minimize the risk of c rdi c perfor tion. The most effective me sure is to use shorter c theters, s recom mended e rlier. The tip of indwelling c theters should be bove the third right cost l c rtil ge (this is the level where the ven c v meets the right trium). If the nterior portion of the third rib c nnot be visu lized, keep the c thete r tip t or bove the tr che l c rin .

Rosen M, L tto P, Ng S. H ndbook of percut neous centr l venous c theteriz tion. 2nd ed. Phil delphi : WB S unders, 1992. P.74 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (22 of 26)03-M y-05 17:31:06

REFERENCES GENERAL TEXTS Benumof JL, ed. Clinic l procedures in nesthesi i : JB Lippincott, 1992.

nd intensive c re. Phil delph

Figure 4.7. Centr l venous c theter in c v . (See lso Figure 5.1).

position to perfor te the superior ven

Ovid: ICU Book REVIEWS 1. Agee KR, B lk RA. Centr l venous c theteriz tion in the critic lly ill p tien t. Crit C re Clin North Am 1992;8:677686. 2. Cl rk VL, Kruse JA. Arteri l c theteriz tion. Crit C re Clin North Am 1992;8: 687698.

PREPARATION 4. Doebbeling BN, St nley GL, Sheetz CT, et l. Comp r tive effic cy of ltern t ive h ndw shing gents in reducing nosocomi l infections in intensive c re units . N Engl J Med 1992;327:8893. 5. Centers for Dise se Control. Guidelines for the prevention of tr nsmission of hum n immunodeficiency virus nd hep titis B virus to he lth-c re nd public-s fety workers. Morbid Mort Weekly Rev 1989;38:137. 6. Centers for Dise se Control Working Group. Guidelines for prevention of intr v scul r infections. In: Guidelines for the prevention of nosocomi l infections. W shington, DC: Dep rtment of He lth & Hum n Services, Public He lth Service, 1 981. 7. Geberding JL. Risks to he lth c re workers from exposure to hep titis B virus , hum n immunodeficiency virus, nd cytomeg lovirus. Infect Dis Clin North Am 19 89;3:735745.

10. Bub ck ME, Reed CE, Fr nsw y AF, et l. Allergic re ctions to l tex mong he lth-c re workers. M yo Clin Proc 1992;67:10751079.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (23 of 26)03-M y-05 17:31:06

9. Needlesticks: pre ching to the seroconverted. L ncet 1992;340:640641 (editori l).

8. M lcolm JA, Dobson PM, Sutherl nd DC. Combin tion chemoprophyl xis lestick injury. L ncet 1993;341:112113.

fter need

3. B r nowski L. Centr l venous 22 references).

ccess devices. J Intr ven Nurs 1993;16:167194 (1

Ovid: ICU Book 11. L rson EL. Guidelines for use of topic l ntimicrobi l gents. APIC Guidelin es for Infection Control Pr ctice. Am J Infect Control 1988;16:253266. 12. Wy tt WJ, Beckett TA, Bonet V, D vis SM. Comp r tive effic cy of surgic l sc rub solutions on control of skin microflor . Infect Surg 1990;9:1721. CATHETERS 13. Rosen M, L tto P, Ng S. H ndbook of percut neous centr l venous c theteriz t ion. 2nd ed. Phil delphi : WB S unders, 1992;1130. 14. L wson M, Vertenstein MJ. Methods for determining the intern l volume of cen tr l venous c theters. J Intr ven Nurs 1993;16:148155. 15. de l Roche MRP, G uthier L. R pid tr nsfusion of p cked red blood cells: ef fects of dilution, pressure, nd c theter size. Ann Emerg Med 1993;22:15511555.

17. Vyskocil JJ, Kruse JA, Wilson RF. Altern tive techniques for g ining venous ccess. J Crit Illness 1993;8:435442. P.75

19. Ros DH, Griffin CC, Fl n g n JJ, M chiedo GW. A comp rison of intr venous ccess sites for bolus injections during shock nd resuscit tion fter emergency room thor cotomy with nd without ortic cross-cl mping. Am Surgeon 1990;56:56757 0. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (24 of 26)03-M y-05 17:31:06

18. Americ n He rt Associ tion. Textbook of dv nced c rdi c life support. D ll s: Americ n He rt Associ tion, 1994.

ANTECUBITAL VEINS 16. H d w y LC. An overview of v scul r ccess devices inserted vi t l re . J Intr ven Nurs 1989;13:297306.

the ntecubi

Ovid: ICU Book ARTERIAL CANNULATION 20. M thers LH. An tomic l consider tions in obt ining rteri l ccess. J Intens ive C re Med 1990;5:110119. 21. Thompson SR, Hirshberg A. Allen's test revisited. Crit C re Med 1988;16:915.

CENTRAL VENOUS CANNULATION 24. Food nd Drug Administr tion. Prec utions necess ry with centr l venous c th eters. FDA Drug Bull 1989;July:1516. 25. Wegener ME. Complic tions of centr l venous line pl cement. Contemp Surg 199 3;42:266 268.

27. McGee WT, Ackerm n BL, Rouben LR, et l. Accur te pl cement of centr l venou s c theters: prospective, r ndomized, multicenter tri l. Crit C re Med 1993;21 :11181123. 28. Seneff MG. Centr l venous c theteriz tion. A comprehensive review. Intensive C re Med 1987;2:163175, 218232. 29. Getzen LC, Poll ck EW. Short-term femor l vein c theteriz tion. Am J Surg 19 79;138:875 877. IMMEDIATE CONCERNS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (25 of 26)03-M y-05 17:31:06

26. Foster PF, Moore LR, S nk ry HN, et l. Centr l venous c theteriz tion in p tients with co gulop thy. Arch Surg 1992;127:273275.

23. Tenholder MF. The pendulum

nd the rteri l line. Chest 1993;104:16501651.

22. Slogoff, S, Ke ts AS, Arlund C. On the s fety of r di l Anesthesiology 1983;59:4247.

rtery c nnul tion.

Ovid: ICU Book 30. Sl den A. Complic tions of inv sive hemodyn mic monitoring in the intensive c re unit. Curr Probl Surg 1988;25:69145. 31. M rino PL. Del yed pneumothor x: complic tion of subcl vi n vein c theteri z tion. J P renter Enter l Nutr 1985;9:232. 32. Tocino IM, Miller MH, F irf x WR. Distribution of pneumothor x in the supine nd semirecumbent critic lly ill dult. Am J R diol 1985;144:901905.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (26 of 26)03-M y-05 17:31:06

33. Collin GR, Cl rke LE. Del yed pneumothor x: c theteriz tion. Surg Rounds 1994;17:589594.

complic tion of centr l venous

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n

ROUTINE CATHETER CARE The following pr ctices re designed to prevent or limit the complic tions cre t ed by indwelling v scul r c theters (1,2). M ny of these preventive pr ctices h ve little or no proven v lue. PROTECTIVE DRESSINGS C theter insertion sites on the skin re covered t ll times s st nd rd nti septic me sure (1,2). The different protective dressings nd their comp r tive f e tures re shown in T ble 5.1. The st nd rd dressing is sterile g uze coverin g th t is nchored to the skin by hypo llergenic t pe nd is repl ced every 48 h ours (1,2). Occlusive dressings m de of tr nsp rent polyureth ne or colloid gels h ve lso become popul r in recent ye rs bec use they llow inspection of the c theter insertion site (for signs of infection) without bre king the protective se l on the skin. However, occlusive dressings h ve one dis dv nt ge th t deserv es emph sis: They promote microbi l coloniz tion on the underlying skin (5,6 nd 7). These dressings block the esc pe of w ter v por from the underlying skin n d cre te moist environment th t enh nces the growth of skin microflor . Some o cclusive dressings re more perme ble to w ter v por th n others (see T ble 5.1) ; however, P.77 ll tr p moisture on the skin to some extent, nd ll occlusive dressings promote skin coloniz tion. However, only the dressings th t re comple tely impervious to w ter v por show higher incidence of c theter-rel ted septi cemi (5,6). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 22)03-M y-05 17:33:40

Ch pter 5 THE INDWELLING VASCULAR CATHETER This ch pter is continu tion of Ch pter 4, nd describes the routine c re dverse complic tions of indwelling v scul r c theters. The first section is oted to the d ily c re of c theters nd insertion sites, (1,2) nd the fin l tions focus on c theter-rel ted complic tions, with emph sis on c theter-rel infections (3,4).

nd dev sec ted

Ovid: ICU Book TABLE 5.1. COMPARATIVE FEATURES OF PROTECTIVE DRESSINGS Thus, occlusive skin dressings produce n effect th t is opposite th t desired f rom protective dressings. B sed on this dverse effect nd the dded cost of occ lusive dressings (which is bout three times the cost of g uze dressings), the r outine use of occlusive dressings is not recommended. ANTIMICROBIAL OINTMENT Another common ntiseptic pr ctice is the pplic tion of polymicrobi l ointmen t (such s polymyxinneomycinb citr cin) to the c theter insertion site on the skin . However, this pr ctice h s not been shown to reduce the incidence of c theterrel ted infections (4), nd therefore is not recommended. REPLACING CATHETERS The incidence of c theter- ssoci ted infections is higher in c theters th t rem in in pl ce for longer th n 3 d ys (8). This observ tion led to the common pr ct ice of repl cing v scul r c theters (usu lly over guidewire) every few d ys to reduce the risk of c theter- ssoci ted infections. However, repl cement of v sc ul r c theters t regul r interv ls, using either guidewire exch nge or new ve nipuncture site, does not reduce the incidence of c theter-rel ted infections (9 ) nd m y ctu lly incre se the risk of c theter- ssoci ted complic tions (both mech nic l nd infectious) (10). Therefore, routine repl cement of indwelling v scul r c theters is not recommended. INDICATIONS FOR REPLACING VASCULAR CATHETERS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 22)03-M y-05 17:33:40

Ovid: ICU Book When there is purulence or spre ding erythem t the c theter insertion site, us e new venipuncture site. P.78 When c theter-rel ted infection is suspected, us e guidewire exch nge. When the tip or intr derm l portion of previously remove d c theter shows signific nt infection (more th n 15 colony-forming units on sem iqu ntit tive culture), use new venipuncture site. When c theter h s been pl ced emergently (e.g. c rdiopulmon ry rrest), without strict septic technique, use guidewire exch nge.

FLUSHING CATHETERS All v scul r c theters re flushed routinely to m int in p tency, nd the st nd rd flush solution is hep rinized s line (hep rin concentr tion r nging from 10 t o 1000 units/mL) (1,2,11). Indwelling c theters th t re used only intermittentl y re c pped nd filled with hep rinized s line when idle. (The c p th t se ls t he proxim l end of c theter cre tes p rti l v cuum th t holds the flush solu tion in pl ce; hence, the term hep rin lock is used to describe this device.) Ar teri l c theters re flushed continuously (usu l r te is 3 mL/hour using press urized b g to drive the flush solution through the c theter (12). Altern tives to Hep rin There re two dis dv nt ges to the use of hep rinized s line s the st nd rd flu sh solution. The first is the dded cost of the hep rin, which c n exceed $65,00 0 per ye r for e ch hospit l (11). The second dis dv nt ge is the risk of hep ri n-induced thrombocytopeni (see Ch pter 45). The dis dv nt ges of hep rinized fl ushes c n be elimin ted by dopting ltern tive flush techniques such s the one s shown in T ble 5.2. S line (without hep rin) h s proved to be s effective s hep rinized P.79 s line for flushing venous c theters (11), nd routine flushes m y not be necess ry for peripher l venous c theters (13). S line is not lw ys n effective ltern tive to hep rinized s line for flushing rteri l c theters ( 12). If there is contr indic tion to the use of hep rinized flushes (e.g., hep rin-induced thrombocytopeni ) 1.4% sodium citr te solution c n be used for fl ushing rteri l c theters (14). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 22)03-M y-05 17:33:40

Ovid: ICU Book TABLE 5.2. ALTERNATIVES TO HEPARINIZED FLUSHES MECHANICAL COMPLICATIONS The mech nic l complic tions of indwelling c theters c n be cl ssified s occlus ive (e.g., c theter or v scul r occlusion) or erosive (e.g., v scul r or c rdi c perfor tion). The following re the more common or prevent ble mech nic l compl ic tions. CATHETER OCCLUSION Sources of c theter occlusion include sh rp ngles or kinks nd loc lized indent tions long the c theter (these re usu lly cre ted during c theter insertion), long with thrombi (from b ckw sh of blood into the c theter) nd insoluble pre cipit tes in the infusion solutions (usu lly medic tions or inorg nic s lts). Si gns of occlusion include limited flow (p rti l occlusion), tot l cess tion of fl ow (complete occlusion), nd unidirection l cess tion of flow (withdr w l occlus ion).

Insoluble Precipit tes The subst nces most likely to form insoluble precipit tes in intr venous infusio n solutions re s follows (15): Medic tions: b rbitur tes, di zep m, digoxin, p henytoin, nd trimethoprim-sulf Anionc tion complexes: c lcium phosph te nd hep rin minoglycoside complexes Precipit tion is most often c used by inherent hydro phobic beh vior in the n tive subst nce (e.g., di zep m or digoxin) nd cid or lk line pH in the infusion solutions (e.g., c lcium nd phosph te http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 22)03-M y-05 17:33:40

Ovid: ICU Book

Restoring P tency Every effort should be m de to relieve c theter occlusion nd prevent c theter r epl cement. Guidewires should never be p ssed through c theters to relieve obstr uction bec use of the risk of dislodging the obstructing m ss nd cre ting n em bolus. When c theters re p rti lly occluded ( nd there is some flow through the c theter), irrig tion with thrombolytic gent or dilute (0.1 N) hydrochloric cid c n be successful in restoring c theter p tency (16,17). Irrig tion with c id is imed prim rily t improving the solubility of c lcium phosph te P.80 prec ipit tes, nd c ses of c theter occlusion refr ctory to thrombolytic gents h ve shown benefici l response to dilute cid (17). A protocol for chemic l dissol ution of c theter occlusions is shown in T ble 5.3. The protocol uses thrombol ytic gent initi lly, but the dilute hydrochloric cid c n be used initi lly if precipit tes re evident in the infusion system or if one of the precipit tion-p rone medic tions is being used. TABLE 5.3. A PROTOCOL FOR RESTORING PATENCY IN PARTIALLY OCCLUDED CATHETERS SUBCLAVIAN VEIN THROMBOSIS Clinic lly pp rent thrombosis of the subcl vi n vein occurs in 3% of p tients w ith subcl vi n vein c theters (18). The h llm rk of subcl vi n vein obstruction is unil ter l rm swelling on the side of the c theter insertion. The ev lu tion for suspected subcl vi n vein thrombosis should begin with http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 22)03-M y-05 17:33:40

form insoluble complexes more re dily

s the pH becomes more lk line).

Ovid: ICU Book noninv sive ppro ch using Doppler ultr sound ex min tion. This method h s h igh success r te for the di gnosis of occlusive thrombosis, nd it c n thus elim in te the need for more inv sive contr st venogr phy in m ny c ses. If subcl v i n vein thrombosis is confirmed, the c theter should be removed. Bec use thromb i from the subcl vi n vein c n extend into the superior ven c v , it seems wise to void ll centr l venous c nnul tion for few weeks, if possible. Hep rin ntico gul tion is often recommended in this setting (18). However, there is no e vidence th t ntico gul tion reduces the incidence of pulmon ry embolism or othe rwise lters the clinic l course in c theter-induced subcl vi n vein thrombosis (13). Symptom tic pulmon ry emboli re uncommon (10%) in subcl vi n vein thrombo sis, nd thus ntico gul tion should not be expected to provide much dded benef it in this situ tion.

VASCULAR EROSIONS C theter-induced perfor tions of the superior ven c v nd right trium re unc ommon, nd m ny of the c ses th t do occur re considered P.81 void ble (19,20) . Attention to proper positioning of centr l venous c theters w s mentioned t t he end of Ch pter 4, nd n ex mple of c theter position th t h s high risk for perfor tion of the superior ven c v w s provided in Figure 4.7. Left-sided c theter insertions re responsible for 70% of superior ven c v perfor tions (20), so voiding left-sided insertions is n import nt step for reducing the ri sk of ven c v perfor tion. Clinic l Fe tures Perfor tion of the superior ven c v c n occur t ny time fter insertion of c entr l venous c theters nd c n lso occur fter repl cement of c theters over guidewire (21). The most common time for perfor tion is in the first 7 d ys ft er c nnul tion, but perfor tions c n be del yed for s long s 2 months fter c nnul tion (19). The clinic l symptoms (substern l chest p in, cough, nd dyspne ) re nonspecific, nd suspicion of perfor tion is prompted by r diogr phic ch n ges such s the ones in Figure 5.1 (19). Note the medi stin l widening nd pleur l effusion, which re the r diogr phic h llm rks of P.82 ven c v perfor tion. In f ct, the sudden nd unexpected ppe r nce of pleur l effusion in p tien t with centr l venous c theter should r ise suspicion of v scul r perfor tion. The pleur l effusions c n be unil ter l (right- or left-sided) or bil ter l. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 22)03-M y-05 17:33:40

Ovid: ICU Book Figure 5.1. The r diogr phic fe tures of superior ven c v perfor tion by cen tr l venous c theter. Note the medi stin l widening nd the pleur l effusion. Ch est film courtesy of Dr. John E. Heffner (from reference 19). Di gnosis Thor centesis is v lu ble di gnostic procedure bec use the chemic l compositio n of the pleur l fluid should be simil r to th t of the intr venous fluid in c s es of v scul r perfor tion. (Pleur l fluid glucose levels c n be useful if the i ntr venous fluids cont in dextrose.) The di gnosis c n be confirmed by injecting r diocontr st dye through the dist l port of the c theter. The presence of dye in the medi stinum (on port ble chest roentgenogr m) confirms the perfor tion. Once the di gnosis is cert in, the perfor ting c theter should be removed immed i tely. Contr ry to wh t one might suspect, c theter remov l does not promote bl eeding in this setting (20). INFECTIOUS COMPLICATIONS Hospit l- cquired (nosocomi l) septicemi occurs 2 to 7 times more often in ICU p tients th n in other hospit lized p tients, nd indwelling v scul r c theters re the second le ding c use of nosocomi l septicemi in the ICU (pneumoni is t he first) (22). The ppe r nce of nosocomi l septicemi c n double the length of st y in the ICU, nd c n incre se the likelihood of f t l outcome by 25 to 35% (22). As these observ tions indic te, infections ssoci ted with v scul r http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 22)03-M y-05 17:33:40

Ovid: ICU Book

C theter Coloniz tion This condition occurs when microorg nism is isol ted from the intr v scul r se gment of the c theter (the c theter tip), but either the org nism is s prophytic or growth is considered too sp rse to c use infection. There is no septicemi nd no evidence of loc l or systemic infl mm tion. C theter-Rel ted Infection In this condition, p thogen is isol ted from the c theter tip, nd growth is c onsidered dense enough to c use infection. This condition is not ccomp nied by septicemi , but it c n be prelude to septicemi . P.83 Signs of loc l infl mm t ion (e.g., erythem or purulence t the insertion site) nd systemic infl mm tio n (e.g., fever or leukocytosis) m y be present or bsent. (This condition is not entirely s tisfying bec use it is possible to h ve c theter-rel ted infection without clinic l signs of infection or infl mm tion.)

PATHOGENESIS The host response to foreign body is imed t degr ding or enc psul ting the f oreign body (23). Bec use c theters re not biodegr d ble, the host response to these devices is to enc psul te them in fibrin she th. This she th is meshwo rk of fibrin str nds th t tr ps microorg nisms nd provides protective environ ment th t f vors microbi l prolifer tion. Routes of Infection http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 22)03-M y-05 17:33:40

C theter-Rel ted Septicemi In this condition, the s me p in the systemic circul tion, e enough to indic te th t the Sp rse growth of the org nism he source of septicemi , with

thogenic org nism is found on the c theter tip nd nd growth of the org nism on the c theter is dens c theter is the prim ry source of the septicemi . on the c theter would indic te dist nt site s t second ry seeding of the c theter tip.

DEFINITIONS The following definitions re

pplied to c theter- ssoci ted infections (3).

c theters h ve

signific nt imp ct on both morbidity nd mort lity in the ICU.

Ovid: ICU Book The common routes of c theter-rel ted septicemi re shown in Figure 5.2 (3,4). These routes re s follows, using the corresponding numbers in Figure 5.2: Figure 5.2. The routes involved in c theter-rel ted infections. q Microbes c n enter the intern l lumen of v scul r c theters through bre k points in the infusion system, such s stopcocks nd c theter hubs. Infections vi thi s route c n be limited by m int ining closed infusion system nd voiding unne cess ry bre ks in the system. P.84 q Microbes on the skin c n migr te long the subcut neous tr ct cre ted by indwell ing c theters. This is considered the predomin nt route of c theter-rel ted infe ctions, but the evidence for this is not convincing ( s discussed in the l st se ction of this ch pter). q Microorg nisms in circul ting blood c n become tr pped in the fibrin meshwork th t surrounds the intr v scul r segment of indwelling c theters. The fibrin she t h thus cts like filter for circul ting blood, like the in-line filters used f or blood tr nsfusions. This route is often ignored, but m y ssume n import nt role in critic lly ill p tients (see end of ch pter). Microbiology http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 22)03-M y-05 17:33:40

Ovid: ICU Book The org nisms involved in c theter-rel ted septicemi re identified by the foll owing list t ken from 13 prospective studies of c theter-rel ted septicemi (24) : St phylococcus epidermidis (27%), S. ureus (26%), C ndid species (17%), Kleb siell Enterob cter (11%), Serr ti (5%), Enterococcus (5%), Pseudomon s species ( 3%), nd others (8%). Thus, bout h lf of the infections involve st phylococci nd h lf involve fungi nd v rious enteric p thogens. This microbi l spectrum is import nt to consider when selecting empiric ntimicrobi l ther py.

CULTURE METHODS The di gnosis of c theter-rel ted septicemi is l bor tory-b sed di gnosis, n d the two culture methods th t provide the most v lu ble di gnostic inform tion re briefly described below (25,26 nd 27).

Qu ntit tive Blood Cultures A qu ntit tive blood culture is obt ined by inocul ting known volume of blood onto culture pl te nd counting the number of colonies th t ppe r fter incub tion. The results re reported s number P.85 of colony-forming units per milli liter (CFU/mL). The fe tures of the qu ntit tive blood culture technique re sho wn in T ble 5.4. Two blood specimens re required: One specimen is withdr wn thr ough the indwelling v scul r c theter, nd the other is obt ined from site dis t l to the c theter insertion site. The blood is not pl ced in the usu l (broth) blood culture bottles, but it must be pl ced in speci lly designed ev cu ted bo ttles th t cont in n ntico gul nt. These bottles re commerci lly v il ble s Isol ter bottles from Dupont Co. (Wilmington, DE), nd come in dult (10 mL) n d children's (1.5 mL) sizes. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 22)03-M y-05 17:33:40

CLINICAL FEATURES C theter-rel ted septicemi usu lly becomes m nifest p tient who h s h d n indwelling v scul r c theter There is often no evidence of loc l infection or infl sertion site on the skin. The presence of erythem t ghten the suspicion of c theter-rel ted septicemi , required to confirm the di gnosis.

s n unexpl ined fever in for longer th n 48 hours. mm tion t the c theter in the insertion site c n hei but the presence of pus is

Ovid: ICU Book TABLE 5.4. THE QUANTITATIVE BLOOD CULTURE METHOD The results of qu ntit tive culture re shown in Figure 5.3. In this c se, the blood s mple obt ined through the indwelling c theter yielded the gre test numb er of colonies on incub tion. This illustr tes how the qu ntit tive culture tech nique uses the differenti l growth in the two blood specimens to identify the so urce of the septicemi . http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 22)03-M y-05 17:33:40

Ovid: ICU Book Figure 5.3. Qu ntit tive cultures of blood spir ted through centr l venous c theter (C theter Blood) nd blood obt ined from dist nt site (Peripher l Blood ). The gre ter growth from the c theter blood specimen is evidence of c theter-r el ted septicemi . (From Curt s S, Tr mposch K. Culture methods to ev lu te cent r l venous c theter sepsis. Nutr Clin Pr ct 1991;6:4351). C theter Tip Cultures The st nd rd di gnostic ppro ch to c theter-rel ted septicemi involves removin g the index c theter nd culturing the intr v scul r segment of the c theter. Wh en c theter tips re pl ced in broth nd cultured qu lit tively (growth reported s present or bsent, s in st nd rd blood cultures), f lse-positive results r e common (25). Therefore, qu lit tive broth cultures re never recommended. The culture method used is semiqu ntit tive technique th t involves rolling the c theter tip cross the surf ce of culture pl te nd counting the number of colo nies (CFUs) on the pl te fter the incub tion period. The princip l fe tures of the semiqu ntit tive c theter culture technique re shown in T ble 5.5 (27). A s t nd rd blood culture must be obt ined from dist l site t the time of c thete r remov l. When the P.86 s me org nism is found in the blood nd on the c theter tip, 15 colonies from the c theter tip is the threshold th t identifies the sou rce of the septicemi . Note th t the sensitivity of this di gnostic method is qu ite low (36%), which me ns th t l rge number of c ses of c theter-rel ted P.87 septicemi will be missed. The sensitivity c n be improved by pl ting the c the ter tips t the bedside, immedi tely fter remov l (28). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 22)03-M y-05 17:33:40

Ovid: ICU Book TABLE 5.5. CATHETER TIP CULTURE: THE SEMIQUANTITATIVE METHOD WHICH METHOD IS PREFERRED? The roll pl te method for culturing c theter tips is the ccepted method for ev lu ting c theterrel ted infections. However, it suffers from sever l dis dv nt g es, including low sensitivity nd in bility to detect infections on the inner lumin l surf ce of c theters; most import nt, it requires remov l of the c thete rs nd insertion of repl cement c theters. More th n 50% of the c theters remove d for suspected c theter-rel ted infections show little growth on subsequent cul tures. This me ns th t c theter remov l nd repl cement is unnecess ry in the m jority of c ses of suspected c theter-rel ted septicemi . This problem c n be el imin ted with the qu ntit tive blood culture method, which does not require c th eter remov l. The qu ntit tive blood culture method is lso much more sensitive th n the c theter culture method. Thus, the qu ntit tive blood culture technique should be preferred to the semiqu ntit tive c theter tip culture. This l tter c ulture technique is ppropri te only when remov l of indwelling c theters is ind ic ted (i.e., when there is erythem or purulence t the c theter insertion site ); when it is used, the c theter tip should be pl ted t the bedside to improve the di gnostic yield. CATHETER GRAM STAIN When c theters re removed for semiqu ntit tive culture, Gr m st in of the c t heter tip c n help in the e rly decisions bout di gnosis nd m n gement (29). T he dist l few centimeters of the c theter c n be don ted to the st in procedure, without compromising the 5-cm segment submitted for culture. The c theter segme nt to be st ined is slit longitudin lly, exposing the inner lumin l surf ce of t he c theter. Holding the c theter with sterile forceps while pplying the st in llows both inner nd outer surf ces of the c theter to be st ined for inspectio n. The benefits of this ppro ch re obvious. The presence of org nisms on the c theter is presumptive evidence of infection, nd the morphology c n help to ide ntify the likely p thogens nd select the ppropri te ntimicrobi l ther py. Mor phology m y be p rticul rly useful for identifying C ndid infections (these org nisms re l rge, Gr m-positive round or ov l forms). Fin lly, identifying org n isms on the inside of the c theter is v lu ble for the interpret tion of the c t heter cultures P.88 bec use the roll pl te method cultures only the outer surf c e of the c theter. EMPIRIC ANTIMICROBIAL THERAPY The decision to institute empiric ntibiotic ther py is determined by the likeli hood of c theterhttp://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 22)03-M y-05 17:33:40

Ovid: ICU Book rel ted infection nd the clinic l condition of the p tient. The clinic l determ in nts of empiric ntimicrobi l ther py re shown in T ble 5.6. In the p tient w ith n unexpl ined fever nd nothing else, ntimicrobi l ther py is not necess r y if the c theter is removed (4). The rem ining conditions in T ble 5.6 w rr nt the empiric ntibiotics indic ted. Remember th t S. epidermidis is often methici llin-resist nt nd thus is lso resist nt to the ceph losporins. V ncomycin is t he drug of choice for this org nism nd should be included in ll empiric ntibi otic regimens (4). Adding n minoglycoside is recommended for p tients with pro sthetic v lves ( minoglycosides nd v ncomycin c n be synergistic for S. epiderm idis endoc rditis), nd ceft zidime is lso dded in neutropenic p tients (for e xtr ntipseudomon l ctivity). TABLE 5.6. EMPIRIC ANTIMICROBIAL THERAPY: INDICATIONS AND REGIMENS

Dissemin ted C ndidi sis Persistent fever despite c theter repl cement nd bro d-spectrum ntimicrobi l her py m y be sign of dissemin ted c ndidi sis. Predisposing P.89 conditions nclude immunosuppression due to chemother py, steroids, hum n immunodeficiency irus infection, bro d-spectrum ntimicrobi l ther py, prosthetic joints nd he t v lves, nd http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 22)03-M y-05 17:33:40

PERSISTENT SEPSIS Evidence of continuing infection or infl mm tion er w y c n sign l the following conditions.

fter ntibiotic ther py is und

t i v r

Ovid: ICU Book v scul r c theters (30). The di gnosis of dissemin ted c ndidi sis is often elus ive bec use blood cultures re neg tive in over 50% of c ses (30,31 nd 32). Oth er l bor tory methods for detecting systemic c ndidi sis include serum levels of enol se ( cytopl smic ntigen) nd l tex gglutin tion test for C ndid ce ll w ll ntigen (titers 1:4 or higher suggest dissemin ted dise se). However, th ese tests re not considered reli ble enough to secure the di gnosis without oth er evidence of dissemin ted c ndidi sis (32). Clinic l m rkers of dissemin ted c ndidi sis include c ndiduri in the bsence of indwelling urethr l c theters, nd endophth lmitis. Therefore, in c ses of suspected c ndidi sis, ex mine or cul ture the urine nd consult n ophth lmologist for funduscopic ex min tion of t he eye. He vy coloniz tion of the urine in high-risk p tients, even in the prese nce of n indwelling Foley c theter, c n be used s n indic tion to initi te em piric ntifung l ther py (31). Endophth lmitis c n occur in up to 20% of c ses o f dissemin ted c ndidi sis (31), nd evidence of ophth lmitis on the eye ex min tion secures the di gnosis of c ndidemi nd provides n indic tion to initi te ntifung l ther py. Empiric ntifung l ther py is usu lly reserved for p tients with persistent unexpl ined fever who re neutropenic or h ve prosthetic he rt v lve, or high-risk p tients who show evidence of progressive sepsis (i.e., org n dysfunction or hemodyn mic inst bility). The ntifung l gent of choice for e mpiric ther py is mphotericin B, 0.5 to 0.7 mg/kg d ily (31). The ddition of f lucon zole (200 to 400 mg/d y) h s no proven benefit in this setting (31).

Suppur tive Thrombophlebitis Thrombosis surrounding the c theter tip is common finding in p tients with c t heter-rel ted septicemi (33), but in most c ses the loc l thrombosis does not t r nsform into suppur tive thrombophlebitis. The l tter complic tion is suggest ed by persistently positive blood cultures despite c theter remov l nd ppropri te ntimicrobi l ther py. (C ndid suppur tive thrombophlebitis m y be ssoci t ed with neg tive blood cultures.) Suppur tive phlebitis in peripher l veins is o ften ccomp nied by loc l infl mm tion nd purulent dr in ge from the c theter i nsertion site on the skin (34). Ther py for this complic tion is surgic l excisi on if peripher l vessel is involved. In the l rge centr l veins, ntimicrobi l ther py combined with hep rin ntico gul tion c n produce s tisf ctory results in 50% of c ses (35). ANOTHER PERSPECTIVE The ccepted pr ctices for the prevention, recognition, nd m n gement of c thet er-rel ted infections re firmly rooted in the notion th t P.90 microorg nisms o n the skin re the culprits in most c ses of c theter-rel ted infections. This n otion http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 22)03-M y-05 17:33:40

Ovid: ICU Book is r rely disputed, yet there is no convincing evidence to support its cl ims. T he rem inder of the ch pter ttempts to show th t there is more likely source of septicemi th n the skin in critic lly ill p tients. Why is this import nt? B ec use we spend lot of time scrubbing nd cle ning the skin, nd we m y be cle ning the wrong surf ce.

A CASE AGAINST THE SKIN The observ tions listed below do not support the notion th t the skin is the pri ncip l source of c theter-rel ted infections. There is poor correl tion betwee n org nisms isol ted from the skin round c theter insertion sites nd org nisms isol ted from the bloodstre m in c ses of c theter-rel ted septicemi (25,36). This is why surveill nce cultures of the skin re not recommended s predictiv e tool for identifying future inh bit nts of the bloodstre m. Pr ctices imed t reducing the migr tion of skin microbes long the c theter tr ct (e.g., ntimic robi l ointments, subcut neous tunnels, nd ntimicrobi l cuffs tt ched to the c theters) h ve h d little or no imp ct on the incidence of c theter-rel ted sep ticemi s (4,37,38 nd 39). H lf of the c ses of c theter-rel ted septicemi invo lve org nisms th t re not commonly found on the skin (i.e., enteric p thogens nd C ndid spp.) (24). The prev lence of st phylococc l isol tes c nnot be used s evidence th t the infections origin te on the skin bec use st phylococci re lso prominent bowel org nisms, p rticul rly in critic lly ill p tients nd p ti ents receiving ntibiotics (40). In f ct, S. epidermidis is one of the most comm on isol tes in the upper g strointestin l tr ct of p tients with multiorg n f il ure (41), so the species term for this org nism (indic ting n epiderm l h bit t ) is misle ding. A Long D y's Journey Another strike g inst the skin is the long dist nces th t must be tr veled in o rder for skin microbes to re ch the intr v scul r segment of indwelling c theter s. St phylococci re nonmotile org nisms nd re only 0.001 mm in di meter. Thus , to re ch the tip of 20-cm centr l venous c theter, these org nisms must tr v el dist nce th t is 2 million times their own length. This is equiv lent to 6-foot hum n being who must tr vel 2272 miles (roughly the dist nce between Atl nt nd Los Angeles) without the id of tr nsport tion or even legs (bec use the st phylococcus is nonmotile). A CASE FOR THE BOWEL The body surf ce with the highest density of p thogens is the bowel mucos , not the skin (remember th t the bowel mucos is outside your P.91 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 22)03-M y-05 17:33:40

Ovid: ICU Book body), nd enteric microbes must tr vel only few millimeters to re ch the bloo dstre m. The f ct th t enteric org nisms re commonly involved in c theter-rel t ed septicemi s suggests th t the bowel route is ctive. The c se report in T ble 5.7 shows how the bowel c n ct s n occult source of c theter-rel ted septice mi (42). The cultures listed were obt ined from p tient in the surgic l ICU w ho developed nosocomi l fever on the seventh d y following n esoph gog strect omy. The s me org nism (Enterob cter clo c e) w s isol ted from multiple blood c ultures nd from semiqu ntit tive centr l venous c theter culture. The colony count from the c theter (more th n 100 CFUs) indic tes di gnosis of c theter-r el ted septicemi . However, cultures were lso obt ined from the skin surroundin g the c theter insertion site nd from n sog stric spir tes, nd these culture results suggest di gnosis other th n c theter-rel ted septicemi . Th t is, the n sog stric spir tes grew the s me type of org nism th t w s isol ted from the blood nd the c theter tip, nd this suggests th t the septicemi origin ted fr om the bowel, with second ry seeding of the indwelling c theter. This illustr te s how the di gnosis of c theter-rel ted septicemi c n be erroneous if the bowel is excluded in the di gnostic ev lu tion. TABLE 5.7. CATHETER-RELATED SEPTICEMIA FROM THE BOWEL The role of the bowel s n occult source of c theter-rel ted sepsis is uncle r t present. However, critic lly ill p tients re p rticul rly prone to developin g septicemi of bowel origin, nd the conditions responsible for this predisposi tion re described in Ch pter 6. REFERENCES REVIEWS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (17 of 22)03-M y-05 17:33:40

Ovid: ICU Book 1. Intr venous Nurses Society. Intr venous Nursing St nd rds of Pr ctice. Belmon t, MA: Intr venous Nurses Society, 1990. 2. Perucc R. Intr venous monitoring nd c theter c re. In: Terry J, B r nowski L, Lonsw y RA, Hedrick C, eds. Intr venous ther py. Phil delphi : WB S unders, 1 995;392399.

4. Bjornson HS. P thogenesis, prevention, nd m n gement of c theter- ssoci ted infections. New Horiz 1993;1:271278. PROTECTIVE DRESSINGS 5. Hoffm n KK, Weber DJ, S ms GP, et l. Tr nsp rent polyureth ne film s intr venous c theter dressing. A met - n lysis of infection risks. JAMA 1992;267:20722 076.

7. M rsh ll DA, Mertz PA, E glestein WH. Occlusive dressings. Arch Surg 1990;125 :1136 1139. REPLACING CATHETERS 8. Ullm n RF, Guerivich I, Schoch PE, Cunh BA. Coloniz tion nd b cteremi rel ted to dur tion of triple-lumen intr v scul r c theter pl cement. Am J Infect Co ntrol 1990;18:201 207. 9. Eyer S, Brummitt C, Crossley K, et l. C theter-rel ted sepsis: prospective, r ndomized study of three methods of long-term c theter m inten nce. Crit C re M ed 1990;18:1073 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (18 of 22)03-M y-05 17:33:40

6. M ki DG, Stolz SS, Wheeler S, Mermi LA. A prospective, r ndomized tri l of g uze nd two polyureth ne dressings for site c re of pulmon ry rtery c theters: implic tions for c theter m n gement. Crit C re Med 1994;22:17291737.

3. Norwood S, Ruby A, Civett J, Cortes V. C theter-rel ted infections nd i ted septicemi . Chest 1991;99:968975. P.92

ssoc

Ovid: ICU Book 1079. 10. Cobb DK, High KP, S wyer RP, et l. A controlled tri l of scheduled repl cem ent of centr l venous nd pulmon ry rtery c theters. N Engl J Med 1992;327:10621 068. CATHETER FLUSHES 11. Peterson FY, Kirchhoff KT. An lysis of rese rch bout hep rinized versus non hep rinized intr v scul r lines. He rt Lung 1991;20:631642. 12. Americ n Associ tion of Critic l C re Nurses. Ev lu tion of the effects of h ep rinized nd nonhep rinized flush solutions on the p tency of rteri l pressur e monitoring lines: the AACN Thunder Project. Am J Crit C re 1993;2:315. 13. W lsh DA, Mellor JA. Why flush peripher l intr venous c nnul e used for inte rmittent intr venous injection? Br J Clin Pr ct 1991;45:3132.

MECHANICAL COMPLICATIONS 15. Trissel LA. Drug st bility nd comp tibility issues in drug delivery. C ncer Bull 1990;42:393398. 16. Monturo CA, Dickerson RN, Mullen J. Effic cy of thrombolytic ther py for occ lusion of long-term c theters. J P rent Ent Nutr 1990;14:312314. 17. Shulm n RJ, Reed T, Pitre D, L ine L. Use of hydrochloric cid to cle r obst ructed centr l venous c theters. J P rent Ent Nutr 1988;12:509510.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (19 of 22)03-M y-05 17:33:40

18. Heffner JE. A 20-ye r-old wom n with respir tory f ilure nd rm. J Crit Illness 1994;9:187192.

swollen right

14. Br nson PK, McCoy RA, Phillips BA, Clifton GD. Effic cy of 1.4% sodium citr te in m int ining rteri l c theter p tency in p tients in medic l ICU. Chest 1993;103:882885.

Ovid: ICU Book

21. Armstrong CW, M yh ll CG. Contr l ter l hydrothor x following subcl vi n c t heter repl cement using guidewire. Chest 1983;84:231233. P.93 INFECTIOUS COMPLICATIONS 22. Pittet D, T r r D, Wenzel RP. Nosocomi l bloodstre m infection in critic ll y ill p tients. JAMA 1994;271:15981601. 23. V ud ux PE, Lew DP, W ldvogel FA. Host f ctors predisposing to foreign body infections. In: Bisno AL, W ldvogel FA, eds. Infections ssoci ted with indwelli ng medic l devices. W shington, DC: Americ n Society for Microbiology, 1989;326. 24. H mpton AA, Sheretz RJ. V scul r- ccess infections in hospit lized p tients. Surg Clin North Am 1988;68:5771. 25. Curt s S, Tr mposch K. Culture methods to ev lu te centr l venous c theter s epsis. Nutr Clin Pr ct 1991;6:4348. 26. Benezr D, Kiehn TE, Gold JWM, et l. Prospective study of infections in ind welling centr l venous c theters using qu ntit tive blood cultures. Am J Med 198 8;85:495498. 27. M ki DG, Weise CE, S r fin HW. A semiqu ntit tive culture method for identif ying intr venous c theter-rel ted infections. N Engl J Med 1977;296:13051309. 28. Hn tiuk OW, Pike J, Stolzfus D, L ne W. V lue of bedside pl ting of semiqu n tit tive cultures for di gnosis of c theter-rel ted infections in ICU p tients. Chest 1993;103:896899. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (20 of 22)03-M y-05 17:33:40

20. Heffner JE. A 49-ye r-old m n with t chypne effusion. J Crit Illness 1994;9:101109.

nd

r pidly enl rging pleur l

19. Duntley P, Siever J, Korwes ML, et theters. Chest 1992;101:16331638.

l. V scul r erosion by centr l venous c

Ovid: ICU Book 29. Cooper GL, Hopkins CC. R pid di gnosis of intr v scul r c theter- ssoci ted infection by direct gr m st ining of c theter segments. N Engl J Med 1985;312:11 421145. 30. M gnussen CR. Dissemin ted C ndid infection: di gnostic clues, ther peutic options. J Crit Illness 1992;7:513525. 31. British Society for Antimicrobi l Chemother py Working P rty. M n gement of deep C ndid infection in surgic l nd intensive c re unit p tients. Intensive C re Med 1994;20:522528.

33. R d II, Lun M, Kh li S-A, et l. The rel tionship between the thrombotic nd infectious complic tions of centr l venous c theters. JAMA 1994;271:10141016. 34. G rrison RN, Rich rdson JD, Frye DE. C theter- ssoci ted septic thrombophleb itis. South Med J 1982;75:917919. 35. Verghese A, Widrich WC, Arbeit RD. Centr l venous septic thrombophlebitis: t he role of ntimicrobi l ther py. Medicine 1985;64:394400. 36. Golledge C, McPherson M. Skin entry site sw bbing poor predictor of c thet er-rel ted sepsis. Infect Control Hosp Epidemiol 1988;9:5462. 37. Sitges-Serr A, Lin res J. Tunnels do not protect g inst venous c theter-re l ted sepsis. L ncet 1984;1:459460. 38. Norwood S, H jj r G, Jenkins L. The influence of n tt ch ble subcut neous cuff for preventing triple lumen c theter infections in critic lly ill surgic l nd tr um p tients. Surg Gynecol Obstet 1992;175:3340. 39. Groeger JS, Luc s AB, Coit D, et l. A prospective, r ndomized ev lu tion of the effect of silver-impregn ted subcut neous cuffs for preventing tunneled chr onic venous ccess http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (21 of 22)03-M y-05 17:33:40

32. Rosemurgy AS, Sweeney JF, Albrink MH, et l. Implic tions of c ndid tests in injured dults. Contemp Surg 1993;42:327332.

ntigen

Ovid: ICU Book c theter infections in c ncer p tients. Ann Surg 1993;218:206210. 40. Altemeier WA, Hummel RP, Hill EO. St phylococc l enterocolitis following nt ibiotic ther py. Ann Surg 1963;157:847858. 41. M rsh ll JC, Christou NV, Horn R, Me kins JL. The microbiology of multiple o rg n f ilure. Arch Surg 1988;123:309315.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (22 of 22)03-M y-05 17:33:40

42. Sing R, M rino PL. B cteri l tr nsloc tion: n occult c use of c theter-rel ted sepsis. Infect Med 1993;10:5457.

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 6 GASTROINTESTINAL PROPHYLAXIS We re told the most f nt stic biologic l t les. For ex mple, th t it is d ngero us to h ve cid in your stom ch. --JBS H ld ne (1939) We re preoccupied with th e thre t of microbi l inv sion through the skin nd tend to neglect simil r th re t of inv sion through our intern l skin in the g strointestin l (GI) tr ct. The GI tr ct is outside the body (like the hole in donut), nd thus the mucos l s urf ce of the GI tr ct must serve s b rrier to microbi l inv sion, much like the skin. Two observ tions suggest th t the re l thre t of microbi l inv sion is from the GI tr ct, not the skin. First, the skin provides sever l l yers of pro tection, where s the GI mucos is single l yer of cells only 0.01 mm thick. Se cond, the microbi l popul tion is f r gre ter in the liment ry tr ct th n on th e skin. In f ct, the number of b cteri in 1 gr m of stool (10 to 100 billion) i s gre ter th n the number of people on E rth (5.7 billion in 1995). This ch pter focuses on conditions th t predispose critic lly ill p tients to microbi l inv sion from the bowel, nd describes some me sures th t c n be used to counter ct this thre t. THE THREAT OF TRANSLOCATION There re 400 to 500 different species of b cteri nd fungi in the hum n limen t ry tr ct (1,2); the distribution of this indigenous microflor is shown in T b le 6.1. The org nisms in the mouth re const ntly being sw llowed in the s liv , but microbi l density in the stom ch is dr m tic lly reduced by the b ctericid l ctions of g stric cid. The density of microbes then rises progressively thro ugh the sm ll bowel nd colon, nd t the dist l end of the GI tr ct, microbes m ke up s much s 40% of the tot l fec l m ss (2). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 16)03-M y-05 17:34:12

Ovid: ICU Book TABLE 6.1. MICROBIAL DENSITY IN THE ALIMENTARY TRACT P.95 PROTECTIVE MECHANISMS Sever l protective mech nisms prevent the microbes on our inner surf ce from inv ding us. First, m ny of the indigenous b cteri nd fungi in the liment ry tr ct re s prophytes nd show little tendency for inv sive infection. Second, s lre dy mentioned, the b ctericid l ction of g stric cid helps to curb the popu l tion of immigr nt microbes from the or l c vity. A third f ctor is the intrins ic b rrier function of the mucos l lining of the bowel. Fin lly, the reticuloend otheli l system in the bowel (i.e., the lymph tic system nd ph gocytic cells) c n tr p nd destroy org nisms th t bre ch the mucos l b rrier. Roughly two-third s of the reticuloendotheli l system in the body is loc ted in the bdomen (3), w hich suggests th t microbi l inv sion cross the bowel mucos m y be common oc currence (3). PREDISPOSING FACTORS A defect in ny of the protective mech nisms described bove f vors the movement , or tr nsloc tion, of enteric p thogens into the systemic circul tion (4). Figu re 6.1 shows three conditions th t f vor bloodstre m inv sion by bowel microbes: microbi l overgrowth in the bowel lumen, disruption of the surf ce mucos , nd defective cle r nce by submucos l lymph tics. The clinic l conditions described in this ch pter re ch r cterized by one or more of these predisposing condition s, nd the go l of g strointestin l prophyl xis is to prevent these conditions f rom developing. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 16)03-M y-05 17:34:12

Ovid: ICU Book Figure 6.1. The triple thre t of tr nsloc tion. This di gr m of n intestin l mi crovillus shows three conditions th t predispose to bloodstre m inv sion by ente ric p thogens. STRESS ULCERS Stress ulcers re superfici l erosions in the g stric mucos th t re commonpl c e in p tients with cute, life-thre tening dise ses (5,6). These erosions re us u lly confined to the surf ce mucos nd re distinct from the lesions of peptic ulcer dise se, which re deeper cr ters th t c n erode through the entire width of the bowel w ll. Thus, the term stress ulcer is somewh t misle ding. P.96 PATHOGENESIS The mucos l lining of the GI tr ct is norm lly shed nd repl ced every 2 to 3 d ys. When nutrient blood flow is in dequ te to support the repl cement process, t he surf ce of the bowel becomes denuded, cre ting the superfici l erosions known s stress ulcers. Although g stric cid c n ggr v te the condition, the prim r y c use of stress ulcer tion is imp ired blood http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 16)03-M y-05 17:34:12

Ovid: ICU Book flow, not g stric cidity (5). CLINICAL FEATURES G stric erosions re present in 10 to 25% of p tients on dmission to medic l nd surgic l intensive c re unit, nd in 90% of p tients by the third d y of the ICU st y (6,7). The risk of mucos l ulcer tion seems to be p rticul rly high in p tients with he d injury nd p tients with therm l injury th t involves t le st 30% of the body surf ce re (5,6). Although g stric erosions re commonpl ce in critic lly ill p tients, they re often clinic lly silent. The two consequen ces of stress ulcer tion re disruption of the mucos l b rrier (with subsequent risk of P.97 tr nsloc tion nd nosocomi l septicemi ) nd hemorrh ge from erosio n through surf ce vessels. Perfor tion is r re with stress ulcer tion bec use of the superfici l loc tion of the lesions. The incidence of nosocomi l septicemi from stress ulcer tion h s not been studied. The incidence of bleeding from str ess ulcers is lmost 100% for clinic lly silent (occult) bleeding (8) but is onl y bout 5% for clinic lly pp rent (overt) hemorrh ge, nd 1 to 2% for clinic ll y import nt hemorrh ge (i.e., hemorrh ge requiring blood tr nsfusions) (8,9 nd 10).

HIGH-RISK CONDITIONS The low incidence of overt or clinic lly import nt bleeding from stress ulcer ti on r ises some question of whether preventive ther py for stress ulcer tion is j ustified in ll p tients. (The risk of tr nsloc tion from stress erosions h s no t been considered in the deb te over the v lue of stress ulcer prophyl xis). In gener l, the following conditions re considered high risk for stress ulcer-rel ted bleeding, nd thus re indic tions for prophyl ctic ther py (9). INDICATIONS FOR PROPHYLAXIS He d injury Therm l injury involving t le st 30% of the body surf ce re Emerg ent or m jor surgery Severe or multisystem tr um Shock or multiorg n f ilure Co gulop thy http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 16)03-M y-05 17:34:12

Ovid: ICU Book Mech nic l ventil tion for more th n 48 hours Ongoing ther py with ulcerogenic d rugs History of ulcer-rel ted bleeding PREVENTIVE STRATEGIES The following me sures re considered effective for reducing the incidence of ov ert bleeding from stress ulcers (10). (Bec use stress ulcers re usu lly sympto m tic, prophyl ctic effic cy is usu lly ev lu ted by the incidence of hemorrh ge from the lesions.) The preventive me sures below re listed in order of prefere nce. Preserving Spl nchnic Blood Flow Bec use imp ired nutrient blood flow is the inciting event in stress ulcer tion, preserving dequ te mesenteric blood flow should be the optim l prophyl xis for stress ulcers. Unfortun tely, few methods re v il ble for monitoring mesenter ic blood flow in the clinic l setting. G stric tonometry (this technique is desc ribed in Ch pter 13) h s been proposed s method of monitoring the dequ cy of g stric mucos l P.98 blood flow in critic lly ill p tients, but this technique h s limit tions nd h s yet to g in widespre d ccept nce. The best str tegy t present is to m int in dequ te levels of systemic blood flow nd oxygen tr nspo rt, using convention l hemodyn mic monitoring with pulmon ry rtery c theters, i f possible. Enter l Nutrition Enter l tube feedings h ve du l benefit: They reduce the risk of stress ulcer bleeding while providing d ily nutrient requirements (11). The mech nism for the protective effect of tube feedings m y be cid neutr liz tion, or m y result fr om the trophic effect of lumin l nutrients on the function l integrity of the bo wel mucos . (The delivery of enter l tube feedings is described in Ch pter 47.) If full enter l nutrition is not possible, there re two ph rm cologic ppro che s to stress ulcer prophyl xis. One ppro ch uses cytoprotective gent th t m i nt ins the function l integrity of the g stric mucos ; the other ppro ch is b s ed on suppression of g stric cid. The ph rm cologic ppro ch to stress ulcer pr ophyl xis is outlined in T ble 6.2. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 16)03-M y-05 17:34:12

Ovid: ICU Book TABLE 6.2. PHARMACOLOGIC APPROACH TO STRESS ULCER PROPHYLAXIS P.99 Sucr lf te Sucr lf te is n luminum s lt of sucrose sulf te th t helps m int in the struct ur l nd function l integrity of the g stric mucos (12). The drug is dminister ed or lly or vi n sog stric tube t the dos ge shown in T ble 6.2. This gent h s proven effective in reducing the incidence of stress ulcers (7) nd reducing the incidence of hemorrh ge from stress ulcers (8,10,12). The mech nism of sucr lf te's protective effect is not known (12). The bulk of evidence indic tes th t sucr lf te should be the preferred method of prophyl xis when full enter l nutr ition is not fe sible. The drug is s fe, inexpensive, nd

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 16)03-M y-05 17:34:12

Ovid: ICU Book e sy to dminister (the only requirement is th t the p tient toler te g stric in st ll tion of 5 to 10 mL of fluid every 4 to 6 hours). The dv nt ges of sucr lf te over other ph rm cologic gents re discussed l ter. Inter ctions. Sucr lf te is c p ble of binding the medic tions listed in T ble 6.3 (13). This c n le d to reduced bsorption nd diminished bio v il bility of the specified d rugs. To minimize potenti l inter ctions, sucr lf te should not be dministered t the s me time s these medic tions (this does not pply to p renter l drug d ministr tion). Bec use sucr lf te cont ins luminum, it c n lso bind nd reduce phosphorous bsorption from the bowel. Although sucr lf te ther py h s report ed ssoci tion with hypophosph temi (14), c use- nd-effect P.100 rel tionship h s not been est blished. Nevertheless, it is wise to consider sucr lf te s possible culprit in ny p tient with unexpl ined hypophosph temi . The luminum content of sucr lf te does not promote elev tions in pl sm luminum levels (15) . http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 16)03-M y-05 17:34:12

Ovid: ICU Book TABLE 6.3. DRUG INTERACTIONS Hist mine H2 Receptor Ant gonists One of the most popul r methods of stress ulcer prophyl xis is to inhibit g stri c cid production with hist mine H2-receptor nt gonists. All hist mine-blocking drugs re equ lly effective in reducing the incidence of stress ulcer bleeding (10). The drugs re usu lly given intr venously ( lthough or l ther py is lso e ffective), nd continuous infusion ther py is more effective in m int ining the t rget pH ( bove 4) th n bolus drug dministr tion (16,17). Dosing regimens for continuous infusion of cimetidine nd r nitidine re shown in T ble 6.2. The dos ge of cimetidine is ltered in the presence of ren l insufficiency (see Ch pter 54 for dos ge djustments); continuous-infusion r nitidine h s not been studied in ren l insufficiency (13). Inter ctions. Cimetidine interferes with the hep tic met bolism of sever l drugs used in the c re of critic lly ill p tients, nd T ble 6.3 cont ins list of the pertinent gents (13). R nitidine h s simil r but less potent ction (13), so r nitidine h s g ined f vor over cimetidine. The potenti l for drug inter ctions must be co nsidered when cimetidine is given concurrently with ny of the drugs in T ble 6. 3. the le st desir ble method of the regimen in T ble 6.2 is t te or hist mine H2 blockers o drug inter ctions (see T ble for stress ulcer prophyl xis.

SUCRALFATE VERSUS H2 BLOCKERS Sucr lf te prophyl xis offers number of dv nt ges over the hist mine H2 block ers. The first is cost. The routine use of sucr lf te inste d of H2 blockers c r ries n estim ted cost s vings of $30,000 per hospit l bed per ye r (8), which t r nsl tes to n estim ted nnu l s vings of $4 billion doll rs ye rly in the Uni ted St tes (8). (This mount is one-third of the tot l NIH budget for 1995!) The second most import nt dv nt ge is the lower risk of g stric coloniz tion with sucr lf te. Suppression of g stric cidity with hist mine H2 blockers or nt cid s results in the overgrowth of b cteri nd fungi in the stom ch, nd this colon iz tion of the stom ch c n be http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 16)03-M y-05 17:34:12

Ant cids Neutr liz tion of g stric cidity with nt cids is stress ulcer prophyl xis. Ant cid titr tion using ime-consuming nd offers no dv nt ge over sucr lf ther th n reduced cost. The potenti l for multiple 6.3) should further discour ge the use of nt cids

Ovid: ICU Book prelude to nosocomi l pneumoni (10,18) P.101 nd tr nsloc tion with nosocomi l septicemi (19). Considering th t stress ulcer tion is itself risk f ctor for tr nsloc tion ( s shown in Figure 6.1), interventions th t promote b cteri l ove rgrowth in this setting dd to the risk of tr nsloc tion, nd thus do not seem wise choice. THE ACID PHOBIA As indic ted in the introductory rem rk by JBS H ld ne ( geneticist, not the ph ysiologist who described the H ld ne effect), our fe r of g stric cid is b sed on legend more th n scientific observ tion. Consider why we secrete g stric cid . It is not n import nt digestive id, bec use p tients with chlorhydri re n ot troubled with m l bsorption (20). The function of g stric cid is reve led in the work of Sir Joseph Lister, who popul rized the ntiseptic ppro ch to surge ry by cle nsing surgic l wounds with c rbolic cid ( lso known s phenol). The b ctericid l ctions of n cid pH re demonstr ted in Figure 6.2. In this c se, the common enteric org nism Escherichi coli is completely er dic ted in 1 hour when pH of the growth medium is reduced to 3; the s me org nism flourishes t slightly higher pH of 5. Note th t the threshold pH t which b cteri l growth oc curs (pH P.102 gre ter th n 4) is the s me s the t rget pH for stress ulcer pro phyl xis with hist mine H2 blockers (see T ble 6.2).

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 16)03-M y-05 17:34:12

Ovid: ICU Book

Thus, g stric cid functions s n ntib cteri l defense mech nism th t prevents immigr nt microbes from g ining ccess to the liment ry tr ct. This function pplies not only to the b cteri th t re sw llowed in the s liv , but lso to er dic ting microbes th t popul te the food we ingest (this f ct expl ins why food ingestion stimul tes g stric cid secretion). Just s cid is used to er dic te S lmonell org nisms in chickens, our own g stric cid serves simil r functio n of er dic ting S lmonell org nisms th t we ingest. This expl ins why g stric cid suppression ther py is ssoci ted with recurrent S lmonell enteritis (21) nd why chlorhydri is ssoci ted with n incre sed risk of enteric infections (20,21 nd 22). In this sense, g stric cid is n intrinsic mech nism for disinf ecting the food we e t. The fe r of g stric cid is b sed l rgely on the presume d ssoci tion between g stric cid nd peptic ulcer dise se ( nd on the m rketin g skills of cert in drug m nuf cturers). However, recent evidence indic tes th t m ny g stric nd duoden l ulcers re the result of loc l infection with Helic ob cter pylori (23). As physici ns become more w re th t g stric cid is not th e culprit in g stric mucos l injury, the unfounded nd d ngerous pr ctice of g s tric cid suppression in critic lly ill p tients m y fin lly be b ndoned.

OCCULT BLOOD TESTING Monitoring for occult blood in g stric secretions is common pr ctice in the pr ophyl xis for http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 16)03-M y-05 17:34:12

Figure 6.2. The influence of pH on the growth of Escherichi J et l. Gut 1972;13:251.)

coli. (From Gi nell

Ovid: ICU Book stress ulcer tion, but it h s limited v lue. N sog stric spir tes lmost lw ys cont in occult blood in the presence of stress ulcer tion (8), nd bec use few c ses of stress ulcer tion re ssoci ted with gross (overt) hemorrh ge, the pre sence of occult blood in n sog stric spir te h s no predictive v lue for sse ssing the risk of signific nt bleeding. Bec use overt hemorrh ge lmost never p pe rs when there is no occult blood in g stric secretions (24), neg tive occul t blood test h s predictive v lue. However, this is unlikely to lter decisions reg rding stress ulcer prophyl xis. When ev lu ting g stric spir tes, the gu i c nd Hemoccult tests re in ppropri te bec use they give f lse-positive nd f l se-neg tive results when the test solution h s pH less th n 4 (25). The G stro ccult test (Smith, Kline L bor tories) is not influenced by pH (25), nd thus is the ppropri te test for occult blood in g stric spir tes. SELECTIVE DIGESTIVE DECONTAMINATION The microbes th t norm lly inh bit the liment ry tr ct seem content with their existence nd show little tendency to inv de us. However, in the presence of dis e se, these microbes re repl ced by more ggressive species, nd this coloniz t ion serves s prelude to nosocomi l P.103 infections (such s pneumoni s, urin ry tr ct infections, nd septicemi ). The im of selective digestive decont min tion (SDD) is to prevent coloniz tion of the liment ry tr ct by inv sive p tho gens, nd thus prevent nosocomi l infections. To ccomplish this purpose, ntibi otics designed to er dic te fungi nd Gr m-neg tive erobic p thogens re pl ced in the or l c vity nd GI tr ct on regul r b sis. The ntibiotics re non bso rb ble, nd thus pose no thre t of systemic toxicity. The norm l, indigenous mic roflor (mostly n erobes) re undisturbed, to limit coloniz tion by resist nt o rg nisms. THE METHOD Sever l ntibiotic regimens h ve been used for SDD, nd one of the effective one s is s follows (26): Or l c vity: A p ste cont ining 2% polymyxin, 2% tobr mycin, nd 2% mphotericin is pplied t o the inside of the mouth with gloved finger, nd the pplic tion is repe ted every 6 hours (the ph rm cy will m ke the p ste). GI tr ct: http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 16)03-M y-05 17:34:12

Ovid: ICU Book A 10-mL solution cont ining 100 mg polymyxin E, 80 mg tobr mycin, nd 500 mg mp hotericin is dministered vi n sog stric tube every 6 hours. This regimen er dic tes most Gr m-neg tive p thogens from the mouth nd GI tr ct in just 1 week. It is then continued for s long s the p tient resides in the ICU. After 10 ye rs of clinic l experience with SDD, there h s been no problem with the emergenc e of ntibiotic resist nce during the course of ther py. RESULTS A recent review of 25 clinic l tri ls showed th t SDD w s effective in reducing the incidence of nosocomi l infections in p tients in the ICU (27). The results of one study re shown in Figure 6.3. In this study, SDD w s ssoci ted with sig nific nt reductions in the incidence of pneumoni , urin ry tr ct infections, nd c theter-rel ted septicemi . (Note th t SDD virtu lly elimin ted c theter-rel t ed septicemi , which supports the notion th t the bowel is n import nt source o f c theter-rel ted septicemi s.) Despite impressive results such s those in Fig ure 6.3, SDD h s not been embr ced with enthusi sm in the United St tes. Much of the hesit ncy is due to the f ct th t SDD h s not been effective in reducing mo rt lity in m ny (but not ll) of the clinic l studies (27). However, the im of SDD is to reduce nosocomi l infections, nd in this reg rd, SDD must be consider ed n effective method of infection control in the ICU. The l ck of ssoci tion between nosocomi l infections nd mort lity r te in the ICU is sep r te issue ( nd one th t dds to the long list of uncert inties in clinic l medicine). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 16)03-M y-05 17:34:12

Ovid: ICU Book Figure 6.3. The effects of selective digestive decont min tion (SDD) on the inci dence of nosocomi l infections in gener l medic l-surgic l ICU p tients. (From U lrich C et l. Intensive C re Med 1989;15:424.) P.104

TRANSLOCATION 1. Simon GL, Gorb ch SL. Intestin l microflor . Med Clin North Am 1982;66:557574. 2. Borriello SP. Microbi l flor of the g strointestin l tr ct. In: Microbi l me t bolism in the digestive tr ct. Boc R ton, FL: CRC Press, 1989;219. 3. L ngk mp-Henken B, Glezer JA, Kudsk KA. Immunologic structure nd function of the g strointestin l tr ct. Nutr Clin Pr ct 1992;7:100108.

STRESS ULCERS 5. Schiessel R, Feil W, Wenzel E. Mech nisms of stress ulcer tion nd implic tio ns for http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 16)03-M y-05 17:34:12

4. Alex nder JW, Boyce ST, B bcock GF, et l. The process of microbi l tr nsloc tion. Ann Surg 1990;212:496510.

M rston A, Bulkley GR, Fiddi n-Green RG, H glund UH, eds. Spl nchnic ischemi d multiple org n f ilure. St. Louis: CV Mosby, 1989.

REFERENCES GENERAL TEXTS Holl nder D, T rnowski AS. G stric cytoprotection: : Plenum, 1989.

clinici n's guide. New York n

Ovid: ICU Book tre tment. G stroenterol Clin North Am 1990;19:101120. P.105 6. Reusser P, Gyr K, Scheidegger D, et l. Prospective endoscopic study of stres s erosions nd ulcers in critic lly ill neurosurgic l p tients. Current incidenc e nd effect of cid-reducing prophyl xis. Crit C re Med 1990;18:270274. 7. Eddleston JM, Pe rson RC, Holl nd J, et l. Prospective endoscopic study of s tress erosions nd ulcers in critic lly ill dult p tients tre ted with either s ucr lf te or pl cebo. Crit C re Med 1994;22:19491954. 8. M ier RV, Mitchell D, Gentiello L. Optim l ther py for stress g stritis. Ann Surg 1994;220:353363. 9. Cook DJ, Fuller MB, Guy tt GH. Risk f ctors for g strointestin l bleeding in critic lly ill p tients. N Engl J Med 1994;330:377381. 10. Cook DJ, Reeve BK, Guy tt GH. Stress ulcer prophyl xis in critic lly ill p t ients. JAMA 1996;275:308314. 11. Pingleton SR. G stric bleeding nd/or enter l feeding. Chest 1986;90:23. 12. McC rthy DM. Sucr lf te. N Engl J Med 1990;325:10161025. 13. McEvoy GK, ed. AHFS Drug Inform tion, 1995. Bethesd , MD: Americ n Society o f He lth System Ph rm cists, 1995:20212065. 14. Miller SJ, Simpson J. Medic tionnutrient inter ctions: hypophosph temi ssoc i ted with sucr lf te in the intensive c re unit. Nutr Clin Pr ct 1991;6:199201. 15. Tryb M, Kurz-Muller K, Donner B. Pl sm luminum concentr tions in long-ter m mech nic lly ventil ted p tients receiving stress ulcer prophyl xis with sucr lf te. Crit http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 16)03-M y-05 17:34:12

Ovid: ICU Book C re Med 1994;22:17691773. 16. Ben-Men chem T, Fogel R, P tel RV, et l. Prophyl xis for stress-rel ted g s tric hemorrh ge in the medic l intensive c re unit. Ann Intern Med 1994;121:56857 5. 17. Morris DL, M rkh m SJ, Beechey A, et l. R nitidine-bolus or infusion prophy l xis for stress ulcer. Crit C re Med 1988;16:229232.

19. G rvey BM, McC mbley JA, Tuxen DV. Effects of g stric lk liz tion on b cter i l coloniz tion in critic lly ill p tients. Crit C re Med 1989;17:211216. 20. Howden CW, Hunt RH. Rel tionship between g stric secretion nd infection. Gu t 1987;28:96107. 21. Wing te DL. Acid reduction nd recurrent enteritis. L ncet 1990;335:222. 22. Cook GC. Infective g stroenteritis nd its rel tionship to reduced cidity. Sc nd J G stroenterol 1985;20(Suppl 111):1721. 23. Soli AH. Medic l tre tment of peptic ulcer dise se. Pr ctice guidelines. JAM A 1996;275:622629. 24. Derrid S, Nury B, Sl m R, et l. Occult g strointestin l bleeding in highrisk intensive c re unit p tients receiving nt cid prophyl xis: frequency nd s ignific nce. Crit C re Med 1989;17:122125. 25. Rosenth l P, Thompson J, Singh M. Detection of occult blood in g stric juice . J Clin G stroenterol 1984;6:119. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 16)03-M y-05 17:34:12

18. Cook DJ, L ine LA, Guy tt GH, et l. Nosocomi l pneumoni stric pH. Chest 1991;100:713.

nd the role of g

Ovid: ICU Book SELECTIVE DIGESTIVE DECONTAMINATION 26. Stoutenbeek CP, v n S ene HKF, Mir nd DR, Z ndstr DF. The effect of select ive decont min tion of the digestive tr ct on coloniz tion nd infection r te in multiple tr um p tients. Intensive C re Med 1984;10:185192. 27. Heyl nd DK, Cook DJ, J eschke R, et l. Selective decont min tion of the dig estive tr ct. An overview. Chest 1994;105:12211229. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 16)03-M y-05 17:34:12

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 7 VENOUS THROMBOEMBOLISM Two words best ch r cterize the mort lity nd morbidity due to venous thromboemb olism in the United St tes: subst nti l nd un ccept ble. --Kenneth M. Moser, Am Rev Resp Dis, 1990 The thre t of venous thrombosis nd cute pulmon ry embolism (i.e., venous thromboembolism) is n everyd y concern in the c re of critic lly ill p tients. A v riety of clinic l conditions, m ny of them common in ICU p ti ents, c n be ccomp nied by venous thrombosis in the lower extremities. Thrombos is involving the deep venous system in the thighs is often silent process, nd becomes evident only when it sends thrombotic projectiles to the lungs. In more th n two-thirds of c ses of cute pulmon ry embolism, the source of the problem (i.e., proxim l leg vein thrombosis) goes unnoticed before the embolism (1,2,3, 4,5 nd 6). Bec use this thrombosis is such n insidious process, emph sis is pl ced on prevention to reduce the risks ssoci ted with venous thromboembolism. T he m teri l in this ch pter is presented in four p rts. The first p rt identifie s p tients who re t risk for thromboembolism, nd the second p rt describes th e ppropri te prophyl xis for e ch group of high-risk p tients. The third p rt d escribes di gnostic lgorithm for p tients with suspected cute pulmon ry embo lism, nd the fin l p rt describes the e rly m n gement of thromboembolism with ntico gul nts nd thrombolytic gents.

PATIENTS AT RISK The clinic l conditions th t predispose to venous thromboembolism re listed by risk c tegory in T ble 7.1. Accomp nying e ch c tegory is the reported incidence of proxim l deep vein thrombosis (proxim l DVT) in the leg nd f t l pulmon ry embolism (f t l PE). The incidence of c lf vein thrombosis is not included bec u se thrombosis P.107 below the knee is not considered n import nt source of pulm on ry emboli (1). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 19)03-M y-05 17:34:06

Ovid: ICU Book TABLE 7.1. RISK CATEGORIES FOR VENOUS THROMBOEMBOLISM SURGERY Sever l f ctors promote thrombosis in the e rly period following m jor surgery. These include venous st sis, v scul r injury, nd gener lized hyperco gul ble st te (c used by thrombopl stin rele se during surgery nd depressed levels of ntithrombin III th t persist for 5 to 7 d ys fter surgery). Orthopedic procedur es involving the hip nd knee represent p rticul rly high risk for thromboembo lism, s does c ncer-rel ted surgery in the bdomen nd pelvis. TRAUMA Tr um c rries the s me risk f ctors for thromboembolism s surgery (surgery is form of controlled tr um ). The high-risk conditions in tr um include multisy stem involvement, cute spin l cord injury, nd fr ctures involving the pelvis nd lower extremities. MEDICAL ILLNESS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 19)03-M y-05 17:34:06

Ovid: ICU Book Rel tively few cute medic l illnesses c rry high risk of venous thromboemboli sm. The most noted high-risk medic l conditions re cute myoc rdi l inf rction, ischemic stroke, lower extremity p r lysis P.108 (either ph rm cologic or dise se rel ted), nd c ncer (p rticul rly pelvic tumors). METHODS OF PROPHYLAXIS A v riety of preventive me sures h ve proven effective in reducing the incidence of thromboembolism in critic lly ill p tients. As result, routine thromboprop hyl xis is recommended for ll p tients in the moder te- nd high-risk c tegorie s in T ble 7.1 (2,7). This point deserves emph sis bec use surveys reve l th t p hysici ns neglect thromboprophyl xis in 70 to 80% of p tients who would benefit from the intervention (8). The preventive me sures recommended in e ch type of c linic l condition re shown in T ble 7.2 (2). E ch of these me sures is describe d briefly in the p r gr phs th t follow. TABLE 7.2. PREFERRED (P) AND ALTERNATIVE (A) METHODS OF THROMBOPROPHYLAXIS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 19)03-M y-05 17:34:06

Ovid: ICU Book GRADED COMPRESSION STOCKINGS Gr ded compression stockings ( lso known s thromboembolism deterrent or TED sto ckings) promote venous flow in the legs by providing 18 mm Hg extern l compressi on t the nkles nd 8 mm Hg extern l compression in the thigh (3). These stocki ngs h ve proven effective in reducing the incidence of thromboembolism ssoci te d with m jor bdomin l surgery nd neurosurgery (9,10). Nevertheless, TED stocki ngs re not recommended s the sole preventive me sure in ny moder te- or highrisk clinic l condition (2). P.109 INTERMITTENT PNEUMATIC COMPRESSION BOOTS Intermittent pneum tic compression boots re infl t ble devices th t provide 35 mm Hg extern l compression t the nkle nd 20 mm Hg extern l compression t the thigh. These devices re considered more effective th n gr ded compression stoc kings, nd c n more th n double the venous flow r te in the legs (2,3). Bec use there is no risk of bleeding with pneum tic boots, they re f vored in neurosurg ic l p tients nd in p tients undergoing prost tectomy (2,3). Pneum tic boots r e lso very effective in p tients undergoing reconstructive surgery of the knee, nd c n be used s the sole prophyl ctic me sure in these p tients if there is no interference from immobiliz tion c sts. LOW-DOSE HEPARIN The m jor ntico gul nt ction of hep rin is to ctiv te ntithrombin III, which then inhibits the conversion of prothrombin to thrombin. In the bsence of cti ve thrombosis, this ction occurs t low doses of hep rin, below the doses th t interfere with other components of the co gul tion process. As result, low dos es of hep rin c n inhibit thrombus form tion without cre ting the risk of hemorr h ge ssoci ted with full ntico gul tion (11). Dosing Regimen The usu l low-dose hep rin regimen is 5000 IU subcut neous every 12 hours. In su rgic l p tients, the initi l dose is given 2 hours before surgery, nd ther py c ontinues through the first postoper tive week, or until the p tient is mbul tor y. L bor tory tests of co gul tion st tus re not monitored. Low-dose hep rin is recommended s effective prophyl xis in m jor bdomin l surgery nd in cute me dic l illnesses with risk of thromboembolism. It does not provide optim l prop hyl xis in high-risk tr um tic nd orthopedic conditions. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 19)03-M y-05 17:34:06

Ovid: ICU Book LOW-MOLECULAR-WEIGHT HEPARIN Convention l prep r tions of hep rin cont in heterogeneous mix of polys cch ri de molecules th t v ry widely in size, nd only 30% of the molecules show ntico gul nt ctivity (11,12). These molecules c n be broken down (enzym tic lly) int o sm ller molecules of more uniform size. The low-molecul r-weight (LMW) hep rin h s more ntico gul nt ctivity nd produces n ntico gul nt effect t lower d os ges th n the convention l unfr ction ted hep rin. The potenti l dv nt ges of LMW P.110 hep rin over unfr ction ted hep rin include less frequent dosing, low er risk of bleeding, nd lower incidence of hep rin-induced thrombocytopeni ( 11,12 nd 13). At present, m ny of these dv nt ges re more theoretic l th n c tu l. The dis dv nt ge of LMW hep rin is the cost. Prophyl xis with LMW hep rin using the regimen shown below is 10 times more costly (per d y) th n low-dose he p rin (13).

Dosing Regimen For enox p rin (Lovenox, Rhone-Poulenc Rorer Ph rm ceutic ls), the dos ge for th romboprophyl xis is 30 mg subcut neous every 12 hours. L bor tory tests of co gu l tion st tus re not monitored. LMW hep rin is more effective th n low-dose hep rin in hip fr ctures, reconstructive surgery of the hip nd knee, nd cute spi n l cord injury with p r lysis (T ble 7.2) (2,12). It is not recommended in p ti ents with ctive bleeding, or in documented c ses of hep rin-induced thrombocyto peni . LOW-DOSE WARFARIN Low-level ntico gul tion with coum din is n effective ltern tive to LMW hep r in in p tients with high risk of thromboembolism. However, prophyl xis with co um din is more cumbersome th n LMW hep rin bec use it requires dos ge titr tion nd monitoring of l bor tory tests of co gul tion st tus. Dosing Regimen The initi l dose of coum din is 10 mg or lly, followed by d ily dose of 2 mg. The dos ge is then djusted to chieve prothrombin time with n intern tion l norm lized r tio (INR) of 2 to 3 (2). (See Reference 25 for description of the INR). Low-dose w rf rin is recommended in the s me high-risk clinic l condition s s low-molecul rhttp://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 19)03-M y-05 17:34:06

Ovid: ICU Book weight hep rin. Bec use it h s been shown to reduce the incidence of f t l pulmo n ry embolism in p tients with hip fr ctures (2), low-dose w rf rin m y be the p referred method of prophyl xis for hip fr ctures. VENA CAVA FILTERS Meshlike filters c n be pl ced in the inferior ven c v to prevent emboliz tion of clots in the leg veins (14,15). The indic tions for pl cement of ven c v l filter re listed below. Indic tions q Documented iliofemor l vein thrombosis nd r Contr indic tion to ntico gul tion (such s ctive hemorrh ge) r Documented pulmon ry emboliz tion during full ntico gul tion r Free-flo ting thrombus P.111 r

No iliofemor l vein thrombosis, but s Long-term prophyl xis is necess ry ( s in p r plegi ) s High risk for both thromboembolism nd hemorrh ge The Greenfield Filter The most widely used ven c v l filter in the United St tes is the Greenfield fi lter (Medi-Tech, W tertown, M ss.) shown in Figure 7.1. The conic l sh pe of thi s filter is structur l http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 19)03-M y-05 17:34:06

High-risk condition for f t l pulmon ry embolus (such r

s severe lung dise se)

Ovid: ICU Book dv nt ge bec use s much s 75% of the b sket c n be filled with thrombus wit hout compromising the cross-section l re of the ven c v . This design limits the risk for ven c v obstruction nd troublesome leg edem .

Figure 7.1. The Greenfield ven c v filter. Note the elong ted sh pe, which ll ows the filter to tr p thrombi without compromising the cross-section l re of the ven c v . Greenfield filters re inserted percut neously, usu lly through the intern l jug ul r vein or femor l vein, nd re pl ced below the ren l veins, if possible. Su pr ren l pl cement is occ sion lly necess ry when the thrombus extends to the le vel of the ren l veins, nd does not imp ir venous dr in ge from the kidneys. Th e Greenfield filter h s been the preferred ven c v filter in the United St tes for l st 20 ye rs, nd h s proven both s fe nd effective. The reported inciden ce of pulmon ry embolism with c v l filter in pl ce is 2 to 5% (14), nd f t l embolism is r re. Complic tions re uncommon. Venous insufficiency occurs in on ly 5% of c ses (14), nd troublesome migr tion of the filters is r re. The cost of inserting these devices deserves mention. In 1992, the p tient ch rge for per cut neous pl cement of Greenfield filter w s $5300 in one hospit l (16). P.112 DIAGNOSTIC APPROACH As mentioned e rlier, thrombosis in the deep veins of the thigh is often clinic lly silent, nd the problem becomes evident only when pulmon ry embolus occurs . Therefore, the di gnostic ev lu tion of thromboembolism usu lly involves p tie nts suspected of h ving n cute http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 19)03-M y-05 17:34:06

Ovid: ICU Book pulmon ry embolism. CLINICAL PRESENTATION The clinic l m nifest tions of cute pulmon ry embolism re neither sensitive no r specific enough to be of di gnostic v lue. As shown in T ble 7.3, no single cl inic l m nifest tion c n reli bly predict whether pulmon ry embolus is present . Of p rticul r note is the observ tion th t hypoxemi m y be bsent in 30% of p tients with cute pulmon ry embolism (i.e., neg tive predictive v lue = 0.70). Although not shown in T ble 7.3, norm l lveol r rteri l PO2 gr dient lso does not exclude the presence of n cute pulmon ry embolism (17). Thus, clinic l pr esent tion c nnot be used to confirm or exclude the presence of pulmon ry emboli . When the clinic l findings re suspicious for n cute pulmon ry embolism, the di gnostic ev lu tion c n proceed ccording to the lgorithm in Figure 7.2. TABLE 7.3. CLINICAL FINDINGS IN ACUTE PULMONARY EMBOLISM http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 19)03-M y-05 17:34:06

Ovid: ICU Book Figure 7.2. Flow di gr m for the di gnostic ppro ch to cute pulmon ry embolism . VENOUS ULTRASOUND The di gnostic ev lu tion in Figure 7.2 begins with se rch for thrombosis in t he proxim l veins of the leg. This ppro ch is b sed on the f ct th t pulmon ry embolism is not prim ry illness, but is second ry m nifest tion of n underl ying venous thrombosis. Bec use most pulmon ry emboli origin te from thrombi in the femor l veins (1), the ev lu tion begins with se rch for thrombosis in the deep veins of the thigh. P.113 Although contr st venogr phy is the most reli bl e method of documenting venous thrombosis in the legs, v scul r ultr sound h s p roven to be reli ble noninv sive method for detecting proxim l vein thrombosis in the legs. V scul r ultr sound offers two complement ry techniques for the de tection of venous thrombosis (18). The first technique is venous compression ult r sound. This method uses two-dimension l brightness modul tion (Bmode) ultr sou nd to obt in cross-section l view of the femor l rtery nd vein, s shown in http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 19)03-M y-05 17:34:06

Ovid: ICU Book p nel 1 of Figure 7.3. When compressive force is cre ted by pushing the ultr s ound probe g inst the overlying skin, the vein becomes obliter ted, s shown in p nel 2 of Figure 7.3. When vein is filled with thrombus, extern l compress ion does not obliter te the vessel. Thus, f ilure to obliter te the femor l vein by extern l compression is indirect evidence of femor l vein thrombosis. Figure 7.3. Ultr sound visu liz tion of the femor l rtery (A) nd femor l vein (V). Extern l compression of the overlying skin in p nel 2 results in obliter ti on of the femor l vein. (Ad pted from Cron n JJ, Murphy TP. A comprehensive revi ew of v scul r ultr sound for intensivists. J Intensive C re Med 1993;8:188.) The second method is Doppler ultr sound, which relies on the well-known Doppler shift to detect the velocity of flow in n underlying vessel. This method is p r ticul rly v lu ble when it is difficult to distinguish P.114 rteries nd veins with two-dimension l B-mode ultr sound. The combin tion of compression nd Doppl er methods is c lled duplex ultr sound. This combined (duplex) http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 19)03-M y-05 17:34:06

Ovid: ICU Book method is more ccur te th n either of the individu l ultr sound methods, nd h s sensitivity of 90 to 100% nd specificity of 80 to 100% for the di gnosis of venous thrombosis in the thigh (18). Ultr sound is less reli ble for the dete ction of c lf-vein thrombosis bec use the veins below the knee re sm ller nd m ore difficult to visu lize with ultr sound techniques. Although most pulmon ry e mboli origin te from thrombi in the proxim l (iliofemor l) veins of the legs (1) , s m ny s 30% of p tients with cute pulmon ry emboli show no evidence of ven ous thrombosis in the legs (19). As result, the bsence of venous thrombosis i n the legs does not exclude the possibility of n cute pulmon ry embolus. When the se rch for leg vein thrombosis is unrew rding, the next step is to obt in r dionuclide lung sc n. P.115 RADIONUCLIDE LUNG SCANS The interpret tion of ventil tionperfusion lung sc ns c n be summ rized s follow s (20): A norm l lung sc n excludes the presence of (clinic lly import nt) pul mon ry embolus. A high-prob bility lung sc n me ns th t the p tient h s n 85% c h nce of h ving pulmon ry embolism. A low-prob bility lung sc n m y not reli b ly exclude the presence of pulmon ry embolism in p tients with underlying c rd iopulmon ry dise se. However, when low-prob bility lung sc n is combined with neg tive ev lu tion for leg vein thrombosis, it is prob bly s fe to observe th e p tient ( nd not proceed with pulmon ry ngiogr phy). An intermedi te-prob bil ity or indetermin te lung sc n h s no v lue in predicting the presence or bsenc e of pulmon ry embolus. Using these interpret tion guidelines, the lung sc n s hould be the fin l di gnostic procedure in the m jority of p tients with suspect ed pulmon ry embolism. Only p tients with indetermin te or intermedi te-prob bil ity lung sc ns should require further di gnostic ev lu tion.

PULMONARY ANGIOGRAPHY Pulmon ry ngiogr phy is the most ccur te (but not inf llible) method of confir ming or excluding the presence of pulmon ry emboli. However, when the di gnostic lgorithm in Figure 7.2 is followed, only minority of c ses of suspected pulm on ry embolism should require this inv sive nd expensive procedure. ANTITHROMBOTIC THERAPY http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 19)03-M y-05 17:34:06

Ovid: ICU Book ANTICOAGULATION The m inst y of cute ther py for thromboembolism is ntico gul tion with intr v enous hep rin. The go l is to chieve n ctiv ted p rti l thrombopl stin time ( PTT) th t is 1.5 to 2.5 times control levels (21). Ther peutic levels of ntico gul tion should be chieved s soon s possible bec use del ys in re ching the t her peutic PTT r nge h ve been shown to incre se the risk of progressive thrombo sis nd recurrent pulmon ry embolism (21). Hep rin dosing b sed on body weight h s proven superior to convention l hep rin dosing in chieving ther peutic level s of ntico gul tion r pidly nd s fely (22). The recommended dos ges of hep rin b sed on ( ctu l) body weight re P.116 shown in T ble 7.4. One point worth not ing is th t weight-b sed hep rin-dosing nomogr ms h ve been derived from p tient s weighing less th n 130 kg (22). In morbidly obese p tients (body weight > 130 kg), weight-b sed hep rin dos ge c n result in excessive ntico gul tion (23), s o it is import nt to monitor ntico gul tion c refully when using weight-b sed h ep rin dosing in the morbidly obese. TABLE 7.4. WEIGHT-BASED HEPARIN DOSING REGIMEN Or l ntico gul tion with coum din c n be st rted on the first d y of hep rin th er py. When the prothrombin time re ches n intern tion l norm lized r tio (INR) of 2 to 3, the http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 19)03-M y-05 17:34:06

Ovid: ICU Book hep rin c n be discontinued. (The INR is more st nd rdized me surement of the prothrombin time, nd h s been dopted to reduce the v ri bility in prothrombin time determin tions in different clinic l l bor tories. See Reference 25 for m ore det iled description of the INR method.) Although the st nd rd dur tion of h ep rin ther py for thromboembolism is 10 to 14 d ys, recent evidence indic tes t h t 5 d ys of hep rin ther py is s effective s the tr dition l 10- to 14-d y r egimen (24). L bor tory Abnorm lities Two nonhemorrh gic side effects of hep rin tre tment deserve mention. The first is elev tion of serum minotr nsfer se levels, which h s been reported in s m n y s 80% of p tients receiving hep rin (26). The incre se in serum tr ns min se levels usu lly occurs 5 to 10 d ys fter onset of hep rin ther py, nd pe k tr n s min se levels c n be 15 times higher th n the norm l r nge. This l bor tory b norm lity is not ssoci ted with liver dysfunction, nd it dis ppe rs fter the drug is discontinued. Therefore, n elev tion in serum tr ns min se P.117 levels during hep rin tre tment should prompt further di gnostic studies unless there is other evidence of liver dysfunction. The other complic tion is hyperk lemi , which occurs in 5 to 10% of p tients receiving hep rin (27), nd is the result o f hep rin-induced ldosterone suppression. The hyperk lemi c n ppe r within few d ys fter the onset of hep rin ther py, nd h s been reported with hep rin dos ges s low s 5000 IU twice d y (the s me dos ge used P.118 in low-dose he p rin thromboprophyl xis). The potenti l for hyperk lemi h s prompted the recom mend tion th t serum pot ssium levels be monitored periodic lly in p tients rece iving hep rin for more th n 3 d ys (27). THROMBOLYTIC THERAPY E rly ther py with thrombolytic gents is ssoci ted with more r pid nd more co mplete resolution of venous thromboemboli th n ntico gul tion ther py with hep rin (28,29). Despite this dv nt ge, thrombolytic ther py h s not g ined widespr e d f vor in the cute m n gement of venous thromboembolism. Sever l f ctors con tribute to this l ck of enthusi sm. Foremost is the risk of hemorrh ge, which co ntr indic tes thrombolytic ther py in l rge number of hospit lized p tients. I n one study, 93% of p tients with leg vein thrombosis h d t le st one contr ind ic tion to thrombolytic ther py (30). (See Ch pter 19 for the contr indic tions to thrombolytic ther py.) In ddition, thrombolytic ther py h s not h d signif ic nt imp ct on mort lity nd (e rly) morbidity ssoci ted with venous thromboem bolism http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 19)03-M y-05 17:34:06

Ovid: ICU Book (28). A fin l dr wb ck is the brief ther peutic window for thrombolytic ther py: Thrombolytic gents re effective only if given within the first 7 d ys fter t he onset of thromboembolism (28). Bec use proxim l leg vein thrombosis is often clinic lly silent ( nd thus the time of onset is not known), it is often not kno wn whether thrombolytic ther py will f ll in the ppropri te ther peutic window. The recommended dos ges of three thrombolytic gents in venous thrombosis nd p ulmon ry embolism re shown in Figure 7.4. All three gents should be considered equiv lent in effic cy nd risk of hemorrh ge (28). If streptokin se is dminis tered, thrombin time should be obt ined few hours fter the onset of drug th er py. Prolong tion of the thrombin time to twice control v lues provides eviden ce for proteolytic st te in the pl sm (29). If the thrombin time is not suffi ciently prolonged, give second lo ding dose th t is double the initi l dose (t o overcome binding by streptokin se ntibodies), nd recheck the thrombin time 2 to 3 hours l ter. F ilure to chieve n dequ te lytic st te fter the second l o ding dose is re son to switch to nother lytic gent. Thrombolytic ther py sho uld lw ys be followed by ntico gul tion ther py with hep rin nd w rf rin. (Fo r more inform tion on thrombolytic ther py, see Ch pter 19.) http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 19)03-M y-05 17:34:06

Ovid: ICU Book Figure 7.4. Thrombolytic ther py for venous thromboembolism. (From Hyers TM et l. Antithrombotic ther py for venous thromboembolic dise se. Chest 1992;102(Supp l):408S.) REFERENCES REVIEWS 1. Moser KM. Venous thromboembolism. Am Rev Respir Dis 1990;141:235249 (199 refer ences). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 19)03-M y-05 17:34:06

Ovid: ICU Book 2. Cl ggett GP, Anderson FA, Levine MN, et l. Prevention of venous thromboembol ism. In: D len JE, Hirsh J, eds. Third ACCP Consensus Conference on Antithrombot ic Ther py. Chest 1992;102(Suppl):391S407S (210 references). P.119 3. Goldh ber SZ, M rpurgo M, for the WHO/ISFC T sk Force on Pulmon ry Embolism. Di gnosis, tre tment nd prevention of pulmon ry embolism. JAMA 1992;268:1727 173 3 (57 references). 4. Weinm nn EE, S lzm n EW. Deep-vein thrombosis. N Engl J Med 1994;331:1630 1641 (215 references). CONCISE REVIEWS 5. Cowen J, Kelley MA. Pulmon ry embolism in the critic lly ill: str tegies for prevention nd m n gement. J Crit Illness 1994;9:988991.

THROMBOPROPHYLAXIS 7. Thromboembolic Risk F ctors (THRIFT) Consensus Group. Risk of nd prophyl xis for venous thromboembolism in hospit lized p tients. BMJ 1992;305:567574. 8. Anderson FA Jr, Wheeler B, Goldberg RJ, et l. Physici n pr ctices in the pre vention of venous thromboembolism. Ann Intern Med 1991;115:591596. 9. Wells PS, Lensing AW, Hirsh J. Gr du ted compression stockings in the prevent ion of postoper tive venous thromboembolism. A met - n lysis. Arch Intern Med 19 94;154:67 72. 10. Turpie AG, Hirsh J, Gent M, et l. Prevention of deep vein thrombosis in pot enti l http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 19)03-M y-05 17:34:06

6. Rosenow EC III. Venous nd pulmon ry thromboembolism: n lgorithmic to di gnosis nd m n gement. M yo Clin Proc 1995;70:4549.

ppro ch

Ovid: ICU Book neurosurgic l p tients. Arch Intern Med 1989;149:679681. 11. Hirsh J, D len JE, Deykin D, Poller L. Hep rin: mech nism of ction, ph rm c okinetics, dosing consider tions, monitoring, effic cy nd s fety. Chest 1992;10 2 (Suppl):337S351S. 12. Hirsh J, Levine MN. Low-molecul r-weight hep rin. Blood 1992;79:117. 13. W rkentin TE, Levine MN, Hirsh J, et l. Hep rin-induced thrombocytopeni in p tients tre ted with low-molecul r-weight hep rin or unfr ction ted hep rin. N Engl J Med 1995;332:13301335. 14. Becker DM, Philbrick JT, Selby B. Inferior ven c v filters. Indic tions, s fety, effectiveness. Arch Intern Med 1992;152:19851994.

16. M gn nt JG, W lsh DB, Jur vsky LI, Cronenwett JL. Current use of inferior ve n c v filters. J V sc Surg 1992;16:701706.

18. Stew rt JH, Grubb M. Underst nding v scul r ultr sonogr phy. M yo Clin Proc 1992;67:11861196. 19. Hull RD, Hirsh J, C rter CJ, et l. Pulmon ry ngiogr phy, ventil tion lung sc nning, nd venogr phy for clinic lly suspected pulmon ry embolism with bnorm l perfusion sc ns. Ann Intern Med 1983;98:891899.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (17 of 19)03-M y-05 17:34:06

DIAGNOSTIC APPROACH 17. Stein PD, Goldh ber SZ, Henry JW. Alveol r rteri l oxygen gr dient in the essment of cute pulmon ry embolism. Chest 1995;107:139143.

15. Smith BA. Ven

c v filters. Emerg Med Clin North Am 1994;12:645656.

ss

Ovid: ICU Book 20. Kelley MA, C rson JL, P levsky HI, Schw rtz S. Di gnosing pulmon ry embolism : new f cts nd str tegies. Ann Intern Med 1991;114:300306. P.120 ANTITHROMBOTIC THERAPY 21. Hull RD, R skob GE, Rosenbloom D, et l. Optim l ther peutic level of hep ri n ther py in p tients with venous thrombosis. Arch Intern Med 1992;152:15891595. 22. R schke RA, Reilly BM, Guidry JR, et l. The weight-b sed hep rin dosing nom ogr m comp red with st nd rd c re nomogr m. Ann Intern Med 1993;119:874 881. 23. Hollid y DM, W tling SM, Y nos J. Hep rin dosing in the morbidly obese p tie nt. Ann Ph rm cother 1994;28:11101111. 24. Hull RD, R skob GE, Rosenbloom D, et l. Hep rin for 5 d ys comp red with 10 d ys in the initi l tre tment of proxim l venous thrombosis. N Engl J Med 1990; 322:12601264. 25. Le DT, Weibert RT, Sevill BK, et l. The intern tion l norm lized r tio (IN R) for monitoring w rf rin ther py: reli bility nd rel tion to other monitoring methods. Ann Intern Med 1994;120:552558. 26. S lomon F, Schmid M. Hep rin. N Engl J Med 1991;325:1585.

28. Hyers TM, Hull RD, Weg JG. Antithrombotic ther py for venous thromboembolic dise se. Chest 1992;102(Suppl):408S425S.

29. Rogers LQ, Lutcher CL. Streptokin se ther py for deep vein thrombosis: http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (18 of 19)03-M y-05 17:34:06

27. Oster JR, Singer I, Fishm n LM. Hep rin-induced ldosterone suppression hyperk lemi . Am J Med 1995;98:575586.

nd

Ovid: ICU Book comprehensive review of the English liter ture. Am J Med 1990;88:389395.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (19 of 19)03-M y-05 17:34:06

30. M rkel A, M nzo RA, Str ndess E Jr. The potenti l role of thrombolytic ther py in venous thrombosis. Arch Intern Med 1992;152:12651267.

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 8 ANALGESIA AND SEDATION Men do not fe r de th, they fe r the p in of dying. --Apsley Cherry-G rr rd Cont r ry to popul r perception, our princip l function in p tient c re is not to s v e lives (bec use this is impossible on consistent b sis), but to relieve suffe ring. And there is no loc tion in the hospit l th t c n m tch the suffering expe rienced by both p tients nd f milies th n the intensive c re unit. To get n id e of how well prep red we re to relieve p in nd suffering in the ICU, look t Figure 8.1. Figure 8.1. Percent ge of house st ff physici ns nd ICU nurses who nswered inc orrectly when sked whether di zep m (V lium) is n n lgesic. (From Loper KA et l. P r lysed with p in: the need for educ tion. P in 1989;37:315.) This ch pter focuses on the v il ble methods for est blishing p tient comfort i n the ICU with http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 25)03-M y-05 07:42:47

Ovid: ICU Book p renter lly dministered n lgesic nd sed tive medic tions. For more inform ti on on this import nt topic, number of gener l reviews re included t the end of the ch pter (1,2,3,4,5,6 nd 7). PAIN IN THE ICU PREVALENCE Although m jority of p tients in the ICU receive p renter l n lgesics routine ly (8,9), p in is prominent p rt of the p tient experience in the ICU. Anywher e from 30 to 70% of p tients re bothered by p in during their ICU st y (10,11 nd 12), nd the p in is described s moder te, severe, or excruci ting in over h lf of the c ses (10,12). In survey of p tients disch rged from n ICU, h lf o f the respondents identified p in s their worst memory of the ICU experience (1 3). (For vivid, first-h nd ccount of the p in involved in being n ICU p tien t, s experienced by critic l c re physici n, see Reference 14.) It ppe rs P. 122 th t the incidence of in dequ te n lgesi in the ICU is the s me tod y s i t w s 30 ye rs go (11). OPIOPHOBIA The problem of in dequ te p in control is c used l rgely by misconceptions bout the ddictive potenti l of n rcotic n lgesics nd bout the ppropri te dos ge needed to chieve effective n lgesi (15). The following st tements re direct ed t these misconceptions. Opi te use in hospit lized p tients does not cre te drug ddicts (16). The effective dos ge of n rcotic n lgesic is determined by the p tient response, not by some perception of wh t n effective dos ge should be (2,3 nd 4). Avoiding irr tion l fe rs bout n rcotic n lgesics is n impor t nt step in le rning to provide dequ te p in relief for ICU p tients. MONITORING PAIN P in severity c n be recorded using v riety of p in intensity sc les such s t he ones shown in Figure 8.2 (1). The uppermost sc le (Adjective R ting Sc le) us es descriptive terms, the middle sc le (Numeric l P.123 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 25)03-M y-05 07:42:47

Ovid: ICU Book R nking Sc le) uses whole numbers, nd the lower sc le (Visu l An log Sc le) rec ords p in intensity s discrete point pl ced long line between two ends of the p in intensity spectrum. The Visu l An log Sc le is the one most commonly us ed in clinic l studies of n lgesic regimens (1), but there is no evidence th t ny one sc le is superior to the others.

Figure 8.2. Three different sc les for recording p in intensity. The recommended length for the numeric sc les (NRS nd VAS) is 10 cm. For more inform tion on r ecording p in intensity, see Reference 1. P in intensity sc les re used to ev lu te the effect of n lgesic regimens in i ndividu l p tients. A numeric l score of 3 or less on the Numeric l R ting Sc le or Visu l An log Sc le c n be used s evidence of effective n lgesi (3). Howe ver, it seems e sier to just sk the p tient whether he or she h s p in, nd whe ther he or she would like something to relieve it. Direct communic tion with p t ients is not only the best method of determining comfort needs, it is itself s ource of comfort to p tients.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 25)03-M y-05 07:42:47

OPIOID ANALGESIA Chemic l deriv tives of opium re c lled opi tes. These subst nces re ed n rcotics (sleep-inducers), but this term h s become

lso c ll

Ovid: ICU Book P.124 stigm tized by l w enforcement gencies (who use the term to refer to illi cit drugs) nd is no longer useful. Opi tes produce their effects by stimul ting discrete receptors in the centr l nervous system c lled opioid receptors. There re t le st three types of opioid receptors: , , and . Stimulation of receptor p roduce analge ia, euphoria, bradycardia, con tipation, and re piratory depre i on. -Receptor timulation produce edation and pupillary con triction. receptor produce dy phoria, delirium, and hallucination . Opiate and other chemical ub tance that timulate opioid receptor are called opioid . Opioid are the mo t commonly admini tered agent for pain relief and edation in the ICU (8,9). The opioid u ed mo t often are morphine and fentanyl, delivered via intravenou , i ntrathecal, or epidural route .

INTRAVENOUS OPIOIDS The intravenou admini tration of morphine and fentanyl i de cribed in Table 8. 1 (1,2,3 and 4,17,18 and 19). The re pon e to opioid varie in individual patie nt (3), and a wide range of do age may be nece ary to achieve effective analg e ia in individual patient . A mentioned earlier, the effective opioid do age hould be determined by patient re pon e, not by the numerical value of the do a ge. The do age hown in Table 8.1 are the u ual effective do age but may not a pply to all patient . TABLE 8.1. INTRAVENOUS OPIOID ANALGESIA (3, 17, 18) http://gateway.ut.ovid.com/gw1/ovidweb.cgi (4 of 25)03-May-05 07:42:47

Fentanyl ver u Morphine In urvey including both medical and urgical ICU , morphine wa the mo t commo n agent given for both analge ia and edation (8,9). Although fentanyl eem le popular, it may be the preferred opioid agent for analge ia in the ICU. One of the mo t noted difference between fentanyl and morphine i their lipid olubil ity. Fentanyl i roughly 40 time more lipid oluble than morphine, o fentanyl i P.125 ta en up into the (lipid-laden) central nervou y tem much more readil y than morphine. Thi ma e fentanyl a more potent analge ic than morphine. In g eneral, fentanyl produce equivalent analge ia at 1/100 the do e of morphine. Fe ntanyl i al o more rapidly acting than morphine, and thu can be titrated more rapidly. Both opioid hould be given by continuou infu ion, preceded by a load ing do e. Intermittent a needed (PRN) do ing i never recommended for pain cont rol in the ICU (2,3 and 4) becau e of the high incidence of inadequate analge ia with PRN do ing chedule . Opioid are metabolized primarily in the liver, and the metabolite are excreted in the urine. Morphine ha an active metabolite tha t can accumulate in renal failure, and for thi rea on, the do age of morphine hould be reduced by 50% in patient with renal failure (i.e., creatinine clearan ce le than 10 mL/min) (19). The recommendation for fentanyl do ing in renal f ailure are incon i tent. One ource recommend a 50% reduction in the fentanyl d o age in renal failure (19), and another ource recommend no change in do age ( 20). For patient with un table or compromi ed hemodynamic function, fentanyl i preferred to morphine for two rea on . The fir t rea on i more rapid titration with fentanyl, a mentioned earlier. The econd rea on i tendency for morphine to promote hi tamine relea e, which promote va odilation and hypoten ion. Fent anyl how much le tendency to timulate hi tamine relea e, o fentanyl u e el iminate the ri of adver e hemodynamic effect produced by hi tamine. Patient-Controlled Analge ia For patient who are awa e and aware, a method nown a patient-controlled analg e ia (PCA) can be u ed for pain control. Thi method involve intermittent intra venou do e of an opioid delivered on-demand by the patient, u ing a elf-activ ating infu ion pump. A mandatory time interval i et between ucce ive do e . Thi time interval i called the loc out interval becau e the patient i unable to deliver a do e of the drug during thi time. The recommended do age regimen fo r PCA are included in Table 8.1. The loc out interval for each drug i clo e to the time required for the on et of drug action. Common loc out interval are 5 m inute for fentanyl and 10 minute for morphine. The intermittent do e and the http://gateway.ut.ovid.com/gw1/ovidweb.cgi (5 of 25)03-May-05 07:42:47

Ovid: ICU Boo

loc out interval mu t be pecified when the order for PCA are written. Patient ati faction i often greater with PCA than with conventional intravenou do ing (21,22), and the technique i particularly popular in po toperative patient .

EPIDURAL OPIOIDS Epidural in tallation of opioid i a popular method of pain control in the earl y po toperative period following thoracic and abdominal urgery. Catheter are p laced in the epidural pace at the cervical, P.126 thoracic, or lower lumbar lev el while the patient i in the operating room, and the analge ic are admini ter ed through the e catheter for the fir t few po toperative day . The do ing regi men for epidural analge ia are hown in Table 8.2 (23). The opioid can be give n a intermittent bolu do e but are more commonly given a a continuou infu i on along with a local ane thetic uch a bupivacaine. The addition of the local ane thetic provide more effective analge ia. To minimize the ri of impaired m otor function and local ympathectomy (with hypoten ion) that accompany epidural ane the ia, only dilute olution of local ane thetic are in tilled. TABLE 8.2. EPIDURAL ANALGESIA AND ANESTHESIA Benefit There i little evidence that the epidural route provide any advantage over the http://gateway.ut.ovid.com/gw1/ovidweb.cgi (6 of 25)03-May-05 07:42:47

Ovid: ICU Boo

intravenou route for opioid analge ia. When a local ane thetic i added to the epidural regimen, the analge ia eem to be more effective than with intravenou opioid (24). However, epidural opioid alone do not provide better analge ia t han intravenou opioid (24). More important, epidural analge ia doe not reduce po toperative morbidity (cardiova cular, pulmonary, or ga trointe tinal) when c ompared with intravenou analge ia (24). Adver e Effect Adver e effect of epidural analge ia are more common with morphine than with fe ntanyl. Epidural opioid can produce re piratory depre ion, but the reaction i delayed. The incidence of re piratory depre ion i 1% for epidural morphine an d 0.9% for intravenou morphine (24). More prevalent ide effect of epidural an alge ia include pruritu (28 to 100%), nau ea (30 to 100%), and urinary retentio n (15 to 90%) (24). The pruritu from epidural opioid i di cu ed in the ecti on of thi chapter on adver e effect of opioid . Intrathecal Opioid Opioid can al o be placed in the ubarachnoid pace around the lower lumbar pi ne (in lower do e ) to produce analge ia. However, P.127 there i a ri of arac hnoiditi if catheter are left in place, o the epidural route i favored. ADVERSE EFFECTS OF OPIOIDS There i a long li t of adver e reaction to opioid ; the mo t important are di cu ed here. (For a more comprehen ive review of opioid ide effect , ee Refere nce 25 and 26.) The opioid antagoni t naloxone i mentioned here, but the u e o f naloxone i de cribed in Chapter 53. Re piratory Depre ion The ame receptor that mediate the analge ic action of opiate al o produce r e piratory depre ion. Therefore, ome degree of re piratory depre ion i expec ted with opioid analge ia. Decrea e in re piratory rate and tidal volume are co mmon, but the e change can repre ent a general edative effect, not pecific in hibition of brain tem re piratory neurone (26). In fact, the u e of opioid doe not eem to increa e the frequency of hypoxemic epi ode in po toperative pati ent (27). Patient with leep-apnea yndrome, however, can develop more frequen t apneic epi ode with evere oxygen de aturation (i.e., SaO2 below 80%) during leep (26). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (7 of 25)03-May-05 07:42:47

Ovid: ICU Boo

Cardiova cular Effect Decrea e in blood pre ure and heart rate are common ide effect of opioid ana lge ia. However, li e re piratory depre ion, the e change often re ult from a generalized decrea e in ympathetic nervou y tem activity and are not patholog ic change . Opioid-induced hypoten ion i well tolerated, at lea t in the upine po ition, and u ually require no intervention (26). The ame i true of the br adycardia produced by opioid , which i u ually a ymptomatic at re t. Inte tinal Motility The depre ion of ga trointe tinal motility a ociated with opioid u e can be a problem in patient in the ICU, who have other condition that depre bowel mot ility (the po toperative tate). Parenteral naloxone can antagonize thi opioid effect, but there i al o lo of the analge ic effect . The oral admini tration of naloxone ha been ucce ful in treating con tipation from chronic opioid u e without producing y temic opioid antagoni m (28). The tarting do e i 1 mg e very 6 hour , titrated upward to a maximum do e of 4 mg every 6 hour . Pruritu A mentioned, pruritu i common with epidural opioid , particularly morphine. T hi effect i re i tant to treatment with antihi tamine P.128 but can be elimin ated with naloxone (29). Naloxone at a do age of 1 to 2 g/ g/hr can bloc the pru ritu without antagonizing the analge ic action of the opioid (Dr. Kenneth Sut in, MD, per onal communication). Low do e of propofol (10 mg) have al o re ulte d in prompt re olution of the pruritu a ociated with epidural opioid (30), by an un nown mechani m. Meperidine Meperidine (Demerol, Pethidine) wa not among the opioid analge ic pre ented he re. The rea on for excluding thi agent i the ri of neurotoxicity. Meperidine i metabolized to normeperidine in the liver, and thi metabolite i excreted v ery lowly in the urine (elimination half-life of 17 hour ). Normeperidine i a neurotoxin, and when it accumulate , it can produce a yndrome characterized by delirium, hallucination , p ycho i , and generalized eizure (31). Normeperidin e can accumulate with repeated do e of meperidine, and the accumulation i more pronounced when renal function i impaired. Becau e there are everal condition in critically ill patient that can impair renal function the ri of neurotox icity hould be particularly high in the e patient . Therefore, it doe not eem wi e to u e meperidine a an analge ic in http://gateway.ut.ovid.com/gw1/ovidweb.cgi (8 of 25)03-May-05 07:42:47

Ovid: ICU Boo

the ICU. NONOPIOID ANALGESIA There are few alternative to the opioid for providing analge ia via the parent eral route. In fact, there i only one alternative, etorolac. KETOROLAC Ketorolac i a non teroidal antiinflammatory drug (NSAID) introduced in 1990 a a parenteral analge ic agent for po toperative pain. Becau e etorolac had none of the trouble ome ide effect of the opioid , it became an overnight hit. By t he end of 1992, over 26 million patient had received the drug. The popularity o f thi agent ha waned ince then, due to a combination of limited efficacy and trouble ome ide effect . Action Ketorolac i li e other NSAID (a pirin-li e drug ) and ha analge ic and antiin flammatory action . It i more of an analge ic agent, and i 350 time more pote nt than a pirin (32). Li e other NSAID , etorolac inhibit platelet adhe ion an d can damage the ga tric muco a. P.129 Ketorolac can be given orally, or by intr amu cular (IM) or intravenou (IV) injection. After IM injection, analge ia beco me evident at 1 hour, pea at 2 hour , and la t 5 to 6 hour . The drug i par tly metabolized in the liver and excreted in the urine. Elimination i prolonged in renal impairment and old age. Efficacy For po toperative analge ia, 30 mg etorolac IM i equivalent to 100 mg meperidi ne IM or 12 mg morphine IM. Ketorolac i u ually not given alone, but it i adde d to opioid analge ia. Often thi allow the opioid do age to be reduced (33). T hi i the opiate- paring effect of etorolac.

http://gateway.ut.ovid.com/gw1/ovidweb.cgi (9 of 25)03-May-05 07:42:47

Do ing Regimen IM or IV (34): 10 mg initial do e, then 10 to 30 mg every 4 to 6 hour

Ovid: ICU Boo

Maximum duration: 2 day Maximum daily do age: 60 mg for the elderly, 90 mg for the other Thi do ing regimen i a re tricted ver ion of the original regimen a nd wa amended becau e of the ri of ga tric muco al injury (34). Although the drug i u ually given by IM injection, thi can re ult in hematoma formation (35 ). The IV route i afer (33). Ketorolac ha been given by continuou IV infu io n (5 mg/hour), re ulting in more effective analge ia than intermittent IV do e (33). Adver e Effect Common ide effect include omnolence (7%), nau ea (7%), dizzine (6%), headac he (4%), and dy pep ia (2%) (32). The bleeding time i prolonged, but there i little evidence of abnormal bleeding. Although ga tric ulceration i a concern, the ri i poorly documented. One ca e of anuric renal failure (on et 1 hour af ter IM etorolac) ha been reported (36). ANXIETY IN THE ICU ANXIETY DISORDERS Anxiety and related di order (agitation and delirium) are almo t univer al in t he ICU patient population, with a many a 90% of patient in the ICU howing ev idence of an anxiety di order (7). Anxiety i characterized by a en e of forebo ding or doom that i magnified P.130 beyond that expected from external circum t ance (7,37). The proce i thu driven more by internal mechani m than by ext ernal event . If the internal proce become autonomou and cannot be controlle d, the anxiety become pathological (i.e., thought proce e become counterprodu ctive). Agitation i a combination of anxiety and increa ed motor activity, and delirium i a pecific yndrome complex that may or may not have anxiety a a co mponent (delirium i de cribed in detail in Chapter 50). SEDATION Sedation i the proce of e tabli hing a tate of relaxation or well-being (38) . It doe not, a often a umed, nece arily involve a depre ed level of con ci ou ne . Survey of edative drug u e in medical ICU indicate that no fewer tha n 18 different pharmacologic agent are u ed for edation in the ICU (9). The dr ug mo t commonly u ed are benzodiazepine and opioid analge ic (8,9). The opio id have a prominent edative effect, and they can be u ed http://gateway.ut.ovid.com/gw1/ovidweb.cgi (10 of 25)03-May-05 07:42:47

Ovid: ICU Boo

alone or in combination with benzodiazepine . The benzodiazepine have no analge ic effect , and thu cannot be u ed alone in patient who have pain a part of their anxiety tate. When benzodiazepine are given to patient with pain, they can produce a dy phoric tate that add to the underlying anxiety.

Monitoring Unfortunately, there i no objective mea ure of anxiety or edation, o it i im po ible to monitor therapy with edative in a quantitative or tandardized man ner. A few clinical coring method are available for anxiety and edation, but they are not con idered reliable enough to be of any value (39). SEDATION WITH BENZODIAZEPINES Benzodiazepine eem well- uited for the ICU becau e they have amne ic effect ( anterograde amne ia) and becau e they are afe to u e. Of the 13 benzodiazepine available for clinical u e, 3 can be given intravenou ly: diazepam, lorazepam, and midazolam. Table 8.3 pre ent ome pertinent information on the intravenou benzodiazepine . TABLE 8.3. SEDATION WITH INTRAVENOUS BENZODIAZEPINES

http://gateway.ut.ovid.com/gw1/ovidweb.cgi (11 of 25)03-May-05 07:42:47

DRUG COMPARISONS The following are zepine .

ome characteri tic

hared by all of the intravenou

benzodia

Ovid: ICU Boo

All are lipid oluble to ome degree. All are metabolized in the liver and excre ted in the urine. P.131 The pharmaco inetic de cription doe not alway correlat e with the clinical performance of the drug. For example, the elimination half-l ife of diazepam i 20 to 50 hour , and i only 2 hour for midazolam, yet the ti me of recovery from edation i the ame with both agent (40). The edative eff ect of benzodiazepine are more pronounced in old age, general debility, heart failure, and hepatic in ufficiency. Although they do not cau e re piratory depre ion in healthy ubject , benzodiazepine can cau e re piratory depre ion in e lderly patient and patient with chronic CO2 retention. Diazepam (Valium) i th e lea t favored of the intravenou benzodiazepine becau e of a tendency for ove r edation with repeated drug admini tration. After a ingle intravenou do e of diazepam, on et of edation occur in 1 to 2 minute , and the edative effect la t 6 to 12 hour (41). Repeated do ing re ult in accumulation of the drug and it active metabolite, and thi accumulation lead to more pronounced edation w ith repeated u e of the drug. Certain precaution are required when admini terin g diazepam intravenou ly. The drug i not water oluble, and nonpolar olvent uch a propylene glycol mu t be added to the commercial preparation to eep the drug in olution. Adding the drug to intravenou fluid can re ult in precipitat ion, o the drug hould be injected a clo e to the hub of the catheter a po i ble (38). The drug i al o very irritating, o fluid hould be flowing through t he catheter during and after the injection. Al o, injection into a large central vein i preferred. The rate of injection hould not exceed 5 mg/minute. Lorazep am (Ativan) ha the lowe t on et and longe t duration of action of the intraven ou benzodiazepine . After a ingle intravenou do e of lorazepam, on et of eda tion occur at 5 to 15 minute , and the edative effect la t 10 to 20 hour (41 ). Becau e of it long duration of action, lorazepam i be t uited for prolonge d edation of table patient (e.g., chronic ventilator-dependent patient ). The ame precaution for drug delivery de cribed for diazepam al o apply to lorazep am. Midazolam (Ver ed) i the benzodiazepine of choice for hort-term edation i n the ICU. Thi agent ha the highe t lipid olubility, P.132 the fa te t on et, and the horte t duration of action of all the intravenou benzodiazepine . Bec au e of it hort elimination half-life (1 to 2 hour ), midazolam i commonly gi ven by continuou infu ion, u ing the do age hown in Table 8.4. However, midaz olam often http://gateway.ut.ovid.com/gw1/ovidweb.cgi (12 of 25)03-May-05 07:42:47

Ovid: ICU Boo

accumulate and produce exce ive edation after prolonged period of infu ion (i.e., longer than 48 hour ) (3,42). Thi tendency to accumulate i cau ed partl y by the high lipid olubility of the drug and partly by prolonged clearance in critically ill and elderly patient (43). Midazolam accumulation i particularly prominent in obe e patient becau e of the lipophilic nature of the drug. There fore, in obe e patient , the ideal body weight, not the actual body weight, hou ld be u ed to determine the appropriate do age of midazolam (2). TABLE 8.4. SEDATION BY CONTINUOUS INFUSION THERAPY The tendency for midazolam to accumulate more readily than u pected i cau ed p artly by a mi conception about the hort elimination half-life of the drug. The rapid elimination of midazolam from the blood partly repre ent reupta e of the drug into ti ue , and thu doe not mean rapid elimination from the body. There fore, the elimination half-life of the drug hould not be ta en a a reflection of excretion. TOXIC EFFECTS The principal toxicity of the benzodiazepine i exce ive edation with depre ed con ciou ne (44). Although ventilation i not uppre ed in normal ubject , over edation with benzodiazepine ha been a ociated with delayed weaning fro m mechanical ventilation (43,45). The manife tation and treatment of benzodiaze pine toxicity are de cribed in Chapter 53. Withdrawal Syndrome http://gateway.ut.ovid.com/gw1/ovidweb.cgi (13 of 25)03-May-05 07:42:47

Ovid: ICU Boo

Abrupt termination of chronic benzodiazepine inta e can produce a withdrawal yn drome con i ting of extreme agitation, di orientation, paranoid delu ion, and ha llucination . Benzodiazepine withdrawal P.133 can be a ource of unexplained del irium in the fir t few day after admi ion to the ICU (46). The ri of withdra wal from long-term benzodiazepine edation in the ICU i not clear. Gradual with drawal ha been recommended after 1 month of midazolam edation to prevent thi complication (38). Drug Interaction Several drug can interfere with the oxidative metaboli m of diazepam and midazo lam in the liver, and the e are li ted in Table 8.5. (The e interaction do not apply to lorazepam, which i metabolized by a different mechani m.) The interact ion between erythromycin and midazolam i ignificant, and erythromycin hould b e avoided when po ible during midazolam edation, or the do age of midazolam h ould be reduced by 55 to 75% during erythromycin therapy (47). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (14 of 25)03-May-05 07:42:47

Ovid: ICU Boo

TABLE 8.5. DRUG INTERACTIONS WITH BENZODIAZEPINES The interaction between theophylline and the benzodiazepine al o de erve menti on. Theophylline antagonize benzodiazepine edation (po ibly by adeno ine bloc ade) (48,49), and intravenou aminophylline (110 mg over 5 minute ) ha been re ported to cau e more rapid awa ening from benzodiazepine edation in po toperati ve patient (48). Thi i li ely to be a ignificant interaction, o it i wi e to avoid theophylline in patient receiving benzodiazepine edation. Thi hould not be a difficult deci ion becau e theophylline i not the mo t effective bron chodilator available ( ee Chapter 25). P.134

OTHER SEDATIVES PROPOFOL Propofol (Deprivan) i a rapid acting edative that wa introduced a an ane the tic induction agent and ha ince been propo ed for hort-term edation in the I CU (50). The major advantage of propofol i rapid awa ening, and the major di ad vantage i expen e. The daily co t of propofol i hown in Table 8.4 (18). Action Propofol i a lipid- oluble agent with a rapid on et and hort duration of actio n. A ingle intravenou do e of propofol produce edation within 1 minute of in jection. The edative effect pea at 2 minute , and la t 4 to 8 minute (50). Recovery from edation u ually occur within 10 minute and i not prolonged by prolonged drug admini tration (51). Propofol i rapidly converted to inactive me tabolite in the liver. Propofol ha no analge ic effect . The edative effect can be accompanied by ignificant re piratory depre ion and va odilation. The v a odilator action can lead to a decrea e in blood pre ure, which occur more o ften after bolu admini tration of the drug, or in patient who are hypovolemic or have a labile blood pre ure (50,51). U e Propofol ha been recommended for hort-term edation in the early po toperative period. Some report how more rapid weaning from mechanical ventilation a oci ated with propofol edation compared with midazolam edation (52). However, thi i not a con i tent ob ervation (53). At pre ent, the propo ed advantage of pro pofol in promoting more rapid http://gateway.ut.ovid.com/gw1/ovidweb.cgi (15 of 25)03-May-05 07:42:47

Ovid: ICU Boo

Preparation and Do age Propofol i not water oluble, and the drug i available a a 1% olution (10 mg /mL) in 10% Intralipid ( oybean oil, glycerol, and egg lecithin). The recommende d do age of propofol are hown in Table 8.4. Due to it hort duration of actio n, the drug i given by continuou infu ion. Becau e it i very lipophilic, prop ofol tend to accumulate in obe e patient . Therefore, a with midazolam, the do age of propofol in obe e patient hould be ba ed on ideal body weight, not act ual body weight (50). P.135 Adver e Effect The major adver e effect of propofol are re piratory depre ion and hypoten ion . Becau e of the ri of re piratory depre ion, propofol hould be u ed only in patient who are receiving mechanical ventilation (38,50). The ri of hypoten ion i greate t after bolu drug admini tration, and for thi rea on the loading do e of the drug i ometime eliminated. Propofol hould not be u ed in patien t who are hemodynamically un table. HALOPERIDOL Haloperidol (Haldol) ha gained popularity a a edative in the ICU becau e of t he low ri of cardiore piratory depre ion. Although the drug i effective for generalized edation, it may be particularly valuable in the patient with deliri um (i.e., confu ional anxiety). The di advantage of the drug include a low on et of action and the inability for rapid titration of drug do age. The intraveno u route ha yet to receive approval by the FDA, but intravenou haloperidol ha been proven afe and effective in repeated clinical trial (54,55). Action Intravenou haloperidol ha a prolonged di tribution time (11 minute or even lo nger in critically ill patient ) (55), o the on et of action i low, but the d rug can be given by rapid intravenou pu h. A edative effect hould be evident within 20 minute after drug injection. The edation i not accompanied by re pi ratory depre ion, and hypoten ion i unu ual unle the patient i hypovolemic or receiving a -receptor antagonist (55). Extrapyramidal reactions are not common with the intravenous route of drug administration (55,56). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 25)03-May-05 07:42:47

weaning from mechanical ventilation i extreme expen e of the drug.

not documented well enough to ju tify the

Ovid: ICU Boo

Ovid: ICU Book Uses Because of its delayed onset of action, haloperidol is not indicated for rapid c ontrol of the anxiety state. A enzodiazepine (e.g., lorazepam, 1 mg) can e add ed to achieve more rapid sedation (56). Haloperidol is often targeted for the pa tient with delirium. However, ecause of the lack of respiratory depression, we use the drug for sedation of ventilator-dependent patients. Because the dosage c annot e titrated rapidly, we reserve the drug for chronic ventilator-dependent patients who are otherwise sta le. Dosage The dosing recommendations for intravenous haloperidol are shown in Ta le 8.6. T he dosages are higher than the traditional intramuscular dosages of haloperidol. Individual patients show a wide variation in serum drug levels after a given do se of haloperidol (57). P.136 If there is no evidence of a sedative response aft er 20 minutes, the dosage should e dou led. If there is a partial response at 2 0 to 30 minutes, a second dose can e given along with 1 mg lorazepam (56). Lack of response to a second dose of haloperidol should prompt a switch to another a gent. TABLE 8.6. INTRAVENOUS HALOPERIDOL Adverse Effects http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 25)03-May-05 07:42:47

Ovid: ICU Book There are two serious ut uncommon reactions to haloperidol. One is the neurolep tic malignant syndrome (NMS) (58), a potentially fatal condition characterized y hyperthermia, muscle rigidity, autonomic dysfunction, and mental confusion. Th e muscle rigidity is widespread, and is often associated with rha domyolysis and myoglo inuric renal failure. This condition is a variant of the malignant hyper thermia syndrome, and responds to early treatment with the muscle relaxant dantr olene (see Chapter 30). NMS is an idiosyncratic response to haloperidol and is i ndependent of drug dosage or duration of therapy (58). Early recognition is mand atory for successful treatment with dantrolene. Haloperidol prolongs the QT inte rval, and this predisposes to the second serious reaction, torsade de pointes (p olymorphic ventricular tachycardia). This is an uncommon complication, with a re ported incidence of 0.4% in patients receiving intravenous haldol in the ICU (59 ). Nevertheless, ecause of the risk of this complication, it is wise to avoid h aloperidol in patients with a prolonged QT interval and in those with a history of torsade de pointes. REFERENCES GENERAL TEXTS Hamill RJ, Rowlingson JC, eds. Hand ook of critical care pain management. New Yo rk: McGraw-Hill, 1994. Park GR, Sladen RN, eds. Sedation and analgesia in the critically ill. Oxford: B lackwell Scientific Pu lications, 1995. SYMPOSIA Hansen-Flaschen J, ed. Improving patient tolerance of mechanical ventilation. Cr itical Care Clinics, Vol. 10. Philadelphia: WB Saunders, 1994 (Octo er). Hoyt JW, ed. Pain management in the ICU. Critical Care Clinics, Vol. 6. Philadel phia: WB Saunders, 1990 (April). P.137 CLINICAL PRACTICE GUIDELINES http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 25)03-May-05 07:42:47

Ovid: ICU Book 1. Acute Pain Management Guideline Panel. Acute pain management: operative or me dical procedures and trauma. Clinical practice guideline. AHCPR Pu . No. 92-0032 . Rockville, MD: Agency for Health Care Policy and Research, Pu lic Health Servi ce, U.S. Department of Health and Human Services, Fe . 1992. REVIEWS 2. Levine RL. Pharmacology of intravenous sedatives and opioids in critically il l patients. Crit Care Clin 1994;10:709731 (98 References). 3. Murray MJ, Plevak DJ. Analgesia in the critically ill patient. New Horiz 1994 ;2:5663 (34 References). 4. Wheeler A. Sedation, analgesia, and paralysis in the intensive care unit. Che st 1993;104:566577 (184 References). 5. Dur in CG. Sedation in the critically ill patient. New Horiz 1994;2:6474 (66 R eferences). 6. Crippen DW. Pharmacologic treatment of rain failure and delirium. Crit Care Clin 1994;10:733766 (110 References). 7. Bone RC, Hayden WR, Levine RL, et al. Recognition, assessment, and treatment of anxiety in the critical care patient. Dis Mon 1995;31:296359 (169 References). THE PAIN EXPERIENCE 8. Dasta JF, Fuhrman TM, McCandles C. Patterns of prescri ing and administering drugs for agitation and pain in patients in a surgical intensive care unit. Crit Care Med 1994;22:974980. 9. Hansen-Flaschen JH, Brazinsky S, Lanken PN. Use of sedating drugs and http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 25)03-May-05 07:42:47

Ovid: ICU Book neuromuscular locking agents in patients requiring mechanical ventilation for r espiratory failure. JAMA 1991;266:28702875. 10. Donovan M, Dillon P, McGuire L. Incidence and characteristics of pain in a s ample of medical-surgical inpatients. Pain 1987;30:6978. 11. Stevens DS, Edwards WT. Management of pain in the critically ill. J Intensiv e Care Med 1990;5:258291. 12. Puntillo KA. Pain experiences in ICU patients. Heart Lung 1990;19:526533. 13. Paiement B, Boulanger M, Jones CW, Roy M. Intu ation and other experiences i n cardiac surgery: the consumer's views. Can Anesth Soc J 1979;26:173180. 14. Hayden WR. Life and near-death in the intensive care unit. Crit Care Clin 19 94;10:651658. 15. Zenz M, Willwe er-Strumpf A. Opiopho ia and cancer pain in Europe. Lancet 19 93;341:10751076. 16. Porter J, Jick H. Addiction rare in patients treated with narcotics. N Engl J Med 1980;302:123128. OPIOID ANALGESIA 17. Willens JS, Myslinski NR. Pharmacodynamics, pharmacokinetics, and clinical u ses of fentanyl, sufentanil, and alfentanil. Heart Lung 1993;22:239251. 18. Armstrong DK, Crisp CB. Pharmacoeconomic issues of sedation, analgesia, and neuromuscular lockade in critical care. New Horiz 1994;2:8593. 19. Bennett WM, Aronoff GR, Golper TA, et al., eds. Drug prescri ing in renal fa ilure. 2nd ed. Philadelphia: American College of Physicians, 1991;63. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 25)03-May-05 07:42:47

Ovid: ICU Book 20. St. Peter WL, Halstenson CE. The pharmacologic approach in patients with ren al failure. In: Chernow B, ed. The pharmacologic approach to the critically ill patient. 3rd ed. Baltimore: Williams & Wilkins, 1994;4179. P.138 21. Smythe M. Patient-controlled analgesia: a review. Pharmacotherapy 1992;12:13 3 141. 22. Eg ert AM, Lampros LL, Parks LL. Effects of patient-controlled analgesia on postoperative anxiety in elderly men. Am J Crit Care 1993;2:118124. 23. Rowlingson JC, Hamill RJ. Techniques of narcotic and local anesthetic admini stration. In: Hamill RJ, Rowlingson JC, eds. Hand ook of critical care pain mana gement. New York: McGraw-Hill, 1994;207228. 24. Liu S, Carpenter RL, Neal JM. Epidural anesthesia and analgesia. Anesthesiol ogy 1995;82:14741506. 25. Buck ML, Blumer JL. Opioids and other analgesics. Adverse effects in the int ensive care unit. Crit Care Clin 1991;7:615637. 26. Schug SA, Zech D, Grond S. Adverse effects of systemic opioid analgesics. Dr ug Safety 1992;7:200213. 27. Bailey PL. The use of opioids in anesthesia is not especially associated wit h nor predictive of postoperative hypoxemia. Anesthesiology 1992;77:1235. 28. Culpepper-Morgan JA, Inturrisi CE, Portenoy RK, et al. Treatment of opioid-i nduced constipation with oral naloxone: a pilot study. Clin Pharmacol Ther 1992; 52:9095. 29. Gowan JD, Hurtig JB, Fraser RA, et al. Naloxone infusion after prophylactic epidural http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 25)03-May-05 07:42:47

Ovid: ICU Book morphine: effects on incidence of postoperative side-effects and quality of anal gesia. Can J Anesth 1988;35:143148. 30. Borgeat A, Wilder-Smith OHG, Salah M, Rifat K. Su hypnotic doses of propofol relieve pruritus induced y epidural and intrathecal morphine. Anesthesiology 1 992;76:510512. 31. Shochet RB, Murray GB. Neuropsychiatric toxicity of meperidine. J Intensive Care Med 1988;3:246252. 32. Buckley M, Brogden RN. Ketorolac. A review of its pharmacodynamic and pharma cokinetic properties and therapeutic potential. Drugs 1990;39:86109. 33. Ready LB, Brown CR, Stahlgren LH, et al. Evaluation of intravenous ketorolac administered y olus or infusion for treatment of postoperative pain. Anesthes iology 1994;80:12771286. 34. Drug Update. Ketorolac doses reduced. Lancet 1993;342:105. 35. Garcha IS, Bostwick J. Postoperative hematomas associated with toradol. Plas t Reconstr Surg 1991;88:919920. 36. Boras-U er LA, Brackett NC. Ketorolac-induced acute renal failure. Am J Med 1992;92:450452. ANXIETY IN THE ICU 37. Hansen-Flaschen J. Improving patient tolerance of mechanical ventilation. Cr it Care Clin 1994;10:659671. 38. Sladen RN. Sedation in the intensive care unit: clinical considerations. In: Reves JG, Green latt DJ, Sladen RV, eds. Drug infusion for sedation in the inte nsive care unit. Boston: Tufts University School of Medicine, 1994;2436. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 25)03-May-05 07:42:47

Ovid: ICU Book 39. Hansen-Flaschen J, Cowan J, Polomano RC. Beyond the Ramsay scale: need for a validated measure of sedating drug efficacy in the intensive care unit. Crit Ca re Med 1994;22:732733. INTRAVENEOUS BENZODIAZEPINES 40. Ariano R, Kassum D, Aronson K. Comparison of sedative recovery time after mi dazolam versus diazepam administration. Crit Care Med 1994;22:14921496. P.139 41. Green latt DJ, Ehren erg DJ, Gunderman J, et al. Kinetic and dynamic study o f intravenous lorazepam: comparison with intravenous diazepam. J Pharmacol Exp T her 1989;250:134140. 42. Byatt CM, Lewis LD, Dawling S, et al. Accumulation of midazolam after repeat ed dosage in patients receiving mechanical ventilation in an intensive care unit . BMJ 1984;289:799800. 43. Malacrida R, Fritz M, Suter P, et al. Pharmacokinetics of midazolam administ ered y continuous infusion to intensive care unit patients. Crit Care Med 1992; 20:11231126. 44. Gaudreault P, Guay J, Thivierge RL, Verdy I. Benzodiazepine poisoning. Drug Safety 1991;6:247265. 45. Shalansky SJ, Naumann TL, Englander FA. Therapy update: effect of flumazenil on enzodiazepine-induced respiratory depression. Clin Pharmacol 1993;12:483487. 46. Moss JH. Sedative and hypnotic withdrawal states in hospitalized patients. L ancet 1991;338:575 (letter). 47. Olkkola KT, Aranko K, Luurila H, et al. A potentially hazardous interaction etween erythromycin and midazolam. Clin Pharmacol Ther 1993;53:298305. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (23 of 25)03-May-05 07:42:47

Ovid: ICU Book 48. Hoegholm A, Steptoe P, Fogh B, et al. Benzodiazepine antagonism y aminophyl line. Acta Anesthesiol Scand 1989;33:164166. 49. Gallen JS. Aminophylline reversal of midazolam sedation. Anesth Analg 1985;6 9:268 (letter). PROPOFOL 50. Barr J. Propofol: a new drug for sedation in the intensive care unit. Int An esthesiol Clin 1993;31:131154. 51. Carrasco G, Molina R, Costa J, et al. Propofol vs. midazolam in short-, medi um-, and long-term sedation of critically ill patients. A cost enefit analysis. C hest 1993;103:557 564. 52. Roekaerts PMHJ, Huygen FJPM, DeLange S. Infusion of propofol versus midazola m for sedation in the intensive care unit following coronary artery surgery. J C ardiothorac Vasc Anesth 1993;7:142147. 53. Higgins TL, Yared J-P, Estafanous FG, et al. Propofol versus midazolam for i ntensive care unit sedation after coronary artery ypass grafting. Crit Care Med 1994;22:1415 1423. HALOPERIDOL 54. Tesar GE, Stern TA. Rapid tranquilization of the agitated intensive care uni t patient. J Intensive Care Med 1988;3:195201. 55. Cassem EH, Lake CR, Boyer WF. Psychopharmacology in the ICU. In: Chernow B, ed. The pharmacologic approach to the critically ill patient. 3rd ed. Baltimore: Williams & Wilkins, 1994;651665. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (24 of 25)03-May-05 07:42:47

Ovid: ICU Book 56. Sanders KM, Minnema AM, Murray GB. Low incidence of extrapyramidal symptoms in the treatment of delirium with intravenous haloperidol and lorazepam in the i ntensive care unit. J Intensive Care Med 1989;4:201204. 57. Lawson GM. Monitoring of serum haloperidol. Mayo Clin Proc 1994;69:189190. 58. Sing RF, Branas CC, Marino PL. Neuroleptic malignant syndrome in the intensi ve care unit. J Am Osteopath Assoc 1993;93:615618. 59. Wilt JL, Minnema AM, Johnson RF, Rosen lum AM. Torsade de pointes associated with the use of intravenous haloperidol. Ann Intern Med 1993;119:391394. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (25 of 25)03-May-05 07:42:47

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 9 ARTERIAL BLOOD PRESSURE It should e clearly recognized that arterial pressure cannot e measured with p recision y means of sphygmomanometers. American Heart Association, Committee fo r Arterial Pressure Recording, 1951 The arterial lood pressure is one of the most common and most unrelia le measur ements in modern medicine. The folly of the lood pressure measurement is demons trated in the following scenario. The most common medical disorder in the United States is hypertension, which affects 20% of Americans over the age of 6 years (1,2). A out 80% of patients with hypertension have no evidence of end-organ inv olvement (1), which means that in the majority of cases of hypertension, the pre sence of the disorder is ased solely on the lood pressure measurement. The sta ndard method for measuring the lood pressure (i.e., sphygmomanometry) is well r ecognized for its inaccuracy (3), as pointed out y the warning issued y the Am erican Heart Association in the introductory statement. Thus, the most common me dical disorder in the United States (high lood pressure) owes its existence to a technique that was disputed y expert opinion almost 50 years ago. This chapte r descri es the methods used to monitor the arterial lood pressure in criticall y ill patients. The first section descri es some of the shortcomings of indirect pressure recordings, and the final section descri es the direct method of recor ding pressures with intravascular catheters. INDIRECT MEASUREMENTS Indirect measurements of arterial pressure are o tained with a device that consi sts of a cloth cuff with an inflata le ru er ladder on the inner surface. The cloth cuff is wrapped around the arm or leg, and the ladder is inflated to gene rate a pressure that compresses the underlying artery and veins. The ladder is then slowly deflated, P.144 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 13)03-May-05 17:16:10

Ovid: ICU Book allowing the compressed artery to open, and the arterial pressure is determined either y detecting sounds that are generated (auscultation method) or y record ing vascular pulsations (oscillometric method). The accuracy of these indirect p ressure determinations is influenced y the size of the inflata le ladder relat ive to the girth of the compressed lim . DIMENSIONS OF THE CUFF BLADDER The inflata le cuff should produce uniform compression of the underlying artery to ensure optimal recordings when the artery is allowed to open. The a ility of cuff inflation to produce uniform arterial compression is a function of the size of the inflata le ladder relative to the size of the lim . Figure 9.1 shows th e optimal dimensions of the cuff ladder for indirect measurements of rachial a rtery pressure. The length of the ladder should e at least 80% of the circumfe rence of the upper arm, and the width of the ladder should e at least 40% of t he upper arm circumference (3,4). If the ladder size is too small for the arm c ircumference, the indirect pressure measurements will e falsely elevated (1,2,3 ,4 and 5). Figure 9.1. Optimal dimensions of the cuff ladder. The width (W) and length (L) of the ladder are expressed in relation to the circumference (C) of the upper arm. The use of inappropriate-sized cuffs is one of the most common sources of error in indirect lood pressure measurements (1,2,3,4 and 5), so attention to this ma tter is important. A simple edside method for evaluating cuff size is descri ed elow. Bedside Method http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 13)03-May-05 17:16:10

Ovid: ICU Book The following maneuver is recommended for each patient efore recording indirect pressures in the arm or leg. First, align the cuff so that its long axis is par allel to the long axis of the arm. Then turn the cuff around so that the inner s urface ( ladder side) is facing outward. Now wrap the cuff around the upper arm. The ladder (width) should P.145 encircle half of the upper arm (circumference) . If the ladder encompasses less than half of the upper arm with this maneuver, select a larger cuff for the pressure measurements. No change in cuff size is n ecessary if the ladder encircles more than half of the upper arm with this mane uver ecause large cuffs (on thin lim s) do not produce considera le errors in lood pressure recording. The following is a rief description of two indirect me thods of lood pressure recording. AUSCULTATORY METHOD The standard method of measuring lood pressure involves manual inflation of an arm cuff placed over the rachial artery. The cuff is then gradually deflated, a nd the pressure is determined y sounds (called Korotkoff sounds) that are gener ated when the artery egins to open. The details of this method are not e prese nted in detail here; see References 3,4 and 5 for a review. The Korotkoff Sounds One of the fundamental pro lems of the auscultatory method is the a ility to hea r the Korotkoff sounds. The threshold frequency for sound detection y the human ear is 16 Hz, and the frequency range of the Korotkoff sounds is just a ove thi s threshold at 25 to 50 Hz (6). (Human speech occurs at frequencies of 120 to 25 0 Hz, whereas optimum sound detection y the human ear occurs at frequencies of 2000 to 3000 Hz.) Thus, the human ear is almost deaf to the sounds it must hear to measure lood pressure. Stethoscope Head Bell-shaped transducer stethoscope heads are designed to detect lower-frequency sounds than flat, diaphragm-shaped transducer heads. Therefore, to optimize dete ction of the lowfrequency Korotkoff sounds, the American Heart Association recom mends that a ell-shaped stethoscope head e used to measure lood pressure (2,4 ). This preference often is not appreciated, as illustrated y the fact that som e stethoscopes are manufactured without a ellshaped head. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 13)03-May-05 17:16:10

Ovid: ICU Book Low Flow States When systemic lood flow is reduced, the auscultatory method can significantly u nderestimate the actual lood pressure. This is illustrated P.146 in Ta le 9.1, which shows the difference etween systolic pressures measured y auscultation o f Korotkoff sounds and systolic pressures recorded with intraarterial catheters in hypotensive patients with low cardiac outputs. In half of the patients, the a uscultatory method underestimated the actual systolic lood pressure y at least 30 mm Hg. According to the American Association for Medical Instrumentation, in direct pressure measurements should e within 5 mm Hg of directly recorded press ures to e considered accurate (7). Thus, there was not a single accurate pressu re recording with the auscultatory method in the study results shown in Ta le 9. 1. TABLE 9.1. BLOOD PRESSURE MEASUREMENT IN SHOCK The poor performance of the auscultatory method in low flow states is not surpri sing ecause Korotkoff sounds are generated y the flow of lood through partial ly constricted arteries. Thus, as flow diminishes, the Korotkoff sounds ecome l ess audi le, and the earliest sounds that signal the systolic pressure might not e detected. The potential for large measurement errors such as those shown in Ta le 9.1 is the reason the auscultatory method should never e used to monitor arterial pressures in hemodynamically unsta le patients. OSCILLOMETRIC METHOD http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 13)03-May-05 17:16:10

Ovid: ICU Book The oscillometric method is ased on the principle of plethysmography to detect pulsatile pressure changes in a near y artery. When an arm cuff is inflated, pul satile pressure changes in an underlying artery produce periodic pressure change s in the inflated cuff. The oscillometric method thus measures periodic pressure changes (i.e., oscillations) in an inflated cuff as an indirect measure of puls atile pressure in an underlying artery (8). The most recognized oscillometric de vice is the Dinamap (device for indirect assessment of mean arterial pressure), first introduced for clinical use in 1976. The original device could detect only mean arterial pressures, ut more modern devices are capa le of measuring oth systolic and diastolic pressures. Relia ility Although considered more relia le than the auscultatory method, the oscillometri c method also suffers from a limited and varia le accuracy. P.147 This is demons trated in Figure 9.2, which shows a comparison of systolic pressures measured wi th an automated oscillometric device and systolic pressures recorded with rachi al artery catheters in patients undergoing cardiopulmonary ypass surgery. The d ark line is the line of unity, where pressures o tained with oth recording tech niques are identical. The lighter lines are placed 5 mm Hg a ove and elow the l ine of unity, and ( ecause indirect pressures should e within 5 mm Hg of direct ly recorded pressures) the area ounded y the lighter lines represents the zone of accepta le accuracy for oscillometric pressure measurements. Note that most of the points fall outside the zone ounded y the lighter lines, (filled square s) indicating that a majority of the oscillometric measurements are inaccurate. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 13)03-May-05 17:16:10

Ovid: ICU Book Figure 9.2. Comparison of direct and oscillometric measurements of systolic pres sure in the rachial artery. (From Gravlee GP, Brockschmidt JK. Accuracy of four indirect methods of lood pressure measurement, with hemodynamic correlations. J Clin Monitor 1990;6:284 298.) Automated oscillometric devices have gained widespread popularity in recent year s, oth in hospitals and in outpatient clinics, so it is important to e aware o f their limitations. INTRAVASCULAR PRESSURES Direct recording of intravascular pressures is recommended for all patients in t he ICU who require careful monitoring of arterial pressure. Unfortunately, direc t arterial pressure monitoring has its own shortcomings. P.148 The following des cription is intended to help reduce errors in interpretation of directly recorde d arterial pressures. PRESSURE VERSUS FLOW WAVES Although there is a tendency to consider arterial pressure an index of lood flo w, pressure and flow are distinct physical entities. Ejection of the stroke volu me generates oth a pressure wave and a flow wave. Under normal conditions, the pressure wave travels 20 times faster than the flow wave (10 m/second versus 0.5 m/second), and thus the pulse pressure recorded http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 13)03-May-05 17:16:10

Ovid: ICU Book in a peripheral artery precedes the corresponding stroke volume y a matter of s econds (9). When vascular impedance (i.e., compliance and resistance) is increas ed, the velocity of transmission of the pressure wave is increased while the vel ocity of transmission of the flow wave is decreased. (When vascular impedance is reduced, pressure can e diminished while flow is enhanced.) Thus, when vascula r impedance is a normal, the arterial pressure is not a relia le index of arteri al flow. This discrepancy etween pressure and flow is one of the major limitati ons of arterial pressure monitoring. THE ARTERIAL PRESSURE WAVEFORM The contour of the arterial pressure waveform changes as the pressure wave moves away from the proximal aorta. This is shown in Figure 9.3. Note that as the pre ssure wave moves toward the periphery, P.149 the systolic pressure gradually inc reases and the systolic portion of the waveform narrows. The systolic pressure c an increase as much as 20 mm Hg from the proximal aorta to the radial or femoral arteries. This increase in peak systolic pressure is offset y the narrowing of the systolic pressure wave, so that the mean arterial pressure remains unchange d. Therefore, the mean arterial pressure is a more accurate measure of central a ortic pressure. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 13)03-May-05 17:16:10

Ovid: ICU Book Figure 9.3. Arterial pressure waveforms at designated points in the arterial cir culation. Systolic Amplification The increase in systolic pressure in peripheral arteries is the result of pressu re waves that are reflected ack from the periphery (10). These reflected waves originate from vascular ifurcations and from narrowed lood vessels. As the pre ssure wave moves peripherally, wave reflections ecome more prominent, and the r eflected waves add to the systolic pressure wave and amplify the systolic pressu re. Amplification of the systolic pressure is particularly prominent when the ar teries are noncompliant, causing reflected waves to ounce ack faster. This is the mechanism for systolic hypertension in the elderly (10). Because a large pro portion of patients in the ICU are elderly, systolic pressure amplification is p ro a ly commonplace in the ICU. RECORDING ARTIFACTS Fluid-filled recording systems can produce artifacts that further distort the ar terial pressure waveform. Failure to recognize recording system artifacts can le ad to errors in interpretation. RESONANT SYSTEMS Vascular pressures are recorded y fluid-filled plastic tu es that connect the a rterial catheters to the pressure transducers. This fluid-filled system can osci llate spontaneously, and the oscillations can distort the arterial pressure wave form (11,12). The performance of a resonant system is defined y the resonant fr equency and the damping factor of the system. The resonant frequency is the inhe rent frequency of oscillations produced in the system when it is distur ed. When the frequency of an incoming signal approaches the resonant frequency of the sy stem, the resident oscillations add to the incoming signal and amplify it. This type of system is called an underdamped system. The damping factor is a measure of the tendency for the system to attenuate the incoming signal. A resonant syst em with a high damping factor is called an overdamped system. WAVEFORM DISTORTION Three waveforms o tained from different recording systems are shown in Figure 9. 4. The waveform in panel A, with the rounded P.150 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 13)03-May-05 17:16:10

Ovid: ICU Book peak and the dicrotic notch, is the normal waveform expected from a recording sy stem with no distortion. The waveform in panel B, with the sharp systolic peak, is from an underdamped recording system. The recording systems used in clinical practice are naturally underdamped, and these systems can amplify the systolic p ressure y as much as 25 mm Hg (13). The systolic amplification can e minimized y limiting the length of the connector tu ing etween the catheter and the pre ssure transducer. Figure 9.4. The rapid flush test. A, Normal test. B, Underdamped system. C, Over damped system. The waveform in panel C of Figure 9.4 shows an attenuated systolic peak with a g radual upslope and downslope and a narrow pulse pressure. This waveform is from an overdamped system. Overdamping reduces the gain of the system and is sometime s the result of air u les trapped in the connector tu ing or in the dome of th e pressure transducer. Flushing the hydraulic system to evacuate air u les sho uld improve an overdamped signal. Unfortunately, it is not always possi le to id entify underdamped and overdamped systems using the arterial pressure waveform. The test descri ed in the next section can help in this regard. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 13)03-May-05 17:16:10

Ovid: ICU Book THE FLUSH TEST A rief flush to the catheter-tu ing system can e applied to determine whether the recording system is distorting the pressure waveform (12,14). P.151 Most com mercially availa le transducer systems are equipped with a one-way valve that ca n e used to deliver a flush from a pressurized source. Figure 9.4 shows the res ults of a flush test in three different situations. In each case, the pressure i ncreases a ruptly when the flush is applied. However, the response at the end of the flush differs in each panel. In panel A, the flush is followed y a few osc illating waveforms. The frequency of these oscillations is the resonant frequenc y (f) of the recording system, which is calculated as the reciprocal of the time period etween the oscillations. When using standard strip-chart recording pape r divided into 1mm segments, f can e determined y measuring the distance etwe en oscillations and dividing this into the paper speed (11); that is, f (in Hz) = paper speed (in mm/second) divided y the distance etween oscillations (in mm ). In the example shown in panel A, the distance (d) etween oscillations is 1.0 mm and the paper speed is 25 mm/second, so f = 25 Hz (25 mm/ second divided y 1.0 mm). Signal distortion is minimal when the resonant frequency of the recordi ng system is five times greater than the major frequency in the arterial pressur e waveform. Because the major frequency in the arterial pulse is approximately 5 Hz (15), the resonant frequency of the recording system in panel A (25 Hz) is f ive times greater than the frequency in the incoming waveform, and the system wi ll not distort the incoming waveform. The flush test in panel B of Figure 9.4 re veals a resonant frequency of 12.5 Hz (f = 25/2). This is too close to the frequ ency of arterial pressure waveforms, so this system will distort the incoming si gnal and produce systolic amplification. The flush test shown in panel C of Figu re 9.4 does not produce any oscillations. This indicates that the system is over damped, and this system will produce a spuriously low pressure recording. When a n overdamped system is discovered, the system should e flushed thoroughly (incl uding all stopcocks in the system) to release any trapped air u les. If this d oes not correct the pro lem, the arterial catheter should e repositioned or cha nged. MEAN ARTERIAL PRESSURE The mean arterial pressure has two features that make it superior to the systoli c pressure for arterial pressure monitoring. First, the mean pressure is the tru e driving pressure for peripheral lood flow. Second, the mean pressure does not change as the pressure waveform moves distally, nor is it altered y distortion s generated y recording systems (11). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 13)03-May-05 17:16:10

Ovid: ICU Book The mean arterial pressure can e measured or estimated. Most electronic pressur e monitoring devices can measure mean arterial pressure y integrating the area under the pressure waveform and dividing this y the duration of the cardiac cyc le. The electronic measurement is preferred to the estimated mean pressure, whic h is derived as the diastolic pressure plus one-third of the pulse pressure. Thi s formula is ased on the assumption that diastole represents two-thirds of the P.152 cardiac cycle, which corresponds to a heart rate of 60 eats/minute. There fore, heart rates faster than 60 eats/minute, which are common in critically il l patients, lead to errors in the estimated mean arterial pressure. Cardiopulmonary Bypass In most circumstances, the mean pressures in the aorta, radial artery, and femor al artery are within 3 mm Hg of each other. However in patients undergoing cardi opulmonary ypass surgery, the mean radial artery pressure can e significantly (more than 5 mm Hg) lower than the mean pressures in the aorta and femoral arter y (16). This condition may e caused y a selective decrease in vascular resista nce in the hand, ecause compression of the wrist often a olishes the pressure d ifference. An increase in radial artery pressure of at least 5 mm Hg when the wr ist is compressed (distal to the radial artery catheter) suggests a discrepancy etween radial artery pressure and pressures in other regions of the circulation (17). REFERENCES INDIRECT MEASUREMENTS 1. Fifth annual report of the Joint National Committee on the Detection, Evaluat ion, and Treatment of High Blood Pressure. Arch Intern Med 1993;153:154183. 2. Reeves RA. Does this patient have hypertension? How to measure lood pressure . JAMA 1995;273:12111218. 3. Pickering TG. Blood pressure measurement and detection of hypertension. Lance t 1994;344:3135. 4. Frolich ED, Grim C, La arthe DR, et al. Recommendations for human lood press ure http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 13)03-May-05 17:16:10

Ovid: ICU Book determination y sphygmomanometers: report of a special task force appointed y the steering committee, American Heart Association. Hypertension 1988;11:210a221a . 5. American Society for Hypertension. Recommendations for routine lood pressure measurement y indirect cuff sphygmomanometry. Am J Hypertens 1992;5:207209. 6. Ellestad MH. Relia ility of lood pressure recordings. Am J Cardiol 1989;63:9 83985. 7. Davis RF. Clinical comparison of automated auscultatory and oscillometric and catheter-transducer measurements of arterial pressure. J Clin Monitor 1985;1:11 4119. 8. Ramsey M. Blood pressure monitoring: automated oscillometric devices. J Clin Monitor 1991;7:5667. INTRAVASCULAR PRESSURES 9. Darovic GO, Vanriper S. Arterial pressure recording. In: Darovic GO, ed. Hemo dynamic monitoring. 2nd ed. Philadelphia: WB Saunders, 1995;177210. P.153 10. Nichols WW, O'Rourke MF. McDonald's lood flow in arteries. 3rd ed. Philadel phia: Lea & Fe iger, 1990;251269. 11. Gardner RM. Direct lood pressure measurement dynamic response requirements. Anesthesiology 1981;54:227236. 12. Darovic GO, Vanriper S, Vanriper J. Fluid-filled monitoring systems. In Daro vic GO, ed. Hemodynamic monitoring. 2nd ed. Philadelphia: WB Saunders, 1995;14917 5. 13. Rothe CF, Kim KC. Measuring systolic arterial lood pressure. Crit Care Med 1980;8:683689. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 13)03-May-05 17:16:10

Ovid: ICU Book 14. Kleinman B, Powell S, Kumar P, Gardner RM. The fast flush test measures the dynamic response of the entire lood pressure monitoring system. Anesthesiology 1992;77:12151220. 15. Bruner JMR, Krewis LJ, Kunsman JM, Sherman AP. Comparison of direct and indi rect methods of measuring arterial lood pressure. Med Instr 1981;15:1121. 16. Rich GF, Lu anski RE Jr, McLoughlin TM. Differences etween aortic and radia l artery pressure associated with cardiopulmonary ypass. Anesthesiology 1992;77 :6366. 17. Pauca AL, Wallenhaupt SL, Kon ND. Relia ility of the radial arterial pressur e during anesthesia. Chest 1994;105:6975. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 13)03-May-05 17:16:10

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 10 THE PULMONARY ARTERY CATHETER A searchlight cannot e used effectively without a fairly thorough knowledge of the territory to e searched. --Fergus Macartney, FRCP The pulmonary artery cath eter is not just important for the specialty of critical care, it is responsi le for the specialty of critical care. This catheter is so much a part of patient care that it is impossi le to function properly in the ICU without a clear under standing of this catheter and the information it provides. The pulmonary artery catheter is very much like a politician: it seems to function in the est intere st of the population it serves, ut you are never certain that what it tells you is completely trustworthy. This chapter introduces the information generated y pulmonary artery catheters (1,2,3 and 4). Most of the hemodynamic varia les in this chapter are descri ed in detail in Chapter 1 and Chapter 2, so it may help to review these chapters efore proceeding further. Remem er that the power of t he pulmonary artery catheter is ased not on its a ility to generate information , ut on the clinician's a ility to understand that information. This may seem t o e a trivial point, ut surveys indicate that physicians have an inadequate un derstanding of the measurements provided y pulmonary artery catheters (5). CATHETER DESIGN The irth of the pulmonary artery (PA) catheter is descri ed here y HJC Swan, t he cardiologist responsi le for the original concept of the catheter. In the fall of 1967, I had occasion to take my (then young) children to the eac h in Santa Monica . It was a hot Saturday, and the sail oats on the P.155 water w ere ecalmed. However, a out half-a-mile offshore, I http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 16)03-May-05 17:16:30

Ovid: ICU Book noted a oat with a large spinnaker well set and moving through the water at a r easona le velocity. The idea then came to put a sail or parachute on the end of a highly flexi le catheter and there y increase the frequency of passage of the device into the pulmonary artery (1). Three years later (in 1970), a PA catheter was introduced with a small alloon a t its tip. When inflated, the alloon serves as a sail that allows the flow of v enous lood to carry the catheter through the right side of the heart and out in to the pulmonary artery. This alloon flotation principle allows a right heart c atheterization to e performed at the edside, without fluoroscopic guidance. BASIC FEATURES A asic PA catheter is illustrated in Figure 10.1. The catheter is 110 cm long a nd has an outside diameter of 2.3 mm (7 French). There are two internal channels : One channel runs the entire length of the catheter and opens at the very tip o f the catheter (the distal port), and the P.156 other channel is shorter, with a n opening 30 cm from the catheter tip (the proximal port). The tip of the cathet er is equipped with a alloon with a 1.5-mL capacity. As shown, the fully inflat ed alloon creates a recess for the tip of the catheter, and this prevents the c atheter tip from perforating the vascular structures as the catheter is advanced . Finally, there is a thermistor (i.e., a transducer device that senses changes in temperature) located on the outer surface of the catheter 4 cm from the cathe ter tip. The thermistor can measure the flow of a cold fluid that is injected th rough the proximal port of the catheter, and this flow rate is equivalent to the cardiac output (see Chapter 12). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 16)03-May-05 17:16:30

Ovid: ICU Book Figure 10.1. A standard pulmonary artery catheter. ADDITIONAL ACCESSORIES Other accessories are availa le on specially designed PA catheters: An extra cha nnel that opens 14 cm from the catheter tip and that can e used as an infusion channel or for passing temporary pacemaker leads into the right ventricle (6) A fi eroptic system that allows continuous monitoring of mixed venous oxygen satur ation (7) A rapid-response thermistor that can measure the ejection fraction of the right ventricle (8) A thermal filament that generates low-energy heat pulses and allows continuous thermodilution measurement of the cardiac output (9) The large variety of gadgets availa le on the PA catheter makes this instrument akin to a Swiss Army knife for the critical care specialist. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 16)03-May-05 17:16:30

Ovid: ICU Book INSERTING THE CATHETER The PA catheter is inserted into the su clavian or internal jugular veins. (A la rge- ore introducer catheter like the one shown in Figure 10.1 is often used to facilitate insertion and removal of PA catheters.) Just efore insertion, the di stal port of the catheter is connected to a pressure transducer and oscilloscope monitor. Pressure tracings are monitored continuously while the catheter is ins erted to help identify the location of the catheter tip. When the catheter tip f irst enters the vessel lumen, oscillations appear on the pressure tracing from t he distal port of the catheter. When this occurs, the alloon should e fully in flated with 1.5 mL of air. The catheter is then advanced while the alloon is in flated. The pressure waveforms encountered as the catheter is advanced through t he right side of the heart are shown in Figure 10.2 (10). Figure 10.2. The pressure waveforms encountered during insertion of a pulmonary artery catheter. q The superior vena cava pressure is identified y the appearance of oscillations in the pressure recording. The pressure recorded from the vena cava remains unch anged when the catheter tip P.157 is advanced into the right atrium. (Normal sup erior vena cava pressure is 1 to 6 mm Hg.) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 16)03-May-05 17:16:30

Ovid: ICU Book When the catheter tip is advanced across the tricuspid valve and into the right ventricle, a pulsatile systolic pressure appears. The diastolic pressure on the pulsatile waveform is equal to the right-atrial pressure. (Normal right-ventricu lar systolic pressure is 15 to 30 mm Hg.) q When the catheter is carried across the pulmonic valve and into the pulmonary ar tery, the diastolic pressure suddenly rises, whereas the systolic pressure remai ns unchanged. (Normal pulmonary artery diastolic pressure is 6 to 12 mm Hg.) q As the catheter is advanced along the pulmonary artery, the systolic component o f the waveform suddenly disappears. The pressure that remains is known as the pu lmonary capillary wedge pressure (PCWP), and it is usually in the same range as the pulmonary artery diastolic pressure. (Normal wedge pressure is 6 to 12 mm Hg .) q When the wedge pressure tracing appears, stop advancing the catheter and deflate the alloon. The pulsatile pulmonary artery pressure should reappear. BALLOON INFLATION The alloon at the tip of the PA catheter should e deflated at all times while the catheter is left in place in the pulmonary artery. Balloon inflation is rese rved for measurements of the pulmonary artery wedge pressure. When the alloon i s inflated for a wedge pressure measurement, do not fully inflate the alloon (w ith 1.5 mL air) all at once. The alloon should e slowly inflated until a wedge pressure tracing P.158 appears on the oscilloscope monitor. Once a satisfactory wedge pressure is recorded, the alloon should e fully deflated. (Detaching th e syringe from the alloon injection port will help prevent undetected alloon i nflation while the catheter is in place.) COMMON PROBLEMS The following are some pro lems that may appear during the placement of a PA cat heter. Catheter Will Not Advance into the Right Ventricle In this instance, no pulsatile pressure (from the right ventricle) is o tained e ven if the catheter http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 16)03-May-05 17:16:30

Ovid: ICU Book is advanced more than 20 cm (see Ta le 4.4 for the cannulation distances to the right atrium). One maneuver that might resolve this pro lem is to fill the allo on with 1.5 mL fluid (sterile) instead of air (11). Then position the patient wi th the left side down and advance the catheter slowly. The fluid adds mass (weig ht) to the alloon, and this can help the alloon to drop into the right ventric le. The added mass also makes it more difficult for the alloon to e repelled y elevated pressures in the right ventricle. When the right ventricle is entered , the fluid should e removed and the alloon should e reinflated with air. Catheter Will Not Advance into the Pulmonary Artery This pro lem is usually caused y coiling of the catheter in the right ventricle . This pro lem can e resolved y simply withdrawing the catheter into the super ior vena cava and readvancing the catheter. When advancing the catheter through the right ventricle, avoid rapid thrusts. Use slower, more continuous motions to advance the catheter. This allows the catheter to float along with the flow of lood into the pulmonary artery. If repeated attempts to advance the catheter ar e unsuccessful, consider giving a olus of calcium sulfate through the distal po rt of the catheter to stimulate ventricular contraction. The only reported exper ience with this maneuver is limited and anecdotal; that is, it has een successf ul in each of three occasions when used y this author. Arrhythmias Atrial and ventricular arrhythmias are common during placement of PA catheters ( arrhythmias appeared in over 50% of catheter placements in one report) (2). Howe ver, the arrhythmias are almost never harmful, and treatment is rarely necessary . If an arrhythmia is noted, withdraw the catheter into the vena cava, and the a rrhythmia should disappear. The only rhythm distur ances that warrant immediate intervention are complete heart lock (which should e treated with a temporary transvenous pacemaker) and sustained ventricular tachycardia P.159 (which should e treated with lidocaine or other suita le antiarrhythmic agent). Fortunately, these malignant arrhythmias are rarely encountered during PA catheter insertion . Una le to O tain a Wedge Pressure In a out one-fourth of PA catheter insertions, the pulsatile PA pressure does no t disappear even when the catheter is advanced the maximum distance along the pu lmonary artery. This finding may e caused y nonuniform alloon inflation, ut the true cause is unknown. When this occurs, the pulmonary artery diastolic pres sure can e used as a su stitute for the pulmonary capillary wedge pressure (the two pressures should e the same in all patients http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 16)03-May-05 17:16:30

Ovid: ICU Book except those with pulmonary hypertension). Reinsertion of a new PA catheter is n ot necessary unless the patient has high PA pressures and the physician wants to determine whether the pro lem is left heart failure. HEMODYNAMIC PARAMETERS The most appealing feature of the PA catheter is its a ility to generate a large num er of (measured and derived) hemodynamic varia les. There are 10 different parameters of cardiovascular performance and 4 parameters of systemic oxygen tra nsport. Each of the physiologic varia les generated y the PA catheter is descri ed in detail in Chapter 1 and Chapter 2 (see Ta le 1.1 and Ta le 2.4).

BODY SURFACE AREA Hemodynamic varia les are often expressed in relation to ody size. Instead of m ass (weight), the index of ody size that is used is the ody surface area (BSA) , which incorporates oth height (Ht) and weight (Wt). The BSA is o tained y us ing standard nomograms, or is calculated using a cum ersome equation known as th e DuBois formula (see Appendix). Either method can e replaced y using this sim ple equation (12): The BSA derived y this formula has a 99% correlation with the BSA derived y th e standard DuBois formula (13). The average-sized adult has a BSA of 1.6 to 1.9 m2. CARDIOVASCULAR PERFORMANCE The parameters of cardiovascular performance are shown in Ta le 10.1 (see also T a le 1.1). Those expressed in relation to ody surface P.160 area (BSA) are deno ted y the term index (e.g., when cardiac output is expressed relative to BSA, i t is called the cardiac index). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 16)03-May-05 17:16:30

Ovid: ICU Book TABLE 10.1. PARAMETERS OF CARDIOVASCULAR PERFORMANCE Central Venous Pressure This pressure is recorded from the proximal port of the PA catheter situated in the superior vena cava or right atrium. Central venous pressure (CVP) is equal t o the pressure in the right atrium. The right-atrial pressure (RAP) should e eq uivalent to right-ventricular end-diastolic pressure (RVEDP) when there is no o struction etween the right atrium and ventricle. Pulmonary Capillary Wedge Pressure This pressure is measured as descri ed in the last section. Because the PCWP is measured when there is no flow etween the catheter tip and the left atrium ( ec ause the alloon on the PA catheter tip is inflated), the PCWP should e the sam e as the left-atrial pressure (LAP). The LAP should also e equivalent to the le ft-ventricular end-diastolic pressure (LVEDP) when there is no o struction etwe en left atrium and ventricle. Cardiac Index The thermistor at the distal end of the PA catheter provides a measure of cardia c output (CO) y recording the change in temperature P.161 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 16)03-May-05 17:16:31

Ovid: ICU Book of lood flowing in the pulmonary artery when the lood temperature is reduced y injecting a volume of cold fluid through the proximal port of the PA catheter in the right atrium. This is known as the thermodilution technique, and it measu res cardiac output as a mean volumetric lood flow. The cardiac output is expres sed as the cardiac index (CI) when it is divided y the BSA. Stroke Volume The stroke volume index (SVI) is the volume ejected y the ventricles during sys tole. It is easily derived as the CI divided y the heart rate (HR). Right-Ventricular Ejection Fraction The ejection fraction is the fraction of ventricular volume that is ejected duri ng systole. The ejection fraction of the right ventricle (RVEF) can e measured with a specialized PA catheter that is equipped with a rapid-response thermistor . The resulting measurement is an index of the ratio etween the stroke volume ( SV) and the right-ventricular end-diastolic volume (RVEDV). Right-Ventricular End-Diastolic Volume When the RVEF is measured, the RVEDV can e determined y rearranging the formul a for RVEF shown a ove. Left-Ventricular Stroke Work Index The left-ventricular stroke work index (LVSWI) is the work performed y the vent ricle to eject the stroke volume into the aorta. Work is determined y force or pressure (Mean arterial pressure PCWP) and the corresponding mass or volume (SV) that is moved. The factor 0.0136 converts pressure and volume to units of work. Right-Ventricular Stroke Work Index The right-ventricular stroke work index (RVSWI) is the work needed to move the s troke volume http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 16)03-May-05 17:16:31

Ovid: ICU Book across the pulmonary circulation. It is P.162 derived as the pressure developed y the right ventricle during systole (Pulmonary artery pressure CVP) to eject t he stroke volume. Systemic Vascular Resistance Index The systemic vascular resistance index (SVRI) is the vascular resistance across the whole of the systemic circulation. It is proportional to the pressure gradie nt from the aorta to the right atrium (MAP CVP), and is inversely related to lo od flow (CI). (The factor of 80 is necessary to convert units.) Pulmonary Vascular Resistance Index The pulmonary vascular resistance index (PVRI) is proportional to the pressure g radient across the entire lungs, from the pulmonary artery (PAP) to the left atr ium (LAP). Because the wedge pressure (PCWP) is equivalent to the LAP, the press ure gradient across the lungs can e expressed as (PAP PCWP). The PVRI can then e derived using the following equation.

SYSTEMIC OXYGEN TRANSPORT The parameters of systemic oxygen transport are shown in Ta le 10.2 (see also Ta le 2.4). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 16)03-May-05 17:16:31

Ovid: ICU Book TABLE 10.2. OXYGEN TRANSPORT PARAMETERS Oxygen Delivery This is the rate of oxygen transport in arterial P.163 lood, and is the product of the cardiac output and the oxygen concentration in arterial lood. The arter ial O2 delivery (DO2) is defined y Equation 10.12, where H is hemoglo in and S aO2 is the arterial O2 saturation (see Chapter 2 for the derivation of this equa tion). Mixed Venous Oxygen Saturation The oxygen saturation in pulmonary artery (mixed venous) lood can e monitored continuously with a specialized PA catheter, or it can e measured in vitro with a lood sample o tained from the distal port of the catheter. (See Chapter 22 f or a description of the O2 saturation measurement.) The SVO2 varies inversely wi th the amount of oxygen extracted from the peripheral microcirculation; that is, SVO2 = 1/O2 extraction. Oxygen Uptake Oxygen uptake (VO2) is the rate of oxygen taken up from the systemic microcircul ation and is the product of the cardiac output and the difference in oxygen conc entration etween arterial and mixed venous lood. The formula for O2 uptake sho wn elow is derived in Chapter 2. Oxygen Extraction Ratio The oxygen extraction ratio (O2ER) is the fractional uptake of oxygen from the s ystemic microcirculation, and is equivalent to the ratio etween O2 delivery and O2 uptake. This ratio can e multiplied y 100 to express it as a percentage. HEMODYNAMIC PROFILES The parameters just descri ed can e organized to provide profiles of specific a spects of http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 16)03-May-05 17:16:31

Ovid: ICU Book cardiovascular performance. Some examples of hemodynamic profiles are presented elow.

Right Heart Failure High RAP Low CI High PVRI Left Heart Failure High PCWP Low CI High SVRI HYPOTENSION The mean arterial pressure is a function of the cardiac output and the systemic vascular resistance: MAP = CI SVRI. The cardiac output, in turn, depends on the venous return. If the CVP is used as an index of venous return, then there are t hree varia les that can e used to evaluate the source of hypotension: CVP, CI, and SVRI. The following are profiles of these varia les in the three classic typ es of hypotension. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 16)03-May-05 17:16:31

HEART FAILURE The performance of the right and left sides of the heart can e evaluated y the relationship etween the ventricular filling pressure, the cardiac output (or s troke volume), and the outflow vascular resistance. The following are examples o f profiles expected in severe heart failure. P.164

Ovid: ICU Book Hypovolemic Low CVP Low CI High SVRI Cardiogenic High CVP Low CI High SVRI Vasogenic Low CVP High CI Low SVRI CLINICAL SHOCK Clinical shock syndromes are all characterized y inadequate tissue oxygenation. The parameters of systemic oxygen transport provide an indirect evaluation of p eripheral (tissue) oxygenation, and thus can help to identify a shock state. An example of how oxygen transport parameters can identify a shock state is shown elow. Heart Failure High CVP Low CI High SVRI Normal VO2 Cardiogenic Shock High CVP Low CI High SVRI Low VO2 Without the VO2 measurement in the a ove profiles, it is impossi le to different iate a low cardiac output state from cardiogenic shock. This illustrates how oxy gen transport monitoring can e used to determine the consequences of hemodynami c a normalities on peripheral http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 16)03-May-05 17:16:31

Ovid: ICU Book

COMPUTERIZED PROFILES The numerous calculations involved in creating hemodynamic profiles make compute r implementation well-suited to this task (14,15). A computer program written in BASIC that generates 10 hemodynamic P.165 parameters, including ody surface ar ea, is availa le from this author on request: REFERENCES GENERAL TEXTS Daily EK, Schroeder JS. Techniques in edside hemodynamic monitoring. 5th ed. St . Louis: CV Mos y, 1994. Darovic GO, ed. Hemodynamic monitoring: invasive and noninvasive clinical applic ation. 2nd ed. Philadelphia: WB Saunders, 1995. REVIEWS 1. Swan HJ. The pulmonary artery catheter. Dis Mon 1991;37:473543. 2. Ermakov S, Hoyt JW. Pulmonary artery catheterization. Crit Care Clin 1992;8:7 73806 (115 References). 3. American Society of Anesthesiologists Task Force on Pulmonary Artery Catheter ization. Practice guidelines for pulmonary artery catheterization. Anesthesiolog y 1993;78:380394 (89 References). 4. Darovic GO. Pulmonary artery pressure monitoring. In: Darovic GO, ed. Hemodyn amic monitoring: invasive and noninvasive clinical application. 2nd ed. Philadel phia: WB http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 16)03-May-05 17:16:31

oxygenation. The uses and limitations of oxygen transport monitoring are descri ed in more detail in Chapter 13.

Ovid: ICU Book Saunders, 1995;253322 (84 References). SELECTED REFERENCES 5. I erti TJ, Fischer EP, Lei owitz AB, et al. A multicenter study of physicians ' knowledge of the pulmonary artery catheter. JAMA 1990;264:29282932. 6. Halpern N, Feld H, Oropello JM, Stern E. The technique of inserting an RV por t PA catheter and pacing pro e. J Crit Illness 1991;6:11531159. 7. Armaganidis A, Dhainaut JF, Billard JL, et al. Accuracy assessment for three fi eroptic pulmonary artery catheters for SvO2 monitoring. Intensive Care Med 19 94;20:484488. 8. Vincent JL, Thirion M, Brimioulle S, et al. Thermodilution measurement of rig ht ventricular ejection fraction with a modified pulmonary artery catheter. Inte nsive Care Med 1986;12:3338. 9. Yelderman M, Ramsay MA, Quinn MD, et al. Continuous thermodilution cardiac ou tput measurement in intensive care unit patients. J Cardiothorac Vasc Anesth 199 2;6:270 274. 10. Amin DK, Shah PK, Swan HJC. The technique of inserting the SwanGanz catheter. J Crit Illness 1993;8:11471156. 11. Venus B, Mathru M. A maneuver for edside pulmonary artery catheterization i n patients with right heart failure. Chest 1982;82:803804. 12. Jaco son B. Medicine and clinical engineering. Englewood Cliffs, NJ: Prentic e Hall, 1977;388. 13. Mattar JA. A simple calculation to estimate ody surface area in adults and its correlation with the Du ois formula. Crit Care Med 1989;846847. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 16)03-May-05 17:16:31

Ovid: ICU Book 14. Marino PL, Krasner J. Computerized interpretation of hemodynamic profiles in the ICU. Crit Care Med 1984;12:601602. 15. Vyskocil JJ, Kruse JA. Hemodynamics and oxygen transport: using your compute r to manage data. J Crit Illness 1994;9:447459. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 16)03-May-05 17:16:31

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 11 CENTRAL VENOUS PRESSURE AND WEDGE PRESSURE It is what we think we know already that often prevents us from learning. --Clau de Bernard Monitoring central venous pressure (CVP) and pulmonary artery occlusi on (wedge) pressure are routine practices in critical care (1,2 and 3). Like all familiar practices, these measurements are not often scrutinized. As a result, they are often misinterpreted (4,5). Attention to the material in this chapter w ill help reduce errors in interpretation of these two measurements. SOURCES OF VARIABILITY BODY POSITION The zero reference point for venous pressures in the thorax is a point on the ex ternal thorax where the fourth intercostal space intersects the midaxillary line (i.e., the line midway etween the anterior and posterior axillary folds). This point (called the phle ostatic axis) corresponds to the position of the right a nd left atrium when the patient is in the supine position. It is not a valid ref erence point in the lateral position, so CVP and wedge pressure should not e re corded when patients are placed in lateral positions (6). CHANGES IN THORACIC PRESSURE The vascular pressure recorded at the edside is an intravascular pressure; i.e. , the pressure in the vessel lumen relative to atmospheric (zero) pressure. Howe ver, the vascular pressure that determines ventricular preload (stretch on the v entricular muscle) and the rate of P.167 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 15)03-May-05 17:19:54

Ovid: ICU Book edema formation is transmural pressure; i.e., the difference etween the intrava scular and extravascular pressures. Changes in thoracic pressures can cause a di screpancy etween intravascular and transmural pressures. This discrepancy is il lustrated y the respiratory variations in the CVP tracing shown in Figure 11.1. The intravascular pressure changes in this tracing are caused y respiratory va riations in intrathoracic pressure that are transmitted into the lumen of the su perior vena cava. If the changes in intrathoracic pressure are completely transm itted across the wall of the vessel, the transmural pressure remains constant th roughout the respiratory cycle. (Because it is not possi le to determine how muc h of the change in thoracic pressure is transmitted into the lood vessel in any individual patient, it is not possi le to determine whether transmural pressure is a solutely constant.) Thus, changes in intravascular pressures in the thorax may not reflect physiologically important (transmural) pressure changes (7). Figure 11.1. Respiratory variations in central venous pressure. Intravascular pr essure is changing, ut transmural pressure may e constant throughout the respi ratory cycle. End-Expiration Intravascular pressures should e equivalent to transmural pressures when the ex travascular pressure is zero. In the thorax, the extravascular pressure should e close to zero (i.e., atmospheric pressure) at the end of expiration. Therefore , intravascular pressures in the thorax should e measured at the end of expirat ion (1,2 and 3,7). The intrathoracic pressure at the end of expiration is positi ve relative to atmospheric pressure in patients who are actively exhaling (e.g., y grunting) and in the presence of positive end-expiratory pressure (PEEP) (8) . In the latter situation, the vascular pressures should e measured while the p atient http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 15)03-May-05 17:19:54

Ovid: ICU Book is riefly removed from PEEP, or the level of PEEP should e su tracted from the pressures measured at end-expiration. The influence of active exhalation on end -expiratory pressures cannot e determined in individual patients without induci ng muscle paralysis. Pressure Monitors If the oscilloscope display screens in the ICU are equipped with horizontal grid s, the CVP and wedge pressures should e measured P.168 directly from the pressu re tracings on the screen. The measurements o tained with this method are more a ccurate than digitally displayed pressures (9). If only digital pressure reading s are availa le, the systolic pressure should e used in patients who are reath ing spontaneously and the diastolic pressure should e used in patients receivin g positivepressure mechanical ventilators. The reason is that the digital displa y on most pressure monitors represents the pressure measured over specific time intervals (usually 4 seconds, or the time for one sweep across the oscilloscope screen). The systolic pressure is the highest pressure, the diastolic pressure i s the lowest pressure, and the mean pressure is the integrated area under the pr essure wave in each time period (Fig. 11.1). During spontaneous reathing, the p ressure at the end of expiration is the highest pressure (i.e., systolic pressur e), and during positive-pressure mechanical ventilation, the end-expiratory pres sure is the lowest pressure (i.e., diastolic pressure). Therefore, systolic pres sure should e used as the endexpiratory vascular pressure in patients who are reathing spontaneously, whereas diastolic pressure should e used in patients re ceiving mechanical ventilation. The mean pressure should never e used as a refl ection of transmural pressure when there are respiratory variations in intravasc ular pressure (1,2 and 3,7). SPONTANEOUS VARIATIONS Like any physiologic varia le, vascular pressures in the thorax can vary spontan eously, without a change in the clinical condition of the patient. The spontaneo us variation in wedge pressure is 4 mm Hg or less in 60% of patients, ut it can e as high as 7 mm Hg in any individual patient (10). In general, a change in C VP or wedge pressure of less than 4 mm Hg should not e considered a clinically significant change. MANOMETER PRESSURES Most vascular pressures are measured with electronic pressure transducers that r ecord the pressure in millimeters of mercury (mm Hg). An alternative method of m easuring pressure (usually reserved for CVP) is water manometry, where pressure is recorded in cm H2O (11). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 15)03-May-05 17:19:54

Ovid: ICU Book Because mercury is 13.6 times more dense than water (see Appendix 1), pressures measured in cm H2O are divided y 1.36 to convert the units to mm Hg: The pressure correlations in cm H2O and mm Hg are shown in Ta le 11.1. TABLE 11.1. CONVERSION OF cm H2O to mm Hg* P.169 WEDGE PRESSURE Few pressures in the ICU are misinterpreted as frequently, and as consistently, as pulmonary capillary wedge pressure (4,5,12). Pro a ly the most important feat ure of the wedge pressure is what it is not: Wedge pressure is not left-ventricu lar preload. Wedge pressure is not pulmonary capillary hydrostatic pressure. Wed ge pressure is not a relia le measure for differentiating cardiogenic from nonca rdiogenic pulmonary edema. These limitations are explained in the description of wedge pressure that follows. WEDGE PRESSURE TRACING http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 15)03-May-05 17:19:54

Ovid: ICU Book When the pulmonary artery catheter is properly positioned, inflation of the all oon at the tip of the catheter causes the pulsatile pressure to disappear. This is demonstrated in Figure 11.2. As mentioned, the P.170 nonpulsatile pressure cr eated y alloon inflation is considered to e the pressure in the pulmonary mic rocirculation; hence it is called pulmonary capillary wedge pressure (PCWP). The wedge pressure shown in Figure 11.2 is lower than the pulmonary artery diastoli c pressure ecause the pressure tracing is from a patient with pulmonary hyperte nsion. In the a sence of pulmonary artery hypertension, the wedge pressure is us ually within a few mm Hg of the pulmonary artery diastolic pressure (13). (The i nflection point in Figure 11.2, indicated as the possi le hydrostatic pressure, is explained later.) Figure 11.2. Pressure tracing showing the transition from a pulsatile pulmonary artery pressure to a alloon occlusion (wedge) pressure. The inflection point ma y represent capillary hydrostatic pressure. PRINCIPLE The rationale for the wedge pressure measurement is illustrated in Figure 11.3 ( 13). Inflation of the alloon at the tip of pulmonary artery catheter o structs lood flow and creates a static column of lood etween the catheter tip and the left atrium. In this situation, the pressure at the tip of the pulmonary artery catheter should e the same as the pressure in the left atrium.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 15)03-May-05 17:19:54

Ovid: ICU Book Figure 11.3. The principle of the wedge pressure measurement. The lung is divide d into three zones ased on the relationship etween alveolar pressure (PA), (me an) pulmonary artery pressure (Pa), and pulmonary capillary pressure (Pc). Wedge pressure is an accurate reflection of left-atrial pressure (PLA) only in zone 3 , where Pc is greater than PA. Q = pulmonary lood flow. This condition is expressed y the hydraulic equation elow (where Pc is capilla ry pressure, PLA is left-atrial pressure, Q is pulmonary lood flow, and Rv is p ulmonary venous resistance). P.171 Thus, alloon inflation allows the pulmonary artery catheter to measure the pressure in the left atrium. Because left-atrial pressure is normally the equivalent of left-ventricular end-diastolic pressure ( LVEDP), the pulmonary capillary wedge pressure can e used as a measure of leftv entricular filling pressure. What the wedge pressure actually measures is the fo cus of the remainder of this chapter. PULMONARY CAPILLARY WEDGE PRESSURE AS PRELOAD The wedge pressure is often used as a reflection of left-ventricular filling dur ing diastole (i.e., ventricular preload). In Chapter 1, preload was defined as t he force that stretched a muscle at http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 15)03-May-05 17:19:54

Ovid: ICU Book rest, and the preload for the intact ventricle was identified as end-diastolic v olume (EDV). However, wedge pressure is a measure of end-diastolic pressure, and end-diastolic pressure may not e an accurate reflection of preload (EDV) when the compliance (distensi ility) of the ventricle is a normal (see Fig. 1.1). The refore, wedge pressure is a reflection of leftventricular preload only when comp liance of the ventricle is normal or constant (13,14). Several conditions can al ter ventricular compliance in patients in the ICU, such as ventricular hypertrop hy, positive pressure ventilation, myocardial ischemia, and myocardial edema (e. g., after cardiopulmonary ypass surgery). Therefore, wedge pressure may not e a relia le index of left-ventricular preload in a large num er of patients in th e ICU. PULMONARY CAPILLARY WEDGE PRESSURE AS LEFTATRIAL PRESSURE The following conditions can influence the accuracy of the wedge pressure as a m easure of left-atrial pressure. Lung Zones If the pressure in the surrounding alveoli exceeds capillary (venous) pressure, the pressure at the tip of the pulmonary artery (PA) catheter may reflect the al veolar pressure more than the pressure in the left atrium. This is illustrated i n Figure 11.3. The lung in this figure is divided into three zones ased on the relationship etween alveolar pressure and the pressures in the pulmonary circul ation (1,2 and 3,13). The most dependent lung zone (zone 3) is the only region w here capillary (venous) pressure exceeds alveolar pressure. Therefore, wedge pre ssure P.172 is a reflection of left-atrial pressure only when the tip of the pul monary artery catheter is located in zone 3 of the lung. Catheter Tip Position Although the lung zones shown in Figure 11.3 are ased on physiologic rather tha n anatomic criteria, the lung regions elow the left atrium are considered to e in lung zone 3 (1,2 and 3). Therefore, the tip of the pulmonary artery catheter should e positioned elow the level of the left atrium to ensure that the wedg e pressure is measuring left-atrial pressure. Most PA catheters are advanced int o lung regions elow the level of the left atrium ( ecause of the higher lood f low in dependent lung regions). However, as many as 30% of PA catheters are posi tioned with their tips a ove the level of the left atrium (13). When patients ar e supine, http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 15)03-May-05 17:19:54

Ovid: ICU Book routine porta le (anteroposterior) chest x rays cannot e used to identify the c atheter tip position relative to the left atrium. Rather, a lateral view of the chest is needed. However, lateral films have not gained favor for this use in mo st ICUs, pro a ly ecause they are too time-consuming for the small percentage o f improperly located catheter tips that will e revealed. Instead, catheter tips can e assumed to e positioned in zone 3 of the lung in all ut the following conditions: when there are marked respiratory variations in the wedge pressure, and when PEEP is applied and wedge pressure increases y 50% or more of the appl ied PEEP (13). Positive End-Expiratory Pressure The presence of PEEP can reduce the area of zone 3 in the lung. In fact, when PE EP is com ined with a low wedge pressure, there may e no zone 3 conditions in t he lung, even in the most dependent lung regions. When this occurs, the wedge pr essure is not an accurate reflection of left-atrial pressure, even when the cath eter tip is elow the level of the left atrium (13). Therefore, when PEEP is ei ng applied, the wedge pressure should e measured while PEEP is temporarily disc ontinued (if this can e done without causing dangerous decreases in arterial ox ygenation). PEEP can also e generated internally y patients who have inadequat e emptying of the lungs during expiration (see Chapter 28). This type of intrins ic or auto-PEEP is common in patients with o structive lung disease, particularl y when they are reathing rapidly or receiving large inflation volumes during me chanical ventilation. A edside maneuver that can help detect auto-PEEP is prese nted in Chapter 28. Wedged Blood Gases As many as 50% of the nonpulsatile pressures produced y alloon inflation repre sent damped PA pressures rather than pulmonary capillary pressures (15). Aspirat ion of lood from the catheter tip during alloon inflation can e used to ident ify a true wedge (capillary) pressure P.173 using the three criteria shown in Ta le 11.2. Although this is a cum ersome practice that is not used routinely, it seems justified when making important diagnostic and therapeutic decisions ased on the wedge pressure measurement. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 15)03-May-05 17:19:54

Ovid: ICU Book TABLE 11.2. CRITERIA FOR WEDGE PRESSURE VALIDATION PULMONARY CAPILLARY WEDGE PRESSURE AS LEFTVENTRICULAR END-DIASTOLIC PRESSURE Even when wedge pressure is an accurate reflection of left-atrial pressure, ther e may e a discrepancy etween wedge (left-atrial) pressure and LVEDP. This can occur under the following conditions (13). Aortic insufficiency: LVEDP can e higher than PCWP ecause the mitral valve closes prematurely while retrograde flow continues to fill the ventricle. Noncompliant ventricle: Atrial contraction against a stiff ventricle produces a rapid rise in end-diasto lic pressure that closes the mitral valve prematurely. The result is a PCWP that is lower than LVEDP. Respiratory failure: PCWP can exceed LVEDP in patients with pulmonary disease. The presumed mechanism is constriction of small veins in lung regions that are hypoxic (17). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 15)03-May-05 17:19:54

Ovid: ICU Book PULMONARY CAPILLARY WEDGE PRESSURE AS A HYDROSTATIC PRESSURE Wedge pressure is often assumed to e a measure of hydrostatic pressure in the p ulmonary capillaries. The pro lem with this assumption is that the wedge pressur e is measured in the a sence of lood flow. When the alloon is deflated and lo od flow resumes, the pressure in the pulmonary capillaries is equivalent to the left-atrial (wedge) pressure only when hydraulic resistance in the pulmonary vei ns is negligi le. This is expressed elow, where Pc is capillary hydrostatic pre ssure, Rv is the hydraulic resistance in the pulmonary veins, Q is lood flow, a nd wedge pressure (PCWP) is su stituted for left-atrial pressure. P.174 Pulmonary Venous Resistance Unlike the systemic veins, the pulmonary veins contri ute a significant fraction to the total vascular resistance across the lungs. (This is a reflection more o f a low resistance in the pulmonary arteries than of a high resistance in the pu lmonary veins.) As shown in Figure 11.4, 40% of the pressure drop across the pul monary circulation occurs on the venous side of the circulation, which means tha t the pulmonary veins contri ute 40% of the total resistance in the pulmonary ci rculation (16). Although this is derived from animal studies, the contri ution i n humans is pro a ly similar in magnitude. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 15)03-May-05 17:19:54

Ovid: ICU Book Figure 11.4. The distinction etween capillary hydrostatic pressure (Pc) and wed ge pressure (PCWP). When the alloon is deflated and flow (Q) resumes, Pc and PC WP are equivalent only when the hydraulic resistance in the pulmonary veins (Rv) is negligi le. Pa = pulmonary artery pressure. If the pulmonary venous resistan ce (Rv) is greater than zero, the capillary hydrostatic pressure (Pc) will e hi gher than the wedge pressure. The contri ution of the hydraulic resistance in the pulmonary veins may e even greater in critically ill patients ecause several conditions that are common in patients in the ICU can promote pulmonary venoconstriction. These conditions in clude hypoxemia, endotoxemia, and the acute respiratory distress syndrome (17,18 ). These conditions further magnify differences etween wedge pressure and capil lary hydrostatic pressure, as demonstrated elow. WedgeHydrostatic Pressure Conversion Equation 11.4 can e used to convert wedge pressure (PCWP) to pulmonary capillar y hydrostatic pressure (Pc). This conversion is ased on the assumption that the pressure drop from the pulmonary capillaries to the left atrium (Pc PLA) repres ents 40% of the pressure drop from the pulmonary arteries to the left atrium (Pa PLA). Su stituting wedge pressure for left-atrial pressure (i.e., PLA = PCWP) y ields the following relationship: P.175 For a normal (mean) pulmonary artery pressure of 15 mm Hg and a wedge pres sure of 10 mm Hg, this relationship predicts the following: Thus, in the normal lung, wedge pressure is equivalent to capillary hydrostatic pressure. However, in the presence of pulmonary venoconstriction and pulmonary h ypertension (e.g., in acute respiratory distress syndrome, ARDS), there can e a considera le difference etween wedge pressure and capillary hydrostatic pressu re. The example elow is ased on a mean PA pressure of 30 mm Hg and a venous re sistance that is 60% of the total pulmonary vascular resistance. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 15)03-May-05 17:19:54

Ovid: ICU Book Unfortunately, pulmonary venous resistance cannot e measured in critically ill patients, and this limits the accuracy of the wedge pressure as a measure of cap illary hydrostatic pressure. OCCLUSION PRESSURE PROFILE The transition from pulsatile pulmonary artery pressure to nonpulsatile wedge pr essure in Figure 11.2 shows an initial rapid phase followed y a slower, more gr adual pressure change. The initial rapid phase may represent the pressure drop a cross the pulmonary arteries, whereas the slower phase represents the pressure d rop across the pulmonary veins. If this is the case, the inflection point markin g the transition from the rapid to the slow phase represents the capillary hydro static pressure. Although this method can provide a more definitive determinatio n of the capillary hydrostatic pressure than the equations used a ove, an inflec tion point is not often recogniza le (19,20). SUMMARY There are numerous sources of error in the interpretation of CVP and wedge press ure. Fortunately, with the a ility to monitor cardiac P.176 output (Chapter 12) and systemic oxygen transport (Chapter 13), these pressures have ecome less imp ortant in the evaluation of hemodynamic status. REFERENCES GENERAL TEXTS Ahrens TS, Taylor LA. Hemodynamic waveform analysis. Philadelphia: WB Saunders, 1992. Daily EK, Schroeder JS. Hemodynamic waveforms. 2nd ed. St. Louis: CV Mos y, 1990 . http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 15)03-May-05 17:19:54

Ovid: ICU Book Darovic GO, ed. Hemodynamic monitoring. Invasive and noninvasive clinical applic ations. 2nd ed. Philadelphia: WB Saunders, 1995. REVIEWS 1. Schwenzer KJ. Venous and pulmonary pressures. In: Lake C, ed. Clinical monito ring. Philadelphia: WB Saunders, 1990;147198 (166 References). 2. Bridges EJ, Woods SL. Pulmonary artery pressure measurement: state of the art . Heart Lung 1993;22:99111 (62 References). 3. Darovic GO. Pulmonary artery pressure monitoring. In: Darovic GO, ed. Hemodyn amic monitoring. Invasive and noninvasive clinical applications. 2nd ed. Philade lphia: WB Saunders, 1995:253322 (71 References). SELECTED REFERENCES 4. Nadeau S, No le WH. Misinterpretation of pressure measurements from the pulmo nary artery catheter. Can Anesth Soc J 1986;33:352363. 5. Komandina KH, Schenk DA, LaVeau P, et al. Intero server varia ility in the in terpretation of pulmonary artery catheter pressure tracings. Chest 1991;100:16471 654. 6. Kee LL, Simonson JS, Stotts NA, et al. Echocardiographic determination of val id zero reference levels in supine and lateral positions. Am J Crit Care 1993;2: 7280. 7. Schmitt EA, Brantigen CO. Common artifacts of pulmonary artery and pulmonary artery wedge pressures. Recognition and management. J Clin Monit 1986;2:4452. 8. Pinsky M, Vincent J-L, De Smet J-M. Estimating left ventricular filling press ure during positive end-expiratory pressure in humans. Am Rev Respir Dis 1991;14 3:2531. 9. Do in K, Wallace S, Ahl erg J, Chulay M. Pulmonary artery pressure measureme nt http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 15)03-May-05 17:19:54

Ovid: ICU Book

10. Nemens EJ, Woods SL. Normal fluctuations in pulmonary artery and pulmonary c apillary wedge pressures in acutely ill patients. Heart Lung 1982;11:393398. 11. Halck S, Walther-Larsen S, Sanchez R. Relia ility of central venous pressure measured y water column. Crit Care Med 1990;18:461462. 12. Morris AH, Chapman RH, Gardner RM. Frequency of wedge pressure errors in the ICU. Crit Care Med 1985;13:705708. P.177 13. O'Quin R, Marini JJ. Pulmonary artery occlusion pressure: clinical physiolog y, measurement, and interpretation. Am Rev Respir Dis 1983;128:319326. 14. Die el LN, Wilson RF, Tagett MG, Kline RA. End-diastolic volume. A etter in dicator of preload in the critically ill. Arch Surg 1992;127:817822.

16. Michel RP, Hakim TS, Chang HK. Pulmonary arterial and venous pressures measu red with small catheters. J Appl Physiol 1984;57:309314. 17. Tracey WR, Hamilton JT, Craig ID, Paterson NAM. Effect of endothelial injury on the responses of isolated guinea pig pulmonary venules to reduced oxygen ten sion. J Appl Physiol 1989;67:21472153. 18. Kloess T, Birkenhauer U, Kottler B. Pulmonary pressureflow relationship and p eripheral oxygen supply in ARDS due to acterial sepsis. Second Vienna Shock For um, 1989:175180. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 15)03-May-05 17:19:54

15. Morris AH, Chapman RH. Wedge pressure confirmation y capillary lood. Crit Care Med 1985;13:756759.

y aspiration of pulmonar

in patients with elevated pressures: effect of measurement. Am J Crit Care 1992;2:6169.

ackrest elevation and method of

Ovid: ICU Book 19. Cope DK, Grim ert F, Downey JM, et al. Pulmonary capillary pressure: a revie w. Crit Care Med 1992;20:10431056. 20. Gil ert E, Hakim TS. Derivation of pulmonary capillary pressure from arteria l occlusion in intact conditions. Crit Care Med 1994;22:986993. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 15)03-May-05 17:19:54

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 12 THERMODILUTION: METHODS AND APPLICATIONS An exact science is dominated y the idea of approximation. --Bertrand Russell A few years after its introduction, the pulmonary artery (PA) catheter was upgrad ed y incorporating a thermistor at the catheter tip to measure lood flow using the thermodilution principle. This single addition to the PA catheter increased its monitoring capacity from 2 parameters (i.e., central venous pressure and we dge pressure) to 10 parameters (see Ta le 10.1 and Ta le 10.2). More recent refi nements in thermodilution methodology have further increased the monitoring capa city of the PA catheter, adding the a ility to measure the ejection fraction of the right ventricle and to monitor cardiac output continuously. This chapter des cri es each of the thermodilution methods for monitoring cardiac performance. BASIC THERMODILUTION METHOD Thermodilution is an indicator-dilution method of measuring lood flow. This met hod is ased on the premise that when an indicator su stance is added to circula ting lood, the rate of lood flow is inversely proportional to the change in co ncentration of the indicator over time. The indicator su stance can e a dye (dy e-dilution method) or a fluid with a different temperature than lood (thermodil ution method). The thermodilution method of measuring cardiac output P.179 with a PA catheter is shown in Figure 12.1 (1,2 and 3). A dextrose or saline solution that is colder than lood is injected through the proximal port of the catheter in the right atrium. The cold fluid mixes with lood in the right heart cham er s, and the cooled lood is ejected into the pulmonary artery and flows past the thermistor on the distal end of the catheter. The thermistor records the change in lood temperature with time and sends this information to an electronic instr ument that records and displays a temperaturetime curve like the one shown in Fig ure 12.1. The area under this curve is inversely proportional to the rate of lo od flow in the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 11)03-May-05 17:19:44

Ovid: ICU Book pulmonary artery. In the a sence of intracardiac shunts, this flow rate is equiv alent to the (average) cardiac output. Figure 12.1. Schematic illustration of the thermodilution method for measuring c ardiac output. THERMODILUTION CURVES Examples of thermodilution curves are shown in Figure 12.2. The low cardiac outp ut curve (upper panel) has a gradual rise and fall, whereas the high output curv e (middle panel) has a rapid rise, an a reviated peak, and a steep downslope. N ote that the area under the low cardiac output curve is greater than the area un der the high output curve; that is, the area under the curves is inversely relat ed to the flow rate. Electronic cardiac monitors integrate the area under the te mperaturetime curves and provide a digital display of the calculated cardiac outp ut. There is a tendency to rely on this digital display of the calculated cardia c output without examining the temperaturetime curve, and this can lead to errors in interpretation. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 11)03-May-05 17:19:44

Ovid: ICU Book Figure 12.2. Thermodilution curves for a low cardiac output (upper panel), a hig h cardiac output (middle panel), and tricuspid insufficiency (lower panel). The sharp inflection in each curve marks the end of the measurement period. CO = car diac output. P.180 TECHNICAL CONSIDERATIONS PATIENT POSITION Cardiac output can e 30% higher in the supine position than in the semierect po sition (4). Therefore, consecutive cardiac output determinations should e perfo rmed with each patient in a uniform position, or the position of the patient sho uld e recorded with each cardiac output determination. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 11)03-May-05 17:19:44

Ovid: ICU Book INJECTING THE INDICATOR Indicator Solution Bolus injection of normal saline (0.9% sodium chloride) or 5% dextrose-in-water produces the most satisfactory measurements (3). Other P.181 injectate solutions can produce varia le results ( ecause of their varying specific heats) and are not recommended. Injectate Volume and Temperature The indicator solution can e cooled in ice or injected at room temperature, and can e administered in a volume of 5 mL or 10 mL. In general, higher-volume, lo wer-temperature injectates produce the highest signal-to-noise ratios, and thus the most accurate measurements (1,2 and 3,5). However, room temperature injectat es (which are less tedious to administer than iced injectates) produce relia le measurements in most critically ill patients (6,7,8 and 9). When the indicator f luid is injected at room temperature, the large (10 mL) injection volume produce s the most relia le results. When using smaller injectate volumes, using iced in jectates increases the relia ility of the measurements. Using small volumes of r oom temperature injectates can yield inaccurate results in low cardiac output st ates (5), and is thus not recommended. Injection Time Optimal results are produced when the olus injection is completed within 2 seco nds (10), ut satisfactory results are o tained with injection times up to 4 sec onds. Longer injection times can produce falsely low measurements. Injection Timing Cardiac output can vary significantly during the respiratory cycle, particularly during mechanical ventilation. Random thermodilution measurements o tained in d ifferent phases of the respiratory cycle can vary y more than 10%, whereas inje ctions that are timed to the end of expiration can reduce the varia ility to 5% (11). This has led to the recommendation that thermodilution cardiac outputs alw ays e recorded at end-expiration. However, it is very difficult to time injecti ons so that the thermodilution curve is recorded at precisely the same time in t he respiratory cycle. In fact, the injection time can e longer than the duratio n of the respiratory cycle in patients with rapid reathing; for example, at a r espiratory rate a ove 15 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 11)03-May-05 17:19:44

Ovid: ICU Book reaths/minute, an injection time of 4 seconds is longer than the duration of ea ch respiratory cycle (less than 4 seconds). It is est in such cases to egin in jecting the indicator solution at the same part of the respiratory cycle for eac h cardiac output measurement.

Alternative Injection Ports If the proximal (right-atrial) port of the PA catheter is o structed, the inject ate can e introduced through an alternative infusion port P.182 on the catheter (if one is availa le) (8) or through the side arm of the introducer catheter (s ee Fig. 10.1) (9). ACCURACY AND RELIABILITY NUMBER OF MEASUREMENTS Serial measurements are recommended for each cardiac output determination. Three measurements are sufficient if they differ y 10% or less. Cardiac output is de termined y averaging the serial measurements. The initial measurement is often falsely elevated (7), so for optimal accuracy, the initial measurement should e discarded. Serial measurements that differ y more than 10% should e considere d unrelia le (12). TRICUSPID REGURGITATION Tricuspid regurgitation causes the cold indicator fluid to e recycled ack and forth across the tricuspid valve. This produces a prolonged, low-amplitude therm odilution curve like the one shown in the lower panel of Figure 12.2. This type of curve is an exaggerated version of a lowoutput curve (with a large area under the curve), so tricuspid regurgitation produces a falsely low thermodilution ca rdiac output (13). Tricuspid regurgitation may e common in mechanically ventila ted patients ecause of the high right-sided cardiac pressures created y positi ve-pressure lung inflations. Therefore, this condition may e a common source of error in thermodilution cardiac outputs in the ICU. LOW-OUTPUT STATES Low cardiac outputs produce low-amplitude temperaturetime curves, and this can af fect the accuracy of thermodilution cardiac outputs y decreasing the signal-tonoise ratio. Accuracy is most affected when room temperature indicator solutions are injected in low (5 mL) volumes. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 11)03-May-05 17:19:44

Ovid: ICU Book In this situation, thermodilution can underestimate cardiac output y as much as 30% (5). In low-output states (cardiac index elow 2.5 L/minute/m2), an iced in jectate gives the most accurate measurements if low injectate volumes are used. INTRACARDIAC SHUNTS Intracardiac shunts produce falsely high thermodilution cardiac output measureme nts. In rightto-left shunts, a portion of the cold indicator fluid passes throug h the shunt, there y creating an a reviated thermodilution curve (similar to th e a reviated high-output curve). In left-to-right shunts, the thermodilution cu rve is a reviated ecause P.183 the shunted lood increases the lood volume in the right heart cham ers, and this dilutes the indicator solution that is injec ted. VARIABILITY Thermodilution cardiac output can vary y as much as 10% without a change in the clinical condition of the patient (14). This means that a aseline cardiac outp ut of 5 L/min can vary spontaneously from 4.5 to 5.5 L/min (or a aseline cardia c index of 3 L/min/m2 can vary from 2.7 to 3.3 L/min/m2) without representing a change in the clinical condition of the patient. Therefore, the thermodilution c ardiac output (or cardiac index) must change y more than 10% for the change to e considered clinically significant.

THERMODILUTION EJECTION FRACTION In the mid-1980s, a PA catheter was introduced with a fast-response thermistor c apa le of measuring the ejection fraction of the right ventricle (15). This adde d the a ility to evaluate right-ventricular function at the edside. METHOD Rapid-response thermistors can record the temperature changes associated with ea ch cardiac cycle. This produces a ramplike thermodilution curve like the one sho wn in Figure 12.3. The change in temperature etween each plateau on the curve i s caused y dilution of the cold indicator fluid y venous lood that fills the ventricle during P.184 diastole. Because the volume that fills the ventricles du ring diastole is equivalent to the stroke volume, the temperature difference T1 T2 is the thermal equivalent of the stroke volume. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 11)03-May-05 17:19:44

Ovid: ICU Book Thus, the points at each end of the temperature change can e taken as the therm al equivalents of the end-diastolic volume (T1) and end-systolic volume (T2), re spectively. Because the ejection fraction is the ratio of stroke volume (SV) to end-diastolic volume (EDV), the right-ventricular ejection fraction (RVEF) can e derived using the appropriate thermal equivalents: Figure 12.3. The thermodilution method for determining ventricular ejection frac tion (EF) using thermal equivalents for end-diastolic volume (EDV), end-systolic volume (ESV), and stroke volume (SV). Normal thermodilution RVEF is 0.45 to 0.50. This is slightly lower than radionuclide RVEF (the gold standard), ut the diffe rence is less than 10% (16). Thermodilution RVEF can e measured relia ly using a room temperature injectate (given as a 10-mL olus) (17).

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 11)03-May-05 17:19:44

End-Diastolic Volume Because the thermodilution PA catheter can measure stroke volume, RVEF can ed to derive the right-ventricular end-diastolic volume (RVEDV):

e us

Ovid: ICU Book This allows for a determination of ventricular preload (end-diastolic volume) at the edside, and ypasses the shortcomings of end-diastolic pressure (e.g., the central venous pressure) as a reflection of preload (18). CONTINUOUS CARDIAC OUTPUT The most recent development in the thermodilution method has led to the introduc tion of a PA catheter that can monitor cardiac output continuously, without the need for intermittent olus injections of indicator fluid (19). This catheter (B axter Edwards Critical Care, Irvine, California) is equipped with a 10-cm therma l filament located 15 to 25 cm from the catheter tip. The filament generates low -energy heat pulses that are transmitted to the surrounding lood. The resulting change in lood temperature is then used to generate a thermodilution curve for determining cardiac output. Although the output is called continuous cardiac ou tput, the measurement is an average cardiac output recorded over 3-minute time i ntervals and updated every 30 to 60 seconds. Continuous thermodilution cardiac o utput monitoring has proven to e oth safe and relia le (19,20). However, it ha s yet to gain widespread popularity, pro a ly ecause of the added cost of the n ewer P.185 technology. Nevertheless, the advantages of a continuous measure of c ardiac output (e.g., more on-line information regarding trends in cardiac output ) seem to justify the additional cost of the technology in patients with unsta l e or severely compromised hemodynamic function. REFERENCES GENERAL TEXTS Ro ertson JIS, Birkenhager WH, eds. Cardiac output measurement. Philadelphia: WB Saunders, 1991. REVIEWS 1. Gardner PE. Cardiac output. Theory, technique, and trou leshooting. Crit Care Nurs Clin North Am 1989;1:577587 (71 References). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 11)03-May-05 17:19:44

Ovid: ICU Book 2. Daily EK, Schroeder JS. Cardiac output measurements. In: Techniques in edsid e hemodynamic monitoring. 5th ed. St. Louis: CV Mos y, 1994;173194 (57 References ). 3. Darovic GO, Yacone-Morton LA. Monitoring cardiac output. In: Darovic GO, ed. Hemodynamic monitoring. Invasive and noninvasive clinical applications. 2nd ed. Philadelphia: WB Saunders, 1995;323346 (27 References). BASIC THERMODILUTION METHOD 4. Driscoll A, Shanahan A, Crommy L, Gleeson A. The effect of patient position o n the reproduci ility of cardiac output measurements. Heart Lung 1995;24:3844. 5. Renner LE, Morton MJ, Sakuma GY. Indicator amount, temperature, and intrinsic cardiac output affect thermodilution cardiac output accuracy and reproduci ilit y. Crit Care Med 1993;21:586597. 6. Nelson LD, Anderson HB. Patient selection for iced versus room temperature in jectate for thermodilution cardiac output determinations. Crit Care Med 1985;13: 182184. 7. Pearl RG, Rosenthal MH, Nielson L, et al. Effect of injectate volume and temp erature on thermodilution cardiac output determinations. Anesthesiology 1986;64: 798801. 8. Pesola GR, Ayala B, Plante L. Room-temperature thermodilution cardiac output: proximal injectate lumen versus proximal infusion lumen. Am J Crit Care 1993;2: 132133. 9. Pesola HR, Pesola GR. Room-temperature thermodilution cardiac output. Central venous versus side port. Chest 1992;103:339341. 10. Conway J, Lund-Johansen P. Thermodilution method for measuring cardiac outpu t. In: Ro ertson JIS, Birkenhager WH, eds. Cardiac output measurement. Philadelp hia: WB Saunders, 1991;1720. 11. Stevens JH, Raffin TA, Mihm FG, et al. Thermodilution cardiac output measure ment. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 11)03-May-05 17:19:44

Ovid: ICU Book Effects of respiratory cycle on its reproduci ility. JAMA 1985;253:22402242. P.18 6

13. Konishi T, Nakamura Y, Morii I, et al. Comparison of thermodilution and Fick methods for measurement of cardiac output in tricuspid regurgitation. Am J Card iol 1992;70:538 540. 14. Sasse SA, Chen PA, Berry RB, et al. Varia ility of cardiac output over time in medical intensive care unit patients. Chest 1994;22:225-232. THERMODILUTION EJECTION FRACTION 15. Vincent JL, Thirion M, Brimioulle S, et al. Thermodilution measurement of ri ght ventricular ejection fraction with a modified pulmonary artery catheter. Int ensive Care Med 1986;12:3338. 16. Vincent JL. The measurement of right ventricular ejection fraction. Intensiv e Care World 1990;7:133136. 17. Safcsak K, Nelson LD. Thermodilution right ventricular ejection fraction mea surements: room temperature versus cold temperature injectate. Crit Care Med 199 4;22:11361141. 18. Die el LN, Wilson RF, Tagett MG, Kline RA. End-diastolic volume. A etter in dicator of preload in the critically ill. Arch Surg 1992;127:817822. CONTINUOUS CARDIAC OUTPUT 19. Yelderman M, Ramsay MA, Quinn MD, et al. Continuous thermodilution cardiac o utput measurement in intensive care unit patients. J Cardiothorac Vasc Anesth http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 11)03-May-05 17:19:44

12. Nadeau S, No le WH. Limitations of cardiac output measurement ion. Can J Anesth 1986;33:780784.

y thermodilut

Ovid: ICU Book 1992;6:270274. 20. Boldt J, Menges T, Woll ruck M, et al. Is continuous cardiac output measurem ent using thermodilution relia le in the critically ill patient? Crit Care Med 1 994;22:19131918. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 11)03-May-05 17:19:44

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 13 TISSUE OXYGENATION No animal can live in an atmosphere where a flame does not urn. --Leonardo da V inci, 1500 Despite good intentions, much of what we do to patients in the name o f aero ic support is without documented need or documented enefit. This is a re flection of the ina ility to o tain a direct measurement of tissue oxygen tensio ns. As a result, decisions a out oxygen are often ased on surrogate measures of tissue oxygenation that are inappropriate (such as arterial lood gases). The i nvasive nature of patient care in the ICU creates approaches to evaluating tissu e oxygenation that are not availa le in other areas of the hospital. This chapte r reviews some of these approaches, along with a more standard marker of tissue oxygen alance (i.e., lood lactate). The oxygen transport varia les in this cha pter are descri ed in more detail in Chapter 2. TISSUE OXYGEN BALANCE The determinants of tissue oxygenation are illustrated in Figure 13.1. The oxyge n supply to tissues is shown as the rate of O2 uptake from the microcirculation (i.e., VO2). The meta olic requirement for oxygen (MRO2) is the rate at which ox ygen is meta olized to water in the mitochondria. Because oxygen is not stored i n tissues, VO2 must match MRO2 for aero ic meta olism to continue undistur ed. W hen this occurs, glucose is completely oxidized, as shown in Figure 13.1, and th e energy yield is 36 moles of ATP (673 kcal) per mole glucose. When VO2 cannot m atch MRO2, a portion of the glucose meta olism is diverted to the production of lactate, with an energy yield of 2 moles ATP (47 kcal) per mole glucose. Thus, w hen the O2 supply is inadequate, the energy yield from su strate meta olism drop s precipitously. This condition, P.188 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 22)03-May-05 17:21:54

Ovid: ICU Book in which the production of ATP is limited y the supply of oxygen, is called dys oxia (1), and when cell dysoxia produces a measura le change in organ function, the condition is commonly known as shock. Figure 13.1. Schematic diagram of the relationship etween oxygen uptake (VO2) a nd the meta olic requirement for oxygen (MRO2). When VO2 matches MRO2, glucose i s completely oxidized to CO2 and water. When VO2 is less than MRO2, glucose form s lactate. Shock is defined as a condition where VO2 is lower than MRO2. OXYGEN TRANSPORT MONITORING As descri ed in Chapter 2, two varia les are used to descri e oxygen transport: the rate of oxygen delivery (DO2) and the rate of oxygen uptake (VO2), also know n as oxygen consumption. These are glo al measures of the oxygen supply (DO2) an d oxygen utilization (VO2) in the systemic circulation. Oxygen uptake (VO2) is e quivalent to the oxygen supply identified in Figure 13.1. Oxygen transport monit oring therefore provides information a out the oxygen supply to tissues (2), ut it provides no information a out the adequacy of tissue oxygenation ( ecause th at requires a measurement of meta olic rate). INTERPRETATIONS The important transport parameter is VO2, which can e interpreted as follows (s ee Ta le 13.1): http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 22)03-May-05 17:21:54

Ovid: ICU Book TABLE 13.1. INDIRECT MEASURES OF TISSUE OXYGEN BALANCE Low VO2: indicates tissue oxygen deficits (oxygen de t) P.189 Normal VO2: A loo d lactate level is required to determine whether tissue oxygenation is adequate. VO2 DEFICIT If a decrease in oxygen uptake is not accompanied y a proportional decrease in meta olic rate, the oxygen supply will e inadequate to support aero ic meta oli sm. Because hypometa olism is uncommon in critically ill patients, a VO2 that fa lls elow the normal range ( elow 100 mL/minute/m2) can e used as evidence of i mpaired tissue oxygenation. An example of a low VO2 that serves as a marker of i nadequate tissue oxygenation is shown in Figure 13.2. The data in this figure ar e taken from a patient who underwent a dominal aortic aneurysm P.190 surgery. At the first postoperative recording (2 hours), VO2 is elow normal, and thereafte r, VO2 never returns to the normal range. At 6 hours, the serum lactate rises a ove the normal range. The lactate levels continue to rise steadily thereafter, r eaching 9 mEq/L at 24 hours http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 22)03-May-05 17:21:54

Ovid: ICU Book after surgery. The progressive rise in lood lactate is evidence that the low VO 2 indicates a widespread tissue oxygen deficit. Note that the cardiac index rema ins normal despite the ongoing ischemia. This highlights the fact that cardiac o utput monitoring does not evaluate tissue oxygenation. Figure 13.2. Serial measurements of cardiac index, oxygen uptake, and lood lact ate levels efore and after a dominal aortic aneurysm surgery. The first postope rative measurement reveals a deficit in oxygen uptake that persists and is accom panied y a progressive rise in lood lactate levels. Dotted lines indicate thre shold values separating normal from a normal measurements. OXYGEN DEBT The area under the dotted line in the VO2 curve reflects the total VO2 deficit o ver time. The cumulative VO2 deficit (derived y integrating the VO2 deficit ove r time) is known as the oxygen de t. Studies of the oxygen de t after resuscitat ion from hemorrhagic shock (3) and in postoperative patients (4) show a direct r elationship etween the magnitude of the oxygen http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 22)03-May-05 17:21:54

Ovid: ICU Book de t and the risk of multiorgan failure and death. These correlations are eviden ce that the VO2 deficit is a marker of tissue ischemia and that early correction of VO2 deficits is warranted to limit the magnitude of the ischemic insult. CORRECTING VO2 DEFICITS The flow diagram in Figure 13.3 shows a management strategy that can e used to correct a VO2 deficit. This approach egins with a focus on lood volume. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 22)03-May-05 17:21:54

Ovid: ICU Book

Step 1: Central Venous Pressure or Wedge Pressure q If low, infuse volume to normalize the filling pressures. q For normal or high pressures, go to Step 2. Correcting volume deficits is essential to maintain cardiac filling volume. Step 2: Cardiac Output q If low, and filling pressures not high, infuse volume until central venous press ure (CVP) is 10 to 12 mm Hg or wedge pressure is 18 to 20 mm Hg. q If low, and filling pressures high, start do utamine infusion at 3 g/kg/minute an d titrate to cardiac index a ove 3.0 L/min/m2. If lood pressure (BP) is low, us e dopamine at starting dose of 5 g/kg/minute. q For cardiac index over 3.0 L/min/m2, go to Step 3. Volume is preferred to adrenergic drugs, so volume is infused to high filling pr essures, if needed. Do utamine is the preferred inotrope P.191 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 22)03-May-05 17:21:54

Figure 13.3. Sequential management of lood volume, , and tissue oxygenation.

lood flow, oxygen transport

Ovid: ICU Book P.192 and is also less thermogenic than the other adrenergic drugs (and thus has less tendency to stimulate meta olism). Step 3: Oxygen Uptake q If VO2 is less than 100 mL/min/m2, use volume (to CVP = 8 to 10 or wedge pressur e = 18 to 20) and inotropic therapy to achieve a cardiac index a ove 4.5 L/min/m 2. Correct anemia if H is less than 8 g/dL. q For VO2 greater than 100 mL/min/m2, go to Step 4. If VO2 does not readily increase when volume is adequate and cardiac index is hi gh, the prognosis is poor. Correcting anemia usually does not increase VO2, ut it can e used as a last resort. When VO2 is normal, the lactate is used to dete rmine whether VO2 is matched to the meta olic rate. Step 4: Blood Lactate Level q If lactate is greater than 4 mmol/L and other signs of shock (organ failure, low BP) are present, options include decreasing meta olic rate (through sedation or stopping feedings) and increasing VO2 a ove 160 mL/min/m2 (if possi le). q For lactate elow 4 mmol/L, o serve. An elevated lactate indicates that VO2 is less than the meta olic rate, so the a pproach is to either decrease the meta olic rate or increase VO2. Achieving a su pranormal VO2 is difficult to accomplish and can lead to unwanted cardiac and me ta olic stimulation. Therefore, decreasing meta olic rate is preferred if possi le. At this point in the management, if the prognosis is poor, there may e noth ing more to do. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 22)03-May-05 17:21:54

Ovid: ICU Book IMPENDING VO2 DEFICIT As descri ed in Chapter 2 and shown in Figure 13.4, VO2 is kept constant when DO 2 is reduced ecause of reciprocal adjustments in O2ER (5,6). When O2ER increase s to a level of 0.5 to 0.6, VO2 ecomes P.193 supply dependent as DO2 is further reduced. When this occurs, cellular energy production is oxygen-limited (dysoxi a). Thus, in the setting of impaired O2 delivery (e.g., low cardiac output or an emia) an O2ER a ove 0.5 indicates a high risk of developing impaired tissue oxyg enation. In this situation, measures aimed at increasing the DO2 are protective. This principle has een applied to patients with normovolemic anemia, where an O2ER a ove 50% is recommended as an indication for lood transfusion (7). Figure 13.4. Graph showing relationship etween O2 delivery (DO2) and O2 uptake (VO2) in normal su jects (solid line) and in critically ill patients (dashed lin e). SUPPLY-DEPENDENT VO2 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 22)03-May-05 17:21:54

Ovid: ICU Book The DO2VO2 relationship in critically ill patients can differ from the normal pat tern, as shown in Figure 13.4. The normal DO2VO2 relationship (descri ed in Chapt er 2) shows a constant VO2 over a wide range of variations in DO2. In critically ill patients, the DO2VO2 curve is predominantly linear, and the slope is reduced (indicating a low O2ER). This covariance was originally attri uted to dysoxia a nd was called pathologic supply dependency. However, it now seems that in most c ases, this phenomenon is not the result of a pathologic derangement in oxygen me ta olism, ut it is a manifestation of the processes descri ed elow (5,6,8,9). P.194 Physiologic Coupling A linear DO2VO2 relationship can occur when there is a primary change in meta oli c rate and DO2 changes proportionally to match the newly created oxygen requirem ents. In this situation, the covariance of DO2 and VO2 is a normal adaptive resp onse and is not a sign of impaired tissue oxygenation. Changes in meta olic rate occur commonly in patients in the ICU. The meta olic response to activity is of ten exaggerated in critically ill patients, and even routine ICU interventions ( e.g., porta le chest x-ray) can cause significant (20% or greater) increases in meta olic rate (9). Mathematical Coupling The a normal supply dependency in patients in the ICU occurs almost exclusively when DO2 and VO2 are calculated, and it disappears 140,300,410 when VO2 is direc tly measured y gas exchange (8,10,11,12 and 13). This indicates that the a norm al link etween DO2 and VO2 is an artifact related to the calculations used to d erive these parameters. One possi le source of the pro lem is mathematical coupl ing ecause the equations for DO2 and VO2 share three varia les (i.e., hemoglo i n, cardiac output, and arterial O2 saturation). Thus, a change in any of these s hared varia les could affect oth calculations and produce an artifactual link. The source of the artifactual link etween DO2 and VO2 is still not clear, ut t he pro lem demonstrates that the calculations used to derive O2 delivery and O2 uptake can affect the relia ility of oxygen transport monitoring. This warrants a rief look at calculated versus measured transport varia les. CALCULATED VERSUS MEASURED VO2 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 22)03-May-05 17:21:54

Ovid: ICU Book The VO2 is usually derived (and not measured) using Equation 13.1. The derivation is ased on four measured varia les: cardiac output (Q), hemoglo in concentration (H ), arterial O2 saturation (SaO2), and mixed venous O2 satura tion (SvO2). Each of these measurements varies, and their summed contri ution ca n lead to considera le varia ility in the calculated VO2. This is shown in Ta le 13.2 (14). TABLE 13.2. VARIABILITY IN THE CALCULATED AND MEASURED VO2 The varia ility of each component is expressed y the coefficient of variation ( CV), which is the standard deviation expressed as a percentage of the mean. Beca use the standard deviation is expressed on either side of the mean, the CV also expresses a range that is 2 CV, and this range is indicated for each measurement . La oratory measurements are considered reproduci le if they have a CV elow 5% , and each of the individual measurements has a CV that is within this range or not far removed. However, the sum of the individual variations creates a large v aria ility in the calculated VO2. Considering the range P.195 of variation, the calculated VO2 could change y 30% without a change in the meta olic condition o f the patient. The calculated VO2 is considered to have a range of error that is 15% on either side of the mean for individual determinations (15). This is cons istent with the varia ility in Ta le 13.2 and forms the asis for the recommenda tion that the calculated VO2 should change at least 15% to e considered a physi ologically significant change. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 22)03-May-05 17:21:54

Ovid: ICU Book Gas Exchange VO2 can also e measured as the oxygen concentration difference in inhaled and e xhaled gas multiplied y the respiratory rate. A num er of instruments are avail a le that can measure VO2 at the edside. Many of these instruments can also mea sure car on dioxide production (VCO2) and are used y nutrition support services to measure daily energy requirements. As shown in Ta le 13.2, the measured VO2 is much less varia le than the calculated VO2, and thus has less of a tendency f or error. Because gas exchange measurements have a CV elow 5%, a change in the measured VO2 that exceeds 5% can e considered physiologically significant. Whole Body versus Systemic VO2 Although the calculated and measured VO2 are often compared, they should not e considered equivalent ecause the gas exchange method measures the whole ody VO 2, whereas the calculated VO2 measures only systemic VO2. Thus, the measured VO2 is higher than the calculated VO2 y the VO2 in the lungs. Normally, the VO2 in the lungs accounts for less than 5% of the whole ody VO2 (16). However, in pat ients with inflammatory lung injury (i. e., acute respiratory distress syndrome) , 20% of the whole ody VO2 can take place in the lungs (17). This corresponds t o a difference of 25 mL/minute/m2 etween whole ody VO2 and calculated VO2. Thi s difference deserves consideration when comparing measured and calculated param eters. P.196 BLOOD LACTATE CONCENTRATION As mentioned, lood lactate levels help determine whether VO2 is adequate for th e needs of aero ic meta olism. Thus, adding lactate determinations to oxygen tra nsport monitoring provides a more complete assessment of tissue oxygen alance. Because lactate levels in whole lood and plasma are equivalent (18), oth measu rements are called lood lactate. BLOOD LACTATE AND SURVIVAL One of the reasons the lood lactate is such popular test is its a ility to pred ict outcome. A comparison of lactate with cardiac output and oxygen uptake is sh own in Ta le 13.3 for patients with septic shock (19). Neither cardiac output no r oxygen uptake differs significantly in http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 22)03-May-05 17:21:54

Ovid: ICU Book survivors and nonsurvivors, whereas the lactate levels are three times as high i n the nonsurvivors. The predictive value of lood lactate levels is consistently etter than any measure of hemodynamics or oxygen transport (20), ut the a ili ty of lactate to predict mortality is limited mostly to patients with shock. TABLE 13.3. CORRELATIONS WITH OUTCOME IN PATIENTS WITH SEPTIC SHOCK Optimal Threshold As shown in Ta le 13.1, the normal lood lactate concentration is less than 2 mm ol/L, ut the threshold for an elevated lood lactate is higher, at 4 mmol/L. Th e reason for this difference is shown in Ta le 13.4 (18). This ta le shows the r elationship etween the cutoff level P.197 for an elevated lood lactate and the relia ility of an elevated lactate level for predicting mortality. The lower th reshold of 2 mmol/L is very sensitive ut not specific, which means that a consi dera le fraction of lactate levels in this range will e false positives. On mov ing to the higher threshold of 4 mmol/L, sensitivity declines y 27% ut specifi city is much (46%) higher, as is the a ility to predict outcome (i.e., the posit ive predictive value). Thus, the higher threshold is preferred for the clinical definition of hyperlactatemia. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 22)03-May-05 17:21:54

Ovid: ICU Book TABLE 13.4. INFLUENCE OF LACTATE THRESHOLD ON MORTALITY PREDICTIONS OTHER SOURCES OF LACTATE Anaero ic meta olism is not the only source of lactate. Other causes of hyperlac tatemia include hepatic insufficiency (caused y impaired clearance of lactate y the liver), thiamine deficiency ( locks pyruvate entry into mitochondria), alk alosis (stimulates glycolysis), and production y enteric micro es (D-lactic aci d). Sepsis Evidence suggests that the lactate accumulation in sepsis is not the result of o xygen deprivation. The culprit may e endotoxin, which locks the actions of the pyruvate dehydrogenase enzyme that moves pyruvate into mitochondria. Pyruvate t hen accumulates in the cell cytoplasm and is converted to lactate. The a ility o f endotoxin to promote lactate formation is shown in the graph in Figure 13.5 (2 1). This graph is taken from a study in which animals were su jected to a 1-hour P.198 infusion of endotoxin. As indicated on the graph, the endotoxin infusion was associated with a progressive rise in lood lactate. After the endotoxin inf usion, the animals were given dichloroacetate, a su stance that activates pyruva te dehydrogenase, ut only in the presence of oxygen. The dichloroacetate was a le to reduce the lactate levels to normal, indicating that oxygen was present in cells to permit the activation of pyruvate dehydrogenase. Furthermore, when hyp oxia was induced y having the animals reathe a low-oxygen gas mixture (on the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 22)03-May-05 17:21:54

Ovid: ICU Book right side of the graph), the lactate levels failed to rise. This study shows th at oxygen deprivation can e unrelated to lactate production in a setting that m imics sepsis. Figure 13.5. Influence of endotoxin, dichloroacetate, and hypoxic challenge on a rterial lactate levels. Response to dichloroacetate indicates that endotoxin-ass ociated lactic acidosis is not caused y anaero ic conditions. (From Curtis SE, Cain SM. Regional and systemic oxygen delivery/uptake relations and lactate flux in hyperdynamic, endotoxin-treated dogs. Am Rev Respir Dis 1992;145:348354.) LACTATE AS A FUEL A final word a out lactate that deserves mention is the possi ility that lactate might serve as an oxidative fuel. The energy yield from the oxidation of lactat e is shown in Ta le 13.5. Also shown is the energy yield from the oxidation of g lucose. The energy yield from glucose oxidation is twice that of lactate, ut gl ucose is twice the size of lactate (i.e., 6 car ons versus 3 car ons, respective ly). Because each mole of glucose produces 2 moles of lactate, the energy yield from glucose meta olism is a out the same when glucose is directly oxidized and when glucose is converted to lactate and the lactate is oxidized. A num er of or gans can oxidize lactate to derive energy, including the heart, rain, liver, an d skeletal muscle (22,23). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 22)03-May-05 17:21:54

Ovid: ICU Book TABLE 13.5. LACTATE AS AN OXIDATIVE FUEL If the lactate generated during periods of oxygen deprivation can undergo oxidat ion at a later time, when tissue oxygenation is restored, then the energy yield of glucose oxidation (i.e., oxidative meta olism) will e preserved. In this con text, lactate production would e a mechanism for preserving nutrient energy dur ing periods of hypoxia or ischemia, when prevailing conditions do not favor oxid ation. GASTRIC TONOMETRY The oxygen transport varia les and lood lactate levels are glo al (whole ody) measures that cannot identify oxygen deficits in individual P.199 organs. This l imitation ecame a concern upon the discovery that splanchnic hypoperfusion is c ommon in critically ill patients and may e the prelude to multiorgan failure (2 4). This led to the development of the method descri ed here for evaluating oxyg enation in the gastrointestinal tract (25,26). METHOD The asic elements of this method are outlined in Ta le 13.6. This method uses a n indirect measurement of the pH in the gastric mucosa to evaluate the adequacy of tissue oxygenation (i.e., oxygen deficits produce a local acidosis). This mea surement is derived using the HendersonHassel ach equation and an indirect measur e of the gastric mucosal car on dioxide pressure (PCO2), o tained with a tonomet er. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 22)03-May-05 17:21:54

Ovid: ICU Book TABLE 13.6. THE GASTRIC TONOMETRY METHOD The tonometer is a CO2-permea le silicone alloon affixed to the distal end of a standard 16French nasogastric tu e. The apparatus is placed in the stomach in t he usual fashion, and the alloon is partially filled with saline (2.5 mL) and l eft that way for at least 30 minutes. During this time, the alloon is in contac t with the gastric mucosa, and the CO2 in the adjacent mucosa moves into the al loon. The CO2 eventually equili rates etween the tissues and the saline in the alloon. When this occurs, the PCO2 in the saline approximates the PCO2 in the g astric mucosa, and thus the saline PCO2 is measured and used as the intramucosal PCO2. The pH calculation also requires a measure of the tissue icar onate, and the icar onate concentration in an arterial lood sample is used for this purp ose. The normal gastric intramucosal pH has a mean of 7.38 and a standard deviat ion (SD) of 0.3, so the range is 7.35 to 7.41 (1 SD on either side of the mean). The threshold for an a normal pH is 7.32, which is 2 SDs from the mean (Ta le 1 3.1). PERFORMANCE The advantage of monitoring pH in the gastric mucosa is illustrated y the case depicted in Figure 13.6. The line plots represent the temporal P.200 changes in the systemic oxygen uptake (VO2) and the gastric intramucosal pH (pHi) in the ea rly postoperative period following renal transplantation. Both measures are in t he normal range during the first postoperative day. However, at approximately 30 hours after surgery, the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 22)03-May-05 17:21:54

Ovid: ICU Book patient developed a sepsis syndrome (indicated y the dashed line). Thereafter, the intramucosal pH dropped precipitously, while the VO2 remained unchanged. The patient was returned to the operating room 12 hours after the onset of sepsis, and an infected renal implant was removed. The sepsis su sequently resolved and the patient survived. At no time during this rocky postoperative course did the VO2 provide any hint of danger, whereas the gastric mucosal pH showed evidence o f progressive ischemia as the sepsis progressed. Figure 13.6. Postoperative changes in oxygen uptake (VO2) and gastric intramucos al pH (pHi) in a patient with postoperative sepsis. Onset of sepsis syndrome is indicated y the dashed line. Patient returned to operating room at 48 hours. (F rom Gutierrez G. Cellular energy meta olism during hypoxia. Crit Care Med 1991;1 9:619626.) In keeping with the o servations in Figure 13.6, the gastric mucosal pH has prov en superior to the glo al measures of tissue oxygenation (O2 transport varia les and lactate) for predicting outcome in critically ill patients (25,26,27 and 28 ). PROBLEMS A num er of shortcomings associated with gastric tonometry deserve mention. Desc riptions of the major ones follow.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 22)03-May-05 17:21:54

Gastric Acid Secretion Acid secretion in the stomach is a confounding varia le that must when using

e eliminated

Ovid: ICU Book gastric mucosal pH as a marker of tissue P.201 oxygenation. Routine doses of his tamine H2 lockers may not achieve adequate acid suppression. Administration of ranitidine 100 mg intravenously 1 hour efore measurements effectively locks ac id secretion for 2 to 4 hours (29). Raising gastric pH carries the risk of gastr ic colonization, which is not a desira le condition in critically ill patients ( as discussed in Chapter 6). AcidBase Disorders Systemic acid ase disorders can also influence the pH of the gastric mucosa (30). This is a particular concern ecause meta olic acidosis can e common in patien ts with circulatory shock, and these patients represent a large target populatio n for gastric mucosal pH monitoring. Also of concern is respiratory alkalosis, w hich is common in mechanically ventilated patients. Arterial versus Mucosal Bicar onate The use of arterial icar onate as a measure of mucosal icar onate is pro lemat ic ecause the two are not equivalent in low flow states ( ecause of local accum ulation of acid) (24). Thus, in low flow states, where accuracy is most importan t, arterial icar onate measurements are likely to produce the least accurate re sults. MANAGEMENT The treatment of a normal mucosal pH is not well defined. The few studies availa le on treatment use volume infusions followed y do utamine. The volume is give n in aliquots, with no clear endpoint. Do utamine has consistently increased muc osal lood flow, ut its effects on mucosal pH are varia le. Individual studies of do utamine show increased pH (31,32), no change in pH (33), and even decrease d pH (34). The response to do utamine therefore must e determined on an individ ual asis. REFERENCES GENERAL TEXTS Edwards JD, Shoemaker WC, Vincent JL, eds. Oxygen transport. Principles and http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 22)03-May-05 17:21:54

Ovid: ICU Book practice. Philadelphia: WB Saunders, 1993. Marston A, Bulkley GB, Fiddian-Green RG, Haglund UH, eds. Splanchnic ischemia an d multiple organ failure. St. Louis: CV Mos y, 1989. TISSUE OXYGEN BALANCE 1. Connett RJ, Honig CR, Gayeski TEJ, Brooks GA. Defining hypoxia: a systems vie w of VO2, glycolysis, energetics, and intracellular PO2. J Appl Physiol 1990;68: 833842. P.202 OXYGEN TRANSPORT 2. Shoemaker WC, Kram HB, Appel PL. Therapy of shock ased on pathophysiology, m onitoring, and outcome prediction. Crit Care Med 1990;18(Suppl):S19S25. 3. Dunham CM, Seigel JH, Weireter L, et al. Oxygen de t and meta olic acidemia a s quantitative predictors of mortality and the severity of the ischemic insult i n hemorrhagic shock. Crit Care Med 1991;19:231243. 4. Shoemaker WC, Appel PL, Krom HB. Role of oxygen de t in the development of or gan failure, sepsis, and death in high-risk surgical patients. Chest 1992;102:20 8215. 5. Leach RM, Treacher DF. The relationship etween oxygen delivery and consumpti on. Dis Mon 1994;30:301368. 6. Dantzker DR, Foresman B, Gutierrez G. Oxygen supply and utilization relations hips. Am Rev Respir Dis 1991;143:675679. 7. Levy P, Chavez RP, Crystal GJ, et al. Oxygen extraction ratio: a valid indica tor of transfusion need in limited coronary vascular reserve? J Trauma 1992;32:7 69774. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 22)03-May-05 17:21:54

Ovid: ICU Book 8. Russell JA, Phang PT. The oxygen delivery/consumption controversy. Am Rev Res pir Crit Care Med 1994;149:533537. 9. Weissman C, Kemper M. The oxygen uptakeoxygen delivery relationship during ICU interventions. Chest 1991;99:430435. 10. Hanique G, Dugernier T, Laterre PF, et al. Significance of pathologic oxygen supply dependence in critically ill patients: comparison etween measured and c alculated methods. Intensive Care Med 1994;20:1218. 11. Ronco JJ, Phang PT, Walley KR, et al. Oxygen consumption is independent of c hanges in oxygen delivery in severe adult respiratory distress syndrome. Am Rev Respir Dis 1991;143:12671273. 12. Romco JJ, Fenwick JC, Wiggs BR, et al. Oxygen consumption is independent of increases in oxygen delivery y do utamine in septic patients who have normal or increased lactate. Am Rev Respir Dis 1993;147:2531. 13. Marik PE, Si ald W. Effect of stored- lood transfusion on oxygen delivery i n patients with sepsis. JAMA 1993;269:30243029. 14. Schneeweiss B, Druml W, Graninger W, et al. Assessment of oxygen-consumption y use of reverse Fick principle and indirect calorimetry in critically ill pat ients. Clin Nutr 1989;8:8993. 15. Bartlett RH, Dechert RE. Oxygen kinetics: pitfalls in clinical research. J C rit Care 1990;5:7780. 16. Nunn JF. Non-respiratory functions of the lung. In: Nunn JF, ed. Applied res piratory physiology. 4th ed. Kent, England: Butterworth, 1993;306317. 17. Jolliet P, Thorens JB, Nicod L, et al. Relationship etween pulmonary oxygen consumption, lung inflammation, and calculated venous admixture in patients wit h acute lung injury. Intensive Care Med 1996;22:277285. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 22)03-May-05 17:21:54

Ovid: ICU Book BLOOD LACTATE 18. Aduen J, Bernstein WK, Khastgir T, et al. The use and clinical importance of a su strate-specific electrode for rapid determination of lood lactate concent rations. JAMA 1994;272:16781685. 19. Mizock BA, Falk JL. Lactic acidosis in critical illness. Crit Care Med 1992; 20:8093. 20. Bakker J, Coffernils M, Leon M, et al. Blood lactate levels are superior to oxygenderived varia les in predicting outcome in septic shock. Chest 1991;99:9569 62. P.203 21. Curtis SE, Cain SM. Regional and systemic oxygen delivery/uptake relations a nd lactate flux in hyperdynamic, endotoxin-treated dogs. Am Rev Respir Dis 1992; 145:348 354. 22. Brooks GA. Lactate production under fully aero ic conditions: the lactate sh uttle during rest and exercise. Fed Proc 1986;45:29242929.

GASTRIC TONOMETRY 24. Fiddian-Green RG. Studies in splanchnic ischemia and multiple organ failure. In: Marston A, Bulkley GB, Fiddian-Green RG, Haglund UH, eds. Splanchnic ischem ia and multiple organ failure. St. Louis: CV Mos y, 1989;349364. 25. Fiddian-Green RG. Tonometry: theory and applications. Intensive Care World 1 992;9:6065. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 22)03-May-05 17:21:54

23. Maran A, Cranston I, Lomas J, et al. Protection ion during hypoglycemia. Lancet 1994;343:1620.

y lactate of cere ral funct

Ovid: ICU Book 26. Gutierrez G, Brown SD. Gastric tonometry: a new monitoring modality in the i ntensive care unit. J Intensive Care Med 1995;10:3444. 27. Marik PE. Gastric intramucosal pH. A etter predictor of multiorgan dysfunct ion syndrome and death than oxygen-derived varia les in patients with sepsis. Ch est 1993;104:225229. 28. Maynard N, Bihari D, Beale R, et al. Assessment of splanchnic oxygenation y gastric tonometry in patients with acute circulatory failure. JAMA 1993;270:120 31210. 29. Kolkman JJ, Groeneveld ABJ, Meuwissen SGM. Effect of ranitidine on asal and icar onate enhanced intragastric PCO2: a tonometric study. Gut 1994;35:737741. 30. Benjamin E, Polokoff E, Oropello JM, et al. Sodium icar onate administratio n affects the diagnostic accuracy of gastrointestinal tonometry in acute mesente ric ischemia. Crit Care Med 1992;20:11811183. 31. Silverman HJ, Tuma P. Gastric tonometry in patients with sepsis: effects of do utamine infusions and packed red lood cell transfusions. Chest 1992;102:18418 8. 32. Gutierrez G, Clark C, Brown SD, et al. Effect of do utamine on oxygen consum ption and gastric mucosal pH in septic patients. Am Rev Respir Crit Care Med 199 4;150:324 329. 33. Uusaro A, Ruokonen E, Takala J. Splanchnic oxygen transport after cardiac su rgery: evidence for inadequate tissue perfusion after sta ilization of hemodynam ics. Intensive Care Med 1996;22:2633. 34. Parviainen I, Ruokonen E, Takala J. Do utamine-induced dissociation etween changes in splanchnic lood flow and gastric intramucosal pH. Br J Anesth 1995;7 4:277 284. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 22)03-May-05 17:21:54

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 14 HEMORRHAGE AND HYPOVOLEMIA The dominant concern in the leeding patient is the human ody's intolerance to acute lood loss. The human cardiovascular system operates with a small volume a nd a steep Starling curve, which seems to e an energy-efficient design, ut doe s not ode well in acute lood loss. Although we are more than half fluid y wei ght, only 12 to 13% of this fluid is located in the loodstream. Because the acu te loss of 40% of the lood volume can e fatal, the loss of only 5% (0.4 12%) o f our total fluid volume can prove fatal. When confronted with acute hemorrhage, the task is to intervene efore this small aliquot of fluid is lost. This chapt er introduces some of the fundamental concerns in the initial management of acut e hemorrhage (1,2,3,4,5,6,7,8 and 9). The next chapter descri es the different t ypes of intravenous fluids. The material in these two chapters provides a fundam ental knowledge of the principles of volume resuscitation. BODY FLUIDS AND BLOOD LOSS The relationship etween total fluid volume and lood volume in adults is shown in Ta le 14.1 (10). The total volume of ody fluids represents 60% of lean ody weight (or 600 mL/kg) in men and 50% of lean ody weight (or 500 mL/kg) in women . The volume of whole lood is 6.0 to 6.6% of lean ody weight (or 60 to 66 mL/k g), which means that only 11 (i.e., 66/600) to 12% (60/500) of the total ody fl uid volume is in the intravascular compartment. Approximately 60% of the lood v olume is in the plasma fraction, and the remaining 40% represents the erythrocyt e volume. The corresponding volumes for an average-size adult man (weighing 80 k g or 176 l s) and adult woman (weighing 60 kg or 132 l s) is shown elow. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 28)03-May-05 17:09:42

Ovid: ICU Book TABLE 14.1. BODY FLUID AND BLOOD VOLUMES P.208

60-kg Woman 30 L 3.6 L 2.2 L 1.4 L Adjustments The American Association of Blood Banks recommends the following adjustments for estimating lood, plasma, and erythrocyte volumes ased on ody weight (10). q http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 28)03-May-05 17:09:42

80-kg Man Total ody fluids Whole L

lood Plasma Erythrocytes 48 L 5.3 L 3.2 L 2.1

Ovid: ICU Book

For patients with a marked weight loss within 6 months, use the premor id weight loss to estimate volumes. RESPONSE TO MILD BLOOD LOSS The principles of volume resuscitation are ased on the normal ody response to hemorrhage. This response has een examined in mild hemorrhage (< 15% loss in l ood volume), when volume resuscitation is not necessary. The response is descri ed in three stages (4). Stage 1. The first hours after the onset of lood loss is characterized y movement of in terstitial fluid into the capillaries. This transcapillary refill helps maintain lood volume ut leaves an interstitial fluid deficit. Stage 2. The loss of ody fluids leads to activation of the reninangiotensinaldosterone sys tem. This results in sodium conservation y the kidneys. Because sodium distri u tes primarily in the interstitial space, the retained sodium replenishes the int erstitial fluid deficit. Stage 3. Within a few hours after the onset of hemorrhage, the marrow egins to increase the production of erythrocytes. This response occurs more gradually, and complet e replacement of erythrocytes can take up to 2 months (4). P.209 According to th e response to mild hemorrhage descri ed a ove, the goal of acute volume resuscit ation for lood loss should e to replenish interstitial fluid deficits (3,4). T his is why crystalloid (electrolyte) fluids are used in the resuscitation of acu te lood loss (3), as is discussed later. CLINICAL CONSEQUENCES http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 28)03-May-05 17:09:42

For o ese and elderly patients, estimate the volumes ased on lean nd reduce the values y 10%. q

ody weight a

Ovid: ICU Book The clinical consequences of hypovolemia are determined y the rapidity and magn itude of volume loss and y the responsiveness of individual patients to volume loss. Most cases of mild lood loss are relatively free of clinical manifestatio ns. In fact, hypovolemia may e clinically silent until the volume loss exceeds 30% of the lood volume. The American College of Surgeons identifies four catego ries of acute lood loss ased on the percent loss of lood volume (11). These a re shown in Ta le 14.2. TABLE 14.2. CLASSIFICATION OF HEMORRHAGE BASED ON EXTENT OF BLOOD LOSS Class I. Loss of 15% or less of the total lood volume. This degree of lood loss is usua lly fully compensated y transcapillary refill. Because lood volume is maintain ed, clinical manifestations of hypovolemia are minimal or a sent. Class II. Loss of 15 to 30% of the lood volume. The clinical findings at this stage may i nclude resting tachycardia and orthostatic changes in heart rate and lood press ure. However, resting tachycardia can e an inconsistent finding, and orthostati c changes in pulse and lood pressure are too insensitive to e considered relia le manifestations (5a,6,12). A positive tilt test, defined as an increase in pu lse rate greater than 30 eats/minute or a drop in systolic pressure greater tha n 30 mm Hg on assuming the upright position (12), can e used as http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 28)03-May-05 17:09:42

Ovid: ICU Book corro orative evidence for lood loss, ut a negative result has no meaning. Whe n the tilt test is performed, the lower legs must e in a dependent position (si tting without legs dangling is inappropriate). Also, ecause changes in pulse an d pressure are varia le within the first minute after changing positions (13), P .210 a waiting period of at least 1 minute after changing positions is recommend ed efore the vital signs are recorded. Class III. Loss of 30 to 40% of the lood volume usually marks the onset of hypovolemic sho ck, with a decrease in lood pressure and urine output. There is evidence that t he tachycardia vasoconstrictor response to hemorrhage can e lost at this stage o f lood loss (5a). When this occurs, the decrease in lood pressure can e sudde n and profound. Class IV. Loss of more than 40% of lood volume is a har inger of circulatory collapse. Th erefore, when hypovolemia is accompanied y marked hypotension, oliguria, or oth er evidence of organ failure, prompt volume resuscitation is mandatory. CLINICAL MONITORING The clinical evaluation of lood loss is far from precise, even when employing i nvasive hemodynamic monitoring. The following are some important points regardin g the clinical parameters used to evaluate patients with suspected or documented lood loss. BLOOD PRESSURE The systemic lood pressure is not a sensitive marker of lood loss, as shown in Ta le 14.2. However, when lood loss is severe enough to produce hypotension, t he lood pressure can e a valua le guide in the resuscitative effort. As mentio ned in Chapter 9, the noninvasive methods of measuring lood pressure often yiel d spuriously low measurements in patients with hypovolemia (presuma ly ecause o f the vasoconstrictor response to volume loss) (6). Therefore, when hypovolemia egins to produce a change in noninvasive lood pressure measurements, the press ure should e monitored y direct intraarterial recordings. CARDIAC FILLING PRESSURES http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 28)03-May-05 17:09:42

Ovid: ICU Book Although cardiac filling pressures (i.e., the central venous pressure [CVP] and wedge pressure) are monitored routinely in patients with acute lood loss, these pressures show a poor correlation with the presence and extent of volume loss ( 14,15). In particular, these pressures often show little change until the lood loss is severe (i.e., greater than 30% of the lood volume). This insensitivity can e explained in two ways. First, the CVP and wedge pressures are normally lo w pressures (particularly the CVP, which is normally less than 5 mm Hg), and thu s there is little margin for a detecta le change in hypovolemia. Second, hypovol emia can e accompanied y a decrease in ventricular distensi ility (presuma ly as a result of sympathetic activation) (16), and when this occurs, the CVP and w edge pressures will e higher P.211 than expected at any given ventricular volum e. In one animal study of hypovolemia, decreases in ventricular compliance resul ted in a two-fold increase in the wedge (left-ventricular enddiastolic) pressure despite a 50% reduction in the end-diastolic volume (16). Positional changes in cardiac filling pressures may e a more sensitive marker of hypovolemia. In one report, hypovolemia failed to produce a change in the CVP when it was measured in the supine position; however, when the patients were placed in an upright pos ition, the CVP decreased 4 to 5 mm Hg (17). Therefore, performing orthostatic ma neuvers may help improve the sensitivity of cardiac filling pressures in hypovol emia. OXYGEN EXTRACTION As explained in Chapter 2, the normal response to a decrease in cardiac output ( O2 delivery) is an increase in oxygen extraction in the systemic microcirculatio n. This is a compensatory response aimed at keeping oxygen uptake normal when ox ygen delivery is compromised. However, there is a limit to the increase in oxyge n extraction, and when this limit is reached, decreases in cardiac output are ac companied y proportional decreases in oxygen uptake into the tissues (see Fig. 2.2). Therefore, an increase in O2 extraction can e a marker of systemic hypope rfusion, and a maximum increases in O2 extraction can e a marker of hypovolemic shock. Oxygen extraction can e monitored without a pulmonary artery catheter y com ining pulse oximetry (for arterial O2 saturation) with measurements of oxy gen saturation in lood samples o tained from a CVP catheter (the O2 saturation in the superior vena cava is normally close to the mixed venous O2 saturation). The expected changes in O2 extraction and mixed venous O2 saturation in progress ive hypovolemia are shown elow. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 28)03-May-05 17:09:42

Ovid: ICU Book Sao2 Normal Hypovolemia Hypovolemic shock >95% >95% >95% Svo2 >65% 5065% <50% Sao2 Svo2 2030% 3050% >50% The transition from compensated hypovolemia to hypovolemic shock takes place whe n the oxygen extraction reaches 50 to 60% and the mixed venous O2 saturation fal ls to 50%. Therefore, an O2 extraction greater than 30% is a marker of hemodynam ically significant hypovolemia, and an O2 extraction greater than 50% indicates possi le hypovolemia shock. When the O2 extraction exceeds 50%, a lood lactate level will help identify hypovolemic shock (a lactate level greater than 4 mmol/ L indicates a shock state). Oxygen extraction can also e increased y hypermeta olism or in response to anemia, and these conditions must also e considered in the interpretation of an increased O2 extraction. P.212

END-EXPIRATORY CO2 A decrease in cardiac output will decrease the PCO2 in exhaled gas(18), and eca use the PCO2 in exhaled gas can e measured noninvasively (see Chapter 22), this provides a potentially useful method for monitoring the severity of hypovolemia . Intu ation is not necessary, ecause exhaled CO2 can e monitored using standa rd nasal cannulas used to deliver supplemental oxygen (descri ed in Chapter 22). The changes in end-expiratory (end-tidal) PCO2 in hypovolemia and volume resusc itation are shown in Figure 14.1 (6). The data in this figure are from a patient who presented with hypovolemic shock. The end-tidal PCO2 efore volume infusion is very low at 10 mm Hg (the end-tidal PCO2 is normally within 3 mm Hg of the a rterial PCO2). After volume resuscitation with 4.5 L of intravenous fluids, the end-tidal PCO2 has risen to 30 mm Hg, indicating that the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 28)03-May-05 17:09:42

Ovid: ICU Book cardiac output has increased in response to the volume infusion. Because the end -tidal PCO2 is measured on a reath- y- reath asis, it provides an on-line meas ure of the success or failure of volume resuscitation. Figure 14.1. Serial changes in the end-tidal PCO2 (ET CO2) during volume resusci tation of a patient with hypovolemic shock. Horizontal axis represents cumulativ e volume infused. (From Falk JL et al. Fluid resuscitation in traumatic hemorrha gic shock. Crit Care Clin 1992;8:323 340.) End-tidal CO2 monitoring has een recommended for evaluating the response to car diopulmonary resuscitation, and it should have a similar role in the evaluation of hypovolemia. The Hematocrit Both physicians and nurses share a common propensity to use the hematocrit as an estimate of acute lood loss. The error of this practice is indicated in the fo llowing statement taken from the Advanced P.213 Trauma Life Support Course stude nt manual, pu lished y the American College of Surgeons: Use of the hematocrit t o estimate acute lood loss is unrelia le and inappropriate (11). Changes in hema tocrit show a poor correlation with lood volume deficits and red cell volume de ficits in acute hemorrhage (15). In fact, loss of whole lood is not expected to change the hematocrit ecause the relative proportions of plasma and red cell http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 28)03-May-05 17:09:42

Ovid: ICU Book volume are unchanged. The decrease in hematocrit occurs when the kidney egins t o conserve sodium (as descri ed previously), which takes 8 to 12 hours to ecome evident. Another factor that drops the hematocrit in acute hemorrhage is the ad ministration of intravenous (asanguinous) fluids. The influence of volume resusc itation on the hematocrit is illustrated in Figure 14.2. Each column in this fig ure is partitioned to indicate the relative proportions of plasma and erythrocyt es in the lood. The columns on the left show that acute lood loss decreases l ood volume ut does not change the hematocrit. The columns on the right show the influence of lood and asanguinous fluids on the hematocrit. Saline infusion in creases the plasma volume selectively and there y decreases the hematocrit. Infu sion of whole lood expands the plasma and erythrocyte fractions proportionately , and thus does not alter the hematocrit. Therefore, in the first hours after th e onset of lood loss, the hematocrit is a reflection of the resuscitation effor t, not the extent of lood loss. The administration of intravenous (asanguinous) fluids P.214 is expected to produce a dilutional decrease in the hematocrit, ev en in the a sence of lood loss (19), and thus a decrease in the hematocrit duri ng volume resuscitation is a dilutional effect, and it is not an indication of o ngoing lood loss. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 28)03-May-05 17:09:42

Ovid: ICU Book Figure 14.2. Influence of acute hemorrhage and resuscitation fluid on hematocrit . Each vertical ar is partitioned to indicated contri ution of plasma and lood cells to total lood volume. Hematocrit indicated at the top of each vertical ar. (See text for further explanation.) RATE OF VOLUME RESUSCITATION The mortality in hypovolemic shock is directly related to the magnitude and dura tion of the ischemic insult (6), and thus prompt replacement of volume deficits is the hallmark of successful management. The a ility to infuse volume rapidly i s thus an important consideration in the management of hypovolemia. The followin g is a rief description of the factors that influence the rate of volume infusi on. VASCULAR ACCESS SITE Although there is a tendency to cannulate the large central veins for volume res uscitation, cannulation of peripheral veins is preferred. The larger size of the central veins is not an important consideration in volume resuscitation ecause the rate of volume infusion is determined y the dimensions of the vascular cat heter, not y the size of the vein. Cannulation of the large central veins requi res catheters that are at least 5 inches in length, whereas cannulation of perip heral veins can e accomplished with catheters that are 2 inches in length. Shor ter catheters permit more rapid rates of volume infusion, and thus cannulation o f peripheral veins is more favora le for rapid volume resuscitation. Central ven ous cannulation is reserved for monitoring cardiac filling pressures and venous O2 saturation unless very-large ore introducer catheters are used for volume res uscitation. Catheter Dimensions The determinants of flow through rigid tu es is descri ed in Chapter 1. Accordin g to the HagenPoiseuille equation (presented in the latter part of Chapter 1 and depicted schematically in Figure 1.4), the rate of laminar or streamlined flow will vary directly with the fourth power of the inner radius of the catheter. Th us, if the radius of a catheter is dou led, the flow rate through the catheter w ill increase sixteenfold; that is, (2r)4 = 16r. Changes in catheter length will have a proportional influence on flow rate; that is, if the length is dou led, t he infusion rate will decrease y one-half. Because central venous catheters are 3 to 4 times longer than peripheral venous catheters, the infusion rate through central catheters will e as much as 75% less than the infusion rate through pe ripheral catheters (of equal diameter). The influence of catheter size on infusi on rate is shown in Figure 14.3 (21). P.215 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 28)03-May-05 17:09:42

Ovid: ICU Book The fluid in this case is water, and the gradient for flow is the force of gravi ty. Note that for catheters of equivalent diameter (16 gauge), flow is 1.5 to 3 times faster in the shorter (2 inch) catheter. This demonstrates why shorter per ipheral catheters are preferred for the resuscitation of hemorrhage. Figure 14.3. The influence of catheter dimensions on the gravity-induced flow ra te of tap water. Catheter dimensions are indicated elow the horizontal axis of the graph. (From Mateer JR et al. Rapid fluid resuscitation with central venous catheters. Ann Emerg Med 1983;12:149152.) Rapid Infusion Because the radius of a tu e has a much stronger influence on flow rate than the length of the tu e, rapid infusion rates are more easily achieved y increasing the diameter of a catheter rather than decreasing its length. Rapid volume infu sion, defined as the infusion of at least 5 L of fluid hourly (20), is thus est accomplished y using large- ore introducer catheters normally used in conjunct ion with multilumen central venous catheters. Introducer catheters are descri ed in Chapter 4 (see Figure 4.3). These devices are 5 to 6 inches in length and ar e availa le in 8.5 French (2.7 mm outer diameter) and 9 French (3 mm outer diame ter) sizes. They are normally used as intravascular conduits or sheaths for mult ilumen central venous and pulmonary artery catheters, and they allow these cathe ters to e inserted and removed without sacrificing the central venous access si te. However, introducer catheters can e used as standhttp://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 28)03-May-05 17:09:42

Ovid: ICU Book alone infusion devices when rapid infusion is desira le. As shown in Ta le 14.3, flow through introducer catheters is almost as rapid as flow through standard i ntravenous P.216 tu ing. However, as also shown in Ta le 14.3, the side infusion ports on introducer catheters create an impediment to flow (22). Thus, when int roducer catheters are used for rapid volume infusion, the side infusion port on the catheter must e ypassed. TABLE 14.3. RAPID VOLUME INFUSION VISCOSITY Viscosity is descri ed in Chapter 1 as the property of a fluid that resists chan ges in flow rate. As shown in Ta le 1.2, the viscosity of lood is primarily a f unction of the erythrocyte density (i. e., hematocrit). The graph in Figure 14.4 shows the influence of cell density (viscosity) on the infusion rate of differe nt resuscitation fluids (23). The force responsi le for flow in this case is gra vity, and the infusion device is a 16-gauge, 2-inch catheter similar to the ones used to cannulate peripheral veins. The acellular fluids (i.e., water and 5% al umin) have P.217 the highest flow rates, whereas the erythrocyte concentrate (p acked RBCs) has the slowest flow rate. This demonstrates the inverse relationshi p etween viscosity and flow rate in the HagenPoiseuille equation (see Chapter 1) . http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 28)03-May-05 17:09:42

Ovid: ICU Book Figure 14.4. Gravity-driven infusion rates of lood products and intravenous flu ids flowing through the same sized catheter. (From Dula DJ et al. Flow rate vari ance of commonly used IV infusion techniques. J Trauma 1981;21:480482.) There is a popular misconception that colloid fluids such as plasma or al umin s olutions flow more sluggishly than water or electrolyte solutions. However, eca use viscosity is primarily a function of cellular density, all acellular fluids should have equivalent flow properties. This is demonstrated y the equivalent f low rates for water and 5% al umin shown in Figure 14.4. Therefore, colloid solu tions (containing large molecular weight su stances) will infuse just as rapidly as crystalloid (electrolyte) solutions. AUTOTRANSFUSION Autotransfusion maneuvers are meant to promote venous return y shifting lood v olume from the legs toward the heart. There are two autotransfusion methods: od y tilt and pneumatic compression. Unfortunately, neither method is successful in achieving the desired effect, as descri ed elow. THE TRENDELENBURG POSITION Elevation of the pelvis a ove the horizontal plane in the supine position was in troduced in the latter part of the 19th century as a method of facilitating surg ical exposure of the pelvic organs. The originator was a surgeon named Friedrich Trendelen urg, who specialized in the surgical correction of vesicovaginal fist ulas (5). The ody position that now ears his name was later adopted as an anti shock position during World War I and gained widespread popularity despite http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 28)03-May-05 17:09:42

Ovid: ICU Book a lack of evidence for its efficacy. This popularity continues, as does the lack of evidence for efficacy (24,25,26 and 27). Hemodynamic Effects The hemodynamic effects of the Trendelen urg position (legs elevated and head e low the horizontal plane) are shown in Ta le 14.4. P.218 The data in this ta le are from a study we performed on postoperative patients with indwelling pulmonar y artery catheters who had evidence of severe hypovolemia (i.e., low cardiac fil ling pressures and hypotension) (24). The hemodynamic measurements were o tained in the supine position, and then repeated after the patients were placed in a p osition with the legs elevated 45 degrees a ove the horizontal plane and the hea d placed 15 degrees elow the horizontal plane. As shown in the ta le, the chang e in position was associated with significant increases in the mean arterial pre ssure, wedge (left-ventricular filling) pressure, and systemic vascular resistan ce, while the cardiac output remained the same. This lack of an effect on the ca rdiac output indicates that the Trendelen urg position does not promote venous r eturn to the heart. The increase in the wedge pressure can e due to an increase in intrathoracic pressure (transmitted into the pulmonary capillaries) caused y cephalad displacement of the diaphragm during the ody tilt. The increase in lood pressure during ody tilt is likely due to systemic vasoconstriction (indic ated y the rise in systemic vascular resistance). These o servations are consis tent with other studies in animals and humans (25,26 and 27). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 28)03-May-05 17:09:42

Ovid: ICU Book TABLE 14.4. HEMODYNAMIC EFFECTS OF THE TRENDELENBURG POSITION IN HYPOVOLEMIC ICU PATIENTS Why the Trendelen urg Position Cannot Work The ina ility of the Trendelen urg position to augment cardiac output is not sur prising, and it is explained y the high capacitance (distensi ility) of the ven ous circulation. To augment cardiac output, the Trendelen urg position must incr ease the pressure gradient from peripheral to central veins, which would then in crease venous return. However, the venous system is a highcapacitance system des igned to a sor pressure and act as a volume reservoir. Thus, when pressure is a pplied to a vein, the vein distends and increases its volume capacity. This dist ensi ility then limits any change in venous pressure, and this, in turn, counter acts any increase in the pressure gradient etween peripheral and central veins. The venous system is more likely to transmit pressure when the veins are volume overloaded and less distensi le. In other words, the Trendelen urg position is more likely to e effective (i.e., to augment venous return) during volume overl oad, not volume depletion. Thus, the Trendelen urg position has not een, and ne ver will e, effective in promoting venous return (cardiac output) in hypovolemi a. As such, this maneuver should e a andoned for the management of hypovolemia. It remains axiomatic that the effective treatment for hypovolemia is volume rep lacement. PNEUMATIC COMPRESSION Pneumatic compression of leg veins has also een used to promote venous return i n acute hemorrhage. However, as descri ed for the Trendelen urg position, pneuma tic compression of peripheral veins seems to augment the lood pressure y incre asing peripheral vascular P.219 resistance (particularly in the a domen), and no t y increasing venous return (28). In fact, this maneuver can actually promote lood loss in penetrating thoracic injuries (28). At the present time, pneumatic antishock trousers are used predominantly for prehospital sta ilization of trauma victims (i.e., when inflated, pneumatic trousers can produce a tourniquet effec t that helps control pelvic and intraa dominal hemorrhage).

RESUSCITATION STRATEGIES The universal goal of resuscitation is to maintain oxygen uptake (VO2) into the vital organs and there y sustain aero ic meta olism (29). The determinants of ox ygen uptake are identified in the equation shown elow (as descri ed in Chapter 2). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 28)03-May-05 17:09:42

Ovid: ICU Book The factors that pose a risk to the VO2 in acute lood loss are the cardiac outp ut (Q) and the hemoglo in concentration (H ). The consequences of a low cardiac output are far more threatening than the consequences of anemia. Therefore, the first priority in acute lood loss is to preserve lood flow (cardiac output), w hile correcting erythrocyte deficits is a secondary goal. FLOW-DIRECTED RESUSCITATION The a ility of different resuscitation fluids to promote the cardiac output is s hown in Figure 14.5 (29). Each of the fluids was infused in P.220 the volume ind icated over 60 minutes, and the height of the columns indicates the change in ca rdiac index recorded at the end of the infusion period. The volumes of whole lo od, packed cells, and dextran-40 are equivalent (500 mL), whereas the infusion v olume of lactated Ringer's (1 L) is dou le that of the other fluids. The most po tent fluid for promoting the cardiac output is the dextran-40, and the least pot ent resuscitation fluid is the erythrocyte concentrate (packed RBCs). Figure 14.5. The a ility of different resuscitation fluids to promote cardiac ou tput. (From Shoemaker WC. Intensive Care Med 1987;13:230243.) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 28)03-May-05 17:09:42

Ovid: ICU Book Blood Products If promoting cardiac output is the first priority in the management of acute hem orrhage, then lood is not the ideal resuscitation fluid for acute lood loss e cause lood products do not promote lood flow as well as some acellular fluids (such as dextran-40 in Figure 14.5). The density of erythrocytes impedes the a i lity of lood products to promote lood flow (a viscosity effect). In fact, the administration of erythrocyte concentrates (packed cells) can reduce lood flow and aggravate tissue oxygen deficits (30,31). Asanguinous Fluids There are two types of fluids other than lood products shown in Figure 14.5. Th e dextran-40 represents one type of fluid, which is characterized y the presenc e of large molecular weight su stances that do not pass easily from one fluid co mpartment to another. These large molecules with limited mo ility prevent the eg ress of water, and this maintains the volume of the fluid compartment. Fluids of this type that restrict water movement are called colloids (from the Greek word for glue). The other type of fluid is the lactated Ringer's solution, which is an electrolyte solution devoid of large molecules that impede water movement. Fl uids of this type that allow water to move freely from one fluid compartment to another are called crystalloids. The graph in Figure 14.5 reveals a marked diffe rence in the a ility of colloid and crystalloid fluids to augment lood flow. Th is difference cannot e explained y viscosity, ecause oth types of fluids are acellular and have negligi le viscosities. The difference is due to the differe nces in volume distri ution. Crystalloid fluids are primarily sodium chloride so lutions, and ecause sodium is distri uted evenly in the extracellular fluid, cr ystalloid fluids will also distri ute evenly in the extracellular fluid. Because plasma represents only 20% of the extracellular fluid, only 20% of an aliquot o f crystalloid fluid will remain in the vascular space, while the remaining 80% w ill add to the interstitial space. On the other hand, colloid fluids, ecause of their limited mo ility, are more prone to remain in the vascular space. In the case of dextran-40, etween 75 and 80% of the infused volume will remain in the plasma. Therefore, the enhanced effect of colloids on the cardiac output is due to the greater tendency of colloid fluids to increase P.221 the plasma volume. T he increase in plasma volume augments cardiac output not only y increasing vent ricular preload (volume effect) ut also y decreasing ventricular afterload (di lutional effect on lood viscosity). The following statements summarize some sal ient features of resuscitation fluids (6,8,9,29). q http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 28)03-May-05 17:09:42

Ovid: ICU Book Colloid fluids are superior to lood products and crystalloid fluids for promoti ng lood flow (cardiac output). q

Crystalloid fluids primarily fill the interstitial space. q To have equivalent effects on cardiac output, the volume of crystalloid fluid in fused must e at least three times greater than the volume of colloid infusion. Despite the superior performance of colloid fluids, crystalloid fluids are more popular for volume resuscitation. This preference is partly due to the lower cos t of crystalloid fluids, and partly due to ha it. Chapter 15 expands further on the various likes and dislikes of colloid and crystalloid fluids. Fluid Administration The standard approach to volume resuscitation in hypovolemic shock is to rapidly administer 2 L of crystalloid fluid as a olus (11), or infuse crystalloid at a rate of 6 mL/min/kg (31a). If a favora le response is seen, then crystalloid fl uids are continued using the endpoints discussed elow. If there is not a favora le response, then colloid fluids and lood products are added to the regimen. T he rate of infusion is dictated y the clinical condition of the patient and wil l vary widely. Infusion rates as high as 2.5 mL/second can e delivered through introducer catheters (22). A rough estimate of the resuscitation volume can e d erived as follows (see Ta le 14.5). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 28)03-May-05 17:09:42

Erythrocyte concentrates (packed cells) do not increase (and can decrease) flow, and thus they should never e used for volume resuscitation. q

lood

Ovid: ICU Book TABLE 14.5. A SIMPLE METHOD FOR DETERMINING RESUSCITATION VOLUME P.222 q

Estimate the percent volume loss using the classification system in Ta le 14.2. Clinical manifestations may not e prominent in mild volume loss, ut aggressive volume infusion is not indicated in that setting. Once lood pressure drops in the supine position, there is at least a 30% decrease in lood volume. q Calculate the volume deficit y multiplying the estimated normal lood volume an d the percent loss. This is a quantitative estimate of the volume needs in each patient. q

Blood replacement is usually not necessary for Class I and Class II hemorrhage ( Ta le 14.2). r Although colloid fluids can differ in their a ility to remain in the vascular http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 28)03-May-05 17:09:42

Determine the resuscitation volume of specific fluids using the rules listed ow. r

Estimate the normal lood volume using Ta le 14.1. Remem er to use lean ght and to adjust for o esity and advanced age. q

ody wei

el

Ovid: ICU Book compartment, a general rule of thum is to assume that no less than 50% and no m ore than 75% of infused colloid will remain in the vascular space. This translat es to a replacement volume for colloid fluids that is 1.5 to 2 times the volume deficit. r The resuscitation volume for crystalloid fluids is 4 times the volume deficit, o r 3 times the resuscitation volume for colloids (8). ENDPOINTS The following are common endpoints of volume resuscitation: q CVP = 15 mm Hg (32) q Wedge pressure = 10 to 12 mmHg (33) q Cardiac index > 3 L/min/m2 q Oxygen uptake (VO2) > 100 mL/min/m2 q Blood lactate < 4 mmol/L q Base deficit 3 to +3 mmol/L These endpoints represent normal hemodynamic parameters for adults. Base Deficit The ase deficit (millimoles of ase needed to correct the pH of 1 L of whole l ood to 7.40) has een shown to correlate with volume deficits and with mortality in trauma victims (34). As a result, this parameter has een recommended as a v alua le guide to volume therapy. The ase deficit is routinely calculated y man y automated lood gas analyzers, and the calculated ase deficit is included in many lood gas reports. The normal range for the ase deficit is 3 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 28)03-May-05 17:09:42

Ovid: ICU Book mmol/L on either side of zero. Elevations in ase deficit can e P.223 classifie d as mild (2 to 5 mmol/L), moderate (6 to 14 mmol/L), or severe (> 15 mmol/L). T he ase deficit that remains elevated during volume infusion is an indication of ongoing tissue ischemia. This measurement is pro a ly a poor man's lactate dete rmination, and it shows promise as a readily availa le and easy-to-use index of ischemic acid production in tissues. ERYTHROCYTE RESUSCITATION In the second stage of management, attention is directed to deficits in oxygen c arrying capacity. The use of lood transfusions to correct normovolemic anemia i s discussed in detail in Chapter 44. The current practice of transfusing red lo od cells ased on hemoglo in determinations has a solutely no scientific asis ( 35,36). A serum hemoglo in concentration provides no information a out tissue ox ygenation, nor is it synonymous with oxygen carrying capacity. To illustrate the latter point, when dehydration increases the serum hemoglo in concentration, do es it also increase the oxygen carrying capacity of lood? The move away from he moglo in and hematocrit is apparent in the Clinical Guideline on Elective Red Ce ll Transfusions pu lished y the American College of Physicians (35). The guidel ine states that for asymptomatic patients with anemia, In the a sence of patient risks (e. g., active coronary disease), transfusion is not indicated, independen t of hemoglo in level(italics mine). Oxygen Transport Varia les A more rational approach to red cell transfusions is to employ the oxygen transp ort varia les and the lood lactate level to assess tissue oxygenation (as descr i ed in Chapter 13) (37). The following conditions would e indications for tran sfusion in normovolemic anemia: q Oxygen uptake (VO2) elow the normal range (indicating an oxygen de t) q Blood lactate greater than 4 mmol/L (regardless of the VO2) q Oxygen extraction ratio (O2ER) greater than 0.5 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 28)03-May-05 17:09:42

Ovid: ICU Book The VO2 can also e used to evaluate the response to transfusion therapy. An inc rease in VO2 after transfusion of one unit of lood or packed cells indicates a eneficial response. Transfusion of single units of lood can then e continued until the VO2 is no longer augmented. (For more information on the correction of anemia see Chapter 44.) P.224

OTHER CONCERNS RESUSCITATION-INDUCED HEMORRHAGE Although the prevailing opinion favors aggressive volume resuscitation for hemor rhage, evidence from oth animal studies (38) and clinical trials (39) indicates that volume resuscitation to normotension can actually promote continued lood loss. This is an important o servation for two reasons. First, it indicates that the lood pressure is not an appropriate endpoint for the resuscitation of hypo volemic shock (not, at least, until holes in lood vessels are sealed). Second, and more important, it implies that therapy aimed at achieving normal clinical p arameters (which is the general approach in modern medicine) is not appropriate when the human ody is su jected to a normal conditions. Normal clinical paramet ers are a desira le goal only when a normal (pathologic) conditions are correcte d. POST-RESUSCITATION INJURY Injury to the major organs can continue una ated following apparently successful resuscitation of hypovolemic shock. This postresuscitation injury can e progre ssive and can involve several organs (the rain and intestinal tract appear to e most suscepti le). Two processes have een implicated in the pathogenesis of t his disorder: the no-reflow phenomenon and reperfusion injury. No-Reflow Phenomenon Defects in microvascular perfusion can persist despite the resuscitation of hypo volemic shock to premor id levels of lood pressure and cardiac output (40,41,42 ,43 and 44). Several mechanisms have een proposed for this phenomenon, includin g calcium-induced vasoconstriction (42), leukocyte plugging (43), and vascular c ompression from the accumulation of edema fluid (43,44). Persistent hypoperfusio n in the splanchnic circulation can lead to translocation of intestinal pathogen s and postresuscitation septicemia (45). At present, there is no therapy that pr events the no-reflow phenomenon. Because its occurrence and severity seem to e related to the duration of ischemia, prompt resuscitation should help to http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 28)03-May-05 17:09:42

Ovid: ICU Book prevent this complication. Reperfusion Injury Postresuscitation injury is also attri uted to toxic meta olites that accumulate during the period of ischemia and are washed away during P.225 reperfusion, cau sing damage to remote tissues. Toxic oxygen meta olites have een implicated in this process (46). Two potential sources of enhanced oxidant production are neut rophil activation and generation of superoxide radicals from the oxidation of hy poxanthine (44,46). Despite the proposed role of oxidant injury in the reperfusi on period, preliminary studies using antioxidants to prevent this phenomenon hav e een disappointing (46). REFERENCES GENERAL TEXTS Committee on Trauma. American College of Surgeons. Early care of the injured pat ient. Philadelphia, B.C. Decker, 1990. Geller ER (ed). Shock and resuscitation. New York, McGraw-Hill, 1993 CLASSIC STUDIES 1. Crowell JW, Smith EE. Oxygen deficit and irreversi le hermorrhagic shock. Am J Physiol. 1964;206:313316. 2. Weil MH, Afifi AA. Experimental and clinical studies on lactate and pyruvate as indicators of the severity of acute circulatory failure (shock). Circulation 1970;41:989 1001. 3. Moore FD. Effects of hemorrhage on ody composition. New Engl J Med 1965;273: 567577. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (23 of 28)03-May-05 17:09:42

Ovid: ICU Book 4. Shires GT, Coln D, Carrico J, Lightfoot S. Fluid therapy in hemorrhagic shock . Arch Surg 1964;88:688693. 5. Trendelen urg F. The elevated pelvic position for operations within the a dom inal cavity. Med Classics 1940;4:964968. (Translation of original manuscript.) REVIEWS 5A. Schadt JC, Lud rook J. Hemodynamic and neurohumoral responses to acute hypov olemia in conscious animals. Am J Physiol 1991;260:H305318. 6. Falk JL, O'Brien JF, Kerr R. Fluid resuscitation in traumatic hemorrhagic sho ck. Crit Care Clin 1992;8:323340. (123 References).

8. Imm A, Carlson RW. Fluid resuscitation in circulatory shock. Crit Care Clin 1 993;9:313 333. (73 References). 9. Domsky MF, Wilson RF. Hemodynamic resuscitation. Crit Care Clin 1993;9:715726 . (47 References). CLINICAL PRESENTATION 10. Walker RH (ed). Technical manual of the American Association of Blood Banks. 10th ed., Arlington, VA, American Association of Blood Banks, 1990:650. P.226 11. Committee on Trauma. Advanced trauma life support student manual. Chicago, A merican College of Surgeons, 1989:4759. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (24 of 28)03-May-05 17:09:42

7. Nacht A. The use of 1 References).

lood products in shock. Crit Care Clin 1992;8:255293. (14

Ovid: ICU Book 12. Williams TM, Knoop R. The clinical use of orthostatic vital signs. In Ro ert s JR, Hedges JR (eds). Clinical procedures in emergency medicine. Philadelphia, W.B. Saunders, 1991:445449. 13. Moore KI, Newton K. Orthostatic heart rates and lood pressures in healthy y oung women and men. Heart & Lung 1986;611617. 14. Shippy CR, Appel PL, Shoemaker WC. Relia ility of clinical monitoring to ass ess lood volume in critically ill patients. Crit Care Med 1984;12:107112. 15. Cordts PR, LaMorte WW, Fisher JB, et al. Poor predictive value of hematocrit and hemodynamic parameters for erythrocyte deficits after extensive vascular op erations. Surg Gynecol O stet 1992;175:243248. 16. Walley KR, Cooper DJ. Diastolic stiffness impairs left ventricular function during hypovolemic shock in pigs. Am J Physiol 1991;260:H702712. 17. Amoroso P, Greenwood RN. Posture and central venous pressure measurements in circulatory volume depletion. Lancet 1989;i:258260. 18. Weil MH, Bisera J, Trevino RP, Rackow EC. Cardiac output and end-tidal car o n dioxide. Crit Care Med 1985;13:907909. 19. Stamler KD. Effect of crystalloid infusion on hematocrit in non leeding pati ents, with applications to clinical traumatology. Ann Emerg Med 1989;18:747749. RATE OF VOLUME REPLACEMENT 20. Buchman TG, Menker JB, Lipsett PA. Strategies for trauma resuscitation. Surg Gynecol O stet 1991;172:812. 21. Mateer JR, Thompson BM, Aprahamian C, Darin JC. Rapid fluid resuscitation wi th central venous catheters. Ann Emerg Med 1983;12:149152. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (25 of 28)03-May-05 17:09:42

Ovid: ICU Book 22. Hyman SA, Smith DW, England R, et al. Pulmonary artery catheter introducers: Do the component parts affect flow rate? Anesth Analg 1991;73:573575. 23. Dula DJ, Muller A, Donovan JW. Flow rate of commonly used IV techniques. J T rauma 1981;21:480482. AUTOTRANSFUSION 24. Sing R, O'Hara D, Sawyer MAJ, Parino PL. Trendelen urg position and oxygen t ransport in hypovolemic adults. Ann Emerg Med 1994;23:564568. 25. Taylor J, Weil MH. Failure of Trendelen urg position to improve circulation during clinical shock. Surg Gynecol O stet 1967;122:10051010. 26. Bivins HG, Knopp R, dos Santos PAL. Blood volume distri ution in the Trendel en urg position. Ann Emerg Med 1985;14:641643. 27. Gaffney FA, Bastian BC, Thal ER, Atkins JM. Passive leg raising does not pro duce a significant autotransfusion effect. J. Trauma 1982;22:190193. 28. Ali J, Vander y B, Purcell C. The effect of pneumatic antishock garment (PAS G) on hemodynamics, hemorrhage, and survival in penetrating thoracic aortic inju ry. J Trauma 1991;31:846851. RESUSCITATION STRATEGIES 29. Shoemaker WC. Relationship of oxygen transport patterns to the pathophysiolo gy and therapy of shock states. Intensive Care Med 1987;213:230243. P.227 30. Marik PE, Si ald WJ. Effect of stored- lood transfusion on oxygen delivery in http://gateway.ut.ovid.com/gw1/ovidwe .cgi (26 of 28)03-May-05 17:09:42

Ovid: ICU Book patients with sepsis. JAMA 1993;269:30243029. 31. Silverman HJ, Tuma P. Gastric tonometry in patients with sepsis. Effects of do utamine infusions and packed red lood cell transfusions. Chest 1992;102:18418 8. 31A. Dula DJ, Lutz P, Vogel MF, et al. Rapid flow rates for the resuscitation of hypovolemic shock. Ann Emerg Med 1985;14:303306. 32. Shoemaker WC, Fleming AW. Resuscitation of the trauma patient. Restoration o f hemodynamic functions using clinical algorithms. Ann Emerg Med 1986;12:14371444 . 33. Packman MI, Rackow EC. Optimum left heart filling pressures during fluid res uscitation of patients with hypovolemic and septic shock. Crit Care Med 1983;11: 165 169. 34. Davis JW, Shackford SR, Hol rook TL. Base deficit as a sensitive indicator o f compensated shock and tissue oxygen utilization. Surg Gynecol O stet 1991;173: 473 478.

36. Consensus Conference. Perioperative red lood cell transfusion. JAMA 1988;26 0:27002703. 37. Levy PS, Chavez RP, Crystal GJ, et al. Oxygen extraction ratio: A valid indi cator of transfusion need in limited coronary vascular reserve. J Trauma 1992;32 :769774. OTHER CONCERNS 38. Stern SA, Dronen SC, Birrer P, Wang X. Effect of lood pressure on hemorrhag e volume and survival in a near-fatal hemorrhage model incorporating a vascular injury. Ann Emerg Med 1993;22:155163. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (27 of 28)03-May-05 17:09:42

35. American College of Physicians. Practice strategies for elective red ell transfusion. Annals Intern Med 1992;116:403406.

lood c

Ovid: ICU Book 39. Bickell WH, Wall MJ, Pepe PE, et al. Immediate versus delayed fluid resuscit ation for hypotensive patients with penetrating torso injuries. New Engl J Med 1 994;331:1105 1109. 40. Ames A, III, Wright RL, Kowada M, et al. Cere ral ischemia; the no-reflow ph enomenon. Am J Pathol 1968;52:437442. 41. Wang P, Hauptman JG, Chaudry IH. Hemorrhage produces depression of microvasc ular lood flow which persists despite fluid resuscitation. Circ Shock 1990;32:3 07318. 42. Wang P, Zheng FB, Dean RE, Chaudry IH. Diltiazem administration after crysta lloid resuscitation restores active hepatocellular function and hepatic lood fl ow after severe hemorrhagic shock. Surgery 1991;110:390397. 43. Carden DL, Smith K, Kortyhuis RJ. Neutrophil-mediated microvascular dysfunct ion in postischemic canine skeletal muscle. Circ Res 1990;66:14361444.

45. Koziol JM, Rush BF, Smith SM, Machiedo GW. Occurrence of acteremia during a nd after hemorrhagic shock. J Trauma 1988;28:1016. 46. Grace PA. Ischemia-reperfusion injury. Br J Surg 1994;81:637647. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (28 of 28)03-May-05 17:09:42

44. Mellow CG, Knight KR, Angel MF. The iochemical Surg Res 1992;52:226232.

asis of secondary edema. J

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 15 COLLOID AND CRYSTALLOID RESUSCITATION In 1861, Thomas Graham's investigations on diffusion led him to classify su stan ces as crystalloids or colloids ased on their a ility to diffuse through a parc hment mem rane. Crystalloids passed readily through the mem rane, whereas colloi ds (from the Greek word for glue) did not. Intravenous fluids are similarly clas sified ased on their a ility to pass through arriers separating ody fluid com partments, particularly the one etween intravascular and extravascular (interst itial) fluid compartments. This chapter descri es the salient features of crysta lloid and colloid fluids, oth individually and as a group. This is a must-know topic in the care of hospitalized patients, and several reviews are included at the end of the chapter to supplement the text (1,2,3 and 4). CRYSTALLOID FLUIDS The principal component of crystalloid fluids is the inorganic salt sodium chlor ide (NaCl). Sodium is the most a undant solute in the extracellular fluids, and it is distri uted uniformly throughout the extracellular space. Because 75 to 80 % of the extracellular fluids are located in the extravascular (interstitial) sp ace, a similar proportion of the total ody sodium is in the interstitial fluids . Exogenously administered sodium follows the same distri ution, so 75 to 80% of the volume of sodium- ased intravenous fluids are distri uted in the interstiti al space. This means that the predominant effect of volume resuscitation with cr ystalloid fluids is to expand the interstitial volume rather than the plasma vol ume. VOLEMIC EFFECTS The change in interstitial volume and plasma volume associated with crystalloid fluid resuscitation is shown in Figure 15.1. As indicated P.229 y the horizonta l ar that is second from the top, infusion of 1 L of 0.9% sodium chloride (isot onic saline) adds 275 mL to the plasma volume and 825 mL to the interstitial vol ume (4). Note that the total volume expansion (1100 mL) is slightly greater than the infused volume. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 17)03-May-05 17:10:10

Ovid: ICU Book This is the result of a fluid shift from the intracellular to extracellular spac e, which occurs ecause isotonic saline is actually hypertonic to the extracellu lar fluids. The salient features of isotonic saline and other crystalloid fluids are summarized in Ta le 15.1 (along with some comparative features of plasma). Figure 15.1. The influence of colloid and crystalloid fluids on the volume of th e extracellular fluid compartments. (Data from Imm A, Carlson RW. Fluid resuscit ation in circulatory shock. Crit Care Clin 1993;9:313.) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 17)03-May-05 17:10:10

Ovid: ICU Book TABLE 15.1. COMPOSITION OF INTRAVENOUS CRYSTALLOID FLUIDS P.230 ISOTONIC SALINE The prototype crystalloid fluid is 0.9% sodium chloride (NaCl), also called isot onic saline or normal saline. The latter term is inappropriate ecause a one nor mal (1 N) NaCl solution contains 58 g NaCl per liter (the com ined molecular wei ghts of sodium and chloride), whereas isotonic (0.9%) NaCl contains only 9 g NaC l per liter. Features As shown in Ta le 15.1, isotonic saline has higher concentrations of sodium and chloride than plasma, and it is slightly hypertonic to plasma. The pH of isotoni c saline is also considera ly lower than the plasma pH. These differences are ra rely of any clinical significance. Disadvantages The chloride content of isotonic saline is particularly high relative to plasma (154 mEq/L versus 103 mEq/L, respectively), so hyperchloremic meta olic acidosis is a potential risk with largevolume isotonic saline resuscitation. Hyperchlore mia has een reported, ut acidosis is rare (1,5). LACTATED RINGER'S Ringer's solution was introduced in 1880 y Sydney Ringer, a British physician a nd research investigator who studied mechanisms of cardiac contraction (6). The solution was designed to promote the contraction of isolated frog hearts, and co ntained calcium and potassium in a sodium chloride diluent. In the 1930s, an Ame rican pediatrician named Alexis Hartmann proposed the addition of sodium lactate uffer to Ringer's solution for the treatment of meta olic acidoses. The lactat ed Ringer's solution, also known as Hartmann's solution, gradually gained in pop ularity and eventually replaced the standard Ringer's solution for routine intra venous therapy. The composition of lactated Ringer's solution is shown in Ta le 15.1. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 17)03-May-05 17:10:10

Ovid: ICU Book Features Lactated Ringer's solution contains potassium and calcium in concentrations that approximate the free (ionic) concentrations in plasma. The addition of these ca tions requires a reduction in sodium concentration for electrical neutrality, so lactated Ringer's solution has less sodium than isotonic saline. The addition o f lactate (28 mEq/L) similarly requires a reduction in chloride concentration, a nd the chloride in lactated Ringer's more closely approximates plasma chloride l evels than does isotonic saline. Despite the differences in composition, there i s no evidence that P.231 lactated Ringer's provides any enefit over isotonic sa line. Furthermore, there is no evidence that the lactate in Ringer's solution pr ovides any uffer effect. Disadvantages The calcium in lactated Ringer's can ind to certain drugs and reduce their ioa vaila ility and efficacy. A list of intravenous products that may e incompati l e with lactated Ringer's solution is shown in Ta le 15.2 (6). Of particular note is calcium inding to the citrated anticoagulant in lood products. This can in activate the anticoagulant and promote the formation of clots in donor lood (7) . For this reason, lactated Ringer's solution is contraindicated as a diluent fo r lood transfusions (7). TABLE 15.2. INCOMPATIBILITIES WITH RINGER'S SOLUTIONS* http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 17)03-May-05 17:10:10

Ovid: ICU Book NORMOSOL OR PLASMA-LYTE Features The major feature of these solutions is the added uffer capacity, which gives t hem a pH that is equivalent to that of plasma. An additional feature is the addi tion of magnesium, which may provide some enefit in light of the high incidence of magnesium depletion in hospitalized patients (see Chapter 42). Disadvantages Magnesium administration can promote hypermagnesemia in renal insufficiency and can counteract compensatory vasoconstriction and promote hypotension in low flow states. P.232 DEXTROSE SOLUTIONS Dextrose is a common additive in intravenous solutions, for reasons that are unc lear. A 5% dextrose-in-water solution is not an effective volume expander, as sh own in Figure 15.1. The use of 5% dextrose solutions was originally intended to supply nonprotein calories and thus provide a protein-sparing effect. However, t otal enteral and parenteral nutrition is now the standard of care for providing daily energy requirements, and the use of 5% dextrose solutions to provide calor ies is o solete. Features A 5% dextrose solution (50 g dextrose per liter) provides 170 kcal per liter (3. 4 kcal/g dextrose). Disadvantages The addition of dextrose to intravenous fluids increases osmolarity (50 g of dex trose adds 278 mosm to an intravenous fluid) and creates a hypertonic infusion w hen 5% dextrose is added to lactated Ringer's solution (525 mOsm/L) or isotonic saline (560 mOsm/L). If glucose use is impaired (as is common in critically ill patients), the infused glucose accumulates and creates an undesira le osmotic fo rce that can promote cell dehydration. Other undesira le effects of glucose infu sions in critically ill patients include enhanced CO2 production (which can e a urden in ventilator-dependent patients) (8), enhanced lactate production (9,10 ), and aggravation of ischemic rain injury (11). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 17)03-May-05 17:10:10

Ovid: ICU Book Lactate Production The proportion of a glucose load that contri utes to lactate formation can incre ase from 5% in healthy su jects to 85% in critically ill patients (9). This can produce an increase in circulating lactate levels, even when infusing 5% dextros e solutions. This is demonstrated in Figure 15.2. In this case, patients undergo ing a dominal aortic aneurysm surgery were given either a Ringer's solution or a 5% dextrose solution intraoperatively to maintain normal cardiac filling pressu res. As shown, the 5% dextrose infusions were associated with a 125% increase in arterial lactate levels (from 1.85 to 4.15 mmol/L). Thus, in patients with circ ulatory compromise, a normal glucose meta olism can transform glucose from a sou rce of useful energy to a source of toxin production. The deleterious meta olic effects of glucose infusions are discussed again in Chapter 17 (on resuscitation of cardiac arrest) and in Chapter 48 (on parenteral nutrition). Figure 15.2. The influence of intravenous fluid therapy with and without dextros e on lood lactate levels in patients undergoing a dominal aortic aneurysm surge ry. Each column represents data (mean and standard deviation) from 10 patients. Total intraoperative volumes for each fluid are indicated in parentheses. (From Degoute CS et al. Intraoperative glucose infusion and lood lactate: endocrine m eta olic relationships during a dominal aortic surgery. Anesthesiology 1989;17:3 55361.) The disadvantages noted a ove, when com ined with a lack of documented P.233 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 17)03-May-05 17:10:10

Ovid: ICU Book enefit, favor the recommendation that the routine use of 5% dextrose infusions e a andoned in critically ill patients.

COLLOID FLUIDS As mentioned earlier, colloids are large molecules that do not pass across diffu sional arriers as readily as crystalloids. Colloid fluids infused into the vasc ular space therefore have a greater tendency to stay put and enhance the plasma volume than do crystalloid fluids. This is illustrated in Figure 15.1. The collo id fluid in this case is 5% al umin, and as demonstrated, the plasma expansion w ith this colloid fluid is nearly twice that produced y an equivalent volume of isotonic saline (500 mL versus 275 mL, respectively). This is the principal ene fit of colloid fluid resuscitation: more effective resuscitation of plasma volum e than that produced y crystalloid fluids. The potency of colloid fluids as pla sma volume expanders can differ with individual fluids, as shown in Ta le 15.3. Much of this potency is related to the colloid osmotic pressure exerted y each fluid. TABLE 15.3. CHARACTERISTICS OF INTRAVENOUS COLLOID FLUIDS* COLLOID OSMOTIC PRESSURE Large solute molecules that do not move freely across arriers separating fluid compartments create a force that draws water into the P.234 large solute compart ment. This force opposes the hydrostatic pressure (which favors the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 17)03-May-05 17:10:10

Ovid: ICU Book movement of water out of a fluid compartment) and is called the colloid osmotic pressure (COP) or oncotic pressure. The COP of the commercially availa le colloi d fluids is shown in Ta le 15.3. As would e expected, the a ility of each fluid to expand the plasma volume is directly related to the COP; that is, the higher the COP, the greater the volume expansion. If the COP of a colloid fluid is gre ater than the COP of plasma (i.e., greater than 25 mm Hg), the plasma volume exp ansion exceeds the infused volume. This is demonstrated in Ta le 15.3 y the 25% al umin solution, which has a COP of 70 mm Hg and a plasma volume expansion tha t is 4 to 5 times the infused volume. ALBUMIN Al umin is a transport protein that is responsi le for 75% of the oncotic pressu re of plasma (1,12,13 and 14). Heat-treated preparations of human serum al umin are commercially availa le in a 5% solution (50 g/L) and a 25% solution (250 g/L ) in an isotonic saline diluent. The 25% solution is given in small volumes (50 to 100 mL) and ecause the accompanying sodium load is small, 25% al umin is als o called salt-poor al umin. Features A 5% al umin solution (50 g/L or 5 g/dL) has a COP of 20 mm Hg and thus is simil ar in oncotic activity to plasma. Approximately half of the infused volume of 5% al umin stays in the vascular space, as shown in Figure 15.1. The oncotic effec ts of al umin last 12 to 18 hours (1,4). The 25% al umin solution has a COP of 7 0 mm Hg and expands the plasma volume y 4 to 5 times the volume infused. Thus, infusion P.235 of 100 mL of 25% al umin can increase the plasma volume 400 to 50 0 mL (1). This plasma volume expansion occurs at the expense of the interstitial fluid volume, so 25% al umin should not e used for volume resuscitation in hyp ovolemia. It is intended for shifting fluid from the interstitial space to the v ascular space in hypoproteinemic conditions, although the wisdom of this applica tion is questiona le (2). Disadvantages Because al umin preparations are heat-treated, there is no risk of viral transmi ssion (including human immunodeficiency virus). Allergic reactions are rare, and although coagulopathies can occur, most are dilutional and not accompanied y leeding (1,4). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 17)03-May-05 17:10:10

Ovid: ICU Book HETASTARCH Hetastarch is a synthetic colloid availa le as a 6% solution in isotonic saline. It contains amylopectin molecules that vary in size from a few hundred to over a million daltons. The average molecular weight of the starch molecules is equiv alent to that of al umin, and the colloid effects are equivalent to those of 5% al umin (1,4). The main advantage of hetastarch over al umin is its lower cost. Features Hetastarch is slightly more potent than 5% al umin as a colloid. It has a higher COP than 5% al umin (30 versus 20 mm Hg, respectively) and causes a greater pla sma volume expansion (up to 30% greater than the infused volume). It also has a long elimination half-life (17 days), ut this is misleading ecause the oncotic effects of hetastarch disappear within 24 hours (1). Disadvantages Hetastarch molecules are constantly cleaved y amylase enzymes in the loodstrea m efore their clearance y the kidneys. Serum amylase levels are often elevated (2 to 3 times a ove normal levels) for the first few days after hetastarch infu sion, and return to normal at 5 to 7 days after fluid therapy (15). This hyperam ylasemia should not e mistaken for early pancreatitis. Serum lipase levels rema in normal, which is an important distinguishing feature (15). Anaphylactic react ions to hetastarch are decidedly rare (incidence as low as 0.0004%) (15). La ora tory test coagulopathy (prolonged partial throm oplastin time from an interactio n with Factor VIII) can occur, ut is not accompanied y leeding (15,16). Coagu lopathy claims have dogged hetastarch for years, without evidence of hetastarchinduced leeding (1,15). P.236 PENTASTARCH Pentastarch is a low-molecular-weight-derivative of hetastarch that is availa le as a 10% solution in isotonic saline. Although it is not currently approved for clinical use in the United States, there is considera le evidence indicating th at pentastarch is an effective and safe plasma volume expander (1,17,18). Features Pentastarch contains smaller ut more numerous starch molecules than hetastarch, and thus http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 17)03-May-05 17:10:10

Ovid: ICU Book has a higher colloid osmotic pressure (see Ta le 15.3). It is more effective as a volume expander than hetastarch, and can increase plasma volume y 1.5 times t he infusion volume (1). The oncotic effects dissipate after 12 hours (1). Pentas tarch shows less of a tendency to interact with coagulation proteins than hetast arch, ut the significance of this tendency is unclear. THE DEXTRANS The dextrans are glucose polymers produced y a acterium (Leuconostoc) incu ate d in a sucrose medium. First introduced in the 1940s, these colloids are not pop ular (at least in the United States) ecause of the perceived risk of adverse re actions. The two most common dextran preparations are 10% dextran-40 and 6% dext ran-70, oth diluted in isotonic saline. Features Both dextran preparations are hyperoncotic to plasma (COP = 40 mm Hg). Dextran-4 0 causes a larger increase in plasma volume than dextran-70, ut the effects las t only a few hours. Dextran-70 is the preferred preparation ecause of its prolo nged action (1). Disadvantages Dextrans produce a dose-related leeding tendency y inhi iting platelet aggrega tion, reducing activation of Factor VIII, and promoting fi rinolysis (15). The h emostatic defects are minimized y limiting the daily dextran dose to 20 mL/kg ( 15). Anaphylactic reactions were originally reported in as many as 5% of patient s receiving dextran infusions. However, this has improved considera ly in the la st 20 years ecause of improvements in antigen detection and desensitization and improvements in preparation purity. The current incidence of anaphylaxis is 0.0 32% (15). Dextrans coat the surface of red lood cells and can interfere with th e a ility to cross-match lood. Red cell preparations must e washed P.237 to el iminate this pro lem. Dextrans also increase the erythrocyte sedimentation rate as a result of their interactions with red lood cells (15). Finally, dextrans h ave een implicated as a cause of acute renal failure (15,18). The proposed mech anism is a hyperoncotic state with reduced filtration pressure. However, this me chanism is unproven, and renal failure occurs only rarely in association with de xtran infusions. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 17)03-May-05 17:10:10

Ovid: ICU Book COLLOIDCRYSTALLOID CONUNDRUM There is considera le disagreement a out the most appropriate fluid for volume r esuscitation in critically ill patients. The following is a rief description of the issues involved in the colloid crystalloid de ate. CRYSTALLOID ORIGINS Because crystalloid fluids fill primarily the interstitial space, these fluids a re not useful for filling the vascular space. The early popularity of crystalloi d fluid resuscitation in hypovolemia stems from two o servations (descri ed in C hapter 14) made a out 40 years ago. The first is the response to mild hemorrhage , which involves a shift of fluid from the interstitial space to the vascular sp ace (19). The second o servation stems from studies in an animal model of hemorr hagic shock, where survival was much improved if a crystalloid fluid was given a long with reinfusion of the shed lood volume (20). The com ination of these two o servations has een interpreted as indicating that the major consequence of h emorrhage is an interstitial fluid deficit, and that replacement of interstitial fluid with crystalloid fluids is important for survival. COLLOID PERFORMANCE The interstitial fluid deficit is predominant only when lood loss is mild (less than 15% of the lood volume), and in this situation, no volume resuscitation i s necessary ( ecause the ody is capa le of fully compensating for the loss of lood volume). When lood loss is more severe, the priority is to keep the vascul ar space filled and there y support the cardiac output. Because colloid fluids a re a out three times more potent than crystalloid fluids for increasing vascular volume and supporting the cardiac output (as shown in Figure 14.5 and Figure 15 .1), colloid fluids are more effective than crystalloid fluids for volume resusc itation in moderate to severe lood loss. Crystalloid resuscitation can achieve the same endpoint as colloid resuscitation, ut larger volumes of crystalloid fl uid (a out three times the volume of colloid fluids) must e used. This latter a pproach is less efficient, yet it is the one favored y crystalloid users. P.238 SURVIVAL Despite the superiority of colloid fluids for expanding plasma volume, colloid f luid resuscitation does not confer a higher survival rate in patients with hypov olemic shock (1,4,21,22). This lack of improved outcomes is a major rallying poi nt for crystalloid users, ut it does not negate the fact that colloid fluids ar e more effective for maintaining lood volume in http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 17)03-May-05 17:10:10

Ovid: ICU Book patients who are actively leeding. EXPENSE The iggest disadvantage of colloid resuscitation is the higher cost of colloid fluids. Ta le 15.4 shows a cost comparison for colloid and crystalloid fluids. U sing equivalent volumes of 250 mL for colloid fluids and 1000 mL for crystalloid fluids, the cost of colloid resuscitation is three times as high (if hetastarch is used) to six times as high (if al umin is used) than volume resuscitation wi th isotonic saline. TABLE 15.4. RELATIVE COST OF INTRAVENOUS FLUIDS EDEMA The risk of edema has een used to discredit each type of fluid. Because crystal loid fluids distri ute primarily in the interstitial space, edema is an expected feature of crystalloid fluid resuscitation. However, edema is also a risk with colloid fluid resuscitation. This is particularly true with al umin-containing f luids; even though al umin is the principal oncotic force in plasma, over half o f the al umin in the human ody is in the interstitial fluid (12,13). Therefore, a large proportion of infused al umin eventually finds its way into the interst itial P.239 fluid and promotes edema. Furthermore, this egress of al umin from t he loodstream is magnified when capillary permea ility is disrupted, which is a common occurrence in critically ill patients. Despite this risk, trou lesome ed ema (e.g., pulmonary edema) is not common with http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 17)03-May-05 17:10:10

Ovid: ICU Book either type of fluid resuscitation when capillary hydrostatic pressure is not ex cessive (23). HOLE-IN-THE-BUCKET ANALOGY The following analogy helped me resolve the colloidcrystalloid conundrum. Assume that the goal is to recreate the performance of crystalloid and colloid fluids i n expanding the plasma volume y filling a ucket. Because the volume of crystal loid fluids needed to expand the plasma volume (fill the ucket) is three times larger than the volume of colloid fluid that fills the ucket, holes will need t o e punched in the ucket while it is filled with crystalloid fluids (to allow the extra fluid to escape). Therefore, the question is this: If the goal is to f ill a ucket with fluid, do you want to punch holes in the ucket (and make the ucket more difficult to fill)? Seen in this light, it is more efficient to use colloid fluid resuscitation to expand the plasma volume.

HYPERTONIC RESUSCITATION An interesting approach to volume resuscitation that has stalled in recent years is the use of small-volume hypertonic saline solutions. A 7.5% sodium chloride solution similar to the one descri ed in Ta le 15.1 is given either in a fixed v olume of 250 mL or in a volume of 4 mL/kg. Figure 15.1 shows the expected change in plasma and interstitial fluid produced y administration of 250 mL 7.5% sodi um chloride. The volume increments in oth fluid compartments are similar to tho se produced y 1 L of 5% al umin. Thus, hypertonic saline resuscitation can prod uce equivalent volume expansion to colloid fluids, ut at one-fourth the infused volume. Note that the total volume expansion (1235 mL) produced y 7.5% saline is far greater than the infused volume (250 mL). The additional volume comes fro m intracellular fluid that moves out of cells and into the extracellular space. This movement of intracellular fluid points to one of the feared complications o f hypertonic resuscitation: cell dehydration. WHAT ROLE? Since the first report of its successful use in 1980, hypertonic saline has een shown repeatedly ( ut not unanimously) to e safe and effective in the early re suscitation of hypovolemia. However, there is little P.240 evidence that hyperto nic resuscitation is superior to standard volume resuscitation. Hypertonic resus citation seems est suited for prehospital resuscitation in cases of trauma, ut studies in trauma resuscitation fail to document a clear enefit with this appr oach in most patients (24,25). Select su groups of patients (e.g., those with pe netrating truncal injuries who required surgery) may enefit from hypertonic res uscitation, ut these su groups are small. Thus, after http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 17)03-May-05 17:10:10

Ovid: ICU Book over 15 years of evaluating this technique, hypertonic resuscitation has few adv ocates. REFERENCES GENERAL WORKS Kaufman BS, ed. Fluid resuscitation of the critically ill. Critical care clinics . Vol. 8, No. 2. Philadelphia: WB Saunders, 1992. Weinstein SM. Plumer's principles and practice of intravenous therapy. 5th ed. P hiladelphia: Lippincott-Raven, 1993. REVIEWS 1. Griffel MI, Kaufman BS. Pharmacology of colloids and crystalloids. Crit Care Clin 1992;8:235254 (118 References). 2. Kaminski MV, Haase TJ. Al umin and colloid osmotic pressure: implications for fluid resuscitation. Crit Care Clin 1992;8:311322 (13 References). 3. Sutin KM, Ruskin KJ, Kaufman BS. Intravenous fluid therapy in neurologic inju ry. Crit Care Clin 1992;8:367408 (181 References). 4. Imm A, Carlson RW. Fluid resuscitation in circulatory shock. Crit Care Clin 1 993;9:313 333 (73 References). CRYSTALLOID & COLLOID FLUIDS 5. Lowery BD, Cloutier CT, Carey LC. Electrolyte solutions in resuscitation in h uman hemorrhagic shock. Surg Gynecol O stet 1971;131:273279. 6. Griffith CA. The family of Ringer's solutions. J Natl Intravenous Ther Assoc http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 17)03-May-05 17:10:10

Ovid: ICU Book 1986;9:480483. 7. American Association of Blood Banks Technical Manual. 10th ed. Arlington, VA: American Association of Blood Banks, 1990:368. 8. Talpers SS, Rom erger DJ, Bunce SB, Pingleton SK. Nutritionally associated in creased car on dioxide production. Chest 1992;102:551555. 9. Gunther B, Jauch W, Hartl W, et al. Low-dose glucose infusion in patients who have undergone surgery. Arch Surg 1987;122:765771. 10. DeGoute CS, Ray MJ, Manchon M, et al. Intraoperative glucose infusion and l ood lactate: endocrine and meta olic relationships during a dominal aortic surge ry. Anesthesiology 1989;71;355361. 11. Sie er FE, Traystman RJ. Special issues: glucose and the rain. Crit Care Me d 1992;20:104114. COLLOID FLUIDS 12. Doweiko JP, Nompleggi DJ. Role of al umin in human physiology and pathophysi ology. J Parent Enter Nutr 1991;15:207211. P.241 13. Guthrie RD Jr, Hines C Jr. Use of al umin in the critically ill patient. Am J Gastroenterol 1991;86:255263. 14. Marik PE. The treatment of hypoal uminemia in the critically ill patient. He art Lung 1993;22:166170. 15. Nearman HS, Herman ML. Toxic effects of colloids in the intensive care unit. Crit Care Clin 1991;7:713723. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 17)03-May-05 17:10:10

Ovid: ICU Book 16. Kapiotis S, Quehen erger P, Eichler H-G, et al. Effect of hydroxyethyl starc h on the activity of lood coagulation and fi rinolysis in healthy volunteers: c omparison with al umin. Crit Care Med 1994;22:606612. 17. Waxman K, Holness R, Tominaga G et al. Hemodynamic and oxygen transport effe cts of pentastarch in urn resuscitation. Ann Surg 1989;209:341345. 18. Strauss RG, Stansfield C, Henriksen RA, et al. Pentastarch may cause fewer e ffects on coagulation than hetastarch. Transfusion 1988;28:257261. 19. Drumi W, Polzleitner D, Laggner AN, et al. Dextran-40, acute renal failure, and elevated plasma oncotic pressure. N Engl J Med 1988;318:252254.

21. Shires T, Carrico J, Lightfoot S. Fluid therapy in hemorrhagic shock. Arch S urg 1964;88:688693. 22. Bisonni RS, Holtgrave DR, Lawler F, et al. Colloids versus crystalloids in f luid resuscitation: an analysis of randomized controlled trials. J Fam Pract 199 1;32:387390. 23. Schaeffer RC, Reeiewicz RA, Chilton SW, et al. Effects of colloid or crystal loid solutions on edemagenesis in normal and throm omicroem olized lungs. Crit C are Med 1987;15:11101115. HYPERTONIC RESUSCITATION 24. Mattox KL. Prehospital hypertonic saline-dextran infusion for post-traumatic hypotension: the USA Multicenter Trial. Ann Surg 1991;213:482486. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 17)03-May-05 17:10:10

COLLOID-CRYSTALLOID WARS 20. Moore FD. The effects of hemorrhage on 73:567577.

ody composition. N Engl J Med 1965;2

Ovid: ICU Book 25. Vassar MJ, Fischer RP, O'Brien PE, et al. A multicenter trial for resuscitat ion of injured patients with 7.5% sodium chloride. Arch Surg 1993;128:10031013. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 17)03-May-05 17:10:10

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 16 ACUTE HEART FAILURES There's no dou t that the proper functioning of our pipes and pumps does have an immediate urgency well eyond that of almost any of our other its and pieces. --Steven Vogel Cardiac pump failure is an ominous sign in the critically ill pat ient that requires prompt recognition and management. Acute heart failure is not a single entity, ut can involve the right or left side of the heart, or can oc cur during diastole or systole. This chapter uses the principles of cardiac perf ormance provided in Chapter 1 to descri e a edside approach to the varieties of heart failure (1,2,3,4,5 and 6). The approach descri ed here requires invasive monitoring with pulmonary artery catheters and centers on the mechanical pro lem rather than the specific illness. The common causes of acute heart failure in p atients in the ICU are indicated in Figure 16.1. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 24)03-May-05 17:12:28

Ovid: ICU Book Figure 16.1. Common causes of acute heart failure in patients in the ICU. HEMODYNAMIC MANIFESTATIONS The diagnosis of heart failure egins with recognition of the early signs of hea rt failure, then identifies the phase of the cardiac cycle and the side of the h eart involved. EARLY RECOGNITION The serial changes in hemodynamic parameters shown in Figure 16.2 are taken from a patient who developed progressive left heart failure following cardiopulmonar y ypass surgery. The sequence of hemodynamic alterations is as follows (the num ers elow correspond to the circled num ers in Figure 16.2):

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 24)03-May-05 17:12:28

Ovid: ICU Book Figure 16.2. Hemodynamic changes during progressive left heart failure in a post operative patient. q The earliest sign of ventricular dysfunction is an increase in pulmonary capilla ry wedge pressure. The stroke volume is maintained P.243 at this stage ecause t he ventricle is still preload-responsive (i.e., the Starling curve is still stee p). q The next stage is marked y a decrease in stroke volume and an increase in heart rate. The tachycardia offsets the reduction in stroke volume, so that the cardi ac output remains unchanged. q The final stage is characterized y a decrease in cardiac output. The point at w hich the cardiac output egins to decline marks the transition from compensated to http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 24)03-May-05 17:12:28

Ovid: ICU Book decompensated heart failure. The decompensated phase of heart failure is charact erized y peripheral vasoconstriction, which initially maintains peripheral loo d flow ut eventually causes a further reduction in cardiac output and periphera l flow. The serial hemodynamic changes shown in Figure 16.2 demonstrate the following im portant point: Cardiac output may not e reduced in the early stages of heart fa ilure. The early recognition of heart failure requires monitoring of the cardiac filling pressures and the ventricular stroke volume. SYSTOLIC VERSUS DIASTOLIC FAILURE Heart failure is not synonymous with contractile failure, and 40% of patients wi th newly diagnosed heart failure have normal systolic function (7,8 and 9). The pro lem in these patients is a decrease in ventricular distensi ility, a disorde r known as diastolic heart failure. In this type of heart failure, inadequate ve ntricular filling compromises cardiac P.244 output, while the force of ventricul ar contraction is normal. Common causes of diastolic heart failure in patients i n the ICU include ventricular hypertrophy, myocardial ischemia, pericardial effu sions, and positive-pressure mechanical ventilation. The distinction etween sys tolic and diastolic heart failure is important, ecause what would e appropriat e management for one type of heart failure can aggravate the other. Routine Hemodynamic Monitoring Routine hemodynamic measurements are incapa le of distinguishing diastolic from systolic heart failure (7,8 and 9). This is illustrated in Figure 16.3. The curv es in this figure are similar to the pressurevolume curves shown in Figure 1.1. T he upper curves in the figure are ventricular function curves relating ventricul ar end-diastolic pressure (EDP) and cardiac stroke volume. These curves demonstr ate that heart failure of either type is associated with an increase in EDP and a decrease in stroke volume. The lower set of curves are diastolic pressurevolume curves, and these curves demonstrate that the increase in EDP in heart failure is associated with opposite changes in the end-diastolic volume P.245 (EDV) in t he two types of heart failure; that is, the EDV is increased in systolic heart f ailure and decreased in diastolic heart failure. Therefore, monitoring cardiac f illing pressures as an index of ventricular preload does not allow a distinction etween systolic and diastolic heart failure. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 24)03-May-05 17:12:28

Ovid: ICU Book Figure 16.3. Changes in end-diastolic pressure (EDP) and end-diastolic volume (E DV) in systolic and diastolic heart failure. See text for explanation. END-DIASTOLIC VOLUME The end-diastolic volume is thus the est measure for identifying systolic and d iastolic heart failure. The EDV can e derived y the following relationship et ween the stroke volume (SV) and ejection fraction (EF): The ejection fraction of the left ventricle can e measured with radionuclide ve ntriculography (10), and the ejection fraction of the right ventricle can e mea sured with a specialized pulmonary artery catheter with a fast-response thermist or, as descri ed in Chapter 12 (see Fig. 12.3). Because edside radionuclide ven triculography is tedious, P.246 and expensive left-ventricular EDV is not a comm on edside measurement. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 24)03-May-05 17:12:28

Ovid: ICU Book RIGHT VERSUS LEFT HEART FAILURE Right heart failure (which is predominantly systolic failure) is more prevalent than considered in patients in the ICU (11), and it may e particularly prominen t in ventilator-dependent patients. The following measurements can prove useful in identifying right heart failure. Cardiac Filling Pressures The relationship etween the central venous pressure (CVP) and the pulmonary cap illary wedge pressure (PCWP) can sometimes e useful for identifying right heart failure. The following criteria have een proposed for right heart failure (12) : CVP > 15 mm Hg and CVP = PCWP or CVP > PCWP. Unfortunately, at least one-third of patients with acute right heart failure do not satisfy these criteria (12). One pro lem is the insensitivity of the CVP; an increase in the CVP is seen only in the later stages of right heart (systolic) failure. Contractile failure of t he right ventricle results in an increase in end-diastolic volume, and only when the increase in volume of the right heart is impeded y the pericardium does th e end-diastolic pressure (CVP) rise (11). Another pro lem with the CVPPCWP relati onship for identifying right heart failure is the interaction etween the right and left sides of the heart. This is shown in Figure 16.4. Both ventricles share the same septum, so enlargement of the right ventricle pushes the septum to the left and compromises the left-ventricular cham er. This interaction etween rig ht and left ventricles is called interventricular interdependence, and it can co nfuse the interpretation of ventricular filling pressures. In fact, as indicated y the diastolic pressures in Figure 16.4, the hemodynamic changes in right hea rt failure can look much like the hemodynamic changes in pericardial tamponade ( 11). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 24)03-May-05 17:12:28

Ovid: ICU Book Figure 16.4. Interventricular interdependence. The mechanism where y right heart failure compromises diastolic filling of the left ventricle and raises the left -ventricular end-diastolic (wedge) pressure. RV = right ventricle, LV = left ven tricle. End-Diastolic Volume The right-ventricular ejection fraction (RVEF) and end-diastolic volume (RVEDV), as determined with pulmonary artery catheters equipped with rapid-response ther mistors (see Chapter 12), are the est measures for identifying right heart fail ure. A decrease in RVEF (normal RVEF is 45 to 50%) and an increase in RVEDV (nor mal RVEDV is 80 to 140 mL/m2) is expected in right heart failure (13). The respo nse to volume infusion may e even more diagnostic. In one study, volume infusio n resulted in a 30% increase in RVEDV in patients with right-ventricular dysfunc tion, while in other patients, there was no increase P.247 in RVEDV after a flui d challenge (14). In another study, volume infusion did not result in an increas e in cardiac output if the RVEDV was in excess of 140 mL/m2 (15). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 24)03-May-05 17:12:28

Ovid: ICU Book Echocardiography Cardiac ultrasound can e useful at the edside for differentiating right from l eft heart failure. Three findings typical of right heart failure are (a) an incr ease in right-ventricular cham er size, ( ) segmental wall motion a normalities on the right, and (c) paradoxical motion of the interventricular septum (12). MANAGEMENT STRATEGIES The primary goal in managing heart failure is to maintain cardiac output, and th e secondary goal is to decrease venous (capillary) pressure P.248 to limit edema formation. The strategies presented here are designed to achieve oth of these goals. LEFT HEART (SYSTOLIC) FAILURE The approach to left-ventricular (systolic) failure presented here centers on th e PCWP). The approach is descri ed graphically in Figure 16.5. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 24)03-May-05 17:12:28

Ovid: ICU Book Figure 16.5. Ventricular function curves for the normal and failing left ventric le. Arrows show the expected hemodynamic changes with various interventions in l eft heart failure. Su optimal Wedge Pressure Correction of inadequate filling pressures is the sine qua non of heart failure management. As stated y the cardiovascular physiologist, Carl Wiggers: It is axi omatic that the heart can pump only as much as it receives. Condition: Low PCWP I ntervention: Volume infusion to optimal PCWP The optimal wedge pressure is the h ighest pressure that augments cardiac output without producing pulmonary edema. This is shown in Figure 16.5 as the highest point on the lower (heart failure) c urve that does not enter the hatched pulmonary edema region. The optimal PCWP is determined y the colloid osmotic pressure (COP) of lood (see Chapter 15 for a description of the COP). When the COP is normal (20 to 25 mm Hg), the optimal P CWP is 20 mm Hg (16). P.249 Optimal Wedge Pressure When the wedge pressure is optimal, therapy is dictated y the lood pressure (B P). The hemodynamic drugs recommended here are those given y continuous IV infu sion (1,3,4 and 5). These drugs are listed in Ta le 16.1, along with their appro priate dose ranges. Each of the drugs in this ta le is descri ed in more detail in Chapter 18. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 24)03-May-05 17:12:28

Ovid: ICU Book TABLE 16.1. PHARMACOTHERAPY OF ACUTE HEART FAILURE1 Condition: Optimal PCWP, low BP Intervention: Dopamine Dopamine stimulates oth receptors (cardio-stimulation and vasodilation) and -receptors (v soconstriction) . The effect increases the cardiac output, and the effect r ises the blood press ure. The effect becomes evident t doses bove 5 /kg/minute, nd v soconstriction is the predomin nt effect t doses bove 10 /kg/minute (1,5). Condition: Optim l PCWP, norm l BP Intervention: Dobut mine, mrinone Dobut mine, synthetic dre nergic gent th t does not c use peripher l v soconstriction, is widely reg rded s the inotropic gent of choice for the cute m n gement of (systolic) he rt f ilure (17). Amrinone is phosphodiester se inhibitor th t h s both positive in otropic nd v sodil tor ctions. This gent c n serve s n effective ltern tiv e to dobut mine, or it c n be dded to dobut mine to enh nce the over ll effect (5). Condition: Optim l PCWP, high BP Intervention: Nitroprusside, nitroglycerin Nitroprusside is popul r v sodil tor in the critic l c re setting. However, c y nide ccumul tion is common during nitroprusside infusions (see Ch pter 18), nd the risk of cy nide toxicity should temper the use of nitroprusside. The risk of cy nide toxicity h s led the Food nd Drug Administr tion to recommend m x imum nitroprusside dose r te of 10 / minute for no more th n 10 minutes. (For mor e inform tion bout nitroprusside use, see reference 18.) Nitroglycerin P.250 is vi ble ltern tive to nitroprusside if dministered in dose r tes th t exceed 50 /minute. Other v sodil tors th t c n be given by continuous intr venous infus ion, such s l bet lol ( combined locker), esmolol (a short-acting locker), and trimethaphan (a ganglionic locker) can decrease the cardiac output, and thus t hese agents are more appropriate for treatment of severe hypertension accompanie d y an adequate cardiac output. High Wedge Pressure http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 24)03-May-05 17:12:28

Ovid: ICU Book If the wedge pressure is high and the patient is at risk for hydrostatic pulmona ry edema, the appropriate management is determined y the cardiac output (CO). C ondition: High PCWP, low CO Intervention: Do utamine, amrinone Therapy with do u tamine and amrinone results in significant reductions in wedge pressure (5,19). Dopamine should e avoided when the wedge pressure is elevated ecause dopamine constricts pulmonary veins and can increase the wedge pressure further (19,20). Vasodilators can e detrimental in pulmonary edema ecause they increase shunt f raction and can aggravate hypoxemia (21). Condition: High PCWP, normal CO Interv ention: Nitroglycerin, ? furosemide A normal cardiac output in the face of a hig h PCWP suggests diastolic heart failure. Aggressive diuresis is not recommended as the first line of therapy in this setting ecause the high filling pressures help maintain cardiac output. Intravenous nitroglycerin (less than 100 /kg/minute ) should e useful here ecause this agent reduces the wedge pressure while also reducing the arterial resistance to maintain cardiac output (21). Su lingual ni troglycerin can e given for immediate results. In the setting of pulmonary edem a, nitroglycerin can increase shunt fraction and decrease the arterial POAB2. Th erefore, arterial gases should e monitored carefully when using nitroglycerin i n pulmonary edema. Furosemide Intravenous furosemide is a popular therapy for acute pulmonary edema, and it is usually given with little regard for the effects on cardiac output. However, it is well esta lished that intravenous furosemide often causes a decrease in card iac output in patients with acute heart failure (22,23,24,25,26,27 and 28). To e mphasize this point, Ta le 16.2 shows the hemodynamic effects of intravenous fur osemide in acute heart failure from all clinical studies reported from 1970 to 1 990 (22,23,24,25,26,27,28,29,30 and 31). A total of 169 su jects are included, a nd in 7 of 10 studies (involving 113 su jects, or 67% of the total study populat ion) intravenous furosemide caused a significant reduction in cardiac output and /or stroke volume. This effect is the result of a decrease in venous return and an increase P.251 in systemic vascular resistance. The latter effect is due to t he a ility of furosemide to stimulate renin release and raise circulating levels of angiotensin, a vasoconstrictor (32). Considering the popularity of lowering angiotensin levels with angiotensin-converting enzyme inhi itors as a therapy in heart failure, the actions of furosemide to promote angiotensin formation seem to e http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 24)03-May-05 17:12:28

Ovid: ICU Book counterproductive. This effect of furosemide should e considered carefully efo re the kneejerk response of administering furosemide in acute heart failure is d eveloped. TABLE 16.2. HEMODYNAMIC EFFECTS OF INTRAVENOUS FUROSEMIDE IN ACUTE LEFT HEART FA ILURE1 Because the diuretic effect of furosemide is more closely related to its urinary excretion rate than to its plasma concentration (33), continuous infusion furos emide has een advocated for more effective diuresis in patients with heart fail ure (33,34). Continuous infusion is usually recommended when more than 80 mg of intravenous furosemide is required to produce the desired diuretic effect. Dose rates range from 2.5 to 160 mg/hr. LEFT HEART (DIASTOLIC) FAILURE The optimal treatment for diastolic heart failure is unknown. Diuretic therapy s hould e avoided, and inotropic therapy should e ineffective. Although vasodila tor therapy should carry a high risk of hypotension in diastolic heart failure ( ecause a ventricle with normal systolic function should e unresponsive to chan ges in afterload), some vasodilator agents (e.g., calcium channel lockers and a ngiotensin converting enzyme or ACE inhi itors) may also have lusitropic actions that enhance myocardial relaxation (7,8). Verapamil has P.252 proven effective in idiopathic hypertrophic cardiomyopathies (7,8,35); however, there is evidence that calcium channel lockers do not improve myocardial relaxation in other http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 24)03-May-05 17:12:28

Ovid: ICU Book conditions associated with diastolic heart failure (36). For now, the management of diastolic heart failure is similar to that for systolic heart failure, ut c areful hemodynamic monitoring is necessary to identify adverse effects in the ma nagement of diastolic heart failure. RIGHT HEART FAILURE Therapeutic strategies for right heart failure are similar in just descri ed. The strategies elow pertain only to primary e (e.g., acute myocardial infarction), and not to right heart to chronic o structive lung disease or to left heart failure. are used as the focal points of management. q If PCWP is elow 15 mm Hg, infuse volume until the PCWP or CVP increases y 5 mm Hg or either one reaches 20 mm Hg (12). q If the RVEDV is less than 140 mL/m2, infuse volume until the RVEDV reaches 140 m L/ m2 (15). q If PCWP is a ove 15 mm Hg or the RVEDV is 140 mL/m2 or higher, infuse do utamine , eginning at a rate of 5 /kg/minute (37,38). q principle to those right heart failur failure secondary The PCWP and RVEDV

The response to volume infusion must e carefully monitored in right heart failu re ecause aggressive volume infusion can overdistend the right ventricle and fu rther reduce cardiac output through interventricular interdependence (see Fig. 1 6.4). Do utamine is an effective agent in right heart failure (37,38). Nitroprus side has een used in right heart failure, ut it is not as effective as do utam ine (38). MECHANICAL SUPPORT A variety of devices are availa le that provide temporary mechanical support of the failing heart. Most of these devices are used after cardiac surgery, where a out 5% of patients require postoperative mechanical assistance to support the c ardiac output (39). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 24)03-May-05 17:12:28

In the presence of AV dissociation or complete heart A-V pacing and avoid ventricular pacing (12).

lock, institute sequential

Ovid: ICU Book INTRAAORTIC BALLOON PUMP (IABP) Intraaortic alloon counterpulsation has een the standard method of providing m echanical circulatory support for over 25 years (39,40). The IABP consists of a 30-cm polyurethane alloon attached to one end of a large- ore catheter. The dev ice is inserted in the femoral artery P.253 at the groin, either percutaneously or via arteriotomy, with the alloon wrapped tightly around the catheter. Once i nserted, the catheter is advanced up the aorta until the tip lies just eyond th e origin of the left su clavian artery. When in place, the alloon wrapping is r eleased to allow periodic alloon inflations. Correct placement does not require fluoroscopy, and the IABP can e placed successfully at the edside. Hemodynamic Effects The intraaortic alloon is rapidly inflated with helium (35 to 40 mL capacity) a t the onset of each diastolic period, when the aortic valve closes. The alloon is then rapidly deflated at the onset of ventricular systole, just efore the ao rtic valve opens. This pattern of alloon inflation and deflation produces two c hanges in the arterial pressure waveform, as illustrated in Figure 16.6 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 24)03-May-05 17:12:28

Ovid: ICU Book Figure 16.6. The influence of intraaortic alloon counterpulsation on arterial p ressure and flow. q Inflation of the alloon increases the peak diastolic pressure and displaces lo od toward the periphery. The increase in diastolic pressure increases the mean a rterial pressure and there y increases mean lood flow in the periphery. Coronar y lood flow should also increase, ecause the ulk of coronary P.254 lood flow occurs during diastole. However, the IABP increases coronary flow only in hypot ensive patients and does not promote coronary flow in normotensive patients (41) . q Deflation of the alloon reduces the end-diastolic pressure, which reduces the i mpedance to flow when the aortic valve opens at the onset of systole. This decre ases ventricular afterload and promotes ventricular stroke output. Indications The usual indications for the IABP include (a) cardiopulmonary ypass ( efore an d after), ( ) cardiac transplantation ( efore and after), (c) acute myocardial i nfarction with cardiogenic shock, (d) acute mitral insufficiency, and (e) unsta le angina Almost half of all alloon insertions occur in the immediate postopera tive period following cardiopulmonary ypass surgery. The IABP has also gained i ncreasing use as a ridge to cardiac transplantation. Contraindications The contraindications to the IABP include aortic regurgitation, aortic dissectio n, and a recently placed (within 12 months) prosthetic graft in the thoracic aor ta (40). Complications The incidence of complications from the IABP ranges from 15 to 45%, with serious complications reported in 5 to 10% of cases (40). The most common complications are leg ischemia (9 to 22%) and septicemia (1 to 22%). Leg ischemia can occur i n the ipsilateral or contralateral leg and can appear either with the device in place or soon after it is removed. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 24)03-May-05 17:12:28

Ovid: ICU Book When distal pulses disappear while the alloon is in place, removal of the devic e is often sufficient to restore flow without any further therapy. A out 20% of patients require surgical interventions for lower leg vascular complications (42 ). Weaning Balloon assistance is usually withdrawn gradually, y either decreasing the freq uency of alloon inflations per cardiac cycle (1:2, 1:3, etc.), or y reducing t he inflation volume gradually to 10% of the original volume (40). The choice of weaning method is a matter of individual preference, and there is no evidence th at one method is superior to the other. The weaning period is also a matter of i ndividual preference and can range from 60 minutes to 24 hours (40). VENTRICULAR ASSIST DEVICES A ventricular assist device (VAD) is a nonpulsatile pump that is placed in paral lel with either the right ventricle (RVAD), the left ventricle P.255 (LVAD), or oth ventricles (BiVAD) (39,43,44). The pump is adjusted to provide a total syst emic flow of 2.0 to 3.0 L/min/m2. These devices are placed intraoperatively in c ases where the IABP fails to provide adequate circulatory support (usually after cardiopulmonary ypass surgery). After 24 hours of operation, attempts to wean pump support are usually initiated y decreasing the pump flow rate until right atrial pressure (RVAD) or left atrial pressure (LVAD) increases to 20 to 25 mm H g (44). The duration of ventricular support is usually 1 to 4 days, ut it can r ange from a few hours to longer than 10 days (44). Complications occur in over 5 0% of patients and most often include leeding or systemic em olism (43,44). Mos t patients can never e weaned from pump support, ut as many as one-third of pa tients survive the ordeal (44).

CARDIOPULMONARY BYPASS SURGERY The immediate period following cardiopulmonary ypass surgery is often marked y hemodynamic insta ility (45). The following are some of the major hemodynamic c oncerns in the early postoperative period. CARDIAC TAMPONADE Cardiac tamponade occurs in 3 to 6% of patients undergoing open-heart surgery (4 6). It often appears in the first few hours after surgery, ut can occur at a la ter time when the pacemaker wires are removed. The pericardium is open after car diac surgery, and this prevents fluid from accumulating evenly around the heart. The most common cause of tamponade after surgery is http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 24)03-May-05 17:12:28

Ovid: ICU Book a lood clot compressing the right heart. Clinical Presentation Cardiac tamponade often has an atypical presentation in the post ypass period (4 6). Two typical manifestations of cardiac tamponade may e a sent. q Pulsus paradoxus (inspiratory drop in systolic lood pressure of at least 10 mm Hg) can e masked in patients receiving mechanical ventilation. Positive-pressur e lung inflation can assist the left ventricle during systole and there y augmen t the systolic lood pressure. The increase in systolic pressure produced y pos itive-pressure mechanical ventilation is called reverse pulsus paradoxus (see Ch apter 26). q Equalization of diastolic pressures (CVP, pulmonary artery diastolic pressure, P CWP) may not e a feature of cardiac tamponade when a clot is compressing the ri ght atrium. In this situation, the superior vena cava pressure (CVP) can increas e while the pulmonary artery and wedge pressures decrease. P.256 Diagnosis Tamponade is often suspected on clinical grounds when there is a sudden decrease in lood drainage from mediastinal chest tu es followed y a rise in cardiac fi lling pressures and a progressive decline in cardiac output. The diagnosis is of ten uncertain (and is a source of much angst in cardiovascular surgeons) and req uires repeat thoracotomy for verification and for ligation of leeding sites. If availa le, transesophageal echocardiography can e a valua le tool for identify ing compression of the right atrium and left-ventricular akinesis (47). POST-BYPASS HEMODYNAMICS The rewarming period after cardiopulmonary ypass is associated with a decrease in the compliance of the ventricles (48). The etiology is unclear, ut myocardia l edema from cooling and reperfusion may play a role. The peripheral vascular re sistance can either increase or decrease in the immediate postoperative period. Systolic function is varia le ut seems to e well maintained in most patients ( 45). Acute infarction is reported in less than 10% of patients (45). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 24)03-May-05 17:12:28

Ovid: ICU Book Management The decrease in ventricular compliance during the rewarming period causes a decr ease in EDV at any given EDP. This means that a normal wedge pressure in the ear ly postoperative period represents a low EDV. Therefore, when cardiac output is low and the PCWP is not elevated, volume infusion is indicated until the PCWP is in the range of 20 mm Hg. Drug therapy can e selected according to the systemi c vascular resistance (SVR). The following scheme may e helpful: SVR High High High Normal Normal Low Low Blood Pressure High Normal Low Normal Low Normal Low Intervention Nitroprusside Do utamine Dopamine, IABP Do utamine Dopamine, IABP D o utamine Dopamine, epinephrine The use of nitroprusside to treat postoperative hypertension carries a particula rly high risk of cyanide accumulation after cardiopulmonary ypass surgery ecau se of the depletion of thiosulfate associated with this procedure (see Chapter 1 8). Nitroglycerin is an effective alternative to nitroprusside (49), and I have een using trimethaphan (a ganglionic locker) to treat severe hypertension in t his situation. For more information on trimethaphan and the other hemodynamic dr ugs in this chapter, see Chapter 18. P.257 REFERENCES GENERAL TEXTS Barnett DB, Pouleur H, Francis GS. Congestive heart failure: pathophysiology and http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 24)03-May-05 17:12:28

Ovid: ICU Book treatment. New York: Marcel Dekker, 1993. Gaash WH, LeWinter MM. Left ventricular diastolic dysfunction and heart failure. Baltimore: Williams & Wilkins, 1994. Khan MG, ed. Cardiac drug therapy. 4th ed. Philadelphia: WB Saunders, 1995.

Moreno-Ca ral C, Mitchell RS, Miller DC. Manual of postoperative management in a dult cardiac surgery. Baltimore: Williams & Wilkins, 1988. REVIEWS 1. Lollgen H, Drexler H. Use of inotropes in the critical care setting. Crit Car e Med 1990;18:S56S60 (32 References). 2. Smith TW, Braunwald E, Kelly RA. The management of heart failure. In: Braunwa ld E, ed. Heart disease. A text ook of cardiovascular medicine. 4th ed. Philadel phia: WB Saunders, 1992;464519 (600 References). 3. Alpert J, Becker JA. Mechanisms and management of cardiogenic shock. Crit Car e Clin 1993;9:205218 (32 References). 4. Snell RJ, Calvin JE. Cardiogenic shock: pathophysiology, management and treat ment. In: Edwards JD, Shoemaker WC, Vincent J-L, eds. Oxygen transport: principl es and practice. Philadelphia: WB Saunders, 1993;246-273 (99 References). 5. Zaloga GP, Prielipp RC, Butterworth JF, Royster RL. Pharmacologic cardiovascu lar support. Crit Care Clin 1993;9:335362 (192 References). 6. Wilson RF. Trauma in patients with pre-existing cardiac disease. Crit Care Cl in http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 24)03-May-05 17:12:28

Maccioli GA. Theory and practice of intra-aortic : Williams & Wilkins, 1996.

alloon pump therapy. Baltimore

Ovid: ICU Book 1994;10:461506 (163 References). DIASTOLIC HEART FAILURE 7. Bonow RO, Udelson JE. Left ventricular diastolic dysfunction as a cause of co ngestive heart failure. Ann Intern Med 1992;117:502510. 8. Goldsmith S, Dick C. Differentiating systolic from diastolic heart failure: p athophysiologic and therapeutic considerations. Am J Med 1993;95:645655. 9. Gaasch WH. Diagnosis and treatment of heart failure ased on left ventricular systolic or diastolic dysfunction. JAMA 1994;271:12761280. 10. Clements IP, Sinak LJ, Gi ons RJ, et al. Determination of diastolic functio n y radionuclide ventriculography. Mayo Clin Proc 1990;65:10071019. RIGHT HEART FAILURE 11. Hurford WE, Zapol WM. The right ventricle and critical illness: a review of anatomy, physiology, and clinical evaluation of its function. Intensive Care Med 1988;14:448457. 12. Isner JM. Right ventricular myocardial infarction. JAMA 1988;259:712718. 13. Ro otham JL, Takala M, Berman M, et al. Ejection fraction revisisted. Anesth esiology 1991;74:172183. P.258 14. Boldt J, Kling D, Moosdorf R, Hempelmann G. Influence of acute volume loadin g on right ventricular function after cardiopulmonary ypass. Crit Care Med 1989 ;17:518521. 15. Reuse C, Vincent JL, Pinsky MR. Measurement of right ventricular volumes dur ing fluid challenge. Chest 1990;98:14501454. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 24)03-May-05 17:12:28

Ovid: ICU Book MANAGEMENT STRATEGIES 16. Franciosa JA. Optimal left heart filling pressure during nitroprusside infus ion for congestive heart failure. Am J Med 1983;74:457464. 17. Chatterjee K, ed. Do utamine. A ten year review. New York: NCM Pu lishers, 1 989. 18. Ro in ED, McCauley R. Nitroprusside-related cyanide poisoning. Time (long pa st due) for urgent, effective interventions. Chest 1992;102:18421845. 19. Te oul J-L. Therapy: effects of vasoactive drugs. In: Edwards JD, Shoemaker WC, Vincent J-L, eds. Oxygen transport. Principles and practice. Philadelphia: W B Saunders, 1993;193208. 20. Gardaz JP, McFarlane PA, Sykes MK. Mechanisms y which dopamine alters lood flow distri ution during lo ar collapse in dogs. J Appl Physiol 1986;60:959964. 21. Milero RR, Fenwell WH, Young JB, et al. Differential systemic arterial and v enous actions and consequent cardiac effects of vasodilator drugs. Prog Cardiova sc Dis 1982;24:353374. 22. Davidson RM. Hemodynamic effects of furosemide in acute myocardial infarctio n. Circulation 1971;54(Suppl II):156. 23. Kiely J, Kelly DT, Taylor DR, Pitt B. The role of furosemide in the treatmen t of left ventricular dysfunction associated with acute myocardial infarction. C irculation 1973;58:581587. 24. Mond H, Hunt D, Sloman G. Haemodynamic effects of frusemide in patients susp ected of having acute myocardial infarction. Br Heart J 1974;36:4453. 25. Nelson GIC, Ahuja RC, Silke B, et al. Haemodynamic advantages of isosor ide http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 24)03-May-05 17:12:28

Ovid: ICU Book dinitrate over frusemide in acute heart failure following myocardial infarction. Lancet 1983a;i:730733. 26. Nelson GIC, Ahula RC, Silke B, et al. Haemodynamic effects of frusemide and its influence on repetitive volume loading in acute myocardial infarction. Eur H eart J 1983 ;4:706711. 27. Tattersfield AE, McNicol MW, Sillett RW. Haemodynamic effects of intravenous frusemide in patients with myocardial infarction and left ventricular failure. Clin Sci Molec Med 1974;46:253264. 28. Larsen FF. Haemodynamic effects of high or low doses of furosemide in acute myocardial infarction. Eur Heart J 1988;9:125131. 29. Dikshit K, Vyden JK, Forrester JS, et al. Renal and extrarenal hemodynamic e ffects of furosemide in congestive heart failure after acute myocardial infarcti on. N Engl J Med 1973;288:10871090. 30. Biddle TL, Yu PN. Effect of furosemide on hemodynamics and lung water in acu te pulmonary edema secondary to acute myocardial infarction. Am J Cardiol 1979;4 3:8690. 31. Nishimura I, Kan e N. The renal and hemodynamic effects of furosemide in acu te myocardial infarction. Crit Care Med 1981;9:829832. 32. Francis GS, Siegel RM, Goldsmith SR, et al. Acute vasoconstrictor response t o intravenous furosemide in patients with chronic congestive heart failure. Ann Intern Med 1986;103:16. P.259 33. van Meyel JJM, Smits P, Russell FGM, et al. Diuretic efficiency of furosemid e during continuous administration versus olus injection in healthy volunteers. Clin Pharmacol Ther 1992;51:440444. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 24)03-May-05 17:12:28

Ovid: ICU Book 34. Martin SJ, Danzinger LH. Continuous infusion of loop diuretics in the critic ally ill: a review of the literature. Crit Care Med 1994;22:13231329. 35. Tam orini G, Pepi M, Susini G, et al. Reversal of cardiogenic shock and seve re mitral regurgitation through verapamil in hypertensive hypertrophic cardiomyo pathy. Chest 1993;104:319324. 36. Nishimura R, Schwartz RS, Holmes DR, Tajik J. Failure of calcium channel lo ckers to improve ventricular relaxation in humans. J Am Coll Cardiol 1993;21:1821 88. 37. Vincent RL, Reuse C, Kahn RJ. Effects on right ventricular function of a cha nge from dopamine to do utamine in critically ill patients. Crit Care Med 1988;1 6:659662. 38. Dell'Italia LJ, Starling MR, Blumhardt R, et al. Comparative effects of volu me loading, do utamine and nitroprusside in patients with predominant right vent ricular infarction. Circulation 1986;72:13271335. MECHANICAL SUPPORT 39. Golding LAR. Postcardiotomy mechanical support. Semin Thorac Cardiovasc Surg 1991;3:2932.

41. Williams DO, Korr KS, Gewirtz H, Most AS. The effect of intra-aortic alloon counterpulsation on regional myocardial lood flow and oxygen consumption in th e presence of coronary artery stenosis with unsta le angina. Circulation 1982;3: 593597. 42. Mackenzie DJ, Wagner WH, Kul er DA, et al. Vascular complications of the int raaortic alloon pump. Am J Surg 1992;164:517521. 43. Killen DA, Piehler JM, Borkon AM, et al. Bio-Medicus ventricular assist devi ce for http://gateway.ut.ovid.com/gw1/ovidwe .cgi (23 of 24)03-May-05 17:12:28

40. Kantrowitz A, Cordona RR, Freed PS. Percutaneous intra-aortic rpulsation. Crit Care Clin 1992;8:819837.

alloon counte

Ovid: ICU Book salvage of cardiac surgical patients. Ann Thorac Surg 1991;52:230235. 44. Lee WA, Gillinov AM, Cameron DE, et al. Centrifugal ventricular assist devic e for support of the failing heart after cardiac surgery. Crit Care Med 1993;21: 11861191. CARDIOPULMONARY BYPASS SURGERY 45. Marino PL, Sink JD. Cardiac performance and systemic oxygen transport after cardiopulmonary ypass surgery. In Salmasi A-M, Iskandrian AS, eds. Cardiac outp ut and regional flow in health and disease. Dordrecht, The Netherlands: Kluwer A cademic Pu lishers, 1993;195212. 46. D'Cruz IA, Callaghan WE. Atypical cardiac tamponade. Clinical and echocardio graphic features. Internal Med Specialist 1988;9:6878. 47. Khoury AF, Afridi I, Quinones MA, et al. Transesophageal echocardiography in critically ill patients: feasi ility, safety, and impact on management. Am Hear t J 1994;127:13631371. 48. Ivanov J, Weisel RD, Mickel orough LL, et al. Rewarming hypovolemia after ao rtocoronary ypass surgery. Crit Care Med 1984;12:10491054. 49. Flaherty JT, Magee PA, Gardner TL, et al. Comparison of intravenous nitrogly cerin and sodium nitroprusside for treatment of acute hypertension developing af ter coronary artery ypass surgery. Circulation 1982;65:10721077. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (24 of 24)03-May-05 17:12:28

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 17 CARDIAC ARREST Medicine cannot, except over a short period, increase the population of the worl d. --Bertrand Russell In 1960, an article was pu lished in the Journal of the Am erican Medical Association that ecame the single most influential study in twen tieth-century medicine. The article, titled Closed-Chest Cardiac Massage, presente d five cases of acute cardiopulmonary arrest, which are summarized in Ta le 17.1 . Although recovery in each case can e attri uted to other interventions (e.g., intu ation and cardioversion), the conclusion of the report stated, Closedchest cardiac massage has een proved to e effective in cases of cardiac arrest (1). T his report represents the irth of what is known today as cardiopulmonary resusc itation (CPR). As shown in Figure 17.1, CPR is far from successful as a life-sav ing intervention (2). Yet despite this poor performance, CPR not only is a unive rsally accepted practice, ut is considered a human right. TABLE 17.1. SUMMARY OF ORIGINAL REPORT ON CLOSED-CHEST CARDIAC MASSAGE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 19)03-May-05 17:15:36

Ovid: ICU Book Figure 17.1. Survival rates for in-hospital cardiac arrest in all clinical repor ts pu lished over a 30-year period from 1960 to 1990. Cases are grouped accordin g to age and initial cardiac rhythm. Total num er (N) of su jects involved is sh own in the upper left corner of the graph. PEA = pulseless electrical activity, V-Tach = ventricular tachycardia. (Data from Schneider AP II et al. In-hospital cardiopulmonary resuscitation: a 30-year review. J Am Board Fam Pract 1993;6:9110 1.) This chapter descri es the mechanical and pharmacologic interventions involved i n the management of cardiac arrest. Also included are recommendations for clinic al monitoring during CPR and some concerns in the early postresuscitation period . A more detailed description of this topic is availa le in the American Heart A ssociation guidelines for asic and advanced cardiac life support (see Suggested Readings at the end of the chapter). BASIC LIFE SUPPORT The ABCs of asic life support are Airway, Breathing, and Circulation. Clearing the airway is achieved y maneuvers such as the Heimlich maneuver that relieve a irway o struction. Breathing is achieved P.261 P.262 y mouth-to-mouth resuscita tion. Circulation involves closed-chest cardiac massage. Circulation is the hear t of asic life support, and it is performed as follows: CHEST COMPRESSIONS q Place the heel of one hand on the lower half of the patient's sternum so that th e long axis of the hand runs perpendicular to the long axis of the sternum. Plac e the other hand, palm down, on top of the sternal hand, and interlace the finge rs while keeping them off the chest. q Lock the el ows so that oth arms are kept straight, and, with the shoulders pos itioned directly a ove the point of contact, depress the sternum y 1.5 to 2 inc hes at a rate of 80 to 100 times per minute. The http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 19)03-May-05 17:15:36

Ovid: ICU Book duration of chest compression should take up half of the total compressionrelease cycle. The traditional ratio of chest compressions to lung inflations is 5:1. q If Steps 1 and 2 do not produce a palpa le carotid or femoral pulse, increase th e force of chest compressions. Pro lems The weak link in CPR is the ina ility of chest compressions to achieve adequate flow to the vital organs. In the original report in 1960, the a ility to achieve a palpa le pulse with chest compressions was mistakenly interpreted as indicati ng that chest compression could achieve adequate systemic lood flow. The pro le m is illustrated in Figure 17.2. The pressure tracings in this figure are taken from a patient P.263 who received standard CPR with rhythmic chest compressions performed as just descri ed. Note that similar pressures were achieved in the ra dial artery and the right atrium (the right-atrial pressure was recorded through a central venous catheter). Therefore, even though the chest compressions produ ced a systolic lood pressure slightly greater than 50 mm Hg, the arteriovenous pressure difference, which is the principal determinant of systemic and regional lood flow, is negligi le. This is why lood flow in oth systemic and regional (e.g., coronary) circulations is less than one-quarter of prearrest levels duri ng closed-chest compressions (4,5 and 6). It also explains why CPR has had such a poor success rate. (If the original investigators had evaluated a venous press ure tracing during chest compressions, we may have een spared many of the false claims a out CPR that predominate today.) Figure 17.2. The influence of chest compressions on arterial and venous (right-a trial) pressure tracings in an adult patient with asystolic cardiac arrest. Note that the difference in arterial and venous (peak) pressures is negligi le. Coronary Perfusion Pressure http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 19)03-May-05 17:15:36

Ovid: ICU Book The difference etween aortic pressure and right-atrial pressure, called coronar y perfusion pressure (CPP), is the pressure gradient that drives coronary lood flow. Studies of CPR outcomes in humans show that a CPP of at least 15 mm Hg is necessary for a satisfactory outcome (7). ACTIVE COMPRESSIONDECOMPRESSION CPR In 1990, a case was reported where a cardiac arrest patient was resuscitated wit h a toilet plunger applied to the anterior chest wall (8). This led to the devel opment of a plunger device that, when applied over the sternum, produces alterna ting chest compression and decompression. Although this device can produce highe r cardiac outputs than standard chest compressions (9), clinical trials with thi s device have not resulted in higher survival rates in either out-of-hospital or in-hospital cardiac arrests (10,11). OPEN-CHEST CARDIAC MASSAGE Emergency thoracotomy with direct cardiac massage can achieve normal and even su pranormal rates of lood flow during CPR (3,4,6). Unfortunately, the role of ope n-chest cardiac massage is limited y the reluctance to perform this procedure o n cardiac arrest patients. ADVANCED LIFE SUPPORT Advanced life support (also called advanced cardiac life support, ACLS) includes maneuvers such as airway intu ation, mechanical ventilation, and adjunctive mea sures (e.g., electric shock and drug administration) to enhance cardiac performa nce and promote lood flow. P.264 The following description focuses on the adjun ctive measures used to promote cardiac output during CPR (intu ation and mechani cal ventilation are presented in Chapter 26, Chapter 27, Chapter 28 and Chapter 29). ACLS ALGORITHMS The flow diagrams in Ta le 17.2, Ta le 17.3, and Ta le 17.4 show the most recent recommendations of the American Heart Association for the management of cardiac arrest due to ventricular fi rillation and pulseless ventricular tachycardia (T a le 17.2); pulseless electrical activity (PEA), which is the new term for elect romechanical dissociation (Ta le 17.3); and ventricular asystole (Ta le 17.4). T he following is a rief description of the resuscitative measures included in th ese diagrams. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 19)03-May-05 17:15:36

Ovid: ICU Book http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 19)03-May-05 17:15:36

Ovid: ICU Book TABLE 17.2. ACLS ALGORITHM FOR VENTRICULAR FIBRILLATION AND PULSELESS VENTRICULA R TACHYCARDIA http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 19)03-May-05 17:15:36

Ovid: ICU Book TABLE 17.3. ACLS ALGORITHM FOR PULSELESS ELECTRICAL ACTIVITY (PEA) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 19)03-May-05 17:15:36

Ovid: ICU Book TABLE 17.4. ACLS ALGORITHM FOR MANAGEMENT OF ASYSTOLE DEFIBRILLATION Direct-current cardioversion is the single most effective resuscitative measure for improving survival in cardiac arrest (2,3,4,5 and 6). In patients with ventr icular tachycardia and ventricular fi rillation, the time from cardiac arrest to defi rillation is the most important factor in determining outcomes. The influe nce of treatment delays on survival is shown in Figure 17.3 (12). The data in th is figure are taken from a study of 1667 cardiac arrest patients with ventricula r fi rillation. Note that survival decreases linearly with increasing time to de fi rillation. Survival decreased from 40% to less than 10% when defi rillation w as delayed 15 minutes (from 5 to 20 minutes after arrest). These results emphasi ze the importance of avoiding delays in initiating defi rillation. Figure 17.3. Survival rate as a function of the time from cardiac arrest to defi rillation in patients with ventricular fi rillation. The num er (N) of su jects studied is shown in the upper left corner of the graph. (Data from Larsen MP et al. Predicting survival from out-of-hospital cardiac arrest: a graphic model. A nn Emerg Med 1993;22:1652.)

Dosage The strength of defi rillation is usually expressed in units of energy (joules) rather than units of electric current (amperes). The recommended energy for thre e successive defi rillations (if necessary) is 200 J, then 300 J, then 360 J. If the initial three defi rillation attempts are unsuccessful, the drugs listed in Ta le 17.2 (e.g., epinephrine and lidocaine) are administered and the sequence of cardioversions is repeated. This pattern of defi rillationsdrugsdefi rillations is the asic management strategy for ventricular tachycardia and fi rillation. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 19)03-May-05 17:15:36

Ovid: ICU Book ROUTES OF DRUG ADMINISTRATION Central versus Peripheral Veins The initial site of venous cannulation for CPR should e the external jugular ve in or the veins in the antecu ital fossa ( ecause these sites do not interfere w ith chest compressions and endotracheal intu ation). P.265 P.266 P.267 Drug admi nistration through peripheral veins should always e olus injection, followed y a 20-mL saline flush (3). If spontaneous circulation does not return after the initial drug injection, central venous cannulation should e performed for su s equent drug administration (3). This latter maneuver reduces the circulation tim e for drug distri ution y at least 2 minutes (3). Endo ronchial Drug Administration When venous access is not readily availa le and an endotracheal tu e is in place , certain drugs can e injected through the endotracheal P.268 tu e (13). The dr ugs that can e given via the endo ronchial route are atropine, epinephrine, and lidocaine. The endo ronchial dose is twice the recommended intravenous dose for each drug (3). Epinephrine seems to e less effective when given via the endo r onchial route, and more than twice the recommended intravenous dose of epinephri ne may e necessary to produce the desired result when the drug is given endo ro nchially (14). All drugs injected into the airways should e diluted in 10 mL sa line or sterile water, and the injection should e made through a long catheter (such as a 20-cm central venous catheter) whose tip extends eyond the tip of th e endotracheal tu e. Drugs should not e injected directly into the endotracheal tu e. Chest compressions should e discontinued while the drug is injected into the upper airways, and the injection should e followed with a few manual lung inflations. This maneuver is effective in promoting drug a sorption from the lun g (13). EPINEPHRINE Intravenous epinephrine is a mainstay in ACLS and is indicated for pulseless ven tricular tachycardiac and ventricular fi rillation, electromechanical dissociati on, and ventricular asystole. The rationale for P.269 epinephrine administration is to promote systemic vasoconstriction and there y direct lood flow to the co ronary and cere ral circulations. Dosage The standard dose of epinephrine in ACLS protocols is 1 mg (10 mL of a 1:10,000 solution), repeated every 3 to 5 minutes if necessary. The optimal dose of epine phrine in CPR may actually e much higher, particularly in larger patients. In a nimal studies, the optimal hemodynamic dose of epinephrine is 0.045 to 2.0 mg/kg (3), which is http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 19)03-May-05 17:15:36

Ovid: ICU Book considera ly higher than the standard epinephrine dose recommended in human ACLS protocols. However, two clinical studies evaluating high-dose epinephrine (7 mg in one study, 0.2 mg/kg in the other) in CPR show no increase in survival with the high-dose regimens compared with standard-dose regimens (15,16). Despite the lack of evidence for improved outcomes with high-dose epinephrine, the American Heart Association now recommends that epinephrine doses can e increased to 5 m g if there is no response to an initial 1-mg dose of the drug (3). ATROPINE Atropine is pro a ly one of the least effective drugs in the ACLS armamentarium. It is most effective for the management of radycardias, ut it is also recomme nded in the management of pulseless electrical activity and ventricular asystole . Dosage The recommended dose of atropine for electromechanical dissociation and asystole is 1 mg y intravenous injection, repeated every 3 to 5 minutes if necessary. A total dose of 3 mg (or 0.04 mg/kg) produces complete vagal lockade, so this do se should not e exceeded. Atropine doses that are less than 0.5 mg can have par asympathomimetic effects (i.e., they can promote radycardia), and thus should e avoided (3). BICARBONATE The recommendations for icar onate administration in CPR have een revised cons idera ly in recent years ecause of an accumulation of studies showing little e nefit and possi le harm associated with icar onate administration in meta olic acidoses (17,18 and 19). Of note is recent evidence showing that icar onate adm inistration in doses recommended for CPR (1 mg/kg) does not result in enhanced v asopressor actions of epinephrine (20). The current recommendations for icar on ate therapy in CPR are P.270 shown in Ta le 17.5. As indicated at the ottom of the ta le, icar onate is no longer recommended in patients with ischemic lactic acidosis. In fact, it is considered potentially harmful in this condition. The effects of icar onate administration are descri ed in Chapter 37. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 19)03-May-05 17:15:36

Ovid: ICU Book TABLE 17.5. RECOMMENDATIONS FOR BICARBONATE ADMINISTRATION CALCIUM Despite the fact that extracellular calcium enhances the contractile force of ca rdiac muscle, there is no evidence to indicate that calcium administration durin g CPR improves cardiac performance. In fact, ischemia promotes the intracellular accumulation of calcium, and this can lead to mem rane disruption and uncouplin g of oxidative phosphorylation (21). Because of the risk of calcium accumulation and su sequent cell injury during periods of tissue ischemia, the indications f or calcium administration during CPR are restricted to cases of acute hyperkalem ia, ionized hypocalcemia, and calcium channel locker overdose. DEXTROSE INFUSIONS Although dextrose is a popular additive in intravenous fluids, dextrose administ ration can have deleterious effects in critically ill patients (22). As mentione d in Chapter 15, dextrose infusions can enhance the production of lactic acid in critically ill patients (see Fig. 15.2). The accumulation of lactic acid can it self promote cell injury, possi ly y promoting the formation of toxic oxygen me ta olites (23). This may explain why hyperglycemia enlarges infarct size in anim al studies of cere rovascular occlusion (24). The impact of car ohydrate infusio ns during CPR is not clear. However, the current recommendations from the Americ an Heart Association are that dextrose infusions are a Class III intervention (h armful) and thus should e avoided (3). CLINICAL MONITORING The num er one pro lem with CPR is the ina ility of chest compressions to mainta in adequate organ lood flow. The num er two pro lem P.271 is the ina ility to m onitor the adequacy of organ perfusion during CPR. The presence of palpa le puls es and arterial pressure waves is not an indication of lood flow (the differenc e etween pressure waves and flow waves is descri ed in Chapter 9). The measurem ents descri ed in this section can provide a more accurate assessment of organ p erfusion than the standard measures used to evaluate the response to CPR. END-TIDAL CO2 PRESSURE The excretion of car on dioxide in exhaled gas is a function of pulmonary lood flow (cardiac output), and thus the level of CO2 in exhaled gas changes in direc t proportion to changes in cardiac output. The CO2 pressure (PCO2 ) in end-expir atory gas (i.e., the end-tidal PCO2 ) is easy to measure at the edside, and cha nges in end-tidal PCO2 can e used as a noninvasive marker of changes in cardiac output (see Chapter 22 for a detailed description of the end-tidal CO2 measurem ent and its applications). End-tidal PCO2 has een used to monitor the cardiac o utput in hypovolemia (see Fig. 14.1) and during cardiopulmonary resuscitation (2 5,26 and 27). Prognostic Value A progressive rise in end-tidal PCO2 during CPR indicates that the resuscitation effort is successful in promoting http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 19)03-May-05 17:15:36

Ovid: ICU Book cardiac output. As such, a steady rise in end-tidal PCO2 during CPR is more like ly to e associated with a successful outcome than a persistently low end-tidal PCO2 . The correlation etween survival and end-tidal PCO2 during CPR is shown i n Figure 17.4. The data in this figure are from a study of P.272 90 cardiac arre st patients with PEA (27). The initial end-tidal PCO2 measurement, o tained at t he onset of CPR, is very low (11 to 12 mm Hg, compared to a normal end-tidal PCO 2 of 40 to 45 mm Hg), and is similar in survivors and nonsurvivors. However, in the survivors, the end-tidal PCO2 increased considera ly (from 12 to 31 mm Hg) a fter 20 minutes of CPR, whereas in the nonsurvivors the end-tidal PCO2 decreased further (from 10.9 to 3.9 mm Hg). These results are supported y the results of another study where CPR that failed to raise the end-tidal PCO2 a ove 10 mm Hg was universally unsuccessful (26). Figure 17.4. Changes in end-tidal PCO2 during CPR in survivors and nonsurvivors of cardiac arrest associated with pulseless electrical activity. Open circles re present the mean values for each group. (Data from Wayne MA et al. Use of end-ti dal car on dioxide to predict outcome in prehospital cardiac arrest. Ann Emerg M ed 1995;25:762 767.) The tendency of the end-tidal PCO2 to rise during CPR can thus e a valua le pro gnostic marker. When end-tidal PCO2 does not rise a ove 10 mm Hg after a resusci tation time of 15 to 20 minutes, the resuscitative effort is unlikely to e succ essful. VENOUS BLOOD GASES The common practice of monitoring arterial lood gases during CPR should e a an doned in favor of monitoring venous lood gases. The rationale for this switch i s the greater propensity for venous lood to represent the oxygenation and acid a se status of peripheral tissues (28,29). The tendency for arterial lood gases t o provide misleading information during CPR is demonstrated y the o servation t hat arterial lood can show a respiratory alkalosis while venous lood shows a m eta olic acidosis during CPR (28,29). The superiority of venous lood gases for monitoring tissue events during CPR (or in any low-flow state) has een ignored for over a decade, resulting in su optimal care for cardiac arrest patients. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 19)03-May-05 17:15:36

Ovid: ICU Book HOW LONG TO RESUSCITATE There is little dou t that CPR is inappropriately prolonged in a significant per centage of resuscitative efforts. The goal of prolonged CPR is to increase the c hance for survival, ut this is not a desira le goal if the survivor is mentally impaired, as is often the case in survivors of prolonged CPR. Ischemic Time and Neurologic Recovery The risk of functional impairment in any of the major organs is directly related to the duration of the ischemic insult. The ischemic time following cardiac arr est includes the time from onset of the arrest to onset of CPR (arrest time) and the duration of the resuscitative effort (CPR time). The influence of these two ischemic times on neurologic recovery is shown in Figure 17.5. The data in this figure are taken from a multicenter study of patients who did not regain consci ousness in the first hour following successful CPR (30). If the arrest time was less than 6 minutes and the CPR time did not exceed 30 minutes, half of the surv ivors had a satisfactory neurologic recovery. However if P.273 the arrest time e xceeded 6 minutes, more than 15 minutes of CPR always produced neurologic impair ment in the survivors. Thus, in witnessed cardiac arrest (when arrest time can e accurately determined), CPR can e continued for 30 minutes if the arrest time is less than 6 minutes, ut if the arrest time is longer than 6 minutes, CPR sh ould e terminated after 15 minutes. Figure 17.5. Time sequence plots showing the incidence of satisfactory neurologi c recovery as a function of oth the duration of the resuscitative effort (CPR t ime) and the time from cardiac arrest to CPR (arrest time). The study group incl uded 262 survivors of cardiac arrest who did not regain consciousness in the fir st hour after CPR. (From A ramson NS et al. Neurologic recovery after cardiac ar rest: effect of duration of ischemia. Crit Care Med 1985;13:930931.) POSTRESUSCITATION CONCERNS When CPR is successful in restoring spontaneous circulation, two concerns deserv e attention in the early postresuscitation period. The first is the potential fo r continued and progressive multiorgan damage (i.e., http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 19)03-May-05 17:15:36

Ovid: ICU Book postresuscitation injury). The second is the likelihood of neurologic recovery i n patients who do not regain consciousness immediately after CPR. POSTRESUSCITATION INJURY The phenomenon of postresuscitation organ damage was descri ed at the end of Cha pter 14. This condition is usually seen after prolonged P.274 ischemic times, an d it is characterized y progressive dysfunction in multiple organs. Other, more familiar terms for this condition are multiple organ failure (31) and multiple organ dysfunction syndrome (32). This condition is often fatal, and there is no effective therapy at present. Several mechanisms have een proposed for this con dition, including persistent vasoconstriction (i.e., the no-reflow phenomenon) a nd the release of toxins produced during the period of ischemia (i.e., reperfusi on injury). (See Chapter 31 for a more detailed description of the multiple orga n failure and its management.) NEUROLOGIC RECOVERY Neurologic impairment is common in cardiac arrest patients who are successfully resuscitated. Many survivors do not regain consciousness immediately after CPR, and the following are some prognostic factors that help identify patients who ar e unlikely to awaken or achieve a satisfactory neurologic recovery. Duration of Coma Failure to regain consciousness in the first few hours after CPR is not a har in ger of prolonged or permanent neurologic impairment, (33). However, coma that pe rsists longer than 4 hours after CPR carries a poor prognosis for full neurologi c recovery. The relationship etween neurologic recovery and coma that persists longer than 4 hours after CPR is shown in Figure 17.6 (34). Although the recover y rates are low for all points on the graph, there is a linear decline in recove ry rate as the duration of coma increases. After 1 day of persistent coma, only 10% of the patients achieved a satisfactory neurologic recovery. The recovery ra te drops elow 5% when the coma lasts 1 week, and no patient recovers neurologic function when the coma persists for 2 weeks. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 19)03-May-05 17:15:36

Ovid: ICU Book Figure 17.6. Graph showing the relationship etween the duration of coma and the incidence of favora le neurologic recovery. The num er (N) of su jects studied is indicated at the top of the graph. (Data from Levy DE et al. Prognosis in non -traumatic coma. Ann Intern Med 1981;94:293301.) The data in Figure 17.6 can e useful for identifying patients who are not likel y to enefit from therapeutic interventions (i.e., futile care). It can also e useful for determining the appropriate time to inform close relatives a out the patient's prognosis. Three days of persistent coma is my threshold for informing families of the poor prognosis for recovery. The actual time selected for infor ming families of a poor prognosis is a matter of individual preference. The impo rtant point is to keep the family informed and to provide guidance in decisions a out future strategies. P.275 Coma Scores Scoring systems such as the Glasgow Coma Scale (GCS) can also provide valua le p rognostic information. (This scoring method is descri ed in Chapter 50.) A GCS elow 5 on the third day of persistent coma is almost always associated with a po or outcome (33). Pupillary Light Reflex Several rainstem reflexes can have prognostic value in patients who do not rega in consciousness after CPR, ut none can match the predictive value of the pupil lary light reflex. The importance of this reflex is its negative predictive valu e (i.e., the a ility to identify a poor outcome). A sence of the pupillary light reflex after one or more days of coma indicates little or no chance for neurolo gic recovery. This reflex has no prognostic value in the first 6 hours after CPR ecause it can e transiently lost and then reappear (35). Finally, the resusci tation drugs atropine and epinephrine can produce pupillary dilation, ut these agents do not interfere with the pupillary response to light (33,36). P.276 REFERENCES GENERAL TEXTS Baskett PJF. Resuscitation hand ook. 2nd ed. London: Mos y Europe, Ltd., 1993. Cummins RO, ed. Text ook of advanced cardiac life support. Dallas: American Hear t Association, 1994. Paradis NA, Halperin HL, Nowak RM. Cardiac arrest. The science and practice of r esuscitation medicine. Baltimore: Williams & Wilkins, 1995. Parr MJA, Craft TM. Resuscitation: key data. Oxford: Bios Scientific Pu lishers, Ltd., 1994. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 19)03-May-05 17:15:36

Ovid: ICU Book HISTORICAL REPORTS 1. Kouwenhoven WB, Ing, Jude JR, Knicker ocker GG. Closed chest cardiac massage. JAMA 1960;173:10641067. 2. Schneider AP II, Nelson DJ, Brown DD. In-hospital cardiopulmonary resuscitati on. A 30-year review. J Am Board Fam Pract 1993;6:91101. GUIDELINES 3. Emergency Cardiac Care Committee and Su committees, American Heart Associatio n. Guidelines for cardiopulmonary resuscitation and emergency cardiac care. JAMA 1992;268:21712241. REVIEWS 4. Weil MH, Gazmuri RJ, Rackow EC. The clinical rationale of cardiac resuscitati on. Dis Mon 1990;36:423 468 (195 References). 5. Barton CW, Manning JE. Cardiopulmonary resuscitation. Emerg Med Clin North Am 1995;13:811830 (101 References). 6. DeBehnke DJ, Swart GL. Cardiac arrest. Emerg Med Clin North Am 1996;14:5782 (1 12 References). BASIC LIFE SUPPORT 7. Paradis NA, Martin GB, Rivers EP, et al. Coronary perfusion pressure and the return of spontaneous circulation in human cardiopulmonary resuscitation. JAMA 1 990;263:11061113. 8. Lurie KG, Lindo C, Chin J. CPR: the P stands for plum er's helper. JAMA 1990; 264:1661. 9. Orliaguet GA, Carli PA, Rozen erg A, et al. End-tidal car on dioxide during o ut-of-hospital cardiac arrest resuscitation: comparison of active compressiondeco mpression and standard CPR. Ann Emerg Med 1995;25:4851. 10. Lurie KJ, Shultz JJ, Callahan ML, et al. Evaluation of active compressiondeco mpression CPR in victims of out-of-hospital cardiac arrest. JAMA 1994;271:1405141 1. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 19)03-May-05 17:15:36

Ovid: ICU Book 11. Tucker KJ, Galli F, Savitt MA, et al. Active compressiondecompression resusci tation: effect on resuscitation success after in-hospital cardiac arrest. J Am C oll Cardiol 1994;24:201209. ADVANCED LIFE SUPPORT 12. Larsen MP, Eisen erg M, Cummins RO, Hallstrom AP. Predicting survival from o ut of hospital cardiac arrest: a graphic model. Ann Emerg Med 1993;22:16521658. 13. Aitkenhead AR. Drug administration during CPR: what route? Resuscitation 199 1;22:191195. P.277 14. McCrirrick A, Kestin I. Haemodynamic effects of tracheal compared with intra venous adrenaline. Lancet 1992;340:868870. 15. Steill IG, He ert PC, Weitzman BN, et al. High-dose epinephrine in adult car diac arrest. N Engl J Med 1992;327:10451050. 16. Brown CG, Martin DR, Pepe P, et al. A comparison of standard-dose and high-d ose epinephrine in cardiac arrest outside the hospital. N Engl J Med 1992;327:10 511055. 17. Hindman BJ. Sodium icar onate in the treatment of su types of acute lactic acidosis: physiologic considerations. Anesthesiology 1990;72:10641076. 18. Arieff AI. Indications for the use of icar onate in patients with meta olic acidosis. Br J Anesth 1991;67:165177. 19. Ritter JM, Doktor HS, Benjamin N. Paradoxical effect of icar onate on cytop lasmic pH. Lancet 1990;335:12431246. 20. Bleske BE, Rice TL, Warren EW, et al. The effect of sodium icar onate admin istration on the vasopressor effect of high-dose epinephrine during cardiopulmon ary resuscitation in swine. Am J Emerg Med 1993;11:439443. 21. Rees AP, Valentino VA, Genton E. Pharmacological adjuncts to cardiopulmonary resuscitation. Intern Med 1991;12:2235. 22. Marino PL, Finnegan MJ. Nutrition support is not eneficial, and can e harm ful, in critically ill patients. Crit Care Clin 1996;12:667676. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 19)03-May-05 17:15:36

Ovid: ICU Book 23. Sie er FE, Traytsman RJ. Special issues: glucose and the rain. Crit Care Me d 1992;20:104114. 24. de Courten-Meyers G, Meyers RE, Schoolfield L. Hyperglycemia enlarges infarc t size in cere rovascular occlusion in cats. Stroke 1988;19:623630. CLINICAL MONITORING 25. Falk JL, Rackow EC, Weil MH. End-tidal car on dioxide concentration during c ardiopulmonary resuscitation. N Engl J Med 1988;318:607611. 26. Sanders AB, Kern KB, Otto CW, et al. End-tidal car on dioxide monitoring dur ing cardiopulmonary resuscitation. JAMA 1989;262:13471351. 27. Wayne MA, Levine RL, Miller CC. Use of end-tidal car on dioxide to predict o utcome in prehospital cardiac arrest. Ann Emerg Med 1995;25:762767. 28. Weil MH, Rackow EC, Trevino R. Difference in acid ase state etween venous an d arterial lood during cardiopulmonary resuscitation. N Engl J Med 1986;315:1531 56. 29. Steedman DJ, Ro ertson CE. Acid ase changes in arterial and central venous l ood during cardiopulmonary resuscitation. Arch Emerg Med 1992;9:169176. POST-RESUSCITATION CONCERNS 30. A ramson NS, Safar P, Detre KM, et al. Neurologic recovery after cardiac arr est: effect of duration of ischemia. Crit Care Med 1985;13:930931. 31. Cerra F. Multiple organ failure syndrome. Dis Mon 1992;12:845878. 32. The ACCP/SCCM Consensus Conference Committee. Definitions for sepsis and org an failure and guidelines for the use of innovative therapies in sepsis. Chest 1 992;101:16441655. 33. Edgren E, Hedstrand U, Kelsey S, et al. Assessment of neurologic prognosis i n comatose survivors of cardiac arrest. Lancet 1994;343:10551059. 34. Levy DE, Caronna JJ, Singer BH, et al. Predicting outcome from hypoxic-ische mic coma. JAMA 1985;253:14201426. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 19)03-May-05 17:15:36

Ovid: ICU Book 35. Steen-Hansen JE, Hansen NM, Vaagenes P, Schreiner B. Pupil size and light re activity during cardiopulmonary resuscitation: a clinical study. Crit Care Med 1 988;16:6970. 36. Goetting MG, Contreras E. Systemic atropine during cardiac arrest does not c ause fixed and dilated pupils. Ann Emerg Med 1991;20:5557. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 19)03-May-05 17:15:36

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 18 HEMODYNAMIC DRUGS This chapter contains a rief description of eight pharmaceutical agents given y continuous intravenous infusion to support the circulation. Each agent is list ed elow in order of presentation. Drugs marked y an asterisk have a dose chart included in the chapter. q Amrinone* q Do utamine* q Dopamine* q Epinephrine q La etalol q Nitroglycerin* q Nitroprusside* q Norepinephrine INFUSION RATES Because the drugs in this chapter are administered y continuous infusion, the r ecommended doses are expressed as dose rates, either in micrograms per minute ( /m in) or micrograms http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 26)03-May-05 17:23:29

Ovid: ICU Book per kilogram ody weight per minute ( /kg/min). To deliver the recommended dose ra te, the concentration of the drug in the infusate must e known. The infusion ra te (i.e., the rate at which the infusate is delivered) is then determined as the ratio of the dose rate to the drug concentration in the infusate. This is shown in Ta le 18.1. In this case, the desired dose rate is R /min, and the drug conce ntration in the infusate is C /mL, so the ratio R/C derives the infusion rate in mL/min. The infusion rate can e converted to microdrops/minute y multiplying m L/min y 60 ( ecause there are 60 microdrops per mL). The conversion to P.279 mi crodrops facilitates control of infusion rates when small volumes of fluid are i nfused. Therefore, all infusion rates included in the drug dosing charts in this chapter are expressed in microdrops/minute. The volumetric equivalent of microd rops/minute is mL/hr (i.e., microdrops/ minute 60/60 = mL/hr). TABLE 18.1. DETERMINING DRUG INFUSION RATES AMRINONE Amrinone is a phosphodiesterase inhi itor that has oth positive inotropic and v asodilator actions (1,2). Despite the potential enefit of its com ined actions (i.e., greater augmentation of cardiac output with less cardiac work), amrinone has not proven superior to single-action cardiotonic agents such as do utamine ( 3,4). ACTIONS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 26)03-May-05 17:23:29

Ovid: ICU Book The com ined actions of amrinone produce an increase in cardiac stoke output wit hout an increase in cardiac stroke work (4). The effects of amrinone on cardiac performance are roughly equivalent to those of do utamine (3,4 and 5). However, ecause amrinone does not stimulate adrenergic receptors, its effects can add to the effects of do utamine (4). Unlike do utamine, the actions of amrinone are n ot attenuated y -receptor antagonists.

Indications Amrinone is effective as single-agent therapy in the management of low output st ates caused y systolic heart failure. However, it is most often used as a secon d agent that is added to do utamine in cases of refractory heart failure (3,4). DRUG ADMINISTRATION A drug dosing chart for amrinone is shown in Ta le 18.2. Amrinone is degraded y dextrose and y light. Therefore, amrinone should not e infused in dextrose-co ntaining fluids, and the infusion solution P.280 should e protected from light. An initial loading dose is used for amrinone therapy. The usual dose is 0.75 mg /kg ut can e as high as 1.5 mg/kg (6). This is followed y a continuous infusi on, usually in the range of 5 to 10 /kg/min. A dose rate of 10 /kg/min will achiev e the desired hemodynamic response in over 80% of patients (3). TABLE 18.2. AMRINONE DOSAGE CHART http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 26)03-May-05 17:23:29

Ovid: ICU Book Incompata ilities Furosemide should not e injected into intravenous lines carrying amrinone infus ions ecause the drug forms a precipitate when added to amrinone solutions (see Hand ook of Injecta le Drugs). ADVERSE EFFECTS Although the oral form of the drug was discontinued y the FDA ecause of a high incidence of complications, short-term therapy with intravenous amrinone is rel atively free of adverse effects. Throm ocytopenia caused y nonimmunogenic plate let destruction was a common side effect of chronic therapy with oral amrinone, ut this complication is reported in only 2 to 3% of patients receiving short-te rm intravenous amrinone (7). The pro lem resolves when the drug is discontinued, and no cases of a normal leeding have een reported. Hypotension caused y exc essive vasodilation has een a common complication of amrinone therapy in some r eports, ut it seems to occur mostly in hypovolemic patients (5,7). Contraindications Amrinone is contraindicated in patients with hypertrophic cardiomyopathy (7). Th rom ocytopenia is not a contraindication to intravenous amrinone, ut it is pro a ly wise to avoid the drug if possi le in patients with platelet counts elow 5 0,000/mL. P.281 DOBUTAMINE Do utamine is a synthetic catecholamine that is generally considered the inotrop ic drug of choice for the acute management of severe (systolic) heart failure (1 ,3). It is primarily a 1receptor agonist (cardiac stimulation), ut it also has m ild -2 effects (vasodilation). ACTIONS As demonstrated in Figure 18.1, do utamine causes a dose-dependent increase in s troke volume (upper graph) accompanied y a decrease in cardiac filling pressure s (lower graph). The increase in stroke output is usually accompanied y a propo rtional decrease in systemic vascular resistance ( aroreceptor-mediated), and th us the arterial lood pressure usually remains unchanged. The drug is effective in oth right- and left-sided heart failure (8,9 and 10). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 26)03-May-05 17:23:29

Ovid: ICU Book Figure 18.1. Effects of do utamine and dopamine on cardiac performance in patien ts with severe heart failure. (Data from Leier CV et al. Comparative systemic an d regional hemodynamic effects of dopamine and do utamine in patients with cardi omyopathic heart failure. Circulation 1978;58:466475.) The inotropic and chronotropic effects of do utamine can vary widely in critical ly ill patients (8). This is partly due to varia le pharmacokinetics (11) and pa rtly due to varia le end-organ responsiveness. Elderly patients are relatively r esistant to do utamine and can have only half the inotropic responsiveness seen in younger patients (12). P.282 The varia le response to do utamine in criticall y ill patients emphasizes the need to guide do utamine therapy y preselected he modynamic end-points, not y preselected dose rates. Indications As mentioned, do utamine is the preferred inotropic agent for the acute manageme nt of low output states due to systolic heart failure. Because do utamine does n ot usually raise the arterial lood pressure, it is not indicated as monotherapy in patients with cardiogenic shock. Do utamine is also used in patients with se ptic shock and multiple organ failure who may have a normal cardiac output. Thes e conditions are often accompanied y hypermeta olism, and in http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 26)03-May-05 17:23:29

Ovid: ICU Book this situation, a normal cardiac output may e not e adequate for the increased oxygen requirements of hypermeta olism. The goal of do utamine therapy in these conditions is to drive the cardiac output to supranormal levels (e.g., > 4.5 L/ min/m2) to meet the increased oxygen consumption of the hypermeta olic state (13 ,14). The use of do utamine to achieve a hyperdynamic state has had an inconsist ent effect on survival (see Chapter 31), and thus is not universally accepted. DRUG ADMINISTRATION A do utamine dose chart is shown in Ta le 18.3. The drug is availa le in 250-mg vials and is infused in a concentration of 1 mg/mL. The usual dose range is 5 to 15 /kg/min, ut doses as high as 200 /kg/min have een used to achieve a hyperdyn amic state in patients with septic shock and multiple organ failure (14). TABLE 18.3. DOBUTAMINE DOSAGE CHART Incompati ilities An alkaline pH inactivates catecholamines such as do utamine (15), and thus sodi um icar onate or other alkaline solutions should not e administered through in travenous tu ing used for do utamine infusions. P.283 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 26)03-May-05 17:23:29

Ovid: ICU Book ADVERSE EFFECTS Do utamine has few serious side effects. As mentioned, tachycardia can develop i n some patients. However, malignant tachyarrhythmias are uncommon (7). Contraindications Do utamine is not indicated for the management of heart failure due to diastolic dysfunction and is contraindicated in patients with hypertrophic cardiomyopathy . DOPAMINE Dopamine is an endogenous catecholamine that serves as a neurotransmitter. As an exogenous agent, it produces a dose-dependent activation of several types of ad renergic and dopaminergic receptors (10). The overall effect of the drug is dete rmined y the pattern of receptor activation, as descri ed elow. ACTIONS When given at low dose rates (0.5 to 3 /kg/min), dopamine selectively activates d opaminespecific receptors in the renal, mesenteric, and cere ral circulations an d increases lood flow in these regions. Dopaminergic activation in the kidneys also produces an increase in urinary sodium and water excretion that is independ ent of the changes in renal lood flow (16). At intermediate dose rates (3 to 7. 5 /kg/min), dopamine stimulates -receptors in the heart and peripheral circulation s, and this produces an increase in cardiac output. The effects of incremental d oses of dopamine on cardiac stroke output are shown in Figure 18.1 (upper graph) . Note that the inotropic response to dopamine is modest when compared to do uta mine. At high dose rates (> 7.5 /kg/min), dopamine produces a dose-dependent acti vation of receptors in the systemic nd pulmon ry circul tions. This results in p rogressive v soconstriction, nd the result nt incre se in ventricul r fterlo d limits the bility of dop mine to ugment c rdi c output. The loss in c rdi c o utput ugment tion t higher doses of dop mine is shown in Figure 18.1 (upper gr ph). The effects of dop mine on the pulmon ry c pill ry wedge pressure re show n in Figure 18.1 (lower gr ph). There is dose-dependent incre se in the wedge pressure, which is independent of the ch nges in stroke volume in the upper gr p h. This effect m y be the result of v soconstriction in pulmon ry veins. Dop min e-induced constriction P.284 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 26)03-M y-05 17:23:29

Ovid: ICU Book of pulmon ry veins is n import nt consider tion, bec use it inv lid tes the pul mon ry c pill ry wedge pressure s me sure of left-ventricul r filling pressur es (see Ch pter 11). The hemodyn mic responses to dop mine re blunted by contin ued drug dministr tion (10). This t chyphyl xis m y be due to dop mine's bilit y to rele se norepinephrine from drenergic nerve termin ls. When t chyphyl xis to dop mine develops, discontinuing the drug for few d ys (if possible) c n re store some of the end-org n responsiveness. Indic tions Dop mine is indic ted for the m n gement of c rdiogenic shock nd ny circul tor y shock syndrome ssoci ted with systemic v sodil tion (e.g., septic shock). The drug is p rticul rly v lu ble for its bility (in intermedi te-to-high dose r t es) to promote v soconstriction while preserving the c rdi c stroke output. Lowdose dop mine is lso used to preserve ren l blood flow nd to promote urine out put in p tients with oliguric cute ren l f ilure, or in those t risk for oligu ric ren l f ilure. Although dop mine does not improve intrinsic ren l function i n this situ tion, it c n promote urine output nd limit fluid retention (16). DRUG ADMINISTRATION A dose ch rt for dop mine dministr tion is shown in T ble 18.4. This ch rt iden tifies three dose r nges b sed on the most prominent clinic l response. At low i nfusion r tes of 0.5 to 3 / kg/min, n triuresis nd diuresis re prominent. As th e infusion r te is incre sed to 4 to 7 /kg/ min, -receptor stimulation and augment ation of cardiac output occurs. At dose rates a ove 8 /kg/min, progressive vasoco nstriction is the dominant feature. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 26)03-May-05 17:23:29

Ovid: ICU Book TABLE 18.4. DOPAMINE DOSAGE CHART P.285 Incompati ilities The precautions for alkaline fluids mentioned for do utamine also apply to dopam ine. ADVERSE EFFECTS Tachyarrhythmias are the most common complication of dopamine administration. Si nus tachycardia is common (8,17) and can occur at -agonist dose rates (i.e., 5 to 7 /kg/min) (8). Malignant tachyarrhythmias (e.g., multifocal ventricular ectopic s, ventricular tachycardia) can also occur, ut are uncommon. The most feared co mplication of dopamine administration is ischemic lim necrosis, which occurs mo re frequently with dopamine than with any other vasoconstrictor agent (7). Lim necrosis has een reported at dopamine doses as low as 1.5 /kg/min (7). Prompt ad ministration of an -receptor blocking gent such s phentol mine (5 mg s n intr venous bolus, followed by continuous infusion t 1 to 2 mg/min) is indic ted t the e rliest signs of limb ischemi . V soconstrictor doses of dop mine should not be given through peripher l veins. Extr v s tion of the drug through peri pher l vein c n be tre ted with loc l injection of phentol mine (5 to 10 mg in 15 mL s line) (7). EPINEPHRINE Epinephrine is n endogenous c techol mine nd the prototype symp thomimetic ge nt. Bec use of its potency nd risk for dverse effects, epinephrine is used sp ringly to support the circul tion in conditions other th n c rdi c rrest. ACTIONS Like dop mine, epinephrine is prim rily -receptor agonist at low doses and an -r eceptor gonist t high doses. However, epinephrine is much more potent th n dop mine, with n effective dose r nge th t is two-to-three orders of m gnitude bel ow the effective dose r nge for dop mine. As shown in T ble 18.5, epinephrine c tiv tes -receptors P.286 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 26)03-May-05 17:23:29

Ovid: ICU Book at doses of only 0.005 to 0.02 /kg/min. -Receptor v soconstriction ppe rs t slig htly higher doses, nd ren l v soconstriction develops e rly (7). Dose r tes bo ve 0.1 /kg/min c n produce severe v soconstriction. TABLE 18.5. EPINEPHRINE DOSAGE RECOMMENDATIONS Antiinfl mm tory Effects Epinephrine blocks the rele se of infl mm tory medi tors by m st cells nd b sop hils in response to n ntigenic ch llenge. This effect m y expl in the s lut ry effects of epinephrine in n phyl ctic re ctions (18). Met bolic Effects Epinephrine h s sever l met bolic effects th t represent d ptive responses in h e lthy subjects but c n be deleterious in the critic lly ill p tient (7). The me t bolic effects th t deserve mention include ( ) hypermet bolism (c lorigenic re sponse), (b) hyperglycemi (enh nced gluconeogenesis nd diminished insulin rele se), (c) n incre se in circul ting keto cids (vi lipolysis), (d) hyperl ct te mi (without ischemi ), nd (e) decre se in serum pot ssium (usu lly < 1 mEq/L ). Indic tions Intr venous epinephrine is indic ted for the m n gement of c rdi c rrest ssoci ted with pulseless ventricul r t chyc rdi nd ventricul r fibrill tion, systo le, nd pulseless electric l http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 26)03-M y-05 17:23:29

Ovid: ICU Book ctivity (see Ch pter 17). It is lso indic ted for severe n phyl ctic re ction s nd n phyl ctic shock. Bec use of the n rrow ther peutic r nge nd risk of d verse re ctions, it is not recommended s first-line gent for the routine m n gement of low c rdi c output or circul tory shock.

DRUG ADMINISTRATION Epinephrine is v il ble s 1:1000 solution (1 mg/mL) nd c n be diluted to cr e te 1:10,000 solution (0.1 mg/mL). As indic ted in T ble 18.5, epinephrine is powerful agonist, with -receptor activation at dose rates of only 0.005 to 0.0 2 /kg/min. The safe range for epinephrine infusions is exceeded at dose rates a o ve 0.1 /kg/min (8). The recommended epinephrine doses for anaphylaxis are shown a t the ottom of Ta le 18.5 (18). Epinephrine is the single most effective drug i n the management of anaphylaxis, and delays in administering the drug can have a dverse consequences (18). P.287 Incompati ilities Like other catecholamines, epinephrine is inactivated y alkaline solutions. ADVERSE EFFECTS Epinephrine is arrhythmogenic, particularly in com ination with halothane or ele ctrolyte a normalities (7). Coronary ischemia can also occur and is not related to dose (7). Although renal vasoconstriction is prominent with epinephrine, isch emic renal failure is seen most often with accidental epinephrine overdose (7). Epinephrine can produce serious hypertension in patients receiving -receptor anta gonists, an effect attri uted to unopposed -receptor stimul tion (7). C lorigenic Effect Ther peutic doses of epinephrine c n produce 35% incre se in resting met bolic r te (19), nd the incre se in tissue oxygen needs c n h ve dverse consequence s in p tients with imp ired or borderline tissue oxygen tion. Dop mine h s sim il r but less pronounced c lorigenic effect (19), where s dobut mine seems to h ve little or no effect on the met bolic r te in critic lly ill p tients (20). LABETALOL http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 26)03-M y-05 17:23:29

Ovid: ICU Book L bet lol is n drenergic-receptor nt gonist th t h s proven effective in cut e m n gement of severe hypertension. P renter l dministr tion of l bet lol c n serve s s fer ltern tive to nitroprusside (21,22 nd 23). ACTIONS L bet lol is nonselective -receptor antagonist that also locks -receptormedi ted v soconstriction. The over ll effect is dose-rel ted decre se in systemic v s cul r resist nce nd blood pressure, without reflex t chyc rdi or incre se in c rdi c output. Unlike nitroglycerin nd nitroprusside, l bet lol does not incr e se intr cr ni l pressure (20). Indic tions L bet lol is indic ted for the cute m n gement of severe hypertension ssoci te d with norm l or dequ te c rdi c output. It m y be p rticul rly effective in hypertension c used by excess circul ting c techol mines, such s the hypertensi on th t occurs in the e rly postoper tive period (23). Bec use the ntihypertens ive ctions of l bet lol re P.288 not ccomp nied by n incre se in c rdi c out put, the drug is p rticul rly useful in the m n gement of ortic dissection (22) . DRUG ADMINISTRATION L bet lol is v il ble in n queous solution (5 mg/mL) th t c n be given intr v enously s bolus injection or by continuous infusion. Bolus Ther py P tients should be pl ced in the supine position for bolus injections of l bet l ol to limit the risk of orthost tic hypotension. The initi l dose is 20 mg, nd repe t doses of 40 mg c n be given t 10-minute interv ls until the desired nti hypertensive effect is chieved. Although the m nuf cturer recommends m ximum cumul tive dose of 300 mg, l rger cumul tive doses of l bet lol h ve been used w ithout ill effects (23). Continuous Infusion Ther py Continuous infusions of l bet lol should be preceded by bolus dose of 20 mg, b ec use the serum h lf-life of l bet lol, which is 6 to 8 hours, indic tes th t 3 0 to 40 hours (5 h lf-lives) http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 26)03-M y-05 17:23:29

Ovid: ICU Book

ADVERSE EFFECTS The most not ble complic tions of intr venous l bet lol include orthost tic hypo tension (block de), myoc rdi l depression (-1 lockade), and ronchospasm (-2 lock ade). Orthostasis should not e a pro lem in the ICU, ecause patients are rarel y am ulatory or in the upright position. The drug should e avoided in patients with heart failure or asthma. NITROGLYCERIN Nitroglycerin is a peculiar chemical ecause it is oth an explosive powder and an effective antianginal agent. It is an organic nitrate (glyceryl trinitrate) t hat relaxes vascular smooth muscle and produces a generalized vasodilation. This action is mediated y nitric oxide, as shown in Figure 18.2 (24,25 and 26). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 26)03-May-05 17:23:29

m y be required to re ch ste dy-st te serum drug levels fter the st nuous infusion ther py. To prep re the infusion solution, 200 mg (40 ol is dded to 160 mL of diluent for fin l drug concentr tion of 1 recommended infusion r te is 2 mL/min, which corresponds to dose r min.

rt of conti mL) l bet l mg/mL. The te of 2 mg/

Ovid: ICU Book Figure 18.2. The iochemical mechanism for the vasodilator actions of nitroglyce rin (NTG) and nitroprusside (NTP). Chemical sym ols: nitroprusside (Fe-NO-CN5), organic nitrate (RONO2), nitrite (NO2), nitric oxide (NO), guanosine triphosphat e (GTP), cyclic guanosine monophosphate (cGMP). P.289 NITRIC OXIDE Nitroglycerin inds to the surface of endothelial cells and undergoes two chemic al reductions to form nitric oxide (NO). The nitric oxide then moves out of the endothelial cell and into an adjacent smooth muscle cell, where it promotes the formation of cyclic guanosine monophosphate (cGMP), which then promotes muscle r elaxation. Vasodilation is a prominent action of nitric oxide, which was known a s endothelium-derived relaxing factor efore its chemical identification (25,26) . ACTIONS Nitroglycerin has a dose-dependent vasodilator effect in arteries and veins and is active in the systemic and pulmonary circulations (27). When the drug is give n y continuous infusion, venous dilator effects are prominent at low dose rates (< 40 /min) and arterial dilator effects predominate at high dose rates (> 200 /m in). As low-dose infusions are titrated upward, the earliest response is a decre ase in cardiac filling pressures (i.e., central venous pressure and wedge pressu re) with little or no change in cardiac output. As the dose rate is increased fu rther, the cardiac output egins to rise as a result of progressive arterial vas odilation. Further increases in the dose rate will eventually produce a drop in lood pressure. The hemodynamic responses to intravenous nitroglycerin have a ra pid onset and short duration, which permits rapid dose titration. P.290 Antiplatelet Effects Nitrates can inhi it platelet aggregation via the mechanism proposed for the vas odilator actions (26). Because platelet throm i are elieved to play an importan t role in the pathogenesis of acute myocardial infarction, the antiplatelet acti ons of nitroglycerin have een proposed as the mechanism for the antianginal eff ects of the drug (26). This may explain why the antianginal efficacy of nitrogly cerin is not shared y other vasodilator agents. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 26)03-May-05 17:23:29

Ovid: ICU Book Indications Intravenous nitroglycerin can e used to decrease left-ventricular filling press ures (low dose), augment cardiac output (intermediate dose), or lower lood pres sure (high dose). It is also useful in relieving anginal chest pain (see Chapter 19). DRUG ADMINISTRATION A nitroglycerin dose chart is shown in Ta le 18.6. The infusion rates in this ch art are ased on a drug concentration of 400 /mL in the infusion solution. TABLE 18.6. NITROGLYCERIN DOSAGE CHART Sorption Nitroglycerin inds to soft plastics such as polyvinylchloride (PVC), which is a common constituent in plastic ags and infusion tu ing. As much as 80% of the d rug can e lost y sorption. Glass and hard plastics do not adsor nitroglycerin , so the pro lem of adsorption can e eliminated y using glass ottles and stif f polyethylene tu ing. P.291 Drug manufacturers often provide specialized infusi on sets to deliver nitroglycerin. (For a comprehensive description of nitroglyce rin adsorption, see the Hand ook on Injecta le Drugs, pp 777781.) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 26)03-May-05 17:23:29

Ovid: ICU Book Nitroglycerin infusions should egin at a rate of 5 /min. The dose rate is then i ncreased in 5- / min increments every 5 minutes until the desired effect is achiev ed. Although effective dose rates vary, the dose requirement should not exceed 4 00 /min in most patients. High dose requirements (e.g., > 350 /min) are often the result of drug loss via adsorption, or nitrate tolerance (see elow). ADVERSE EFFECTS Nitroglycerin can produce three types of adverse reactions. One is flow related, another is related to oxidant stress, and the final one is related to the way t he drug is administered. Flow-Related Effects Excessive flow in the cere ral and pulmonary circulations can create complicatio ns. Nitroglycerin seems adept at increasing cere ral lood flow (headache is a c ommon complaint), and this can increase intracranial pressure and produce sympto matic intracranial hypertension (28). Because of this effect, nitroglycerin is a voided in patients with increased intracranial pressure. Increases in pulmonary lood flow can ecome a pro lem when the augmented flow occurs in areas of the l ung that are poorly ventilated. This increases shunt fraction and can lead to hy poxemia. This effect can e prominent in the acute respiratory distress syndrome (29), where much of the lung is poorly ventilated.

Methemoglo inemia Nitroglycerin meta olism produces inorganic nitrites (see Fig. 18.2), and accumu lation of nitrites can result in the oxidation of heme- ound iron in hemoglo in, as shown elow. The oxidation of iron from the Fe(II) to Fe(III) state creates methemoglo in (me tH ). Oxidized iron does not carry oxygen effectively, and thus metH accumulati on can impair tissue oxygenation. Clinically significant methemoglo inemia is no t a common complication of nitroglycerin therapy and usually occurs only at very high dose rates (28). MetH accumulation has few specific manifestations other than the characteristic rown discoloration of lood (due to the rown color of metH ). MetH can e detected y light reflection (oximetry), a technique descri ed in Chapter 22. Pulse oximeters do not relia ly detect metH (30), and the me asurement should e performed y more sophisticated oximeters (called co-oximete rs) in the clinical la oratory. P.292 MetH levels a ove 3% (fraction of total h emoglo in) are a normal. Levels a ove 40% can http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 26)03-May-05 17:23:29

Ovid: ICU Book produce tissue ischemia, and levels a ove 70% are lethal (28). If there is no ev idence of tissue hypoxia, discontinuing nitroglycerin is all that is required. I f tissue oxygenation is impaired (e. g., lactic acidosis), metH can e chemical ly converted ack to normal hemoglo in with methylene lue (a reducing agent), 2 mg/kg IV over 10 minutes. Solvent Toxicity Nitroglycerin does not readily dissolve in aqueous solutions, and nonpolar solve nts such as ethanol and propylene glycol are required to keep the drug in soluti on. These solvents can accumulate during continuous infusion. Ethanol intoxicati on has een reported in association with nitroglycerin infusions (31). Manifesta tions include a change in mental status and gar led speech. Hypotension can also occur. A lood ethanol level will confirm the diagnosis. Propylene glycol toxic ity has also een reported (32), and ecause some commercial nitroglycerin prepa rations contain 30 to 50% propylene glycol (28), clinical toxicity may e more c ommon than suspected. Toxic manifestations include altered mental status that ca n progress to coma, meta olic acidosis, and hemolysis. The propylene glycol leve l in lood confirms the diagnosis. In patients who develop a change in mental st atus during prolonged or high-dose nitroglycerin infusions, the serum osmolal ga p (i.e., the difference etween measured and calculated serum osmolality) might e a valua le screening test for possi le solvent toxicity. The osmolal gap shou ld e elevated (> 10 mOsm/kg) y the presence of either solvent in the loodstre am. An elevated gap would then prompt more specific assays to identify the toxin . NITRATE TOLERANCE Tolerance to the vasodilator and antiplatelet actions of nitroglycerin is more c ommon than suspected and can appear after only 24 hours of continuous drug admin istration (27). No single mechanism seems to explain this phenomenon. One possi le cause is the depletion of reducing agents in the vascular endothelium, which impairs the conversion of nitroglycerin to nitric oxide. This mechanism is suppo rted y the association of nitrate tolerance with decreased nitric oxide product ion (33). However, the administration of reducing agents (e.g., Nacetylcysteine) does not consistently restore responsiveness in nitrate tolerance. At present, the most effective method for restoring nitroglycerin responsiveness is to disco ntinue drug administration for at least 6 to 8 hours each day (27). P.293 NITROPRUSSIDE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 26)03-May-05 17:23:29

Ovid: ICU Book Nitroprusside is a vasodilator agent that shares many features with nitroglyceri n. One of these is the participation of nitric oxide in the vasodilator actions of the drug. The nitroprusside molecule is shown in Figure 18.3. It contains one nitrosyl group (NO), and this is released in the loodstream as nitric oxide, w hich then can move into the vessel wall and move along the pathway shown in Ta l e 18.2. As a result of their shared mechanisms of action, nitroprusside and nitr oglycerin are classified as nitrovasodilators (26). Figure 18.3. The sodium nitroprusside molecule, and the chemical reaction used t o remove free cyanide from the loodstream. There is one feature that distinguishes nitroprusside from nitroglycerin: Nitrop russide is a dangerous drug and may e responsi le for over 1000 deaths yearly ( 34). Because the major actions of nitroprusside may e its toxic effects, the ad verse effects will e presented first. The description of nitroprusside toxicity in this chapter is rief; Chapter 53 covers this topic in detail. TOXICOLOGY The toxic nature of nitroprusside is due to its molecular composition, which is shown in Figure 18.3. The nitroprusside molecule contains five cyanide ions, and almost half of the molecular weight of the parent molecule is cyanide. When nit roprusside disrupts to release nitric oxide and exert its actions, the cyanide i s released into the loodstream. The chemical reaction shown in Figure 18.3 desc ri es how the free cyanide is removed. Sulfur from a donor source com ines with the free cyanide and forms thiocyanate (SCN), which is cleared y the kidneys. T he sulfur donor for this reaction is thiosulfate. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 26)03-May-05 17:23:29

Ovid: ICU Book P.294 Cyanide The capacity of the human ody to clear cyanide was grossly overestimated when n itroprusside was first introduced (28). The limiting factor is thiosulfate, whic h is stored in limited quantities and is easily depleted. The result is early an d frequent accumulation of cyanide during nitroprusside infusions. The consequen ces of this cyanide accumulation is descri ed in Chapter 53. Thiocyanate The clearance of thiocyanate y the kidneys is impaired when renal function or r enal lood flow is compromised. The accumulation of thiocyanate produces a toxic syndrome that is distinct from cyanide intoxication. Thus, nitroprusside toxici ty can e expressed as either cyanide or thiocyanate intoxication. Both of these toxic syndromes are descri ed in Chapter 53. ACTIONS Nitroprusside has een favored ecause the vascular responses are prompt and sho rt lived and this allows for rapid dose titration. Vasodilator effects are often evident at low dose rates (0.5 /kg/min), and the sequence of hemodynamic respons es is the same as descri ed for nitroglycerin. Blood pressure does not usually d ecline at dose rates elow 1 /kg/min. An immediate drop in lood pressure at low dose rates can e a sign of hypovolemia. Indications Because of the toxic potential of the drug, nitroprusside should e used only wh en there is no alternative availa le. The drug seems est suited for the managem ent of severe hypertension com ined with a low cardiac output. DRUG ADMINISTRATION A dose chart for nitroprusside is shown in Ta le 18.7. Note the recommendation t o add thiosulfate to the nitroprusside infusate. This provides the sulfur needed to detoxify cyanide and should e a mandatory practice. A proportional dose of 500 mg thiosulfate per 50 mg nitroprusside should e used (34). Thiosulfate is p rovided as sodium thiosulfate (290 mg sodium per gram thiosulfate) and is commer cially availa le as a 10% solution (5 mL = 500 mg thiosulfate). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 26)03-May-05 17:23:29

Ovid: ICU Book TABLE 18.7. NITROPRUSSIDE DOSAGE CHART The FDA recommends starting nitroprusside at a low dose rate (0.2 /kg/min) and ti trating the dose upward in 5-minute increments (35). The maximum allowa le dose rate is 10 /kg/min for no longer than 10 minutes. P.295 ADVERSE EFFECTS In addition to cyanide and thiocyanate intoxication, nitroprusside has adverse h emodynamic effects that are identical to those descri ed for nitroglycerin. Nitr oprusside increases intracranial pressure, and thus it is not advised for patien ts with intracranial hypertension. Because hypertensive encephalopathy is associ ated with a raised intracranial pressure, nitroprusside seems ill-advised for ma naging hypertensive encephalopathy. NOREPINEPHRINE Norepinephrine is an -receptor gonist th t promotes widespre d v soconstriction. As result of e rly reports of ren l f ilure from norepinephrine, combined wit h gener l decre se in enthusi sm for v soconstrictor drugs, norepinephrine is no longer considered first-line drug for the m n gement of circul tory shock. In c ses of hypotension refr ctory to dop mine, it c n be dded s second gen t. There is some renewed interest in norepinephrine bec use of reports showing t h t there is less v soconstriction nd even improved org n perfusion in response to norepinephrine in http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (20 of 26)03-M y-05 17:23:29

Ovid: ICU Book

ACTIONS Norepinephrine produces dose-dependent incre se in systemic v scul r resist nc e. Although the drug c n stimul te c rdi c -receptors P.296 over a wide range, th e cardiac output is increased only at low doses. Over the remainder of the thera peutic dose range, the inotropic response to norepinephrine is overshadowed y t he vasoconstrictor response. At high dose rates, the cardiac output decreases in response to the vasoconstriction and increased afterload. Indications In cases of septic shock where the desired vasoconstriction is not achieved y a dopamine infusion norepinephrine can e added as a second drug (36). DRUG ADMINISTRATION One milligram of norepinephrine is added to a diluent volume of 250 mL (4 /mL). T he infusion should e egun at 1 /min (15 microdrops/min) and titrated to the des ired effect. The usual dose rate is 2 to 4 /min, with a range of 1 to 12 /min. The effective dose of norepinephrine can vary widely, and in clinical reports invol ving patients with septic shock, the effective norepinephrine dose has varied fr om 0.7 to 210 /min (36). ADVERSE EFFECTS The administration of any vasoconstrictor agent carries a risk of hypoperfusion and ischemia involving any tissue ed or vital organ. For any condition that req uires vasoconstrictor drugs to maintain a lood pressure, it may e difficult to distinguish adverse drug effects and adverse disease effects. Furthermore, if a n adverse drug effect is identified or suspected, there may e little or no room for therapeutic manipulations. REFERENCES GENERAL TEXTS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 26)03-May-05 17:23:29

p tients with septic shock (36,37). However, it seems foolish to expect th t witch to norepinephrine will improve the clinic l outcome in septic shock.

Ovid: ICU Book 1. Trissel LA (ed). Hand ook on injecta le drugs. 8th ed. Bethesda: American Soc iety of Hospital Pharmacists, 1994. 2. Kahn MG. Cardiac drug therapy. 4th ed. Philadelphia: W.B. Saunders, Co., 1995 . REVIEWS 3. Zaloga GP, Prielipp RC, Butterworth JF, Royster RL. Pharmacologic cardiovascu lar support. Crit Care Clin 1993;9:335362 (192 references). 4. Trujillo MH, Arai K, Bellorin-Font E. Practical guide for drug administration y intravenous infusion in intensive care units. Crit Care Med 1994;22:10491063 (80 references). 5. Zaritsky AL. Catecholamines, inotropic medications, and vasopressor agents. I n Chenow B, (ed). The pharmacologic approach to the critically ill patient. 3rd ed., Baltimore: Williams & Wilkins, 1994 (224 references). P.297 AMRINONE 6. DiBianco B. Acute positive inotropic intervention: the phosphodiesterase inhi itors. Am Heart J 1991;121:18711875.

7. Levy JH, Bailey JM. Amrinone: Its effect on vascular resistance and capacitan ce in human su jects. Chest 1994;105:6264. 8. Marcus RH, Raw K, Patel J, et al. Comparison of intravenous amrinone and do u tamine in congestive heart failure due to idiopathic dilated cardiomyopathy. Am J Cardiol 1990;66:11071112. 9. Sundram P, Reddy HK, McElroy PA, et al. Myocardial energetics and efficiency in http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 26)03-May-05 17:23:29

Ovid: ICU Book patients with idiopathic cardiomyopathy: Response to do utamine and amrinone. Am J Cardiol 1990;119:891898. 10. Butterworth JF, Royster RL, Prielipp RC, et al. Amrinone in cardiac surgical patients with left-ventricular dysfunction. Chest 1993;104:16601667. 11. Notterman DA. Inotropic agents. In Blumer JL, Bond R, (eds). Toxic effects o f drugs used in the ICU. Crit Care Clin 1991;7:583614. 12. Scalea TM, Donovan R. Amrinone as an inotrope in managing hypermeta olic sur gical stress. J Trauma 1992;32:372379. DOBUTAMINE 13. Vincent J-L. Do utamine in the intensive care setting. In Chatterjee K (ed). Do utamine. A ten year review. New York: NCM Pu lishers, 1989:109121. 14. Klem C, Dasta JF, Reilley TE, Flanc aum LJ. Varia ility in do utamine pharma cokinetics in unsta le critically ill surgical patients. Crit Care Med 1994;22:1 926 1932. 15. Rich MW, Im urgia M. Inotropic response to do utamine in elderly patients wi th decompensated congestive heart failure. Am J Cardiol 1990;65:519521. 16. Hayes MA, Yau EHS, Timmins AC, et al. Response of critically ill patients to treatment aimed at achieving supranormal oxygen delivery and consumption. Relat ionship to outcome. Chest 1993;103:886895. DOPAMINE 17. Duke GJ, Briedis JH, Weaver RA. Renal support in critically ill patients: Lo w-dose dopamine or low-dose do utamine? Crit Care Med 1994;22:19191925. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (23 of 26)03-May-05 17:23:29

Ovid: ICU Book 18. Oung CM, English M, Chiu RCJ, Hinchey J. Effects of hypothermia on hemodynam ic responses to dopamine and do utamine. J Trauma 1992;33:671678. EPINEPHRINE 19. Hollingsworth HM, Giansiracusa DF, Upchurch KS. Anaphylaxis. J Intensive Car e Med 1991;6:5570. 20. Chiolero R, Flatt J-P, Revelly J-P, Jequier E. Effects of catecholamines on oxygen consumption and oxygen delivery in critically ill patients. Chest 1991;10 0:16761684. 21. Ronco JJ, Fenwick JC, Wiggs BR, et al. Oxygen consumption is independent of increases in oxygen delivery y do utamine in septic patients who have normal or increased plasma lactate. Am Rev Respir Dis 1993;147:2531. LABETALOL 22. Kaplan NM. Management of hypertensive emergencies. Lancet 1994;344:13351338. 23. DeVault GA Jr. Therapy in hypertensive emergencies: A disease-specific appro ach. J Crit Illness 1990;6:477484. P.298 24. Cosentino F, Vidt DG, Orlowski JP, et al. The safety of cumulative doses of la etalol in perioperative hypertension. Clev Clin J Med 1989;56:371376. NITROGLYCERIN 25. Anggard E. Nitric oxide: mediator, murderer, and medicine. Lancet 1994;343:1 199 1206. 26. Anderson TJ, Meredith IT, Ganz P, et al. Nitric oxide and nitrovasodilators: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (24 of 26)03-May-05 17:23:29

Ovid: ICU Book similarities, differences and potential interactions. J Am Coll Cardiol 1994;24: 555566. 27. Stamler JS, Loscalzo J. The antiplatelet effects of organic nitrates and rel ated nitroso compounds in vitro and in vivo and their relevance to cardiovascula r disorders. J Am Coll Cardiol 1991;18:15291536. 28. Elkayam U. Nitrates in heart failure. Cardiol Clin 1994;12:7385. 29. Radermacher P, Santak B, Becker H, Falke KJ. Prostaglandin F1 and nitroglyce rin reduce pulmonary capillary pressure ut worsen ventilationperfusion distri ut ion in patients with adult respiratory distress syndrome. Anesthesiology 1989;70 :601606. 30. Curry SC, Arnold-Cappell P. Nitroprusside, nitroglycerin, and angiotensin-co nverting enzyme inhi itors. In: Blumer JL, Bond GR, (eds). Toxic effects of drug s used in the ICU. Crit Care Clin 1991;7:555582. 31. Barker SJ, Kemper KK, Hyatt J. Effects of methemoglo inemia on pulse oximetr y and mixed venous oximetry. Anesthesiology 1989;70:112117. 32. Korn SH, Comer JB. Intravenous nitroglycerin and ethanol intoxication. Ann I ntern Med 1985;102:274. 33. Demey HE, Daelemans RA, Verpooten GA, et al. Propylene glycol-induced side e ffects during intravenous nitroglycerin therapy. Intensive Care Med 1988;14:22122 6. 34. Husain M, Adrie C, Ichinose F, et al. Exhaled nitric oxide as a marker for o rganic nitrate tolerance. Circulation 1994;89:24982502. NITROPRUSSIDE 35. Fenichel RR. A quick (high pressure) tour of nitroprusside (NTP). Washington , DC: Food and Drug Administration Center for Drug Evaluation and Research. DHHS , 1990 (April). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (25 of 26)03-May-05 17:23:29

Ovid: ICU Book 36. Ro in ED, McCauley R. Nitroprusside-related cyanide poisoning. Time (long pa st due) for urgent, effective interventions. Chest 1992;102:18421845. 37. Hall VA, Guest JM. Sodium nitroprusside-induced cyanide intoxication and pre vention with sodium thiosulfate prophylaxis. Am J Crit Care 1992;2:1927. NOREPINEPHRINE 38. Dasta JF. Norepinephrine in septic shock: Renewed interest in an old drug. D ICP, Ann Pharmacother 1990;24:153156. 39. Desairs P, Pinaud M, Bugnon D, Tasseau F. Norepinephrine therapy has no dele terious renal effects in human septic shock. Crit Care Med 1989;17:426429. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (26 of 26)03-May-05 17:23:29

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 19 EARLY MANAGEMENT OF ACUTE MYOCARDIAL INFARCTION The management strategies for acute myocardial infarction (MI) shifted focus in the early 1980s, following the discovery that acute (transmural) MI is the direc t result of an occlusive throm osis. The original focus of early management, whi ch included preventing malignant arrhythmias (with lidocaine) and maintaining a satisfactory alance etween myocardial oxygen supply and oxygen consumption, wa s su sequently shifted to the current emphasis on esta lishing and maintaining r eperfusion in the infarct-related artery (1,2). The elements of the reperfusion strategy for acute MI are presented in this chapter. When executed appropriately and promptly, this approach can result in fewer deaths in patients with acute M I. BACKGROUND CORONARY THROMBOSIS A causal relationship etween coronary artery throm osis and acute MI was sugges ted in the original description of the acute MI syndrome in 1912. However, autop sy studies in the first half of the century showed only sporadic evidence of thr om osis (presuma ly missing the ones that lysed spontaneously). The first defini tive report was an angiographic study pu lished in 1980, which showed that 87% o f patients with Q-wave infarctions had complete occlusion of the infarct-related arteries y an o structing throm us when studied in the first 4 hours after the onset of symptoms (4). The incidence of complete o struction was much lower (24 %) in patients with nonQ-wave infarctions (4). These results were confirmed in se veral studies over the next 5 years. Morphologic studies have identified rupture d atherosclerotic plaques as the nidus for coronary throm osis (5). The interior of the P.302 plaque first undergoes liquefaction necrosis, and the process erod es outward and ruptures into the loodstream. When lood comes in contact with t he throm ogenic lipids, clot formation http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 20)03-May-05 07:43:36

Ovid: ICU Book egins. The trigger for plaque disruption is not known, and the process is indep endent of the size or age of the plaque (3). The degree of throm otic o structio n determines the clinical syndrome. Persistent o struction lasting more than 24 hours results in Q-wave (transmural) infarctions, whereas transient o struction for 1 to 2 hours results in nonQ-wave (su endocardial) infarctions. Free-floating throm i attached y a stalk may e the source of unsta le angina. THROMBOLYTIC THERAPY The evaluation of drugs that stimulate fi rinolysis egan immediately after the 1980 discovery of occlusive throm osis in Q-wave infarctions. Intracoronary inst allation of fi rinolytic agents resulted in effective clot lysis in 80% of o str ucted arteries (6), and one agent (streptokinase) proved equally effective when given intravenously. In 1986, the first clinical trial of intravenous streptokin ase in acute MI was completed. Nicknamed GISSI, the study included almost 12,000 su jects, and the results showed fewer deaths in the patients who received thro m olytic therapy (7). Survival Benefit The survival enefit of throm olytic therapy is demonstrated in Ta le 19.1. Incl uded in this ta le are the pooled results of nine clinical trials comparing thro m olytic therapy (with different lytic agents) to place o in patients with acute MI (8). TABLE 19.1. SURVIVAL BENEFIT OF LYTIC THERAPY IN NINE CLINICAL TRIALS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 20)03-May-05 07:43:36

Ovid: ICU Book Lytic therapy given within 12 hours after the onset of chest pain is included in this analysis, and only Q-wave infarctions (identified y ST elevation on the E CG) are represented. Note that the total num er of patients studied is close to 60,000. The mortality (1 month after randomization) is lower in the patients giv en lytic therapy (9.6%) than P.303 in the patients given place o (11.5%). The a solute decrease in mortality rate is 1.9%, with a proportional decrease in morta lity of 18% (i.e., 18% of 11.7%). Because the proportional change is a percentag e of a percent, the 18% proportional reduction in mortality is equivalent to 1.8 % fewer deaths. This can e expressed as a survival enefit of 1.8 lives saved p er 100 patients treated (or 18 per 1000). Thus, for every 1000 patients who rece ived throm olytic therapy, 18 patients survived ecause of the therapy. Timing If the average survival enefit in Ta le 19.1 is analyzed according to the time when throm olytic therapy was initiated, the result is shown in Figure 19.1. Thi s graph shows the survival enefit of throm olytic therapy (in lives saved per 1 000) in relation to the time elapsed from the onset of chest pain to the initiat ion of therapy (7). The earlier the throm olytic therapy is initiated, the great er the survival enefit. After 12 hours following the onset of chest pain, the s urvival enefit of throm olytic therapy is lost. Over the 12-hour effective trea tment period, the decline in survival enefit averages 1.6 lives lost per 1000 p atients per hour of delay. In the first 6 hours, the hourly decline in survival (2.6 per 1000) is greater than in the final 6 hours (0.6 per 1000), so treatment delays are four times more costly in the first 6 hours than in the second 6 hou rs of the effective treatment period. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 20)03-May-05 07:43:36

Ovid: ICU Book Figure 19.1. The survival enefit (in lives saved per 1000 patients treated) of throm olytic therapy in relation to the time after the onset of chest pain that therapy is initiated. The height of each column represents the average survival enefit reported over the time period indicated on the horizontal axis. (Data fr om Fi rinolytic Therapy Trialists Colla orative Group [7].)

The time dependence shown in Figure 19.1 illustrates one of the P.304 most impor tant features of throm olytic therapy: Time lost is lives lost. This sets the st age for the description that follows of the initial evaluation of patients with suspected acute MI where the goal is to identify candidates for throm olytic the rapy and initiate therapy as quickly as possi le. THE INITIAL ENCOUNTER In the initial encounter with patients suspected of having an acute MI, there ar e four principal tasks: Relieve the chest pain. Identify and treat life-threaten ing conditions. Initiate throm olytic therapy in appropriate candidates. Complet e the first three tasks as quickly as possi le. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 20)03-May-05 07:43:36

Ovid: ICU Book These individual tasks are performed simultaneously, and in no particular order. RELIEVE CHEST PAIN If not used efore presentation, su lingual nitroglycerin (2 ta lets, 3 minutes apart) can e used as an initial attempt to relieve the chest pain. This is usua lly not successful in chest pain from acute MI (1,2). Persistent chest pain shou ld e treated with morphine instead of intravenous nitroglycerin. This latter in tervention is est reserved for patients with hypertension or left-heart failure (1,2). Aggressive use of nitroglycerin via either route is not recommended in r ight-ventricular infarction ecause of the risk of hypotension (2). Morphine is the drug of choice for relieving the chest discomfort in acute MI (1,2). The usu al dose is 4 mg y slow intravenous push (e.g., 1 mg/minute), which can e repea ted every 5 minutes if necessary. Morphine administration may e followed y a d ecrease in lood pressure. This is usually the result of a decrease in sympathet ic nervous system activity and is not a pathologic process. If the systolic pres sure falls to a level of concern ( elow 100 mm Hg), volume infusion is usually e ffective in raising the pressure. If hypotension persists and is associated with radycardia, atropine (0.5 to 1 mg intravenously) may e required (2). Do not u se pressor agents to reverse the decrease in lood pressure induced y morphine (2). LIFE-THREATENING CONDITIONS Masquerading Illness Do not forget to consider other serious conditions that can produce su sternal c hest pain, such as acute aortic dissection, acute pulmonary em olus, pericarditi s, and esophageal rupture. Aortic dissection is descri ed at the end of this cha pter. P.305 Complications A out one-quarter of patients with acute MI have evidence of heart failure on in itial presentation, and another 3 to 4% are in cardiogenic shock (6). Throm olyt ic therapy has no documented enefit in acute MI accompanied y heart failure or cardiogenic shock (10), so the presence of either complication excludes a patie nt from throm olytic therapy. The management of acute MI accompanied y cardiac pump failure is discussed later in this chapter. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 20)03-May-05 07:43:36

Ovid: ICU Book SELECTION CRITERIA FOR LYTIC THERAPY Patients who are candidates for throm olytic therapy must e identified as soon as possi le after presentation using the eligi ility criteria shown in Ta le 19. 2. Unfortunately, only 20% of patients with acute MI receive throm olytic therap y. An additional 15% are eligi le ut never receive the therapy. TABLE 19.2. SELECTION CRITERIA FOR THROMBOLYTIC THERAPY Indications Patients are candidates for lytic therapy if they have chest pain for at least 3 0 minutes ut less than 12 hours (as in Figure 19.1) and have a 12-lead ECG that shows ST elevation of 0.1 mm or more in two contiguous leads, or a new left un dle ranch lock. As mentioned P.306 earlier, patients with heart failure or car diogenic shock are not eligi le for throm olytic therapy. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 20)03-May-05 07:43:36

Ovid: ICU Book Contraindications The a solute contraindications for throm olytic therapy are listed in Ta le 19.2 . Relative contraindications are not included ecause the survival enefit from throm olytic therapy should justify the risks associated with relative contraind ications (3). Note that neither advanced age nor the presence of indwelling vasc ular catheters (including central venous pressure and pulmonary artery catheters ) are reasons to exclude a patient from throm olytic therapy. Although leeding complications are more common in the elderly, the survival enefit of throm olyt ic therapy is also greater in the elderly (3). Therefore, the enhanced survival cancels the negative effect of enhanced leeding in the elderly population. Time to Therapy Any patient who has all the indications for throm olytic therapy, and no contrai ndications, should receive lytic therapy as soon as possi le. According to the A merican Heart Association, the time from presentation to initiation of throm oly tic therapy should not exceed 1 hour (2). Longer delays in candidates for throm olytic therapy will result in unnecessary deaths. THROMBOLYTIC THERAPY The throm olytic agents and dosing regimens for patients with suspected acute MI are included in Ta le 19.3. In terms of overall clinical P.307 performance, no one throm olytic agent has proven consistently superior to the others. Therefore , the important issue in throm olytic therapy is not which agent to use, ut how quickly to use it. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 20)03-May-05 07:43:36

Ovid: ICU Book TABLE 19.3. THROMBOLYTIC THERAPY FOR SUSPECTED ACUTE MI MECHANISM OF ACTION All throm olytic agents promote fi rinolysis y converting plasminogen to plasmi n, which reaks fi rin strands into smaller su units. Streptokinase (SK) is a a cterial protein (from streptococci) that acts on plasminogen in the circulating lood and produces a disseminated or systemic lytic state. Tissue plasminogen ac tivator (TPA) is the molecular clone of an endogenous fi rinolytic su stance wit h the same name. This agent inds to fi rin in lood clots and converts plasmino gen to plasmin locally, in the vicinity of the throm us. This produces a clot-sp ecific activation of fi rinolysis, with less of a systemic lytic effect. Anistre plase, also known as anisolated plasminogenstreptokinase activator complex (APSAC ), is a clot-specific version of streptokinase. However, in the dosages needed f or clot lysis, this agent also produces a systemic lytic effect. Lytic Effect The most rapid lysis is produced y TPA (60 to 70% lysis at 90 minutes), and SK and APSAC are tied for second place (50% lysis at 90 minutes) (9). However, the final lytic effect is the same with all agents (80% lysis at 180 minutes). BENEFITS AND RISKS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 20)03-May-05 07:43:36

Ovid: ICU Book Survival Benefit All throm olytic agents improve survival when compared with place o (Ta le 19.1) . There are three clinical trials comparing survival with different lytic agents . Two trials (GISSI-2 and ISIS3) show no difference in survival with any of the agents (11,12). The third trial (GUSTO) shows a 1% (a solute) reduction in morta lity with TPA given in an accelerated regimen (the same regimen as in Ta le 19.3 ) compared to SK (13). Although TPA is eing marketed as the est lytic agent a sed on this 1% difference, it is difficult to take the sum of all megatrials of lytic agents over the last 10 years and identify the superior agent ased on a 1 % margin in one trial. Complications The most feared complication of lytic therapy is intracere ral hemorrhage, which is reported in 0.5 to 1% of cases (8,9). This complication occurs more often wi th TPA than with the other agents, as documented in three studies (GISSI-2, ISIS -3, and GUSTO-1) (9). Extracranial leeding P.308 that requires lood transfusio ns occurs in 5 to 15% of patients, regardless of the agent used (9). The risk of leeding is not related to the extent (systemic versus clot-specific) of the fi rinolysis produced y each agent. Complications other than leeding are common with SK, which y virtue of eing a acterial product can act as an antigen (APS AC contains SK, and thus has similar antigenic actions). Common antigen-related reactions are fever (20 to 40%), allergic-type reactions (5%), and the productio n of neutralizing anti odies (14). The anti ody response to SK can persist for u p to 8 months (15), so SK should not e readministered for at least 6 months. Th e anti ody response to streptococcal infections (e.g., pharyngitis) can also neu tralize SK, so SK should not e used for 6 months after a documented streptococc al infection. Hypotension of unclear etiology can occur with any of the throm ol ytic agents ut is more common with SK (14). The hypotension usually resolves wh en the infusion is stopped, although volume infusion is sometimes necessary. The infusion can e resumed, ut the infusion rate should e reduced y 50%. Cost A cost comparison for each of the lytic agents is shown at the ottom of Ta le 1 9.3. Compared to SK, the additional cost to the patient is $1800 for APSAC and $ 2200 for TPA. Considering that no individual throm olytic agent has proven consi stently superior to the others, http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 20)03-May-05 07:43:36

Ovid: ICU Book streptokinase may e the preferred agent y virtue of its reduced cost. PREVENTING REOCCLUSION Initial clot lysis is followed y reocclusion in 15 to 30% of infarct-related ar teries. The rate of reocclusion is particularly high with TPA (30%) (8). Agents that inhi it throm osis (aspirin and heparin) are often added to lytic therapy t o help reduce the risk of reocclusion. Aspirin is used to inhi it platelet aggre gation. In one study, aspirin added to the survival enefit of throm olytic ther apy, and also showed a survival enefit independent of lytic therapy (16). Dosag e: 160 mg (chewa le ta lets) at the time of lytic therapy, then 160 to 325 mg da ily. Heparin is routinely added to lytic therapy with TPA ecause of the high re occlusion rate with TPA. This regimen should e started in the final hour of the TPA infusion. Dosage: 5000 U as intravenous olus, then 1000 U/hour. Check part ial throm oplastin time (PTT) in 6 hours (target PTT is 60 to 85 seconds). The h eparin infusion is continued for 24 to 72 hours (1). Heparin does not add to the enefit of aspirin with the other two lytic agents (SK and APSAC), so heparin i s not indicated when these lytic agents are used (17). P.309 REPERFUSION INJURY Clot lysis and reperfusion can itself e a source of injury. One manifestation o f reperfusion injury is a form of diastolic dysfunction known as stunned myocard ium (18). Oxidant injury has een implicated in this process, and peroxidation o f mem rane lipids (descri ed in Chapter 3) may e involved. Damage to smooth mus cle mem ranes would lead to calcium influx, and the resulting intracellular calc ium accumulation could retard muscle relaxation and produce the myocardial stunn ing. Although antioxidants have not een effective in preventing reperfusion inj ury, magnesium may e protective. THERAPEUTIC OPTIONS The following therapy can e used with or without throm olytic therapy. Each age nt or class of agents included here has shown a survival enefit in clinical tri als, or is recommended y the American Heart Association as a useful interventio n (1,2). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 20)03-May-05 07:43:36

Ovid: ICU Book -BLOCKERS When given within a few hours of onset of acute MI, - lockers have een shown to reduce infarct size, decrease the incidence of arrhythmias, and improve survival (3,19,20 and 21). The com ined trials of - locker therapy in acute MI (which inc lude a total of 27,000 patients) show an average survival enefit of 13 per 1000 (3). When to Use -Blockers should e used for all patients with acute MI who have no specific cont raindication to their use. They are particularly useful for patients who develop a hyperdynamic state with tachycardia and hypertension (e.g., younger patients with anterior MI). Contraindications include heart rate elow 50 eats/minute, A V lock (first-, second-, or third-degree), systolic lood pressure elow 100 mm Hg, severe (systolic) heart failure, and reactive airway disease. Dosing Regimens The drugs and dosing schedules in Ta le 19.4 have proven successful in clinical trials of lockers in acute MI (19,20 and 21). Optimal results are o tained when therapy egins within 4 hours after the onset of chest pain. TABLE 19.4. -BLOCKERS IN ACUTE MI P.310 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 20)03-May-05 07:43:36

Ovid: ICU Book ANTITHROMBOTIC AGENTS Aspirin has een shown to improve survival in acute MI independent of throm olyt ic therapy (16). Although the evidence of a survival enefit is from a single st udy, aspirin is universally recommended for patients with suspected acute MI. Th e dosage is similar to that descri ed for throm olytic therapy: 160 mg as early as possi le after onset of chest pain, followed y 160 to 325 mg daily. Heparin therapy has een (consistently) inconsistent in improving outcomes in acute MI. However, the American Heart Association continues to recommend full heparinizati on as a routine measure for all cases of acute transmural MI, presuma ly to decr ease the risk of mural throm osis (1,2). For dosing information on heparin, see the weight- ased heparin dosing nomogram in Chapter 7 (Ta le 7.4). NITROGLYCERIN Intravenous nitroglycerin is effective in relieving anginal chest pain, ut the enefits in acute MI have een varia le. Clinical trials performed efore 1990 ( 3000 patients) show improved survival, whereas trials performed after 1990 (79,0 00 patients) show no survival enefit (3).

When to Use Intravenous nitroglycerin is not recommended for routine use ecause of the risk of hypotension and lack of survival enefit. It is est used to relieve chest p ain or manage hypertension. Nitroglycerin impairs the lytic actions of TPA (22), so com ined therapy should e avoided. P.311 Dosing Regimen Chapter 18 has a dose chart for intravenous nitroglycerin (Ta le 18.6). After 24 hours of continuous infusion, switch to a nitroglycerin patch that delivers 10 mg/day (23). The patch should e removed for 8 to 10 hours during each 24-hour p eriod to prevent nitrate tolerance. MAGNESIUM Magnesium has several actions that should e salutary in acute MI: It dilates co ronary arteries, inhi its platelet aggregation, suppresses arrhythmias, and can protect against reperfusion injury y locking the intracellular accumulation of calcium (24,25). In addition, magnesium depletion is common in the setting of a cute MI (24), pro a ly as a result of diuretic therapy http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 20)03-May-05 07:43:36

Ovid: ICU Book (which enhances renal magnesium excretion). Despite its potential enefit, intra venous magnesium has produced conflicting results in clinical trials. In one tri al where intravenous magnesium was started within 4 hours after onset of chest p ain, there was a survival enefit of 23 per 1000 (26). However, in a larger tria l where magnesium was started 8 to 12 hours after onset of chest pain, there was no survival enefit (3). Because lytic therapy was used in oth trials, one sug gestion is that the earlier use of magnesium was eneficial ecause it protected against reperfusion injury from the lytic therapy. For now, the discrepancy is not explained. When to Use Intravenous magnesium should e used in acute MI associated with hypokalemia, hy pocalcemia, and arrhythmias, and in patients receiving daily diuretics without m agnesium supplements. Because magnesium depletion is common in MI patients (even with normal serum magnesium levels), routine magnesium infusion seems warranted . Contraindications to intravenous magnesium include renal insufficiency, hypote nsion, severe heart failure, and atrioventricular lock. Dosing Regimen The dose used in the clinical trials is 2 g MgSO4 intravenously over 5 minutes, then 8 g intravenously y continuous infusion over 24 hours. This dose is design ed to dou le the serum magnesium level. However, in the presence of normal renal function, the extra magnesium is cleared quickly and the hypermagnesemia is tra nsient. A popular dosing regimen is 2 g MgSO4 in 100 mL saline infused over 30 t o 60 minutes, which is given routinely on admission to the ICU. P.312 EARLY COMPLICATIONS ARRHYTHMIAS The American Heart Association no longer recommends routine prophylaxis for vent ricular tachycardia or ventricular fi rillation, or treatment of asymptomatic war ning arrhythmias (such as a run of six premature ventricular contractions) (1,2). Ventricular ectopic depolarizations are treated only if they are associated wit h heart failure or hypotension. The recognition and management of arrhythmias ar e covered in Chapter 20. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 20)03-May-05 07:43:36

Ovid: ICU Book PUMP FAILURE Survival in acute MI declines as ventricular function declines. This is illustra ted in the first lytic therapy trial, where mortality rates were 7% for uncompli cated MI, 16% for heart failure, 39% for pulmonary edema, and 70% for cardiogeni c shock (6). The management of progressive heart failure and cardiogenic shock i n acute MI involves the same measures descri ed in Chapter 16. Two features of t he management in acute MI deserve mention. Angioplasty Patients with pump failure and ECG evidence of Q-wave infarction (ST elevations) are candidates for angioplasty. A review of 14 reports of angioplasty in cardio genic shock shows a decrease in mortality from 80 to 44% (10). Angioplasty is th erefore appealing in this setting, ut unfortunately is not availa le on an emer gency asis in many hospitals. Limiting Cardiac Work When cardiac pump failure is accompanied y ischemia, the goal of cardiac suppor t is to augment cardiac output without increasing myocardial oxygen consumption. The influence of cardiac support interventions on myocardial oxygen consumption (VO2) can e estimated using the four determinants of myocardial work: preload, contractility, afterload, and heart rate (these four varia les also determine m yocardial VO2). Ta le 19.5 shows the effects of different cardiac support interv entions on the four determinants of myocardial VO2, along with the net effect on myocardial VO2. Vasodilator therapy is the most eneficial (i.e., it produces t he greatest decrease in myocardial VO2), whereas dopamine creates the most harm (i.e., produces the greatest increase in myocardial VO2). Vasodilators are prefe rred to do utamine for treating heart failure, and the intraaortic alloon pump is preferred to dopamine for managing cardiogenic shock. The intraaortic alloon pump is also preferred to do utamine for managing heart failure accompanied P.3 13 y a orderline or la ile lood pressure (see Chapter 16 for more information on the intraaortic alloon pump). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 20)03-May-05 07:43:36

Ovid: ICU Book TABLE 19.5. CARDIAC SUPPORT AND MYOCARDIAL OXYGEN CONSUMPTION ACUTE AORTIC DISSECTION Acute aortic dissection involving the ascending aorta is a surgical emergency th at produces su sternal chest pain that can e confused with pain of cardiac orig in. CLINICAL PRESENTATION Aortic dissection occurs in all age groups ut is three times more common in men . A out 75% of su jects have a history of hypertension. Some have connective tis sue disorders (e.g., Marfan syndrome). Chest Pain Proximal dissection (ascending aorta) typically produces su sternal chest pain t hat can radiate to the neck, jaw, or shoulders. The character of the pain tends to e very sharp, with descriptive terms such as tearing or ripping that also de scri e the pathologic process in the aorta. The chest pain tends to remain sever e while it is present, ut it can also su side spontaneously for hours to days. The disease is not silent during this period, and when the pain returns, it is m ore severe. Clinical Findings The hallmark of this disorder is aortic insufficiency, which is present in over 50% of cases of http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 20)03-May-05 07:43:36

Ovid: ICU Book proximal dissection. Proximal dissection can extend into proximal lim vessels ( uneven pulses) and coronary arteries (ischemia/infarction) and can rupture into the pericardium (tamponade). P.314 Diagnosis The chest x ray is usually a normal, ut the diagnosis requires one of several i maging modalities. In order of sensitivity, these include the following (28): (a ) magnetic resonance imaging (MRI) (sensitivity and specificity 98%), ( ) transe sophageal echocardiography (sensitivity 98%, specificity 77%), (c) contrast-enha nced computed tomography (sensitivity 94%, specificity 87%), and (d) aortography (sensitivity 88%, specificity 94%). MRI and transesophageal ultrasound are the most sensitive diagnostic modalities availa le (28). Aortography is the least se nsitive, ut provides valua le information for the operating surgeon. MANAGEMENT The appropriate treatment for proximal aortic dissection is immediate surgery (2 7). Medical management is used to control pain and manage hypertension during th e diagnostic evaluation. (Uncomplicated dissections in the descending aorta can often e managed medically.) Because increases in flow rate produce higher shear forces and promote dissection, the goal of antihypertensive therapy in aortic d issection is to decrease the lood pressure without increasing the cardiac outpu t. This can e accomplished with the antihypertensive regimens listed in Ta le 1 9.6 (29). The traditional regimen for aortic dissection involves a com ination o f vasodilator therapy with nitroprusside (to decrease lood pressure) and intrav enous administration of - lockers (to prevent the increase in cardiac output that accompanies vasodilator therapy). -Blockade must e esta lished first, efore ni troprusside administration. Single-agent therapy is easier to use, and may e pr eferred (29). Continuous infusion of the ganglionic locker trimethaphan camsyla te (Arfonad) produces generalized vasodilation while locking any reflex increas e in cardiac output. Effective monotherapy is also possi le with la etalol, a co m ined - nd -receptor locker that can e given intravenously as intermittent ol us injections or y continuous infusion. Ganglionic lockade with P.315 trimetha phan can e associated with trou lesome side effects (such as urinary retention, ileus, constipation, and tachyphylaxis), so la etalol may e the preferred choi ce for monotherapy (except in patients with asthma or atrioventricular lock, wh ere la etalol is contraindicated). See Chapter 18 for more information on la eta lol. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 20)03-May-05 07:43:36

Ovid: ICU Book TABLE 19.6. TREATING HYPERTENSION IN ACUTE AORTIC DISSECTION REFERENCES REVIEWS 1. Guidelines for the early management of patients with acute myocardial infarct ion: a report of the American College of Cardiology/American Heart Association T ask Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures. J Am Coll Cardiol 1990;16:249292. 2. Cummins RO, ed. Text ook of advanced cardiac life support. Dallas: American H eart Association, 1994:9.19.16 (171 References)6. 3. Rogers WJ. Contemporary management of acute myocardial infarction. Am J Med 1 995;99:195206 (62 References). CORONARY THROMBOSIS 4. DeWood MA, Spores J, Notske R, et al. The prevalence of total coronary occlus ion during the early hours of transmural myocardial infarction. N Engl J Med 198 0;303:897 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 20)03-May-05 07:43:36

Ovid: ICU Book 902. 5. Davies MJ, Thomas AC. Plaque fissuring: the cause of acute myocardial infarct ion. Br Heart J 1985;53:363373. THROMBOLYTIC THERAPY 6. Anderson HV, Willerson JT. Throm olysis in acute myocardial infarction. N Eng l J Med 1993;329:703725. 7. Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico (G ISSI). Effectiveness of intravenous throm olytic treatment in acute myocardial i nfarction. Lancet 1986;1:397401. 8. Fi rinolytic Therapy Trialists Colla orative Group. Indications for fi rinoly tic therapy in suspected acute myocardial infarction: colla orative overview of early mortality and major mor idity results from all randomized trials of more t han 1000 patients. Lancet 1994;343:311322. 9. Young GP, Hoffman JR. Throm olytic therapy. Emerg Med Clin 1995;13:735759. 10. Bates ER, Topol EJ. Limitations of throm olytic therapy for acute myocardial infarction complicated y congestive heart failure and cardiogenic shock. J Am Coll Cardiol 1991;18:10771084. 11. ISIS-3 (Third International Study of Infarct Survival) Colla orative Group. ISIS-3: a randomized comparison of streptokinase vs. tissue plasminogen activato r vs. anistreplase and of aspirin plus heparin vs. aspirin alone among 41,299 ca ses of suspected acute myocardial infarction. Lancet 1992;339:753770. 12. Reeder GS, Kopecky SL. Throm olysis in acute myocardial infarction: t-PA for everyone? Mayo Clin Proc 1994;69:796799. P.316 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 20)03-May-05 07:43:36

Ovid: ICU Book 13. GUSTO Investigators. An international randomized trial comparing four throm olytic strategies for acute myocardial infarction. N Engl J Med 1993;329:673682. 14. Guidry JR, Raschke R, Morkunas AR. Anticoagulants and throm olytics. In: Blu mer JL, Bond GR, eds. Toxic effects of drugs in the ICU. Critical care clinics. Vol. 7. Philadelphia: WB Saunders, 1991;533554. 15. Jalihal S, Morris GK. Antistreptokinase titres after intravenous streptokina se. Lancet 1990;335:184185. 16. ISIS-2 (Second International Study of Infarct Survival) Colla orative Group. Randomized trial of intravenous streptokinase, oral aspirin, oth, or neither a mong 17,187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988; 2:349360. 17. Fuster V. Coronary throm olysis: a perspective for the practicing physician. N Engl J Med 1993;329:723725. 18. Reeder GS. Acute myocardial infarction: enhancing the results of reperfusion injury. Mayo Clin Proc 1995;70:11851190. BETA BLOCKERS 19. ISIS-1 (First International Study of Infarct Survival) Colla orative Group. Randomized trial of intravenous atenolol among 16,027 cases of or suspected case s of acute myocardial infarction. Lancet 1986;2:5766. 20. The MIAMI Trial Research Group. Metoprolol in acute myocardial infarction (M IAMI): a randomized, place o-controlled international trial. Eur Heart J 1985;6: 199226. 21. The International Colla orative Study Group. Reduction of infarct size with the early use of timolol in acute myocardial infarction. N Engl J Med 1984;310:91 5. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 20)03-May-05 07:43:36

Ovid: ICU Book NITROGLYCERIN 22. Nicolini FA, Ferrini D, Ottani F, et al. Concurrent nitroglycerin therapy im pairs tissuetype plasminogen activatorinduced throm olysis in patients with acute myocardial infarction. Am J Cardiol 1994;74:662666. 23. Gruppo Italiano per lo Studio della Sopravivenza nell'Infarto Miocardico. GI SSI-3: effects of lisinopril and transdermal glyceryl trinitrate singly and toge ther on 6-week mortality and ventricular function after acute myocardial infarct ion. Lancet 1994;343:11151122. 24. Heesch CM, Eichhorn EJ. Magnesium in acute myocardial infarction. Ann Emerg Med 1994;24:11541160. 25. du Toit EF, Opie L. Modulation of severity of reperfusion stunning in the is olated rat heart y agents altering calcium influx at onset of reperfusion. Circ Res 1992;70:960967. 26. Woods KL, Fletcher S, Roffe C, Haider Y. Intravenous magnesium sulfate in su spected acute myocardial infarction: results of the second Leicester Intravenous Magnesium Intervention Trial (LIMIT-2). Lancet 1992;339:15531558. 27. Crawford ES. The diagnosis and management of aortic dissection. JAMA 1990;26 4:25372541. 28. Zegel HG, Chmielewski S, Freiman DB. The imaging evaluation of thoracic aort ic dissection. Appl Radiol 1995;(June):1525. 29. Asfoura JY, Vidt DG. Acute aortic dissection. Chest 1991;99:724729. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 20)03-May-05 07:43:36

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 20 TACHYARRHYTHMIAS Acute arrhythmias are the gremlins of the ICU ecause they pop up unexpectedly, create some havoc, and are often gone again in a flash. This chapter descri es t he acute management of the most common and trou lesome rhythm distur ances in th e ICU: the tachyarrhythmias. CLASSIFICATION Tachycardias (heart rate a ove 100 eats/minute) can e the result of increased automaticity in pacemaker cells (e.g., sinus tachycardias), triggered activity ( e.g., ectopic impulses), or a process known as re-entry, where a triggered impul se encounters a pathway that locks propagation in the forward direction ut all ows the impulse to pass in the return (retrograde) direction (1,2). Such retrogr ade transmission allows a triggered impulse to propagate continually, creating a self-sustaining tachycardia. Re-entry is the most common cause of tachycardias that are clinically significant. Tachycardias are classified according to the si te of impulse generation in relation to the atrioventricular (AV) conduction sys tem (1,2 and 3). Supraventricular tachycardias (SVTs) originate a ove the AV con duction system and have a normal QRS duration. Ventricular tachycardias (VTs) or iginate elow the AV conduction system and have a prolonged QRS duration (more t han 0.12 second). As shown in Figure 20.1, oth types of tachycardias can e fur ther su divided y their regularity (i.e., y the regularity of the RR interval o n the ECG). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 26)03-May-05 07:46:02

Ovid: ICU Book Figure 20.1. Classification of the tachycardias ased on the QRS duration and th e regularity of the RR intervals. CLUES FROM THE RHYTHM STRIP The initial encounter with arrhythmias in the ICU is often the moment the nurse notices an irregularity on a single-lead rhythm strip. P.318 If the rhythm strip shows a tachycardia, the major arrhythmias of concern are as follows. Narrow-Complex Tachycardia If the QRS duration is 0.12 seconds or less, the arrhythmias to consider include sinus tachycardia, atrial tachycardia, AV nodal tachycardia, atrial flutter, an d atrial fi rillation. The following clues may help identify the pro lem. Rhythm and Rate http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 26)03-May-05 07:46:02

Ovid: ICU Book

AV nodal re-entrant tachycardia Possi le atrial flutter with 2:1 AV lock Atrial fi rillation or multifocal atrial tachycardia Atrial Activity Uniform P waves, fixed PR interval Multiform P waves, varia le PR interval Inverte d P waves Sawtooth waves No atrial activity Sinus tachycardia Multifocal atrial tachycardia AV nodal tachycardia Atrial flutter AV nodal re-entrant tachycardia P.319 Wide-Complex Tachycardia A tachycardia with a prolonged QRS duration is either a VT or an SVT with prolon ged AV conduction. The following signs help to differentiate etween these two a rrhythmias. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 26)03-May-05 07:46:02

Regular rhythm, rate > 150 regular

pm Regular rhythm, rate = 150 pm Rhythm markedly ir

Ovid: ICU Book

SVT VT VT When all RR intervals are irregular (irregularly irregular rhythm), the pro lem i s either atrial fi rillation or multifocal atrial tachycardia. The presence of f usion eats or AV dissociation efore the onset of the tachycardia is evidence o f VT. A fusion eat is caused y a ventricular ectopic impulse that merges (fuse s) with a normal QRS complex. This results in a QRS complex that is intermediate in morphology etween the normal complex and the ectopic impulse. A fusion eat is thus evidence of ventricular ectopic activity. It may e difficult to distin guish VT from an SVT with prolonged AV conduction, as illustrated in Figure 20.2 . In this case, the tracing in the upper panel shows a wide complex tachycardia that looks like VT. However, the tracing in the lower panel reveals that when th e arrhythmia converts to a normal sinus rhythm, the QRS duration remains prolong ed. This indicates that the tachycardia is an SVT superimposed on a undle ranc h lock. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 26)03-May-05 07:46:02

Pronounced irregularity Fusion

eats AV dissociation

Ovid: ICU Book Figure 20.2. An SVT with prolonged AV conduction masquerading as VT. (Tracings c ourtesy of Dr. Richard M. Green erg, M.D.) SINUS TACHYCARDIA Increased automaticity in the pacemaker cells of the sinoatrial node produces a regular, narrow-complex tachycardia with a gradual onset P.320 and rate of 100 t o 140 eats/minute. The ECG shows uniform P waves and a fixed PR interval. Sinus tachycardia can also e the result of re-entry into the sinus node. This variant sinus tachycardia has an a rupt onset, ut is otherwise indistinguisha le from the increased automaticity type of sinus tachycardia (2). MANAGEMENT Sinus tachycardia is usually a response to a systemic or noncardiac condition an d is often an adaptive response (as in exercise). It is usually well tolerated ( cardiac filling is usually not compromised until the heart rate rises a ove 180 eats/minute) (4) and does not require primary treatment. The primary goal of ma nagement is to identify and treat the associated condition. Potential sources of sinus tachycardia in the ICU are hypoxemia, sepsis, hypovolemia, and adrenergic drugs. The main indication for slowing a sinus tachycardia is the presence of m yocardial ischemia or infarction. In this situation, -receptor antagonists can e used to slow the heart rate. Because

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 26)03-May-05 07:46:02

Ovid: ICU Book these agents also depress ventricular function, they are not recommended for sin us tachycardias associated with systolic heart failure. The use of - lockers in m yocardial ischemia and infarction is descri ed in Chapter 19 (see Ta le 19.4). ATRIAL FLUTTER AND FIBRILLATION With the possi le exception of sinus tachycardia, the most common tachycardia in ICU patients is atrial fi rillation. Atrial flutter is an uncommon and transien t arrhythmia that often turns into atrial fi rillation. Atrial flutter and fi ri llation are oth a reviated as AF in this chapter. PREVALENCE There are an estimated 2.2 million adults in the United States with AF (5). Most are elderly (median age 75 years) and have either valvular heart disease, coron ary artery disease, or dilated cardiomyopathy. Contrary to popular perception, f ew have hyperactive thyroid disease (6). Additional predisposing factors in the ICU include recent cardiac surgery (particularly valvular surgery), resectional lung surgery, and acute myocardial infarction. POSTOPERATIVE AF A out 30% of patients undergoing cardiopulmonary ypass surgery develop AF in th e second to fourth postoperative day (7). Patients treated with - locker drugs e fore surgery, and those undergoing valvular surgery, are at highest risk. The et iology is unclear, ut the P.321 delayed onset suggests electrolyte (magnesium, potassium) shifts. These arrhythmias are more of a nuisance than a life-threaten ing complication, ut they often prolong the hospital stay y a few days. -Blocke rs are popular agents for the prevention and treatment of postoperative AF. ADVERSE CONSEQUENCES The loss of atrial contraction in AF, plus the decrease in diastolic filling tim e that accompanies the tachycardia, oth impair cardiac filling. Contraction of the atria is responsi le for 25% of the ventricular end-diastolic volume (preloa d) in the normal heart (4), and this contri ution is lost when the ventricle ec omes noncompliant and resists diastolic filling. The loss of the atrial componen t of cardiac filling is well tolerated in the normal heart, ut it can significa ntly impair cardiac performance in the noncompliant (e.g., hypertrophied) heart. The other nota le complication of AF is mural throm osis in the left atrium and cere ral http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 26)03-May-05 07:46:02

Ovid: ICU Book em olic stroke. Atrial throm osis can e demonstrated in 15% of patients who hav e AF for longer than 3 days (5), and 4 to 5% of patients with chronic AF suffer an em olic stroke each year without warfarin anticoagulation (5). Although it is possi le for a mural throm us to appear within 3 days after the onset of AF (5) , the throm oem olism associated with AF is usually a su acute or chronic compli cation, and is not discussed here. (For an informative description of throm oem olism in AF, see Reference 5.) MONITORING AND MANAGEMENT GOALS The acute management of AF is usually aimed at decreasing the ventricular rate r ather than conversion to a normal sinus rhythm. In other words, the management i s designed for rate control rather than rhythm control. The goals of rate contro l should include all of the following: Normal stroke volume A sence of a pulse d eficit Ventricular rate elow 100 eats/minute Because of the direct relationshi p etween preload and ventricular stroke output (see Fig. 1.2), the stroke volum e is the single est parameter to follow for determining the impact of AF on car diac performance. If invasive hemodynamic monitoring is not availa le, then the next est parameter to monitor is the pulse deficit (i.e., the difference etwee n the precordial heart rate and the pulse rate in a peripheral artery). The pres ence of a pulse deficit in AF indicates that cardiac stroke output is significan tly impaired, so disappearance of a pulse deficit should e a goal of rate-contr ol management in AF. The most commonly monitored parameter is the heart rate, wh ich should e kept within the normal range (less than 100 eats/minute). P.322 ELECTRICAL CARDIOVERSION Recommendations for direct-current (DC) cardioversion, taken from the most recen t advanced cardiac life support (ACLS) guidelines, are summarized in Ta le 20.1 (1). Electrical cardioversion should e reserved for cases where lood flow to t he major organs is impaired (such as hypotension or a change in mental status). Although shocks are normally synchronized to the R wave of the QRS complex (to p revent shocks from eing delivered during the vulnera le period of ventricular r epolarization) synchronization has no proven enefit, and can delay the delivery of the DC countershocks (1). Thus, unsynchronized shocks are not only accepta l e, they may e preferred when immediate cardioversion is desira le. The strength of DC shocks is expressed in joules (J), which is a unit of energy (see Appendi x 1). The recommended strength of the shocks for different arrhythmias is shown in Ta le 20.1. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 26)03-May-05 07:46:02

Ovid: ICU Book The energy of the initial shock should e 100 J for atrial fi rillation and 50 J for atrial flutter (flutter is easier to convert). If these initial shocks are not successful, then successive shocks of increasing strength are delivered, if necessary, using the energy sequences shown in Ta le 20.1. At these energy level s, DC cardioversion should e successful in 90% of patients with recent-onset AF . TABLE 20.1. DIRECT-CURRENT CARDIOVERSION ALTERNATIVE STRATEGY DC cardioversion can e a very distressing and painful experience for patients w ho are awake and aware. In the patient who is hypotensive ut awake, lood flow is adequate for at least one vital organ, so immediate cardioversion is not nece ssary. In this situation, volume P.323 infusion can e used in an attempt to rai se the lood pressure and o viate electrical cardioversion. As long as the centr al venous and wedge pressures are elow 20 mm Hg, infuse 100-mL aliquots of 5% a l umin or 6% hetastarch until the systolic lood pressure rises a ove 100 mm Hg. If this is successful, proceed with the conventional (pharmacologic) management of AF. Colloid fluids are preferred here ecause they provide more rapid expans ion of the intravascular volume than crystalloid fluids (see Chapter 15). PHARMACOTHERAPY The pharmacologic agents listed in Ta le 20.2 have all proven successful in the acute http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 26)03-May-05 07:46:02

Ovid: ICU Book management of AF. Only one of the agents (procainamide) P.324 is capa le of conv erting AF to a normal sinus rhythm. The remainder act y slowing the ventricular rate. TABLE 20.2. ACUTE PHARMACOTHERAPY OF ATRIAL FIBRILLATION Verapamil http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 26)03-May-05 07:46:02

Ovid: ICU Book

Dosage: 0.075 mg/kg intravenously over 2 minutes. If no response after 10 minutes, give a second dose of 0.15 mg/kg. The response is evident within minutes after drug i njection, and the effect lasts up to 1 hour (2). Because the response is short-l ived, the effective dose of verapamil should e followed y continuous infusion (5 to 20 mg/hour) (9,10). Verapamil is meta olized in the liver, and the dosage should e reduced 50% in patients with hepatic insufficiency. The major draw ack s with verapamil are hypotension (due to vasodilator actions) and worsening syst olic heart failure (due to the negative inotropic actions) (10). The risk of hyp otension is reduced y pretreatment with intravenous calcium (11). Calcium pretreatment: Give 10 mL of 10% calcium gluconate intravenously over 5 minutes, just efore ve rapamil. This regimen provides 90 mg of elemental calcium. Calcium is also avail a le as 10% calcium chloride, ut this salt contains three times more elemental calcium than 10% calcium gluconate (i.e., 10 mL of 10% calcium chloride provides 270 mg of elemental calcium). Therefore, do not ask for an amp of calciumspecify t he gluconate salt. The drug interactions involving verapamil are included in Ta le 20.3. Verapamil increases serum digoxin levels (mechanism not clear), so it i s wise to monitor serum digoxin levels when instituting therapy with verapamil. Verapamil can also precipitate cardiovascular collapse when given in com ination with - lockers, and com ined therapy is contraindicated. Other contraindications to verapamil infusion include second- or third-degree AV lock, systolic heart failure, hypotension, and WolfParkinsonWhite (WPW) syndrome (2). The pro lem in th e WPW syndrome is the tendency for verapamil (and all agents that prolong AV con duction) to paradoxically increase the ventricular rate in AF. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 26)03-May-05 07:46:02

Intravenous verapamil, a calcium channel l in 70% of patients with rapid AF (2).

locker, provides effective rate contro

Ovid: ICU Book TABLE 20.3. ANTIARRHYTHMIC DRUG INTERACTIONS Diltiazem Another calcium channel locker, diltiazem, provides the same rapid and effectiv e rate control in AF as verapamil. However, diltiazem produces less myocardial d epression than verapamil (12). For this reason, diltiazem is the preferred agent for patients with (systolic) heart failure. The dosage shown here (and in Ta le 20.2) has een used safely in patients with moderate to severe heart failure (1 3). Dosage: 0.25 mg/kg intravenously over 2 minutes. If no response after 10 minutes, give a second dose of 0.35 mg/kg. P.325 Like those of verapamil, diltiazem's effects a re short-lived, so the effective dose should e followed y a continuous infusio n of the drug (10 to 15 mg/hour) (12,13). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 26)03-May-05 07:46:02

Ovid: ICU Book -Receptor Antagonists -Blockers are used for rate control in AF when it is accompanied y hyperadrenerg ic states (such as acute MI and after cardiac surgery). The cardioselective - loc kers (which preferentially lock -1 receptors) have largely replaced the nonselec tive lockers such as propranolol. Ta le 20.2 includes two cardioselective agent s: esmolol and metoprolol. Esmolol (Brevi loc) has received much attention ecau se it is an ultrashort-acting agent (serum halflife of 9 minutes) that can e tit rated rapidly (14). However, as demonstrated in the recommended dosing regimen h ere, therapy with esmolol can e a tedious process. Esmolol dosing regimen: Start with a 30-second loading dose of 0.5 mg/kg and follow with continuous infu sion at 50 g/ kg/minute for 4 minutes. If the rate is not adequately controlled, repeat the loading dose and increase the infusion to 100 g/kg/minute for another 4 minutes. If the rate is still not controlled, increase the dosing rate in incr ements of 50 mg/kg/minute every 4 minutes to a maximum dosing rate of 300 g/kg/mi nute. Each increase in dosing rate should e preceded y a loading dose of 0.5 m g/kg (15). Of the - lockers, metoprolol (Lopressor) is much more convenient to us e. Com ined therapy with - lockers and calcium channel P.326 lockers should e a voided, ecause this com ination can produce severe cardiac depression (see Ta l e 20.3) (8). Magnesium Intravenous magnesium has proven safe and effective for rate control in AF (16,1 7) and for prophylaxis of postoperative AF (18). Dosage: 2 g MgSO4 over 5 to 15 minutes. Then infuse 6 g MgSO4 in 500 mL saline over 6 ho urs (1 g/ hour). Although the clinical experience with magnesium in AF is limite d, there is reason to elieve that magnesium will e a valua le intervention for AF in the ICU. First, magnesium is a calcium channel locker, and other agents with a similar action (i.e., verapamil and diltiazem) provide effective rate con trol in AF. Second, magnesium is a necessary cofactor for the proper function http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 26)03-May-05 07:46:02

Ovid: ICU Book of the mem rane pump for sodium and potassium. Because this pump maintains a hyp erpolarized state in cells, magnesium could act as an antiarrhythmic agent y vi rtue of its a ility to sta ilize excita le cell mem ranes. Third, ecause magnes ium deficiency is common in ICU patients (see Chapter 42), magnesium infusion co uld reduce the ventricular rate in AF y correcting an underlying magnesium defi ciency. Digoxin Digoxin is a popular agent for chronic rate control in AF, ut ecause of a dela yed onset of action, digoxin is not indicated for the acute management of AF. Th e delayed action of intravenous digoxin is illustrated in Figure 20.3. This figu re is taken from a study of patients who presented to the emergency room with re cent-onset AF and were randomized to receive either intravenous diltiazem (same dosage as in Ta le 20.2) or intravenous digoxin (0.5 mg in two divided doses ove r 30 minutes) (19). As demonstrated, the patients who received diltiazem showed adequate rate control (i.e., a heart rate elow 100 eats/minute) y 15 minutes after the onset of drug therapy, whereas the patients who received digoxin were not adequately controlled after 3 hours. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 26)03-May-05 07:46:02

Ovid: ICU Book Figure 20.3. Acute rate control of AF with IV diltiazem (dose same as in Ta le 2 0.2) and IV digoxin (0.5 mg in two divided doses over 6 hours). Height of the ve rtical columns indicates the mean heart rate for all patients in each treatment group. (Data from Schreck DM et al. Emergency treatment of atrial fi rillation a nd flutter: comparison of IV digoxin versus IV diltiazem. Ann Emerg Med 1995;25: 127.) Procainamide The goal of intravenous procainamide therapy is not to control the ventricular r ate, ut to convert AF to a normal sinus rhythm. Dosage: 10 mg/kg intravenously at a maximum rate of 25 mg/minute. Follow with a continuo us infusion at 1 to 6 mg/minute (20). Half of a procainamide dose is cleared unc hanged y the kidneys, and 15% is first converted in plasma to an active meta ol ite, to N-acetyl procainamide (NAPA), which is then cleared y the kidneys. P.32 7 The dosage of procainamide should e reduced y 50% in elderly patients and in patients with renal dysfunction, and the dosage should e reduced y 25% in hea rt failure (20). Procainamide transiently facilitates AV conduction, so the drug should e given in com ination with an agent that locks AV conduction. Because procainamide has negative inotropic actions, digitalis is favored over calcium lockers or - lockers to provide the AV lock. Procainamide also prolongs the QT i nterval, and can e proarrhythmic in patients with a prolonged QT interval. For t his reason, procainamide is contraindicated in patients with prolonged QT interva l. The drug should also e avoided in patients with renal failure.

WOLFPARKINSONWHITE SYNDROME Patients with WPW syndrome (short PR interval and waves before the QRS) have an a ccessory, re-entrant pathway in the AV node that predisposes them to recurrent t achycardias (both narrow complex and wide complex), including atrial fibrillatio n. Agents that block AV conduction and slow the rate in conventional AF, such as digoxin and verapamil, can paradoxically accelerate the ventricular rate (by bl ocking the wrong pathway) in the AF associated with WPW syndrome. In patients wi th a history of WPW syndrome who develop AF, http://gateway.ut.ovid.com/gw1/ovidweb.cgi (14 of 26)03-May-05 07:46:02

Ovid: ICU Book P.328 calcium channel blockers and digoxin are contraindicated, and the treatmen t of choice is electrical cardioversion or procainamide infusion. ATRIAL AN JUNCTIONAL TACHYCAR IAS Some atrial tachycardias are caused by ectopic foci, but most are the result of re-entry through the AV node. The following arrhythmias are commonly encountered in the ICU. MULTIFOCAL ATRIAL TACHYCAR IA Multifocal atrial tachycardia (MAT) is characterized by multiple P wave morpholo gies and a variable PR interval (21). The ventricular rate is highly irregular, a nd MAT is easily confused with atrial fibrillation when atrial activity is not c learly displayed on a single-lead rhythm strip. MAT is most often seen in patien ts with chronic lung disease and has also been linked to theophylline therapy (2 2). Other predisposing conditions include hypokalemia, magnesium deficiency, acu te pulmonary embolism, acute MI, and congestive heart failure (1,2,21). Acute Management MAT can be a difficult arrhythmia to manage. In patients with severe lung diseas e, aggravating conditions, such as hypoxemia, should be corrected. Then proceed as follows. q iscontinue ongoing therapy with theophylline. In one study, this maneuver resul ted in conversion to sinus rhythm in half the patients with MAT (22). q In the absence of contraindications to magnesium therapy, give intravenous magne sium using the dosing regimen in Table 20.2 (23). In addition to its antiarrhyth mic actions, magnesium also dilates pulmonary arteries (which might help in pati ents with MAT secondary to chronic obstructive pulmonary disease, COP ). q If the serum potassium is low, give intravenous magnesium first (2 g MgSO4 in 50 mL saline over 15 minutes), then infuse 40 mg potassium over 1 hour. The magnes ium pretreatment is often necessary for correcting the hypokalemia (see Chapter 42). q If the above measures are ineffective, consider therapy with verapamil or metopr olol http://gateway.ut.ovid.com/gw1/ovidweb.cgi (15 of 26)03-May-05 07:46:02

Ovid: ICU Book (see dosages in Table 20.2). If the patient has a history of reactive airway dis ease, verapamil (which is a mild bronchodilator) is preferred to - lockers. If th e patient has COPD without reactive airways, selective -1 receptor antagonists su ch as metoprolol can e used safely. Metoprolol has een P.329 reported to conve rt MAT to sinus rhythm in 70 to 100% of cases (2). AV NODAL RE-ENTRANT TACHYCARDIA Re-entry of impulses through the AV node (triggered y an ectopic pulse) is the most common cause of narrow-complex tachycardias with a regular rate (other than sinus tachycardia). These tachycardias are also known as paroxysmal SVTs, and a re characterized y an a rupt onset and the a sence of identifia le atrial activ ity (P waves hidden in the QRS complex). These tachycardias are faster than a si nus tachycardia, with rates usually etween 140 and 220 eats/minute. Acute Management Maneuvers aimed at increasing vagal tone were once popular for AV nodal re-entra nt tachycardia (AVNRT). However, most of these maneuvers are ineffective, and so me (such as eye all compression) are downright dangerous. The standard approach for terminating AVNRT is to administer drugs that lock the re-entrant pathway i n the AV node. The most effective agents for this purpose are the calcium channe l lockers, (verapamil and diltiazem), and adenosine. These agents are equally e ffective for terminating AVNRT, ut adenosine has less cardiovascular depressant effect than the calcium antagonists. The salient features of adenosine administ ration are presented in Ta le 20.4. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 26)03-May-05 07:46:02

Ovid: ICU Book TABLE 20.4. INTRAVENOUS ADENOSINE FOR SVT Adenosine (Adenocard) Adenosine is an endogenous su stance that dilates coronary arteries, slows the s inus rate, prolongs AV conduction, and locks the positive inotropic actions of catecholamines (24). When given y rapid intravenous injection, adenosine termin ates AVNRT in 90 to 100% of cases (24,25 and 26). However, ecause adenosine is an ultrashort-acting agent (effects last 1 or 2 minutes), the cardiac depressant actions of adenosine are too short-lived to produce significant myocardial depre ssion (24,25 and 26). For this reason, adenosine is the drug of choice for termi nating AVNRT in patients with heart failure. The intravenous dosing recommendati ons for adenosine are included in Ta le 20.4. Note that the dosage of adenosine should e reduced y 50% when the drug is injected through a central venous (CVP ) catheter instead of a peripheral vein (27). This recommendation is ased on re ports of ventricular asystole when adenosine in standard doses is given through CVP catheters (27). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 26)03-May-05 07:46:02

Ovid: ICU Book The drug interactions with adenosine are shown in Ta le 20.3. Note that theophyl line antagonizes the actions of adenosine y locking adenosine receptors. As a result, therapeutic doses of adenosine may not e effective in patients receivin g theophylline, so com ined therapy with adenosine and theophylline is not advis ed. P.330 Side effects are common after adenosine injection, and the most common ones are listed in Ta le 20.4 (along with the incidence of each in parentheses) (26). Fortunately, these side effects are short-lived. The anginal-type chest p ain is often a source of concern, ut is not the result of myocardial ischemia. Adenosine also provokes ronchoconstriction in asthmatic su jects (28,29), and t he drug is contraindicated in patients with asthma. VENTRICULAR TACHYCARDIA The identity of a wide (QRS) complex tachycardia is easily confused on a singlelead ECG recording, as demonstrated in Figure 20.2. Although it is possi le to d ifferentiate etween VT and SVT using a 12-lead ECG (30), this type of detailed analysis is more suited to the cardiac electrophysiology la than the ICU. As st ated in the most recent ACLS guidelines, Faced with urgent care of an ill patient , the physician should ignore detailed criteria for ECG analysis and attend to t he patient (1). In fact, detailed ECG criteria may not e necessary, ecause as m any as 95% of wide complex tachycardias in patients with underlying cardiac dise ase represent VT (31). Because most patients P.331 in the ICU have cardiac disea se, a wide complex tachycardia in an ICU patient should e treated as pro a le V T. ACUTE MANAGEMENT The acute management of pro a le ventricular tachycardia should proceed as follo ws (1,32). This approach is outlined in Ta le 20.5. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 26)03-May-05 07:46:02

Ovid: ICU Book TABLE 20.5. STAGED APPROACH TO VENTRICULAR TACHYCARDIA q If there is evidence of hemodynamic compromise, initiate DC cardioversion immedi ately with an initial shock of 100 J, followed y repetitive shocks of 200, 300, and 360 J, if necessary (see Ta le 20.1). q If there is no evidence of hemodynamic compromise, administer lidocaine at the d osages shown in Ta le 20.5. If olus doses of lidocaine are successful in termin ating the arrhythmia, start a continuous infusion at 2 to 4 mg/minute. Prolonged infusions of lidocaine can cause an excitatory neurotoxic syndrome, particularl y in elderly patients and in those receiving - lockers or cimetidine P.332 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 26)03-May-05 07:46:02

Ovid: ICU Book (see Ta le 20.3). Therefore, lidocaine infusion should not e continued eyond 6 to 12 hours. q If there is no response to lidocaine, check the QT interval on a recent ECG recor ding. The QT interval is usually measured in lead II, and is corrected for the he art rate as follows: Corrected QT (QTc) = measured QT divided y the square root of the RR interval. A QTc > 0.44 second is considered a normal (33). If the QT interv al is normal (QTc < 0.44 second), administer procainamide. As indicated in Ta le 20.5, procainamide is not given as a olus injection, ut is infused slowly at a rate of 25 mg/ minute. As a result, the response to procainamide can e delayed (up to 15 minutes after the onset of drug administration). The procainamide inf usion should e continued until cardioversion, or until the cumulative dose reac hes a maximum of 17 mg/kg (1 to 1.5 g) (34). If the QRS duration increases y mo re than 50%, or if hypotension develops, the infusion should e stopped temporar ily. Procainamide is an appealing agent for wide complex tachycardias ecause the dru g can convert oth VT and SVT. As mentioned earlier, the drug is contraindicated in patients with a prolonged QT interval (can e proarrhythmic). If the QT interv al is prolonged (QTc > 0.44 second), intravenous magnesium is recommended (except in the presence of renal failure). Magnesium not only is effective in suppressi ng ventricular arrhythmias, ut can e effective in VT refractory to lidocaine ( 35). Because a prolonged QT interval can e the result of magnesium depletion, ma gnesium infusions may e particularly effective in patients with a prolonged QT i nterval. The ACLS guidelines recommend retylium for VT that is refractory to li docaine (1). However, retylium has not proven more effective in suppressing VT than lidocaine (32). In addition, retylium can cause profound hypotension (the drug was originally developed as an antihypertensive agent). The lack of improve d efficacy, along with the potential for serious toxicity, suggests that the rec ommendations for retylium in VT must e re-examined. Precipitating Factors Cardioversion of VT (either electrical or pharmacologic) should e followed y a search for precipitating conditions. The standard approach should include an EC G (ischemia), arterial lood gases (hypoxemia and alkalosis), serum electrolytes (hypokalemia, hypocalcemia, hypomagnesemia), and serum levels of select drugs ( e.g., digitalis and theophylline). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 26)03-May-05 07:46:02

Ovid: ICU Book TORSADE DE POINTES Torsade de pointes (twisting around the points) is a polymorphic VT characterized y phasic changes in the amplitude and polarity of P.333 the ventricular complex es (Fig. 20.4). This arrhythmia is often, ut not always, associated with a prol onged QT interval (36). A variety of drugs and electrolyte deficiencies can predi spose to this arrhythmia. Predisposing drugs include antiarrhythmic agents (quin idine, procainamide), antimicro ial agents (erythromycin, pentamidine), and psyc hotropic agents (haloperidol, phenothiazines). Predisposing electrolyte disorder s include calcium, magnesium, and potassium deficiency. Figure 20.4. Torsade de pointes, which means twisting around (the isoelectric) po ints. (Tracing courtesy of Dr. Richard M. Green erg, M.D.) Management The management of this arrhythmia is guided y the QT interval. When the QT interv al is prolonged, the arrhythmia is often resistant to traditional antiarrhythmic agents. The treatment of choice is temporary ventricular pacing to increase the heart rate (100 to 120 eats/minute) and there y shorten the QT interval. When t he QT interval is normal, traditional antiarrhythmic therapy (e.g., with lidocain e or procainamide) is usually effective. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 26)03-May-05 07:46:02

Ovid: ICU Book After this arrhythmia is terminated, potentially offending drugs should e disco ntinued and electrolyte deficiencies should e corrected. REFERENCES GENERAL TEXTS Podrid PJ, Kowey PR, eds. Hand ook of cardiac arrhythmia. Baltimore: Williams & Wilkins, 1996. REVIEWS 1. Emergency Cardiac Care Committee and Su committees, American Heart Associatio n. Guidelines for cardiopulmonary resuscitation and emergency cardiac care. JAMA 1992;268:21992241 (430 References). 2. Collier WW, Holt SE, Wellford LA. Narrow complex tachycardias. Emerg Med Clin North Am 1995;13:925954 (99 References). 3. Dell ridge TR, Yealy DM. Wide complex tachycardia. Emerg Med Clin North Am 19 95;13:903924 (69 References). P.334 ATRIAL FLUTTER AND FIBRILLATION 4. Guyton AC. The relationship of cardiac output and arterial pressure control. Circulation 1981;64:10791088. 5. Blackshear JL, Kopecky SL, Litin SC, et al. Management of atrial fi rillation in adults: prevention of throm oem olism and symptomatic treatment. Mayo Clin P roc 1996;71:150160. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 26)03-May-05 07:46:02

Ovid: ICU Book 6. Sie ers MJ, Drinka PJ, Vergauwen C. Hyperthyroidism as a cause of atrial fi r illation in long-term care. Arch Intern Med 1992;152:20632064. 7. Creswell LL, Schuessler RB, Rosen loom M, Cox JL. Hazards of postoperative at rial arrhythmias. Ann Thorac Surg 1993;56:539549. 8. Kuhn M, Schriger DL. Verapamil administration to patients with contraindicati ons: is it associated with adverse outcomes? Ann Emerg Med 1991;20:10941099. 9. Edwards JD, Kishen R. Significance and management of intracta le supraventric ular arrhythmias in critically ill patients. Crit Care Med 1986;14:280282. 10. I erti TJ, Benjamin E, Paluch TA, et al. Use of constant-infusion verapamil for the treatment of postoperative supraventricular tachycardia. Crit Care Med 1 986;14:283284. 11. Dolan DL. Intravenous calcium efore verapamil to prevent hypotension. Ann E merg Med 1991;20:588589. 12. Ellen ogen KA, Dias VC, Plum VJ, et al. A place o-controlled trial of conti nuous intravenous diltiazem infusion for 24-hour heart rate control during atria l fi rillation and atrial flutter: a multicenter study. J Am Coll Cardiol 1991;1 8:891897. 13. Golden erg IF, Lewis WR, Dias VC, et al. Intravenous diltiazem for the treat ment of patients with atrial fi rillation or flutter and moderate to severe cong estive heart failure. Am J Cardiol 1994;74:884889. 14. Gray RJ. Managing critically ill patients with esmolol. An ultra-short-actin g adrenergic locker. Chest 1988;93:398404. 15. Brevi loc (esmolol HCL) Dosage Chart. Pharmaceutical Products Division. Ohme da, Inc., Li erty Corner, NJ: 1993. 16. Brodsky MA, Orlov MV, Capparelli EV, et al. Magnesium therapy in acute-onset atrial http://gateway.ut.ovid.com/gw1/ovidwe .cgi (23 of 26)03-May-05 07:46:02

Ovid: ICU Book fi rillation. Am J Cardiol 1994;73:12271229. 17. Hays JV, Gilman JK, Ru al BJ. Effect of magnesium sulfate on ventricular rat e control in atrial fi rillation. Ann Emerg Med 1994;24:6164. 18. Fanning WJ, Thomas CS Jr, Roach A, et al. Prophylaxis of atrial fi rillation with magnesium sulfate after coronary artery ypass grafting. Ann Thorac Surg 1 991;52:529 533. 19. Schreck DM, Rivera AR, Zacharias D. Emergency treatment of atrial fi rillati on and atrial flutter: comparison of IV digoxin versus IV diltiazem. Ann Emerg M ed 1995;25:127 (a stract). 20. Marcus FI, Opie LH. Antiarrhythmic drugs. In: Opie LH, ed. Drugs for the hea rt. 4th ed. Philadelphia: WB Saunders, 1995;207246. ATRIAL & JUNCTIONAL TACHYCARDIAS 21. Kastor J. Multifocal atrial tachycardia. N Engl J Med 1990;322:17131720. 22. Levine J, Michael J, Guanieri T. Multifocal atrial tachycardia: a toxic effe ct of theophylline. Lancet 1985;1:116. 23. Iseri LT, Fairshter RD, Hardeman JL, Brodsky MA. Magnesium and potassium the rapy in multifocal atrial tachycardia. Am Heart J 1985;312:2126. 24. Shen W-K, Kurachi Y. Mechanisms of adenosine-mediated actions on cellular an d clinical cardiac electrophysiology. Mayo Clin Proc 1995;70:274291. P.335 25. Rankin AC, Brooks R, Ruskin JM, McGovern BA. Adenosine and the treatment of supraventricular tachycardia. Am J Med 1992;92:655664. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (24 of 26)03-May-05 07:46:02

Ovid: ICU Book 26. Chronister C. Clinical management of supraventricular tachycardia with adeno sine. Am J Crit Care 1993;2:4147. 27. McCollam PL, U er W, Van Bakel AB. Adenosine-related ventricular asystole. A nn Intern Med 1993;118:315316.

29. Bjorck T, Gustafsson LE, Dahlen S-E. Isolated ronchi from asthmatics are hy perresponsive to adenosine, which apparently acts indirectly y li eration of le ukotrienes and histamine. Am Rev Respir Dis 1992;145:10871091. VENTRICULAR TACHYCARDIA 30. Brugada P, Brugada J, Mont L, et al. A new approach to the differential diag nosis of a regular tachycardia with a wide QRS complex. Circulation 1991;83:16491 659. 31. Akhtar M, Shenasa M, Jazayeri M, et al. Wide QRS complex tachycardia. Ann In tern Med 1988;109:905912. 32. Slovis CM, Wrenn KD. The technique of managing ventricular tachycardia. J Cr it Illness 1993;8:731741. 33. Garson A Jr. How to measure the QT interval: what is normal? Am J Cardiol 19 93;72:14B16B. 34. Sharma AD, Purves P, Yee R, et al. Hemodynamic effects of intravenous procai namide during ventricular tachycardia. Am Heart J 1990;119:10341041. 35. Roden D. Magnesium treatment of ventricular arrhythmias. Am J Cardiol 1989;6 3:43G46G. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (25 of 26)03-May-05 07:46:02

28. Cushley MJ, Tattersfield AE, Holgate ST. Adenosine-induced on in asthma. Am Rev Respir Dis 1984;129:380384.

ronchoconstricti

Ovid: ICU Book 36. Vukmir RB. Torsades de pointes: a review. Am J Emerg Med 1991;9:250262. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (26 of 26)03-May-05 07:46:02

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 21 HYPOXEMIA AND HYPERCAPNIA There is a distinct fondness for arterial lood gas analysis in ICU patients. In fact, arterial lood gases are the most commonly performed la oratory test in I CU patients (1,2); and with the advent of edside methods for measuring lood ga ses (point of patient care la oratory testing) (3) and continuous in vivo lood gas monitoring (4), it is likely that lood gas measurements in days to come wil l e even more inundating. This chapter focuses on what to do when the arterial lood gas shows a low PO2 or a high PCO2. PULMONARY GAS EXCHANGE The adequacy of gas exchange in the lungs is determined y the alance etween p ulmonary ventilation and capillary lood flow (5,6,7 and 8). This alance is com monly expressed as the ventilationperfusion (V/Q) ratio. The influence of V/Q rat ios on pulmonary gas exchange can e descri ed using a schematic alveolarcapillar y unit, as shown in Figure 21.1. The upper panel shows a perfect match etween v entilation and perfusion; that is, a V/Q ratio of 1.0. This is the reference poi nt for defining the a normal patterns of gas exchange. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 20)03-May-05 17:04:07

Ovid: ICU Book

DEAD SPACE VENTILATION A V/Q ratio a ove 1.0 (Fig. 21.1, middle panel) descri es the condition where ve ntilation is excessive relative to capillary lood flow. The excess ventilation, known as dead space ventilation, does not participate in gas exchange with the lood. There are two types of dead space ventilation. Anatomic dead space is the gas in the large conducting airways that does not come in contact with capillar ies. Approximately 50% of the anatomic dead space is in the pharynx. Physiologic dead space is the alveolar gas that does not equili rate P.340 fully with capil lary lood. This represents excess alveolar ventilation relative to capillary l ood flow. In normal su jects, dead space ventilation (Vd) accounts for 20 to 30% of the total ventilation (Vt); that is, Vd/Vt = 0.2 to 0.3 (5,7). An increase i n Vd/Vt results in oth hypoxemia and hypercapnia (analogous to what would happe n if you held your reath). The hypercapnia usually appears when the Vd/Vt rises a ove 0.5 (7). Pathophysiology http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 20)03-May-05 17:04:07

Figure 21.1. Ventilationperfusion (V/Q) relationships and associated lood gas a normalities.

Ovid: ICU Book Dead space ventilation increases when the alveolarcapillary interface is destroye d (e.g., emphysema), when lood flow is reduced (e.g., heart failure, pulmonary em olism), or when alveoli are overdistended y positive-pressure ventilation. INTRAPULMONARY SHUNT A V/Q ratio elow 1.0 (Fig. 21.1, lower panel) descri es the condition where cap illary lood flow is excessive relative to ventilation. The excess lood flow, k nown as intrapulmonary shunt, does not participate in pulmonary gas exchange. Th ere are two types of intrapulmonary shunt. True shunt indicates the total a senc e of exchange etween capillary P.341 lood and alveolar gas (V/Q = zero), and i s equivalent to an anatomic shunt etween the right and left sides of the heart. Venous admixture represents the capillary flow that does not equili rate comple tely with alveolar gas (V/Q a ove zero ut less than 1.0). As the venous admixtu re increases, the V/Q im alance approaches true shunt conditions (V/Q = 0). The fraction of the cardiac output that represents intrapulmonary shunt is known as the shunt fraction. In normal su jects, intrapulmonary shunt flow (Qs) represent s less than 10% of the total cardiac output (Qt); that is, the shunt fraction (Q s/Qt) is less than 10% (5,6,8). Pathophysiology Intrapulmonary shunt fraction increases when the small airways are occluded (e.g ., asthma, chronic ronchitis), when alveoli are filled with fluid (e.g., pulmon ary edema, pneumonia) or when alveoli collapse (e.g., atelectasis), and when cap illary flow is excessive (e.g., nonem olized regions of lung in pulmonary em oli sm). Arterial Blood Gases The influence of shunt fraction on arterial oxygen and car on dioxide tensions ( PAO2, PaCO2, respectively) is shown in Figure 21.2. The PAO2 falls progressively as shunt fraction increases, ut the PaCO2 remains P.342 constant until the shu nt fraction exceeds 50% (8). The PaCO2 is often elow normal in patients with in creased intrapulmonary shunt as a result of hyperventilation from the disease pr ocess (e.g., pneumonia) or from the accompanying hypoxemia. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 20)03-May-05 17:04:07

Ovid: ICU Book Figure 21.2. The influence of shunt fraction on arterial PO2 (PAO2) and arterial PCO2 (PaCO2). (From D'Alonzo GE, Dantzger DR. Med Clin North Am 1983;67:557571.) The shunt fraction also determines the influence of inspired oxygen on the arter ial PO2. This is shown in Figure 21.3 (8). As intrapulmonary shunt increases fro m 10 to 50%, the increase in fractional concentration of inspired oxygen (FIO2) produces less of an increment in the arterial PO2. When the shunt fraction excee ds 50%, the PAO2 is independent of changes in FIO2. This latter condition mimics the ehavior of a true (anatomic) shunt. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 20)03-May-05 17:04:07

Ovid: ICU Book Figure 21.3. The influence of shunt fraction on the relationship etween the ins pired oxygen (FIO2) and the arterial PO2 (PAO2). (From D'Alonzo GE, Dantzger DR. Med Clin North Am 1983;67:557571.) The diminished influence of inspired oxygen on arterial PO2 as shunt fractions r ises has important implications for limiting the risk of pulmonary oxygen toxici ty. That is, in conditions associated with a high shunt fraction (e.g., acute re spiratory distress syndrome), the FIO2 can often e reduced to levels considered nontoxic (i.e., FIO2 elow 50%) without further compromising arterial oxygenati on. QUANTITATIVE DETERMINATIONS The following determinations can e used to identify and quantitate ventilationpe rfusion a normalities. As will e demonstrated, these determinations can prove u seful in oth the diagnosis and management of respiratory failure. P.343 DEAD SPACE (VD/VT) The determination of dead space ventilation (Vd/Vt) is ased on the difference etween the PCO2 in exhaled gas and the PCO2 in end-capillary (arterial) lood. I n the normal lung, the capillary lood equili rates fully with alveolar gas, and the exhaled PCO2 (PECO2) is roughly equivalent to the arterial PCO2 (PaCO2). As dead space ventilation (Vd/Vt) increases, the PECO2 falls elow the PaCO2. The Bohr equation shown elow (derived y Cristian Bohr, father of Neils Bohr) is a sed on this principle. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 20)03-May-05 17:04:07

Ovid: ICU Book Thus, as the exhaled PCO2 decreases relative to the arterial PCO2, the calculate d Vd/Vt rises. The PECO2 is measured y collecting expired gas in a large collec tion ag and using an infrared CO2 analyzer to measure the PCO2 in a sample of t he gas. This is usually done on request y the respiratory therapy department. SHUNT FRACTION (QS/QT) The shunt fraction (Qs/Qt) is not as easily determined as the Vd/Vt. The Qs/Qt i s derived y the relationship etween the O2 content in arterial lood (CaO2), m ixed venous lood (CvO2), and pulmonary capillary lood (CcO2). The pro lem here is the ina ility to measure the capillary O2 content (CcO2) dir ectly. As a result, pure oxygen reathing has een recommended (to produce 100% oxyhemoglo in saturation in pulmonary capillary lood) for the shunt calculation . However in this situation, Qs/Qt measures only true shunt. THE A-a PO2 GRADIENT The PO2 difference etween alveolar gas and arterial lood (A-a po2 gradient) is used as an indirect measure of ventilationperfusion a normalities (9,10 and 11,1 3). The A-a po2 gradient is determined with the alveolar gas equation shown elo w. This equation descri es the relationship etween the alveolar PO2 (PAO2), the PO 2 in inspired gas (PIO2), the alveolar (arterial) PCO2, and the respiratory quot ient (RQ). The latter varia le defines the proportional P.344 exchange of O2 and CO2 across the alveolarcapillary interface; that is, RQ = VCO2/VO2. The PIO2 is a function of the fractional concentration of inspired oxygen (FIO2), the arome tric pressure (PB), and the partial pressure of water vapor (PH2O) in humidified gas; that is, PIO2 = FIO2 (PB PH2O). The PH2O is 47 mm Hg at ody temperature. In a healthy su ject reathing room air at sea level, where FIO2 = 0.21, PB = 76 0 mm Hg, PH2O = 47 mm Hg, PAO2 = 90 mm Hg, PaCO2 = 40 mm Hg, and RQ = 0.8: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 20)03-May-05 17:04:07

Ovid: ICU Book Because the arterial PO2 is 90 mm Hg, the A-a po2 gradient in this example is 10 mm Hg. This represents an idealized rather than normal A-a po2 gradient, ecaus e the A-a po2 gradient varies with age and with the concentration of inspired ox ygen. Influence of Age As shown in Ta le 21.1, the normal A-a po2 gradient rises steadily with advancin g age (10). Assuming that most patients in an adult ICU are 40 years of age or o lder, a normal A-a po2 gradient in an ICU patient's reathing room air may e as high as 25 mm Hg. However, few ICU patients reathe room air, and as demonstrat ed next, the normal A-a po2 gradient is increased during supplemental oxygen re athing. TABLE 21.1. NORMAL ARTERIAL BLOOD GASES Influence of Inspired Oxygen The influence of inspired oxygen on the A-a po2 gradient is shown in Figure 21.4 (11). The A-a po2 gradient increases from 15 to 60 mm Hg as the FIO2 increases from room air to pure oxygen. According to this relationship, the normal A-a pO2 gradient increases 5 to 7 mm P.345 Hg for every 10% increase in FIO2. This effe ct is presuma ly caused y the loss of regional hypoxic vasoconstriction in the lungs. Hypoxic vasoconstriction in poorly ventilated lung regions can http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 20)03-May-05 17:04:07

Ovid: ICU Book serve to maintain V/Q alance y diverting lood to more adequately ventilated l ung regions. Loss of this regional hypoxic vasoconstriction during supplemental oxygen reathing increases lood flow in poorly ventilated lung regions. This in creases the intrapulmonary shunt fraction and there y increases the A-a po2 grad ient. Figure 21.4. The influence of FIO2 on the alveolar-arterial PO2 gradient (A-a po 2) and the arterialalveolar PO2 ratio (a/A PO2) in normal su jects. (From Gil er t R, Kreighley JF. Am Rev Respir Dis 1974;109:142145.) Positive-Pressure Ventilation Positive-pressure mechanical ventilation elevates the pressure in the airways a ove the am ient arometric pressure. Therefore, when determining the A-a po2 gra dient in a ventilator-dependent patient, the mean airway pressure should e adde d to the arometric pressure (12). In the example presented previously, a mean a irway pressure of 30 cm H2O would increase the A-a po2 gradient from 10 to 16 mm Hg (a 60% increase in the A-a po2 gradient). Although this correction is necess ary for optimal accuracy in determining the A-a po2 gradient, the clinical signi ficance is unproven. THE a/A PO2 RATIO Unlike the A-a po2 gradient, the a/A PO2 ratio is relatively unaffected y the F IO2. This is demonstrated in Figure 21.4 (11). The independence http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 20)03-May-05 17:04:07

Ovid: ICU Book

FIO2 0.21 1.0 Normal a/A PO2 0.74 to 0.77 0.80 to 0.82 THE PaO2/FIO2 RATIO Another simple determination that can correlate with shunt fraction is the ratio of arterial PO2 to FIO2. The following correlations have een reported (13). PaO2/FIO2 < 200 > 200 Qs/Qt > 20% < 20% http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 20)03-May-05 17:04:07

Because the alveolar PO2 is in n, the influence of FIO2 on the mathematical manipulation that ient. The normal a/A PO2 ratios own elow (11).

oth the numerator and denominator of the equatio PAO2 is eliminated. Thus, the a/A PO2 ratio is a eliminates the influence of FIO2 on the A-a grad during room air and pure oxygen reathing are sh

P.346 of the a/A PO2 gradient in relation to the FIO2 is explained on elow.

y the equati

Ovid: ICU Book The major limitation of the PAO2/FIO2 ratio is the varia ility of the FIO2 in pa tients receiving oxygen via nasal prongs or face mask (see Chapter 23). This lim itation also applies to the A-a po2 gradient. BLOOD GAS VARIABILITY The very first step in the approach to managing hypoxemia and hypercapnia is def ining what constitutes an a normal change in the arterial PO2 and PCO2. The info rmation in Ta le 21.2 is relevant to this issue. The data in this ta le are from a study of 26 clinically sta le, ventilatordependent trauma victims who each ha d a series of four arterial lood gas measurements performed over a 1-hour perio d (14). The varia ility in the arterial PO2 and PCO2 for all patients is present ed P.347 in the ta le. The arterial PO2 varied y as much as 36 mm Hg, whereas t he arterial PCO2 varied y as much as 12 mm Hg. This varia ility is similar to t hat reported in another study involving patients in a medical ICU (15). TABLE 21.2. SPONTANEOUS BLOOD GAS VARIABILITY The varia ility in measured lood gas varia les shown in Ta le 21.2 highlights t he following two important points. q http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 20)03-May-05 17:04:07

Ovid: ICU Book Routine monitoring of arterial lood gases (without a change in the clinical con dition of the patient) is not warranted and often yields misleading information. q A change in arterial PO2 and PCO2 on routine lood gas analysis is not necessari ly a normal if the clinical condition of the patient has not changed. This shoul d e considered efore a lengthy and time-consuming search for something that ma y not e found is initiated. HYPOXEMIA When a lood gas showing a significant reduction in the arterial PO2 is encounte red, there are three principal disorders to consider (5,8,16,17). These are show n in Ta le 21.3, along with the measurements that will help identify each disord er. Note that one of the causes of hypoxemia is an im alance etween oxygen deli very and oxygen uptake in the systemic circulation (DO2/VO2 im alance). In this situation, the peripheral O2 extraction is increased (due to a low O2 delivery o r an increased O2 uptake) and the mixed venous PO2 is reduced. The relationship etween the mixed venous PO2 and the arterial PO2 is descri ed elow. TABLE 21.3. EVALUATION OF HYPOXEMIA MIXED VENOUS PO2 The oxygen in arterial lood represents the sum of the oxygen in mixed venous (p ulmonary artery) lood and the oxygen added from alveolar gas. When gas exchange is normal, the PO2 in alveolar gas is the major determinant of the arterial PO2 . However when gas exchange is impaired, the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 20)03-May-05 17:04:07

Ovid: ICU Book contri ution of the alveolar PO2 declines and the contri ution of the mixed veno us PO2 rises (17). If gas exchange declines to zero, the venous PO2 is the sole determinant of the arterial PO2. P.348 The diagram in Figure 21.5 illustrates ho w hypoxemia can result from either a gas exchange a normality or a low mixed ven ous PO2. The graph shows the transition from venous PO2 to arterial PO2 in norma l, increased shunt fraction, and low venous PO2 conditions. The slope of the tra nsition curve is determined y the efficiency of oxygen exchange from alveoli to capillary lood. A normal gas exchange therefore reduces the slope of the curve , as indicated y the curve for increased intrapulmonary shunt. When the mixed v enous PO2 is reduced and the slope of the transition curve is normal (i.e., norm al gas exchange), the arterial PO2 is decreased to the same degree as seen when intrapulmonary shunt is increased. This illustrates how a low mixed venous PO2 c an e the source of decreased arterial PO2. Figure 21.5. Influence of ventilationperfusion a normality (shunt) and low mixed venous PO2 (PvO2) on transition from venous to arterial PO2. Shunt Fraction The influence of the mixed venous PO2 on the arterial PO2 is determined y the d egree of intrapulmonary shunt. In the normal lung, decreases in venous PO2 have relatively little effect on the arterial PO2. However, as shunt fraction increas es, changes in venous PO2 egin to affect the arterial PO2. If shunt fraction is increased to 100%, the venous PO2 is the sole determinant of the arterial PO2. Thus, in pulmonary conditions associated with a high shunt fraction, such as pul monary edema or pneumonia, the mixed venous PO2 is an important consideration in the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 20)03-May-05 17:04:07

Ovid: ICU Book evaluation of hypoxemia (5,8,17). DIAGNOSTIC EVALUATION The search for the source of hypoxemia in the individual patient can proceed acc ording to the flow diagram in Figure 21.6. This approach P.349 requires a measur ement of the PO2 in superior vena cava or pulmonary artery lood, and thus it ap plies only to patients who have indwelling central venous or pulmonary artery ca theters. Figure 21.6. Flow diagram for the evaluation of hypoxemia. Step 1: A-a PO2 Gradient The first step in the approach involves a determination of the A-a po2 gradient. After correcting for age and FIO2, the A-a po2 gradient can e interpreted as f ollows: Normal A-a po2: Indicates a generalized hypoventilation disorder rather than a cardiopulmonary disorder. In this situation, the most likely pro lems are drug-induced respiratory depression and neuromuscular weakness. The latter cond ition can e uncovered y measuring the maximum inspiratory pressure (PImax). Th is measurement is descri ed in the upcoming section on hypercapnia. Increased Aa po2: Indicates a V/Q a normality (cardiopulmonary disorder) and/or a systemic DO2/ http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 20)03-May-05 17:04:07

Ovid: ICU Book VO2 im alance. When the A-a po2 gradient is increased, a mixed venous PO2 (or ce ntral venous PO2) is needed to identify a systemic DO2/VO2 im alance. Step 2: Mixed Venous PO2 When the A-a gradient is increased, o tain a lood sample from the distal port o f a pulmonary artery catheter or from a central venous catheter (superior vena c ava lood). The PO2 in either lood sample can e interpreted as follows: P.350 Normal venous PO2: Indicates that the pro lem is solely a V/Q a normality in the lungs. If the venous PO2 is 40 mm Hg or higher, then the lungs may e the sourc e of the hypoxemia. If a chest x-ray examination is unrevealing, an acute pulmon ary em olism should e considered. Low venous PO2: Indicates a systemic DO2/VO2 im alance (i.e., a low DO2 or a high VO2). A mixed venous PO2 elow 40 mm Hg ind icates either a decreased rate of oxygen delivery (anemia, low cardiac output) o r an increased rate of oxygen consumption (hypermeta olism). Thus, the approach outlined a ove and shown in Figure 21.7 uses three varia les (A-a po2 gradient, PImax, and PvO2) to pinpoint the source of hypoxemia. Although this approach doe s not identify the disease process, it helps focus the search for the responsi l e illness. Figure 21.7. Flow diagram for the evaluation of hypercapnia. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 20)03-May-05 17:04:07

Ovid: ICU Book HYPERCAPNIA The evaluation of hypercapnia (PaCO2 a ove 46 mm Hg) proceeds in a very similar fashion to the hypoxemia evaluation. Before eginning the evaluation, the increa se in arterial PCO2 must not e a compensatory response to meta olic alkalosis ( see Chapter 36). If this is not the case, then proceed as descri ed elow. CAUSES OF HYPERCAPNIA The car on dioxide level in arterial lood (PaCO2) is directly proportional to t he rate of CO2 production y oxidative meta olism (VCO2) and inversely proportio nal to the rate of CO2 elimination y alveolar P.351 ventilation (VA) (5,18). Th erefore, PaCO2 = k (VCO2/VA), where k is the proportionality constant. Alveolar ventilation is the fraction of the total expired ventilation (VE) that is not de ad space ventilation (Vd/Vt); that is, VA = VE (1 Vd/Vt). Therefore, the a ove r elationship can e rewritten as follows: This equation identifies three major sources of hypercapnia: (a) increased CO2 p roduction (VCO2), ( ) hypoventilation (VE), and (c) increased dead space ventila tion (Vd/Vt). Increased CO2 production (e.g., from hypermeta olism) is normally accompanied y an increase in minute ventilation. The ventilatory response serve s to eliminate the excess CO2 and maintain a constant arterial PCO2. Therefore, excess CO2 production does not normally cause hypercapnia. However when CO2 excr etion is impaired y an increase in dead space ventilation, an increase in CO2 p roduction can result in an increase in the arterial PCO2. Thus, increased CO2 pr oduction is an important factor in promoting hypercapnia only in patients with u nderlying lung disease. DIAGNOSTIC EVALUATION The edside evaluation of hypercapnia is strikingly similar to the evaluation of hypoxemia. The flow diagram for the evaluation is shown in Figure 21.7 (note th e similarities etween Figure 21.6 and Figure 21.7). As was the case with hypoxe mia, the evaluation egins with the A-a po2 gradient (19). An increased A-a grad ient indicates an increase in dead space ventilation (i.e., a pulmonary disorder ), possi ly complicated y an increase in CO2 production. A normal or unchanged A-a po2 gradient indicates that the pro lem is alveolar hypoventilation. CO2 Production http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 20)03-May-05 17:04:07

Ovid: ICU Book The rate of CO2 production (VCO2) can e measured at the edside with specialize d meta olic carts that are normally used to perform nutritional assessments. The se instruments use infrared light to measure the CO2 in expired gas (much like t he end-tidal CO2 monitors descri ed in Chapter 22), and they provide a measure o f total CO2 excreted per minute. In steady-state conditions, the rate of CO2 exc retion is equivalent to the VCO2. The normal VCO2 is 90 to 130 L/minute/m2, whic h is roughly 80% of the VO2 (see Ta le 2.4). An increased VCO2 is evidence for o ne of the following a normalities: generalized hypermeta olism, overfeeding (exc ess calories, or organic acidoses. TABLE 21.4. ALVEOLAR HYPOVENTILATION IN THE ICU Overfeeding, or the provision of calories in excess of daily needs, is a recogni zed cause of hypercapnia in patients with severe lung disease and acute respirat ory failure (20). Nutritionassociated hypercapnia has een reported predominantl y in ventilator-dependent patients and can delay weaning from mechanical ventila tion (22). Thus, overfeeding P.352 should e considered as a possi le cause of C O2 retention in any patient in the ICU with respiratory impairment, and particul arly in patients who require mechanical ventilation. This possi ility can e inv estigated y asking the nutrition support team to measure the VCO2 at the edsid e. ALVEOLAR HYPOVENTILATION The causes of alveolar hypoventilation that are most likely to e encountered in the ICU are listed in Ta le 21.4. The most common causes in the ICU are drug-in duced respiratory depression and neuromuscular weakness. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 20)03-May-05 17:04:07

Ovid: ICU Book Respiratory Muscle Weakness Possi le causes of neuromuscular weakness in patients in the ICU include shock, multiorgan failure, prolonged neuromuscular lockade, electrolyte a normalities, and cardiac surgery (phrenic nerve injury). Critical illness itself may damage peripheral nerves and produce a syndrome known as critical illness polyneuropath y (21) that can e associated with profound generalized weakness and delayed wea ning from mechanical ventilation. This and other sources of neuromuscular weakne ss in patients in the ICU are descri ed in Chapter 51. The standard measurement for evaluating respiratory muscle strength is the maximum inspiratory pressure ( PImax), which is o tained y having the patient make a maximum inspiratory effor t from functional residual capacity (FRC) against a closed valve. The normal PIm ax depends on oth age and sex, and it can vary widely in individual patients. H owever, most healthy adults have a PImax a ove 80 cm H2O (22). Car on dioxide re tention develops when the PImaxfalls to less than 40% of normal values (23). Central Hypoventilation Syndromes Hypoventilation without evidence of respiratory muscle weakness indicates possi le drug-induced respiratory depression. Opiates (and P.353 enzodiazepine in the elderly) are the most likely offenders. Hypoventilation without apparent cause in an o ese patient may e the result of a chronic and poorly understood disorde r known as o esityhypoventilation syndrome (OHS). Hypoventilation without appare nt cause in a lean patient, known as primary alveolar hypoventilation, is rare. REFERENCES GENERAL TEXTS Grippi MA , ed. Pulmonary pathophysiology. Philadelphia: JB Lippincott, 1995. Sutton JR, Coates G, Remmers JE, eds. Hypoxia: the adaptations. Philadelphia: BC Decker, 1990. West JB. Ventilation/ lood flow and gas exchange. 5th ed. Philadelphia: JB Lippi ncott, 1990. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 20)03-May-05 17:04:07

Ovid: ICU Book

2. Maukkassa FF, Rutledge R, Fakhry SM, et al. ABGs and arterial lines: the rela tionship to unnecessarily drawn arterial lood gas samples. J Trauma 1990;30:108 71095. 3. Shapiro BA, Mahutte CK, Cane RD, Gilmour IJ. Clinical performance of an arter ial lood gas monitor. Crit Care Med 1993;21:487494. 4. Zimmerman JL, Dellinger RP. Initial evaluation of a new intra-arterial lood gas system in humans. Crit Care Med 1993;21:495500. PULMONARY GAS EXCHANGE 5. Dantzger DR. Pulmonary gas exchange. In: Dantzger DR, ed. Cardiopulmonary cri tical care. 2nd ed. Philadelphia: WB Saunders, 1991;2543. 6. Lanken PN. Ventilation-perfusion relationships. In: Grippi MA, ed. Pulmonary pathophysiology. Philadelphia: JB Lippincott, 1995;195210. 7. Buohuys A. Respiratory dead space. In: Fenn WO, Rahn H, eds. Hand ook of phys iology: respiration. Bethesda: American Physiological Society, 1964;699714. 8. D'Alonzo GE, Dantzger DR. Mechanisms of a normal gas exchange. Med Clin North Am 1983;67:557571. QUANTITATIVE DETERMINATIONS 9. To in MJ. Respiratory monitoring during mechanical ventilation. Crit Care Cli n 1990;6:679 709. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 20)03-May-05 17:04:07

INTRODUCTION 1. Raffin TA. Indications for

lood gas analysis. Ann Intern Med 1986;105:390398.

Zander R, Mertzlufft F, eds. The oxygen status of arterial land: S. Karger Pu lishers, 1991.

lood. Basel, Switzer

Ovid: ICU Book 10. Harris EA, Kenyon AM, Nis et HD, et al. The normal alveolar-arterial oxygen tension gradient in man. Clin Sci 1974;46:89104. P.354 11. Gil ert R, Kreighley JF. The arterial/alveolar oxygen tension ratio. An inde x of gas exchange applica le to varying inspired oxygen concentrations. Am Rev R esp Dis 1974;109:142145. 12. Carroll GC. Misapplication of the alveolar gas equation. N Engl J Med 1985;3 12:586. 13. Covelli HD, Nessan VJ, Tuttle WK. Oxygen derived varia les in acute respirat ory failure. Crit Care Med 1983;11:646649. BLOOD GAS VARIABILITY 14. Hess D, Agarwal NN. Varia ility of lood gases, pulse oximeter saturation, a nd end-tidal car on dioxide pressure in sta le, mechanically ventilated trauma p atients. J Clin Monit 1992;8:111115. 15. Sasse SA, Chen P, Mahutte CK. Varia ility of arterial lood gas values over time in sta le medical ICU patients. Chest 1994;106:187193. HYPOXEMIA 16. Kreimeier U, Mesmer K. The differential diagnosis of arterial hypoxemia. In: Zander R, Mertzlufft F, eds. The oxygen status of arterial lood. Basel, Switze rland: S. Karger Pu lishers, 1991;196202. 17. Rossaint R, Hahn S-M, Pappert D, et al. Influence of mixed venous PO2 and in spired oxygen fraction on intrapulmonary shunt in patients with severe ARDS. J A ppl Physiol. 1995;78:15311536. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 20)03-May-05 17:04:07

Ovid: ICU Book HYPERCAPNIA 18. Wein erger SE, Schwartzstein RM, Weiss JW. Hypercapnia. N Engl J Med 1989;32 1:12231230. 19. Gray BA, Blalock JM. Interpretation of the alveolar-arterial oxygen differen ce in patients with hypercapnia. Am Rev Respir Dis 1991;143:48. 20. Talpers SS, Rom erger DJ, Bunce SB, Pingleton SK. Nutritionally associated i ncreased car on dioxide production. Chest 1992;102:551555. 21. Wijdicks EFM, Litchy WJ, Harrison BA, Gracey DR. The clinical spectrum of cr itical illness polyneuropathy. Mayo Clin Proc 1994;69:955959. 22. Bruschi C, Cerveri I, Zoia MC, et al. Reference values for maximum respirato ry mouth pressures: A population- ased study. Am Rev Respir Dis 1992;146:790793. 23. Baydur A. Respiratory muscle strength and control of ventilation in patients with neuromuscular disease. Chest 1991;99:330338. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 20)03-May-05 17:04:07

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 22 OXIMETRY AND CAPNOGRAPHY The noninvasive detection of arterial lood gases using optical and colorimetric techniques (1,2,3,4,5 and 6) is the most significant advance in critical care m onitoring in the last quartercentury. This chapter descri es the techniques that have ecome an integral part of daily patient care in the ICU. Despite their ro utine use, surveys indicate that 95% of staff physicians and staff nurses workin g in ICUs have little or no understanding of how these techniques work (7). DEFINITIONS All atoms and molecules reflect specific wavelengths of light (this is the sourc e of color in the lighted world). Spectrophotometry is an optical detection meth od that uses the light reflection properties of molecules to measure the concent ration of chemical species in a gaseous or liquid environment. When spectrophoto metry is applied to the detection of oxygenated and deoxygenated hemoglo in, the method is known as oximetry. The optical detection of car on dioxide is known a s capnometry. The term capnography refers to the optically recorded pattern of C O2 excretion in single- reath exhalations. OXIMETRY Hemoglo in (like all proteins) changes its structural configuration when it part icipates in a chemical reaction, and each of the configurations has a different pattern of light reflection. The patterns of light reflection associated with ox ygenated hemoglo in (H O2) and deoxygenated hemoglo in (H ) are shown in Figure 22.1. At wavelengths of 660 nanometers (nm), which corresponds to the red region of the light spectrum, oxygenated hemoglo in (H O2) reflects light more effecti vely than deoxygenated hemoglo in (H ). (This explains why oxygenated lood is m ore intensely red than deoxygenated lood.) P.356 This relationship is reversed at 940 nm (the infrared spectrum), where H reflects light more effectively than H O2. Thus, when oth wavelengths of light are passed through a sample of lood , the intensity of light transmission at 660 nm is primarily a function of the c oncentration of H O2 in the sample, whereas the transmission at 940 nm is determ ined primarily y the concentration of H . http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 20)03-May-05 17:04:44

Ovid: ICU Book The concentrations of H O2 and H are expressed in relative terms, that is, as t he fraction of hemoglo in that is in the oxygenated form. This is known as the p ercent oxyhemoglo in saturation (% saturation), and is derived as follows: Figure 22.1. The pattern of light reflection y oxygenated hemoglo in (H O2) and deoxygenated hemoglo in (H ). The vertical lines represent the two wavelengths of light (660 nm and 940 nm) used y pulse oximeters. (From the Ohmeda Biox 3700 Pulse Oximeter Operating and Maintenance Manual. Louisville, CO: BOC Health Car e, 1988;12.) Limitations The use of two wavelengths of light to derive the fractional concentration of H O2 is ased on the assumption that other forms of hemoglo in such as methemoglo in (metH ) and car oxyhemoglo in (COH ) have a negligi le contri ution to the to tal hemoglo in pool. In most situations, less than 5% of hemoglo in is present a s COH and metH (3,4,6), so the assumption is valid. However, in conditions P.3 57 associated with an increase in COH (e.g., smoke inhalation) or metH (e.g., high-dose nitroglycerin), the exclusion of these forms of hemoglo in leads to fa lsely high estimates of the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 20)03-May-05 17:04:44

Ovid: ICU Book prevalence of H O2 (% saturation). TYPES OF OXIMETRY Oximetry can e performed oth in vivo and in vitro. There are two in vivo techn iques. One uses pro es placed on the surface of the skin to measure the O2 satur ation in the underlying lood vessels. The other uses a pulmonary artery cathete r to measure the O2 saturation in mixed venous (pulmonary artery) lood. Ear Oximetry Oximeters capa le of continuous, on-line monitoring of arterial oxygen saturatio n (SaO2) at the edside were introduced in the 1960s. The original devices were placed on the earlo es. The oximeter pro e consisted of a phototransmitter situa ted on one side of the earlo e that emitted red and infrared light waves, and a photodetector situated on the opposite skin surface to record light transmission through the intervening tissue. These early oximeters suffered from two major l imitations. First, the transmission of light was affected y factors other than hemoglo in, such as earlo e thickness and skin pigments. The second pro lem was the ina ility to differentiate etween hemoglo in in arteries and veins. Pulse Oximetry The pro lems encountered with the early ear oximeters were largely eliminated wi th the introduction of pulse oximetry in the mid-1970s. The principle of pulse o ximetry is illustrated in Figure 22.2 (3,4,5 and 6). Arterial pulsations are ass ociated with changes in lood volume that produce phasic changes in the intensit y of transmitted light. The photodetectors in pulse oximeters are designed to se nse only light of alternating intensity (analogous to AC amplifiers, which proce ss only alternatingcurrent electrical impulses). This eliminates errors created y light reflection in nonpulsatile structures such as extravascular tissues and (nonpulsating) veins. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 20)03-May-05 17:04:44

Ovid: ICU Book Figure 22.2. The principle of pulse oximetry. The photodetector senses only ligh t of alternating intensity (analogous to an AC amplifier). Co-Oximeters In vitro oximetry is performed with instruments called co-oximeters that transmi t four wavelengths of light through a lood sample. These devices are capa le of detecting methemoglo in and car oxyhemoglo in (in addition to H and H O2), ut do not provide the continuous monitoring availa le with pulse oximeters. P.358 Mixed Venous Oximetry The O2 saturation in mixed venous (pulmonary artery) lood can e measured conti nuously with specialized pulmonary artery catheters that emit red and infrared l ight from the catheter tip and record the light reflected ack from the hemoglo in in circulating erythrocytes. This technique is called reflectance spectrophot ometry, whereas pulse oximeters and co-oximeters use transmission spectrophotome try. PULSE OXIMETRY As mentioned previously, pulse oximeters record light transmission through pulsa ting arteries only. Like the early ear oximeters, pulse oximeters use two wavele ngths of light, one in the red spectrum (660 nm) and the other in the infrared s pectrum (940 nm). Pulse oximetry pro es are http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 20)03-May-05 17:04:44

Ovid: ICU Book usually placed on the fingers. Accuracy At clinically accepta le levels of arterial oxygenation (SaO2 a ove 70%), the O2 saturation recorded y pulse oximeters (SpO2) differs y less than 3% from the actual SaO2 (measured y cooximeters) (3,4,5 and 6). SpO2 also shows a high degr ee of precision (consistency of repeated measurements). This is demonstrated in Ta le 22.1, which shows the spontaneous variation in SpO2 over a 1-hour period i n ventilator-dependent patients who appeared clinically sta le (8). The variatio n in P.359 SpO2 was 2% or less in 95% of cases, indicating that there is little random varia ility (or random error) in pulse oximeter recordings. TABLE 22.1. VARIABILITY IN OXIMETRY AND CAPNOMETRY RECORDINGS COMMON CONCERNS ABOUT SpO2 A num er of conditions are cited as sources of erroneous pulse oximetry recordin gs, and some of these are listed in Ta le 22.2. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 20)03-May-05 17:04:44

Ovid: ICU Book TABLE 22.2. ACCURACY OF PULSE OXIMETRY RECORDING (SpO2) Dyshemoglo inemias As mentioned, oximeters that use two wavelengths of light are not relia le when hemoglo in forms other than H O2 and H are elevated. Car oxyhemoglo in (COH ) r eflects as much red light as H O2, which explains the cherry red color of lood in car on monoxide intoxication. Pulse oximeters therefore record COH as H O2, and SpO2 then overestimates the actual SaO2 (3,4,6). The same error occurs with high levels of methemoglo in (metH ): SpO2 rarely falls elow 85% in methemoglo inemia despite much lower levels of SaO2 (6). Thus, pulse oximetry P.360 should not e used when methemoglo inemia or car on monoxide intoxication is suspected. Accurate measurements of COH , metH , and SaO2 are provided only y fourwavelen gth co-oximeters. Hypotension Although pulse oximetry is ased on the presence of pulsatile lood flow, SpO2 i s an accurate reflection of SaO2 down to lood pressures as low as 30 mm Hg (9). Damped pulsations also do not affect the accuracy of fingertip SpO2 recordings when the radial artery is cannulated (10). Anemia In the a sence of hypoxemia, pulse oximetry is accurate down to hemoglo in level s as low as 2 to 3 g/dL (11). With lesser degrees of anemia (H etween 2.5 and 9 g/dL), SpO2 underestimates SaO2 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 20)03-May-05 17:04:44

Ovid: ICU Book y only 0.5% (11).

WHEN TO USE PULSE OXIMETRY Pulse oximetry offers several advantages over conventional arterial lood gases for monitoring arterial O2 saturation: More accurate determination of arterial O 2 saturation (15) Superior detection of hypoxemic episodes (6) Noninvasive, less involved Less mor idity, greater patient satisfaction Less expensive (16) Pulse oximetry is thus superior to lood gases in every aspect of performance. In fac t, SaO2 from arterial lood gases is not a measured varia le, ut rather is deri ved from a nomogram, and this derived parameter has een shown to e much less a ccurate than the measurement provided y pulse oximetry (15). The value of pulse oximetry P.361 has een recognized y the American Society of Anesthesiologists in a formal statement recommending pulse oximetry as a routine measure during a nd immediately after general anesthesia (17,18). A similar recommendation seems warranted for all patients receiving supplemental oxygen in the ICU. Detection of Hypoxemia At least 15 clinical studies attest to the superiority of pulse oximetry over pe riodic lood gas measurements for detecting episodes of significant hypoxemia in critically ill patients (6). However, these studies also reveal that the enhanc ed detection of hypoxemic episodes with pulse http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 20)03-May-05 17:04:44

Pigments Skin pigmentation ( iliru in or melanin) usually does not affect the accuracy of pulse oximetry. However, SpO2 can e spuriously low in patients with very dark skin (12). Fingernail polish containing lue and lack colors can result in spur ious decreases of 3 to 5% in SpO2 (6,13). Onychomycosis (a fungal infection of t he fingernails) can result in a similar 3- to 5% decrease in SpO2 relative to Sa O2 (14). The largest pigment effect is produced y methylene lue, which can pro duce a 65% decrease in SpO2 when injected intravenously (6). Because methylene lue is used to treat methemoglo inemia (see Chapter 18), this is another reason to avoid pulse oximetry in patients with methemoglo inemia.

Ovid: ICU Book oximetry has no impact on either mor idity or mortality (6). This is not the fau lt of pulse oximetry, ut may e the fault of those who continually emphasize th e largely undocumented dangers of hypoxemia (discussed in more detail in Chapter 24). Limitation One limitation of pulse oximetry that deserves mention is the insensitivity of t he SaO2 for detecting a normalities in pulmonary gas exchange (19). This is expl ained y the shape of the oxyhemoglo in dissociation curve. When SaO2 exceeds 90 % (arterial oxygen pressure [PAO2] a ove 60 mm Hg), the curve is flat, and large changes in PAO2 are associated with small changes in SaO2. Thus, SaO2 is not a sensitive marker of changes in pulmonary gas exchange in the range where SaO2 is usually maintained. MIXED VENOUS OXIMETRY The continuous measurement of O2 saturation in mixed venous (pulmonary artery) lood is performed with specialized PA catheters (manufactured y Baxter Health C are, Santa Ana, CA and A ot Critical Care, North Chicago, IL) equipped with fi eroptic undles that can transmit light to and from the catheter tip. The optica l detection of mixed venous O2 saturation (SvO2) is performed y reflection spec trophotometry. Wavelengths of light similar to those used in pulse oximetry are passed along fi eroptic undles in the PA catheter and out from the catheter tip . The light eam is transmitted through the circulating lood, and the light tha t comes in contact with hemoglo in in the circulating erythrocytes is reflected ack to the catheter tip. This light is then transmitted ack through the cathet er to a photodetector and microprocessor that record the average SvO2 at 5second intervals.

Relia ility SvO2 measured y PA catheters is within 1.5% of the SvO2 measured y co-oximeter s (the latter eing the gold standard) (20). Although P.362 accuracy seems accep ta le, SvO2 can vary considera ly without a change in hemodynamic status. The sp ontaneous varia ility of SvO2 is shown in Ta le 22.1. The average variation over a 2-hour period is 6%, ut it can e as high as 19% (21). For practical purpose s, a greater than 5% variation in SvO2 that persists for longer than 10 minutes is considered a significant change (22). SvO2 is normally etween 65 and 75% (21 ). Clinical Uses http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 20)03-May-05 17:04:44

Ovid: ICU Book SvO2 is a marker of the alance etween (whole ody) O2 delivery and O2 consumpt ion; that is, SvO2 = DO2/VO2. If DO2 and VO2 are roken down into their componen t parts (see Chapter 2), the determinants of SvO2 can e expressed as follows: (where Q is cardiac output and H is the hemoglo in concentration in whole lood ). Thus, SvO2 varies in the same direction with changes in cardiac output, hemog lo in, and arterial O2 saturation, and varies in the opposite direction to chang es in VO2. A change in SvO2 thus signals a change in any one or more of the four varia les in the SvO2 equation. Identifying the culprit then requires the appro priate measurements. Thus, continuous monitoring of SvO2 provides a general scre ening method for monitoring a group of four varia les (Q, H , SaO2, and VO2). A change in the SvO2 cannot e interpreted in isolation, ut should serve as a tri gger to perform other measurements to identify the pro lem. Dual Oximetry The predictive value of SvO2 can e increased y adding the SpO2 (SaO2) measurem ent provided y pulse oximetry. This creates a continuous measure of whole ody O2 extraction (i.e., SaO2 SvO2), known as dual oximetry (23). The SvO2 equation can e rearranged as follows to define the determinants of whole ody O2 extract ion: Thus, SaO2SvO2 changes in the same direction as the changes in the meta olic rate (VO2), and changes in the opposite direction to the changes in cardiac output a nd hemoglo in. This is the asis for the interpretations shown in Ta le 22.3. Fo r more on O2 extraction, see Chapter 2 and Chapter 11. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 20)03-May-05 17:04:44

Ovid: ICU Book TABLE 22.3. DUAL OXIMETRY COLORIMETRIC CO2 DETECTOR The colorimetric detection of CO2 in exhaled gas with a disposa le device, such as the one shown in Figure 22.3, has ecome a standard P.363 method for determin ing the success or failure of endotracheal intu ation (24,25). The central ovoid area of the device contains filter paper impregnated with a pH-sensitive indica tor that changes color as a function of pH. When exhaled gas passes over the fil ter paper, the CO2 in the gas is hydrated y a liquid film on the filter paper, and the resulting pH is detected y a color change. The outer perimeter of the d evice contains color-coded areas that indicate different concentrations of CO2. The a sence of CO2 is indicated y a purple color (the color of the check area on the device), and a yellow color indicates a CO2 concentration in excess of 2% (a rea C on the device). Thus, when the device is attached to an endotracheal tu e immediately after intu ation, a yellow color in the central area indicates succe ssful tu e placement in the upper airways, whereas a purple color indicates that the tu e is pro a ly in the esophagus (the pro a ly will e explained shortly). Figure 22.3. A disposa le device (Easy Cap, Nellcor Puritan Bennett) for the col orimetric detection of CO2 in exhaled gas. PREDICTIVE VALUE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 20)03-May-05 17:04:44

Ovid: ICU Book The accuracy of this colorimetric device for predicting the success of endotrach eal intu ation is shown in Ta le 22.4 (24). For patients who are not in cardiac arrest, failure of the central area to turn yellow always indicates that the tu e is not in the airways (i.e., sensitivity of P.364 the color change = 100%). Ho wever, in the setting of a cardiac arrest, the CO2 in exhaled gas may e too low to produce a yellow color change (the sensitivity is only 72%). Therefore, for intu ations during cardiac arrest, the lack of an appropriate color change (to y ellow) does not always indicate improper tu e placement. TABLE 22.4. PREDICTIVE VALUE OF COLORIMETRIC CO2 DETECTION FOR ENDOTRACHEAL INTU BATION According to Ta le 22.4, a color change to yellow does not always indicate prope r endotracheal tu e placement. Although gas in the esophagus usually has negligi le amounts of CO2 ( ecause it is am ient air that is swallowed), the CO2 in eso phageal gas can e elevated following either mouthto-mouth resuscitation (i.e., exhaled gas insufflation) or the ingestion of car onated everages. This CO2 usu ally clears after a few inflations, so at least 4 inflations should e performed efore checking the exhaled CO2. INFRARED CAPNOMETRY Infrared light a sorption provides a more quantitative measure of exhaled CO2 th an colorimetric methods (26,27). Infrared CO2 analyzers can e placed along the expiratory lim of ventilator tu ing, as depicted in Figure 22.4. A light-emitti ng diode generates a steady eam of infrared light that passes across the stream of exhaled gas, and a photodetector on the other side measures the intensity of light that is transmitted. The intensity of light transmission is inversely rel ated to the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 20)03-May-05 17:04:44

Ovid: ICU Book concentration of CO2 in the intervening gas. Figure 22.4. Infrared device for monitoring the CO2 in exhaled gas. The capnogra m shows the normal pattern of CO2 elimination in a single reath. CAPNOGRAM Infrared CO2 detectors have a rapid response and can measure changes in CO2 duri ng a single exhalation. These changes are recorded as a waveform called a capnog ram, which normally appears like the one shown in Figure 22.4. The shape of the normal capnogram has een descri ed as the outline of a snake that has swallowed an elephant (28). The PCO2 at the onset of exhalation is negligi le ecause the gas in the upper airways is first to leave the lungs. As exhalation proceeds, g as from the alveoli egins to contri ute to the exhaled gas, and the PCO2 egins to rise steadily. Near the end of exhalation, the PCO2 reaches a plateau and re mains there until the onset of inhalation. When gas exchange in the lungs is nor mal, the PCO2 at the end of exhalation is equivalent to the PCO2 in end-capillar y (arterial) lood. The end-expiratory PCO2 is more commonly called end-tidal PC O2 (PETCO2). P.365 PaCO2 VERSUS PETCO2 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 20)03-May-05 17:04:44

Ovid: ICU Book Normally, the difference etween the arterial and end-tidal PCO2 (called the PaC O2PETCO2 gradient) is less than 5 mm Hg (26,27). However, when gas exchange in th e lungs is impaired, PETCO2 decreases relative to PaCO2, so the PaCO2PETCO2 gradi ent increases. This occurs in the following conditions: Increased anatomic dead space Open ventilator circuit Shallow reathing Increased physiologic dead space O structive lung disease Low cardiac output Pulmonary em olism Excessive lung i nflation (e.g., PEEP) Virtually any pulmonary disorder, or any cardiac disorder where cardiac output is reduced, produces an increase in the PaCO2PETCO2 gradient . Thus, in patients with cardiopulmonary disease, a lood gas measurement is nec essary at the outset to determine the equivalence of PaCO2 and PETCO2. If there is a PaCO2PETCO2 gradient, changes in end-tidal CO2 can e monitored ecause in s teady-state conditions, changes in PETCO2 should e equivalent to changes in PaC O2 (i.e., the PaCO2PETCO2 gradient will not change). However, any pertur ation th at affects gas exchange (e.g., changing ventilator settings) can change the PaCO 2PETCO2 gradient (29), so when a change in gas exchange is P.366 suspected, anoth er lood gas is needed to determine the PaCO2PETCO2 gradient. PETCO2 Higher Than PaCO2 Although uncommon, end-tidal CO2 can e higher than arterial PCO2 in the followi ng conditions (27,30): Excessive CO2 production, plus Low inspired volume High c ardiac output High inspired O2 (CO2 displaced from H ) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 20)03-May-05 17:04:44

Ovid: ICU Book NONINTUBATED PATIENTS Although usually reserved for intu ated patients, end-tidal CO2 monitoring is po ssi le in nonintu ated patients as well (31,32 and 33). Nasal prongs adapted for end-tidal CO2 monitoring are commercially availa le (Datascope Corp., Paramus, NJ: Salter La s, Arvin, CA), and Figure 22.5 shows how to modify a nasal cannula for exhaled CO2 monitoring (31). The trick is to occlude the tu ing etween the two nasal prongs (either with a cotton all inserted through one of the nasal p rongs or with a small screw clamp). This allows one nasal prong to e used for o xygen inhalation while the other nasal prong is used to transmit exhaled gas. A 14-gauge intravascular catheter (2 inches long) is inserted into the exhalation side of the nasal prong apparatus to transmit gas to the CO2 detector. A sidestr eam CO2 detector (i.e., one that applies suction to draw gas from the tu ing) is est suited for this application. If one of these is not availa le, a mainstrea m CO2 detector (such as the one shown in Fig. 22.3) can e used with a suction P .367 pump to draw gas samples from the cannula (at 150 mL/minute). The respirato ry therapy department should help with this adaptation. Figure 22.5. Nasal cannula modification for end-tidal CO2 monitoring. CLINICAL APPLICATIONS Infrared capnometry has several uses in the ICU. The following are some situatio ns where it can prove valua le. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 20)03-May-05 17:04:44

Ovid: ICU Book Cardiac Output Monitor As mentioned in Chapter 14 and Chapter 17, changes in PETCO2 can e used to moni tor changes in cardiac output during volume resuscitation in hypovolemic shock ( Fig. 14.1) and during cardiopulmonary resuscitation (Fig. 17.4). The close corre lation etween changes in PETCO2 and changes in cardiac output is demonstrated i n Figure 22.6 (34). Figure 22.6. Correlation etween changes in end-tidal PCO2 and changes in cardia c output. Recordings from nine patients undergoing a dominal aortic aneurysm sur gery. Correlation coefficient indicated as r. (From Shi utani K et al. Changes i n cardiac output affect PETCO2, CO2 transport, and O2 uptake during unsteady sta te in humans. J Clin Monit 1992;8:175.) Because of the a ility to monitor trends in cardiac output, end-tidal CO2 monito ring has several potential applications in the ICU (e.g., monitoring the effects of weaning intraaortic alloon counterpulsation). One of the more important app lications is the a ility of PETCO2 to predict the outcome of cardiopulmonary res uscitation (see Chapter 17). P.368 Ventilator-Related Mishaps Infrared CO2 monitors are equipped with alarms that can signal the separation of a patient from the ventilator. In this situation, the ventilator disconnection causes an a rupt drop in exhaled CO2, causing the alarm to trigger (35). A drop in PETCO2 can also signal migration of the endotracheal http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 20)03-May-05 17:04:44

Ovid: ICU Book tu e into a mainstem ronchus (35).

Ventilator Weaning During weaning from mechanical ventilation, end-tidal CO2 monitoring can serve s everal purposes (36). In uneventful weaning (e.g., following surgery), it serves as a noninvasive measure of PaCO2. In difficult or complicated weaning, it can help determine the success or failure of the wean attempt. For example, a progre ssive rise in PETCO2 can signal an increased work of reathing (a sign of wean f ailure), whereas a decline in PETCO2 with a rising PaCO2PETCO2 gradient can signa l respiratory muscle weakness with shallow reathing (another sign of wean failu re). Controlled Hyperventilation When hyperventilation is used to reduce intracranial pressure in patients with h ead injuries, monitoring end-tidal CO2 helps maintain the target PaCO2. In this setting, the PaCO2PETCO2 gradient must e checked periodically to select the appr opriate target PETCO2. REFERENCES GENERAL TEXTS Gravenstein JS, Paulus DA, Hayes TJ. Capnography in clinical practice. Boston: B utterworthHeinemann, 1989. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 20)03-May-05 17:04:44

Early Detection of Nosocomial Disorders A decrease in PETCO2 accompanied y an increase in the PaCO2PETCO2 gradient can e an early manifestation of any of the following conditions (26,27): Acute pulmo nary em olism Atelectasis Low cardiac output Pneumonia Pulmonary edema An increa se in the PaCO2PETCO2 gradient can thus serve as an early warning signal for any of these conditions. This can e particularly valua le in ventilator-dependent p atients, who are at risk for nosocomial complications.

Ovid: ICU Book P.369 GENERAL REVIEWS 1. To in MJ. Respiratory monitoring. JAMA 1990;264:244251 (122 references).

PULSE OXIMETRY: REVIEWS 3. Tremper KK, Barker SJ. Pulse oximetry. Anesthesiology 1989;70:98108 (59 refere nces). 4. Severinghaus JW, Kelleher JF. Recent developments in pulse oximetry. Anesthes iology 1992;76:10181038 (249 references). 5. Council on Scientific Affairs, American Medical Association. The use of pulse oximetry during conscious sedation. JAMA 1993;270:14631468 (69 references). 6. Wahr JA, Tremper KK. Noninvasive oxygen monitoring techniques. Crit Care Clin 1995;11:199217 (110 references). PULSE OXIMETRY: SELECTED REFERENCES 7. Stoneham MD, Saville GM, Wilson IH. Knowledge a out pulse oximetry among medi cal and nursing staff. Lancet 1994;344:13391342. 8. Hess D, Agarwal NN. Varia ility of lood gases, pulse oximeter saturation, an d end-tidal car on dioxide pressure in sta le, mechanically ventilated trauma pa tients. J Clin Monit 1992;8:111115. 9. Severinghaus JW, Spellman MJ. Pulse oximeter failure thresholds in hypotensio n and vasoconstriction. Anesthesiology 1990;73:532537. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 20)03-May-05 17:04:44

2. Clark JS, Votteri B, Ariagno RL, et al. Noninvasive assessment of . Am Rev Respir Dis 1992;145:220232 (149 references).

lood gases

Ovid: ICU Book 10. Morris RW, Nairn M, Beaudoin M. Does the radial arterial line degrade the pe rformance of a pulse oximeter? Anesth Intensive Care 1990;18:107109. 11. Jay GD, Hughes L, Renzi FP. Pulse oximetry is accurate in acute anemia from hemorrhage. Ann Emerg Med 1994;24:3235. 12. Ries AL, Prewitt LM, Johnson JJ. Skin color and ear oximetry. Chest 1989;96: 287290. 13. Ru in AS. Nail polish color can affect pulse oximeter saturation. Anesthesio logy 1988;68:825. 14. Ezri T, Szmuk P. Pulse oximeters and onychomycosis. Anesthesiology 1992;76:1 53154. 15. Johnson PA, Bihari DJ, Raper RF, et al. A comparison etween direct and calc ulated oxygen saturation in intensive care. Anesth Intensive Care 1993;21:7275. 16. Roizen MF, Schreider B, Austin W, et al. Pulse oximetry, capnography, and l ood gas measurements: reducing cost and improving the quality of care with techn ology. J Clin Monit 1993;9:237240. 17. Standards for intra-operative monitoring. Park Ridge, IL: American Society o f Anesthesiologists, 1991. 18. Standards for post-anesthesia care. Park Ridge, IL: American Society of Anes thesiologists, 1990.

MIXED VENOUS OXIMETRY 20. Armaganidis A, Dhinaut JF, Billard JL, et al. Accuracy assessment for three fi eroptic http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 20)03-May-05 17:04:44

19. Hutton P, Clutton-Brock T. The ed J 1993;307:457458. P.370

enefits and pitfalls of pulse oximetry. Be M

Ovid: ICU Book pulmonary artery catheters for SvO2 monitoring. Intensive Care Med 1994;20:484488 . 21. Noll ML, Fountain RL, Duncan CA, et al. Fluctuation in mixed venous oxygen s aturation in critically ill medical patients: a pilot study. Am J Crit Care 1992 ;3:102106. 22. Krafft P, Steltzer H, Heismay M, et al. Mixed venous oxygen saturation in cr itically ill septic shock patients. Chest 1993;103:900906. 23. Bongard FS, Leighton TA. Continuous dual oximetry in surgical critical care. Ann Surg 1992;216:6068. COLORIMETRIC CO2 DETECTOR 24. MacLeod BA, Heller MB, Gerard J, et al. Verification of endotracheal tu e pl acement with colorimetric end-tidal CO2 detection. Ann Emerg Med 1991;20:267270. 25. Deem S, Bishop MJ. Evaluation and management of the difficult airway. Crit C are Clin 1995;11:127. CAPNOGRAPHY: REVIEWS 26. Szaflarski NL, Cohen NH. Use of capnography in critically ill adults. Heart Lung 1991;20:363374 (45 references). 27. Stock MC. Capnography for adults. Crit Care Clin 1995;11:219232 (32 reference s). CAPNOGRAPHY: SELECTED REFERENCES 28. Gravenstein JS, Paulus DA, Hayes TJ. Capnography in clinical practice. Bosto n Butterworth-Heinemann, 1989;11. 29. Hoffman RA, Kreiger PB, Kramer MR, et al. End-tidal car on dioxide in critic ally ill patients during changes in mechanical ventilation. Am Rev Respir Dis 19 89;140:12651268. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 20)03-May-05 17:04:44

Ovid: ICU Book 30. Moorthy SS, Losasso AM, Wilcox J. End-tidal PCO2 greater than PaCO2. Crit Ca re Med 1984;12:534535. 31. Wright SW. Conscious sedation in the emergency department: the value of capn ography and pulse oximetry. Ann Emerg Med 1992;21:551555. 32. Roy J, McNulty SE, Torjman MC. An improved nasal prong apparatus for end-tid al car on dioxide monitoring in awake, sedated patients. J Clin Monit 1991;7:2492 52. 33. Liu SY, Lee TS, Bongard F. Accuracy of capnography in nonintu ated surgical patients. Chest 1992;102:15121515. 34. Shi utani K, Shirasaki S, Braatz T, et al. Changes in cardiac output affect PETCO2, CO2 transport, and O2 uptake during unsteady state in humans. J Clin Mon it 1992;8:175176.

36. Healey CJ, Fedullo AJ, Swin urne AJ, Wahl GW. Comparison of noninvasive meas urements of car on dioxide tension during weaning from mechanical ventilation. C rit Care Med 1987;15:764767. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 20)03-May-05 17:04:44

35. Gandhi SK, Munshi CA, Bardeen-Henschel A. Capnography for detection of endo ronchial migration of an endotracheal tu e. J Clin Monit 1991;7:3538.

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 23 ACUTE RESPIRATORY DISTRESS SYNDROME Physicians think they do a lot for a patient when they give his disease a name. --Immanuel Kant The condition descri ed in this chapter is considered to e the leading cause of acute respiratory failure in the United States. It was first de scri ed formally in the 1960s and has had several names, including shock lung, w et lung, and leaky-capillary pulmonary edema. The most popular name is adult res piratory distress syndrome (ARDS), recently changed to acute respiratory distres s syndrome ecause it is not limited to adults (1). Despite the propensity for n aming and renaming this condition over the years, there has een little improvem ent in the poor clinical outcomes associated with this condition (2). In fact, A RDS is not really a primary disease, ut is more of a complication that arises w hen other diseases produce a severe and progressive form of systemic inflammator y response (3). After finishing this chapter, look at Chapter 31 (Infection, Inf lammation, and Multiorgan Injury) and note the similarities etween ARDS and the un ridled systemic inflammatory response. PATHOGENESIS The first clinical report of ARDS included 12 patients with diffuse infiltrates on chest roentgenogram and hypoxemia that was resistant to supplemental oxygen ( 4). Seven patients died (mortality 60%), and autopsy findings revealed dense inf iltration of the lungs with leukocytes and proteinaceous material. These microsc opic findings suggested that ARDS was a diffuse inflammatory injury in the lungs . P.372

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 22)03-May-05 17:07:00

INFLAMMATORY INJURY The asic pathology of ARDS is a diffuse inflammatory process that involves lungs. The

oth

Ovid: ICU Book photomicrograph in Figure 23.1 is from a patient in the early stages of ss than 24 hours in duration). The alveolar spaces on oth sides of the y located septum contain erythrocytes, leukocytes, and proteinaceous de the condition P.373 progresses, this exudative material accumulates and ly o literates the alveolar airspaces. ARDS (le centrall ris. As eventual

Figure 23.1. Microscopic changes in the lungs in the early stage of ARDS. The ve rtical column of tissue in the center of the photomicrograph is an interalveolar septum. The alveoli (A) on either side of the septum contain leukocytes (LC), e rythrocytes (EC), pulmonary macrophages (M), and strands of fi rin (F). The lung consolidation in ARDS is elieved to originate from a systemic activati on of circulating neutrophils (5). The activated neutrophils ecome sticky and a dhere to the vascular endothelium in the pulmonary capillaries (6). They then re lease the contents of their cytoplasmic granules (i.e., proteolytic enzymes and toxic oxygen meta olites), and this http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 22)03-May-05 17:07:00

Ovid: ICU Book damages the endothelium and leads to a leaky-capillary type of exudation into th e lung parenchyma. Neutrophils and other inflammatory mediators can thus gain ac cess to the lung parenchyma and carry on the inflammatory process. The inflammat ion then produces the lung injury. Thus, although ARDS is often referred to as a type of pulmonary edema, ARDS is an inflammatory process, not an accumulation o f watery edema fluid. This is an important distinction for the diagnosis and man agement of ARDS. PREDISPOSING CONDITIONS Many conditions predispose to ARDS, and the more common ones are indicated on th e ody map in Figure 23.2. The common feature in these conditions is the activat ion of neutrophils in the pulmonary or systemic circulation. Thus, the condition s that produce a systemic inflammatory response are the ones that predispose to ARDS. The most common predisposing condition is sepsis, which is a systemic infl ammatory response (such as fever or leukocytosis) due to an infection. P.374 The distinction etween inflammation and infection is an important one ecause infl ammation, not infection, produces the lung injury in ARDS. This distinction eco mes important again in Chapter 30 and Chapter 31. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 22)03-May-05 17:07:00

Ovid: ICU Book Figure 23.2. Common conditions that predispose to ARDS. The risk of developing ARDS in some high-risk conditions is shown in Ta le 23.1. The data in this ta le are from a study of 695 patients with predisposing condi tions for ARDS who were admitted to the ICU (7). Overall, one of every four pati ents developed ARDS (see ottom of the ta le), and ARDS was most common in those with sepsis syndrome and those given multiple lood transfusions (defined in th is study as 15 or more units transfused over 24 hours). Note the negative impact of ARDS on survival. For all patients with high-risk conditions, the mortality rate was tripled (from 19 to 62%) y the appearance of ARDS. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 22)03-May-05 17:07:00

Ovid: ICU Book TABLE 23.1. PREDISPOSING FACTORS AND MORTALITY IN ARDS CLINICAL FEATURES The earliest clinical signs of ARDS include tachypnea and progressive hypoxemia in a patient with a condition that predisposes to ARDS. The hypoxemia is often r efractory to supplemental oxygen. The chest roentgenogram can e unrevealing in the first few hours of the illness. However within 24 hours, the chest roentgeno gram egins to reveal ilateral pulmonary infiltrates (8). The infiltrates may e more prominent in the peripheral lung fields (9), like the pattern of infiltra tion in Figure 23.3. Progression to mechanical ventilation often occurs in the f irst 48 hours of the illness (2). Figure 23.3. Porta le chest roentgenogram from a 36-year-old woman who presented with fever and respiratory failure. The patient died two days later. Blood cult ures su sequently grew Escherichia coli. Experts from Europe and the United States have proposed the clinical criteria fo r ARDS shown in Ta le 23.2 (1). The hallmarks of ARDS are diffuse pulmonary infi ltrates, refractory http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 22)03-May-05 17:07:00

Ovid: ICU Book hypoxemia (PAO2/FIO2 less than 200), the presence of a predisposing condition, a nd no evidence of left-heart failure (wedge pressure less than 18 mm Hg). Howeve r, the clinical features of ARDS are shared y other disorders that can cause ac ute respiratory failure. TABLE 23.2. CLINICAL CRITERIA FOR ARDS P.375 DIAGNOSTIC DILEMMAS The other conditions that can mimic the clinical presentation of ARDS include pn eumonia, acute pulmonary em olism, and cardiogenic (hydrostatic) pulmonary edema . The overlapping features of these disorders are shown in Ta le 23.3. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 22)03-May-05 17:07:00

Ovid: ICU Book TABLE 23.3. OVERLAPPING FEATURES OF ARDS AND OTHER CAUSES OF ACUTE RESPIRATORY F AILURE Physical Examination The physical findings in ARDS (e.g., fever, tachypnea, and rales) are nonspecifi c, and can e seen in any of the conditions listed in Ta le 23.3. Note that feve r can e a feature of cardiogenic pulmonary edema P.376 (i.e., edema resulting f rom an acute myocardial infarction or inflammatory myocarditis). Severity of Hypoxemia The severity of the hypoxemia can sometimes help distinguish ARDS from cardiogen ic pulmonary edema. In the early stages of ARDS, the hypoxemia is often more pro nounced than the chest roentgenogram a normalities, whereas in the early stages of cardiogenic pulmonary edema, the roentgenogram a normalities are often more p ronounced than the hypoxemia. However, there are exceptions, and severe hypoxemi a can occur in cardiogenic pulmonary edema if the mixed venous oxygen pressure ( PO2) is reduced from a low cardiac output (see Figure 21.5). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 22)03-May-05 17:07:00

Ovid: ICU Book Figure 23.5. Multiorgan failure and survival in ARDS. (Results of the multicente r study from Reference 17. Results of the single center study from Reference 18. ) The Chest Roentgenogram When a chest roentgenogram shows ilateral pulmonary infiltrates, the principal concern is to differentiate ARDS from cardiogenic pulmonary edema. This can e a difficult task (8). To illustrate the difficulty, compare the chest films in Fi gure 23.3 and Figure 23.4. Both are from patients with fever and acute respirato ry failure, and oth show ilateral infiltrates that are more prominent in the p eripheral lung fields. Despite the similarities in the chest roentgenograms (and in the clinical presentation), one represents ARDS (Fig. 23.3) and the other re presents cardiogenic pulmonary edema (Fig. 23.4). This illustrates why the conse nsus is that chest roentgenograms are not relia le for distinguishing ARDS from cardiogenic pulmonary edema (2,8). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 22)03-May-05 17:07:00

Ovid: ICU Book Figure 23.4. Porta le chest roentgenogram of a 46-year-old woman who presented w ith fever and acute respiratory failure. This patient had an acute anterior myoc ardial infarction with cardiogenic pulmonary edema. The patient survived. Predisposing Conditions The presence of high-risk conditions for ARDS may e the est means of distingui shing ARDS from cardiogenic edema. However, many of the conditions that predispo se to ARDS also predispose to pneumonia and pulmonary em olism, so the predictiv e value of these conditions is limited if pneumonia or pulmonary em olism is a c onsideration. P.377 DIAGNOSTIC AIDS Despite the high prevalence of ARDS, diagnostic recognition can e a pro lem. Th e following diagnostic approaches are availa le. WEDGE PRESSURE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 22)03-May-05 17:07:00

Ovid: ICU Book The standard method of distinguishing ARDS from cardiogenic pulmonary edema is t o measure the pulmonary capillary wedge pressure (PCWP) with a pulmonary artery catheter. The PCWP is interpreted as follows (Ta le 23.2): PCWP 18 mm Hg = ARDS PCWP > 18 mm Hg = Cardiogenic pulmonary edema Wedge versus Hydrostatic Pressure The use of the wedge pressure to identify ARDS is ased on the assumption that P CWP is a measure of pulmonary capillary hydrostatic pressure (Pc). However, PCWP is a measure of left-atrial pressure, and left-atrial pressure cannot e equiva lent to pulmonary capillary pressure in the presence of lood flow (see Chapter 11). That is, P.378 if the wedge (left-atrial) pressure were equivalent to the p ressure in the pulmonary capillaries, there would e no pressure gradient for fl ow in the pulmonary veins. Thus, PCWP must e lower than Pc. Estimating Capillary Pressure Pc can e estimated from PCWP and mean pulmonary artery pressure (Pa) according to the following relationship (from Chapter 11): (where 0.4 is the fractional decrease in pressure across the lungs that takes pl ace across the pulmonary veins, as shown in Fig. 11.4). This adjustment produces little difference etween Pc and PCWP in normal patients. However, as descri ed in Chapter 11, Pc can e dou le the PCWP in severe ARDS with pulmonary hyperten sion. For a more detailed description of the difference etween Pc and PCWP, see Chapter 11. COLLOID OSMOTIC PRESSURE The value of wedge pressure in distinguishing ARDS from cardiogenic pulmonary ed ema is limited further ecause it neglects other factors that govern the tendenc y for edema formation. These factors are identified y the Starling relationship (10): where Q is the rate of fluid movement out of the capillaries, Kp is the capillar y permea ility http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 22)03-May-05 17:07:00

Ovid: ICU Book coefficient, Pc is the capillary hydrostatic pressure, and COP is the colloid os motic pressure of lood. The colloid osmotic pressure (also called oncotic press ure) is a force created y large plasma proteins that do not move freely across capillaries. This force acts to draw fluid into the vascular space, and therefor e opposes the capillary hydrostatic pressure. Most (60% to 80%) of COP is produc ed y the al umin in plasma; fi rinogen and the immunoglo ulins account for the remainder (11,12). COP Measurement Plasma COP can e measured with an oncometer (see Reference 11 for a description of how an oncometer works). Normal COP varies with ody position (12): Normal C OP (upright): 2229 mm Hg (mean = 25 mm Hg) Normal COP (supine): 1724 mm Hg (mean = 20 mm Hg) The decrease in COP in the supine position is due to the mo ilization of protein-free fluid into the central circulation and takes a out 4 hours to d evelop. Because most patients in the ICU are in the supine position, normal COP is taken as 20 mm Hg in the ICU. The measured COP in P.379 patients in the ICU i s as low as 10 mm Hg (12). This means that hydrostatic pulmonary edema can occur when Pc is in the normal range (less than 18 mm Hg) in ICU patients. Estimating COP If an oncometer is not availa le, COP can e estimated using the total protein ( TP) concentration in plasma (11,12): Note that the last two terms of this equation contri ute little to the final COP , so they can e eliminated without sacrificing accuracy. Although the relia ili ty of this calculation varies somewhat (6), it should still provide an adequate approximation of COP. One source of error arises when nonprotein colloids (hetas tarch or the dextrans) are used as plasma expanders. These colloids dilute the s erum proteins and decrease the calculated COP ut not the measured COP. Therefor e, COP should e measured and not calculated when hetastarch or dextran is used for volume expansion. Diagnostic Use of COP http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 22)03-May-05 17:07:00

Ovid: ICU Book The addition of COP should improve the relia ility of capillary hydrostatic pres sure in identifying cardiogenic (hydrostatic) pulmonary edema. The following rul es seem reasona le: If the adjusted PCWP (Pc) is at least 4 mm Hg less than COP, then hydrostatic pulmonary edema is unlikely. If the adjusted PCWP (Pc) is more than 4 mm Hg a ove COP, then hydrostatic pulmonary edema is likely. BRONCHOALVEOLAR LAVAGE Bronchoscopy is generally not warranted as a diagnostic tool in ARDS. However, ronchoscopy may e performed occasionally for other reasons in a patient with su spected ARDS (e.g., to identify infection). In this situation, the ronchoscopis t can perform a ronchoalveolar lavage for the measurements descri ed elow (13) . This procedure can e performed safely in patients with ARDS (14). Neutrophils In normal su jects, neutrophils make up less than 5% of the cells recovered in l ung lavage fluid, whereas in patients with ARDS, as many as 80% of the recovered cells are neutrophils (13). The neutrophil preponderance in lung lavage fluid c an help distinguish ARDS from cardiogenic edema (2,13), ut may not e helpful i f pneumonia is a consideration. P.380 Total Protein The protein concentration in edema fluid and lood can e used to distinguish wa tery hydrostatic edema from inflammatory edema. The following criteria have een suggested (15): Protein (edema/serum) < 0.5 = Hydrostatic edema Protein (edema/ serum) > 0.7 = Inflammatory edema Once again, this may help distinguish ARDS fro m cardiogenic edema, ut it is of little help if pneumonia is a consideration. MANAGEMENT STRATEGIES Because there is no specific therapy that halts the inflammatory lung injury in ARDS, the management of ARDS is usually designed with the following goals in min d: (a) preventing http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 22)03-May-05 17:07:00

Ovid: ICU Book iatrogenic lung injury, ( ) reducing lung water, and (c) maintaining tissue oxyg enation. These goals can e applied to any patient with acute respiratory failur e. MANAGEMENT FOCUS Before descri ing the management of ARDS, it is important to point out that alth ough the ulk of management is focused on the lungs, only 15 to 40% of deaths in ARDS are caused y respiratory failure (16,17,18 and 19). The majority of death s are attri uted to multiple organ failure. Age is also an important factor, wit h mortality eing as much as five times higher in patients over 60 years of age (20). The influence of multiorgan failure on survival in ARDS is shown in Figure 23.5. Included in this graph are the results of a multicenter P.381 study (17) and the results of a study conducted at a single hospital (18). Both show a stea dy rise in mortality as more organs fail in addition to the respiratory failure. This demonstrates that ARDS is often just one part of a multiorgan illness, and it emphasizes the limitations of managements strategies that focus primarily on the lungs. PREVENTING IATROGENIC INJURY Ventilator Management There is now considera le evidence indicating that the large tidal volumes used during conventional mechanical ventilation (10 to 15 mL/kg) can damage the lungs (21,22). The pathologic changes in ARDS are not distri uted uniformly throughou t the lungs. Rather, there are regions of lung infiltration interspersed with re gions where the lung architecture is normal. These normal lung regions (which ma y make up only 30% of the lung) receive most of the delivered tidal volume. This results in overdistension of normal lung regions, which leads to alveolar ruptu re, surfactant depletion, and disruption of the alveolarcapillary interface. Reco gnition of the risk of lung injury at high inflation volumes and pressures has l ed to an alternative strategy where peak inspiratory pressures are kept elow 35 cm H2O y using tidal volumes of 7 to 10 mL/kg (21,22). According to this strat egy, mechanical ventilation is started at inflation volumes of 10 mL/kg. If the resulting peak inspiratory pressure (PIP) is a ove 35 cm H2O, the inflation volu me is reduced in increments of 2 mL/kg until PIP falls elow 35 cm H2O. When low inflation volumes are used, external positive end-expiratory pressure (PEEP, 5 to 10 cm H2O) is added to prevent compression atelectasis and to limit phasic co llapse of the distal airways (23,24). Inflation volumes of 5 to 8 mL/kg can resu lt in http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 22)03-May-05 17:07:00

Ovid: ICU Book CO2 retention, ut in the a sence of adverse effects, the hypercapnia is allowed to continue (permissive hypercapnia) (25). This type of pressure-limited ventil atory support is descri ed again in Chapter 26. Pulmonary Oxygen Toxicity The fractional concentration of inspired oxygen (FIO2) should e kept at 50% or lower to minimize the risk of oxygen toxicity. Arterial oxygen saturation (SaO2) should e monitored instead of arterial PO2 ecause SaO2 determines the oxygen content in arterial lood. An SaO2 a ove 90% should e sufficient to maintain ox ygen delivery to peripheral tissues. If the FIO2 cannot e reduced to elow 60%, external PEEP is added to help reduce the FIO2 to nontoxic levels. Oxygen toxic ity is descri ed in Chapter 24. REDUCING LUNG WATER The two measures that are advocated for reducing lung water are diuretics and PE EP. Unfortunately, neither measure is likely to e effective in ARDS, as explain ed elow. P.382 Diuretics Diuretic therapy can reduce lung water y decreasing capillary hydrostatic press ure and increasing colloid osmotic pressure (increased plasma protein concentrat ion). Although this should work in watery hydrostatic edema, the situation is di fferent in ARDS. The lung infiltration in ARDS is an inflammatory process, and d iuretics do not reduce inflammation. Thus, it is no surprise that diuretics have not een shown to consistently reduce lung water in ARDS (26). Considering the pathology of ARDS, the use of diuretics to reduce lung infiltration in ARDS does not seem warranted as a routine measure (27). The use of diuretics to minimize or reduce fluid overload seems a more reasona le measure, ut only when renal wa ter excretion is impaired (otherwise, the est way to prevent fluid overload is to maintain an adequate cardiac output). Whenever diuretics are used, some form of hemodynamic monitoring must e used to make sure that the diuresis is not adv ersely affecting cardiac output. Positive End-Expiratory Pressure The use of PEEP was popularized in ARDS ecause of the presumption that lung wat er could e reduced y this maneuver. However, as demonstrated in Figure 23.6, t he application of http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 22)03-May-05 17:07:00

Ovid: ICU Book PEEP does not reduce extravascular lung water in ARDS (28,29). Furthermore, the prophylactic use of PEEP does not reduce the incidence of ARDS in high-risk pati ents (30). In fact, high levels of PEEP can actually increase lung P.383 water ( 31). This latter effect may e the result of alveolar overdistension mentioned e arlier, or may e the result of PEEP-induced impairment of lymphatic drainage fr om the lungs. Figure 23.6. Effect of PEEP on extravascular lung water in ARDS. (Adapted from S aul GM et al. Effect of graded administration of PEEP on lung water in noncardio genic pulmonary edema. Crit Care Med 1982;10:667669.) Thus, the ulk of evidence indicates that PEEP is not a therapy for ARDS. Instea d, PEEP is a measure that helps to reduce iatrogenic lung injury y allowing ven tilation with low inflation volumes, and y allowing the FIO2 to e reduced to l ess toxic levels. MAINTAINING TISSUE OXYGENATION The ultimate goal of management in respiratory failure is an adequate level of o xygenation in the vital organs. The evaluation of tissue oxygenation is descri e d in Chapter 13. The est measurements availa le for evaluating tissue oxygenati on at the edside are systemic oxygen uptake (VO2), venous lactate level, and ga stric intramucosal pH (measured indirectly y gastric tonometry). Tissue oxygena tion is considered to e inadequate if whole ody VO2 is less than 100 mL/minute /m2, venous lactate is greater than 4 mmol/L, or gastric intramucosal http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 22)03-May-05 17:07:00

Ovid: ICU Book pH is less than 7.32 (Ta le 13.1). If there is evidence for impaired tissue oxyg enation, the sequence of management options shown in Figure 13.4 may e followed . A few points worth mentioning are included elow. Cardiac Filling Pressures Central venous pressure (CVP) and wedge pressures tend to overestimate cardiac f illing volumes during positive-pressure mechanical ventilation, particularly whe n PEEP is applied. This is partly ecause of transmission of intrathoracic press ures into the lumen of intrathoracic lood vessels, which increases intravascula r pressure without changing transmural pressure (the pressure that determines ve ntricular stretch and edema formation). Thus, a normal CVP or wedge pressure doe s not necessarily indicate normal cardiac filling volumes during positive-pressu re mechanical ventilation. In this setting, CVP and wedge pressure are interpret a le only if they are reduced or are less than the level of applied PEEP. Cardiac Output If the cardiac output is inadequate (e.g., a cardiac index elow 3 L/min/m2) and the CVP or wedge pressures are not elevated, volume infusion is indicated. Even though ARDS is called leaky-capillary pulmonary edema, the lung infiltration in ARDS is an inflammatory exudate, so volume infusion is not different here from volume infusion in a patient with a pneumonia. If volume infusion is not indicat ed, do utamine is used to augment the cardiac output (see Ta le 18.3 for a do ut amine dose chart) (26). Dopamine should e avoided ecause of its propensity to constrict P.384 pulmonary veins, which will raise the wedge pressure while reduc ing the left-ventricular enddiastolic volume (see Fig. 18.1). Vasodilators shoul d also e avoided ecause of their propensity to increase intrapulmonary shunt, which can add to the primary gas exchange a normality in ARDS (vasodilator prost aglandins are the exception). Blood Transfusions Transfusion is often recommended to keep the H a ove 10 g/dL, ut there is no asis for this recommendation. In fact, given the propensity for lood transfusio ns to cause ARDS (Ta le 23.1), it seems wise to avoid transfusing lood products in ARDS. If there is no evidence of inadequate tissue oxygenation, there is no need to correct anemia. SPECIFIC THERAPIES http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 22)03-May-05 17:07:00

Ovid: ICU Book The following therapeutic measures are aimed at reversing the pathologic lung in jury in ARDS. Unfortunately, there is little reason for excitement here (32). Steroids High-dose steroids have een evaluated for their a ility to reduce the inflammat ory lung injury in ARDS. Unfortunately, the results do not favor the use of ster oids in ARDS, at least not in the early stages of the illness. The following is a rief summary of the availa le studies. High-dose methylprednisolone (30 mg/kg intravenously every 6 hours for 4 doses) given to patients within 24 hours of t he diagnosis of ARDS has not improved outcome or reduced mortality (33,34). In f act, one study showed a higher mortality associated with steroid therapy in ARDS (34). High-dose methylprednisolone (same dose as a ove) given as prophylaxis to patients with sepsis syndrome did not reduce the incidence of ARDS (34). Second ary infections are increased in patients receiving high-dose methylprednisolone for ARDS (34). High-dose methylprednisolone (2 to 3 mg/kg/day) given to 25 patie nts with late ARDS (2 weeks duration) who had evidence of active fi rinoprolifer ation (leading to irreversi le pulmonary fi rosis) resulted in a eneficial resp onse in 21 patients and an 86% survival in the responders (35). This study sugge sts a possi le role for steroids late in the course of ARDS, ut corro orative e vidence is required. Surfactant Aerosolized surfactant (Exosurf, Burroughs Wellcome, and Survanta, Ross La orato ries) has proven effective in improving outcomes P.385 in the neonatal form of r espiratory distress syndrome, ut it has not met with similar success in adults with ARDS (36). Antioxidants Considering that oxygen meta olites play an important role in neutrophil-mediate d tissue injury (see Chapter 3) and that neutrophil-mediated tissue injury may p lay an important role in the pathogenesis of ARDS, it is no surprise that there is considera le interest in the possi le role of antioxidants as a specific ther apy for ARDS. Although nitric oxide can improve oxygenation http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 22)03-May-05 17:07:00

Ovid: ICU Book and reduce pulmonary artery pressures in ARDS, mortality is unchanged (37). Ther e is one report of improved survival in patients with ARDS treated with N-acetyl cysteine (a glutathione surrogate descri ed in Chapter 3) (38). However, this st udy stands alone at present. REFERENCES REVIEWS 1. Bernard GR, Artigas A, Brigham KL, et al. The AmericanEuropean Consensus Confe rence on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial co ordination. Am Rev Respir Crit Care Med 1994;149:818824. 2. Kollef MH, Schuster DP. The acute respiratory distress syndrome. N Engl J Med 1995;332:2737 (179 references). 3. Barie PS. Organ-specific support in multiple organ failure: pulmonary support . World J Surg 1995;19:581591 (140 references). PATHOGENESIS 4. Petty TL. The acute respiratory distress syndrome. Historical perspective. Ch est 1994;105(Suppl):44S46S. 5. Windsor ACJ, Mullen PG, Fowler AA, Sugerman HJ. Role of the neutrophil in adu lt respiratory distress syndrome. Br J Surg 1993;80:1017. 6. Donnelly SC, Haslett C, Dransfield I, et al. Role of selectins in development of adult respiratory distress syndrome. Lancet 1994;344:215219. 7. Hudson LD, Mil erg JA, Anardi D, Maunder RJ. Clinical risks for development o f the acute respiratory distress syndrome. Am Rev Respir Crit Care Med 1995;151: 293301. CLINICAL FEATURES http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 22)03-May-05 17:07:00

Ovid: ICU Book 8. A erle DR, Brown K. Radiologic considerations in the adult respiratory distre ss syndrome. Clin Chest Med 1990;11:737754. 9. Chiles C, Putman CE. Techniques for interpreting pulmonary opacities in the I CU. J Crit Illness 1994;9:198206. P.386 DIAGNOSTIC AIDS 10. Michel CC. Fluid movement through capillary walls. In: Renkin EM, Michel CC, Geiger SR, eds. The cardiovascular system. Vol. 4. The microcirculation, Part 1 . The hand ook of physiology. Bethesda, MD: American Physiological Society, 1984 ;375410. 11. Sinclair S, We AR. Colloid osmotic pressure measurement in critically ill patients. Intensive Care World 1991;8:120122. 12. Weil MH, Henning RJ. Colloid osmotic pressure. Significance, methods of meas urement, and interpretation. In: Weil MH, Henning RJ, eds. Hand ook of critical care medicine. Chicago: Year Book, 1979;7381. 13. Idell S, Cohen AB. Bronchoalveolar lavage in patients with the adult respira tory distress syndrome. Clin Chest Med 1985;6:459471. 14. Stein erg KP, Mitchell DR, Maunder RJ, et al. Safety of ronchoalveolar lava ge in patients with adult respiratory distress syndrome. Am Rev Respir Dis 1993; 148:556561. 15. Sprung CL, Long WM, Marcial EH, et al. Distri ution of proteins in pulmonary edema. The value of fractional concentrations. Am Rev Respir Dis 1987;136:957963 . MANAGEMENT FOCUS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 22)03-May-05 17:07:00

Ovid: ICU Book 16. Montgomery AB, Stager MA, Carrico J, Hudson LD. Causes of mortality in patie nts with the adult respiratory distress syndrome. Am Rev Respir Dis 1985;132:4854 89. 17. Bartlett RH, Morris AH, Fairley B, et al. A prospective study of acute hypox ic respiratory failure. Chest 1986;89:684689. 18. Gillespie DJ, Marsh HMM, Divertie MB, Meadows JA III. Clinical outcome of re spiratory failure in patients requiring prolonged (> 24 hours) mechanical ventil ation. Chest 1986;90:364369. 19. Suchyta MR, Clemmer TP, Elliott CG, et al. The adult respiratory distress sy ndrome. A report of survival and modifying factors. Chest 1992;101:10741079. 20. Gee M, Gottlie JE, Al ertine KH, et al. Physiology of aging related to outc ome in the adult respiratory distress syndrome. J Appl Physiol 1990;69:822829. MANAGEMENT 21. Hickling KG. Ventilatory management in ARDS: can it affect outcome? Intensiv e Care Med 1990;16:219226. 22. Marini JJ. Pressure-targeted, lung-protective ventilatory support in acute l ung injury. Chest 1994;105(Suppl):109S115S. 23. Gattinoni L, D'Andrea L, Pelosi P, et al. Regional effects and mechanism of positive end-expiratory pressure in early adult respiratory distress syndrome. J AMA 1993;269:21222127. 24. Muscedere JG, Mullen JBM, Slutsky AS. Tidal ventilation at low airway pressu res can augment lung injury. Am Rev Respir Crit Care Med 1994;149:13271334. 25. Hickling KG, Walsh J, Henderson S, Jackson R. Low mortality rate in adult re spiratory distress syndrome using low-volume, pressure-limited ventilation with http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 22)03-May-05 17:07:00

Ovid: ICU Book permissive hypercapnia: a prospective study. Crit Care Med 1994;22:15681578. P.38 7 26. Broaddus VC, Berthiaume Y, Biondi JW, et al. Hemodynamic management of the a dult respiratory distress syndrome. J Intensive Care Med 1987;2:190213. 27. Hyers Tm. ARDS: the therapeutic dilemma. Chest 1990;97:1025. 28. Saul GM, Feeley TW, Mihm FG. Effect of graded PEEP on lung water in noncardi ogenic pulmonary edema. Crit Care Med 1982;10:667669. 29. Hel ert C, Paskanik A, Breden erg CE. Effect of positive end-expiratory pres sure on lung water in pulmonary edema caused y increased mem rane permea ility. Ann Thorac Surg 1984;36:4248. 30. Pepe PE, Hudson LD, Carrico J. Early application of positive end-expiratory pressure in patients at risk for the adult respiratory distress syndrome. N Engl J Med 1984;311:281286. 31. Demling RH, Stau NC, Edmunds LH. Effect of end-expiratory pressure on accum ulation of extravascular lung water. J Appl Physiol 1975;38:907912. 32. Messent M, Griffiths MJD. Pharmacotherapy in lung injury. Thorax 1992;47:6516 56. 33. Bernard GR, Luce JM, Sprung CL, et al. High-dose corticosteroids in patients with adult respiratory distress syndrome. N Engl J Med 1987;317:15651570. 34. Bone RC, Fischer CJ Jr, Clemmer TP, et al. Early methylprednisolone treatmen t for septic syndrome and the adult respiratory distress syndrome. Chest 1987;92 :10321036. 35. Meduri GU, Chinn A. Fi rinoproliferation in late adult respiratory distress syndrome. Chest 1994;105(Suppl):127S129S. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 22)03-May-05 17:07:00

Ovid: ICU Book 36. Anzueto A, Baughman RP, Guntupalli KK, et al. Aerosolized surfactant in adul ts with sepsis-induced acute respiratory distress syndrome. N Engl J Med 1996;33 4:14171421. 37. Lunn RJ. Inhaled nitric oxide therapy. Mayo Clin Proc 1995;70:247255. 38. Suter PM, Domenighetti G, Schaller MD, et al. N-acetylcysteine enhances reco very from acute lung injury in man: a randomized, dou le- lind, place o-controll ed clinical study. Chest 1994;105:190194. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 22)03-May-05 17:07:00

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 24 OXYGEN INHALATION THERAPY For as a candle urns out much faster in dephlogisticated than in common air, so we might, as may e said, live out too fast in this pure kind of air. --Joseph P riestley One of the rare sights in any ICU is a patient who is not receiving sup plemental oxygen to reathe. Considering the destructive nature of oxygen descri ed in Chapter 3, the tendency to shower ICU patients with oxygen is a risky ven ture. This chapter egins y examining the indications and goals of supplemental oxygen administration (1). This is followed y a rief description of the diffe rent oxygen inhalation devices (2,3 and 4). The final section of the chapter is an addendum to Chapter 3 and focuses on the toxic risk of inhaled oxygen. THE NEED FOR SUPPLEMENTAL OXYGEN Despite the fact that oxygen inhalation is a therapeutic intervention designed t o correct tissue hypoxia, oxygen administration seems to e more of a knee-jerk response to the presence of a life-threatening condition. This is supported y a recent survey showing that over 50% of hospitalized patients were receiving sup plemental oxygen without a written order for oxygen therapy (5). INDICATIONS FOR OXYGEN Arterial Hypoxemia versus Tissue Hypoxia Clinical studies reveal a poor correlation etween arterial hypoxemia and tissue hypoxia. This is demonstrated in Ta le 24.1, which shows the arterial PO2 and w hole lood lactate level in seven patients with severe hypoxemia (PAO2 elow 40 mm Hg) secondary to acute exacer ation of chronic o structive lung disease (6). If hyperlactatemia http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 17)03-May-05 17:08:09

Ovid: ICU Book P.389 ( lood lactate a ove 4 mmol/L) is used as a marker of tissue hypoxia, ther e is no evidence of tissue hypoxia in any patient with severe hypoxemia, even at arterial PO2 levels as low as 22 mm Hg. This o servation has een corro orated in patients with acute respiratory distress syndrome (7). TABLE 24.1. SEVERE HYPOXEMIA WITHOUT EVIDENCE OF TISSUE HYPOXIA The availa le evidence thus indicates that arterial hypoxemia is not a marker of tissue hypoxia (at least not when using indirect markers of tissue hypoxia). Th e following statement from the study in Ta le 24.1 is relevant in this regard: I n the resting patient, even the most severe clinical hypoxemia due to pulmonary insufficiency does not itself lead to generalized tissue anaero iasis (6). Remem er this statement when giving supplemental oxygen ased on measures of arterial oxygenation. WHAT TO MONITOR Arterial oxygenation can also e misleading as an endpoint of oxygen inhalation therapy. This point is demonstrated in Figure 24.1. The graphs in this figure sh ow the discrepancy etween changes in arterial PO2 and changes in systemic oxyge n transport during supplemental oxygen administration (8). Note that arterial PO 2 increases from 61 to 83 mm Hg (36% change, P < 0.01) while the rate of oxygen transport decreases from 12.8 to 12.1 mL/minute/ kg (5% change, not significant) . Thus, an increase in arterial PO2 during oxygen inhalation should not e used as evidence for an increase in tissue oxygen availa ility (8,9). This is http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 17)03-May-05 17:08:09

Ovid: ICU Book consistent with the o servation that oxygen inhalation does not protect against myocardial ischemia (10). Figure 24.1. Effects of oxygen inhalation (FIO2 = 0.26) on arterial oxygenation and systemic oxygen transport. Points on the graph represent mean values for the group of patients studied. (Data from DeGaute JP et al. Oxygen delivery in acut e exacer ation of chronic o structive pulmonary disease. Effects of controlled o xygen therapy. Am Rev Respir Dis 1981;124:26.) Oxygen and Systemic Blood Flow The lack of improvement in systemic oxygen transport during oxygen inhalation is explained y the tendency for oxygen to reduce systemic lood flow. There are t wo mechanisms for this effect. First, P.390 oxygen acts as a vasoconstrictor in all vascular eds except the pulmonary circulation (where it acts as a vasodilat or) (11,12). Second, oxygen inhalation is often associated with a decrease in ca rdiac output (8,9,13). Although this is caused partly y reversal of the cardiac stimulatory effects of hypoxemia, oxygen also has negative inotropic effects on the heart, and oxygen http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 17)03-May-05 17:08:09

Ovid: ICU Book inhalation can reduce cardiac output in the a sence of hypoxemia (13). The tende ncy for oxygen to reduce cardiac output emphasizes the value of invasive hemodyn amic monitoring to evaluate the response to oxygen inhalation. METHODS OF OXYGEN INHALATION Oxygen delivery systems are classified as low-flow or high-flow systems. Low-flo w delivery systems (nasal prongs, face masks, and masks with reservoir ags) pro vide a reservoir of oxygen for the patient to inhale. When total ventilation exc eeds the capacity of the oxygen reservoir, room air is inhaled. The final concen tration of inhaled oxygen (FIO2) is determined y the size of the oxygen reservo ir, the rate at which the reservoir is filled, and the ventilatory demands of th e patient. In contrast to the varia le FIO2 of the low-flow systems, high-flow o xygen delivery systems provide a constant FIO2. This is P.391 achieved y delive ring oxygen at flow rates that exceed the patient's peak inspiratory flow rate, or y using devices that entrain a fixed proportion of room air. NASAL PRONGS Nasal prongs deliver a constant flow of oxygen to the nasopharynx and oropharynx , which acts as an oxygen reservoir (average capacity = 50 mL, or a out one-thir d of the anatomic dead space) (2,14). The relationship etween the oxygen flow r ate and the FIO2 in normal su jects is shown in Ta le 24.2 (14). As the oxygen f low rate increases from 1 to 6 L/min, the FIO2 increases from 0.24 to 0.46. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 17)03-May-05 17:08:09

Ovid: ICU Book TABLE 24.2. LOW-FLOW OXYGEN DELIVERY SYSTEMS It is important to note that the FIO2 values in Ta le 24.2 pertain only to patie nts with a normal ventilatory pattern (see footnote). Any variation in the venti latory pattern will also vary the FIO2. This effect is illustrated in Figure 24. 2 (4). In this example, the respiratory rate is varied from 10 to 40 reaths/min ute while the oxygen flow rate through the nasal cannula and the patient's tidal volume are oth held constant. A quadrupling of the respiratory rate results in a 48% reduction in FIO2. This demonstrates the limitations of oxygen delivery t hrough nasal prongs in patients who are reathing rapidly or otherwise have a hi gh minute ventilation. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 17)03-May-05 17:08:09

Ovid: ICU Book Figure 24.2. The relationship etween respiratory rate and FIO2 during oxygen ad ministration at 6 L/minute through a nasal cannula. Assumes a constant tidal vol ume (Vt) of 500 mL. (Adapted from O'Connor BS, Vender JS. Oxygen therapy. Crit C are Clin 1995;11:67.) Advantages and Disadvantages Nasal prongs are easy to use and well tolerated y most patients. The major disa dvantages of nasal prongs are the tendency for the FIO2 to vary as the patient's ventilatory pattern changes and the ina ility to achieve a high FIO2 in patient s with a high ventilatory demand. The continuous flow of oxygen through nasal pr ongs has een reported P.392 as a cause of spontaneous gastric rupture (15), ut this iatrogenic complication is rare. LOW-FLOW OXYGEN MASKS Face masks add 100 to 200 mL to the capacity of the oxygen reservoir. These devi ces fit loosely on the face, which allows room air to e inhaled, if needed. Sta ndard face masks deliver oxygen at flow rates etween 5 and 10 L/min. The minimu m flow rate of 5 L/min is needed to clear exhaled gas from the mask. Low-flow ox ygen masks can achieve a maximum FIO2 of approximately 0.60. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 17)03-May-05 17:08:09

Ovid: ICU Book Advantages and Disadvantages Standard face masks can provide a slightly higher maximum FIO2 than nasal prongs . However, this difference may e small. In general, face masks have the same dr aw acks as nasal prongs. MASKS WITH RESERVOIR BAGS The addition of a reservoir ag to a standard face mask increases the capacity o f the oxygen reservoir y 600 to 1000 mL (depending on the size of the ag). If the reservoir ag is kept inflated, the patient P.393 will inhale only the gas c ontained in the ag. There are two types of maskreservoir ag devices. The one sh own in Figure 24.3 is a partial re reathing system. This device allows the gas e xhaled in the initial phase of expiration to return to the reservoir ag. As exh alation proceeds, the expiratory flow rate declines, and when the expiratory flo w rate falls elow the oxygen flow rate, exhaled gas can no longer return to the reservoir ag. The initial part of expiration contains gas from the upper airwa ys (anatomic dead space), so the gas that is re reathed is rich in oxygen and la rgely devoid of CO2. Partial re reather devices can achieve a maximum FIO2 of 70 to 80%. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 17)03-May-05 17:08:09

Ovid: ICU Book Figure 24.3. Partial re reathing system. The initial 100 to 150 mL of exhaled ga s (anatomic dead space) is returned to the reservoir ag for re reathing. The modified device in Figure 24.4 is a nonre reathing system. This device has a one-way valve that prevents any exhaled gas from returning to the reservoir ag . Nonre reather devices permit inhalation of pure oxygen (FIO2 = 1.0). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 17)03-May-05 17:08:09

Ovid: ICU Book Figure 24.4. Nonre reathing system. A one-way valve prevents exhaled gas from re turning to the reservoir ag. Advantages and Disadvantages The principal advantage of the reservoir ags is the greater a ility to control the composition of inhaled gas. However, ecause the masks must create a tight s eal on the face, it is not possi le to feed patients y P.394 mouth or nasoenter al tu e when these devices are in use. Aerosolized ronchodilator therapy is als o not possi le with reservoir ag devices. HIGH-FLOW OXYGEN MASKS High-flow oxygen inhalation devices provide complete control of the inhaled gas mixture and deliver a constant FIO2 regardless of changes in ventilatory pattern . The operation of a highflow oxygen mask is shown in Figure 24.5 (16). Oxygen i s delivered to the mask at low flow rates, ut at the inlet of the mask, the oxy gen is passed through a narrowed orifice, and this http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 17)03-May-05 17:08:09

Ovid: ICU Book creates a high-velocity stream of gas (analogous to the effect created y a nozz le on a garden hose). This high-velocity jet stream generates a shearing force k nown as viscous drag that pulls room air into the mask. The volume of room air t hat moves into the mask (which determines the FIO2) can e varied y varying the size of the openings (called entrainment ports) on the mask. These masks can in crease the FIO2 to a maximum of 0.50. At any given FIO2, the proportion of inhal ed gas provided y entrained room air remains constant; that is, FIO2 remains fi xed regardless of changes in oxygen flow rate or changes in inspiratory flow rat e. Figure 24.5. High-flow oxygen inhalation device. RA = room air. P.395 Advantages and Disadvantages The major advantage of high-flow oxygen masks is the a ility to deliver a consta nt FIO2. This feature is desira le in patients with chronic hypercapnia ecause an inadvertent increase in FIO2 in these patients can lead to further CO2 retent ion. The major draw ack with these masks is the ina ility to deliver high concen trations of inhaled oxygen. TOXICITY OF INHALED OXYGEN The damaging effects of oxygen are descri ed in detail in Chapter 3. The oxygen in inhaled gas is an unrelenting threat to the functional integrity of the lungs . Fortunately, the lungs are http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 17)03-May-05 17:08:09

Ovid: ICU Book endowed with a concentrated supply of endogenous antioxidants that protect again st the destructive actions of oxygen. However, when these antioxidant defenses a re overwhelmed ( y too much oxygen or not enough antioxidants), oxygen inhalatio n can produce widespread damage in the lungs (17). PULMONARY OXYGEN TOXICITY Inhalation of pure oxygen can produce a progressive and lethal form of lung inju ry that is very similar to the acute respiratory distress syndrome (ARDS) descri ed in Chapter 23 (17). This similarity is not surprising ecause ARDS is the re sult of inflammatory cell injury, and oxygen meta olites play an important role in the damaging effects of inflammation. P.396 Comparative Physiology The tendency to develop pulmonary oxygen toxicity varies widely in different spe cies. For example, 5 to 7 days of reathing pure oxygen is fatal in la oratory r ats, whereas sea turtles can reathe pure oxygen indefinitely without harm (18). This difference in suscepti ility to oxygen toxicity is important ecause the e xperimental studies of pulmonary oxygen toxicity have een conducted almost sole ly in animals. As a result, little is known a out the tendency to develop pulmon ary oxygen toxicity in humans. In healthy volunteers, inhalation of 100% oxygen for rief periods of time (6 to 12 hours) results in a tracheo ronchitis and a d ecrease in vital capacity (17). The latter effect is considered to e the result of a sorption atelectasis, ut this is unproven. The longest exposure to 100% o xygen in humans includes 5 patients with irreversi le coma (3 to 4 days) and one healthy volunteer (4.5 days) (19,20). In all these cases, prolonged exposure to oxygen resulted in a pulmonary syndrome very much like ARDS. SAFE VERSUS TOXIC FIO2 Based on the o servation that oxygen inhalation does not reduce the vital capaci ty when the FIO2 is elow 0.60 (17), an FIO2 of 0.60 was esta lished as the thre shold FIO2 separating safe from toxic levels of inhaled oxygen. Although this th reshold FIO2 was esta lished in healthy adults, it has also een applied to pati ents in the ICU. The consensus is that inhalation of a gas mixture with an FIO2 a ove 0.60 for longer than 48 hours is a toxic exposure to inhaled oxygen. There fore, when an FIO2 a ove 0.60 is required for longer than a few days, other meas ures should e instituted (such as mechanical ventilation or PEEP) to reduce the FIO2 to less toxic levels. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 17)03-May-05 17:08:09

Ovid: ICU Book Antioxidant Depletion The recommendation of a universal FIO2 threshold separating safe from toxic oxyg en inhalation is inappropriate ecause it neglects the contri ution of endogenou s antioxidants to the risk of oxygen toxicity. The principal antioxidants are in cluded in Ta le 3.1. If these antioxidants ecome depleted, oxygen toxicity is p ossi le at an FIO2 that is much lower than 0.60. Antioxidant depletion may e co mmon in patients in the ICU, particularly during a prolonged ICU stay. This effe ct is illustrated in Figure 24.6 (21). The antioxidant in this case is selenium, a cofactor for the glutathione peroxidase enzyme (Figure 3.3). In the first wee k after admission to the ICU, the selenium level (in whole lood) is 37% lower t han in the control su jects, and y the fourth week of the ICU stay, the seleniu m level is a out 50% of the control level. Reductions in other antioxidants (glu tathione and vitamin E) has also een reported in ICU patients (22,23). Figure 24.6. Reduced levels of an antioxidant (selenium) in ICU patients. Data ( means and standard deviations) from 57 healthy lood donors and 175 consecutive ICU admissions. (From Hawker FH et al. Effects of acute illness on selenium home ostasis. Crit Care Med 1990;18:442.) P.397 Optimal FIO2 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 17)03-May-05 17:08:09

Ovid: ICU Book Because antioxidant protection may e defective in ICU patients, the recommendat ion that an FIO2 elow 0.6 is safe does not apply to them. A more reasona le app roach is to assume that any FIO2 a ove 0.21 (room air) can represent a toxic exp osure to oxygen in ICU patients. Based on this assumption, the optimal FIO2 for safe oxygen inhalation is the lowest tolera le (rather than any) FIO2 elow 0.60 . PREVENTIVE MEASURES The measures presented in Ta le 24.3 are aimed at reducing the risk of pulmonary oxygen toxicity. These preventive measures are organized according to three goa ls. TABLE 24.3. PREVENTIVE MEASURES FOR OXYGEN TOXICITY Limit O2 Inhalation As stressed repeatedly in this chapter, the routine administration of supplement al oxygen is inappropriate and should e a andoned. The principal indications fo r oxygen inhalation http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 17)03-May-05 17:08:09

Ovid: ICU Book therapy are listed in Ta le 24.3. They include clinical evidence of inadequate t issue oxygenation and a high risk of tissue hypoxia. Adherence to these indicati ons P.398 should reduce the risk of oxygen toxicity without creating any added r isk of tissue hypoxia. Limit the FIO2 The second measure is a corollary of the first and involves using the lowest tol era le FIO2. First, FIO2 should not e kept a ove 0.6 for longer than 48 hours. If the FIO2 is elow 0.6 and there are no indications for supplemental oxygen, F IO2 should e reduced gradually to the lowest tolera le level. Reduction of the FIO2 can proceed as long as there is no evidence of (or high risk of) tissue hyp oxia. Limiting supplemental oxygen as descri ed here may not e an easy task; me m ers of the ICU staff who are accustomed to the unregulated use of oxygen may r esist. When opposition is encountered, remem er to point out that oxygen is a th erapy (in fact, oxygen is often classified as a drug), and like all therapies, i t is not appropriate if not justified. Support Antioxidant Protection Because antioxidant depletion may e common in ICU patients, maintaining the poo l of endogenous antioxidants is an important measure for preventing oxygen toxic ity. Unfortunately, it is not yet possi le to evaluate the capacity for antioxid ant protection. However, the P.399 measures indicated in Ta le 24.3 may help mai ntain the ody stores of two endogenous antioxidants: selenium and vitamin E. Fo r more information on the evaluation and administration of these antioxidants, s ee Chapter 3. REFERENCES REVIEWS 1. Fulmer JD, Snider GL. ACCP-NHLBI National Conference on Oxygen Therapy. Chest 1984;86:234247 (34 references). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 17)03-May-05 17:08:09

Ovid: ICU Book 2. Kacmarek RM. Supplemental oxygen and other medical gas therapy. In: Pierson D J, Kacmarek RM, eds. Foundations of respiratory care. New York: Churchill Living stone, 1992;859890 (45 references). 3. Carlton TJ, Anthonisen NR. A guide for judicious use of oxygen in critical il lness. J Crit Illness 1992;7:17441757 (36 references). 4. O'Connor BS, Vender JS. Oxygen therapy. Crit Care Clin 1995;11:6778 (7 referen ces). THE NEED FOR SUPPLEMENTAL OXYGEN 5. Small D, Duha A, Wieskopf B, et al. Uses and misuses of oxygen in hospitalize d patients. Am J Med 1992;92:591595. 6. Eldridge FE. Blood lactate and pyruvate in pulmonary insufficiency. N Engl J Med 1966;274:878883. 7. Lundt T, Koller M, Kofstad J. Severe hypoxemia without evidence of tissue hyp oxia in the adult respiratory distress syndrome. Crit Care Med 1984;12:7576. 8. Corriveau ML, Rosen BJ, Dolan GF. Oxygen transport and oxygen consumption dur ing supplemental oxygen administration in patients with chronic o structive pulm onary disease. Am J Med 1989;87:633636. 9. Lejeune P, Mols P, Naeije R, et al. Acute hemodynamic effects of controlled o xygen therapy in decompensated chronic o structive pulmonary disease. Crit Care Med 1984;12:10321035. 10. Kavanaugh PB, Cheng DCH, Sandler AN, et al. Supplemental oxygen does not red uce myocardial ischemia in premedicated patients with critical coronary artery d isease. Anesth Analg 1993;76:950956. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 17)03-May-05 17:08:09

Ovid: ICU Book 11. Packer M, Lee WH, Medina N, Yushak M. Systemic vasoconstrictor effects of ox ygen administration in o literative pulmonary vascular disorders. Am J Cardiol 1 986;57:853 858. 12. Bongard O, Bounameaux H, Fagrell B. Effects of oxygen on skin microcirculati on in patients with peripheral arterial occlusive disease. Circulation 1992;86:8 78886. 13. Daly WJ, Bondurant S. Effects of oxygen reathing on the heart rate, lood p ressure and cardiac index of normal menresting, with reactive hyperemia, and afte r atropine. J Clin Invest 1962;41:126132. METHODS OF OXYGEN INHALATION 14. Shapiro BA, Kacmarek RM, Cane RD, et al. Clinical application of respiratory care. 4th ed. St. Louis: CV Mos y, 1991;123134. P.400 15. van der Loos TLJM, Lustermans FAT. Rupture of the normal stomach after thera peutic oxygen administration. Intensive Care Med 1986;12:5253. 16. Scacci R. Air entrainment masks: jet mixing is how they work. The Bernoulli and Venturi principles is how they don't. Respir Care 1979;24:928931. TOXICITY OF INHALED OXYGEN 17. Lodato RF. Oxygen toxicity. Crit Care Clin 1990;6:749765. 18. Fan urg BL. Oxygen toxicity: why can't a human e more like a turtle? Intens ive Care Med 1988;3:134136. 19. Bar er RE, Hamilton WK. Oxygen toxicity in man. N Engl J Med 1970;283:1478 14 83. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 17)03-May-05 17:08:09

Ovid: ICU Book 20. Winter PM, Smith G. The toxicity of oxygen. Anesthesiology 1972;37:210212. 21. Hawker FH, Stewart PM, Switch PJ. Effects of acute illness on selenium homeo stasis. Crit Care Med 1990;18:442446. 22. Cor ucci GG, Gasparetto A, Candiani A, et al. Shock-induced damage to mitoch ondrial function and some cellular antioxidant mechanisms in humans. Circ Shock 1985;15:1526. 23. Pincemail J, Bertrand Y, Hanique G, et al. Evaluation of vitamin E deficienc y in patients with adult respiratory distress syndrome. Ann N Y Acad Sci 1989;57 0:498500. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 17)03-May-05 17:08:09

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 25 RESPIRATORY PHARMACOTHERAPY This chapter focuses on the use of pharmaceutical agents to enhance flow in the conducting airways of the lungs. Three distinct strategies are descri ed. The fi rst involves drugs that increase airway cali er y relaxing ronchial smooth mus cle ( ronchodilators). The second involves drugs that reduce airflow o struction from inflammation in the walls of the small airways (corticosteroids), and the third strategy involves measures that facilitate the removal of o structing secr etions (mucokinetic therapy). BEDSIDE MONITORING Interventions aimed at enhancing airflow in the lungs must e guided y measurem ents of airway function. Because ICU patients are not easily transported to the pulmonary function la oratory, these measurements must e o tained at the edsid e. The following are three measurements of airway function that are easily perfo rmed at the edside. PEAK EXPIRATORY FLOW The peak expiratory flow rate, the highest flow achieved during a maximum expira tory effort, is a measurement of the total flow capacity of the airways. The pea k expiratory flow is easily measured at the edside with a hand-held device such as the one in Figure 25.1. The device shown (the MiniWright peak flowmeter) is a out 7 inches in length, and weighs only 11 ounces. The patient holds the devic e in the palm of the hand and inhales as much air as possi le (to total lung cap acity), then exhales as forcefully as possi le into the mouthpiece of the device . The flow of exhaled gas follows a contour like the one illustrated in the figu re, with the peak flow occurring early in expiration (when the elastic recoil of the lungs is the highest and the cali er of the airways P.402 is the greatest). The flow of exhaled gas displaces a spring-loaded piston in the peak flowmeter, and a pointer attached to the piston records the displacement on a cali rated s cale etched on the outer surface of the device. This scale records the displacem ent as the maximum (peak) flow rate, in liters per minute (L/min). This maneuver is repeated twice, and the highest of the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 22)03-May-05 17:24:28

Ovid: ICU Book three flows is recorded as the peak expiratory flow at that time (1). Figure 25.1. A hand-held device for measuring the peak expiratory flow rate (the highest point on the expiratory flow contour). The peak expiratory flow is an effort-dependent parameter and is a relia le inde x of flow resistance only when the expiratory effort is maximal (and constant). Therefore, it is important to o serve the patient during the peak flow measureme nt to determine whether a maximum effort is eing expended. If not, the measurem ent should e discarded (2,3 and 4). Varia ility The peak expiratory flow varies with age, sex, and height, and thus reference ta les are needed to interpret the peak flow measurement in individual patients (t hese ta les are included in the Appendix at the end of the text). The influence of sex and age on the peak expiratory flow rate (PEFR) is demonstrated elow (5) . http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 22)03-May-05 17:24:28

Ovid: ICU Book Age (yr) Male (height 70 in) 20 70 Female (height 65 in) 20 70 Normal PEFR 690 L/min 550 L/min 460 L/min 400 L/min At 20 years of age, the male has a PEFR that is 50% higher than the P.403 female , and in oth sexes, the PEFR is 15 to 20% slower in the elderly. The PEFR also has a diurnal variation of 10 to 20%, with the nadir in the early morning (6). H owever, the importance of this diurnal variation in hospitalized patients (where circadian rhythms may e a sent) has not een determined. Clinical Applications The PEFR can e used to evaluate the severity of airways o struction (1,7,8) or to determine ronchodilator responsiveness (9). The criteria for these applicati ons are shown in Ta le 25.1 (1). According to the Expert Panel Report of the Nat ional Asthma Education Program (1), the PEFR should e monitored on a regular a sis in patients with asthma. The management of acute asthma in the emergency dep artment can e guided y the PEFR as shown in Figure 25.4. The PEFR can also e used to determine the need for ronchodilator therapy. This is accomplished y r ecording the PEFR just efore, and again 20 minutes after, a ronchodilator aero sol treatment (the respiratory therapy department will perform the peak flow mea surements on request). If the PEFR increases y 15% or more after the treatment, then therapy with ronchodilator aerosols is justified. This represents a more rational approach to ronchodilator therapy than the current practice of orderin g ronchodilator aerosol treatments routinely for all hospitalized patients with o structive airways disease. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 22)03-May-05 17:24:28

Ovid: ICU Book TABLE 25.1. PEAK EXPIRATORY FLOW INTERPRETATIONS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 22)03-May-05 17:24:28

Ovid: ICU Book Figure 25.4. Protocol for the early management of acute asthma in the emergency department. PEFR = peak expiratory flow rate. (Adapted from the National Asthma Education Program, Expert Panel Report [1].) PEAK INSPIRATORY PRESSURE Aerosol ronchodilator treatments are given routinely to ventilator-dependent pa tients, often without documented need or enefit. One method of assessing ronch odilator responsiveness during mechanical ventilation is to monitor changes in t he peak inspiratory pressure (PIP) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 22)03-May-05 17:24:28

Ovid: ICU Book P.404 measured in the endotracheal tu e at the end of lung inflation. This press ure is determined y the inflation volume, the distensi ility (compliance) of th e lungs, and the resistance to flow in the endotracheal tu e and airways. If a ronchodilator treatment has effectively decreased airways resistance, the PIP sh ould also decrease. Therefore, assuming all other varia les are constant, a decr ease in PIP after an aerosol ronchodilator treatment can e used as evidence fo r ronchodilator responsiveness (9). The PIP is descri ed in more detail in Chap ter 26. AUTO-PEEP When flow in the airways is reduced, exhaled gas does not escape completely from the lungs, and the gas that is trapped in the distal airspaces creates a positi ve pressure at the end of expiration. This positive end-expiratory pressure (PEE P) is known as intrinsic PEEP or autoPEEP, and it is an indirect measure of the severity of airways o struction. (This phenomenon is an expression of the hyperi nflation that occurs in o structive airways disease.) Thus, in the presence of a uto-PEEP, a favora le response to a ronchodilator treatment should e accompani ed y a decrease in the level of auto-PEEP. The measurement of auto-PEEP is desc ri ed in Chapter 28. -RECEPTOR AGONISTS The most effective ronchodilators are pharmaceutical agents that stimulate -adre nergic receptors in ronchial smooth muscle (-2 su type) and promote smooth muscl e relaxation. These -receptor agonists are most effective and least toxic when th ey are given as an aerosol that is inhaled (1,2,3 and 4,10). The drugs and dose recommendations for -agonist aerosol treatments are shown in Ta le 25.2. All trea tments in this ta le should e considered equivalent in efficacy and toxicity. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 22)03-May-05 17:24:28

Ovid: ICU Book TABLE 25.2. AEROSOL THERAPY WITH -RECEPTOR AGONISTS P.405 AEROSOL THERAPY There are two devices used to generate ronchodilator aerosols, and each is depi cted schematically in Figure 25.2 (see reference 2 for a detailed description of aerosol generators). Figure 25.2. Small-volume aerosol generators. Jet Ne ulizer The jet ne ulizer operates on the same principle as the high-flow oxygen mask sh own in Figure 24.5. One end of a narrow capillary tu e is su merged in the drug solution, and a rapidly flowing stream of gas is passed over the other end of th e tu e. This gas jet creates a viscous drag that draws the drug solution up the capillary tu e, and when the solution reaches the gas jet, it is pulverized to f orm the aerosol spray, which is then carried to the patient with the inspiratory gas flow. Small-volume jet ne ulizers use a reservoir volume of 3 to 6 mL (drug solution plus saline diluent) and can completely aerosolize the reservoir volum e in less than 10 minutes. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 22)03-May-05 17:24:28

Ovid: ICU Book Metered-Dose Inhaler The metered-dose inhaler operates in much the same way as a hair spray canister. This device has a pressurized canister that contains a drug solution with a oi ling point elow room temperature. When the canister is squeezed etween the thu m and fingers, a valve opens that releases a fixed volume of the drug solution. The liquid immediately vaporizes when it emerges from the cannister, and a liqu id propellant P.406 in the solution creates a high-velocity spray. The spray eme rging from the canister can have a velocity in excess of 30 meters per second (o ver 60 miles per hour) (11). Because of the high velocity of the emerging spray, when a metered-dose inhaled is placed in the mouth, most of the aerosol spray i mpacts on the posterior wall of the oropharynx and is not inhaled. This inertial impaction is reduced somewhat y using a spacer device to increase the distance etween the canister and mouth (this reduces the velocity of the spray reaching the mouth). Ne ulizer versus Metered-Dose Inhaler As indicated in Ta le 25.2, the therapeutic dose delivered y metered-dose inhal ers (usually 2 or 3 sprays) is only a fraction of the dose delivered y ne ulize d aerosols. Yet despite this dose difference, the ronchodilator response to met ered dose inhalers and ne ulized aerosols is often indistinguisha le. This is de monstrated in Figure 25.3. In this case, patients with severe asthma (PEFR less than 50% predicted) were given al uterol aerosol treatments with a hand-held ne ulizer (2.5 mg al uterol per treatment) or metered-dose inhaler (4 puffs or 0.36 mg al uterol per treatment) (12). After two treatments with each type of aeroso l therapy, the increase in peak expiratory flow is equivalent. Thus, despite an almost tenfold difference in total drug dose (5 mg via ne ulizer versus 0.7 mg v ia metered-dose inhaler), the ronchodilator response to each type of aerosol th erapy is the same. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 22)03-May-05 17:24:28

Ovid: ICU Book

Equivalent ronchodilator responses to ne ulizer and metered-dose aerosols have een documented in oth spontaneously reathing and P.407 ventilator-dependent p atients (2,12,13 and 14). The likely reason for this is that very little of the drug dose (less than 10%) is deposited in the lungs with either type of aerosol therapy (most of the dose is lost via condensation on the equipment). Because dr ug therapy with a metered-dose inhaler is less expensive and less la or-intensiv e than ne ulizer treatments, there has een a general trend to adopt metered-dos e inhalers for ronchodilator treatments in hospitalized patients. At one hospit al, the annual savings from such a change were estimated at $83,000 for the hosp ital and $300,000 for the patients (15). Ventilator-Dependent Patients Although ne ulizers have een the standard aerosol generator for ventilator-depe ndent patients, metered-dose inhalers have also een used to deliver ronchodila tor aerosols during mechanical ventilation (see reference 2 for examples of how to connect the metered-dose inhaler to the ventilator circuit). The est results are o tained when the aerosol spray from the metered-dose inhaler is passed thr ough a narrow-gauge catheter extending eyond the tip of the tracheal tu e (16). The following steps achieve optimal delivery with either type of aerosol http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 22)03-May-05 17:24:28

Figure 25.3. Bronchodilator responses to al uterol delivered y mg per treatment) and a metered-dose inhaler (MDI) (0.4 mg per tients with acute exacer ation of asthma. Each point represents or all patients in each treatment group. PEFR = peak expiratory from Idris AH, et al. Emergency department treatment of severe 93;103:665672.)

a ne ulizer (2.5 treatment) in pa the mean value f flow rate. (Data asthma. Chest 19

Ovid: ICU Book therapy (13,17): (a) turn the humidifier off, ( ) decrease the inspiratory flow rate, and (c) increase inspiratory time. MANAGEMENT OF ACUTE ASTHMA Inhaled agonists are the drugs of choice for the initial management of patients with asthma. The initial management of patients with acute exacer ation of asthm a can e organized as shown in Figure 25.4. Treatment egins with a series of -ag onist aerosol treatments given at 20-minute intervals. Each treatment can e giv en y ne ulizer or metered-dose inhaler, and each treatment is followed y a pea k flow measurement. If a satisfactory response is not o tained after three aeros ol treatments (which covers a 1-hour period), a corticosteroid is administered, and su sequent aerosol treatments can e given at 1-hour intervals until a satis factory response is achieved. After the patient has sta ilized, the -agonist aero sol treatments are given every 6 hours. SIDE EFFECTS The most commonly reported side effects of -agonist aerosol therapy are tachycard ia (cardiac receptors), tremors (skeletal muscle receptors), a decrease in serum potassium (transcellular potassium shift), and a decrease in arterial PO2 (incr eased intrapulmonary shunt) (18,19). The latter effect is transient (lasting a o ut 30 minutes) and usually is seen during aggressive ronchodilator therapy in a cute asthma (18). Other side effects include hyperglycemia and a decrease P.408 in serum magnesium and serum phosphorous (19). Cardiac ischemia has een reporte d, ut is rare (18). Hypokalemia The usual therapeutic doses of inhaled agonists are associated with only mild de creases in serum potassium (0.5 mEq/L or less) (19). P.409 However, high doses o f inhaled agonists (e.g., 20 mg al uterol) have een used as a treatment for hyp erkalemia (20), so aggressive therapy with inhaled agonists could result in sign ificant hypokalemia. THEOPHYLLINE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 22)03-May-05 17:24:28

Ovid: ICU Book Theophylline has een a popular ronchodilator in the United States for over 50 years, ut the drug has little value in the acute care setting (21). Theophyllin e is less effective and more likely to produce undesira le side effects than the -agonist aerosols (1,22,23). This is reiterated in the statements shown in Ta le 25.3, which are taken from the National Asthma Education Program (1). As shown in Figure 25.5, theophylline adds little when it is given in com ination with -ag onist aerosols for acute asthma (22,23). TABLE 25.3. EXPERT COMMENTARY ON THEOPHYLLINE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 22)03-May-05 17:24:28

Ovid: ICU Book Figure 25.5. Changes in peak expiratory flow rate (PEFR) in patients with acute exacer ation of asthma treated with inhaled metaproterenol and intravenous hydro cortisone, with and without intravenous aminophylline. Each point represents the mean value for all patients in each treatment group. (From Wrenn K, et al. Amin ophylline therapy for acute ronchospastic disease in the emergency room. Ann In tern Med 1991;115:241247.) Theophylline ecomes even more undesira le when you consider the cost of monitor ing serum theophylline levels (a out $50 per determination). There are an estima ted 500,000 hospital admissions annually for asthma (24), and if serum theophyll ine levels were measured at each admission and discharge, the yearly cost would e $25 million. DRUG ADMINISTRATION Although there are no clear indications for theophylline in ICU patients, the dr ug continues to e used, so I have included some dose recommendations in Ta le 2 5.4. TABLE 25.4. INTRAVENOUS AMINOPHYLLINE DOSING Aminophylline When theophylline is given intravenously, a solvent must e added to maintain so lu ility. The solvent is ethylenediamine (also found in shellacs and pesticides) , and the theophylline ethylenediamine preparation is called aminophylline. Hyper sensitivity reactions to ethylenediamine have een reported (25), and for sensit ized patients, there is a solution of theophylline in dextrose availa le for int ravenous use (Travenol La s). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 22)03-May-05 17:24:28

Ovid: ICU Book P.410 Drug Doses The dose regimen in Ta le 25.4 is meant to achieve a serum theophylline level of approximately 10 mg/L, which is the low end of the therapeutic range. The thera peutic range for serum theophylline is 10 to 20 mg/L, or 55.5 to 111 mol/L (1 mg/ L = 5.5 mol/L). The loading dose in o ese patients is usually given according to ideal ody weight, ut the appropriate dosing weight in o esity is somewhere et ween ideal and actual ody weight (26). Therefore, monitoring serum theophylline levels is very important in o ese patients. Note also the marked reduction in t heophylline dose for patients with heart failure. Dose adjustments may also e n ecessary in patients receiving drugs that can alter serum theophylline levels (s ee Chapter 54 for the drugs that interact with theophylline). THEOPHYLLINE TOXICITY Theophylline is the fifth leading cause of (pharmaceutical) drug-induced toxicit y, and theophylline toxicity is the fifth leading cause of death from toxic drug exposure (27). One report revealed that physicians' prescri ing errors were res ponsi le for 68% of the cases of in-hospital theophylline toxicity (28). Manifestations Theophylline toxicity is defined as any serum theophylline level a ove 20 mg/L. Clinical manifestations of toxicity are present in most patients with serum leve ls of 20 to 30 mg/L, and in virtually all patients with drug levels a ove 30 mg/ L (29). Common manifestations P.411 in mild toxicity include tachycardia, tremor s, nausea, diarrhea, and agitation. More severe cases of toxicity are characteri zed y confusion, hypotension, hypokalemia, and supraventricular tachycardia. Th e severest cases are characterized y generalized seizures, coma, and ventricula r fi rillation. The severity of the clinical manifestations does not always corr elate with the degree of elevation in serum theophylline levels (29,30). Management The management of theophylline toxicity can e divided into three stages. Stage 1: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 22)03-May-05 17:24:28

Ovid: ICU Book The first step is to discontinue drug administration. In mild cases of toxicity, therapy can e restarted at a lower dose. Stage 2. When the toxic manifestations are more severe (e.g., change in mental status), o ral activated charcoal is used to enhance theophylline clearance. Charcoal in th e lumen of the gastrointestinal (GI) tract can actually speed up theophylline cl earance from the loodstream ecause the charcoal a sor s some of the drug that circulates ack and forth across the GI mucosa (GI dialysis). The dose of charco al is 20 g every 2 hours, with 75 mL of 70% sor itol added to every other dose t o facilitate passage of the charcoal (30). Stage 3: In cases of life-threatening theophylline toxicity (e.g., serious arrhythmias, s eizures, or a serum theophylline level a ove 100 mg/L), charcoal or resin hemope rfusion is recommended to enhance theophylline clearance. If these measures are not availa le, hemodialysis is an effective alternative. Peritoneal dialysis doe s not enhance theophylline clearance. More specific measures may e necessary in addition to the measures aimed at removing theophylline. Although seizures are usually not recurrent (29), short-term therapy with anticonvulsants is recommend ed until serum theophylline is reduced to safe levels. Cardiac arrhythmias can e treated with selective -1 receptor antagonists (e.g., esmolol) without precipit ating ronchospasm (31). Theophylline-induced hypotension may e refractory to c onventional vasopressors. CORTICOSTEROIDS Corticosteroids have enjoyed a long-standing popularity in the management of ast hma that is refractory to ronchodilator therapy. The relative features of the a ntiinflammatory corticosteroids are shown in Ta le 25.5. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 22)03-May-05 17:24:28

Ovid: ICU Book TABLE 25.5. A COMPARISON OF ANTIINFLAMMATORY CORTICOSTEROIDS ASTHMA Acute asthma is no longer viewed as a purely ronchospastic disorder. The ronch ospasm is an early manifestation, ut lasts only 30 to P.412 60 minutes. A secon d episode of airway o struction occurs 3 to 8 hours later, and is caused y infl ammation and edema in the walls of the small airways (32). Thus, ronchodilators should e effective in the early stages of acute asthma, whereas anti-inflammat ory corticosteroids should e effective in relieving the o struction that occurs later in the disorder. The effect of steroids on the delayed o struction would explain why steroid effects in acute asthma often require 6 to 8 hours to ecome evident. Which Agent at What Dose? As demonstrated in Figure 25.4, acute asthma that is refractory to three success ive -agonist aerosol treatments should e treated with an intravenous corticoster oid. The recommendations of the National Asthma Education Program (1) are methyl prednisolone, 80 to 125 mg iv push (then 80 mg iv every 6 to 8 hours), or hydroc ortisone, 2 mg/kg iv push (then 2 mg/kg iv every 4 hours). Although there is no evidence that one corticosteroid is superior to the others for the management of acute asthma, methylprednisolone might e favored over hydrocortisone ecause o f its superior anti-inflammatory activity (Ta le 25.5). The appropriate dose of corticosteroids is also unknown, ut most recommendations favor high-dose regime ns. When patients sta ilize on intravenous steroids, they can e switched to ora l prednisone starting at a http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 22)03-May-05 17:24:28

Ovid: ICU Book

What to Expect Do not expect miracles. In fact, despite the overwhelming popularity of steroid use in acute asthma, the reported experience with steroid therapy in acute asthm a is far from an endorsement of merit. A num er of clinical studies have shown n o enefit with steroid therapy in acute asthma (see references 33,34,35 and 36). These studies provide a more alanced perspective of steroid therapy in acute a sthma. P.413 CHRONIC OBSTRUCTIVE PULMONARY DISEASE Corticosteroids are also popular in the management of patients with severe COPD. However, there is no evidence that corticosteroids are eneficial in COPD (36). Because one of the characteristics of COPD is the lack of airway response to r onchodilators, it is unreasona le to expect an airway response to pharmaceutical manipulation with corticosteroids in COPD. SIDE EFFECTS Although chronic steroid therapy is associated with multiple side-effects, short -term administration of steroids (i.e., 1 month or less) is relatively free of a dverse reactions. Mental Status Acute therapy with high-dose steroids is often accompanied y a sense of euphori a and well eing. Although psychotic reactions are often mentioned as a complicat ion of high-dose steroids (18), this phenomenon is not well documented. Myopathy A myopathy has een reported in ventilator-dependent asthmatic patients treated with high-dose steroids and neuromuscular locking agents (37). Unlike the tradi tional steroid myopathy, which is characterized y proximal muscle weakness, the myopathy in ventilator-dependent asthmatics involves oth proximal and distal m uscles, and is often associated with rha domyolysis (37). The etiology of this d estructive myopathy is unknown, ut the com ination of steroids and paralyzing d rugs is somehow involved. The muscle weakness can e prolonged and can hamper we aning from mechanical ventilation. Once the disorder is suspected, rapid taper o f the paralyzing agents and steroids is advised. Fortunately, this disorder is r eversi le. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 22)03-May-05 17:24:28

daily dose of 60 mg. The oral dose is then tapered as dictated ituation.

y the clinical s

Ovid: ICU Book MUCOKINETIC THERAPY The goal of mucokinetic therapy is to facilitate the clearance of respiratory se cretions and remove any luminal arriers to airflow (38). This approach is limit ed primarily to patients who require prolonged mechanical ventilation. BLAND AEROSOLS The principal function of aerosols in clinical medicine is to introduce a pharma ceutical agent into the upper airways. Aerosols that do not P.414 contain a phar maceutical agent are called land aerosols (38). Most land aerosols are made up of saline solutions (hypotonic, isotonic, and hypertonic) or distilled water. B land aerosols are used to induce coughing (hypertonic aerosols) and to liquefy t hick and tenacious secretions. The latter practice is common ut not justified, as descri ed elow. Sol versus Gel The respiratory secretions create a lanket that covers the mucosal surface of t he airways. This lanket has a hydrophilic region (sol phase) and a hydropho ic region (gel phase). The hydrophilic region faces inward, and helps keep the muco sal surface moist. The hydropho ic region faces the lumen of the airways. This r egion contains a meshwork of mucoprotein strands (called mucus threads) that is held together y disulfide ridges. This meshwork traps particles and de ris in the airways, and the com ination of the mucoprotein mesh and the trapped de ris is what determines the viscoelastic ehavior of the secretions. Because water ca nnot enter the phase of secretions that is responsi le for their viscoelastic e havior, the use of land aerosols to liquefy and disrupt secretions is futile. MUCOLYTIC THERAPY The disruption of tenacious secretions is accomplished y disrupting the onds t hat link the mucoprotein strands together. Disruption of disulfide ridges etwe en mucoprotein strands is the asis for mucolytic therapy with N-acetylcysteine (Mucomyst) (39), a sulfhydryl-containing tripeptide that is etter known as the antidote for acetaminophen overdose (see Fig. 53.1). Nacetylcysteine (NAC) is av aila le in a liquid preparation (10 or 20% solution) that can e given as an aer osol spray, or injected directly into the airways (Ta le 25.6). P.415 Aerosolize d NAC should e avoided when possi le ecause it is irritating to the airways an d can provoke coughing and ronchospasm (particularly in asthmatics) and ecause it has a http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 22)03-May-05 17:24:28

Ovid: ICU Book disagreea le taste ( ecause of the sulfur content of NAC) and can provoke nausea and vomiting in patients who are not intu ated. Direct instillation of NAC into the tracheal tu e (two or three times a day) can e used in select patients wit h pro lems clearing secretions (I use this approach for postoperative patients w ho require ronchoscopy to clear o structing secretions). Excess volumes of intr atracheal NAC should e avoided ecause the solution is hypertonic (even with th e saline additive) and can provoke ronchorrhea. TABLE 25.6. MUCOLYTIC THERAPY WITH N-ACETYLCYSTEINE (NAC) REFERENCES REVIEWS 1. National Asthma Education Program Expert Panel Report. Guidelines for the dia gnosis and management of asthma. Bethesda, MD: U.S. Department of Health and Hum an Services, Pu lic Health Service. Pu . No. 91-3042, August, 1991. 2. Kacmarek RM. Humidity and aerosol therapy. In: Pierson DJ, Kacmarek RM, eds. Foundations of respiratory care. New York: Churchill Livingstone, 1992;793824 (17 5 references). 3. McFadden ER Jr, Hejal R. Asthma. Lancet 1995;345:12151220 (60 references). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 22)03-May-05 17:24:28

Ovid: ICU Book 4. Manthous CA. Management of severe exacer ations of asthma. Am J Med 1995;99:2 98308 (89 references). P.416 BEDSIDE MONITORING 5. Leiner GC. Expiratory peak flow rate. Standard values from normal su jects. A m Rev Respir Dis 1963;88:644. 6. Quacken oss JJ, Le owitz MD, Kryzyzanowski M. The normal range of diurnal cha nges in peak expiratory flow rate. Am Rev Respir Dis 1991;143:323330. 7. Mendoza GR. Peak flow monitoring. J Asthma 1991;28:161177. 8. Li JTC. Home peak expiratory flow rate monitoring in patients with asthma. Ma yo Clin Proc 1995;70:649656. 9. Gay PC, Rodarte JR, Tayya M, Hu mayr RD. Evaluation of ronchodilator respon siveness in mechanically ventilated patients. Am Rev Respir Dis 1987;136:880885. BETA-RECEPTOR AGONISTS 10. Salmeron S, Brochard L. Mal H, et al. Ne ulized versus intravenous al uterol in hypercapnic acute asthma. Am J Respir Crit Care Med 1994;149:14661470. 11. Clarke SW, Newman SP. Differences etween pressurized aerosol and sta le dus t particles. Chest 1981;80(Suppl):907908. 12. Idris AH, McDermott MF, Raucci JC, et al. Emergency department treatment of severe asthma. Metered-dose inhaler plus holding cham er is equivalent in effect iveness to ne ulizer. Chest 1993;103:665672. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 22)03-May-05 17:24:28

Ovid: ICU Book 13. Manthous CA, Hall JB. Update on using therapeutic aerosols in mechanically v entilated patients. J Crit Illness 1996;11:457468. 14. Gay PC, Patel HG, Nelson SB, et al. Metered dose inhalers for ronchodilator delivery in intu ated, mechanically ventilated patients. Chest 1991;99:6671. 15. Bowton DL, Goldsmith WM, Haponik EF. Su stitution of metered-dose inhalers f or hand-held ne ulizers. Chest 1992;101:305308. 16. Taylor RH, Lerman J. High-efficiency delivery of sal utamol with a metered-d ose inhaler in narrow tracheal tu es and catheters. Anesthesiology 1991;74:360363 . 17. O'Riordan TG, Palmer LB, Smaldone GC. Aerosol deposition in mechanically ven tilated patients. Am Rev Respir Crit Care Med 1994;149:214219. 18. Truwit JD. Toxic effect of ronchodilators. Crit Care Clin 1991;7:639657. 19. Bodenhamer J, Bergstrom R, Brown D, et al. Frequently ne ulized eta-agonist s for asthma: effects on serum electrolytes. Ann Emerg Med 1992;21:13371342. 20. Allon M, Dunlay R, Copkney C. Ne ulized al uterol for acute hyperkalemia in patients on hemodialysis. Ann Intern Med 1989;110:426429. THEOPHYLLINE 21. Johnson ID. Theophylline in the management of airflow o struction. Difficult drug to use, few clinical indications. Br Med J 1990;300:929931. 22. Self TH, A ou-Shala N, Burns R, et al. Inhaled al uterol and oral prednisone therapy in hospitalized adult asthmatics. Does aminophylline add any enefit? C hest 1990;98:1317 1321.

23. Rodrigo C, Rodrigo G. Treatment of acute asthma: lack of therapeutic and http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 22)03-May-05 17:24:28

enefit

Ovid: ICU Book increase in toxicity from aminophylline given in addition to high dose sal utamo l delivered y metered-dose inhaler with spacer. Chest 1994;106:10711076. P.417 24. Weiss KB. Seasonal trends in U.S. asthma hospitalizations and mortality. JAM A 1990;263:23232328. 25. Terzian CG, Simon PA. Aminophylline hypersensitivity apparently due to ethyl enediamine. Ann Emerg Med 1992;21:312317. 26. Rizzo A, Mira ella A, Bonanno A. Effect of ody weight on the volume of dist ri ution of theophylline. Lung 1988;166:269276. 27. Litovitz TL. 1992 Annual Report of the American Association of Poison Contro l Centers Toxic Exposure Surveillance System. Am J Emerg Med 1993;11:494555. 28. Schiff GD, Hegde HK, LaCloche L, Hryhorczuk DO. Inpatient theophylline toxic ity: preventa le factors. Ann Intern Med 1991;114:748752. 29. Sissler CN. Theophylline toxicity: clinical features of 116 consecutive case s. Am J Med 1990;88:567576. 30. Cooling DS. Theophylline toxicity. J Emerg Med 1993;11:415425. 31. Seneff M, Scott J, Friedman B, et al. Acute theophylline toxicity and the us e of esmolol to reverse cardiovascular insta ility. Ann Emerg Med 1990;19:671673. CORTICOSTEROIDS 32. Kay AB. Asthma and inflammation. J Allergy Clin Immunol 1991;87:893945. 33. Stein LM, Cole RP. Early administration of corticosteroids in emergency room http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 22)03-May-05 17:24:28

Ovid: ICU Book treatment of asthma. Ann Intern Med 1990;112:822827. 34. Bowler SD, Mitchell CA, Armstrong JG. Corticosteroids in acute severe asthma : effectiveness of low doses. Thorax 1992;47:584587. 35. Morrell F, Orriols R, de Gracia J, et al. Controlled trial of intravenous co rticosteroids in severe acute asthma. Thorax 1992;47:588591.

37. Griffin D, Fairman N, Coursin D, et al. Acute myopathy during treatment of s tatus asthmaticus with corticosteroids and steroidal muscle relaxants. Chest 199 2;102:510-514. MUCOKINETIC THERAPY 38. Le ovitz DJ, Reed MD. Clinical pharmacology of mucokinetic drugs. In: Cherno w B, ed. The pharmacologic approach to the critically ill patient. 3rd ed. Balti more: Williams & Wilkins, 1994;605613. 39. Holdiness MR. Clinical pharmacokinetics of N-acetylcysteine. Clin Pharmacoki net 1991;20:123134. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 22)03-May-05 17:24:28

36. Eliasson O, Hoffman J, True 9:484 489.

D, et al. Corticosteroids in COPD. Chest 1986;8

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 26 PRINCIPLES OF MECHANICAL VENTILATION an opening must e attempted in the trunk of the trachea, into which a tu e of re ed or cane should e put; you will then low into this, so that the lung may ris e again and the heart ecomes strong --Andreas Vesalius (1555) Vesalius is credited with the first description of positive-pressure ventilation, ut it took 400 ye ars to apply his concept to patient care. The occasion was the polio epidemic of 1955, when the demand for assisted ventilation outgrew the supply of negative-p ressure tank ventilators (known as iron lungs). In Sweden, all medical schools s hut down and medical students worked in 8-hour shifts as human ventilators, manu ally inflating the lungs of afflicted patients. In Boston, the near y Emerson Co mpany made availa le a prototype positive-pressure lung inflation device, which was put to use at the Massachusetts General Hospital, and ecame an instant succ ess. Thus egan the era of positive-pressure mechanical ventilation (and the era of intensive care medicine). CONVENTIONAL MECHANICAL VENTILATION The first positive-pressure ventilators were designed to inflate the lungs until a preset pressure was reached. This type of pressure-cycled ventilation fell ou t of favor ecause the inflation volume varied with changes in the mechanical pr operties of the lungs. In contrast, volumecycled ventilation, which inflates the lungs to a predetermined volume, delivers a constant alveolar volume despite ch anges in the mechanical properties of the lungs. For this reason, volume-cycled ventilation has ecome the standard method of positive-pressure mechanical venti lation (1,2,3,4 and 5). P.422 INFLATION PRESSURES http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 17)03-May-05 08:25:47

Ovid: ICU Book The waveforms produced y volume-cycled ventilation are shown in Figure 26.1. Th e lungs are inflated at a constant flow rate, and this produces a linear increas e in lung volume. The pressure in the proximal airways (Pprox) shows an a rupt i nitial rise, followed y a more gradual rise through the remainder of lung infla tion. However, the pressure in the alveoli (PALV) shows only a gradual rise duri ng lung inflation. Figure 26.1. Waveforms produced y constant-flow, volume-cycled ventilation. V = inspiratory flow rate, V = inspiratory volume, R = flow resistance in the airwa ys, Pprox = proximal airway pressure, PALV = alveolar pressure. The early, a rupt rise in proximal airway pressure is a reflection of flow resis tance in the airways. An increase in airways resistance magnifies the initial ri se in proximal airway pressure, while the alveolar pressure at the end of lung i nflation remains unchanged. Thus, when resistance in the airways increases, high er inflation pressures are needed to deliver the inflation volume, ut the alveo li are not exposed to the higher inflation pressures. This is not the case when the distensi ility (compliance) of the lungs is reduced. In this latter conditio n, there is an increase in oth the proximal airways pressure and the alveolar p ressure. Thus, when lung distensi ility (compliance) decreases, the higher infla tion pressures http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 17)03-May-05 08:25:47

Ovid: ICU Book needed to deliver the P.423 inflation volume are readily transmitted to the alve oli. The increase in alveolar pressure in noncompliant lungs can lead to pressur e-induced lung injury (see later in the chapter). CARDIAC PERFORMANCE The influence of positive-pressure ventilation on cardiac performance is complex and involves changes in preload and afterload for oth the right and left sides of the heart (6). To descri e these changes, it is important to review the infl uence of intrathoracic pressure on transmural pressure, which is the pressure th at determines ventricular filling (preload) and the resistance to ventricular em ptying (afterload). TRANSMURAL PRESSURE The transmission of intrathoracic pressure into the lumen of intrathoracic lood vessels is descri ed riefly in Chapter 11 (see Fig. 11.1). The influence of lu ng mechanics on this pressure transmission is illustrated in Figure 26.2. The pa nel on the left shows what happens when a normal lung is inflated with 700 mL fr om a positive-pressure source. In this situation, the increase in alveolar press ure is completely transmitted into the pulmonary capillaries, and there is no ch ange in transmural pressure (Ptm) across the capillaries. However, when the same lung inflation occurs in lungs that are not easily distended (panel on the righ t), the increase in alveolar pressure is not completely transmitted P.424 into t he capillaries and the transmural pressure increases. This increase in transmura l pressure acts to compress the capillaries. Therefore, in conditions associated with a decrease in lung compliance (e.g., pulmonary edema, pneumonia), positive -pressure lung inflation tends to compress the heart and intrathoracic lood ves sels (7,8 and 9). This compression can e eneficial or detrimental, as descri e d elow. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 17)03-May-05 08:25:47

Ovid: ICU Book Figure 26.2. Alveolarcapillary units showing the transmission of alveolar pressur e (PALV) to the pulmonary capillaries in normal and noncompliant (stiff) lungs. Pc = capillary hydrostatic pressure, Ptm = transmural pressure across the capill ary wall, VT = tidal volume delivered y the ventilator. Preload Positive-pressure lung inflation can reduce ventricular filling in several ways, as indicated in Figure 26.3. First, positive intrathoracic pressure decreases t he pressure gradient for venous inflow into the thorax. In addition, any increas e in positive pressure on the outer surface of the ventricles will reduce ventri cular distensi ility, and this can reduce ventricular filling during diastole. F inally, compression of pulmonary lood vessels can reduce left ventricular filli ng y decreasing venous inflow to the left side of the heart, or y impeding rig ht heart ejection. In this latter situation, the right ventricle can dilate and push the interventricular septum toward the left ventricle and reduce left ventr icular cham er size. This phenomenon, known as ventricular interdependence, is o ne of the mechanisms where y right heart failure can impair the performance of t he left side of the heart (see Fig. 16.4). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 17)03-May-05 08:25:47

Ovid: ICU Book Figure 26.3. The mechanisms where y positive-pressure mechanical ventilation can decrease ventricular filling (preload). Afterload Whereas ventricular compression during positive-pressure lung inflation impedes ventricular filling during diastole, this same compression facilitates ventricul ar emptying during systole. This latter effect is P.425 easy to visualize (like a hand squeezing the ventricles during systole) and can also e explained in ter ms of ventricular afterload. That is, ventricular afterload, or the impedance to ventricular emptying, is a function of the peak systolic transmural wall pressu re (see Fig. 1.3). Incomplete transmission of positive intrathoracic pressure in to the ventricular cham ers will decrease the transmural pressure across the ven tricles during systole, and this decrease ventricular afterload. CARDIAC OUTPUT Positive-pressure lung inflation tends to reduce ventricular filling during dias tole ut enhances ventricular emptying during systole. When ventricular filling is reduced, cardiac output is similarly reduced. However, when ventricular filli ng is not compromised, positive-pressure lung inflation can increase cardiac str oke output. The increase in stroke volume causes an increase http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 17)03-May-05 08:25:47

Ovid: ICU Book in systolic lood pressure during positive-pressure lung inflation; a phenomenon known as reverse pulsus paradoxus. The favora le influence of positive intratho racic pressure on cardiac output is one mechanism that explains the eneficial a ctions of closed chest compressions in cardiac arrest (see Chapter 17). INDICATIONS FOR MECHANICAL VENTILATION The decision to intu ate and initiate mechanical ventilation has always seemed m ore complicated than it should e. Instead of the usual lists of clinical and ph ysiologic indications for mechanical ventilation, the following simple rules sho uld suffice. Rule 1. The indication for intu ation and mechanical ventilation is thinking of it. There is a tendency to delay intu ation and mechanical ventilat ion as long as possi le in the hopes that it will e unnecessary. However, elect ive intu ation carries far fewer dangers than emergent intu ation, and thus dela ys in intu ation create unnecessary dangers for the patient. If the patient's co ndition is severe enough for intu ation and mechanical ventilation to e conside red, then proceed without delay. Rule 2. Intu ation is not an act of weakness. H ousestaff tend to apologize on morning rounds when they have intu ated a patient during the evening, almost as though the intervention was an act of cowardice. What they must understand is that they will never e faulted for esta lishing co ntrol of the airways. Rule 3. Endotracheal tu es are not a disease, and ventilat ors are not an addiction. The assumption that once on a ventilator, always on a v entilator is a fallacy and should never influence the decision to initiate mechan ical ventilation. Endotracheal tu es and ventilators do not create the need for mechanical ventilation; cardiopulmonary and neuromuscular diseases do. P.426 VENTILATORY STRATEGIES Thirty years after positive-pressure mechanical ventilation was introduced on a large scale, it ecame evident that the traditional methods of assisted ventilat ion can actually add to the underlying cardiopulmonary derangements. This realiz ation has led to a revised strategy for mechanical ventilation (3), which was in troduced in Chapter 23 and is outlined in Ta le 26.1. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 17)03-May-05 08:25:47

Ovid: ICU Book TABLE 26.1. STRATEGIES FOR MECHANICAL VENTILATION LARGE INFLATION VOLUMES In the early days of positive-pressure mechanical ventilation, large inflation v olumes were recommended to prevent alveolar collapse (10), and these large infla tion volumes ecame standard. Thus, whereas the tidal volume during unassisted v entilation in adults is normally 5 to 7 mL/kg (ideal ody weight), the standard inflation volumes during volumecycled ventilation are 10 to 15 mL/kg (i.e., at l east twice normal). The volume discrepancy is even greater with the addition of mechanical sighs, which are 1.5 to two times greater than standard inflation vol umes (or 15 to 30 mL/kg) and are delivered 6 to 12 times per hour (Ta le 26.1). The large inflation volumes used in conventional mechanical ventilation can redu ce cardiac output (8,9). Large inflation volumes can also damage the lungs, as d escri ed elow. Ventilator-Induced Lung Injury In lung diseases that most often require mechanical ventilation (e.g., acute res piratory distress syndrome [ARDS], pneumonia), the pathologic changes are not un iformly distri uted throughout the lungs (11). Because inflation volumes are dis tri uted preferentially to regions of normal lung function, inflation volumes te nd to overdistend the normal regions of diseased lungs. This tendency to overdis tend normal lung regions is exaggerated when large inflation volumes are used. P .427 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 17)03-May-05 08:25:47

Ovid: ICU Book The hyperinflation of normal lung regions during mechanical ventilation can prod uce stress fractures in the walls of alveoli and adjacent pulmonary capillaries (12,13). The alveolar damage can lead to alveolar rupture, with accumulation of alveolar gas in the pulmonary parenchyma (pulmonary interstitial emphysema), med iastinum (pneumomediastinum), or pleural cavity (pneumothorax). The damage to th e pulmonary capillaries can result in a leaky-capillary type of pulmonary edema (14). These complications may e the result of excessive alveolar pressures ( ar otrauma) or excessive alveolar volumes (volutrauma) (15). LUNG-PROTECTIVE VENTILATION Because of the risk of lung injury with large inflation volumes, an alternative approach evolved in recent years where reduced inflation volumes (5 to 10 mL/kg) are advocated (Ta le 26.1) (3). This strategy also eliminates mechanical sighs and uses positive end-expiratory pressure (PEEP) to prevent collapse of alveoli and small airways (16). The goal is an end-inspiratory plateau pressure that is less than 35 cm H2O (this pressure is descri ed later in this chapter). Inflatio n volumes of 8 mL/kg or lower can result in CO2 retention, which is allowed if n o evidence of harm exists. This latter strategy is known as permissive hypercapn ia (17). MONITORING LUNG MECHANICS During spontaneous reathing, the mechanical properties of the lungs (i.e., elas tic recoil of the lungs and resistance to flow in the airways) can e monitored with pulmonary function tests. However, these tests are not easily performed dur ing mechanical ventilation. In this setting, the proximal airways pressures can e used to assess pulmonary function (18,19). PROXIMAL AIRWAY PRESSURES Positive-pressure mechanical ventilators have a pressure gauge that monitors the proximal airway pressure during each respiratory cycle. The components of this pressure are illustrated in Figure 26.4. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 17)03-May-05 08:25:47

Ovid: ICU Book Figure 26.4. Proximal airway pressures at the end of a positive-pressure lung in flation. The plateau pressure is produced y occluding the tracheal tu e or vent ilator tu ing to prevent lung deflation. These pressures provide a qualitative a ssessment of lung mechanics at the edside (see Fig. 26.5). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 17)03-May-05 08:25:47

Ovid: ICU Book Figure 26.5. The use of proximal airway pressures to evaluate an acute change in ventilatory status at the edside. End-Inspiratory Peak Pressure The peak pressure at the end of inspiration (Ppeak) is a function of the inflati on volume, the flow resistance in the airways, and the elastic recoil force of t he lungs and chest wall. At a constant inflation volume, the peak pressure is di rectly related to airflow resistance and to the elastic recoil force (elastance) of the lungs and chest wall: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 17)03-May-05 08:25:47

Ovid: ICU Book P.428 Therefore when the inflation volume is constant, an increase in peak inspi ratory pressure indicates an increase in either airway resistance or elastance o f the lungs and chest wall, or oth. End-Inspiratory Plateau Pressure The contri ution of resistance and elastance to the peak inspiratory pressure ca n e distinguished y occluding the expiratory tu ing at the end of inspiration, as shown in Figure 26.4. When the inflation volume is held in the lungs, the pr oximal airway pressure decreases initially and then reaches a steady level, whic h is called the end-inspiratory plateau pressure. Because no airflow is present when the plateau pressure is created, the pressure is not a function of flow res istance in the airways. Instead, the plateau pressure (Pplateau) is directly pro portional to the elastance of the lungs and chest wall. P.429 Therefore, the difference etween end-inspiratory peak and plateau pressur es is proportional to the flow resistance in the airways. Practical Applications The flow diagram in Figure 26.5 demonstrates how the proximal airways pressures can e applied to patient care. In this case, the pressures are used to evaluate an acute deterioration in respiratory status. q If the peak pressure is increased ut the plateau pressure is unchanged, the pro lem is an increase in airways resistance. In this situation, the major concerns are o struction of the tracheal tu e, airway o struction from secretions, and a cute ronchospasm. Therefore, airways suctioning is indicated to clear secretion s, followed y an aerosolized ronchodilator treatment if necessary. q If the peak and plateau pressures are oth increased, the pro lem is a decrease in distensi ility of the lungs and chest wall. In this situation, the major conc erns are http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 17)03-May-05 08:25:47

Ovid: ICU Book pneumothorax, lo ar atelectasis, acute pulmonary edema, and worsening or ARDS. Active contraction of the chest wall and increased a dominal an also decrease the distensi ility of the thorax. Finally, a patient uctive lung disease who ecomes tachypneic can develop auto-PEEP, and ases the peak and plateau pressures as well (see Chapter 28). q pneumonia pressure c with o str this incre

If the peak pressure is decreased, the pro lem may e an air leak in the system (e.g., tu ing disconnection, cuff leak). In this situation, the lungs should e manually inflated with an Am u ag and a cuff leak should e listened for during lung inflation. A decrease in peak pressure can also e due to hyperventilation , when the patient is generating enough of a negative intrathoracic pressure to p ull air into the lungs. q If no change in peak pressure occurs, it does not necessarily mean that there ha s een no change in lung mechanics. The sensitivity of the proximal airways pres sures in detecting changes in lung mechanics is unknown. When the pressures do n ot change, the evaluation should proceed as it would without the aid of proximal airway pressures. Bronchodilator Responsiveness Ventilator-dependent patients often receive aerosolized ronchodilator treatment s routinely, without documented need or enefit. The proximal airway pressures p rovide a more rational approach to the use of ronchodilators in ventilator-depe ndent patients. If an aerosol ronchodilator treatment does not produce a decrea se in peak inspiratory pressures (indicating ronchodilation), there is little j ustification for continuing the aerosol treatments. P.430 THORACIC COMPLIANCE The compliance, or distensi ility, of the lungs and chest wall (called thoracic compliance) can e determined quantitatively as the ratio of a change in lung vo lume (i.e., inflation volume) to a change in elastic recoil pressure (i.e., plat eau pressure). For a tidal volume (VT) of 800 mL and a plateau pressure (Ppl) of 10 cm H2O, the static compliance (Cstat) of the thorax is as follows: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 17)03-May-05 08:25:47

Ovid: ICU Book P.431 The thoracic compliance in intu ated patients with no known lung disease i s etween 0.05 to 0.08 L/cm H2O (or 50 to 80 mL/cm H2O) (18). In patients with s tiff lungs, the thoracic compliance is much lower at 0.01 to 0.02 L/cm H2O (20). Thus, the compliance determination provides an o jective measure of the severit y of illness in pulmonary disorders associated with a change in lung compliance. Considerations The following factors influence the static compliance measurement. q PEEP increases the plateau pressure. Therefore, the level of PEEP (either extern ally applied or auto-PEEP) should e su tracted from the plateau pressure for th e compliance determination. q The connector tu ing etween the ventilator and patient expands during positivep ressure lung inflations, and the volume lost in this expansion reduces the infla tion volume reaching the patient. The volume lost is a function of the peak infl ation pressure and the inherent compliance of the tu ing. The usual compliance o f connector tu ing is 3 mL/cm H2O, which means that 3 mL of volume is lost for e very 1 cm H2O increase in inflation pressure. Thus, if the inflation volume from the ventilator is 700 mL and the peak inspiratory pressure is 40 cm H2O, then ( 3 40) 120 mL will e lost to expansion of the ventilator tu ing and the inflatio n volume reaching the patient will e (700 120) 580 mL. Because of this discrepa ncy, the predetermined volume setting on the ventilator should not e used as th e inflation volume for the compliance calculation. Instead, the exhaled volume, which is usually displayed digitally on the ventilator panel, should e used. q Because the proximal airways pressures are transthoracic pressures (i.e., measur ed relative to atmospheric pressure) and not transpulmonary pressures (i.e., mea sured relative to intrapleural pressure), the compliance measurement includes th e chest wall as well as the lungs. Because contraction of the chest wall muscles can reduce the compliance (distensi ility) of the chest wall, the compliance de termination should e performed only during passive ventilation. However during passive ventilation, the chest wall can account for 35% of the total thoracic co mpliance (19, 20). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 17)03-May-05 08:25:47

Ovid: ICU Book AIRWAY RESISTANCE The resistance to airflow during inspiration (Rinsp) can e determined as the ra tio of the pressure gradient needed to overcome airways P.432 resistance (Ppk Pp l) and the inspiratory flow rate (Vinsp): Rinsp = Ppk Ppl/Vinsp. This resistance represents the summed resistances of the connector tu ing, the tracheal tu e, a nd the airways. Because the resistance of the connector tu ing and tracheal tu e should remain constant (assuming the tracheal tu e is clear of secretions), cha nges in Rinsp should represent changes in airways resistance. A sample calculati on of inspiratory resistance is shown in Equation 26.5 for an inspiratory flow o f 60 L/min (1.0 L/sec), a peak pressure of 20 cm H2O, and a plateau pressure of 10 cm H2O. The minimal flow resistance in large- ore endotracheal tu es is 3 to 7 cm H2O/L/ sec (21), so nonpulmonary resistive elements can contri ute a considera le fract ion of the total inspiratory resistance. Limitations The major limitations of the inspiratory resistance measurement are the contri u tions of resistive elements not in the lung and the relative insensitivity of th e resistance measured during inspiration. Airflow o struction is usually measure d during expiration, when the airways have the greatest tendency to collapse. Th e distending pressures delivered y the ventilator during lung inflation keep th e airways open, and thus the resistance to flow during inspiration does not meas ure the tendency for the airways to collapse during expiration (19). REFERENCES GENERAL TEXTS Grenvik A, Downs J, Rasanen J, Smith R, eds. Mechanical ventilation and assisted respiration. Contemporary management in critical care. New York: Churchill Livi ngstone, http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 17)03-May-05 08:25:47

Ovid: ICU Book 1991;1(1). Hess DR, Kacmarek RM. Essential of mechanical ventilation. New York: McGraw-Hill , 1996. REVIEWS 1. Slutsky AS (chairman). American College of Chest Physicians' Consensus Confer ence on Mechanical Ventilation. Chest 1993;104:18331859 (158 references). 2. To in MJ. Mechanical ventilation. N Engl J Med 1994;330:10561061 (58 reference s). P.433 3. Marini JJ. Pressure-targeted, lung protective ventilatory support in acute lu ng injury. Chest 1994;105(Suppl):109S115S (58 references). 4. Gammon RB, Strickland JH Jr, Kennedy KI Jr, Young KR Jr. Mechanical ventilati on: a review for the internist. Am J Med 1995;99:553562. 5. Shapiro BA, Peruzzi WT. Changing practices in ventilator management: a review of the literature and suggested clinical correlations. Surgery 1995;117:121133. CARDIAC PERFORMANCE 6. Pinsky MR. Cardiovascular effects of ventilatory support and withdrawal. Anae sth Analg 1994;79:567576. 7. Versprille A. The pulmonary circulation during mechanical ventilation. Acta A nesthesiol Scand 1990;34(Suppl):5162. 8. Venus B, Cohen LE, Smith RA. Hemodynamics and intrathoracic pressure transmis sion during controlled mechanical ventilation and positive end-expiratory http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 17)03-May-05 08:25:47

Ovid: ICU Book pressure in normal and low compliant lungs. Crit Care Med 1988;16:686690. 9. Kiiski R, Takala J, Kari A, Milic-Emili J. Effect of tidal volume on gas exch ange and oxygen transport in the adult respiratory distress syndrome. Am Rev Res pir Dis 1992;146:11311135. VENTILATORY STRATEGIES 10. Bendixen HH, Eg ert LD, Hedley-White J, et al. Respiratory care. St. Louis: Mos y, 1965;137153. 11. Gattinoni L, Bom ino M, Pelosi P, et al. Lung structure and function in diff erent stages of severe adult respiratory distress syndrome. JAMA 1994;271:1772177 9. 12. Costello ML, Mathieu-Costello OA, West JB. Stress failure of alveolar epithe lial cells studied y scanning electron microscopy. Am Rev Respir Dis 1992;145:1 4461455. 13. Mathieu-Costello OA, West JB. Are pulmonary capillaries suscepti le to mecha nical stress? Chest 1994;105(Suppl):102S107S. 14. Tim y J, Reed C, Zeilander S, Glauser F. Mechanical causes of pulmonary edema. Chest 1990;98:973979. 15. Bray JC, Cane RD. Mechanical ventilatory support and pulmonary parenchymal i njury: positive airway pressure or alveolar hyperinflation? Intensive Crit Care Digest 1993;12:3336. 16. Muscedere JG, Mullen JBM, Gan K, Slutsky AS. Tidal ventilation at low airway pressures can augment lung injury. Am J Respir Crit Care Med 1994;149:13271334. 17. Bidani A, Tzounakis AE, Cardenas VJ, Zwischen erger JB. Permissive hypercapn ia in acute respiratory failure. JAMA 1994;272:957962. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 17)03-May-05 08:25:47

Ovid: ICU Book MONITORING LUNG MECHANICS 18. To in MJ. Respiratory monitoring. JAMA 1990;264:244251. 19. Marini JJ. Lung mechanics determinations at the edside: instrumentation and clinical application. Respir Care 1990;35:669696. 20. Katz JA, Zinn SE, Ozanne GM, Fairley BB. Pulmonary, chest wall, and lungthora x elastances in acute respiratory failure. Chest 1981;80:304311. 21. Marini JJ. Strategies to minimize reathing effort during mechanical ventila tion. Crit Care Clin 1990;6:635662. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 17)03-May-05 08:25:47

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 27 PATTERNS OF ASSISTED VENTILATION Development in most fields of medicine appears to occur according to sound scien tific principles. However, exceptions can e found, and the development of mecha nical ventilatory support is one of them. --J. Rasanen In the half century since positive-pressure ventilation first appeared, at least 15 different methods of assisted reathing have een introduced, each with claims of superiority over th e others (1,2 and 3). However, these claims usually dissolve in clinical trials. There is a tendency to forget that mechanical ventilation is a support measure, not a therapy for cardiopulmonary disease. To improve outcomes in ventilator-de pendent patients, less attention should e paid to the kno s on ventilators, and more attention should e given to the diseases that prompt ventilator dependenc y. ASSIST-CONTROL VENTILATION The standard method of positive-pressure mechanical ventilation involves volumecycled lung inflation (i.e., each machine reath delivers a preselected inflatio n volume). The patient can initiate each mechanical reath (assisted ventilation ), ut when this is not possi le, the ventilator provides machine reaths at a p reselected rate (controlled ventilation). This pattern is called assist-control ventilation (ACV). VENTILATORY PATTERN The upper panel in Figure 27.1 shows the changes in airway pressure produced y ACV. The tracing egins with a negative-pressure deflection, followed y a posit ive-pressure machine reath. The negative P.435 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 18)03-May-05 08:26:24

Ovid: ICU Book pressure represents a spontaneous inspiratory effort, which opens a pressure-act ivated valve that allows the machine reath to e delivered. The second machine reath in the tracing is identical to the first, ut it is not preceded y a spo ntaneous ventilatory effort. The first reath is an example of assisted ventilat ion, and the second reath is an example of controlled ventilation. Figure 27.1. Airway pressure patterns in assist-control ventilation (ACV) and sy nchronized intermittent mandatory ventilation (SIMV). Spontaneous reaths are in dicated y dashed lines, and machine reaths are indicated y solid lines. The m echanical reath in the upper panel indicates the period of lung inflation (I) a nd deflation (E). Respiratory Cycle As mentioned in Chapter 26, conventional volume-cycled ventilation uses large in flation volumes (approximately twice the normal tidal volume during spontaneous reathing). To allow the lungs to empty these large volumes, much more time is a llowed for lung deflation than for lung inflation (Fig. 27.1). During convention al volume-cycled ventilation, the I:E ratio (the ratio of inspiratory-to-expirat ory duration) is maintained at 1:2 to 1:4 (1,2 and 3). This is accomplished y a djusting the inspiratory flow rate (e.g., an increase in the inspiratory flow ra te decreases the time for lung inflation and increases the I:E ratio). When the I:E ratio falls elow 1:2, the lungs may not empty completely after each inflati on. This can result in progressive hyperinflation. WORK OF BREATHING http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 18)03-May-05 08:26:24

Ovid: ICU Book A popular misconception a out mechanical ventilation is that the diaphragm rests when the ventilator delivers a machine reath. However, the contraction of the diaphragm is dictated y rainstem respiratory P.436 neurons that fire in period ic ursts throughout life (4). When the diaphragm contracts and triggers a machi ne reath, the contraction does not cease, ut continues for the duration of the urst activity in the rainstem respiratory neurons. Because the diaphragm is n ot a voluntary muscle that ceases to contract during assisted machine reaths, t he work of reathing can e considera le during assisted ventilation (5). ADVERSE EFFECTS The undesira le features of ACV occur primarily in patients who are reathing ra pidly. In this situation, the increased frequency of machine reaths can lead to overventilation and severe respiratory alkalosis, and the decreased time for ex halation can result in hyperinflation. Hyperinflation from inadequate lung empty ing is accompanied y an intrinsic form of PEEP called auto-PEEP, and this type of PEEP can have the same adverse consequences as externally-applied PEEP (which are descri ed later in the chapter). Hyperinflation and autoPEEP from excessive positive-pressure ventilation is a source of electromechanical dissociation dur ing CPR (6). INTERMITTENT MANDATORY VENTILATION The pro lems created y rapid reathing during mechanical ventilation with ACV l ed to the introduction of intermittent mandatory ventilation (IMV). This method, introduced in 1971, was designed to ventilate neonates with respiratory distres s syndrome, who typically have respiratory rates in excess of 40 reaths/minute. IMV delivers periodic volume-cycled reaths at a preselected rate, ut allows s pontaneous reathing etween machine reaths (see Fig. 27.1, lower panel). Becau se each spontaneous reath does not trigger a machine reath, there is a reduced risk of respiratory alkalosis and hyperinflation with IMV. Shortly after its in troduction, IMV was proposed as a valua le method for gradually withdrawing vent ilatory support (7). Since then, IMV has ecome a popular method of weaning pati ents from mechanical ventilation. The description of IMV in this chapter is limi ted to its use as a mode of ventilation. (The use of IMV in weaning from mechani cal ventilation is descri ed in Chapter 29.) BASIC FEATURES http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 18)03-May-05 08:26:24

Ovid: ICU Book The asic IMV circuit is illustrated in Figure 27.2. The patient is connected to a source of oxygen through two parallel pathways. One pathway contains a volume -cycled ventilator, and the other contains a reservoir ag filled with the desir ed inspiratory gas mixture. The ventilator delivers machine reaths at a presele cted rate and inflation volume. However, a unidirectional valve in the circuit a llows the patient to reathe spontaneously from the reservoir ag when a machine reath is not eing delivered. Figure 27.2. Schematic diagram of the circuit used for intermittent mandatory ve ntilation (IMV). P.437 Ventilatory Pattern The lower portion of Figure 27.1 shows the pattern of ventilation associated wit h IMV. The initial negative and positive deflection in the tracing represents a spontaneous reath (dashed line). The second spontaneous reath triggers a machi ne reath (solid line). This pattern is called synchronized IMV ecause the mach ine reaths are synchronized to coincide with spontaneous lung inflations. There is also an asynchronous IMV, when machine reaths are delivered at any time dur ing the spontaneous respiratory cycle. This asynchronous pattern is not well tol erated y patients (e.g., the lung inflation can e delivered while the patient is attempting to exhale), and it can produce uneven ventilation. Therefore, the synchronized method of IMV is always favored. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 18)03-May-05 08:26:24

Ovid: ICU Book DISADVANTAGES The principal disadvantages of IMV are an increase in the work of reathing and a tendency for the cardiac output to e reduced. Work of Breathing The spontaneous reathing during IMV takes place through a high-resistance circu it (the tracheal tu e and ventilator tu ing) and requires a valve to e opened. Both of these factors can result in an increased work of reathing during IMV (8 ). In fact, the tendency for less respiratory alkalosis during IMV may e a refl ection of an increased work of reathing (which is accompanied y an increase in CO2 production) and not a decrease in alveolar ventilation (9). The increased w ork of reathing during IMV could lead to respiratory muscle fatigue, which coul d then promote further ventilator dependency. To limit the increased P.438 work of reathing during IMV, pressure-support can e added during the period of spon taneous reathing (see later for a description of pressure-support). Cardiac Output Positive-pressure mechanical ventilation can reduce cardiac output y impeding v entricular filling, and can also increase cardiac output y reducing ventricular afterload (10). In patients with left-ventricular dysfunction, the favora le ef fects of mechanical ventilation may outweigh the adverse effects (as long as the lood volume is maintained) (9). This may explain why the cardiac output can de crease during IMV in patients with left-ventricular dysfunction (11). IMV VERSUS ACV The a ove descriptions of ACV and IMV suggest the following considerations. q In patients who are reathing rapidly during ACV and have evidence of overventil ation (respiratory alkalosis) or hyperinflation (auto-PEEP), a switch to IMV sho uld prove eneficial. q In patients with evidence of respiratory muscle weakness or with a history of le ftventricular dysfunction, ACV should e favored over IMV. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 18)03-May-05 08:26:24

Ovid: ICU Book Otherwise, there is little proven advantage with either mode of ventilation (1,3 ,12,13), and personal preference largely dictates which mode of ventilation is p opular in a particular ICU. PRESSURE-CONTROLLED VENTILATION The risk for ventilator-induced lung injury due to large inflation volumes (see Chapter 26) has led to renewed interest in pressure-cycled ventilation. Because peak airway pressures are lower in pressure-cycled than in volume-cycled machine reaths, there is less risk of arotrauma with pressure-cycled ventilation. Pre ssure-controlled ventilation (PCV) is pressurecycled reathing that is completed controlled y the ventilator, with no participation y the patient (similar to the control mode of volume-cycled ventilation). The pattern of ventilation produ ced y PCV is shown in Figure 27.3. Figure 27.3. Airway pressure patterns in pressure-controlled ventilation (PCV), inverse ratio ventilation (IRV) and pressure-support ventilation (PSV). Spontane ous and machine reaths indicated y dashed and solid lines, respectively. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 18)03-May-05 08:26:24

Ovid: ICU Book BENEFITS AND RISKS The major advantage of PCV relates to the inspiratory flow pattern. In volume-cy cled reathing, the inspiratory flow rate is constant throughout lung inflation, whereas in pressurecycled reathing, the inspiratory flow decreases exponential ly during lung inflation (to keep the airway pressure at the preselected value). The decreasing inspiratory P.439 flow pattern reduces peak airway pressures and can improve gas exchange (13,14). The major disadvantage of PCV is the tendency for inflation volumes to vary with changes in the mechanical properties of the lungs. This is illustrated in Figure 27.4. The inflation volume increases as the P.440 peak inflation pressure increases (dashed lines). However, at a constant peak inflation pressure, the inflation volume decreases as the airway resistance increases or the lung compliance decreases. Therefore, any change in lung mecha nics during PCV can lead to a change in inflation volumes. Because of the influe nce of lung mechanics on inflation volumes during PCV, this method of mechanical ventilation seems est suited for patients with neuromuscular disease (and norm al lung mechanics). Figure 27.4. The determinants of inflation volume during pressure-cycled ventila tion. INVERSE RATIO VENTILATION http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 18)03-May-05 08:26:24

Ovid: ICU Book When PCV is com ined with a prolonged inflation time, the result is inverse rati o ventilation (IRV) (15,16). The ventilatory pattern produced y IRV is shown in the middle panel of Figure 27.3. A decrease in inspiratory flow rate is used to prolong the time for lung inflation, and the usual I:E ratio of 1:2 to 1:4 is r eversed to a ratio of 2:1. The prolonged inflation time can help prevent alveola r collapse. However, prolonged inflation times also increase the tendency for in adequate emptying of the lungs, which can lead to hyperinflation and auto-PEEP. The tendency to produce auto-PEEP can lead to a decrease in cardiac output durin g IRV (17), and this is the major draw ack with IRV. The major indication for IR V is for patients with ARDS who have refractory hypoxemia or hypercapnia during conventional modes of mechanical ventilation (17). PRESSURE-SUPPORT VENTILATION Pressure-augmented reathing that allows the patient to determine the inflation volume and respiratory cycle duration is called pressure-support ventilation (PS V) (18). This method of ventilation is used to augment spontaneous reathing, no t to provide full ventilatory support. VENTILATORY PATTERN The ventilatory pattern produced y PSV is shown in the lower panel of Figure 27 .3. At the onset of each spontaneous reath, the negative pressure generated y the patient opens a valve that delivers the inspired gas at a preselected pressu re (usually 5 to 10 cm H2O). The patient's inspiratory flow rate is adjusted y the ventilator as needed to keep the inflation pressure constant, and when the p atient's inspiratory flow rate falls elow 25% of the peak inspiratory flow, the augmented reath is terminated. By recognizing the patient's inherent inspirato ry flow rate, PSV allows the patient to dictate the duration of lung inflation a nd the inflation volume. This should result in a more physiologically accepta le method of positive-pressure lung inflation. P.441 CLINICAL USES PSV can e used to augment inflation volumes during spontaneous reathing or to overcome the resistance of reathing through ventilator circuits. The latter app lication is the most popular and is used to limit the work of reathing during w eaning from mechanical ventilation. The goal of PSV in this setting is not to au gment the tidal volume, ut merely to provide enough pressure to overcome the re sistance created y the tracheal tu es and ventilator tu ing. Inflation pressure s of 5 to 10 cm H2O are appropriate for this purpose. PSV has also ecome popula r as a noninvasive method of mechanical ventilation (19). In this situation, PSV is http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 18)03-May-05 08:26:24

Ovid: ICU Book delivered through specialized face masks or nasal masks, using inflation pressur es of 20 cm H2O. POSITIVE END-EXPIRATORY PRESSURE Under normal circumstances, the volume of gas inhaled is exhaled completely. As a result, the expiratory airflow ceases y the end of expiration, and the alveol ar pressure at end-expiration is equivalent to atmospheric pressure (zero refere nce point). When the alveolar pressure at end-expiration is a ove atmospheric pr essure, it is referred to as positive end-expiratory pressure (PEEP). There are two ways to create PEEP during mechanical ventilation. One is to add a pressurelimiting device that stops exhalation at a preselected pressure. This is called extrinsic PEEP. The other is to ventilate patients at high inflation volumes and rapid rates and promote hyperinflation (20). This is called intrinsic PEEP or a uto-PEEP. The description of PEEP in this chapter pertains mostly to extrinsic P EEP. Auto-PEEP is descri ed in Chapter 28. EXTRINSIC PEEP Extrinsic PEEP is applied y placing a pressure-limiting valve in the expiratory lim of the ventilator circuit. This valve exerts a ack pressure, and exhalati on proceeds until this ack pressure is reached, whereupon flow ceases. A simila r effect would e created y placing the distal end of the expiratory tu ing und er water. The ack pressure would then e equivalent to the distance that the tu e is su merged. The first time I saw PEEP used, this is how it was done.) The e ffect of extrinsic PEEP on the phasic airway pressures is shown in Figure 27.5. Note that the entire pressure waveform is displaced upward y the applied PEEP. This means that PEEP increases not only the end-expiratory pressure, ut also th e mean intrathoracic pressure. This is important in understanding the effects of PEEP on cardiac performance. The effect of PEEP on the proximal airway pressure s is an important consideration when using the end-inspiratory peak and plateau pressures to evaluate lung mechanics (see Chapter 26). When doing so, the amount of applied PEEP must e su tracted from the P.442 end-inspiratory pressures to determine the actual pressures generated y intrinsic lung mechanics. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 18)03-May-05 08:26:24

Ovid: ICU Book Figure 27.5. Airway pressure patterns in volume-cycled ventilation with positive endexpiratory pressure (PEEP), and in spontaneous reathing with continuous pos itive airway pressure (CPAP). THE PHYSIOLOGY OF PEEP The distal airspaces tend to collapse at the end of expiration, and this tendenc y is exaggerated when the lungs are stiff (e.g., in ARDS). Alveolar collapse imp airs gas exchange and makes the lungs stiffer. PEEP prevents the alveoli from co llapsing at the end of expiration and can open alveoli that have already collaps ed (21). This improves gas exchange (decreases intrapulmonary shunt) and makes t he lungs less stiff (increases lung compliance). The improved gas exchange raise s the arterial PO2, which allows the inspired oxygen (FIO2) to e reduced to les s toxic levels. This latter effect (a decrease to less toxic inhaled O2 concentr ations) is one of the major indications for extrinsic PEEP. Cardiac Performance Because PEEP displaces the entire positive-pressure waveform upward, the effects of positivepressure ventilation on cardiac performance are magnified (see Chapt er 26). Thus, the tendency for positive-pressure ventilation to reduce cardiac f illing and cardiac output is much greater with PEEP, and this tendency is magnif ied even further P.443 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 18)03-May-05 08:26:24

Ovid: ICU Book y hypovolemia and cardiac dysfunction. The tendency for PEEP ventilation to red uce cardiac output is not a function of the PEEP level, ut is a function of the PEEP-induced increase in mean intrathoracic pressure. Thus the often-heard proc lamation it's only 5 of PEEP, so the cardiac output should e OK is meaningless. L ow levels of PEEP can e deleterious for the cardiac output if the mean intratho racic pressure is high. Oxygen Transport The tendency for PEEP to reduce the cardiac output is an important consideration , as demonstrated y the determinants of systemic oxygen delivery (DO2): Thus, the effect of PEEP on cardiac output (Q) will determine whether a PEEP-ind uced increase in arterial oxygenation (SaO2) is associated with a similar increa se in systemic oxygenation. This is demonstrated in Figure 27.6. When PEEP does not change the cardiac output, the increase in arterial O2 saturation is associa ted with an increase in systemic O2 delivery. However, when PEEP reduces the car diac output, the increase in arterial oxygenation is associated with a decrease in systemic oxygenation. Thus, the effects of PEEP on systemic oxygenation are d etermined y the change in cardiac output, not the change in arterial oxygenatio n. The point at which PEEP est improves systemic oxygen transport is sometimes called est PEEP. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 18)03-May-05 08:26:24

Ovid: ICU Book Figure 27.6. The effects of incremental positive end-expiratory pressure (PEEP) on arterial and systemic oxygenation in a patient with severe ARDS. The effects of PEEP shown in Figure 27.6 emphasize the following important point : An increase in arterial oxygenation should never e used as the end-point of P EEP ventilation. Rather, the cardiac output must e monitored when PEEP is appli ed. There is a waiting period of at least 15 minutes after PEEP is applied efor e the effects on cardiac output are fully expressed (22). If the cardiac output is not availa le, the venous oxygen saturation can e useful. A decrease in veno us O2 saturation after PEEP can e used as evidence for a PEEP-induced decrease in cardiac output. CLINICAL USES The following conditions are considered appropriate indications for extrinsic PE EP. Toxic Level of Inhaled O2 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 18)03-May-05 08:26:24

Ovid: ICU Book In patients who require a toxic level of inhaled oxygen (FIO2 greater than 0.60) to maintain adequate oxygenation, the addition of PEEP can increase arterial an d systemic oxygenation, and allow the inhaled oxygen to e reduced to less toxic levels. PEEP is most effective in this regard when the lungs are stiff and the pathologic process is diffuse, P.444 as in ARDS). When lung disease is localized (as in pneumonia), PEEP can overdistend alveoli in normal lung regions and redi rect lood ack to the diseased areas. In this situation, PEEP can worsen arteri al oxygenation (23). Low-Volume Ventilation As mentioned in Chapter 26, PEEP is recommended for volume-cycled ventilation wi th low tidal volumes (5 to 10 mL/kg) (24). The PEEP is meant to prevent repetiti ve opening and closing of small airways, which is elieved to e a source of fur ther lung injury. The level of PEEP should e a ove the inflection point in the pressure-volume curve (the point of airway closure), ut in the a sence of such a curve, a PEEP level of 10 cm H2O may e satisfactory (24). P.445 O structive Lung Disease The small airways tend to collapse at the end of expiration in patients with o s tructive lung disease (i.e., asthma and chronic o structive pulmonary disease). In this situation, extrinsic PEEP can keep the airways open at end-expiration an d reduce the tendency for air trapping (25). The use of extrinsic PEEP in o stru ctive lung disease is discussed in more detail in Chapter 28. CLINICAL MISUSES The following uses of PEEP are either inappropriate or unjustified. Reducing Lung Edema As descri ed in Chapter 23, PEEP does not reduce lung edema in ARDS (see Fig. 23 .6). In fact, positive intrathoracic pressures can promote water accumulation in the lungs (26), possi ly y impeding lymphatic drainage from the lungs. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 18)03-May-05 08:26:24

Ovid: ICU Book Routine PEEP Because glottic closure at the end of expiration creates low levels of PEEP (phy siologic PEEP), the practice of applying PEEP to all intu ated patients has eco me popular. However, there is no evidence that adults produce PEEP from glottic closure (grunting neonates do), and routine PEEP has no documented enefits in i ntu ated patients (26). Mediastinal Bleeding The application of PEEP is a common practice for preventing or controlling media stinal leeding after cardiopulmonary ypass surgery. This practice is ased on a misconception a out the effect of PEEP on transmural pressure. Because PEEP is transmitted across the walls of lood vessels, PEEP may not reduce the transmur al pressure. Thus, there is no mechanism for PEEP to reduce mediastinal leeding from intrathoracic lood vessels. This is confirmed y clinical reports showing that PEEP does not reduce postoperative mediastinal leeding (27). CONTINUOUS POSITIVE AIRWAYS PRESSURE Spontaneous reathing in which a positive pressure is maintained throughout the respiratory cycle is called continuous positive airway pressure (CPAP). The airw ay pressure pattern with CPAP is shown in Figure 27.5. Note that the patient doe s not have to generate a negative airway pressure to receive the inhaled gas. Th is is made possi le y P.446 a specialized inhalation valve that opens at a pres sure a ove atmospheric pressure. This eliminates the extra work involved in gene rating a negative airway pressure to inhale. CPAP should e distinguished from s pontaneous PEEP. In spontaneous PEEP, a negative airway pressure is required for inhalation. Spontaneous PEEP has een replaced y CPAP ecause of the reduced w ork of reathing with CPAP. CLINICAL USES The major uses of CPAP are in nonintu ated patients. CPAP can e delivered throu gh specialized face masks equipped with adjusta le, pressurized valves. CPAP mas ks have een used successfully to postpone intu ation in patients with acute res piratory failure (28). However, these masks must e tight-fitting, and they cann ot e removed for the patient to eat. Therefore, they are used only as a tempora ry measure. Specialized nasal masks may e etter tolerated. CPAP delivered thro ugh nasal masks (nasal CPAP) has ecome popular in patients with o structive sle ep-apnea (29). In this situation, the CPAP is used as a stint to http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 18)03-May-05 08:26:24

Ovid: ICU Book prevent upper airway collapse during negative pressure reathing. Nasal CPAP has also een used successfully in patients with acute exacer ation of chronic o st ructive lung disease (30). Airway Pressure Release A specialized form of CPAP is availa le in which the positive airway pressure is released temporarily during exhalation, to facilitate exhalation. This method, called airway pressure release ventilation, uses higher pressures than conventio nal CPAP, and has een used as an alternative to mechanical ventilation in patie nts with acute respiratory failure (31). REFERENCES REVIEWS 1. Sassoon CSH, Mahutte K, Light RW. Ventilator modes: old and new. Crit Care Cl in 1990;6:605634 (205 references). 2. Rasanen J. Mechanical ventilatory support: time for appraisal. Int Crit Care Digest 1991;10:35 (editorial; 13 references). 3. Slutsky AS. American College of Chest Physicians' Consensus Conference on Mec hanical Ventilation. Chest 1993;104:18331857 (158 references). ASSIST-CONTROL VENTILATION 4. Fernandez R, Blanch L, Antigas A. Respiratory center activity during mechanic al ventilation. J Crit Care 1991;6:102111. 5. Marini JJ. Strategies to minimize reathing effort during mechanical ventilat ion. Crit Care Clin 1990;6:635661. P.447 6. Rogers RL, Schlichtig R, Miro A, Pinsky M. Auto-PEEP during CPR: an occult caus e of electromechanical dissociation. Chest 1991;99:492493. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (15 of 18)03-May-05 08:26:24

Ovid: ICU Book INTERMITTENT MAN ATORY VENTILATION 7. owns JB, Klein EF, esautels , et al. IMV: a new approach to weaning patien ts from mechanical ventilation. Chest 1973;64:331335. 8. Sassoon CSH, el Rosario N, Fei R, et al. Influence of pressure- and flow-tri ggered synchronous intermittent mandatory ventilation on inspiratory muscle work . Crit Care Med 1994;22:19331941. 9. Hudson L , Hurlow RS, Craig KC, Pierson J. oes intermittent mandatory venti lation correct respiratory alkalosis in patients receiving assisted mechanical v entilation? Am Rev Respir is 1985;132:10711074. 10. Pinsky MR. Cardiovascular effects of ventilatory support and withdrawal. Ane sth Analg 1994;79:567576. 11. Mathru M et al. Hemodynamic responses to changes in ventilatory patterns in patients with normal and poor left ventricular reserve. Crit Care Med 1982;10:42 3426. 12. Sternberg R, Sahebjami H. Hemodynamic and oxygen transport characteristics o f common ventilator modes. Chest 1994;105:17981803. PRESSURE-CONTROLLE VENTILATION 13. Shelledy C, Rau JL, Thomas-Goodfellow L. A comparison of the effects of ass istcontrol, SIMV, and SIMV with pressure-support on ventilation, oxygen consumpt ion, and ventilatory equivalent. Heart Lung 1995;24:6775. 14. Rappaport SH, Shipner R, Yoshihara G, et al. Randomized, prospective trial o f pressure-limited versus volume controlled ventilation in severe respiratory fa ilure. Crit Care Med 1994;22:2232. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (16 of 18)03-May-05 08:26:24

Ovid: ICU Book 15. Marcy TW, Marini JJ. Inverse ratio ventilation in AR S. Rationale and implem entation. Chest 1991;100:494504. 16. Papadakos PJ, Halloran W, Hessney JI, et al. The use of pressure-controlled inverse ratio ventilation in the surgical intensive care unit. J Trauma 1991;31: 12111215. 17. Chan K, Abraham E. Effects of inverse ratio ventilation on cardiorespiratory parameters in severe respiratory failure. Chest 1992;102:15561561. PRESSURE-SUPPORT VENTILATION 18. MacIntyre NR. Pressure-support ventilation. In: Grenvik A, owns J, Rasanen J, Smith R, eds. Mechanical ventilation and assisted respiration. Contemporary m anagement in critical care. New York: Churchill Livingstone, 1991;1:5162. 19. Brochard L, Mancebo J, Wysocki M, et al. Noninvasive ventilation for acute e xacerbations of chronic obstructive pulmonary disease. N Engl J Med 1995;338:817 822. POSITIVE EN -EXPIRATORY PRESSURE 20. Ligas JR, Mosiehi JR, Epstein MAF. Occult positive end-expiratory pressure w ith different types of mechanical ventilators. J Crit Care 1990;5:95100. P.448 21. Ranieri VM, Eissa NT, Corbeil C, et al. Effects of positive end-expiratory p ressure on alveolar recruitment and gas exchange in patients with the adult resp iratory distress syndrome. Am Rev Respir is 1991;144:544551. 22. Patel M, Singer M. The optimal time for measuring the cardiorespiratory effe cts of positive end-expiratory pressure. Chest 1993;104:139142. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (17 of 18)03-May-05 08:26:24

Ovid: ICU Book 23. Hawker FH, Torzillo PJ, Southee AE. PEEP and reverse mismatch. A case where le ss PEEP is best. Chest 1991;99:10341036. 24. Marini JJ. Pressure-targeted, lung-protective ventilatory support in acute l ung injury. Chest 1994;105(Suppl):109S115S. 25. Pinsky MR. Through the past darkly: ventilatory management of patients with chronic obstructive pulmonary disease. Crit Care Med 1994;22:17141717. 26. Petty TL. The use, abuse, and mystique of positive end-expiratory pressure. Am Rev Respir is 1988;138:475478. 27. Zurick AM, Urzua J, Ghattas M, et al. Failure of positive end-expiratory pre ssure to decrease postoperative bleeding after cardiac surgery. Ann Thorac Surg 1982;34:608 611.

29. Takasaki Y, Orr , Popkin J, et al. Effect of nasal continuous positive airw ay pressure in sleep apnea in congestive heart failure. Am Rev Respir is 1989;1 40:15781584. 30. de Lucas P, Tarancon C, Puente L, et al. Nasal continuous positive airway pr essure in patients with COP in acute respiratory failure. Chest 1993;104:1694169 7. 31. Cane R, Peruzzi WT, Shapiro BA. Airway pressure release ventilation in acute respiratory failure. Chest 1991;100:460463. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (18 of 18)03-May-05 08:26:24

CONTINUOUS POSITIVE AIRWAY PRESSURE 28. Miro AM, Shivaram U, Hertig I. Continuous positive airway pressure in COP atients in acute hypercapnic respiratory failure. Chest 1993;103:266268.

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 28 THE VENTILATOR- EPEN ENT PATIENT This chapter describes some of the practices and daily concerns in patients rece iving mechanical ventilation. The focus is on practices and complications that a re directly related to artificial airways (endotracheal and tracheostomy tubes) and positive-pressure ventilation. The nosocomial illnesses that can appear in v entilator-dependent patients, such as pneumonia and progressive systemic inflamm ation, are described elsewhere. ARTIFICIAL AIRWAYS Positive-pressure mechanical ventilation usually (but not always) requires trach eal intubation, which can be accomplished by translaryngeal (endotracheal) intub ation, percutaneous cricothyroidotomy (coniotomy), or tracheotomy (1,2 and 3). T he tubes used for tracheal intubation in adults are equipped with inflatable bal loons that are used to seal the trachea and isolate the airways from the larynx and oral cavity. (Because of the narrow diameter of the trachea in young childre n, pediatric tracheal tubes do not always have inflatable balloons to create a t racheal seal.) The complications of tracheal intubation are related to the route of intubation and to the airway damage created by the pressurized seal of the t rachea. The principal complications of tracheal intubation are shown in Figure 2 8.1 (2,4,5). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (1 of 16)03-May-05 08:29:14

Ovid: ICU Book Figure 28.1. Complications of endotracheal intubation (nasal and oral routes) an d tracheostomy. EN OTRACHEAL INTUBATION Endotracheal tubes can be inserted through the nose or mouth. The principal feat ures of each insertion route are listed in Table 28.1. The nasal route is often preferred for intubating patients who are awake and cooperative, whereas the ora l route is preferred in comatose or P.450 P.451 uncooperative patients, or when immediate intubation is necessary (e.g., cardiac arrest). The complications pecu liar to nasotracheal intubation include epistaxis and paranasal sinusitis, and n ecrosis of the nasal mucosa. The complications peculiar to orotracheal intubatio n include dental trauma and tube occlusion in awake patients (who tend to bite d own on the oral tubes). In addition, orotracheal tubes follow an acute angle as they pass from the oropharynx toward the larynx; this angulation creates a torqu e that can damage the posterior larynx. TABLE 28.1. NASAL AN ORAL ROUTES OF EN OTRACHEAL INTUBATION Physiologic Implications The volume of the anatomic dead space in adults is approximately 1 mL/kg, wherea s the volume of most endotracheal tubes is 35 to 45 mL (6). Therefore, in the av erage-sized adult, endotracheal intubation reduces the anatomic dead space by ap proximately 50%. This, however, has minor clinical significance. The most signif icant physiologic consequence of tracheal intubation is an increase in the resis tance to airflow. As predicted by the HagenPoiseuille equation (see Chapter 1), the radius of tracheal tubes has a much greater influence on resistance to airfl ow than the length of the tubes. Therefore, small-diameter tracheal tubes (both endotracheal tubes and tracheostomy tubes) can cause a significant increase in t he work of breathing. This is particularly marked at high flow rates (such as th ose in patients with respiratory distress) (6). The diameter of adult tracheal t ubes should be no less than 7 mm, and preferably 8 mm, to minimize the extra wor k of breathing through the tubes when weaning from mechanical ventilation (6). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (2 of 16)03-May-05 08:29:14

Ovid: ICU Book Intracranial Pressure Endotracheal intubation can be accompanied by an increase in intracranial pressu re, regardless of the intubation route. The mechanism for this effect is unclear , but it can be blocked by lidocaine. Intravenous lidocaine (1.5 mg/kg) produces a more effective blockade than endotracheal lidocaine (7). If intracranial hype rtension is a concern, lidocaine is recommended just before endotracheal intubat ion. Rapid sequence intubation with fentanyl (3 to 5 g/ kg) and succinylcholine ( 1.5 mg/kg IV) can also reduce the intracranial pressure response to endotracheal intubation (4,8). Succinylcholine can increase intracranial pressure, but this response can be blocked by prior administration of vecuronium (0.01 mg/kg) or pa ncuronium (0.01 mg/kg) (8). Tracheal Tube Position In 5 to 10% of endotracheal intubations, the tip of the endotracheal tube is inc orrectly positioned in one of the mainstem bronchi (9). Because the right mainst em bronchus runs a straighter course from the trachea, selective intubation of t he right lung is most common. The P.452 consequences of right lung intubation ar e illustrated in Figure 28.2. Note the overinflation of the right lung, with her niation into the left hemithorax (arrowheads). The risk for selective intubation of one lung can be minimized by advancing endotracheal tubes no further than 21 cm from the teeth in women and 23 cm in men (10). Although an established pract ice, listening for bilateral breath sounds is not a reliable method for determin ing endotracheal tube position. Specifically, the presence of bilateral breath s ounds does not rule out either selective lung intubation or esophageal intubatio n (9,11). Therefore, routine chest films are required to determine tube position . Figure 28.2. Portable chest x-ray film showing the tip of an endotracheal tube i n the right mainstem bronchus, with herniation of the right lung into the left h emithorax (arrowheads). The thin arrows point to a collection of pleural air at the left lung base. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (3 of 16)03-May-05 08:29:14

Ovid: ICU Book Radiographic Landmarks Some useful radiographic landmarks for evaluating endotracheal tube position are shown in Table 28.2 (12). The vocal cords are usually P.453 situated at the int erspace between the fifth and sixth cervical vertebrae (C5-C6). If not visible, the main carina is usually over the interspace between the fourth and fifth thor acic vertebrae (T4-T5) on a portable chest x-ray film. When the head is in a neu tral position, the tip of the tracheal tube should be 3 to 5 cm above the carina , or midway between the carina and vocal cords. The head is in the neutral posit ion when the inferior border of the mandible projects over the lower cervical sp ine (C5-C6). Note that flexion or extension of the head and neck causes a 2 cm d isplacement of the tip of the endotracheal tube (12). The total displacement of 4 cm with changes in head position represents about one-third of the length of t he trachea. TABLE 28.2. RA IOGRAPHIC EVALUATION OF TRACHEAL INTUBATION Sinusitis Nasotracheal (and nasoenteral) tubes can obstruct the ostia that drain the paran asal sinuses, which can lead to a purulent sinusitis (13,14). The maxillary sinu s is almost always involved (see Fig. 30.3). The presence of an air-fluid level by sinus radiography suggests the diagnosis, but confirmation requires aspiratio n of infected material from the involved sinus. Paranasal sinusitis is an import ant consideration in the evaluation of fever in intubated patients, and it may b e just as common in patients with orotracheal tubes (15). This complication is d escribed in more detail in Chapter 30. Laryngeal amage The risk for laryngeal injury from translaryngeal intubation is the major reason for performing tracheostomies when prolonged intubation P.454 is anticipated (1 ,2). The spectrum of laryngeal damage includes ulceration, granulomas, vocal cor d paresis, and http://gateway.ut.ovid.com/gw1/ovidweb.cgi (4 of 16)03-May-05 08:29:14

Ovid: ICU Book laryngeal edema. Some type of laryngeal damage is usually evident after 72 hours of translaryngeal intubation (2), and laryngeal edema is reported in 5% of case s. Fortunately, most cases of laryngeal injury do not result in significant airw ays obstruction, and the injury resolves within weeks after extubation (16). The problem of laryngeal edema after tracheal decannulation is described in Chapter 29. TRACHEOSTOMY Tracheostomy is preferred in patients who require prolonged mechanical ventilati on. The advantages of tracheostomy tubes over endotracheal tubes include greater patient comfort and more effective clearing of secretions (17). Tracheostomy pa tients can also ingest food orally and converse using special tracheal tubes. Ho wever, the complications of tracheostomy can overshadow those associated with en dotracheal intubation. Complications Surgical tracheotomy is accompanied by serious complications in 5% of cases, and the procedure has a reported mortality as high as 2% (compared with a reported mortality of 1:5000 endotracheal intubations) (18). Immediate postoperative comp lications include pneumothorax (5%), stomal hemorrhage (5%), and accidental deca nnulation (17,18). The latter complication can be troublesome because the trache ostomy tract closes quickly, and attempts to reinsert the tubes can create false tracts. If a tracheostomy tube has to be reinserted in the first week after sur gery, a 12-French suction catheter should be used as a guidewire to reduce the r isk for trauma and unsuccessful reintubation. Tracheal stenosis is the most fear ed complication of tracheostomy. The stenosis usually occurs at the tracheotomy site and not at the site where the inflated balloon creates the tracheal seal. S tenosis is a late complication that usually appears days to weeks after tracheal decannulation. When to Perform a Tracheostomy The optimal time for performing a tracheostomy in patients with prolonged endotr acheal intubation is a time-honored controversy. However, the consensus at prese nt seems to favor the following approach. After 1 week of endotracheal intubatio n, if little chance for extubation in the ensuing week exists, proceed to trache ostomy (2,17,18). Because of the increased risk for tracheal stenosis associated with repeated tracheotomy incisions, tracheostomy should not be performed in pa tients who have had prior tracheostomy until a waiting period of more than 1 wee k has passed, if possible. P.455 CUFF-RELATE COMPLICATIONS As mentioned earlier, positive-pressure mechanical ventilation in adults require s tracheal tubes that are equipped with an inflatable balloon (called a cuff) th at seals the trachea and prevents gas from escaping back through the larynx. Fig ure 28.3 shows two tracheostomy tubes with inflated cuffs at their distal ends. The tube on the left has a balloon that is normally deflated and must be inflate d by injecting a volume of air into the balloon with an air-filled syringe. The tube on the right has a foam rubber cuff that is normally inflated at atmospheri c pressure. This cuff is manually deflated by applying suction from a syringe du ring insertion of the tube; once the tube is in place, the cuff is allowed to in flate and create the tracheal seal. This cuff design, therefore, allows the trac hea to be sealed with minimum risk of pressure-induced injury to the tracheal mu cosa. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (5 of 16)03-May-05 08:29:14

Ovid: ICU Book Figure 28.3. Tracheostomy tubes with different types of inflatable balloons (cuf fs). The tube on the left has a cuff that requires positive-pressure inflation, and the (Fome-Cuff) tube on the right has a foam rubber cuff that is normally in flated at atmospheric (zero) pressure. The foam rubber cuff can create a trachea l seal without the risk of pressure necrosis of the tracheal mucosa. Aspiration Contrary to popular belief, cuff inflation to create a tracheal seal does not pr event aspiration of mouth secretions and tube feedings into the lower airways. A spiration of saliva and liquid tube feedings has been documented in more than 50 % of ventilator-dependent patients with tracheostomies (19). Furthermore, in ove r threefourths of the cases, the aspiration is clinically silent (19). Consideri ng that approximately 1 L of saliva is normally produced each day (the normal ra te of saliva production is 0.6 mL/min) and that each microliter of saliva contai ns approximately 1 billion microorganisms, the danger associated with aspiration of mouth secretions seems considerable (20). This emphasizes the value of routi ne tracheal suctioning to clear secretions from the airways. Cuff Leaks Just as liquid can pass around inflated cuffs and move into the lower airways, g as in the airways can pass around inflated cuffs and move out of the lungs. Cuff leaks are usually detected by sounds generated during lung inflation (created b y gas flowing through the vocal cords). When a cuff leak becomes audible, the vo lume of the leak can be determined by noting the difference between the desired inflation volume and the exhaled volume recorded by the ventilator. Cuff leaks a re rarely due to disruption of the cuff itself (21). The most common cause of cu ff leaks is nonuniform contact between the cuff and the tracheal wall. Another s ource of cuff leaks is faulty function of a one-way valve at the air injection i nlet that normally keeps the cuff inflated. These valves can become leaky and al low air to escape from the cuff. If a cuff leak is suspected, the patient should be separated from the ventilator and the lungs should be manually inflated with an anesthesia bag. If a leak is audible, the cuff should be inflated until the sounds http://gateway.ut.ovid.com/gw1/ovidweb.cgi (6 of 16)03-May-05 08:29:14

Ovid: ICU Book P.456 P.457 disappear, and the resultant cuff pressure should be checked. If the cuff pressure is above 25 cm H2O, the tracheal tube should be replaced. If the cuff pressure gradually diminishes over time, the one-way valve mentioned earlie r may be defective. In this situation, recurrent cuff leaks can be prevented by clamping the cuff inflation tubing to isolate the cuff from the faulty valve. Tracheal Necrosis The original tracheal tubes had low-compliance cuffs that promoted pressure-indu ced tracheal necrosis when inflated. This complication was reduced significantly when larger, more compliant cuffs were introduced in the mid-1970s. The newer c uffs generate lower pressures when inflated, and the larger size of the cuff dis perses the cuff pressure over a wider area of the tracheal mucosa. The systolic pressure in the mucosal vessels of the trachea is normally 20 to 25 mm Hg (17), so the goal of cuff inflation is to create a seal at an inflation pressure below 20 cm H2O. In patients with clinical shock and hypotension, pressure necrosis o f the trachea is possible at cuff pressures far below 20 cm H2O. In this situati on, the tracheal tube on the right in Figure 28.3, which creates a tracheal seal at atmospheric (zero) pressure, is my preference. ALVEOLAR RUPTURE One of the constant concerns in the ventilator-dependent patient is the risk of alveolar rupture, which occurs in up to one-fourth of patients receiving positiv e-pressure mechanical ventilation (22). As mentioned in Chapter 26, alveolar rup ture can be the result of excessive pressure (barotrauma) or overdistension (vol utrauma). CLINICAL PRESENTATION Escape of gas from the alveoli can produce a variety of clinical manifestations (23). The alveolar gas can dissect along tissue planes and produce pulmonary int erstitial emphysema, and can move into the mediastinum and produce pneumomediast inum. Mediastinal gas can move into the neck to produce subcutaneous emphysema, or can pass below the diaphragm to produce pneumoperitoneum. Finally, if the rup ture involves the visceral pleura, gas will collect in the pleural space and pro duce a pneumothorax. Each of these entities can occur alone or in combination wi th the others (22,23). PNEUMOTHORAX Radiographic evidence of pneumothorax occurs in 5 to 15% of ventilator-dependent patients (22,23). Risk factors include high inflation P.458 pressures and infla tion volumes, positive end-expiratory pressure (PEEP), and diffuse lung injury. The highest risk of pneumothorax is in patients with acute respiratory distress syndrome (AR S), where the incidence of pneumothorax is as high as 60% (23). Clinical Presentation Clinical manifestations are either absent, minimal, or nonspecific. The most val uable clinical sign is subcutaneous http://gateway.ut.ovid.com/gw1/ovidweb.cgi (7 of 16)03-May-05 08:29:14

Ovid: ICU Book emphysema in the neck and upper thorax, which is pathognomonic of alveolar ruptu re. Breath sounds are unreliable in ventilator-dependent patients because sounds transmitted from the ventilator tubing can be mistaken for airway sounds. Radiographic etection The radiographic detection of pleural air can be difficult in portable chest fil ms, particularly when the patient is in the supine position, when pleural air do es not collect at the lung apex (23,24). Figure 28.4 illustrates some of the aty pical features of pneumothorax in the supine position. The hyperlucent area at t he right lung base represents a collection of pleural air that is anterior to th e right lung (which is the P.459 most superior region of the pleural space in th e supine position). A similar collection of pleural air is shown at the left lun g base in Figure 28.2. Basilar and subpulmonic collections of air are characteri stic of pneumothorax in the supine position (24). Figure 28.4. Atypical features of pneumothorax in the supine position. The hyper lucent area at the base of the right lung is a collection of pleural air that is anterior to the right lung. In the supine position, this is the highest region of the pleural space. The sharp line demarcating the descending aorta (thin arro ws) represents air behind the inferior pulmonary ligament. Redundant Skin Folds When the film cartridge used for portable chest x-ray examinations is placed und er the patient, the skin on the back can fold over on itself, and the edge of th is redundant skin fold creates a radiographic shadow that can be mistaken for a pneumothorax. The radiographic appearance of a redundant skin fold is shown in F igure 28.5. Note that there is a gradual increase in radiodensity that suddenly ends as a wavy line. The increase in density is produced by the skin that is fol ded back on itself. A pneumothorax would appear as a sharp white line with dark shadows (air) on either side. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (8 of 16)03-May-05 08:29:14

Ovid: ICU Book Figure 28.5. Portable chest x-ray film showing a wavy line in the left hemithora x. This line is the edge of a redundant skin fold, not a pneumothorax. When a redundant skin fold is suspected, a repeat chest film should be obtained (it may be wise to alert the x-ray technician to make sure the film cartridge is flush on the patient s back). The shadow should disappear if it is due to a red undant skin fold. PLEURAL EVACUATION Evacuation of pleural air is accomplished by inserting a chest tube through the fourth or fifth intercostal space along the mid-axillary line. The tube should b e advanced in an anterior and superior direction (because this is where pleural air collects in the supine position). The pleural space is drained of fluid and air using a three-chamber system like the one shown in Figure 28.6 (25). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (9 of 16)03-May-05 08:29:14

Ovid: ICU Book Figure 28.6. A standard pleural drainage system for evacuating air and fluid fro m the pleural space. Collection Chamber The first bottle in the system collects fluid from the pleural space and allows air to pass through to the next bottle in the series. Because the inlet of this chamber is not in direct contact with the fluid, the pleural fluid that is colle cted does not impose a back pressure on the pleural space. Water-Seal Chamber The second bottle acts as a one-way valve that allows air to escape from the ple ural space but prevents air from entering the pleural space. This one-way valve is created by submerging the inlet tube under water. This imposes a back-pressur e on the pleural space that is equal to the depth that the tube is submerged. Th e positive pressure in the pleural space then prevents atmospheric air (at zero pressure) from entering the pleural space. The water thus seals the pleural space from the surrounding atmosphere. This water-seal pressure is usually 2 cm H2O. P .460 P.461 Air that is evacuated from the pleural space passes through the water in the second bottle and creates bubbles. Thus, the presence of bubbles in the water-seal chamber (called bubbling) is used as evidence for a continuing bronch opleural air leak. Suction-Control Chamber The third bottle in the system is used to set a maximum limit on the negative su ction pressure that is imposed on the pleural space. This maximum pressure is de termined by the height of the water column in the air inlet tube. Negative press ure (from wall suction) draws the water down the air inlet tube, and when the ne gative pressure exceeds the height of the water column, air is entrained from th e atmosphere. Therefore, the pressure in the bottle can never become more negati ve than the height of the water column in the air inlet tube. Water is added to the suction-control chamber to achieve a water level of 20 cm. The wall suction is then activated and slowly increased until bubbles appear in the water. This b ubbling indicates that atmospheric air is being http://gateway.ut.ovid.com/gw1/ovidweb.cgi (10 of 16)03-May-05 08:29:14

Ovid: ICU Book entrained, and thus the maximum negative pressure has been achieved. The continu ous bubbling causes water evaporation, so it is imperative that the height of th e water in this chamber is checked periodically, and more water is added when ne cessary. Why Use Suction? The practice of using suction to evacuate pleural air is both unnecessary and po tentially harmful. Although there is a perception that suction will help the lun gs reinflate, the lungs will reinflate without the P.462 use of suction. Further more, creating a negative pressure in the pleural space also creates a higher tr anspulmonary pressure (the pressure difference between alveoli and the pleural s pace), and this increases the rate of the air flowing through the bronchopleural fistula. Thus applying suction to the pleural space increases bronchopleural ai r leaks, and this can keep bronchopleural fistulas patent. If a persistent air l eak is present when suction is applied to the pleural space, the suction should be discontinued. Any air that collects in the pleural space will continue to be evacuated when the pleural pressure becomes more positive than the water-seal pr essure. OCCULT PEEP As mentioned in Chapter 27, mechanical ventilation with high inflation volumes a nd rapid rates can (and probably often does) produce PEEP as a result of incompl ete alveolar emptying during expiration (26,27 and 28). The illustration in Figu re 28.7 helps explain this phenomenon. Figure 28.7. The features of intrinsic PEEP produced by inadequate alveolar empt ying. The presence of airflow ([V with dot above]) at end-expiration creates a p ressure drop from the alveolus (PALV) to the proximal airways (Pprox). As shown at the top of the figure, the proximal airway pressure returns to zero at end-ex piration while the alveolar pressure remains positive; hence the term occult PEE P. The upper left panel illustrates the end-expiratory occlusion method for dete cting occult PEEP. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (11 of 16)03-May-05 08:29:14

Ovid: ICU Book PATHOGENESIS Under normal circumstances, there is no airflow at the end of expiration, and th us end-expiratory pressure is the same in the alveoli and P.463 proximal airways . On the other hand, when alveolar emptying is incomplete during expiration, air flow is present at the end of expiration, and this creates a pressure drop from the alveoli to the proximal airways at end-expiration. Thus in this situation, a lveolar pressure will be positive relative to atmospheric (zero) pressure at end -expiration (i.e., PEEP). As shown in the upper portion of Figure 28.7, the posi tive alveolar pressure is not evident on the proximal airways pressure tracing, and hence this is an occult PEEP (26). Other terms for this pressure are intrins ic PEEP, auto-PEEP, and dynamic hyperinflation (27). The latter term is usually reserved for PEEP produced by obstructive airways disease (see below). Predisposing Factors The factors that predispose to hyperinflation and occult PEEP can be separated i nto ventilation-associated and disease-associated factors. The ventilatory facto rs that promote hyperinflation include high inflation volumes, rapid breathing, and a relative decrease in exhalation time relative to inhalation time. The dise ase-related factor that promotes hyperinflation is airways obstruction (as occur s in asthma and chronic obstructive pulmonary disease [COP ]). All three factors are operative in ventilator-dependent patients with obstructive airways disease . In patients with asthma and COP , occult PEEP is probably universal during con ventional volume-cycled mechanical ventilation (27,28 and 29). In one study of v entilator-dependent patients with COP , the level of occult PEEP varied from 2.5 to 15 cm H2O (27). Occult PEEP may also be common in ventilator-dependent patie nts with AR S (30). CONSEQUENCES Cardiac Performance Occult PEEP can have the same effects on cardiac performance as extrinsic PEEP, which are described in Chapter 27. The detrimental effects of occult PEEP on car diac performance have been implicated as a cause of electromechanical dissociati on during cardiopulmonary resuscitation (31). Alveolar Rupture Occult PEEP is a manifestation of hyperinflation, and thus there is an increased risk of alveolar rupture (volutrauma) with occult PEEP. Hyperinflation-induced alveolar rupture and pneumothorax are particular concerns in ventilatordependent asthmatic patients (29). Thoracic Compliance Occult PEEP increases the end-inspiratory pressures in the proximal airways (bot h peak and plateau pressures). When occult PEEP goes P.464 undetected, the incre ase in plateau pressure is misinterpreted as a decrease in the compliance of the lungs and chest wall. Therefore, failure to consider occult PEEP can result in an underestimation of thoracic compliance (32). When respiratory compliance is b eing monitored (see Chapter 27), the level of occult PEEP should be subtracted http://gateway.ut.ovid.com/gw1/ovidweb.cgi (12 of 16)03-May-05 08:29:14

Ovid: ICU Book from the measured plateau pressure and the adjusted plateau pressure should be u sed for the compliance calculation. Work of Breathing Occult PEEP can be accompanied by an increased work of breathing. Two mechanisms are involved. First, the hyperinflation places the lungs on a flatter portion o f their pressure-volume curve, so higher pressures are needed to inhale the tida l volume. (To appreciate this effect, take a deep breath, and then try to breath e in further.) The hyperinflation also flattens the diaphragm, and when this occ urs, the diaphragm muscle must generate a greater tension to achieve a given cha nge in thoracic pressure. (This effect is predicted by the Law of Laplace: T = P r.) The increased work of breathing with occult PEEP can impede efforts to wean from mechanical ventilation (see Chapter 29). MONITORING OCCULT PEEP Unfortunately, occult PEEP is difficult to monitor, at least in a quantitative m anner. The following methods have been used to monitor occult PEEP. End-Expiratory Occlusion uring controlled ventilation (i.e., when the patient is completely passive), oc cult PEEP can be uncovered by occluding the expiratory tubing at the end of expi ration (32). This maneuver blocks airflow and allows the pressure in the proxima l airways to equilibrate with alveolar pressure. Thus, a sudden rise in proximal airways pressure with end-expiratory occlusion is evidence of occult PEEP (see the upper left panel in Fig. 28.7). To quantitate the level of occult PEEP, the occlusion must be timed to the very end of expiration. Expiratory flow can persi st until the end of expiration, so occlusions performed before end-expiration wi ll overestimate the prevalence and magnitude of occult PEEP. Unfortunately, timi ng the occlusion to the end of expiration is a difficult task. End-expiratory oc clusion becomes more reliable when it is performed electronically, which is anti cipated in future models of mechanical ventilators. Extrinsic PEEP The application of extrinsic PEEP normally increases the peak inspiratory airway pressure by an equivalent amount. However, in the presence P.465 of occult PEEP , the peak inspiratory pressure may not change when extrinsic PEEP is applied (t his is explained below). Therefore, failure of extrinsic PEEP to produce an incr ease in peak inspiratory airway pressure is evidence of occult PEEP (33,34 and 3 5). Furthermore, the level of extrinsic PEEP that first produces a rise in peak inspiratory pressures can be taken as the quantitative level of occult PEEP (33) . Management The methods used to prevent or reduce hyperinflation and occult PEEP are all dir ected at promoting alveolar emptying during expiration. Ventilator-induced hyper inflation can be minimized by avoiding excessive inflation volumes, and by optim izing the time allowed for exhalation. These measures are described in Chapter 2 6. Extrinsic PEEP When hyperinflation is due to small airway collapse at end-expiration, as occurs in patients with asthma and COP , http://gateway.ut.ovid.com/gw1/ovidweb.cgi (13 of 16)03-May-05 08:29:14

Ovid: ICU Book the application of extrinsic PEEP can help keep the small airways open at end-ex piration. The level of extrinsic PEEP must be enough to counterbalance the press ure causing small airways collapse (the critical closing pressure) but should no t exceed the level of intrinsic PEEP (so that it does not impair expiratory flow ) (34). To accomplish this, the level of extrinsic PEEP should match the level o f occult PEEP. The problem then is to obtain an accurate measure of occult PEEP. When this is impossible, an alternative method is to observe the changes in pea k inspiratory pressure when extrinsic PEEP is applied. As mentioned earlier, the level of extrinsic PEEP that first causes an increase in peak inspiratory airwa y pressures can be taken as the level of occult PEEP. Therefore, the extrinsic P EEP can be increased gradually as long as it does not cause an increase in peak inspiratory pressure. When used properly, extrinsic PEEP should reduce the work of breathing created by intrinsic or occult PEEP (35). The clinical significance of this, however, is unclear. REFERENCES 1. National Association of Medical irectors of Respiratory Care (NAM RC) Consen sus Conference on Artificial Airways in Patients Receiving Mechanical Ventilatio n. Chest 1989;96:178180. 2. Gallagher TJ. Endotracheal intubation. Crit Care Clin 1992;8:665676. 3. Heffner JE. The technique of tracheotomy and cricothyroidotomy. J Crit Illnes s 1995;10:561568. 4. Kharasch M, Graff J. Emergency management of the airway. Crit Care Clin 1995; 11:5366. 5. Stauffer JL, Olson E, Petty TL. Complications and consequences of endotrache al intubation and tracheostomy. 1981;70:6576. P.466 6. Habib MP. Physiologic implications of artificial airways. Chest 1989;96:180184 . 7. Hamill JF, Bedford RF, Weaver C, et al. Lidocaine before endotracheal intuba tion: intravenous or laryngotracheal? Anesthesiology 1981;55:578581. 8. Walls RM. Rapid-sequence intubation in head trauma. Ann Emerg Med 1993;22:100 81013. 9. Brunel W, Coleman L, Schwartz E, et al. Assessment of routine chest roentge nograms and the physical examination to confirm endotracheal tube placement. Che st 1989;96:10431045. 10. Owen RL, Cheney FW. Endotracheal intubation: a preventable complication. Ane sthesiology 1987;67:255 257. 11. Mizutani AR, Ozaki G, Benumof JL, Scheller ML. Auscultation cannot distingui sh esophageal from tracheal http://gateway.ut.ovid.com/gw1/ovidweb.cgi (14 of 16)03-May-05 08:29:14

Ovid: ICU Book passage of tube. J Clin Monit 1991;7:232236. 12. Goodman LR. Pulmonary support and monitoring apparatus. In: Goodman LR, Putm an CE, eds. Critical care imaging. 3rd ed. Philadelphia: WB Saunders, 1992;3559. 13. Pedersen J, Schurizek BA, Melsen NC, Juhl B. The effect of nasotracheal intu bation on the paranasal sinuses. Acta Anesthesiol Scand 1991;35:1113. 14. Rouby J-J, Laurent P, Gosnach M, et al. Risk factors and clinical relevance of nosocomial maxillary sinusitis in the critically ill. Am J Respir Crit Care M ed 1994;150:776783. 15. Holzapfel L, Chevret S, Madinier G, et al. Influence of long-term oro- or na sotracheal intubation on nosocomial maxillary sinusitis and pneumonia: results o f a prospective, randomized clinical trial. Crit Care Med 1993;21:11321138. 16. Colice GL. Resolution of laryngeal injury following translaryngeal intubatio n. Am Rev Respir is 1992;145:361364. 17. Heffner JE, Miller S, Sahn SA. Tracheostomy in the intensive care unit. Part s 1 and 2. Chest 1986;90:269 274, 430436. 18. Marsh HM, Gillespie J, Baumgartner AE. Timing of tracheostomy in the critic ally ill patient. Chest 1989;96:190192. 19. Elpern EH, Scott MG, Petro L, Ries MH. Pulmonary aspiration in mechanically ventilated patients with tracheostomies. Chest 1994;105:563566. 20. Estes RJ, Meduri GU. The pathogenesis of ventilator-associated pneumonia: me chanisms of bacterial transcolonization and airway inoculation. Int Care Med 199 5;21:365383. 21. Kearl RA, Hooper RG. Massive airway leaks: an analysis of the role of endotr acheal tubes. Crit Care Med 1993;21:518521. ALVEOLAR RUPTURE 22. Gammon RB, Shin MS, Buchalter SE. Pulmonary barotrauma in mechanical ventila tion. Chest 1992;102:568572. 23. Marcy TW. Barotrauma: detection, recognition, and management. Chest 1993;104 :578584. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (15 of 16)03-May-05 08:29:14

Ovid: ICU Book 24. Tocino IM, Miller MH, Fairfax WR. istribution of pneumothorax in the supine and semirecumbent critically ill adult. Am J Radiol 1985;144:901905. 25. Kam AC, O Brien M, Kam PCA. Pleural drainage systems. Anesthesia 1993;48:1541 61. OCCULT PEEP 26. Pepe P, Marini JJ. Occult positive end-expiratory pressure in mechanically v entilated patients with airflow obstruction. Am Rev Respir is 1982;126:166170. P .467 27. Gottfried SB, Rossi A, Milic-Emili J. ynamic hyperinflation, intrinsic PEEP , and the mechanically ventilated patient. Crit Care igest 1986;5:3033. 28. Ligas JR, Mosiehi F, Epstein MAF. Occult positive end-expiratory pressure wi th different types of mechanical ventilators. J Crit Care 1990;5:95100. 29. Weiner C. Ventilatory management of respiratory failure in asthma. JAMA 1993 ;269:21282131. 30. Tantucci C, Corbeil C, Chasse M, et al. Flow and volume dependence of respir atory system flow resistance in patients with adult respiratory distress syndrom e. Am Rev Respir is 1992;145:355360. 31. Rogers PL, Schlichtig R, Miro A, Pinsky M. Auto-PEEP during CPR. An occult cau se of electromechanical dissociation. Chest 1991;99:492493. 32. Tobin MJ. Respiratory monitoring. JAMA 1990;264:244251. 33. Slutsky AS. Mechanical ventilation. Chest 1993;104:18331859. 34. Tobin MJ, Lodato RF. PEEP, auto-PEEP, and waterfalls. Chest 1989;96:449451. 35. Pinsky MR. Through the past darkly: ventilatory management of patients with chronic obstructive pulmonary disease. Crit Care Med 1994;22:17141717. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (16 of 16)03-May-05 08:29:14

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 29 ISCONTINUING MECHANICAL VENTILATION Attempts to discontinue mechanical ventilation account for approximately 40% of the time that a patient spends on a ventilator (1). When a ventilator is require d because of brainstem respiratory depression (e.g., general anesthesia or a dru g overdose), removal of ventilatory assistance is easily accomplished when the p atient awakens. However, when a ventilator is required because of cardiopulmonar y insufficiency, withdrawing ventilatory support is a more involved and more gra dual process. This more gradual withdrawal of mechanical ventilation, commonly k nown as weaning, is the subject of this chapter (2,3 and 4). MISCONCEPTIONS The practice of weaning from mechanical ventilation has been dominated by a numb er of misconceptions, which are listed in Table 29.1. Many of these misconceptio ns are related to the perception that mechanical ventilation weakens the respira tory muscles. Although this may be true for the accessory muscles of respiration , it does not necessarily apply to the diaphragm. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (1 of 17)03-May-05 08:29:43

Ovid: ICU Book TABLE 29.1. COMMON MISCONCEPTIONS ABOUT WEANING THE IAPHRAGM The perceived need for weaning (i.e., gradual withdrawal) is based on the notion that the respiratory muscles become weak during mechanical ventilation, and thu s must be strengthened gradually to allow a return to spontaneous breathing. Thi s is what I call the armin-a-cast approach to weaning. This analogy may be appro priate for the accessory muscles of respiration but not for the diaphragm. That is, as mentioned in Chapter 26, the diaphragm is not a voluntary muscle P.469 th at will stop contracting during mechanical ventilation. The contraction of the d iaphragm is dictated by the activity of brainstem respiratory neurons, and these neurons continue to fire periodically throughout life. In fact, these neurons m ust continue to fire periodically for life to continue. Thus, the diaphragm cont racts periodically throughout life, regardless of the presence or absence of mec hanical ventilation. When the diaphragm contracts and triggers a ventilator brea th, the contraction does not cease, but continues for the duration of inspiratio n set by the brainstem respiratory neurons (5). Therefore, the nature of diaphra gm contraction implies that the diaphragm does not necessarily become weak durin g mechanical ventilation. However, other conditions in ventilator-dependent pati ents in the ICU (e.g., shock, low cardiac output, and hypophosphatemia) could pr omote diaphragm weakness. Weaning Because the diaphragm does not necessarily become weak during mechanical ventila tion, the focus on diaphragm weakness as a cause of failure to wean may be misdi rected. In fact, in a study we performed, there was no evidence that diaphragm w eakness was the cause of failure to wean from mechanical ventilation (6). Failur e to document diaphragm weakness as a cause of failure to wean is also noted by others (7). BE SI E WEANING PARAMETERS The initial step in the wean process is to identify patients who are candidates for weaning. All candidates should have adequate arterial oxygenation at an FIO2 of 0.5 or less and extrinsic positive end-expiratory pressure (PEEP) of 5 cm H2 O or less. Once this condition is satisfied, a number of easily obtained bedside parameters can be used to predict the likelihood of http://gateway.ut.ovid.com/gw1/ovidweb.cgi (2 of 17)03-May-05 08:29:43

Ovid: ICU Book removing the patient from ventilatory support. These parameters are listed in Ta ble 29.2. Although easily obtained, many of the popular bedside weaning paramete rs have a poor predictive value, particularly when applied to individual patient s (8). The poor predictive value of the minute ventilation measurement is shown in Figure 29.1 (8). In this study, the minute ventilation incorrectly predicted the outcome of the wean in 40 to 50% of the patients, P.470 which is about the p redictive value of a coin toss. The most predictive of the bedside parameters ar e described below. TABLE 29.2. BE SI E WEANING PARAMETERS http://gateway.ut.ovid.com/gw1/ovidweb.cgi (3 of 17)03-May-05 08:29:43

Ovid: ICU Book Figure 29.1. The predictive value of selected weaning parameters. VE = minute ve ntilation, Pmax = maximin inspiratory pressure, RR/VT = ratio of respiratory rat e to tidal volume. ( ata from Yang K, Tobin MJ. A prospective study of indexes p redicting the outcome of weaning from mechanical ventilation. N Engl J Med 1991; 324:1445.) MAXIMUM INSPIRATORY PRESSURE The strength of the diaphragm and other muscles of inspiration can be evaluated by having the patient exhale to residual lung volume and then inhale as forceful ly as possible against a closed valve (9). The airway pressure generated by this maneuver is called the maximum inspiratory pressure (PImax). Normal men can gen erate a negative PImax over 100 cm H2O, whereas women achieve slightly lower lev els (Table 29.2). The threshold PImax for predicting a successful wean is 20 to 30 cm H2O (negative pressure). P.471 Predictive Value The predictive value of the PImax . When the PImax was less than 20 wever, 40% of the patients with a fore, a low PImax is valuable for r, an acceptable PImax has little in one large study is shown in Figure 29.1 (8) cm H2O, no patient was successfully weaned. Ho PImax above 20 cm H2O also did not wean. There identifying patients who will not wean. Howeve value for predicting a successful wean.

FREQUENCYVOLUME RATIO Rapid and shallow breathing is common in patients who fail to wean from mechanic al ventilation. An index of rapid and shallow breathing is provided by the ratio of respiratory rate to tidal volume (RR/VT). This ratio is normally less than 5 0 breaths/minute/L, and is often above 100 breaths/minute/L in patients who do n ot wean from the ventilator. Predictive Value The predictive value of the frequencyvolume ratio in one study is shown in Figure 29.1 (8). When the RR/VT ratio was above 105 breaths/minute/L, 95% of the wean attempts were unsuccessful. However when the RR/VT ratio was below 105, 80% of t he wean attempts were a success. Thus, the respiratory ratetidal volume ratio has a high predictive value for identifying patients who will and will not wean. Un fortunately, this predictive value is not http://gateway.ut.ovid.com/gw1/ovidweb.cgi (4 of 17)03-May-05 08:29:43

Ovid: ICU Book consistently high in all studies (10). Nevertheless, this parameter is one of th e most predictive bedside weaning parameters. METHO S OF WEANING There are two general approaches to withdrawing ventilator support. One involves periods of spontaneous breathing interspersed with periods of ventilatory suppo rt. The other involves a gradual reduction in the fraction of total ventilation contributed by the ventilator. The abrupt withdrawal method is known as T-piece weaning because the original breathing apparatus was shaped like the letter T. T he gradual withdrawal method uses intermittent mandatory ventilation (IMV), a mo de of ventilation described in Chapter 27. T-PIECE WEANING This method is like an onoff toggle switch that switches between periods on and o ff the ventilator. The original circuit design for the P.472 spontaneous breathi ng periods is shown in Figure 29.2 (note the T-shaped arrangement of the tubing) . The inhaled gas is delivered at a high flow rate through the upper arm of the apparatus. The high flow rate serves two purposes. First, it prevents the patien t from inhaling room air from the exhalation side of the apparatus. Second, it c arries the exhaled gas from the patient out of the apparatus, and prevents rebre athing. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (5 of 17)03-May-05 08:29:43

Ovid: ICU Book Figure 29.2. iagram of the T-shaped circuit used for spontaneous breathing tria ls. A continuous flow of gas passes along the horizontal upper arm of the T-circ uit, and this prevents the patient from inhaling room air or rebreathing exhaled gas. Protocols A variety of protocols are used for T-piece weaning. Each uses a specified perio d for spontaneous breathing and ventilator support (e.g., 4 hours off, 4 hours o n) and gradually increases the time off the ventilator. The method I prefer is t o keep the patient off the ventilator as long as tolerated. If spontaneous breat hing is tolerated for 24 hours, the wean is considered successful. If the patien t fails the T-piece trial, ventilatory support is resumed until the patient appe ars comfortable, and then another T-piece trial begins and lasts as long as tole rated. With this method, the patient is never placed back on the ventilator as l ong as spontaneous breathing is tolerated. Minimum Pressure Support T-piece breathing trials are now usually conducted while the patient is attached to the ventilator (this allows breath-to-breath monitoring P.473 of tidal volum e and respiratory rate). The only drawback with this arrangement is the increase d resistance to breathing created by the ventilator tubing and the actuator valv es in the circuit. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (6 of 17)03-May-05 08:29:43

Ovid: ICU Book To overcome this resistance, pressure-support ventilation (described in Chapter 27) is often used during spontaneous breathing trials. The goal is to provide en ough pressure to overcome the resistance of the ventilator circuit without augme nting the patient s spontaneous inflation volumes, so this method is called mini mum pressure-support. The minimum pressure needed to overcome resistance in the ventilator circuit can be determined as follows (11): where PIFR is the peak inspiratory flow rate during spontaneous breathing (measu red while the patient breathes spontaneously through the ventilator circuit) and R is the resistance to inspiratory flow during mechanical ventilation. The resi stance to flow is derived as descried in Chapter 26; that is, R = Ppk Ppl/[V wit h dot above]insp, where Ppk and Ppl are the endinspiratory peak and plateau pres sures, respectively, and [V with dot above]insp is the inspiratory flow rate del ivered by the ventilator. The minimum pressure derived in this fashion is then u sed for the periods of spontaneous breathing and is continued until the patient is extubated. IMV WEANING IMV provides a predetermined number of machine breaths each minute and allows th e patient to breathe spontaneously between the machine breaths (see Fig. 27.1). IMV can be used to wean ventilatory support by gradually decreasing the number o f machine breaths delivered each minute. The time over which this takes place is extremely variable and depends on the condition of the patient and the preferen ces of the ICU staff. Pressure-support ventilation is often added to assist the spontaneous breathing efforts during IMV, for the same reason as described for T -piece weaning. The False Security of IMV There is a tendency to view IMV as safer method of weaning because IMV provides a certain number of backup machine breaths. However, the IMV mode does not adjus t to changes in a patient s ventilatory demands (i.e., it is not a closed-loop f eedback system). If the minute ventilation of the patient declines, there is no compensation by the ventilator to maintain a constant minute ventilation. Theref ore, the patient can crash just as easily during IMV weaning as during a T-piece trial. The false sense of security with IMV weaning is dangerous because P.474 it leads to the perception that patients weaned with IMV can be left unattended more than http://gateway.ut.ovid.com/gw1/ovidweb.cgi (7 of 17)03-May-05 08:29:43

Ovid: ICU Book patients weaned with a T-piece. WHICH METHO ? The early experience with weaning suggested that all methods were equivalent. Ho wever, more recent studies indicate that daily T-piece trials result in the most rapid weaning from mechanical ventilation (12,13). Furthermore, the superiority of daily T-piece trials seems not to be a function of the method itself, but is rather a result of early recognition of patients capable of spontaneous breathi ng (13). Remember that the end-point of IMV weaning is a T-piece trial, so the t ime involved in tapering the machine breaths during an IMV wean is wasted time f or patients who are capable of unassisted breathing. This leads to a very import ant point about weaning: The important factor in successful weaning is not the m ethod used, but the ability to recognize when a patient is capable of spontaneou s unassisted breathing. Most patients require ventilator assistance because thei r lungs do not work, and weaning does not improve lung function. When lung funct ion is adequate to support spontaneous breathing, the patient will wean, regardl ess of the weaning technique used. COMPLICATING FACTORS The following factors can make weaning more difficult; attention to these factor s is necessary to ensure an optimal attempt at weaning. YSPNEA Anxiety and dyspnea are common during weaning, regardless of the weaning method used (14). yspnea itself can be detrimental because it causes rapid breathing, which can produce intrinsic or occult PEEP (see Chapter 28). The best way to avo id or limit anxiety and dyspnea is to provide adequate sedation. Morphine works well for relieving the sense of dyspnea during weaning. For patients with chroni c CO2 retention, haloperidol (which does not depress ventilation) is an appropri ate substitute. CAR IAC OUTPUT The transition from positive-pressure ventilation to negative-pressure breathing can result in a decrease in cardiac output (caused by an increase in left-ventr icular afterload) (15). This decrease can impair weaning, not only by promoting pulmonary congestion, but also by impairing diaphragm function. The diaphragm de pends heavily on cardiac output (like the heart, the diaphragm maximally extract s oxygen), P.475 http://gateway.ut.ovid.com/gw1/ovidweb.cgi (8 of 17)03-May-05 08:29:43

Ovid: ICU Book and a drop in cardiac output can decrease the strength of diaphragmatic contract ions (16). Therefore, the cardiac output should be monitored if possible (i.e., with an indwelling pulmonary artery catheter) in patients with cardiac dysfuncti on. If necessary, dobutamine can be used for temporary support of the cardiac ou tput. OVERFEE ING An increase in daily caloric intake results in an increase in metabolic CO2 prod uction, as illustrated in Figure 29.3 (17). Excess calories therefore result in excess CO2 production, and this can impair the ability to wean from mechanical v entilation (17,18). For this reason, patients receiving prolonged mechanical ven tilation should have their daily energy needs measured with indirect calorimetry (not estimated with conventional formulas) and the daily caloric intake should not exceed daily needs (see Chapter 46). Figure 29.3. The influence of daily caloric intake on metabolic CO2production (V CO2) in mechanically ventilated patients. REE is the resting energy expenditure (in kcal/24 hours) determined by indirect calorimetry. Each data point represent s a mean value for all measurements obtained. ( ata from Talpers S, et al. Nutri tionally associated increased carbon dioxide production. Chest 1992;102:551.) ELECTROLYTE EPLETION epletion of magnesium and phosphorous can impair respiratory muscle strength (1 8,19), and http://gateway.ut.ovid.com/gw1/ovidweb.cgi (9 of 17)03-May-05 08:29:43

Ovid: ICU Book deficiencies in these electrolytes should be corrected for optimal weaning. P.47 6 THE PROBLEM WEAN The following problems are common sources of difficulty in discontinuing ventila tor support. RAPI BREATHING One of the most common problems in weaning is an increase in the respiratory rat e. In this situation, the task is to distinguish anxiety from respiratory muscle fatigue or cardiopulmonary insufficiency. The flow diagram in Figure 29.4 shows a simple approach to this problem using the spontaneous tidal volume. Anxiety i s accompanied by hyperventilation, where the tidal volume is usually increased, whereas the pathologic causes of wean failure (i.e., muscle fatigue and cardiopu lmonary P.477 disease) usually cause rapid shallow breathing. Therefore, an incr ease in tidal volume suggests anxiety, whereas a decrease in tidal volume sugges ts true wean failure. When the tidal volume is unchanged or increased, arterial blood gases can further help identify the problem; i.e., a decrease in arterial PCO2 suggests anxiety (hyperventilation), whereas an unchanged or rising arteria l PCO2 indicates true wean failure. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (10 of 17)03-May-05 08:29:43

Ovid: ICU Book Figure 29.4. An approach to rapid breathing during weaning from mechanical venti lation. AB OMINAL PARA OX The respiratory movements of the abdomen can provide information about the funct ional integrity of the diaphragm (20). When the diaphragm contracts, it descends into the abdomen and increases intraabdominal pressure. This pushes the anterio r wall of the abdomen outward. However, when the diaphragm is weak, the negative intrathoracic pressure created by the accessory muscles of respiration pulls th e diaphragm upward into the thorax. This decreases the intraabdominal pressure a nd causes a paradoxical inward displacement of the abdomen during inspiration. T his phenomenon is called abdominal paradox. Therefore, the appearance of abdomin al paradox during weaning is a possible sign of a diaphragmatic weakness. Predictive Value Abdominal paradox is a reliable sign of bilateral diaphragm weakness only during quiet breathing. When breathing is labored, contraction of the accessory muscle s of inspiration can http://gateway.ut.ovid.com/gw1/ovidweb.cgi (11 of 17)03-May-05 08:29:43

Ovid: ICU Book overcome the contractile force of the diaphragm and pull the diaphragm up into t he thorax during inspiration. This results in abdominal paradox in the face of a n actively contracting diaphragm. Because quiet breathing may be uncommon during a difficult wean, the appearance of abdominal paradox during weaning is not nec essarily a sign of diaphragm weakness. However, it is a sign of labored breathin g, and should prompt immediate return to full ventilatory support. HYPOXEMIA The approach to hypoxemia is presented in Chapter 21 (see Fig. 21.6). One potent ial cause of hypoxemia that deserves mention in this setting is a decrease in ca rdiac output. The hypoxemia in this case is caused by a decrease in mixed venous oxygen saturation (SVO2), so SVO2 can be an important measurement in the evalua tion of hypoxemia during weaning. HYPERCAPNIA The approach to hypercapnia is described in Chapter 21 (see Fig. 21.7). The appe arance of hypercapnia is an ominous sign, and should P.478 prompt immediate retu rn to full ventilatory support. If the end-tidal PCO2 (PETCO2) is being monitore d during the wean, the gradient between end-tidal and arterial PCO2 can help ide ntify the problem. An increase in the PaCO2 PETCO2 gradient indicates an increase in dead space ventilation (from a decrease in cardiac output or hyperinflation with intrinsic PEEP), whereas an unchanged gradient suggests respiratory muscle fatigue or enhanced CO2 production. TRACHEAL ECANNULATION When the patient no longer requires ventilator support, the next step is to remo ve the tracheal tube, if possible. Successful weaning from mechanical ventilatio n is not synonymous with tracheal decannulation. When a patient has successfully weaned but is not fully awake, or is unable to clear respiratory secretions, th e tracheal tube should be left in place. WORK OF BREATHING The cross-sectional area of the glottis (the narrowest portion of the upper airw ay) is 66 mm3 in the average-sized adult, whereas the cross-sectional area of an average-sized adult tracheal tube (8 mm internal diameter) is 50 mm3 (21). The narrower cross-sectional area of the http://gateway.ut.ovid.com/gw1/ovidweb.cgi (12 of 17)03-May-05 08:29:43

Ovid: ICU Book tracheal tube creates an increased resistance to airflow, so the work of breathi ng should be greater when breathing through tracheal tubes (both endotracheal an d tracheostomy tubes). This observation has led to the perception that tracheal decannulation results in a decreased work of breathing. However, as shown in Fig ure 29.5, tracheal decannulation P.479 is accompanied by an increased work of br eathing (11). This is presumably the result of laryngeal edema produced by endot racheal tubes. Although this obstruction rarely requires reintubation, the incre ased work of breathing has implications for the occasional practice of removing tracheal tubes to reduce the work of breathing. That is, tracheal decannulation should never be performed to reduce the work of breathing. Figure 29.5. The work of breathing in joules/liter (J/L) after removal of endotr acheal tubes (size 8). Each column shows the mean and standard error of the mean . (From Nathan S et al. Prediction of minimal pressure support during weaning f rom mechanical ventilation. Chest 1993;103:1215.) EN OTRACHEAL TUBES Although some degree of laryngeal edema occurs during more than 90% of endotrach eal intubations (22), the incidence of significant upper airway obstruction (i.e ., obstruction requiring reintubation) is less than 2% (22,23). The signs of upp er airway obstruction following extubation are labored and stridorous breathing. Because the obstruction is extrathoracic, the stridorous sounds should occur du ring inspiration (when the negative pressure of inspiration leads to narrowing i n the extrathoracic portion of the upper airways). Inspiratory stridor is a http://gateway.ut.ovid.com/gw1/ovidweb.cgi (13 of 17)03-May-05 08:29:43

Ovid: ICU Book sign of severe obstruction (more than 80%), and should prompt immediate reintuba tion. The management of severe upper airway obstruction after extubation include s reintubation followed by tracheostomy. Steroids do not reduce the severity of postextubation laryngeal edema (23). Aerosolized epinephrine (2.5 mL of 1% epine phrine or 2.25% racemic epinephrine) has proven effective in reducing postextuba tion laryngeal edema in children (24). However, epinephrine treatments should ne ver delay reintubation. TRACHEOSTOMY TUBES Removal of tracheostomy tubes is a two-step process. The first step is to replac e the cuffed tracheostomy tube with an uncuffed, fenestrated tube of the same si ze. This tube is then plugged, and the fenestration (hole) in the tube allows th e patient to breathe through the normal upper airway route. This is done to iden tify an upper airway obstruction caused by laryngeal edema. Contrary to popular belief, tracheostomy may not reduce, and can even aggravate, the laryngeal damag e caused by endotracheal intubation (although the mechanism is unclear) (22). Th erefore, evaluation for upper airway obstruction is important before removal of tracheostomy tubes. If there is no evidence of upper airway obstruction after 24 hours of breathing with the tracheostomy tube plugged, then the tube can be rem oved. Tracheal stenosis following tracheostomy usually occurs at the site of the tracheal incision, not at the site of cuff inflation. This complication usually becomes evident after the tracheal stoma closes, and is thus a late complicatio n of tracheostomy (usually appearing after the patient leaves the ICU). P.480 REFERENCES 1. Esteban A, Alfa I, Ibanez J, et al. Modes of mechanical ventilation and weani ng: a national survey of Spanish hospitals. Chest 1994;106:11881193. REVIEWS 2. Schuster P. A physiologic approach to initiating, maintaining, and withdrawi ng mechanical ventilatory support during acute respiratory failure. Am J Med 199 0;88:268 278 (148 references). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (14 of 17)03-May-05 08:29:43

Ovid: ICU Book 3. Tobin MJ, Alex CG. iscontinuation of mechanical ventilation. In: Tobin MJ, e d. Principles and practice of mechanical ventilation. New York: McGraw-Hill, 199 4;1177 1206. 4. Patel RG, Petrini MF, Norman JR. Strategies for maximizing your chances for w eaning success. J Crit Illness 1995;10:411423. MISCONCEPTIONS 5. Marini JJ. Strategies to minimize breathing effort during mechanical ventilat ion. Crit Care Clin 1990;6:635661. 6. Swartz MA, Marino PL. iaphragm strength during weaning from mechanical venti lation. Chest 1985;88:736739. 7. Hubmayr R , Rehder K. Respiratory muscle failure in critically ill patients. Semin Respir Med 1992;13:1421. BE SI E WEANING PARAMETERS 8. Yang K, Tobin MJ. A prospective study of indexes predicting the outcome of tr ials of weaning from mechanical ventilation. N Engl J Med 1991;324:14451450. 9. Marini JJ, Smith TC, Lamb V. Estimation of inspiratory muscle strength in mec hanically ventilated patients: the measurement of maximal inspiratory pressure. J Crit Care 1986;1:3238. 10. Lee KH, Hui KP, Chan TB, et al. Rapid shallow breathing (frequencytidal volum e ratio) did not predict extubation outcome. Chest 1994;105:540543. METHO S OF WEANING 11. Nathan S , Ishaaya AM, Koerner SK, et al. Prediction of minimal pressure sup port http://gateway.ut.ovid.com/gw1/ovidweb.cgi (15 of 17)03-May-05 08:29:43

Ovid: ICU Book during weaning from mechanical ventilation. Chest 1993;103:12151219. 12. Esteban A, Frutos F, Tobin MJ, et al. A comparison of four methods of weanin g patients from mechanical ventilation. N Engl J Med 1995;332:345350. 13. Ely W, Baker AM, unagen P, et al. Effect of duration of mechanical ventila tion of identifying patients capable of breathing spontaneously. N Engl J Med 19 96;335:1864 1869. P.481 COMPLICATING FACTORS 14. Bouley GH, Froman R, Shah H. The experience of dyspnea during weaning. Heart Lung 1992;21:471476. 15. Pinsky M. Cardiovascular effects of ventilatory support and withdrawal. Anes th Analg 1994;79:567576. 16. Nishimura Y, Maeda H, Tanaka K, et al. Respiratory muscle strength and hemod ynamics in heart failure. Chest 1994;105:355359. 17. Talpers SS, Romberger J, Bunce SB, Pingleton SK. Nutritionally associated i ncreased carbon dioxide production. Chest 1992;102:551555. 18. Benotti PN, Bistrian B. Metabolic and nutritional aspects of weaning from me chanical ventilation. Crit Care Med 1989;17:181185. 19. Malloy W, hingra S, Solren F, et al. Hypomagnesemia and respiratory muscle power. Am Rev Respir is 1984;129:427431. THE PROBLEM WEAN http://gateway.ut.ovid.com/gw1/ovidweb.cgi (16 of 17)03-May-05 08:29:43

Ovid: ICU Book 20. Mier-Jedrzejowicz A, Brophy C, Moxham J, Geen M. Assessment of diaphragm wea kness. Am Rev Respir is 1988;137:877883. TRACHEAL ECANNULATION 21. Kaplan J , Schuster P. Physiologic consequences of tracheal intubation. Cli n Chest Med 1991;12:425432. 22. Colice C, Stukel T, ain B. Laryngeal complications of prolonged intubation. Chest 1989;96:877884. 23. Gaussorgues P, Boyer F, Piperno , et al. o corticosteroids prevent postint ubation laryngeal edema? A prospective study of 276 adults. Crit Care Med 1988;1 6:649652. 24. Nutman J, Brooks LJ, eakins K, et al. Racemic versus l-epinephrine aerosol in the treatment of postextubation laryngeal edema: results from a prospective, randomized, double-blind study. Crit Care Med 1994;22:15911594. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (17 of 17)03-May-05 08:29:43

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 30 THE FEBRILE PATIENT Humanity has but three great enemies: Fever, famine and war. Of these, by far th e greatest, By far the most terrible, is fever. --Sir William Osler espite Osle r s harsh comments about fever, the appearance of fever in a patient in the ICU may not be a sign of impending disaster. However, it is a sign that requires att ention. This chapter presents some of the most likely causes of nosocomial fever (i.e., onset after 48 hours of hospitalization) in patients in the ICU (1,2 and 3) and describes some early diagnostic and therapeutic approaches to the febril e patient. The specific concerns in patients who are immunocompromised are prese nted in Chapter 34. BO Y TEMPERATURE Two scales (Celsius and Fahrenheit) are conversion from one scale to the other s on the Celsius scale are often called d for the degrees on a compass, not for or temperatures on the Celsius scale is used to record body temperature, and the is shown in Table 30.1. Although reading degrees centigrade, this unit is intende temperatures (4). The appropriate term f degrees Celsius.

http://gateway.ut.ovid.com/gw1/ovidweb.cgi (1 of 21)03-May-05 08:12:57

Ovid: ICU Book TABLE 30.1. TEMPERATURE CONVERSIONS NORMAL BO Y TEMPERATURE espite the fact that the body temperature is one of the most common measurement s performed in clinical medicine, there is some disagreement about what the norm al body temperature is in healthy adults. The following points illustrate some o f the confusion regarding the normal body temperature. P.486 The traditional nor m of 37 C (98.6 F) is a mean value derived from a study of axillary temperatures i n 25,000 healthy adults, conducted in the late 19th century (5). A more recent s urvey of oral temperatures in 148 healthy subjects (ages 18 to 40 years) reveale d a mean temperature of 36.8 C (98.2 F) (6). Elderly subjects have a mean body tem perature that is approximately 0.5 C (0.9 F) lower than that of younger adults (5, 7). The normal body temperature has a diurnal variation, with the nadir in the e arly morning (between 4 and 8 a.m.) and the peak in the late afternoon (between 4 and 6 p.m.). The range of diurnal variation varies in individual subjects, wit h the highest reported range in an individual subject being 1.3 C (2.4 F) (6). In the original 19th-century report, fever was defined as an axillary temperature o f 38 C (100.4 F) or greater at any time of the day (5). In the smaller, more recen t survey, fever is defined as an oral temperature of 37.2 C (99 F) or greater in t he early morning, and 37.8 C (100 F) or greater in the late afternoon (6). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (2 of 21)03-May-05 08:12:57

Ovid: ICU Book The core body temperature is approximately 1.0 C (1.8 F) higher than axillary temp eratures (8) and 0.5 C (0.9 F) higher than oral temperatures (9). The rectal tempe rature is usually higher than the oral or axillary temperatures, but the relatio nship is variable and unpredictable (10). These observations indicate that the n ormal body temperature is not a single temperature, but rather is a range of tem peratures that is influenced by age, time of day, and measurement site. EFINITION OF FEVER The observations just presented illustrate the problem with using a single tempe rature threshold to define the febrile state. Fever is best defined as a tempera ture that exceeds the normal daily temperature range for an individual subject. However, the operational definition P.487 of fever proposed by a consensus confe rence on sepsis and inflammation is a body temperature above 38 C (100.4 F) (11). Just remember that this definition has several shortcomings (i.e., it does not t ake into account the measurement site, time of the day, or age of the patient), and its validity is easily questioned. THE FEBRILE RESPONSE Fever is an adaptive process where the normal body temperature is set at a highe r level in response to circulating pyrogens (12). This is distinct from hyperthe rmia (also called hyperpyrexia), which is not an adaptive response but represent s a condition where the body is unable to control the core body temperature (13) . The following aspects of the febrile response deserve attention. Fever is a si gn of inflammation, not infection. The febrile response is initiated by inflamma tory cytokines (e.g., tumor necrosis factor and interleukin 1-B). Therefore, fev er is not a specific response to infection, but is a response to whatever incite s the production of inflammatory cytokines. In many instances, the inciting even t is tissue injury, not infection. In fact, approximately 50% of fevers in patie nts in the ICU are not the direct result of infection (14). The distinction betw een inflammation and infection is an important one, not only for the evaluation of fever, but also for the popular practice of using antimicrobial agents to tre at a fever. The severity of the febrile response is not an indication of the pre sence or severity of infection (1,3). High fevers and rigors can be associated w ith noninfectious processes (e.g., drug fever), and minor temperature elevations can accompany life-threatening infections. Therefore, the clinical presentation should not dictate the evaluation of fever or the decision to start antimicrobi al therapy. The usual causes of nosocomial fever in patients in the ICU are indi cated on the body map in Figure 30.1 (1,2 and 3). In some instances, the search for a cause of nosocomial fever is influenced by the clinical setting. This is t he case for postoperative fevers and fevers associated with specific http://gateway.ut.ovid.com/gw1/ovidweb.cgi (3 of 21)03-May-05 08:12:57

Ovid: ICU Book procedures. These situational fevers are described next. Figure 30.1. Common causes of nosocomial fever in the ICU. POSTOPERATIVE FEVER Surgery is a form of controlled tissue injury (or, in the words of r. John Mill ili, a surgeon and close friend, surgery is like being hit with a baseball bat). Because inflammation and fever are the normal response to tissue injury, it is not surprising that infection may be absent in as many as two-thirds of fevers t hat occur in the early postoperative period (15,16). Most noninfectious fevers o ccur as a single episode. For the first episode of postoperative fever, a thorou gh physical examination should be sufficient to identify an underlying infection (16). If the physical examination is unrevealing, no further clinical evaluatio n is necessary. P.488 MALIGNANT HYPERTHERMIA Malignant hyperthermia is an uncommon disorder that occurs in approximately 1 in 15,000 episodes of general anesthesia and affects approximately 1 in 50,000 adu lts (17,18). It is an inherited disorder and is characterized by an excessive re lease of calcium from the sarcoplasmic reticulum in skeletal muscle in response to general anesthetics and muscle relaxants (18). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (4 of 21)03-May-05 08:12:57

Ovid: ICU Book Clinical features include early-onset muscle rigidity, often appearing in the op erating room or postanesthesia recovery area, followed within minutes to hours b y fever, hyperpyrexia (i.e., core temperatures above 40 C or 104 F), and depressed consciousness. In 20% of cases, the temperature elevation is not accompanied by observable muscle rigidity (19). The acute syndrome is followed by progression to rhabdomyolysis, myoglobinuric renal failure, and autonomic instability. Morta lity is above 80% in untreated cases (18). Treatment The mainstay of treatment is the immediate administration of the skeletal muscle relaxant dantrolene. osing Regimen antrolene should be given in a dose or 1 to 2 mg/kg as an intravenous bolus. Th is dose should be repeated every 15 minutes if necessary P.489 to a total dose o f 10 mg/kg. A dose of 1 to 2 mg/kg orally four times a day for 3 days should fol low. antrolene works by blocking the release of calcium by the sarcoplasmic ret iculum. Early drug administration is necessary to halt the progression to rhabdo myolysis and myoglobinuric renal failure. Therapy with dantrolene has reduced mo rtality in this disorder to less than 10% (18). All patients who survive an epis ode of malignant hyperthermia should be given a Medi-Alert bracelet to prevent f urther episodes. In addition, because it is a genetic disorder (with an autosoma l dominant inheritance pattern), family members should be informed of their gene tic predisposition. WOUN INFECTIONS Surgical wounds are classified as clean (abdomen and chest unopened), contaminat ed (abdomen or chest opened), or dirty (direct contact with pus or bowel content s) (20). Most wound infections are uncomplicated (i.e., only the skin and subcut aneous tissues are involved) and the treatment is debridement. Antimicrobial the rapy (to cover streptococcus, staphylococcus, and anaerobes) should be reserved for cases of persistent erythema or for evidence of deep tissue involvement (15) . In fever that follows median sternotomy, sternal wound infection with spread t o the mediastinum is a prominent concern (21). In this situation, sternal instab ility can be an early sign of infection. Necrotizing wound infections are produc ed by Clostridia or -hemolytic streptococci. Unlike other wound infections, which appear 5 to 7 days after surgery, necrotizing infections are evident in the fir st few postoperative days. There is often marked edema around the incision, and the skin may have crepitance and fluid-filled ullae. Spread to deeper structure s is rapid and produces http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 21)03-May-05 08:12:57

Ovid: ICU Book progressive rha domyolysis and myoglo inuric renal failure. Treatment involves e xtensive de ridement and intravenous penicillin. The mortality is high (a ove 60 %) when treatment is delayed. ATELECTASIS Atelectasis is considered a common cause of fever in the early postoperative per iod (15). However, no correlation exists etween the presence of atelectasis and the appearance of postoperative fever (22). For example, atelectasis is univers al after upper a dominal surgery (where the functional residual capacity decreas es y 40 to 70% during the first postoperative week) (23), yet fever develops in only 15% of patients following major a dominal surgery (16). Postoperative atel ectasis is a radiographic diagnosis, with linear densities and crowded markings in dependent lung zones. The nonspecific nature of this finding is illustrated i n Figure 30.2. The chest film on the left (arrow pointing down) shows reduced lu ng volumes with coalescence of markings at the left lung ase. This film might e interpreted as showing atelectasis at the left lung ase. However, this film w as o tained only minutes after the chest radiograph on the right, P.490 using th e same su ject. The difference etween the two radiographs is the ody position; the film on the right is in the upright position, whereas the one on the left w as o tained in the supine position. Thus, what might e interpreted as atelectas is is merely a consequence of ody position. This is an important consideration ecause most chest radiographs in the early postoperative period are taken with the patient in the supine position. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 21)03-May-05 08:12:57

Ovid: ICU Book Figure 30.2. The influence of ody position on the appearance of the chest radio graph. Both films were o tained within minutes of each other, in the same su jec t. The arrow pointing upward and downward indicates upright and supine positions , respectively. For these reasons, it is unlikely that atelectasis is an important source of pos toperative fever. Because most of the fevers attri uted to atelectasis resolve w ithin a day or two, it is likely that these fevers are produced y an inflammato ry reaction to the surgical trauma. THROMBOEMBOLISM Throm oem olism can produce a fever that lasts for up to 1 week (24). In one aut opsy study of postoperative deaths, pulmonary em oli were present in one-third o f the patients ut were rarely the cause of death (25). This confirms the high r isk of throm oem olism in postoperative patients. The risk is highest following orthopedic procedures on the hips and knees and is lowest following cardiopulmon ary ypass surgery (26). Throm oem olism can e associated with a normal leg exa mination in more than half of the cases (27), so the a sence of leg swelling and leg tenderness does not rule out the possi ility of throm oem olism as a cause of postoperative fever. The approach to suspected throm oem olism is descri ed i n Chapter 7. ABDOMINAL ABSCESSES Localized collections of infected material in the a domen can occur with trauma to the a dominal viscera or after laparotomy. Septicemia P.491 occurs in approxi mately 50% of cases (28). Computed tomography of the a domen can uncover the col lection in more than 95% of cases (28). Initial antimicro ial therapy should e directed at Gram-negative enteric pathogens, including anaero es (e.g., Bacteroi des fragilis). Definitive treatment requires surgical or percutaneous drainage. PROCEDURES Fever associated with the following procedures may e causally linked to the pro cedure. HEMODIALYSIS Fe rile reactions during hemodialysis are attri uted to endotoxin contamination of the dialysis equipment, ut acteremia occurs on occasion (29). If toxicity i s suspected, the dialysis should e stopped, lood cultures should e o tained, and anti iotics should e started immediately. The dialysis need not e stopped if toxicity is not suspected, ut lood cultures should always e o tained. Empi ric anti iotic therapy should cover oth Gram-positive and Gram-negative http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 21)03-May-05 08:12:57

Ovid: ICU Book pathogens. Vancomycin plus aztreonam should suffice pending culture results. Bec ause of their nephrotoxic actions, aminoglycosides should e avoided if there is a chance for recovery of renal function. ENDOSCOPY Flexi le fi eroptic ronchoscopy is followed y fever in 5% of cases (30). The c ause may e release of tumor necrosis factor (an endogenous pyrogen) into the sy stemic circulation. Pneumonia and acteremia are rare (30,31). Thus, there is us ually no need for lood cultures or empiric antimicro ial therapy for post ronch oscopic fever, unless the patient shows signs of progressive sepsis (e.g., menta l status changes or hypotension) or is asplenic (32). Gastrointestinal endoscopy is also an uncommon source of infection. The highest risk seems to occur with r etrograde cholangiopancreatography, which is complicated y infection in 1% of c ases (31). Offending organisms are usually enteric pathogens. BLOOD TRANSFUSION Fe rile reactions occur in as many as 5% of patients receiving lood products. T he fever is usually the result of antileukocyte anti odies (not infection), and appears during or shortly after the transfusion. For more information on fe rile transfusion reactions, see Chapter 44. IATROGENIC FEVER Faulty thermal regulators in water mattresses and aerosol humidifiers can cause fever y transference (33). It takes only a minute to check P.492 the temperatur e settings on heated mattresses and ventilators, ut it can take far longer to e xplain why such a simple cause of fever was overlooked. COMMON INFECTIONS Common infectious sources of fever in patients in the ICU are pneumonia, urinary tract infections, and sepsis from intravascular catheters. These infections are descri ed in detail elsewhere in the ook, and are mentioned here only riefly. PNEUMONIA Pneumonia is an overdiagnosed cause of nosocomial fever in patients in the ICU. In one study of ventilator-dependent patients with fever and pulmonary infiltrat es, pneumonia was the source of fever in only 19% of the patients (34). The diag nosis of pneumonia requires quantitative cultures of http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 21)03-May-05 08:12:57

Ovid: ICU Book deep specimens o tained y ronchoalveolar lavage or ronchoscopic (protected) rush specimens. Culture of unscreened nasotracheal aspirates is misleading and s hould e a andoned. Only nasotracheal aspirates that are screened microscopicall y (to eliminate upper airway contamination) should e su mitted for culture. See Chapter 32 for more information on the diagnosis of pneumonia. The approach to pulmonary infiltrates in the immunocompromised patient is presented in Chapter 3 4. UROSEPSIS Urinary tract infections are responsi le for 35 to 45% of nosocomial infections (35). In many cases, the predisposing factor is an indwelling ladder catheter. The diagnostic and therapeutic approach to this pro lem is presented in Chapter 33. CATHETER SEPSIS Infections caused y indwelling vascular catheters should e suspected in any ca se of unexplained fever when a catheter has een in place for more than 48 hours . Some guidelines for confirming suspected catheter-related sepsis are presented in Ta le 30.2. For more information on catheterrelated sepsis, see Chapter 5. TABLE 30.2. GUIDELINES FOR CONFIRMING SUSPECTED CATHETER SEPSIS P.493 LESS COMMON INFECTIONS When the common infections are not evident, the following infections are potenti al sources of http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 21)03-May-05 08:12:57

Ovid: ICU Book nosocomial fever. PARANASAL SINUSITIS Nasogastric and nasotracheal tu es can lock the ostia that drain the paranasal sinuses; this can lead to the accumulation of infected secretions in the paranas al sinuses (36,37). The maxillary sinuses are almost always involved, and the re sulting acute sinusitis can e an occult source of fever. This complication is r eported in 15 to 20% of patients with nasal tu es (36,37). Surprisingly, paranas al sinusitis is also found in patients who are intu ated via the oral route and have free nares (36,37). Diagnosis Purulent drainage from the nares may e a sent, and the diagnosis is suggested y radiographic features of sinusitis (i.e., opacification or airfluid levels in t he involved sinuses). Radiographic detection of sinusitis can e accomplished y conventional radiography, as in Figure 30.3. The maxillary sinuses can e viewe d with a single occipitomental radiograph, called a Waters view, o tained at the edside (38). Computed tomography is also commonly used (36,37), ut seems unne cessary. It is important to emphasize that 30 to 40% of patients with radiograph ic evidence of sinusitis do not have an infection documented y culture of aspir ated material from the involved sinus (36,37). Therefore, radiographic evidence of sinusitis is not sufficient for the diagnosis of purulent sinusitis. The diag nosis must e confirmed y sinus puncture and isolation of one or more pathogens y quantitative culture, that is, at least 103 colony forming units per millili ter (CFU/ mL) (36,37). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 21)03-May-05 08:12:57

Ovid: ICU Book Figure 30.3. Opacification of the left maxillary and frontal sinuses in a patien t with nasotracheal and nasogastric tu es. Acute sinusitis was su sequently conf irmed y maxillary sinus puncture with isolation of 103 CFU/mL S. epidermidis on quantitative culture. Treatment Responsi le pathogens include streptococci, staphylococci (including Staphylococ cus epidermidis), enteric pathogens, and yeasts (mostly Candida al icans) (36,37 ). Polymicro ial infections are common. Although local irrigation of the sinuses with antimicro ial solutions has een recommended as treatment (37), a rief co urse of systemic anti iotics seems wise ecause septicemia is a complication of purulent sinusitis in critically ill patients (unlike outpatients, where septice mia is rare). When a sinus aspirate is purulent or shows organisms on Gram stain , empiric antimicro ial therapy can e started with vancomycin plus aztreonam or an aminoglycoside pending results of sinus cultures and lood cultures. Nasal t u es should also e removed when possi le. In cases of recurrent sinusitis, deco ntamination of the nares with topical antimicro ial paste should e considered ( see Chapter 6 P.494 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 21)03-May-05 08:12:57

Ovid: ICU Book for a description of anti iotic paste used for oropharyngeal decontamination). ACALCULOUS CHOLECYSTITIS Acalculous cholecystitis is an uncommon ut serious disorder reported in up to 1 .5% of critically ill patients (39). It is most common P.495 in postoperative pa tients, trauma victims, and patients receiving parenteral nutrition. The incitin g event is edema of the cystic duct, which locks drainage of the gall ladder. T he resulting clinical syndrome includes fever (70 to 95% of cases) and right upp er quadrant tenderness (60 to 100% of cases) (39). Diagnosis is often possi le w ith right upper quadrant ultrasound. Perforation of the gall ladder can occur wi thin 48 hours after onset. The treatment of choice is cholecystectomy, or percut aneous cholecystostomy in patients who are too ill to undergo surgery. (For more information on this disorder, see Chapter 33.) PSEUDOMEMBRANOUS ENTEROCOLITIS Intestinal overgrowth of Clostridium difficile is common in patients undergoing antimicro ial therapy, and this can lead to a toxin-mediated enterocolitis chara cterized y fever and diarrhea (40). The diagnosis requires documentation of C. difficile toxin in stool samples or evidence of pseudomem ranes on proctosigmoid oscopy. Therapy includes oral or intravenous metronidazole (500 mg every 6 hours ) or oral vancomycin (500 mg orally every 6 hours). This disorder is descri ed i n more detail in Chapter 33. OTHER INFECTIONS Other infections that should e considered in selected patient populations are e ndocarditis (in patients with prosthetic valves), meningitis (in neurosurgical p atients and those with human immunodeficiency virus infection), and spontaneous acterial peritonitis (in patients with cirrhosis and ascites).

NONINFECTIOUS CAUSES OF FEVER The following are some noninfectious causes of fever that can e encountered in the ICU. DRUG FEVER A variety of pharmacologic agents are capa le of causing fever, and some of the likely offenders in the ICU are listed in Ta le 30.3. The P.496 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 21)03-May-05 08:12:57

Ovid: ICU Book fever can appear as an isolated finding or can e accompanied y the other manif estations listed in the ta le (41). Note the high incidence of rigors, indicatin g that drug fever can e mistaken for sepsis. Note also the 18% incidence of hyp otension, indicating that patients with a drug fever can appear to e seriously ill. Finally, note that evidence of a hypersensitivity reaction (i.e., eosinophi lia and rash) is often a sent. TABLE 30.3. DRUG-ASSOCIATED FEVER IN THE ICU The diagnosis of drug fever usually is suggested y excluding all other potentia l causes of fever. When drug fever is a likely diagnosis, as many drugs as possi le should e discontinued and replaced with suita le alternative agents. The fe ver should disappear in 2 to 3 days if a drug is responsi le. The diagnosis can e confirmed y drug rechallenge, which should produce a fever within hours. Neuroleptic Malignant Syndrome A variant of malignant hyperthermia that is triggered y neuroleptic agents (e.g ., haloperidol) is called the neuroleptic malignant syndrome (42). The clinical presentation and management are the same as descri ed for malignant hyperthermia . Haloperidol is a popular sedative in many ICUs, so awareness of this syndrome is important in selected patients (For more information on this syndrome, see Re ference 42.) SYSTEMIC INFLAMMATORY RESPONSE SYNDROME Evidence of systemic inflammation (i.e., fever and leukocytosis) in the a sence of documented infection or other noninfectious source of fever may represent a c linical entity known as the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 21)03-May-05 08:12:57

Ovid: ICU Book systemic inflammatory response syndrome (SIRS). This syndrome can e triggered y traumatic tissue injury or translocation of endotoxin across the owel mucosa. Evidence of multiorgan dysfunction may e present. This syndrome can lead to pr ogressive multiorgan failure and death (this is descri ed in more detail in Chap ter 30). BOWEL INFARCTION Mesenteric ischemia and infarction are often accompanied y a dominal pain and t enderness, ut these may e masked in patients with altered mental status. Unfor tunately, the diagnosis is often difficult ecause physical findings and la orat ory tests are neither sensitive nor specific. Plain radiographs of the a domen c an reveal gas in the owel wall or portal venous gas, as shown in Figure 30.4. H owever, radiographic findings are often nonspecific. The diagnosis often require s laparotomy. Figure 30.4. Supine film of the a domen showing air in wall of the small owel ( outlined y arrows) and air in the hepatic veins (arrows in upper left). The pat ient died a few hours after this film was o tained. OTHER CAUSES Other noninfectious causes of fever that may e encountered in the ICU are throm oem olism, pancreatitis, and adrenal insufficiency. Each of these disorders is descri ed elsewhere. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 21)03-May-05 08:12:57

Ovid: ICU Book P.497 EARLY MANAGEMENT DECISIONS BLOOD CULTURES Blood cultures should e o tained in all cases of nosocomial fever where an infe ctious source is suspected. In the ICU, this pertains to virtually every nosocom ial fever, except those that occur in the early postoperative period. No more th an one set of lood cultures should e o tained from any one venipuncture site ( 43). The num er of sets of lood cultures (venipuncture sites) is determined y assessment of the likelihood of septicemia in individual cases. A low or moderat e pro a ility of acteremia, such as that expected from a pneumonia or urinary t ract infection, merits two sets of cultures. A high pro a ility of acteremia, a s expected in catheter-related sepsis, requires at least three sets of lood cul tures. If the patient has received antimicro ial agents, at least four sets of lood cultures are recommended for high-pro a ility septicemia cases (41). P.498 Volume of Blood The volume of lood cultured is one of the most important factors in determining the yield from lood cultures. As much as 20 to 30 mL per set of lood cultures is recommended as the optimal volume (41), ut the volume ratio of lood to ro th (culture medium) should e kept at 1:5 for each culture ottle. EMPIRIC ANTIMICROBIAL THERAPY In the a sence of an identifia le infection, empiric anti iotic therapy should e reserved for patients with evidence of septic shock (e.g., hypotension or vita l organ dysfunction) and patients who are immunocompromised (e.g., those who are neutropenic). The empiric anti iotic therapy of neutropenic patients is present ed in Chapter 34. If life-threatening sepsis is suspected in nonimmunocompromise d patients, initial therapy with vancomycin and aztreonam provides adequate cove rage for most potential offenders pending culture results. Aminoglycosides have no documented enefit over other agents such as aztreonam for covering Gram-nega tive sepsis in the nonimmuno-compromised patient. Considering the potential for nephrotoxicity with aminoglycosides, it seems wise to avoid these agents wheneve r possi le. See Chapter 35 for more information on antimicro ial agents. ANTIPYRETIC THERAPY http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 21)03-May-05 08:12:57

Ovid: ICU Book There seems to e general agreement that the ina ility to mount a fever has adve rse consequences in serious infections (44). However, antipyretic therapy is a s tandard practice in many ICUs. The following are some reasons to avoid antipyret ic therapy in the ICU. Fever as a Host Defense Mechanism High temperatures inhi it acterial growth and viral replication. The effect of ody temperature on the growth of acteria in the circulating plasma is illustra ted in Figure 30.5. Note that the temperatures used in this study correspond to the normal afe rile and fe rile temperatures of the study animals. This clearly demonstrates the enefits of fever as a host defense mechanism. Considering that serious infections are a major source of mor idity and mortality in the ICU, pr eservation of the fe rile response seems desira le to optimize host defenses aga inst micro ial invasion.

Figure 30.5. The influence of ody temperature on the growth of Pasteurella mult ocida in the lood of infected la oratory animals. The temperatures correspond t o the normal afe rile and fe rile temperatures for the study animals (ra its). (Data from Kluger M, Rothen urg BA. Fever and reduced iron: their interaction as a host defense response to acterial infection. Science 1979;203:374376.) Cooling Blankets The common practice of using cooling lankets to suppress fever is contrary to t he physiology of fever. That is, the fe rile response is an adaptive one that ra ises the operative range of the ody temperature: P.499 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 21)03-May-05 08:12:57

Ovid: ICU Book The ody ehaves as if it were responding to a cold environment. Therefore, exte rnal cooling will support the fe rile response, not inhi it it (44). For this re ason, cooling lankets are inappropriate for fever suppression. External cooling is more appropriate for hyperthermia, where the a ility to lower ody temperatu re is impaired. REFERENCES GENERAL TEXTS Mackowiak PA, ed. Fever: asic mechanisms and management. New York: Raven Press, 1991. REVIEWS 1. Clarke DE, Kimelman J, Raffin TA. The evaluation of fever in the intensive ca re unit. Chest 1991;100:213230 (58 references). 2. Holtzclaw BJ. The fe rile response in critical care: state of the science. He art Lung 1992;21:482501 (71 references). 3. Ar o MJ, Fine MJ, Hanusa BH, et al. Fever of nosocomial origin: etiology, ris k factors, and outcomes (32 references). BODY TEMPERATURE 4. Stimson HF. Celsius versus centigrade: the nomenclature of the temperature sc ale of science. Science 1962;136:254255. P.500 5. Wunderlich CA, Sequine E. Medical thermometry and human temperature. New York : William Wood, 1871. 6. Mackowiak PA, Wasserman SS, Levine MM. A critical appraisal of 98.6 F, the upp er limit http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 21)03-May-05 08:12:57

Ovid: ICU Book of the normal ody temperature, and other legacies of Carl Reinhold August Wunde rlich. JAMA 1992;268:15781580. 7. Marion GS, McGann KP, Camp DL. Core ody temperature in the elderly and facto rs which influence its measurement. Gerontology 1991;37:225232. 8. Mellors JW, Horwitz RI, Harvey MR, et al. A simple index to identify occult acterial infection in adults with acute unexplained fever. Arch Intern Med 1987; 147:666671. 9. Tand erg D, Sklar D. Effect of tachypnea on the estimation of ody temperatur e y an oral thermometer. N Engl J Med 1983;308:945946.

11. The ACCP/SCCM Consensus Conference Committee. Definitions for sepsis and org an failure and guidelines for the use of innovative therapies in sepsis. Chest 1 992;101:1644 1655. 12. Saper CB, Breder CB. The neurologic asis of fever. N Engl J Med 1994;330:18 801886. 13. Simon HB. Hyperthermia. N Engl J Med 1993;329:483487. 14. Rangel-Frausto MS, Pittet D, Costigan M, et al. The natural history of the s ystemic inflammatory response syndrome (SIRS). JAMA 1995;273:117123.

16. Freischlag J, Busuttil RW. The value of postoperative fever evaluation. Surg ery 1983;94:358363. 17. Strazis KP, Fox AW. Malignant hyperthermia: a review of pu lished cases. Ane sth Analg http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 21)03-May-05 08:12:57

POSTOPERATIVE FEVER 15. Fry DE. Postoperative fever. In: Mackowiak PA, ed. Fever: nd management. New York: Raven Press, 1991;243254.

10. Hirschman JV. Normal

ody temperature. JAMA 1992;267:414.

asic mechanisms a

Ovid: ICU Book 1993;77:297304. 18. MacLennan DH, Phillips MS. Malignant hyperthermia. Science 1992;256:789794. 19. Shannon KM, Bleck TP. How to detectand managecatastrophic thermoregulatory dis orders. J Crit Illness 1988;3:1324. 20. Ehrenkranz NJ, Meakins JL. Surgical infections. In: Bennet JV, Brachman PS, eds. Hospital infections. 3rd ed. Boston: Little, Brown, 1992;685710. 21. Loopp FD, Lytle BW, Cosgrove DM, et al. Sternal wound complications after is olated coronary artery ypass grafting: early and late mortality, mor idity, and cost of care. Ann Thorac Surg 1990;49:179187. 22. Ro erts J, Barnes W, Pennock M, Browne G. Diagnostic accuracy of fever as a measure of postoperative pulmonary complications. Heart Lung 1988;17:166169. 23. Meyers JK, Lem eck L, O'Kane H, Baue AE. Changes in functional residual capa city of the lung after operation. Arch Surg 1975;110:576583. 24. Murray HW, Ellis GC, Blumenthal DS, et al. Fever and pulmonary throm oem oli sm. Am J Med 1979;67:232235. 25. Lind lad B, Eriksson A, Bergqvist D. Autopsy-verified pulmonary em olism in a surgical department: analysis of the period from 1951 to 1988. Br J Surg 1991; 78:849852. 26. Deep vein throm osis. Implications after open heart surgery. Chest 1991;99:2 84288. 27. Weinmann EE, Salzman EW. Deep-vein throm osis. N Engl J Med 1994;331:16301641 . P.501 28. Stilwell M, Caplan ES. The septic multiple-trauma patient. Crit Care Clin 19 88;4:345373. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 21)03-May-05 08:12:57

Ovid: ICU Book PROCEDURES 29. Pollack VE. Adverse effects and pyrogenic reactions during hemodialysis. JAM A 1988;260:21062107. 30. Silver MR, Balk RA. Bronchoscopic procedures in the intensive care unit. Cri t Care Clin 1995;11:97109. 31. Spach DH, Silverstein FE, Stamm WE. Transmission of infection y gastrointes tinal endoscopy and ronchoscopy. Ann Intern Med 1993;118:117128. 32. Gillis S, Dann EJ, Berkman N, et al. Fatal Hemophilus influenzae septicemia following ronchoscopy in a splenectomized patient. Chest 1993;104:16071609. 33. Gonzalez EB, Suarez L, Magee S. Nosocomial (water ed) fever. Arch Intern Me d 1990;150:687 (letter). INFECTIONS 34. Meduri GU, Mauldin GL, Wunderink RG, et al. Causes of fever and pulmonary de nsities in patients with clinical manifestations of ventilator-associated pneumo nia. Chest 1994;106:221 235. 35. Stamm WE. Nosocomial urinary tract infections. In: Bennet JV, Brachman PS, e ds. Hospital infections. 3rd ed. Boston: Little, Brown, 1992;597610. 36. Holzapfel L, Chevret S, Madinier G, et al. Influence of long-term oro- or na sotracheal intu ation on nosocomial maxillary sinusitis and pneumonia: results o f a prospective, randomized, clinical trial. Crit Care Med 1993;21:11321138. 37. Rou y J-J, Laurent P, Gosnach M, et al. Risk factors and clinical relevance of nosocomial maxillary sinusitis in the critically ill. Am Rev Respir Dis 1994; 150:776783.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 21)03-May-05 08:12:57

38. Diagnosing sinusitis n Med

y x-ray: is a single Waters view adequate? J Gen Inter

Ovid: ICU Book 1992;7:481485. 39. Walden DT, Urrutia F, Soloway RD. Acute acalculous cholecystitis. J Intensiv e Care Med 1994;9:235243. 40. Kelley CP, Pothoulakis C, Lamont JT. Clostridium difficile colitis. N Engl J Med 1994;330:257262. NONINFECTIOUS SOURCES 41. Mackowiak PA, LeMaistre CF. Drug fever: a critical appraisal of conventional concepts. Ann Intern Med 1987;106:728733. 42. Prager LM, Millham FH, Stern TA. Neuroleptic malignant syndrome: a review fo r intensivists. J Intensive Care Med 1994;9:227234. EARLY MANAGEMENT DECISIONS 43. Aronson MD, Bor DH. Blood cultures. Ann Intern Med 1987;106:246253. 44. Styrt B, Sugarman B. Antipyresis and fever. Arch Intern Med 1990;150:15891597 . http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 21)03-May-05 08:12:57

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 31 INFECTION, INFLAMMATION, AND MULTIORGAN INJURY Inflammation is not itself considered to e a disease ut a salutary operation ut when it cannot accomplish that salutary purpose it does mischief. --John Hunt er Becoming an astute clinician requires avoiding ad ha its in clinical reasoni ng, and one of these ad ha its is the tendency to consider signs of inflammatio n (e.g., fever and leukocytosis) as evidence of infection. As is emphasized repe atedly in this chapter, inflammation and infection are two distinct entities (1) . Failure to recognize this distinction can result in errors in the management o f two of the most lethal conditions in the ICU: septic shock and multiorgan fail ure. Errors in management have een identified as a significant source of mortal ity in multiorgan failure (30% of deaths were attri uted to management errors in one report) (2), so equating inflammation and infection can e a lethal error. THE INFLAMMATORY RESPONSE The inflammatory response is a complex affair that is initiated y some insult t o the host. This insult can e any process capa le of injuring the host (e.g., m icro ial invasion, physical trauma, and urns). The purpose of the inflammatory response is to protect the host from the damaging effects of the insult, as illu strated in Figure 31.1. However, as pointed out y the astute surgeon, John Hunt er, over 200 years ago, the inflammatory response can also injure the host. The inflammatory response generates a variety of noxious su stances (e.g., proteolyt ic enzymes and oxygen meta olites) that can damage the tissues of the host (3). This damage is prevented y endogenous su stances (e.g., antioxidants) capa le o f locking or inactivating the noxious products P.503 of inflammation. However, when the inflammatory response overwhelms the normal protective http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 19)03-May-05 08:13:52

Ovid: ICU Book mechanisms availa le to the host, the inflammatory response ecomes a source of tissue damage. This is illustrated in the panel on the right in Figure 31.1. Whe n this occurs, the inflammatory response itself ecomes an insult to the host, a nd triggers further inflammation. This creates a chain reaction that leads to wi despread and progressive inflammatory injury. The clinical expression of this wi despread inflammatory injury is multiple organ dysfunction, which can progress t o multiple organ failure (1,4). Figure 31.1. The protective and injurious actions of the inflammatory response. CLINICAL SYNDROMES The relationships etween infection, inflammation, and organ injury just descri ed are the asis for a new nomenclature proposed y a Consensus Conference spons ored y the American College of Chest Physicians and the Society of Critical Car e Medicine (1). Signs of a systemic inflammatory response (e.g., fever and leuko cytosis) are called the systemic inflammatory response syndrome (SIRS). When inf ection is the underlying cause of SIRS, the condition is called sepsis. When sep sis is accompanied y dysfunction in one or more vital organs, the condition is called severe sepsis. When severe sepsis is accompanied y hypotension that is r efractory to volume infusion, the condition is called septic shock. This nomencl ature can e summarized as follows: Fever + Leukocytosis = SIRS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 19)03-May-05 08:13:52

Ovid: ICU Book SIRS + Infection = Sepsis Sepsis + Multiorgan dysfunction = Severe sepsis Severe sepsis + Refractory hypotension = Septic shock P.504 The major value of this no menclature is to highlight the distinction etween inflammation and infection; t hat is, signs of inflammation (SIRS) are not evidence of infection. SYSTEMIC INFLAMMATORY RESPONSE SYNDROME The diagnosis of SIRS requires the presence of two or more clinical signs of a s ystemic inflammatory response (1,5,6). These signs are listed in Ta le 31.1. Bec ause some of the signs (e.g., tachypnea and tachycardia) are common in patients in the ICU, SIRS can e a u iquitous condition. In one survey of 170 patients in a surgical ICU, 93% of the patients satisfied the criteria for SIRS (5). Infect ion is identified in only 25 to 50% of patients with SIRS (5,6). The remainder o f patients have one of the noninfectious conditions in Ta le 31.1, and some have no identifia le source. The important message here is that the clinical signs o f inflammation should not e assumed to indicate infection. TABLE 31.1. THE SYSTEMIC INFLAMMATORY RESPONSE SYNDROME MULTIPLE ORGAN DYSFUNCTION SYNDROME The appearance of functional a normalities in more than one vital organ system i n patients http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 19)03-May-05 08:13:52

Ovid: ICU Book with SIRS (with or without infection) is called multiple organ dysfunction syndr ome (MODS) (1). Several organ systems can e involved, and each is listed in Ta le 31.2 along with its associated clinical syndrome. The organs most often invol ved are the lungs, kidneys, cardiovascular system, and central nervous system. M ultiorgan dysfunction can progress to multiorgan failure (7,8). The most common organ failure in this setting is the acute respiratory distress syndrome descri ed in Chapter 25. TABLE 31.2. MULTIPLE ORGAN DYSFUNCTION SYNDROME (MODS) PATHOGENESIS As mentioned earlier, MODS is an inflammatory-mediated injury that occurs when t he host defenses against the inflammatory response are overwhelmed (4). Infectio n is not necessary to produce the syndrome. Activated neutrophils in the circula ting lood play an important P.505 role in the inflammatory organ injury (3,9). Specialized adhesion molecules on the surface of endothelial cells and neutrophi ls promote the adhesion of neutrophils to the endothelium (10). Once adherent, t he activated neutrophils can then release the contents of their cytoplasmic gran ules (which contain proteolytic enzymes and oxygen meta olites) and damage the e ndothelium (11). The damaged endothelium then permits infiltration of the tissue parenchyma with plasma products and inflammatory mediators; this latter process produces the organ dysfunction. These events are an important consideration in devising strategies for the management of multiorgan failure. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 19)03-May-05 08:13:52

Ovid: ICU Book MORTALITY The mortality in MODS is directly related to the num er of organ systems that fa il. This relationship is demonstrated in Figure 23.5 (see Chapter 23). When four or more vital organs have failed, mortality exceeds 80% (7,8). A scoring system (the Multiple Organ Dysfunction Score) has een developed to predict the ICU mo rtality rate of individual patients with MODS (12) (this scoring system is prese nted in Appendix 3). The relationship etween organ failures and mortality is no t surprising ecause the chances for survival should diminish as more vital orga ns fail. In fact, multiorgan failure can e viewed as an expression of the dying process; that is, failure of the vital organs is a necessary feature of death. This may seem a trivial point, ut it emphasizes that MODS is not a primary illn ess ut is a manifestation of some other disease process. SEVERE SEPSIS AND SEPTIC SHOCK As mentioned earlier, severe sepsis and septic shock are conditions in which mul tiorgan dysfunction is the result of infection. The only difference etween the two conditions is the presence of volume-resistant hypotension in septic shock. In surveys of patients in the ICU P.506 with SIRS, severe sepsis and septic shoc k occurred in 18% and 4% of the patients, respectively (5,6). Both these conditi ons can e present at the initial presentation of SIRS, or can develop days afte r the onset of SIRS (5,6). PATHOGENESIS The spectrum of microorganisms involved in disseminated infections in patients i n the ICU are shown in Ta le 31.3. The most common isolates are Gram-negative en teric pathogens (e.g., Kle siella-Entero acter spp., Pseudomonas aeruginosa, and Escherichia coli), staphylococci (Staphylococcus epidermidis and Staphylococcus aureus), and Enterococcus spp. (13,14). These organisms are also the most commo n isolates in cases of severe sepsis and septic shock (15). Although Gram-negati ve septicemia is considered the most common cause of infection-induced multiorga n injury, infection with any microorganism can result in multiorgan damage. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 19)03-May-05 08:13:52

Ovid: ICU Book TABLE 31.3. NOSOCOMIAL SEPTICEMIA IN CRITICALLY ILL PATIENTS Recent evidence indicates that the tendency to develop severe sepsis and septic shock is not a function of the organism involved ut rather is determined y the host response to infection (14,16,17 and 18). That is, the more exaggerated the inflammatory response to infection, the greater the chances of developing multi organ damage. This correlation is consistent with notion that MODS is a result o f inflammatory injury to the vital organs. HEMODYNAMIC ALTERATIONS The initial stages of severe sepsis and septic shock are often characterized y hypovolemia (19), either relative (from venous pooling) or a solute (from transu dation of fluid). This tends to produce a hypodynamic state (i.e., low cardiac o utput). When the intravascular volume is adequate, the cardiac output usually is elevated (19,20). However, P.507 intrinsic cardiac function (systolic and diast olic) is impaired in sepsis (20) and the increase in cardiac output is the resul t of tachycardia rather than an increase in stroke output. Despite the increase in cardiac output, peripheral lood can e diminished, as demonstrated in Figure 31.2 (21). In fact, contrary to the popular notion that sepsis is a hyperdynami c circulatory condition (i. e., increased lood flow and vasodilation), the hemo dynamic changes in advanced stages of sepsis more closely resem le a hypodynamic state (i.e., reduced lood flow and vasoconstriction). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 19)03-May-05 08:13:52

Ovid: ICU Book Figure 31.2. Hypodynamic circulatory changes in sepsis and septic shock. (Data f rom Astiz ME et al. Peripheral vascular tone in sepsis. Chest 1991;99:10721075.) Oxygen Transport Severe sepsis and septic shock are characterized y a defect in the peripheral e xtraction of oxygen (22). Thus, as peripheral lood flow is reduced, the normal a ility to increase peripheral oxygen extraction is impaired, and this results i n a decreased oxygen uptake (VO2) from the microcirculation. This defect is eli eved to e the principal cause of impaired tissue oxygenation in septic shock. H yperlactatemia is a characteristic finding in severe sepsis and septic shock, u t this may e the result of impaired pyruvate meta olism (23), not tissue dysoxi a (i.e., oxygen-limited energy production). The etiology of sepsis-induced hyper lactatemia is descri ed in Chapter 13 (see Fig. 13.5). HEMODYNAMIC MANAGEMENT In light of the tendency for relative and a solute hypovolemia in sepsis, the ma instay of early hemodynamic management is aggressive volume infusion (19,24). Be cause sepsis is often accompanied y hypoal uminemia (25), initial resuscitation with colloid fluids may e P.508 preferred. The end-point of volume infusion is a right-atrial (central venous) pressure of 8 to 10 mm Hg or a pulmonary capill ary wedge pressure of 18 to 20 mm Hg (see Chapter 11). If hypotension persists d espite high-normal cardiac filling pressures, dopamine is the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 19)03-May-05 08:13:52

Ovid: ICU Book hemodynamic drug most often used to raise the lood pressure. The infusion is st arted at 5 g/ kg/minute and titrated upward (see Chapter 18). However, dopamine c an actually impair tissue oxygenation in regional eds such as the splanchnic ci rculation (26), so gastric intramucosal pH should e monitored (if possi le) dur ing dopamine infusions (see Chapter 13 for more information on gastric mucosal p H monitoring). Oxygen Transport As in any case of shock, the goal of hemodynamic management should e a normal w hole ody oxygen uptake (i.e., VO2 a ove 100 mL/min/m2). However, ecause sepsis is characterized y hypermeta olism and a heightened meta olic oxygen requiremen t, some have recommended achieving a supranormal level of systemic oxygen delive ry (DO2 a ove 600 mL/min/m2 VO2 a ove 170 mL/min/m2) in patients with septic sho ck. Following adequate volume resuscitation, do utamine is started at 5 g/kg/minu te and titrated upward to achieve the desired end-point. Some studies suggest an improved outcome with this strategy (28), ut others do not (29). Furthermore, it is impossi le to achieve supranormal levels of systemic oxygen transport in 4 0 to 50% of patients (28,29), regardless of the do utamine infusion rate. Theref ore, the practice of striving for supranormal levels of systemic oxygen transpor t is not advised as a general strategy. However, if signs of continuing and prog ressive organ injury are present at normal levels of DO2 and VO2, attempts to in crease the DO2 and VO2 to supranormal levels with volume infusion and do utamine are justified. EMPIRIC ANTIMICROBIAL THERAPY When sepsis is suspected ut not yet documented, empiric antimicro ial therapy s hould e directed at the common organisms listed in Ta le 31.3. A com ination of vancomycin (covers staphylococci, enterococci, and streptococci) plus aztreonam or an aminoglycoside (to cover Gram-negative enteric organisms) should suffice, pending culture results. When sepsis of intestinal origin is suspected, coverag e for Bacteroides fragilis is advised. This can e accomplished y the addition of clindamycin or metronidazole to the a ove antimicro ial regimen, or y monoth erapy with imipenem. Empiric antimicro ial therapy is presented in more detail i n Chapter 35. STEROIDS In the 1960s, high-dose intravenous steroids emerged as a popular therapy for se ptic shock. However, steroids have not proven eneficial P.509 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 19)03-May-05 08:13:52

Ovid: ICU Book and may e harmful in patients with severe sepsis and septic shock (30,31 and 32 ). As a result, the Infectious Disease Society of America advises that steroids should e avoided in patients with disseminated sepsis and septic shock (33). NOVEL THERAPIES Neither hemodynamic nor antimicro ial therapy has had a tremendous impact on red ucing the mortality in severe sepsis and septic shock (15,17,28,29 and 30). With this in mind and considering that inflammatory tissue injury plays an important role in the pathogenesis of septic shock, newer therapeutic strategies have ee n directed at the inflammatory response. The following two approaches have recei ved the most attention. ANTI-INFLAMMATORY ANTIBODIES Fueled y dreams of a multi illion-dollar industry, drug manufacturers have unde rtaken enormous (and expensive) efforts to manufacture and use anti odies agains t various inflammatory mediators in patients with severe sepsis and septic shock . Most of the anti odies have een directed against endotoxin, tumor necrosis fa ctor, or interleukin-1 (17,34,35,36 and 37) ecause these su stances are elieve d to play an important role in the hemodynamic alterations in sepsis. Clinical Experience Large clinical trials have failed to document improved outcomes in patients with severe sepsis and septic shock who were treated with anti odies against endotox in (35), tumor necrosis factor (36), or interleukin-1 (37). In light of these re sults and the enormous expense of anti ody therapy, little enthusiasm exists for this approach to sepsis at present. Theoretical Pro lems There are two pro lems with the anti ody approach to inflammatory organ injury. First, more than 30 cytokines are involved in the inflammatory response (34), an d creating anti odies for many (or all) of these su stances is clearly not feasi le. Second, as demonstrated in Figure 31.3, inhi iting the inflammatory respons e can promote tissue injury from infection. The consequences of an inhi ited imm une response are readily demonstrated in patients who receive chemotherapy. For these two reasons, the future of immunotherapy for inflammatory organ injury is leak. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 19)03-May-05 08:13:52

Ovid: ICU Book Figure 31.3. The theoretical consequences of novel strategies for limiting infla mmatory organ injury. Note that anti-inflammatory (anti ody) therapy can e asso ciated with tissue injury, whereas antioxidant therapy is not. ANTIOXIDANT THERAPY Toxic oxygen meta olites play an important role in the inflammatory response (9, 11,38) and may also play an important role in inflammatory-mediated P.510 tissue injury (39,40). This latter possi ility has led to the consideration of antioxi dant therapy as a means of limiting inflammatory organ injury (41,42). In fact, as illustrated in Figure 31.3, antioxidant therapy has an theoretical advantage over anti-inflammatory therapy ecause antioxidants do not inhi it the inflammat ory response, ut merely protect the host from it. Another advantage of antioxid ant therapy derives from the tendency of oxidation reactions to produce a chain reaction (see Chapter 3), thus creating a self-sustaining source of pathologic t issue injury. Clinical Experience Unlike anti ody-directed therapy, the clinical experience with antioxidants is e xtremely limited. Intravenous N-acetylcysteine has shown promising results in im proving outcome in acute respiratory distress syndrome (43), ut confirmatory st udies are needed. Given the theoretical advantages of antioxidants, this should e a fecund area of clinical investigation. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 19)03-May-05 08:13:52

Ovid: ICU Book TOXIC SHOCK SYNDROME Toxic shock syndrome (TSS) is a type of multiorgan dysfunction syndrome caused y colonization or infection with a toxigenic strain of S. aureus (43). The respo nsi le organism typically inha its mucocutaneous regions (such as nasal passages and the vagina), and it releases P.511 an exotoxin that is a sor ed into the sy stemic circulation and produces SIRS and MODS. More than 80% of cases of TSS inv olve tampon use in menstruating females (the tampon may disrupt the vaginal muco sa). Other predisposing conditions are child irth, pelvic infections, sinusitis, and influenza. Although mucocutaneous colonization or infection predominates, a ny staphylococcal infection can produce TSS if the strain is toxigenic. In this context, the role of influenza as a predisposing condition for TSS can e explai ned y the increased risk of staphylococcal pneumonia following influenza (45). CLINICAL PRESENTATION The initial presentation of TSS is nonspecific, with fever, headache, and diarrh ea predominating. This is followed within 24 to 48 hours y rapid progression to multiorgan dysfunction, often accompanied y a rash (43). The early rash is a d iffuse erythematous rash that lanches on pressure. Another desquamative rash in volving the palms and soles can appear 1 to 2 weeks later, in the convalescent p hase of the illness (44). The diagnosis of TSS is ased on the clinical presenta tion and on isolation of a toxigenic strain of S. aureus from suspected mucocuta neous regions. Blood cultures are usually sterile. MANAGEMENT The management of TSS is very similar to that for septic shock (i.e., volume inf usion followed y dopamine for persistent hypotension). Anti iotic therapy has l ittle impact on the acute illness ( ecause disseminated infection is uncommon), ut it reduces the recurrence of TSS that can occur in one-third of menstruating patients (44). Vancomycin, nafcillin, or firstgeneration cephalosporins (e.g., cephalothin) are all effective. In menstruating patients, immediate tampon remov al is mandatory. Vaginal douches to clear local toxin are of unproven value. Wit h appropriate management, the mortality in TSS is less than 5% (44). ANAPHYLAXIS Anaphylaxis is a form of inflammatory-mediated organ dysfunction that is produce d y an http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 19)03-May-05 08:13:52

Ovid: ICU Book a normal inflammatory response to chemical agents. Common offenders are antimicr o ial agents, anesthetics, radiocontrast dyes, nutrients, and insect venoms (46) . Radiocontrast dyes are the most common source of serious anaphylactic reaction s, with an incidence of 1 in 1000 to 14,000 injections (47). A out 10% of these reactions are fatal (47). CLINICAL PRESENTATION Anaphylactic reactions vary in presentation and severity. Clinical manifestation s appear within minutes to a few hours after exposure P.512 to the offending age nt. In general, the more severe the reaction, the more rapid the onset of clinic al manifestations. Less serious reactions include flushing, erythematous rashes, urticaria, a dominal cramping, and diarrhea. More severe reactions include angi oedema, laryngeal edema, ronchospasm, and hypotension. The most life-threatenin g reaction is a rapid-onset cardiovascular collapse known as anaphylactic shock. MANAGEMENT The management of anaphylaxis is dictated y the clinical presentation (Ta le 31 .4). For severe reactions, prompt intervention is the key to a successful outcom e. The following are the common treatment modalities for anaphylaxis. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 19)03-May-05 08:13:52

Ovid: ICU Book TABLE 31.4. MANAGEMENT OF ANAPHYLACTIC REACTIONS P.513 Histamine lockers have a limited value in the management of anaphylaxis, even though histamine release is elieved to play a major role in the pathogenes is of anaphylaxis. Histamine H1 lockers (such as diphenhydramine) can e used t o alleviate the pruritus associated with cutaneous reactions. Histamine H2 lock ers (such as cimetidine) provide no enefit over the H1 lockers. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 19)03-May-05 08:13:52

Ovid: ICU Book Epinephrine is the drug of choice for severe anaphylactic reactions and is capa le of locking inflammatory mediator release from sensitized cells. It is given su cutaneously for most reactions, ut can e aerosolized for laryngeal edema or given intravenously for anaphylactic shock (see Ta le 31.4 for the doses of epi nephrine) (48). Volume infusion is important in hypotension ecause in dissemina ted anaphylaxis, fluid loss from the vascular space is rapid and profound. This is one setting where colloid fluids are preferred to crystalloid fluids for volu me expansion. Aminophylline is reserved for patients with persistent ronchospas m after epinephrine is administered. This drug may aggravate arrhythmias when gi ven with epinephrine and should e used with caution. Steroids are used for pers istent hypotension, despite little evidence of enefit. Hydrocortisone is favore d ecause of its mineralocorticoid effects, although there is no proven superior ity over methylprednisolone. REFERENCES CONSENSUS CONFERENCE 1. American College of Chest Physicians/Society of Critical Care Medicine Consen sus Conference Committee. Definitions of sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest 1992;101:16441655. THE SYSTEMIC INFLAMMATORY RESPONSE 2. Pellicane JV, Byrne K, DeMaria EJ. Preventa le complications and death from m ultiple organ failure among geriatric trauma victims. J Trauma 1992;33:440444. 3. Fujishima S, Aikawa N. Neutrophil-mediated tissue injury and its modulation. Intensive Care Med 1995;21:277285. 4. Pinsky MR, Matuschak GM. Multiple systems organ failure: failure of host defe nse mechanisms. Crit Care Clin 1989;5:199220. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 19)03-May-05 08:13:52

Ovid: ICU Book 5. Pittet D, Rangel-Frausto S, Li N, et al. Systemic inflammatory response syndr ome, sepsis, severe sepsis, and septic shock: incidence, mor idities and outcome s in surgical ICU patients. Intensive Care Med 1995;21:302309. 6. Rangel-Frausto MS, Pittet D, Costigan M, et al. Natural history of the system ic inflammatory response syndrome (SIRS). JAMA 1995;273:117123. MULTIPLE ORGAN DYSFUNCTION SYNDROME 7. Deitch EA. Multiple organ failure: pathophysiology and potential future thera py. Ann Surg 1992;216:117133. P.514 8. Beal AL, Cerra FB. Multiple organ failure in the 1990s. Systemic inflammatory response and organ dysfunction. JAMA 1994;271:226233. 9. Windsor ACJ, Mullen PG, Fowler AA, Sugerman HJ. Role of the neutrophil in the adult respiratory distress syndrome. Br J Surg 1993;80:1017. 10. Donnelly SC, Haslett C, Dransfield I, et al. Role of selectins in the develo pment of the adult respiratory distress syndrome. Lancet 1994;344:215219. 11. Rivkind AI, Siegel JH, Littleton M, et al. Neutrophil oxidative urst activa tion and the pattern of respiratory physiologic a normalities in the fulminant p ost-traumatic adult respiratory distress syndrome. Circ Shock 1991;33:4862. 12. Marshall JC, Cook DJ, Christou NV, et al. Multiple Organ Dysfunction Score: a relia le descriptor of a complex clinical outcome. Crit Care Med 1995;23:163816 52. SEVERE SEPSIS AND SEPTIC SHOCK 13. Pittet D, Tarara D, Wenzel RP. Nosocomial loodstream infection in criticall y ill patients. JAMA 1994;271:15981601. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 19)03-May-05 08:13:52

Ovid: ICU Book

15. Brun-Buisson C, Doyon F, Carlet J, et al. Incidence, risk factors, and outco me of severe sepsis and septic shock in adults. JAMA 1995;274:968974. 16. Guillou PJ. Biological variation in the development of sepsis after surgery or trauma. Lancet 1993;342:217220. 17. Natanson C, Hoffman WD, Suffredini AF, Eichacker PO. Selected treatment stra tegies for septic shock ased on proposed mechanisms of pathogenesis. Ann Intern Med 1994;120:771783. 18. Parillo JE. Pathogenetic mechanisms of septic shock. N Engl J Med 1993;328:1 471 1477. 19. Rackow EC, Astiz ME. Mechanisms and management of septic shock. Crit Care Cl in 1993;9:219228. 20. Snell RJ, Parillo JE. Cardiovascular dysfunction in septic shock. Chest 1991 ;99:1000 1009. 21. Astiz ME, Tilly E, Rackow ED, Weil MH. Peripheral vascular tone in sepsis. C hest 1991;99:10721075. 22. Vincent J-L, van der Linden P. Septic shock: particular type of acute circul atory failure. Crit Care Med 1990;18(Suppl):S70S74. 23. Curtis SE, Cain SM. Regional and systemic oxygen delivery/uptake relations a nd lactate flux in hyperdynamic, endotoxin-treated dogs. Am Rev Respir Dis 1992; 145:348 354. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 19)03-May-05 08:13:52

14. Dunn DL. Gram-negative Am 1994;74:621635.

acterial sepsis and sepsis syndrome. Surg Clin North

Ovid: ICU Book 24. Carcillo JA, Davis AL, Zaritsky A. Role of early fluid resuscitation in pedi atric septic shock. JAMA 1991;266:12421245. 25. Marik PE. The treatment of hypoal uminemia in the critically ill patient. He art Lung 1993;22:166170. 26. Marik PE, Mohedin M. The contrasting effects of dopamine and norepinephrine on systemic and splanchnic oxygen utilization in hyperdynamic sepsis. JAMA 1994; 272:13541357. 27. Yu M, Levy M, Smith P, et al. Effect of maximizing oxygen delivery on mor id ity and mortality rates in critically ill patients: a prospective, randomized, c ontrolled study. Crit Care Med 1993;21:830838. 28. Hayes MA, Timmins AC, Yau EHS, et al. Elevation of systemic oxygen delivery in the treatment of critically ill patients. N Engl J Med 1994;330:17171722. 29. Gattinoni L, Brazzi L, Pelosi P, et al. A trial of goal-oriented hemodynamic therapy in critically ill patients. N Engl J Med 1995;333:10251032. P.515 30. Natanson C, Danner RL, Reilly JM, et al. Anti iotics versus cardiovascular s upport in a canine model of human septic shock. Am J Physiol 1990;259:H1440H1447. 31. Bone RC, Fisher CJ, Clemmer TP. A controlled clinical trial of high-dose met hylprednisolone in the treatment of severe sepsis and septic shock. N Engl J Med 1987;317:653658. 32. VA Systemic Sepsis Cooperative Study Group. Effect of high-dose glucocortico id therapy on mortality in patients with clinical signs of systemic sepsis. N En gl J Med 1987;317:659665. 33. McGowan JR Jr, Chesney PJ, Crossley KB, LaForce FM. Guidelines for the use o f http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 19)03-May-05 08:13:52

Ovid: ICU Book systemic glucocorticosteroids in the management of selected infections. J Infect Dis 1992;165:113. NOVEL THERAPIES 34. Dinarello CA, Gelfand JA, Wolff SM. Anticytokine strategies in the treatment of the systemic inflammatory response syndrome. JAMA 1993;269:18282835. 35. McCloskey RV, Strau e RC, Sanders C, et al. Treatment of septic shock with h uman monoclonal anti ody HA-1A. Ann Intern Med 1994;121:15. 36. A raham E, Wunderink R, Silverman H, et al. Efficacy and safety of monoclona l anti ody to human tumor necrosis factor alpha in patients with sepsis syndrome . JAMA 1995;273:934941. 37. Fisher CJ, Dhainaut J-F, Opal SM, et al. Recom inant human interleukin 1 rec eptor antagonist in the treatment of patients with sepsis syndrome. JAMA 1994;27 1:1836 1843. 38. Hurst JK, Barrette WC Jr. Leukocyte oxygen activation and micro icidal oxida tive toxins. Crit Rev Biochem Molec Biol 1989;24:271328. 39. Ward PA. Mechanisms of endothelial cell injury. J La Clin Med 1991;118:42142 6.

41. Said SI, Foda HD. Pharmacologic modulation of lung injury. Am Rev Respir Dis 1989;139:15531564. 42. Henderson A, Hayes P. Acetylcysteine as a cytoprotective antioxidant in pati ents with severe sepsis: potential new use for an old drug. Ann Pharmacother 199 4;28:10861088. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 19)03-May-05 08:13:52

40. Cochrane CG. Cellular injury

y oxidants. Am J Med 1991;91(Suppl 3C):23S30S.

Ovid: ICU Book 43. Suter PM, Domenighetti G, Schaller MD, et al. N-acetylcysteine enhances reco very from acute lung injury in man: a randomized, dou le- lind, place o-controll ed clinical study. Chest 1994;105:190194. TOXIC SHOCK 44. Ciesielski CA, Broome CV. Toxic shock syndrome: still in the differential. J Crit Illness 1986;1:2640. 45. Conway EE, Ha er RS, Gumprecht J, Singer LP. Toxic shock syndrome following influenza A in a child. Crit Care Med 1991;19:123125. ANAPHYLAXIS 46. Hollingsworth HM, Giansiracusa DF, Upchurch KS. Anaphylaxis. J Intensive Car e Med 1991;6:5570. 47. Crnkovich DJ, Carlson RW. Anaphylaxis: an organized approach to management a nd prevention. J Crit Illness 1993;8:332246. 48. Fisher M. Treating anaphylaxis with sympathomimetic drugs. Br Med J 1992;305 :11071108. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 19)03-May-05 08:13:52

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 32 NOSOCOMIAL PNEUMONIA Everything hinges on the matter of evidence. --Carl Sagan If nosocomial (hospita l-acquired) pneumonia had to e characterized in one word, an appropriate term w ould e pro lematic. The major pro lems in nosocomial pneumonia are determining when it is present and identifying the responsi le pathogen (1,2 and 3). For exa mple, although fever and new pulmonary infiltrates are considered evidence of pn eumonia, 50 to 60% of patients with these clinical manifestations do not have an identifia le lung infection (4,5). As a result of the pro lem in identifying pn eumonia, the literature on the prevalence, causes, and consequences of nosocomia l pneumonia is confusing. This chapter summarizes the current knowledge on nosoc omial pneumonias in immunocompetent patients, with emphasis on the ventilator-de pendent patient in the ICU. The focus here is on the diagnostic methods that opt imize the a ility to recognize a parenchymal lung infection and identify the res ponsi le pathogen. The approach to nosocomial pneumonias in immunocompromised pa tients is descri ed in Chapter 34. PATHOGENESIS Unlike community-acquired pneumonias, in which gram-positive cocci such as Strep tococcus pneumoniae are the most common offenders, the most common isolates in n osocomial pneumonias are gram-negative enteric pathogens. The pathogens responsi le for nosocomial pneumonia in ward patients and ventilator-dependent patients are P.517 listed in Ta le 32.1, along with the frequency that each pathogen is i solated (3). Note that only organisms isolated from lood cultures, empyema flui d, or lower respiratory tract specimens are included in the ta le. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 19)03-May-05 08:17:16

Ovid: ICU Book TABLE 32.1. MICROBIAL ISOLATES IN NOSOCOMIAL PNEUMONIAS Nosocomial pneumonias are often polymicro ial, so the percentages listed in the ta le exceed 100%. In ventilator-dependent patients, the most common isolates ar e gram-negative enteric pathogens (particularly Pseudomonas species) and Staphyl ococcus aureus. As descri ed elow, the micro ial spectrum in nosocomial pneumon ias is a reflection of a change in the resident microflora in the oropharynx. COLONIZATION OF THE OROPHARYNX The aspiration of mouth secretions into the upper airways is the inciting event in most cases of pneumonia. An average of 1 illion acteria are found in each m illiliter of saliva (6), so aspiration of even microliters of saliva will introd uce large num ers of acteria into the airways. In healthy adults, the resident microflora of the oropharynx is mostly made up of commensal organisms such as an aero ic acteria and -hemolytic streptococci. However in hospit lized p tients, t he oroph rynx is often colonized with ( erobic) gr m-neg tive enteric p thogens (3). The prev lence of this coloniz tion in community-b sed nd hospit l-b sed s ubjects is shown in Figure 32.1. Note th t the princip l determin nt of coloniz tion is the presence nd severity of illness nd th t the environment (i.e., com munity versus hospit l) h s no effect. This ch nge in the resident microflor of the oroph rynx expl ins the prev lence of gr m-neg tive p thogens in nosocomi l pneumoni . http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 19)03-M y-05 08:17:16

Ovid: ICU Book Figure 32.1. Coloniz tion of the oroph rynx with gr m-neg tive erobic b cilli ( GNAB) in different groups of subjects. (D t from Joh nson WG et l. Ch nging ph rynge l b cteri l flor of hospit lized p tients. N Engl J Med 1969;281:11371140 .) P.518 B cteri l Adherence The illness-induced ch nge in the microflor of the mouth is due to ch nge in the bility of gr mneg tive org nisms to dhere to the or l mucos . In he lthy v olunteers, subst nce c lled fibronectin co ts the epitheli l cells in the mout h nd prevents gr m-neg tive p thogens from dhering to the cells. This protecti ve co ting is lost in severe illness, nd this llows gr m-neg tive p thogens to bind to the epitheli l cells in the oroph rynx. This coloniz tion is the b sis for the pplic tion of ntimicrobi l p ste to the or l mucos s method of pro phyl xis for nosocomi l pneumoni . GASTRIC COLONIZATION As described in Ch pter 6, the cid pH in the stom ch h s b ctericid l effect on microorg nisms th t re sw llowed with food or s liv (see Fig. 6.2), nd thi s ction helps m int in sterile environment in the upper g strointestin l tr c t. When g stric cidity is suppressed (with nt cids, hist mine H2 blockers, hyd rogen ion pump inhibitors, or enter l tube feedings), the b cteri th t re sw l lowed c n thrive nd colonize the g stric contents nd this provides reservoir for b cteri to seed the lungs vi regurgit tion or tr nsloc tion (3,7). As r esult, p tients who receive g stric cid suppression ther py for prophyl xis of stress ulcer bleeding h ve n incre sed risk of nosocomi l P.519 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 19)03-M y-05 08:17:16

Ovid: ICU Book pneumoni (8). On the other h nd, stress ulcer prophyl xis with sucr lf te, whic h does not decre se g stric pH in most p tients, c rries lower risk of nosocom i l pneumoni (9). For more inform tion on the risks of g stric cid suppression , see Ch pter 6. CLINICAL PRESENTATION The clinic l m nifest tions of pneumoni (i.e., fever, new pulmon ry infiltr tes , t chypne , nd dyspne ) re nonspecific. The Centers for Dise se Control nd P revention (CDC) h s proposed the following criteri for identifying pneumoni in dults (10). CDC Criteri for Pneumoni According to the CDC, pneumoni is present when the chest x-r y film shows n ew or progressive infiltr te, c vit tion, or pleur l effusion nd ny of the f ollowing occur: q A ch nge occurs in ch r cter of the sputum. q A p thogen is isol ted from blood cultures, lung biopsy, or specimens obt ined f rom the dist l irw ys (i.e., bronchi l w shings or brushings). q A virus or vir l ntigen is isol ted in respir tory secretions. q

Histologic evidence of pneumoni exists. Note th t these criteri do not include sputum gr m st ins or sputum cultures, d espite the f ct th t sputum n lysis is common in the ev lu tion of pneumoni . THE CHEST RADIOGRAPH Sensitivity According to the CDC criteri for pneumoni , the presence of new infiltr te on chest r diogr ph h s sensitivity of 100% (i.e., lw ys present). However, thi s is n ssumption, nd n unlikely one t th t. For ex mple, pulmon ry edem is not evident on chest r diogr ph until extr v scul r lung w ter h s incre sed 30 to 50% (11), nd this insensitivity is likely to pply to infl mm tory infilt r tion of the lungs http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 19)03-M y-05 08:17:16

There is di gnostic serum ntibody titer (IgM) or titers (IgG) in p ired serum s mples. q

fourfold rise in ntibody

Ovid: ICU Book s well. The sensitivity of the chest r diogr ph for p renchym l infiltr tes m y be further reduced by hyperinfl tion during mech nic l ventil tion. Therefore, in the ev lu tion of nosocomi l fever, cle r or unch nged chest r diogr ph sho uld not be t ken s bsolute evidence g inst pneumoni ( t le st in the e rly s t ges of the investig tion). If the c use of the fever is not evident on initi l ev lu tion, seri l chest films re dvised. P.520

Specificity As mentioned in the introductory p r gr ph, p tients with fever nd new pulmon r y infiltr tes often h ve no microbi l evidence of pneumoni (4,5). In most clini c l studies, the presence of new infiltr tes on chest r diogr ph h s specifici ty of 40 to 60% for the presence of pneumoni (4,5). This me ns th t in 40 to 60 % of p tients with fever, new pulmon ry infiltr tes re c used by process othe r th n lung infection. The low specificity of the chest r diogr ph is illustr ted by the c se shown in Figure 32.2. This chest film w s obt ined from n elder ly p tient who presented to the emergency dep rtment with fever, obtund tion, n d cute respir tory f ilure. No evidence of lung infection existed in cultures obt ined from the lower respir tory P.521 tr ct (by bronchoscopy). However, the urine culture grew gre ter th n 100,000 colonies of Escherichi coli, nd the s me org nism w s isol ted from blood cultures. Therefore, the chest r diogr ph i n this c se represents cute respir tory distress syndrome (ARDS) second ry to u rosepsis with gr mneg tive septicemi .

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 19)03-M y-05 08:17:16

Ovid: ICU Book

In some c ses, the p ttern of lung infiltr tion c n be helpful in identifying lung infection. For ex mple, the presence of c vit ry lesions with ir-fluid lev els is highly suggestive of necrotizing lung infection. Another ex mple of d i gnostic p ttern on chest x-r y film is shown in Figure 32.3. The chest film in this c se shows peripher l infiltr tes with rounded density in the right lowe r lung field. This p ttern is suggestive of septic pulmon ry emboli, nd subsequ ent ev lu tion in this p tient reve led evidence of tricuspid v lve endoc rditis . http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 19)03-M y-05 08:17:16

Figure 32.2. Port ble chest x-r y film from n elderly p tient with fever ute respir tory f ilure.

nd c

Ovid: ICU Book Figure 32.3. Chest film showing peripher l infiltr tes nd circul r re of de nse consolid tion in the right lower lung field. The presumptive di gnosis in th is c se w s septic pulmon ry emboli, nd echoc rdiogr phy subsequently showed ve get tions on the tricuspid v lve.

DIAGNOSTIC EVALUATION The di gnostic ev lu tion of suspected pneumoni should include cultures of bloo d, pleur l effusions, nd specimens obt ined from the lower respir tory tr ct. EXPECTORATED OR ASPIRATED SECRETIONS The import nce of voiding b d h bits in clinic l pr ctice w s mentioned in the introduction to Ch pter 31. The worst h bit to void in the ev lu tion of suspec ted pneumoni is the pr ctice of obt ining cultures on nonselected s mples of ex pector ted sputum or tr che l spir tes. This pr ctice yields erroneous inform t ion in 40 to 60% of c ses (12,13). In other words, the ch nces of obt ining ccu r te di gnostic inform tion from (expector ted or spir ted) respir tory secreti ons is equiv lent to the ch nces of predicting he ds or t ils on coin toss. To improve on this b d h bit, the specimens must first be screened microscopic lly to identify their site or origin (i.e., proxim l versus dist l irw ys). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 19)03-M y-05 08:17:16

In summ ry, the presence of n infiltr te on chest r diogr ph should not be reg rded s definitive proof of lung infection, but should prompt further studies imed t documenting infection nd identifying the responsible p thogen. P.522

Ovid: ICU Book Some criteri for identifying the origin of respir tory secretions re shown in T ble 32.2 (14,15,16 nd 17). TABLE 32.2. SPUTUM MORPHOLOGY IN THE EVALUATION OF PNEUMONIA Squ mous Epitheli l Cells The epitheli l cells th t line the or l c vity re l rge, fl ttened cells with bund nt cytopl sm nd sm ll nucleus. The morphologic ppe r nce of these cells is shown in Figure 32.4. The presence of more th n 25 squ mous epitheli l cells per low-power field ( 100) indic tes th t the specimen is cont min ted with mout h secretions (15). If evidence of cont min tion with mouth secretions exists, th e specimen should not be submitted for culture. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 19)03-M y-05 08:17:16

Ovid: ICU Book Figure 32.4. Microscopic ppe r nce (m gnific tion 400) of bronchi l brushings fr om ventil tordependent p tient. The l rge, fl ttened cells re squ mous epithe li l cells. The ov l cells with eccentric nuclei re m croph ges. The sm llest c ells re neutrophils. P.523 Lung M croph ges Lung m croph ges re l rge, ov l-sh ped cells with gr nul r cytopl sm nd sm ll, eccentric nucleus (Fig. 32.4). The size of the nucleus in m croph ge is r oughly the s me size s neutrophil. Although m croph ges c n inh bit the irw ys (16), the predomin nt home of the lung m croph ge is the dist l irsp ces. Th erefore, the presence of m croph ges indic tes th t the specimen is from the low er respir tory tr ct. As such, sputum s mple or tr che l spir te th t cont in s even one m croph ge c n be considered suit ble for culture. El stin Fibers In necrotizing lung infections, el stin fibers from the lung p renchym c n be r ecovered in expector ted or spir ted irw y secretions. These el stin fibers r e fil mentous str nds th t c n be visu lized by pl cing drop of 40% pot ssium hydroxide on sputum sme r nd pl cing cover slip over the sme r (17). When l ow-power m gnific tion is used, el stin fibers re seen s clumps of interl cing fil ments. The presence of these el stin str nds indic tes not only th t the sp ecimen is from the dist l irsp ces, but lso th t ongoing lung necrosis is pres ent. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 19)03-M y-05 08:17:16

Ovid: ICU Book P.524 Neutrophils Abund nt neutrophils in expector ted or spir ted lung secretions indic te prob ble infection, but they do not loc lize the site of the infection (i.e., tr cheo bronchitis versus pneumoni ). The term bund nt deserves emph sis here bec use n eutrophils c n m ke up 20% of the cells recovered from routine mouthw sh (16). More th n 25 neutrophils per low-power field ( 100) c n be used s evidence of i nfection (14). However, this is not to be t ken s evidence of pneumoni . To m r k the loc tion of n infection, the criteri listed in T ble 32.2 must be used. PROTECTED SPECIMEN BRUSHING (PSB) Bronchoscopic spir tes nd brushings tend to produce f lse-positive cultures be c use the ch nnel through which the specimens re obt ined often picks up cont m in nts s the bronchoscope is p ssed through the upper respir tory tr ct (18). T o elimin te the problem of cont min tion, speci lized brush device simil r to the one shown in Figure 32.5 h s been developed for collecting specimens from th e lower irw ys. The brush sits in c theter th t is housed in l rger outer c nnul th t is plugged t the dist l end. When the entire device is dv nced thr ough the bronchoscope, the se l on the outer c theter protects the brush from co nt min tion with upper irw ys secretions. When the device is p ssed out of the bronchoscope nd into the lower irw ys, the inner c theter is dv nced until it knocks off the dist l occluding plug of the outer c theter (this plug is m de o f gel tin or some other dissolv ble m teri l). The brush is then dv nced f rthe r into the lower irw ys to collect the specimen. After the brushing is P.525 ob t ined, the brush is retr cted into the inner c nnul , the inner c nnul is retr cted into the outer c nnul , nd the entire device is retr cted through the bro nchoscope. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 19)03-M y-05 08:17:16

Ovid: ICU Book Figure 32.5. The protected specimen brush (PSB) technique for obt ining uncont m in ted lower irw ys secretions. Processing the Specimen Once the device is withdr wn from the bronchoscope, the brush is pl ced in 1 mL of tr nsport medium. The 1-mL mixture of diluent nd brush specimen is then cult ured qu ntit tively, nd growth of 103 colonies (per mL) or higher is used s positive culture result (3,19,20). This threshold corresponds to b cteri l den sity of 105 to 106 colonies per mL in undiluted secretions (3). The relev nt vol umes nd b cteri l counts for PSB cultures re shown in T ble 32.2. Di gnostic V lue The di gnostic yield from qu ntit tive PSB cultures is strongly influenced by th e presence of ongoing ntibiotic ther py (3,19,20). This is illustr ted in Figur e 32.6 (20). In this study, the incidence of f lseneg tive P.526 PSB cultures in cre sed from 24 to 46% when the PSB w s obt ined during ongoing ntibiotic ther py. Bec use of the incre sed incidence of f lse-neg tive PSB cultures during nt ibiotic ther py, the PSB technique is not recommended for p tients receiving nt ibiotic ther py. Therefore, in c ses of suspected pneumoni , PSB cultures should be obt ined before empiric ntimicrobi l ther py is instituted. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 19)03-M y-05 08:17:16

Ovid: ICU Book Figure 32.6. The influence of ntibiotic ther py on the di gnostic yield of prot ected specimen brush cultures. All cultures re qu ntit tive, with threshold o f 103 cfu/mL for positive result. (From Meduri GU, Ch stre J. The st nd rdiz t ion of bronchoscopic techniques for ventil tor- ssoci ted pneumoni . Chest 1993; 102(Suppl):557S564S.) BRONCHOALVEOLAR LAVAGE Broncho lveol r l v ge (BAL) is performed by wedging the bronchoscope in dist l irw y nd l v ging the wedged lung segment with isotonic s line. A speci lize d du l c theter device with plugged tip (simil r in design to the c theter in Figure 32.5) is lso v il ble. When this c theter is used, the technique is c l led protected BAL (3). A minimum l v ge volume of 120 mL is recommended for deq u te s mpling of the l v ged lung segment (20), but sm ller volumes re lso use d (when l v ge volumes of 10 to 20 mL re used, the technique is c lled mini-BAL ). Only 25% or less of the volume instilled will be returned vi spir tion, whi ch is then submitted for qu ntit tive culture, s is done for PSB specimens. How ever for BAL, the threshold for positive culture is 104 colonies per mL (3). T he BAL threshold is 10 times higher th n the PSB threshold, but both thresholds correspond to the s me b cteri l count in undiluted secretions (T ble 32.3). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 19)03-M y-05 08:17:16

Ovid: ICU Book TABLE 32.3. QUANTITATIVE CULTURES OR PROTECTED SPECIMEN BRUSHINGS (PSB) AND BRON CHOALVEOLAR LAVAGE (BAL) Di gnostic V lue The sensitivity nd specificity of BAL cultures r nges from 70 l r to PSB cultures, the di gnostic yield from BAL cultures is by ongoing ntibiotic ther py (21). Therefore like PSB, BAL is s high yield procedure in p tients receiving ntibiotic ther

PARAPNEUMONIC EFFUSIONS Pleur l effusions re present in up to 50% of b cteri l pneumoni s (22), nd the se p r pneumonic effusions should be ev lu ted when possible. If the effusion is locul ted (i.e., it does not move with ch nge in body position), bedside ultr s ound c n be used to m rk the loc tion P.527 nd depth of the fluid. In ddition to Gr m's st in nd culture, the pleur l fluid glucose concentr tion nd pH shou ld be me sured. Cl ssific tion of the fluid s tr nsud te or exud te (by pleur l fluid protein nd LDH levels) is unnecess ry bec use this does not reli bly i dentify infected versus sterile effusions. Indic tions for Dr in ge The indic tions for immedi te dr in ge of p r pneumonic effusions re listed in T ble 32.4. The r diogr phic criteri for dr in ge include the presence of n i rfluid level in the effusion, or hydropneumothor x (both re signs of broncho pleur l fistul ). Chemic l criteri for dr in ge include pleur l fluid glucose concentr tion below 40 g/dL (2.4 mol/L) or pleur l fluid pH below 7.0 (22). If p tient with p r pneumonic effusion improves clinic lly on ntibiotics, the re is little need for further ev lu tion or dr in ge of the effusion. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 19)03-M y-05 08:17:16

to 100% (3). Simi m rkedly reduced not recommended py.

Ovid: ICU Book TABLE 32.4. MANAGEMENT OF PARAPNEUMONIC EFFUSIONS ANTIMICROBIAL THERAPY EARLY ANTIBIOTIC THERAPY E rly ntibiotic ther py is guided by microscopic inspection of lower respir tor y tr ct secretions nd p r pneumonic effusions. Some initi l ntibiotic regimens b sed on the results of the Gr m's st in re shown in T ble 32.5. These regimen s re me nt for the initi l 24 to 48 hours of ther py, pending culture results. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 19)03-M y-05 08:17:16

Ovid: ICU Book TABLE 32.5. EARLY ANTIMICROBIAL THERAPY OF NOSOCOMIAL PNEUMONIA IN NONIMMUNOCOMP ROMISED PATIENTS Gr m-Neg tive B cilli When gr m-neg tive b cilli re the sole or predomin nt org nisms, the likely p t hogens re the enteric p thogens listed in T ble 32.1. The princip l concern is to provide dequ te cover ge for Pseudomon s species. Combin tion ntibiotic the r py h s been shown to reduce mort lity comp red with monother py in p tients wi th b cteremic Pseudomon s pneumoni s (23), nd thus combin tion ther py is dvoc ted P.528 for suspected Pseudomon s pneumoni s. Suit ble regimens include n m inoglycoside plus n ntipseudomon l penicillin (c rbenicillin or tic rcillin) o r ceft zidime. When gr m-neg tive b cilli show multiple morphologies, the org ni sms m y represent Hemophilus influenz . Bec use of the rising incidence of mpic illin-resist nt H. influenz isol tes, e rly cover ge for this org nism is best chieved with trimethoprimsulf methox zole or cefuroxime (24).

Gr m-Positive Cocci When gr m-positive cocci re the predomin nt org nism of Gr m's st in, the most likely p thogens re http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 19)03-M y-05 08:17:16

Ovid: ICU Book Streptococcus nd St phylococcus species. V ncomycin is the f vored gent in thi s situ tion, p rticul rly if there is risk of methicillin-resist nt st phyloco cci or penicillin-resist nt streptococci (25). Although v ncomycin-resist nt ent erococci re becoming more prev lent (see Ch pter 35), enterococci re not commo n offenders in nosocomi l pneumoni (3). Mixed Flor The presence of mixed flor , with no one predomin nt org nism, is ch r cteristic of n erobic infections, but might lso represent polymicrobi l erobic infec tion. In both situ tions, monother py with imipenem will cover most of the poten ti l offenders (the exceptions P.529 being methicillin-resist nt st phylococci nd some str ins of Pseudomon s species). Combin tion ther py with ztreon m or n minoglycoside (for Enterob cteri ce e) plus metronid zole or clind mycin (for n erobic cover ge) lso provides dequ te bro d-spectrum cover ge. An erobic o rg nisms will not grow on routine cultures of respir tory secretions, so steri le culture report does not elimin te the possibility of n n erobic infection. ANTIBIOTIC PROPHYLAXIS A seemingly effective yet neglected ppro ch to nosocomi l pneumoni involves th e topic l pplic tion of n ntimicrobi l p ste to the or l mucos to prevent co loniz tion of the oroph rynx with p thogenic org nisms (26). The prep r tion mos t often used is methylcellulose p ste (Or b se, Squibb Ph rm ceutic ls) cont i ning 2% polymyxin, 2% tobr mycin, nd 2% mphotericin B, which is pplied to the inside of the mouth with gloved finger every 6 hours (26,27). This me sure is p rt of l rger ppro ch to preventing nosocomi l infections known s selectiv e digestive decont min tion (SDD) (27), which is described t the end of Ch pter 6. The effects of SDD on the incidence of nosocomi l pneumoni re shown in Fig ure 6.3. Considering these results, it seems th t this ppro ch to nosocomi l pn eumoni deserves much more ttention th t it currently receives. REFERENCES CONSENSUS CONFERENCE 1. Meduri GU, Joh nson WG, eds. Intern tion l Consensus Conference: Clinic l inv estig tion of ventil tor- ssoci ted pneumoni . Chest 1992;102(Suppl 1):551S588S. REVIEWS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 19)03-M y-05 08:17:16

Ovid: ICU Book 2. Griffin JG, Meduri GU. New ppro ches in the di gnosis of nosocomi l pneumoni . Surg Clin North Am 1994;78:10911122 (141 references).

3. Estes RJ, Meduri GU. The p thogenesis of ventil tor- ssoci ted pneumoni : I. Mech nisms of b cteri l tr nscoloniz tion nd irw y inocul tion. Intensive C re Med 1995;21:365383 (93 references). PATHOGENESIS 4. Meduri GU, M uldin GL, Wunderink RG, et l. C uses of fever nd pulmon ry den sities in p tients with clinic l m nifest tions of ventil tor- ssoci ted pneumon i . Chest 1994;106:221235.

7. Fiddi n-Green RG, B ker S. Nosocomi l pneumoni in the critic lly ill: produc t of spir tion or tr nsloc tion? Crit C re Med 1991;19:763769. 8. Cook DJ, Reeve BK, Guy tt GH. Stress ulcer prophyl xis in critic lly ill p ti ents. JAMA 1996;275:308314. 9. Driks MR, Cr ven DE, Celli BR, et l. Nosocomi l pneumoni in intub ted p tie nts given sucr lf te s comp red with nt cids or hist mine type-2 blockers. N E ngl J Med 1987;317:1376 1382. CLINICAL PRESENTATION 10. G rner JS, J rvis WR, Emori TG, et l. CDC definitions for nosocomi l infect ions, 1988. Am J Infect Control 1988;16:128140. 11. Pistolesi M, Mini ti M, Milne ENC, Giuntini C. Me surement of extr v scul r lung w ter. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (17 of 19)03-M y-05 08:17:16

6. Higuchi JH, Joh nson WG. Coloniz tion est Med 1982;3:133142. P.530

nd bronchopulmon ry infection. Clin Ch

5. B tes JH, C mpbell D, B rron AL, et l. Microbi l etiology of cute pneumoni in hospit lized p tients. Chest 1992;101:10051012.

Ovid: ICU Book Intensive C re World 1991;8:1621. DIAGNOSTIC EVALUATION 12. Berger R, Ar ngo L. Etiologic di gnosis of b cteri l nosocomi l pneumoni in seriously ill p tients. Crit C re Med 1985;13:833836.

14. Wong LK, B rry AL, Horg n S. Comp rison of six different criteri for judgin g the ccept bility of sputum specimens. J Clin Microbiol 1982;16:627631. 15. W shington JA. Techniques for noninv sive di gnosis of respir tory tr ct inf ections. J Crit Illness 1996;11:5562. 16. R nkin JA, M rcy T, Rochester CL, et l. Hum n irw y m croph ges. Am Rev Re spir Dis 1992;145:928933. 17. S l t RA, Lederm n MM, Shl es DM. Di gnosis of nosocomi l pneumoni in intu b ted intensive c re unit p tients. Am Rev Respir Dis 1987;135:426432.

19. Allen RM, Dunn WF, Limper AH. Di gnosing ventil tor- ssoci ted pneumoni : th e role of bronchoscopy. M yo Clin Proc 1994;69:962968. 20. Meduri GU, Ch stre J. The st nd rdiz tion of bronchoscopic techniques for ve ntil tor ssoci ted pneumoni . Chest 1992;102(Suppl):557S564S. 21. C ntr l DE, T pe TG, Reed EC, et l. Qu ntit tive culture of broncho lveol r l v ge fluid for the di gnosis of b cteri l pneumoni . Am J Med 1993;95:601607. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (18 of 19)03-M y-05 08:17:16

18. Ov ss pi n A, R ndel GI. The role of the fiberscope in the critic lly ill p tient. Crit C re Clin 1995;11:2952.

13. Fine M, Orloff J, Rihs JD, et l. Ev lu tion of housest ff physici ns' prep r tion nd interpret tion of sputum gr m st ins for community cquired pneumoni . J Gen Intern Med 1991;6:189198.

Ovid: ICU Book

ANTIMICROBIAL THERAPY 23. Hilf M, Yu VL, Sh rp J, et l. Antibiotic ther py for Pseudomon s eruginos b cteremi : outcome correl tions in prospective study of 200 p tients. Am J M ed 1989;87:540546. 24. Doern GV. Trends in ntimicrobi l susceptibility of b cteri l p thogens of t he respir tory tr ct. Am J Med 1995;99(Suppl 6B):37. 25. Wilhelm MP. V ncomycin. M yo Clin Proc 1991;66:11651170. 26. Rodriguez-Rold n JM, Altun -Cuest A, Lopez A, et l. Prevention of nosocomi l lung infection in ventil ted p tients: use of n ntimicrobi l ph rynge l non bsorb ble p ste. Crit C re Med 1990;18:12391242. 27. Heyl nd DK, Cook DJ, J eschke R, et l. Selective digestive decont min tion: n overview. Chest 1994;105:12211229. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (19 of 19)03-M y-05 08:17:16

22. Light RW, Meyer RD, S hn SA, et l. P r pneumonic effusions n Chest Med 1985;6:5562.

nd empyem . Cli

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 33 SEPSIS FROM THE ABDOMEN AND PELVIS One of the recurring themes in this book is the import nce of the g strointestin l (GI) tr ct s source of infection in critic lly ill p tients. This ch pter g in returns to th t theme nd describes the infectious risks t both ends of t he GI tr ct, including the neighboring bili ry tree. The l st section of this ch pter describes nosocomi l infections in the urin ry tr ct, with emph sis on inf ections ssoci ted with indwelling urethr l c theters.

ACALCULOUS CHOLECYSTITIS Ac lculous cholecystitis is condition th t could be described s n ileus of t he g llbl dder. Although n uncommon condition, it c n be f t l if not recognize d nd tre ted promptly (1). PATHOGENESIS There re number of conditions th t predispose to c lculous cholecystitis, s shown in T ble 33.1 (1,2 nd 3). Most c ses occur in ssoci tion with multiple tr um nd bdomin l (nonbili ry) surgery. A number of mech nisms m y be involve d in the p thogenesis of c lculous cholecystitis, including ischemi (e.g., mul tiple tr um , shock), st sis (e.g., p renter l nutrition), nd reflux of p ncre tic secretions (e.g., opioid n lgesics). In p tients with hum n immunodeficienc y virus (HIV) infection, c lculous cholecystitis is most often due to opportuni stic infections with cytomeg lovirus nd Cryptosporidium (3). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 19)03-M y-05 08:23:25

Ovid: ICU Book TABLE 33.1. ACUTE ACALCULOUS CHOLECYSTITIS CLINICAL FEATURES The clinic l m nifest tions of c lculous cholecystitis include fever, n use n d vomiting, bdomin l p in, nd right upper qu dr nt tenderness P.532 (T ble 33. 1). Abdomin l findings c n be minim l or bsent, nd fever m y be the only prese nting m nifest tion. Elev tions in serum bilirubin, lk line phosph t se, nd m yl se c n occur but re v ri ble (1,2). Di gnosis An ultr sound of the right upper qu dr nt often provides di gnostic inform tion. G llbl dder sludge nd distension of the g llbl dder re common findings but c n be nonspecific. More specific findings include g llbl dder w ll thickness of t le st 3.5 mm nd submucos l edem (1,2). If ultr sound visu liz tion is h mp ered, computed tomogr phy (CT) sc nning c n provide useful inform tion (1). M n gement Prompt intervention is necess ry to prevent progressive distension nd rupture o f the g llbl dder. The l tter complic tion h s been reported in 40% of c ses whe n di gnosis nd tre tment is del yed for 48 hours or longer fter the onset of s ymptoms (1). The tre tment of choice is cholecystectomy. In p tients who re too moribund for surgery, percut neous cholecystostomy is suit ble ltern tive. E mpiric ntibiotic ther py should be st rted when the di gnosis is first http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 19)03-M y-05 08:23:25

Ovid: ICU Book

GASTRIC COLONIZATION The GI tr ct c n become source of sepsis when overgrowth of p thogenic org nis ms is present in the bowel lumen. This occurs in the upper GI tr ct in p tients receiving g stric cid suppression ther py nd in the lower GI tr ct in p tients receiving ntimicrobi l ther py (see C. difficile colitis l ter in this ch pter ). P.533 PATHOGENESIS G stric cid suppression nd subsequent coloniz tion of the upper GI tr ct is di scussed in Ch pter 6. The p thogens th t most commonly colonize the stom ch re the s me p thogens most commonly involved in nosocomi l infections (4). This cor rel tion is shown in Figure 33.1. Although it does not prove c us l rel tionsh ip between g stric coloniz tion nd nosocomi l infections, it does show th t the upper GI tr ct serves s reservoir for p thogens th t re commonly involved i n nosocomi l sepsis. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 19)03-M y-05 08:23:25

confirmed. The combin tion of v ncomycin intestin l p thogens.

nd imipenem is suit ble

nd covers ll

Ovid: ICU Book

PREVENTIVE MEASURES The me sures imed t preventing coloniz tion of the upper GI tr ct re describe d in Ch pter 6. Two b sic ppro ches exist. The first is to void the use of ge nts th t suppress g stric cidity (i.e., nt cids, hist mine H2 nt gonists, nd hydrogen ion pump inhibitors). Prophyl xis for stress ulcer bleeding c n be cc omplished with sucr lf te, cytoprotective gent th t does not incre se the pH of g stric secretions. P.534 The second me sure imed t preventing coloniz tion of the GI tr ct is selective digestive decont min tion (SDD) with non bsorb ble ntibiotics (see Ch pter 6 for the ntibiotic regimen used in SDD). The success of SDD in reducing the incidence of nosocomi l infections is shown in Figure 6. 3. Not only do these results demonstr te the effic cy of SDD, but they lso prov ide evidence th t the GI tr ct is indeed n import nt source of nosocomi l sepsi s. Tr nsloc tion Another f ctor th t c n promote sepsis of GI origin is the tendency for disrupti on of the bowel mucos in critic lly ill p tients. This imp irs the norm l b rri er function of the bowel w ll, nd llows enteric p thogens to g in ccess to th e systemic circul tion (see Ch pter 6). Two conditions c n predispose to microbi l tr nsloc tion cross the bowel mucos : mesenteric hypoperfusion nd prolonged bowel rest. Therefore, voiding these two conditions will help limit the risk o f sepsis from the GI tr ct. Although it is not lw ys possible to void mesenter ic hypoperfusion, the use of g stric intr mucos l pH monitoring (described in Ch pter 13) will help m int in mesenteric perfusion in p tients with clinic l shoc k. Avoiding prolonged bowel rest to m int in the function l integrity of the GI mucos is discussed in Ch pter 47. CLOSTRIDIUM DIFFICILE COLITIS Coloniz tion with p thogenic org nisms c n lso occur in the lower regions of th e GI tr ct. The most troublesome intruder is Clostridium difficile, spore-form ing gr m-positive n erobic b cillus th t is now the most common c use of nosoco mi l infections involving the GI tr ct (5). Although not prominent bowel inh b it nt in he lthy subjects, this org nism flourishes when the norm l microflor o f the lower GI tr ct is ltered by ntibiotic ther py. C. difficile is not n in v sive org nism, but it el bor tes enterotoxins th t incite infl mm tion in the bowel mucos . Severe c ses of mucos l infl mm tion re ccomp nied by r ised pl que-like lesions on the mucos l surf ce http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 19)03-M y-05 08:23:25

Figure 33.1. The most commonly oloniz tion of the upper GI tr , Christou NV, Me kins JL, et ss of multiple org n f ilure.

isol ted p thogens in nosocomi l infections nd c ct in p tients in the ICU. (D t from M rsh ll JC l. The g strointestin l tr ct: the undr ined bsce Ann Surg 1993;218:111119.)

Ovid: ICU Book c lled pseudomembr nes. These lesions re responsible for the term pseudomembr n ous colitis, which is used to describe dv nced c ses of C. difficile enterocoli tis. EPIDEMIOLOGY Although C. difficile is found in fewer th n 5% of he lthy dults in the communi ty, it c n be seen in s m ny s 40% of hospit lized p tients (6). More th n h l f of the p tients who h rbor C. difficile in their stool re symptom tic (7,8). The org nism is found prim rily in p tients receiving ongoing or recent (within 2 weeks) ntibiotic ther py nd in p tients who re in close proximity to other p tients who h rbor P.535 the org nism. C. difficile is re dily tr nsmitted fro m p tient to p tient by cont ct with cont min ted objects (e.g., toilet f ciliti es) nd by the h nds of hospit l personnel (7). Strict dherence to the use of d ispos ble gloves c n signific ntly reduce the nosocomi l tr nsmission of C. diff icile (9). CLINICAL MANIFESTATIONS The most common m nifest tions of symptom tic C. difficile infection re fever, bdomin l p in, nd w tery di rrhe . Bloody di rrhe is seen in 5 to 10% of c se s (5). R rely, the enterocolitis c n progress to toxic meg colon, which presents with bdomin l distension, ileus, nd clinic l shock. This l tter complic tion c n be f t l nd requires emergent subtot l colectomy (10).

DIAGNOSTIC TESTS The di gnosis of C. difficile colitis is often b sed on the results of the l bor tory tests listed in T ble 33.2. The org nism c n be cultured from fresh stool s mples. However, bec use growth often requires 48 hours or longer to become evi dent, cultures do not provide immedi te di gnostic inform tion. R pid confirm ti on of the presence of C. difficile in the stool is possible with l tex ggluti n tion test (11). This test uses l tex p rticles co ted with n ntibody to C. d ifficile ntigens, nd when mixed with stool s mple, produces visible clumping when C. difficile is present. The most popul r di gnostic test is tissue cult ure ss y for cytotoxin th t is el bor ted by C. difficile. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 19)03-M y-05 08:23:25

Ovid: ICU Book TABLE 33.2. LABORATORY TESTS FOR C. DIFFICILE ENTEROCOLITIS The sensitivity nd specificity of these l bor tory tests re shown in T ble 33. 2. As is evident, no single l bor tory test is sufficient to confirm or exclude the di gnosis of C. difficile colitis (12). Bec use nontoxigenic str ins of C. d ifficile c n be isol ted in the stool of symptom tic p tients (3,6), tests th t document the presence of C. difficile in stool (i.e., stool culture or l tex g glutin tion test) l ck specificity. Furthermore, s m ny s 30% of p tients with presumed C. difficile colitis (positive cultures nd pseudomembr nes) c n h ve neg tive cytotoxin ss y on single stool specimen (13,14), so the cytotoxin P.536 ss y l cks sensitivity. Performing seri l stool cytotoxin ss ys is recom mended to improve the sensitivity of the C. difficile cytotoxin ss y (14). Anot her ppro ch to optimizing the di gnostic ccur cy for C. difficile colitis is t o combine tests to document both the org nism nd the cytotoxin in stool (15). Computed Tomogr phy When l bor tory tests re equivoc l or un v il ble, CT sc nning c n provide v lu ble di gnostic inform tion (16). The ch r cteristic CT findings in C. difficile colitis re shown in Figure 33.2. There is m rked thickening in the w ll of the colon, nd the edem fluid in the bowel w ll produces low density stripe betw een the mucos nd seros , cre ting t rget sign. These signs re highly sugges tive of n ctive infl mm tory enterocolitis. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 19)03-M y-05 08:23:25

Ovid: ICU Book Figure 33.2. Contr st-enh nced CT sc n of p tient with C. difficile enterocoli tis. Note the m rked thickening of the w ll of the colon (C), but not the sm ll bowel (SB). The edem fluid produces low-density stripe between the more r dio p que mucos nd seros , cre ting t rget sign. (From Br ley SE et l. Overview of di gnostic im ging in sepsis. New Horiz 1993;1:214230.) Lower GI Endoscopy Direct visu liz tion of the lower GI mucos is v lu ble but underused di gnost ic test for C. difficile colitis. The presence of pseudomembr nes P.537 on the m ucos l surf ce confirms the di gnosis of C. difficile colitis. Colonoscopy is pr eferred to proctosigmoidoscopy for optim l results. TREATMENT The st nd rd tre tment for C. difficile enterocolitis is to dminister ntibioti cs to er dic te the org nism from the stool. The recommended ntibiotic regimens re shown below. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 19)03-M y-05 08:23:25

Ovid: ICU Book q Or l v ncomycin (250 to 500 mg every 6 hours for t le st 7 d ys) is the tr diti on l tre tment of choice (but m y no longer be, s described below). V ncomycin is not bsorbed cross the bowel mucos , nd thus the drug re ches high levels i n the stool. Intr venous v ncomycin is ineffective. Despite lmost complete er d ic tion of the org nism within 1 week of ther py, rel pses occur within the firs t month fter ther py in 20% of c ses (17). q Or l or intr venous metronid zole (500 mg every 6 hours for t le st 7 d ys) is s effective s or l v ncomycin for er dic ting the org nism. Or l metronid zole h s tr dition lly been considered n ltern tive to or l v ncomycin but is beco ming the preferred ther py to limit the emergence of v ncomycin-resist nt entero cocci. Intr venous metronid zole is reserved for p tients who c nnot toler te or l medic tions. Although r rely necess ry, surgic l intervention is required when C. difficile c olitis is ssoci ted with progressive sepsis nd multiorg n f ilure, or signs of peritonitis, despite ntibiotic ther py (10). The procedure of choice is subtot l colectomy.

Preventive Ther py The best method for preventing C. difficile colitis is to limit the use of ntib iotics, p rticul rly clind mycin (18). Another preventive me sure is the or l in gestion of the ye st S cch romyces boul rdii (which is norm lly found on lychee fruits). This org nism produces prote se th t destroys the receptor site for t he C. difficile enterotoxin, nd thus coloniz tion of the lower GI tr ct with th is org nism c n prevent the infl mm tory response to toxigenic str ins of C. dif ficile. One gr m of lyophilized ye st prep r tion given or lly e ch d y c n pr oduce 50% decre se in the incidence of ntibiotic- ssoci ted di rrhe (19). De spite its effic cy, this ppro ch h s not g ined popul rity in the United St tes .

CLINICAL FEATURES Abdomin l bscesses re difficult to detect on routine clinic l ev lu tion, nd they often present with fever of uncle r etiology. The low yield of the routine clinic l ev lu tion is demonstr ted in T ble 33.3. In this series of 143 p tient s with bdomin l bscesses, loc lized bdomin l tenderness http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 19)03-M y-05 08:23:25

ABDOMINAL ABSCESSES Abdomin l bscesses most often occur s complic tions of blunt or bdomin l surgery (20,21). P.538

bdomin l tr um

Ovid: ICU Book w s present in only one-third of c ses nd p lp ble bdomin l m ss w s present in fewer th n 10% of c ses (21). Although bdomin l films t ken with the p tien t in the supine, upright, nd l ter l decubitus positions re obt ined routinely , they provide little useful inform tion. TABLE 33.3. ROUTINE CLINICAL EVALUATION IN 143 PATIENTS WITH INTRA-ABDOMINAL ABS CESSES CT sc nning of the bdomen is the single most v lu ble di gnostic test for suspe cted intr bdomin l bscess. Sensitivity nd specificity for detection of bsces ses is 90% or higher (20,21). CT sc nning in the e rly postoper tive period c n result in f lse-positive sc ns bec use of collections of blood or irrig nt solut ions in the peritone l c vity from the oper tive procedure. For optim l results, CT sc nning should be performed fter the first postoper tive week (when perito ne l fluid collections h ve resorbed) (21). MANAGEMENT Immedi te dr in ge is m nd tory for ll intr - bdomin l bscesses (22). Precise loc liz tion with CT sc nning llows m ny bscesses to be dr ined percut neously with r diogr phic lly directed dr in ge c theters. Empiric ntibiotic ther py s hould be st rted while w iting the culture results from the bscess fluid. Sing le-drug ther py h s been shown to be s effective s multiple drug regimens (23) . Popul r gents for monother py include mpicillinsulb ct m (Un syn), cefoxitin, nd imipenem (23). P.539 UROSEPSIS Urin ry tr ct infection (UTI) ccounts for 30% of ll nosocomi l infections (24) . The single most http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 19)03-M y-05 08:23:25

Ovid: ICU Book import nt predisposing f ctor is urethr l c theteriz tion. Bec use most p tients in the ICU h ve indwelling urethr l c theters, the following description is lim ited to UTIs in the c theterized p tient. PATHOGENESIS The presence of n indwelling urethr l c theter cre tes 4 to 7% risk of develo ping UTI per d y (25). B cteri l migr tion long the c theter is the presumed mech nism for the link between c theters nd UTIs; however, there is more to the puzzle. The rem ining question is why b cteri th t migr te up the urethr nd into the bl dder re not w shed out of the bl dder by the urine flow. The flushi ng ction of urine is defense mech nism th t protects the bl dder from retrogr de inv sion by skin p thogens. This protective ction expl ins why direct injec tion of b cteri into the bl dder will not produce UTI in he lthy subject (2 6).

B cteri l Adherence The missing piece in the UTI puzzle is the dherence of b cteri to the bl dder epithelium (27). The epitheli l cells of the bl dder norm lly re co ted with L ctob cillus org nisms, s shown in Figure 33.3. These org nisms re neither inv sive nor virulent, nd thus by covering the surf ce of the bl dder epithelium, t hey prevent more inv sive nd virulent org nisms from tt ching to the bl dder w ll nd est blishing residence in the bl dder. Loss of this protective co ting nd dherence or urin ry p thogens to the bl dder mucos is the fin l prelude to coloniz tion nd infection in the lower urin ry tr ct (27). However, the events th t link urethr l c theteriz tion to b cteri l dherence in the bl dder re unk nown. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 19)03-M y-05 08:23:25

Ovid: ICU Book Figure 33.3. Photomicrogr ph showing L ctob cillus org nisms bl nketing bl dde r epitheli l cell. (From Sobel JD. P thogenesis of urin ry tr ct infections: hos t defenses. Infect Dis Clin North Am 1987;1:751772.) MICROBIOLOGY The common p thogens responsible for nosocomi l UTIs in the United St tes re li sted in T ble 33.4. Two surveys from the N tion l Nosocomi l Infections Surveill nce System re included; one is from the e rly 1980s (28), nd the other is fro m the e rly 1990s (29). Note the simil rities between the two surveys, indic tin g th t the spectrum of p thogens in nosocomi l UTI ch nged little over the dec d e sep r ting the two surveys. The predomin nt p thogens re enteric gr m-neg tiv e erobic b cilli, nd the most common offender is Escherichi coli. Most commun ity- cquired UTIs involve single p thogen, where s nosocomi l, c theter- ssoci ted UTIs c n be polymicrobi l (p rticul rly with chronic indwelling c theters) (24). Polymicrobi l P.540 growth from urine in the bsence of n indwelling uret hr l c theter is most likely cont min ted specimen. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 19)03-M y-05 08:23:25

Ovid: ICU Book TABLE 33.4. MICROBIAL ISOLATES IN NOSOCOMIAL UTI CLINICAL PRESENTATION Asymptom tic b cteriuri is not considered n infection in c theterized p tients (24), nd thus the presence of clinic l m nifest tions is P.541 the h llm rk of UTI. However, the clinic l m nifest tions re few in number. Fever nd leukocyt osis (i.e., signs of systemic infl mm tory response) m y be the only m nifest tions of UTI in c theterized p tients (the indwelling c theters elimin te compl ints of urgency, frequency, nd dysuri ). Approxim tely 50% of elderly p tients will lso develop ch nge in ment l st tus in ssoci tion with UTIs (30). Sever e c ses of urosepsis c n be ccomp nied by multiorg n dysfunction progressing to multiorg n f ilure (31). DIAGNOSIS The di gnosis of UTI is not lw ys str ightforw rd in c theterized p tients, p r ticul rly in p tients with prolonged c theteriz tion (i.e., longer th n 4 weeks) . Prolonged c theteriz tion is often ssoci ted with pyuri nd b cteriuri s b seline condition, nd this complic tes the use of urine cultures nd urine mi croscopy to identify infection. In this setting, seri l ch nges in the urine re more reli ble th n ev lu tion of n isol ted urine s mple. Urine Microscopy Urine microscopy h s limited v lue in the c theterized p tient. The usu l criter i for pyuri (gre ter th n 10 leukocytes per high-power field) nd b cteriuri (2 or more org nisms per oil-immersion http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 19)03-M y-05 08:23:25

Ovid: ICU Book field) in spun urine sediment re intended for the ev lu tion of suspected UTI in outp tients (32), not c theterized p tients. As mentioned previously, c thet erized p tients c n h ve microscopic criteri for UTI chronic lly, nd thus th e criteri h ve little me ning in these p tients. One potenti l v lue of urine m icroscopy is the bility to identify cert in microbes on Gr m's st in (e.g., gr mpositive cocci versus gr m-neg tive rods), which c n be used to select e rly n timicrobi l ther py. Urine Cultures Urine cultures c n lso be misle ding. The tr dition l threshold for signific nt b cteriuri in c theterized p tients is 105 colony-forming units per mL (cfu/mL ). However, colony counts bove this threshold c n represent coloniz tion, nd c olony counts s low s 100 cfu/mL c n represent infection (24). Microbi l growth in single urine specimen m y not be s import nt s the p ttern of growth in seri l specimens. However, the time required for the results of seri l cultures to become v il ble is too long for the inform tion to be useful in m king cute m n gement decisions for suspected urosepsis. In f ct, even single urine cult ure is of little help in m king cute m n gement decisions for the p tient with suspected sepsis. EARLY ANTIBIOTIC THERAPY E rly ntibiotic ther py is recommended for p tients with suspected UTI who re immunocompromised, h ve evidence of multiorg n dysfunction, P.542 or h ve pros thetic or d m ged he rt v lve. As mentioned e rlier, the urine Gr m's st in c n be helpful in selecting the ppropri te ntibiotic. The following re some sugge stions (see lso T ble 35.1). Gr m-Neg tive B cilli If gr m-neg tive b cilli re the predomin nt org nisms on urine Gr m's st in, e rly ntibiotic ther py should include one of the following: third-gener tion c eph losporin (e.g., ceftri xone or ceft zidime), n minoglycoside ( mik cin, ge nt micin, or tobr mycin), ztreon m, or imipenem. Trimethoprimsulf methox zole is n effective gent for community-b sed UTIs but is not f vored in nosocomi l UT Is bec use of emerging resist nce in E. coli nd Klebsiell species (33). Gr m-Positive Cocci A preponder nce of gr m-positive cocci on the urine Gr m's st in suggests th t e nterococcus is the responsible p thogen (st phylococci re uncommon offenders in nosocomi l UTIs). If the p tient is not seriously ill, enterococc l UTI c n be tre ted effectively with ciproflox cin. If the p tient is seriously ill or h s prosthetic or d m ged he rt v lve, mpicillin or v ncomycin plus gent micin is the preferred regimen. Ampicillin resist nce is reported in 10 to 15% of nosocom i l enterococc l http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 19)03-M y-05 08:23:25

Ovid: ICU Book infections (34), so v ncomycin m y be preferred for combin tion ther py. CANDIDURIA The presence of C ndid in the urine c n represent coloniz tion, infection in th e urin ry tr ct, or dissemin ted c ndidi sis with ren l involvement (35,36). Mos t c ses represent coloniz tion of the lower urin ry tr ct, which is common con sequence of ntibiotic ther py nd prolonged urethr l c theteriz tion. Coloniz t ion is distinguished from infection by the presence or bsence of fever, leukocy tosis, nd other signs of sepsis. Urine colony counts re unreli ble s discri min tory test. Loc l Ther py Coloniz tion of the urin ry tr ct with C ndid is not precursor to dissemin te d c ndidi sis, nd thus ther py to er dic te the org nism is unnecess ry. Howeve r, when evidence of C ndid cystitis (i.e., fever, lower bdomin l tenderness) e xists, bl dder irrig tion with mphotericin is indic ted to er dic te the org ni sm from the urin ry tr ct (35). Amphotericin bl dder irrig tion is lso used in p tients who h ve urine cultures showing 100,000 cfu/mL, lthough the v lue of t his pr ctice is unproven. Bec use indwelling urethr l c theters prevent P.543 re tention of the drug solution in the bl dder, continuous infusion technique is used for bl dder irrig tion. The regimen is s follows: Continuous bl dder irrig tion: Add 50 mg mphotericin to 1 L of sterile w ter (not s line) nd infuse the solut ion t 40 mL/hour using three-w y bl dder c theter. Continue the irrig tion fo r 3 to 4 d ys. This regimen is effective in er dic ting the org nism in most c s es, unless the c ndiduri is second ry to dissemin ted c ndidi sis. Amphotericin is not bsorbed cross the bl dder mucos , so there is no risk of systemic toxi city. Once ther py is completed, the urethr l c theter should be removed if poss ible to prevent recurrent coloniz tion. Dissemin ted C ndidi sis Although c ndid does not usu lly dissemin te from the urin ry tr ct, dissemin t ed c ndidi sis c n seed the kidneys nd produce second ry c ndiduri . Dissemin t ed c ndidi sis should be considered when c ndiduri is ssoci ted with signs of progressive sepsis, multiorg n dysfunction, or clinic l shock. The di gnosis c n be elusive bec use blood cultures re sterile in more th n 50% of c ses of diss emin ted c ndidi sis (36). C ndid infiltr tion of the retin c n produce n end ophth lmitis, nd the lesions in this condition c n be p thognomonic for dissemi n ted c ndidi sis (36). Therefore, funduscopic ex min tion (by n ophth lmolog ist) is m nd tory when http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 19)03-M y-05 08:23:25

Ovid: ICU Book dissemin ted c ndidi sis is suspected. If this is unreve ling, evidence of C ndi d t multiple sites should be looked for. Isol tion of the org nism from three sep r te sites (e.g., urine, sputum, nd v scul r c theter) c n be used s presu mptive evidence of dissemin ted c ndidi sis (36). When dissemin ted c ndidi sis is suspected, empiric ther py c n begin with intr venous flucon zole (400 mg/d y ). This ntifung l gent m y be s effective s mphotericin for tre ting dissem in ted C ndid infections ( t le st in nonneutropenic p tients), while being muc h less toxic th n mphotericin (37). However, if the clinic l condition worsens during flucon zole ther py, intr venous mphotericin should be st rted. (For mor e inform tion on flucon zole nd mphotericin, see Ch pter 35.) REFERENCES ACALCULOUS CHOLECYSTITIS 1. W lden D, Urruti F, Solow y RD. Acute c lculous cholecystitis. J Intensive C re Med 1994;9:235243. 2. Imhof M, R unest J, Ohm nn Ch, Rohrer H-D. Acute c lculous cholecystitis com plic ting tr um : prospective sonogr phic study. World J Surg 1992;11601166. P. 544 3. Bon cini M. Hep tobili ry complic tions in p tients with hum n immunodeficien cy virus infection. Am J Med 1992;92:404411. GASTRIC COLONIZATION 4. M rsh ll JC, Christou NV, Me kins JL. The g strointestin l tr ct: the undr ine d bscess of multiple org n f ilure. Ann Surg 1993;218:111119. 5. Kelly CP, Pothoul kis C, L mont JT. Clostridium difficile colitis. N Engl J M ed 1994;330:257262.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 19)03-M y-05 08:23:25

6. Fekety R, Kim F-H, Brown D, et l. Epidemiology of tis. Am J Med 1981;70:906908.

ntibiotic ssoci ted coli

Ovid: ICU Book 7. S more MH, Venk t r m n L, DeGirol mi, et l. Clinic l nd molecul r epidemio logy of spor dic nd clustered c ses of nosocomi l Clostridium difficile di rrhe . Am J Med 1996;100:3240.

8. Johnson H, Hom nn SR, Bettin KM, et l. Tre tment of symptom tic Clostridium difficile c rriers (fec l excreters) with v ncomycin nd metronid zole. Ann Int ern Med 1992;117:297 302. 9. Johnson S, Gerding DN, Olson MM, et l. Prospective, controlled study of viny l glove use to interrupt Clostridium difficile nosocomi l tr nsmission. Am J Med 1990;88:137140.

11. Biddle WL, H rms JL, Greenberger NJ, Miner PB. Ev lu tion of ntibiotic- sso ci ted di rrhe with l tex gglutin tion test nd cell culture cytotoxicity ss y for Clostridium difficile. Am J G stroenterol 1988;84:279283. 12. Gerding DN, Br zier JS. Optim l methods for identifying Clostridium difficil e infections. Clin Infect Dis 1993;16(Suppl 4):439442. 13. Gerding DN. Di gnosis of Clostridium difficile- ssoci ted dise se: p tient s election nd test perfection. Am J Med 1996;100:485486. 14. M n be YC, Vinetz JM, Moore RD, et l. Clostridium difficile colitis: n eff icient clinic l ppro ch to di gnosis. Ann Intern Med 1995;123:835840. 15. Gerding DN, Johnson S, Peterson LR, et l. Clostridium difficile- ssoci ted di rrhe nd colitis. Infect Control Hosp Epidemiol 1995;16:459477. 16. Fishm n EK, K vuru M, Jones B, et l. Pseudomembr nous colitis: CT ev lu tio n of 26 c ses. R diology 1991;180:5760. 17. Fekety R, Silv J, K ufm nn C, et l. Tre tment of ntibiotic- ssoci ted Clo stridium difficile colitis with or l v ncomycin: comp rison of two dos ge regime ns. Am J Med http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 19)03-M y-05 08:23:25

10. Lipsett PA, S m nt r y DK, T m ML, et l. Pseudomembr nous colitis: l dise se? Surgery 1994;116:491496.

surgic

Ovid: ICU Book 1989;86:1519. 18. Pe r SM, Willi mson TH, Bettin KM, et l. Decre se in nosocomi l Clostridium difficile ssoci ted di rrhe by restricting clind mycin use. Ann Intern Med 19 94;120:272277. 19. Sur wicz C. Prevention of ntibiotic- ssoci ted di rrhe by S cch romyces bo ul rdii: prospective study. G stroenterology 1989;96:981988. ABDOMINAL ABSCESSES 20. Mirvis SE, Sh nmug nth n K. Tr um r diology: p rt I. Computerized tomogr ph ic im ging of bdomin l tr um . J Intensive C re Med 1994;9:151163. 21. Fry DE. Noninv sive im ging tests in the disgnosis nd tre tment of intr - b domin l bscesses in the postoper tive p tient. Surg Clin North Am 1994;74:693709 . P.545 22. Oglevie SB, C sol G, v n Sonnenberg E, et l. Percut neous bscess dr in ge : current pplic tions for critic lly ill p tients. J Intensive C re Med 1994;9: 191206. 23. Mosdell DM, Morris DM, Voltur A, et l. Antibiotic tre tment for surgic l p eritonitis. Ann Surg 1991;214:543549. UROSEPSIS 24. St mm WE, Hooten TM. M n gement of urin ry tr ct infection in dults. N Engl J Med 1993;329:13281334.

26. How rd RJ. Host defense g inst infection-P rt 1. Curr Probl Surg 1980;27:26 7316. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (17 of 19)03-M y-05 08:23:25

25. Amin M. Antib cteri l prophyl xis in urology: ppl 4A):114 117.

review. Am J Med 1992; 92(Su

Ovid: ICU Book 27. D ifuku R, St mm WE. B cteri l dherence to bl dder uroepitheli l cells in c theter ssoci ted urin ry tr ct infection. N Engl J Med 1986;314:12081213.

28. J rvis JR, White JM, Munn VP, et l. Nosocomi l infections surveill nce, 198 3. MMWR 1985;33:14SS. 29. Emori TG, G ynes RP. An overview of nosocomi l infections, including the rol e of the microbiology l bor tory. Clin Microbiol Rev 1993;6:428442. 30. McCue JD. How to m n ge urin ry tr ct infections in the elderly. J Crit Illn ess 1996;11 (Suppl):S30S40. 31. Bone RC, L rson CB. Gr m-neg tive urin ry tr ct infections nd the developme nt of SIRS. J Crit Illness 1996;11(Suppl):S20S29. 32. B chm n JW, Heise RH, N essens JM, Timmerm n MG. A study of v rious tests to detect symptom tic urin ry tr ct infections in n obstetric popul tion. JAMA 1 993;270:1971 1974. 33. Cockerill FR, Edson RS. Trimethoprimsulf methox zole. M yo Clin Proc 1991;66: 1260 1269. 34. Jenkins SG. Ch nging spectrum of urop thogens: implic tions for tre ting com plic ted UTIs. J Crit Illness 1996;11(Suppl):S7S13. CANDIDURIA 35. Gubbins PO, Piscitelli SC, D nziger LH. C ndid l urin ry tr ct infections: comprehensive review of their di gnosis nd m n gement. Ph rm cother py 1993;13 :110127. 36. British Society for Antimicrobi l Chemother py Working P rty. M n gement of deep C ndid infection in surgic l nd intensive c re unit p tients. Intensive C re Med 1994;20:522528. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (18 of 19)03-M y-05 08:23:25

Ovid: ICU Book 37. Rex J, Bennett JE, Sug r AM. A r ndomized tri l comp ring flucon zole with mphotericin B for the tre tment of c ndidemi in p tients without neutropeni . N Engl J Med 1994;331:13251330. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (19 of 19)03-M y-05 08:23:25

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 34 THE IMMUNOCOMPROMISED PATIENT The c re of critic lly ill p tients is l bor-intensive, time-consuming, nd me nt lly exh usting experience, nd e ch of these spects of critic l c re re ches its zenith in the c re of the immunocompromised p tient. This ch pter presents some of the m jor concerns in the c re of three different groups of immunocompro mised p tients: p tients with hum n immunodeficiency virus (HIV) infection, neut ropenic p tients, nd org n tr nspl nt recipients. Ple se underst nd th t this i s very fluid re , nd the recommend tions in this ch pter could ch nge t ny m oment. INFECTION CONTROL The pr ctice of infection control involves number of prec ution ry me sures de signed to limit or prevent the tr nsmission of infectious gents between p tient s nd hospit l st ff. One of the most import nt go ls of infection control is to prevent the tr nsmission of the HIV to those involved in direct p tient c re (1 ,2). HIV TRANSMISSION Infection with HIV c n be tr nsmitted in the body fluids indic ted in T ble 34.1 (3). In the hospit l setting, tr nsmission of HIV occurs vi needlestick injuri es or by direct cont ct of mucous membr nes with blood from HIV-infected p tient s. Fortun tely, HIV tr nsmission to hospit l st ff is n uncommon event. As of 1 993, only 39 c ses of HIV tr nsmission ttributed to direct p tient c re cont ct h d been reported (2). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 20)03-M y-05 08:20:09

Ovid: ICU Book TABLE 34.1. THE RISK OF HIV TRANSMISSION IN BODY FLUIDS Needlestick Injuries A needlestick puncture tr nsfers n ver ge of 1 L of blood (4). During the virem ic st ges of HIV infection, there re s m ny s 5 P.547 infectious p rticles pe r L of blood (3). Therefore, direct cont ct with less th n one-thous ndth of mi lliliter of blood from n HIV-infected p tient c n result in tr nsmission of inf ectious vir l p rticles. In comp rison, 1 L of blood from p tients with vir l hep titis c n cont in s m ny s one million infectious p rticles (3), so the risk of tr nsmission of the hep titis virus vi needlestick injuries is f r gre ter t h n the risk of HIV tr nsmission. A single needlestick injury with blood from n HIV-infected p tient c rries 0.25% risk of HIV tr nsmission (2). This me ns t h t for every 10,000 needlestick injuries with HIVinfected blood, 25 c ses of HI V tr nsmission occur. There re n estim ted 16,000 needlestick exposures to HIV -infected blood e ch ye r in the United St tes (4), which tr nsl tes to 38 c ses of possible HIV tr nsmission nnu lly from ccident l needlestick punctures.

Mucous Membr ne Exposure The risk of HIV tr nsmission from direct exposure of broken skin or mucous membr nes to cont min ted body fluids is lower th n the risk of tr nsmission from nee dlestick injuries. A single exposure of broken skin or mucous membr nes to blood from n HIV-infected p tient c rries 0.09% risk of HIV tr nsmission (2). This me ns th t for every 10,000 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 20)03-M y-05 08:20:09

Ovid: ICU Book mucous membr ne exposures to cont min ted body fluids, 9 c ses of HIV tr nsmissi on occur. UNIVERSAL PRECAUTIONS The discovery th t HIV infection is tr nsmitted in blood nd body fluids prompte d the immedi te doption of blood nd body fluid prec utions s univers l pr c tice. The result nt guidelines re known s univers l prec utions. The b sis of this pr ctice is the ssumption th t blood nd body fluids from ll p tients re potenti lly infectious. Therefore, the summ ry of univers l prec utions describ ed below is me nt for ll p tients (5). P.548 Gloves Gloves should be worn for procedures th t involve direct cont ct with blood, muc ous membr nes, nd the following body fluids: mniotic fluid, peric rdi l fluid, peritone l fluid, pleur l fluid, synovi l fluid, semen, v gin l secretions, nd ny body fluid th t is cont min ted with blood. Gloves re lso recommended for the insertion of v scul r c theters nd for ll inv sive procedures. Gloves re not required (but re not discour ged) for routine phlebotomy procedures. H ndw shing H ndw shing is recommended fter ny cont ct with blood nd body fluids, even if gloves h ve been worn. A simple so p- nd-w ter scrubbing is sufficient, nd spe ci lized ntimicrobi l so ps re unnecess ry. Protective B rriers The mucous membr nes should be protected by b rriers such s m sks, gowns, f ci l shields, nd protective gl sses during ny procedure th t could gener te spr y of blood or body fluids.

Needlestick Prec utions An estim ted 800,000 needlestick injuries occur e ch ye r (2000 e ch d y) in the United St tes (4). To prevent these injuries, needles should not be rec pped by h nd or m nipul ted by h nd in ny w y (this includes the remov l of needles fr om the hub of syringe). Used needles should be left tt ched to the syringe n d disc rded in speci l puncture-resist nt sh rps boxes (which re in every room) . If used needle is rec pped, the one-h nded scoop http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 20)03-M y-05 08:20:09

Ovid: ICU Book technique is recommended. Using the h nd th t gr sps the syringe, the needle c p is first scooped up over the needle, nd then the tip of the needle c p is pres sed g inst fl t surf ce to secure the c p firmly to the hub of the needle. Th is elimin tes the risk of puncturing the other h nd with the needle. These needl estick prec utions re recommended for ll used needles, including those used to inject medic tions into intr venous infusion sets. Exud tive Derm titis Any member of the hospit l st ff who h s n exud tive or weeping derm titis should not p rticip te in direct p tient c re or h ndle p tient-c re equipment until t he problem resolves. Univers l prec utions re considered m nd tory in the c re of ll hospit lized p tients. Remember th t m ny of the pr ctices re designed t o protect the st ff s well s p tients. THE HIV-INFECTED PATIENT Most p tients with HIV infection who re dmitted to the ICU h ve severe pneumon i with cute respir tory f ilure. Fewer re dmitted P.549 with severe c ses of cryptococc l meningitis or toxopl smic enceph litis. The description th t follo ws is restricted to these problems. For inform tion on the m ny other disorders ssoci ted with HIV infection, see S nde nd Volberding (1995).

INCIDENCE OF HIV INFECTION HIV infection is worldwide epidemic. The dise ses directly rel ted to HIV infe ction (known collectively s cquired immunodeficiency syndrome, or AIDS) h ve b ecome more prev lent e ch ye r in the United St tes, s indic ted in T ble 34.2. (The l rge incre se in the number of reported c ses of AIDS in 1993 is c used b y the exp nded number of AIDS-defining dise ses proposed in th t ye r.) Close to three-qu rters of the reported c ses of AIDS in the United St tes h ve h d f t l outcome. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 20)03-M y-05 08:20:09

Ovid: ICU Book TABLE 34.2. THE AIDS EXPERIENCE IN THE UNITED STATES Cell-Medi ted Immunity HIV h s propensity for inv ding lymphocytes nd m croph ges. Infection with HI V is followed by long l tency period th t ver ges 10 to 11 ye rs (3). During this time, the virus multiplies in lymphoid tissues, which eventu lly results in depression of cell-medi ted immunity. When this occurs, there is n incre sed r isk of infection, p rticul rly with enc psul ted microorg nisms. These enc psul ted microbes include b cteri (Streptococcus pneumoni e, Hemophilus influenz ), fungi (Cryptococcus neoform ns), nd protozo ns (Pneumocystis c rinii, Toxopl sm gondii). The depression of cell-medi ted immunity in HIV infection is ccomp n ied by decre se in the density of cert in immunologic lly distinct lymphocytes in blood. One of these is the CD4 lymphocyte, which is decre sed below the norm l level of 800 to 1200 cells per microliter (or mm3) of blood. The number of ci rcul ting CD4 lymphocytes is therefore used s m rker of the severity of immun odepression in HIV-infected p tients (3). P.550 PNEUMONIA The most common c use of pneumoni in HIV-infected p tients h s been the protozo n org nism P. c rinii. However, recent evidence indic tes th t b cteri l pneumo ni m y be more common th n Pneumocystis pneumoni in this p tient popul tion (6 ,7). Enc psul ted b cteri such s S. pneumoni e (pneumococcus) nd H. influenz re common offenders, long with http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 20)03-M y-05 08:20:09

Ovid: ICU Book St phylococcus ureus (7). Other not ble c uses of lung infection in HIV-infecte d p tients re cytomeg lovirus (CMV), Mycob cterium tuberculosis, nd Mycob cter ium vium. Lung infections c used by fungi (e.g., Aspergillus) re usu lly seen only in the dv nced st ges of HIV infection (i.e., when CD4 lymphocyte counts f ll below 50/mm3) (3,6).

Clinic l Fe tures The clinic l present tion of pneumoni is nonspecific nd does not llow identif ic tion of the responsible p thogen (6). In p rticul r, the p ttern of infiltr t ion on the chest r diogr ph is not p thogen-specific. To illustr te this point, review the port ble chest r diogr phs shown in Figure 34.1 nd Figure 34.2. Both r diogr phs re from HIV-infected p tients who presented with fever, nonproduct ive cough, nd dyspne . Neither p tient could produce sputum s mple for n lys is, nd the results of blood cultures were either unreve ling or pending. C n yo u identify the p thogen b sed on the v il ble inform tion? http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 20)03-M y-05 08:20:09

Ovid: ICU Book Figure 34.1. Port ble chest r diogr ph from n HIV-infected p tient who presente d with fever, nonproductive cough, nd dyspne . Note the p tchy infiltr tes in b oth lungs. (Chest film courtesy of Dr. Rich rd K tz, M.D.) Figure 34.2. Port ble chest r diogr ph of n HIV-infected p tient who presented with fever, nonproductive cough, nd dyspne . Note infiltr te in left lower lung field. (Chest film courtesy of Dr. Rich rd K tz, M.D.) The clinic l present tion of the p tients just described is common scen rio. I n the bsence of sputum or pleur l fluid to ex mine, bronchoscopy is usu lly req uired to identify the responsible p thogen. Bronchoscopy Bronchoscopy provides the di gnosis in 85% of lung infections c used by P. c rin ii (the org nisms re pp rent on microscopy using speci lized st ins), nd in 9 5% of infections c used by Mycob cterium tuberculosis (6,8,9 nd 10) Bronchoscop y c n lso h ve high yield in the b cteri l pneumoni s, s described in Ch pte r 32. B cteri l cultures re performed qu ntit tively on specimens obt ined by s peci lized protected-brush devices or http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 20)03-M y-05 08:20:09

Ovid: ICU Book broncho lveol r l v ge. However, the threshold for identifying infection on qu n tit tive cultures m y be higher in HIV-infected p tients. The usu l threshold of 103 colony-forming units per milliliter (CFU/mL) does not lw ys indic te p ren chym l infection in p tients with HIV infection, nd higher threshold of 104 C FU/mL h s been recommended for the di gnosis of pneumoni in HIV-infected p tien ts (9). In the p tient with the r diogr ph shown in Figure 34.1, bronchoscopy re ve led numerous P. c rinii org nisms in broncho lveol r l v ge fluid. In the oth er p tient with the r diogr ph shown in Figure 34.2, bronchoscopy w s unreve lin g. P.551 Open-Lung Biopsy When one or more bronchoscopies do not provide the di gnosis for suspected pneum oni , open-lung biopsy is consider tion. However, lthough open-lung biopsy c n provide ddition l inform tion, this inform tion often does not le d to improv ed outcomes (8). In ddition, severely ill p tients with HIV infection m y do po orly fter open-lung biopsy (10). For these re sons, open-lung biopsy is not p opul r procedure in the ev lu tion of pneumoni in HIV-infected (or otherwise im munocompromised) p tients. In the p tient with the r diogr ph shown in Figure 34 .2 who h d n unreve ling bronchoscopy, open-lung biopsy w s not performed. Ther efore, the responsible p thogen w s never identified. (The p tient improved on e mpiric ntibiotic ther py.) This c se is included to emph size th t in most c se s of pneumoni in HIV-infected p tients, the responsible p thogen is never ident ified. In one study of pp rent pneumoni in HIV infection, p thogen w s isol ted in only 40% of the p tients (7). P.552 P. CARINII PNEUMONIA P. c rinii pneumoni (PCP) is the most common c use for dmission of HIV-infecte d p tients to the ICU (11). In p tients with PCP who re dmitted to the ICU, th e chest r diogr ph often looks like the one shown in Figure 34.3 (comp re this f ilm to the one in Fig. 34.1, which shows PCP in less dv nced st ge). The r di ogr phic ch nges in dv nced st ges of PCP re simil r to the ones seen in the cute respir tory distress syndrome (ARDS). In f ct, except for ntimicrobi l the r py, the m n gement of p tients with PCP nd cute respir tory f ilure is the s me s th t of p tients with ARDS. This m n gement str tegy is described in Ch p ter 23. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 20)03-M y-05 08:20:09

Ovid: ICU Book Figure 34.3. Port ble chest r diogr ph of p tient with severe P. c rinii pneum oni nd cute respir tory f ilure. Note the simil rity with the r diogr phic p pe r nce of ARDS. TrimethoprimSulf methox zole As shown in T ble 34.3, the initi l ntibiotic of choice for PCP is trimethoprim sulf methox zole (TMPSMX). The recommended dose is 20 mg/kg of TMP nd 100 mg/kg of SMX d ily, dministered P.553 P.554 in three or four divided doses. Although TMPSMX c n be given or lly, intr venous ther py is dvised for p tients with resp ir tory f ilure. A f vor ble response m y not be pp rent for 3 to 5 d ys, nd t here m y be n initi l period of deterior tion. If f vor ble response is not e vident fter 5 d ys of ther py, tre tment is considered f ilure. If evidence o f improvement exists in the first few d ys of ther py, tre tment is continued fo r tot l of 3 weeks (11). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 20)03-M y-05 08:20:09

Ovid: ICU Book TABLE 34.3. RECOMMENDED TREATMENT FOR OPPORTUNISTIC INFECTIONS Adverse re ctions to TMPSMX develop in 30 to 50% of HIV-infected p tients (11,12, 13 nd 14). These re ctions usu lly ppe r during the second week of tre tment, nd they re often severe enough to w rr nt discontinuing the drug. The most com mon side effects re neutropeni (45 to 50%), fever (45 to 50%), skin r sh (35 t o 40%), elev ted hep tic tr ns min se enzymes (30 to 35%), hyperk lemi (30%), nd thrombocytopeni (10 to 15%). A c se of f t l p ncre titis h s lso been link ed to TMPSMX (15). Only 35 to 45% of p tients who receive TMPSMX re ble to compl ete the full course of ther py. The high incidence of dverse re ctions to TMPSMX is specific for HIV infection. In other groups of p tients, dverse re ctions t o TMPSMX develop in only 10% of the p tients (11).

Pent midine When ther py with TMPSMX f ils or is not toler ted, the preferred second-line ge nt is pent midine isethion te. The recommended dose is 4 mg/kg s single d ily dose. Pent midine c uses sterile bscesses when given by intr muscul r injectio n, so intr venous ther py is preferred. The response time nd dur tion of ther p y re the s me s for TMP SMX. Tre tment f ilures occur in one-third of p tients (11). Adverse re ctions re lso common with intr venous pent midine (11,16,17 nd 18). These side effects include neutropeni (5 to 30%), hyperglycemi nd hyp oglycemi (10 to 30%), prolonged QT interv l (3 to 35%), tors de de pointes (up t o 20%), ren l insufficiency (3 to 5%), nd p ncre titis (up to 1%). Of ll p tie nts who receive intr venous pent midine, 40 to 50% re un ble to complete ther p y bec use of dverse re ctions (11,16). In p tients who re un ble to complete t her py with either TMPSMX or pent midine, v riety http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 20)03-M y-05 08:20:09

Ovid: ICU Book of other gents (e.g., d psone nd trimetrex te) re v il ble. In this situ tio n, you should consult n infectious dise se expert pronto. Steroids Steroid ther py st rted t the time of ntimicrobi l ther py h s been reported t o improve outcomes in p tients with PCP (19). Most tri ls h ve involved or l pre dnisone, but intr venous methylprednisolone (40 mg every 6 hours for t le st 7 d ys) h s lso been recommended (11). Del y of tre tment for 72 hours fter the st rt of ntimicrobi l ther py neg tes ny possible benefit from steroids (19). The response to steroids seems to v ry in different clinic l reports, nd f vor ble responses c n be short-lived (20). P.555 CRYTOCOCCAL MENINGITIS Cryptococc l meningitis is the most common life-thre tening fung l infection in HIV-infected p tients (21,22). It is expected in 10% of p tients with HIV infect ion, nd usu lly ppe rs in the dv nced st ges of immunosuppression (i.e., when CD4 lymphocyte counts f ll below 50/ mm3). Clinic l Fe tures The most common m nifest tions re fever nd he d che, e ch reported in pproxim tely 85% of c ses (21). Other findings include meninge l signs (35 to 40%), lt ered ment l st tus (10 to 15%), nd seizures (less th n 10%) (21). Cryptococc l infections t other sites (e.g., pneumoni nd skin r sh) re seen in 20% of c s es (22). Di gnosis Lumb r puncture is required for the di gnosis. St nd rd me surements in cerebros pin l fluid (CSF), such s glucose, protein, nd leukocyte count, c n be norm l in up to 50% of c ses (21). The org nism c n be demonstr ted on indi ink st ins of CSF in 75% of c ses (which is higher th n the yield from indi ink st ins in nonHIV-infected p tients) (21). CSF cultures nd cryptococc l ntigen titers re positive in over 90% of c ses (21). Tre tment The st nd rd tre tment of cryptococc l meningitis is intr venous mphotericin (T ble 34.3). The dministr tion of this ntifung l gent is described in Ch pter 35 (T ble 35.3). Ther py http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 20)03-M y-05 08:20:09

Ovid: ICU Book with flucon zole (200 to 400 mg d ily) h s been reported to be s effective s mphotericin (23), but more experience is needed before flucon zole c n be recomm ended s the preferred (i.e., less toxic) first-line gent for cryptococc l meni ngitis. (See Ch pter 35 for more inform tion on flucon zole.) Despite ntifung l ther py, pproxim tely one-third of HIVinfected p tients with cryptococc l meni ngitis succumb to the infection (21). TOXOPLASMIC ENCEPHALITIS T. gondii enceph litis is the most common neurologic disorder in HIV-infected p tients. Clinic l evidence of toxopl smic enceph litis is reported in 5 to 15% of HIV-infected p tients, nd utopsy evidence of the dise se is present in up to 30% of p tients (21). Clinic l Fe tures Toxopl smic enceph litis is ch r cterized by foc l br in lesions. Hemip resis n d other foc l neurologic deficits re seen in 60% of P.556 c ses, nd seizures re reported in 15 to 30% of p tients (21). Other m nifest tions include fever (5 to 55%), confusion (60 to 65%), nd choreiform movements (considered by some to be p thognomonic of toxopl smic enceph litis) (21). Although extr neur l dise s e is not common, dissemin ted toxopl smosis with septic shock h s been reported (24). Di gnosis Computerized tomogr phy (CT) usu lly reve ls solit ry or multiple hypodense, con tr stenh ncing m ss lesions in the b s l g ngli nd frontop riet l regions of t he cerebr l hemispheres (21). M gnetic reson nce im ging (MRI) is more sensitive nd c n reve l lesions when CT sc ns re neg tive (25). Lumb r puncture usu lly reve ls bnorm l findings, but these re nonspecific. Definitive di gnosis requ ires excision l br in biopsy (needle biopsy specimens re often unreve ling). Th e org nism is demonstr ted by immunoperoxid se st ining. Tre tment The tre tment for toxopl smic enceph litis is shown in T ble 34.3. The preferred regimen is combin tion of pyrimeth mine (200 mg lo ding dose, then 75 mg d il y) nd clind mycin (600 mg every 6 hours). Bec use pyrimeth mine is fol te nt gonist, folinic cid (10 mg) is given with e ch dose of pyrimeth mine to reduce the incidence of bone m rrow suppression. All gents re given or lly. Approxim tely 70% of c ses show f vor ble response to this http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 20)03-M y-05 08:20:09

Ovid: ICU Book regimen, nd improvement is usu lly evident within the first week of ther py (26 ). The condition is considered uniformly f t l without ppropri te ther py. THE NEUTROPENIC PATIENT Unlike the depression in cell-medi ted immunity th t occurs in HIV infections, i sol ted neutropeni (neutrophil count less th n 500/mm3) results in depression of humor l immunity. This results in n incre sed risk of infections with b cte ri l r ther th n opportunistic p thogens. The princip l concern in the neutropen ic p tient is Gr m-neg tive b cteremi , which c n be ccomp nied by r pid nd f t l downhill course. Fortun tely, b cteremi is uncommon in neutropenic p tien ts unless the neutrophil count f lls below 100/mm3 (27). In f ct, most neutropen ic p tients with new-onset fever do not h ve documented infection (28). Theref ore, the rush to tre tment in this situ tion is prob bly overkill in m ny inst n ces. EMPIRIC ANTIBIOTICS A number of empiric ntibiotic regimens h ve proven successful in febrile neutro peni , nd these re shown in T ble 34.4 (27,28,29,30,31 nd 32). No one regimen h s proven to be superior to the others. TABLE 34.4. EMPIRIC ANTIBIOTIC THERAPY FOR FEBRILE NEUTROPENIA

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 20)03-M y-05 08:20:09

P.557 Monother py with ceft zidime is ored regimen t the

popul r regimen

nd is currently the f v

Ovid: ICU Book N tion l C ncer Institute (30). Monother py with imipenem is equ lly effective, but the popul rity of this regimen h s been hindered by the risk for seizures wi th imipenem (31), which is lower th n perceived. The combin tion regimens re us ed when Gr m-neg tive septicemi , p rticul rly with Pseudomon s org nisms, is su spected. The once popul r minoglycoside regimen is no longer f vored bec use of the potenti l for toxicity with minoglycosides. Fin lly, v ncomycin c n be dd ed if Gr m-positive p thogen is suspected (e.g., c theter-rel ted sepsis) or i f methicillin-resist nt st phylococci re prev lent t your hospit l. Pulmon ry Infiltr tes If the fever is ccomp nied by pulmon ry infiltr tes (p rticul rly diffuse infil tr tes), the p thogens to consider re Legionell pneumophil , Mycopl sm pneumo ni e, Pneumocystis, nd CMV. In this setting, empiric ther py with erythromycin nd TPMSMX h s been recommended, pending bronchoscopy (8). Bec use only few dos es of ntibiotics c n reduce the di gnostic yield from bronchoscopic cultures (s ee Ch pter 32), the bronchoscopy should be performed s soon s possible fter ntibiotic ther py is st rted.

Persistent Fever Continued fever fter 1 week of empiric ntibiotic ther py suggests possible inv sive fung l dise se (p rticul rly dissemin ted c ndidi sis). In this situ tion, intr venous mphotericin (0.5 mg/kg d ily) is recommended (8). As mentioned in the Ch pter 33, blood cultures re neg tive in more th n 50% of c ses of dissemi n ted c ndidi sis. The di gnosis of dissemin ted c ndidi sis is suggested by iso l ting C ndid P.558 t three sep r te sites (e.g., urine, sputum, nd v scul r c theter tips), nd the di gnosis is confirmed by evidence of C ndid ophth lmit is (see Reference 36 in Ch pter 33). THE TRANSPLANT RECIPIENT A number of org n tr nspl nts c n be encountered in the ICU, including of kidney , he rt, liver, nd bone m rrow tr nspl nts. Infections in the first few weeks f ollowing org n tr nspl nt tion re usu lly c used by b cteri l p thogens (e.g., S. ureus), but infections th t ppe r l ter (1 to 3 months) re often c used by opportunistic p thogens such s CMV, C ndid , or other fungi (8,33,34). The ch nge in microbi l spectrum is expl ined by the chemother py used to prevent org n rejection. CYTOMEGALOVIRUS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 20)03-M y-05 08:20:09

Ovid: ICU Book CMV is herpesvirus th t c n c use number of different infections in immunoco mpromised p tients, including pneumoni , retinitis, hep titis, p ncre titis, nd esoph ge l ulcers. The virus is most p thogenic in tr nspl nt recipients nd HI V-infected p tients. Tr nspl nt recipients who h ve no CMV ntibodies before sur gery re most likely to develop CMV infections fter tr nspl nt tion. Di gnosis The di gnosis of CMV infection requires histologic evidence of cell inv sion by the virus. Cultures of CMV in blood nd urine re considered unreli ble (35). Th e cytop thic ction of CMV produces ch r cteristic cell with intr nucle r incl usion bodies (36). The presence of this cell in tissue s mples or bronchoscopy s pecimens is considered p thognomonic of CMV infection. Cytop thic ch nges in ir w y epitheli l cells (from bronchoscopy) indic tes CMV infection in the irw ys (tr cheobronchitis). The di gnosis of CMV pneumoni requires cytop thic ch nges in lveol r m croph ges or lung biopsies. CMV pneumoni is ccomp nied by diffus e infiltr tes in both lungs nd c n be confused with severe Pneumocystis pneumon i ( s shown in Fig. 34.3). In f ct, prroxim tely 50% of HIVinfected p tients w ith P. c rinii pneumoni lso h ve evidence of CMV infection (37). However, this is believed to represent coloniz tion r ther th n infection in most c ses. The situ tion is different in tr nspl nt recipients, in whom CMV is considered to be common c use of pneumoni (38). Tre tment The preferred gent for tre ting CMV infections is intr venous g nciclovir (T bl e 34.3). The recommended dose is 5 mg/kg every 8 to 12 hours (36). Lower doses ( 5 mg/kg/d y) re used to prevent CMV P.559 infection in tr nspl nt recipients (3 9). Adverse re ctions to g nciclovir, including neutropeni , thrombocytopeni , nd skin r sh, develop in 10% of p tients. In CMV infections th t re refr ctory to g nciclovir, the ntivir l gent fosc rnet c n be effective (40). The recomme nded dose is 200 mg/kg/d y by continuous intr venous infusion. Adverse re ctions to fosc rnet develop in 20% of p tients nd include hypom gnesemi , hypoc lcemi , hypophosph temi , nemi , nd ren l insufficiency (36,40). Bec use of the gre ter incidence of dverse re ctions, fosc rnet is considered second-line gent for the tre tment of CMV infections. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 20)03-M y-05 08:20:09

Ovid: ICU Book REFERENCES SUGGESTED READING Greenb um DM, ed. M n gement of the AIDS p tient in the ICU. Critic l c re clini cs. Vol. 9. Phil delphi : WB S unders, 1993. S nde MA, Volberding PA, eds. The medic l m n gement of AIDS. 4th ed. Phil delph i : WB S unders, 1995. INFECTION CONTROL 1. Americ n College of Physici ns nd Infectious Dise se Society of Americ . Hum n immunodeficiency virus (HIV) infection. Ann Intern Med 1994;120:310319.

2. Geberding JL. Limiting the risks of he lth c re workers. In: S nde MA, Volber ding PA, eds. The medic l m n gement of AIDS. 4th ed. Phil delphi : WB S unders, 1995;89 101.

4. Berry AJ, Greene ES. The risk of needlestick injuries nd needlestick-tr nsmi tted dise ses in the pr ctice of nesthesiology. Anesthesiology 1992;77:100710021 . 5. G rner JS. Univers l prec utions nd isol tion systems. In: Bennett JV, Br ch m n PS, eds. Hospit l infections. 3rd ed. Boston: Little, Brown, 1992;231244. THE HIV-INFECTED PATIENT 6. Rosen MJ. Pneumoni in p tients with HIV infection. Med Clin North Am 1994;78 :10671079. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 20)03-M y-05 08:20:09

3. Levy JA. The tr nsmission of HIV Am J Med 1993;95:86100.

nd f ctors influencing progression to AIDS.

Ovid: ICU Book 7. Hirschtick RE, Gl ssroth J, Jord n MC, et l. B cteri l pneumoni in persons infected with the hum n immunodeficiency virus. N Engl J Med 1995;333:845851. 8. Shelh mmer JH, Toews GB, M sur H, et l. Respir tory dise se in the immunosup pressed p tient. Ann Intern Med 1992;117:415431. 9. Ferrer M, Torres A, X ubet A, et l. Di gnostic v lue of telescoping plugged c theters in HIV-infected p tients with pulmon ry infiltr tes. Chest 1992;102:768 3. P.560 10. Tr chiotis GD, H fner GH, Hix WR, A ron BL. Role of open lung biopsy in di g nosing pulmon ry complic tions of AIDS. Ann Thor c Surg 1992;54:898902. PNEUMOCYSTIS CARINII PNEUMONIA 11. Brooks KR, Ong R, Spector RS, Greenb um DM. Acute respir tory f ilure due to Pneumocystis c rinii pneumoni . Crit C re Clin 1993;9:3148. 12. Johnson MP, Goodwin D, Sh nds JW. Trimethoprim-sulf methox zole n phyl ctoi d re ctions in p tients with AIDS: c se reports nd liter ture review. Ph rm cot her py 1990;10:423426. 13. v n der Ven AJAM, Koopm ns PP, Vree TB, v n der Meer JWM. Adverse re ctions to co-trimox zole in HIV infection. L ncet 1991;338:431433. 14. Greenberg S, Reiser IW, Chou S-Y, Porush JG. Trimethoprim-sulf methox zole i nduces reversible hyperk lemi . Ann Intern Med 1993;119:291295. 15. Jost R, Stey C, S lomon F. F t l drug-induced p ncre titis in HIV. L ncet 19 93;341:1412. 16. Dohn MN, Weinberg WG, Torres RA, et l. Or l tov quone comp red with http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (17 of 20)03-M y-05 08:20:09

Ovid: ICU Book intr venous pent midine for Pneumocystis c rinii pneumoni in p tients with AIDS . Ann Intern Med 1994;121:174180. 17. Eisenh uer MD, Eli sson AH, T ylor AJ, et l. Incidence of c rdi c rrhythmi s during intr venous pent midine ther py in HIV-infected p tients. Chest 1994;1 05:389 394. 18. Foisey MM, Sl yter KL, Morse GD. P ncre titis during intr venous pent midine ther py in n AIDS p tient with prior exposure to did nosine. Ann Ph rm cother 1994;28:10251028. 19. N tion l Institutes of He lthUniversity of C liforni Expert P nel for Cortic osteroids s Adjunctive Ther py for Pneumocystis Pneumoni . Consensus st tement on the use of corticosteroids s djunctive ther py for Pneumocystis pneumoni i n the cquired immunodeficiency syndrome. N Engl J Med 1990;323:15001504.

CRYPTOCOCCAL MENINGITIS 21. Levy RM, Berger JR. Neurologic critic l c re in p tients with hum n immunode ficiency virus 1 infection. Clin Crit C re 1993;9:4972. 22. Ennis DM, S g MS. Cryptococc l meningitis in AIDS. Hosp Pr ct 1993;28:99112. 23. S g MS, Powderly WG, Cloud GA, et l. Comp rison of mphotericin B with flu con zole in the tre tment of cute AIDS- ssoci ted cryptococc l meningitis. N En gl J Med 1992;326:8389. TOXOPLASMIC ENCEPHALITIS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (18 of 20)03-M y-05 08:20:09

20. Schiff MJ, F rber BF, K pl n MH. Steroids for Pneumocystis c rinii pneumoni nd respir tory f ilure in the cquired immunodeficiency syndrome. Arch Intern Med 1990;150:18191821.

Ovid: ICU Book 24. Lucet J-C, B illey M-P, Bedos J-P, et l. Septic shock due to toxopl smosis in p tients infected with the hum n immunodeficiency virus. Chest 1993;104:105410 58. 25. Knobel H, Gr us F, Miro JM, et l. Toxopl smic enceph litis with norm l CT s c n nd p thologic MRI. Am J Med 1995;99:220221. P.561 26. Luft BJ, H fner R, Korzun AH, et l. Toxopl smic enceph litis in p tients wi th the cquired immunodeficiency syndrome. N Engl J Med 1993;329:9951000. THE NEUROPENIC PATIENT 27. Shenep JL. Empiric ntimicrobi l tre tment in febrile neutropenic c ncer p t ients. Infect Med 1992;April:3947. 28. R nph l R, Guc lp R, Rotstein C, et l. Clinic l experience with single gen t nd combin tion regimens in the m n gement of infection in the febrile neutrop enic p tient. Am J Med 1996;100:83S89S. 29. Hughes WT, Armstrong D, Bodey GP, et l. Guidelines for the use of ntimicro bi l gents in neutropenic p tients with unexpl ined fever. J Infect Dis 1990;16 1:381396. 30. Pizzo PA. Choosing empiric ther py for febrile neutropenic p tients. J Crit Illness 1995;10:165168. 31. Winston DJ, Ho WG, Bruckner DA, Ch mplin RE. Bet -l ct m ntibiotic ther py in febrile gr nulocytopenic p tients. Ann Intern Med 1991;115:849859. 32. De P uw BE, Deresinski SC, Feld R, et l. for the Intercontinent l Antimicro bi l Study Group. Ceft zidime comp red with piper cillin nd tobr mycin for the empiric tre tment of fever in neutropenic p tients with c ncer. A multicenter r ndomized tri l. Ann Intern Med 1994;120:833844. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (19 of 20)03-M y-05 08:20:09

Ovid: ICU Book THE TRANSPLANT RECIPIENT 33. Kob sig w JA, Stevenson LW. M n ging complic tions in he rt tr nspl nt reci pients. J Crit Illness 1993;8:678689. 34. How rd RJ. Infections in the immunocompromised p tient. Surg Clin North Am 1 994;74:609620. CYTOMEGALOVIRUS INFECTIONS 35. Zurlo JJ, O'Neill D, Polis M, et l. L ck of clinic l utility of cytomeg lov irus blood nd urine cultures in p tients with HIV infection. Ann Intern Med 199 3;118:1217. 36. Goodg me RW. G strointestin l cytomeg lovirus dise se. Ann Intern Med 1993;1 19:924935. 37. J cobson MA, Mills J, Rush J, et l. Morbidity nd mort lity of p tients wit h AIDS nd first-episode Pneumocystis c rinii pneumoni un ffected by concomit n t pulmon ry cytomeg lovirus infection. Am Rev Respir Dis 1991;144:69. 38. Sommer SE, Em nuel D, Groeger J, et l. Successful m n gement of CMV pneumon i in mech nic lly ventil ted p tient. Chest 1991;100:856858. 39. Goodrich JM, Bowden RA, Fisher L, et l. G nciclovir prophyl xis to prevent cytomeg lovirus dise se fter llogenic m rrow tr nspl nt. Ann Intern Med 1993;1 18:173178. 40. Lietm n PS. Clinic l ph rm cology: fosc rnet. Am J Med 1992;92(Suppl):8S11S. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (20 of 20)03-M y-05 08:20:09

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 35 ANTIMICROBIAL THERAPY The d nger with germ-killing drugs is th t they m y kill the p tient s well s the germ. --JBS H ld ne Antimicrobi l ther py is p rt of everyd y life in the IC U. The gents listed here cover most of the infections encountered in the ICU, nd e ch is presented briefly in the order listed. At the end of the ch pter is t ble listing the estim ted d ily cost of ther py with these gents (see T ble 35.8). TABLE 35.8. COST OF ANTIMICROBIAL THERAPY Aminoglycosides Antifung l gents ( mphotericin, flucon zole) Aztreon m Ceph los porins http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 17)03-M y-05 08:21:34

Ovid: ICU Book Imipenem Penicillins Quinolones V ncomycin A list of preferred ntibiotics for s pecific p thogens is shown in T ble 35.1 TABLE 35.1. PARENTERAL ANTIBIOTIC SELECTIONS AMINOGLYCOSIDES The minoglycosides (gent micin, tobr mycin, nd mik cin) re the tr dition l gents used for serious Gr mneg tive infections (1). However bec use of the risk of nephrotoxicity, there h s been m rked decline in minoglycoside use in rece nt ye rs. At one l rge te ching hospit l, the use of gent micin decre sed 70% in the l te 1980s (1). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 17)03-M y-05 08:21:34

Ovid: ICU Book ANTIBACTERIAL SPECTRUM The minoglycosides re ctive g inst ll Enterob cteri ce e, including Pseudom on s eruginos . There is no emerging resist nce to the minoglycosides, P.563 w hich is one of the re sons for their continued popul rity. DOSING Aminoglycoside dosing in critic lly ill p tients is summ rized in T ble 35.2 (2) . Drug doses re determined ccording to ide l body P.564 weight, except in p ti ents who re morbidly obese; then the ppropri te dosing weight is the ide l bod y weight plus h lf the difference between the ide l nd ctu l body weight (3). The initi l lo ding doses shown in T ble 35.2 re l rger th n those recommended norm lly bec use of the l rger volume of distribution for minoglycosides in cri tic lly ill p tients (4). The d ily m inten nce dose is tr dition lly given in t hree divided doses, but c n lso be given once d ily. Bec use minoglycosides r e elimin ted by the kidneys, dose djustments re required when ren l function i s imp ired. There re two methods for djusting the dose when ren l function is imp ired: TABLE 35.2. AMINOGLYCOSIDE THERAPY IN THE ICU The dose interv l c n be djusted by multiplying the norm l dose interv l by the serum cre tinine (in mg/dL). The d ily dose c n be djusted by dividing the nor m l d ily dose by the serum cre tinine (in mg/dL). Once-D ily Dosing The b ctericid l effect of the minoglycosides is concentr tion-dependent, so hi gher drug concentr tions in the blood nd tissues produce more b cteri l killing (5). This h s led to the recommend tion th t minoglycosides be given in one d ily dose, which produces higher drug concentr tions in the tissues th n the divi ded-dose regimen. The once-d ily regimen h s proven equiv lent in effic cy nd t oxicity to the divided-dose regimen (2,6), while http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 17)03-M y-05 08:21:34

Ovid: ICU Book being less costly (bec use it elimin tes the cost of extr intr venous infusion sets nd reduces the need for monitoring serum drug levels). However, despite th e potenti l dv nt ges, once-d ily dosing h s been dopted in only 15 to 20% of hospit ls in the United St tes (6). Monitoring Routine monitoring of serum minoglycoside levels is recommended for three re so ns (1,2). First, pe k serum levels re directly rel ted to clinic l effic cy. Se cond, minoglycoside toxicity shows correl tion with the minim l (trough) seru m levels t the end of the dosing interv l. Third, equiv lent doses of minoglyc osides result in v ri ble serum drug levels in individu l p tients. The recommen ded serum drug levels re shown in T ble 35.2. The pe k levels in this t ble h v e been shown to result in improved outcomes (2), where s the trough levels re ssoci ted with reduced clinic l toxicity. There is one report of spurious elev t ions in serum minoglycoside levels when blood s mples were dr wn through centr l venous Sil stic c theters (mech nism unknown). TOXICITY Nephrotoxicity Approxim tely 20% of p tients who receive minoglycosides develop some degree of ren l imp irment, reg rdless of the gent used (7). P.565 The onset is usu lly 3 to 7 d ys fter the onset of ther py, nd e rly signs include cylindric l c st s in the urine, proteinuri , nd in bility to concentr te urine. This is followe d by rising serum cre tinine concentr tion th t culmin tes in cute ren l f il ure. Nephrotoxicity is enh nced by hypovolemi , dv nced ge, preexisting ren l imp irment, hypok lemi , nd hypom gnesemi (7). The cute ren l f ilure is usu lly nonoliguric nd reversible. Ototoxicity Aminoglycosides produce dose-rel ted nd irreversible he ring loss for high-fr equency sounds. Bec use the he ring loss occurs bove the frequency r nge of nor m l hum n convers tion, udiometry is required to document the ototoxicity (1). Neuromuscul r Block de In high doses, the minoglycosides block cetylcholine rele se t presyn ptic ne rve termin ls nd reduce postsyn ptic responsiveness to cetylcholine. These ge nts c n ggr v te the neuromuscul r block de ssoci ted with my stheni gr vis nd nondepol rizing muscle rel x nts (1,8,9). However, little evidence exists of this effect when the drugs re given in ther peutic doses (10). COMMENT The risk of minoglycoside nephrotoxicity is re son to void these gents whenev er possible. A number of less toxic gents re now v il ble for tre ting Gr m-n eg tive infections, including the third-gener tion ceph losporins, ztreon m, n d imipenem. Ph rm ceutic l comp nies pp rently re lize the ble k future for mi noglycosides bec use no new minoglycoside gents h ve been introduced in over dec de. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 17)03-M y-05 08:21:34

Ovid: ICU Book

Infusion-Rel ted Side Effects Amphotericin infusions re ccomp nied by fever, chills, n use , vomiting, nd r igors in pproxim tely 70% of c ses (12). Premedic tion with cet minophen nd d iphenhydr mine, given 30 minutes before the infusions, c n reduce the incidence of infusion-rel ted side P.566 effects (T ble 35.3). Infusion-rel ted rigors c n be tre ted by premedic tion with meperidine (Demerol). An infusion-rel ted phle bitis c n ppe r d ys fter mphotericin ther py is st rted, nd it is p rticul rly prominent when the drug is given through peripher l vein. Hep rin is often dded to the infus te (1 U/mL) to prevent this complic tion, but no evidence ex ists th t this is effective (11). TABLE 35.3. GUIDELINES FOR AMPHOTERICIN ADMINISTRATION Dosing Some recommend tions for mphotericin dministr tion re shown in T ble 35.3 (12 ). A test dose is commonly given t the st rt of ther py, but this is unnecess r y if premedic tions re given routinely. If fever persists despite http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 17)03-M y-05 08:21:34

ANTIFUNGAL AGENTS AMPHOTERICIN Amphotericin is the most effective r, ther py with this gent c rries prec utions re necess ry.

ntifung l gent in clinic l use (11). Howeve consider ble risk of toxicity, nd sever l

Ovid: ICU Book premedic tions, hydrocortisone c n be dded to the infus te. The drug is given o nce d ily, st rting t dose of 0.25 mg/kg nd incre sing to fin l d ily dose of 0.5 to 1.0 mg/kg (higher d ily doses re used for more severe infections). T he initi l dur tion of drug infusion is 4 hours, nd this c n be reduced to 1hour infusion fter 5 d ys of ther py if ren l imp irment or hyperk lemi does n ot occur. D ily infusions re continued until the cumul tive dose re ches spec ified level. The tot l mphotericin dose is dict ted by the severity of the infe ction nd c n be s little s 500 mg (for c theter-rel ted c ndidemi ) or s muc h s 4 g (for life-thre tening fung l infections). Nephrotoxicity The m jor complic tion of mphotericin ther py is imp ired ren l function. Nephr otoxic effects include ren l v soconstriction nd P.567 ren l tubul r cidosis (dist l type) with incre sed urin ry excretion of pot ssium nd m gnesium (13). The l tter complic tion le ds to pot ssium nd m gnesium depletion. The inciden ce of nephrotoxicity is uncle r bec use mphotericin is often used in conditions th t themselves c use ren l imp irment. Preventive m n gement of mphotericin n ephrotoxicity includes routine monitoring of serum cre tinine, pot ssium, nd m gnesium. If the serum cre tinine rises bove 3.0 mg/dL, the mphotericin infusio n should be tempor rily discontinued (T ble 35.3) D ily m gnesium supplements (3 00 to 600 mg element l m gnesium d ily) should be given to counterb l nce urin r y losses. Serum m gnesium is not sensitive m rker of m gnesium depletion (see Ch pter 42), so d ily m gnesium supplement tion is import nt. Sodium lo ding (i. e., infusion of 1 L of isotonic s line) h s been recommended before mphotericin infusions to m int in ren l blood flow (13). However, sodium lo ding does not p revent mphotericin-induced ren l f ilure, nd it ggr v tes the ren l tubul r d efect nd promotes urin ry pot ssium w sting (13). Thus, sodium lo ding should b e viewed with c ution. Wh t m y be more import nt is to void hypovolemi during mphotericin ther py. FLUCONAZOLE Flucon zole (Difluc n) is member of the imid zole cl ss of ntifung l gents ( the other members re itr con zole nd ketocon zole) nd w s introduced for clin ic l use in 1990. The benefici l fe tures of flucon zole include bro d spectru m of ntifung l ctivity nd fewer toxic side effects th n mphotericin. Bec use of these f vor ble fe tures, flucon zole could become s fer ltern tive to m photericin for the tre tment of severe fung l infections (14). Indic tions Flucon zole h s proven effective in tre ting infections c used by C ndid nd Cr yptococcus org nisms in p tients who re immunocompromised or h ve n int ct imm une system (15,16). It h s lso proven successful for preventing fung l infectio ns in immunocompromised p tients (17). In the ICU, the m jor use of flucon zole is in p tients with documented or suspected dissemin ted c ndidi sis. In p tient s who re not immunocompromised, flucon zole (400 mg d ily) c n be used s n l tern tive to mphotericin (16). In those who re immunocompromised, combin tion ther py with flucon zole nd mphotericin might be the wiser choice. Dosing http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 17)03-M y-05 08:21:34

Ovid: ICU Book Flucon zole is given once d ily. The dose for serious infections is 400 mg/d y. In p tients with ren l f ilure, the d ily dose should be P.568 reduced by 50% (1 4). Serum drug levels c n be used to guide drug dosing in ren l f ilure. Trough levels should be m int ined between 6 nd 20 g/mL (14). Drug Inter ctions Flucon zole c n interfere with the met bolism of phenytoin nd w rf rin nd pote nti te the ctions of both drugs. These inter ctions c n be clinic lly signific nt, so it is import nt to monitor serum phenytoin levels nd prothrombin times w hen flucon zole is given in concert with these drugs (18,19). Hep totoxicity Severe nd even f t l hep tic injury ssoci ted with flucon zole ther py h s bee n reported (20). Most c ses h ve been in p tients with liver dise se or hum n im munodeficiency virus (HIV) infection. Although this seems to be r re event, li ver enzymes prob bly should be monitored periodic lly during flucon zole ther py in p tients who re HIV-positive or h ve liver dise se (20).

DOSING The usu l intr venous dose of ztreon m in dults is 1 g every 8 hours (21). Dou bling the dose h s been recommended for seriously ill p tients, but no evidence exists th t this improves effic cy. A dose reduction of 50 to 75% is recommended in p tients with ren l f ilure (22). TOXICITY Aztreon m is rel tively s fe drug. In survey of 2700 p tients receiving ztr eon m, dverse re ctions were noted in 7% of p tients. Most of these re ctions w ere mild nd nonspecific (e.g., n use nd di rrhe ) (21). COMMENT Aztreon m provides s fe nd effective ltern tive to the minoglycosides for t he m n gement of serious Gr mneg tive infections. P.569 CEPHALOSPORINS A sm ll rmy of ceph losporins is v il ble for clinic l use. These gents re d ivided into gener tions b sed on the historic l sequence of their introduction i nto clinic l pr ctice. Some of the p renter l gents in e ch gener tion http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 17)03-M y-05 08:21:34

AZTREONAM Aztreon m (Az ct m) is synthetic ntimicrobi l gent th t h s simil r ntib cteri l spectrum to the minoglycosides, but is devoid of nephrotoxicity (21).

Ovid: ICU Book re shown in T ble 35.4 (23).

TABLE 35.4. THE GENERATIONS OF PARENTERAL CEPHALOSPORINS THE FAMILY OF CEPHALOSPORINS The first-gener tion ceph losporins re prim rily ctive g inst erobic Gr m-po sitive cocci, but re not ctive g inst St phylococcus epidermidis or methicill in-resist nt str ins of S. ureus. The popul r intr venous gent in this group i s cef zolin (Ancef). The second-gener tion ceph losporins exhibit stronger ntib cteri l ctivity g inst Gr m-neg tive erobic nd n erobic b cilli of enteric origin. The popul r p renter l gents in this group re cefoxitin (Mefoxin) nd cef m ndole (M ndol). The third-gener tion ceph losporins h ve gre ter ntib ct eri l ctivity g inst Gr m-neg tive erobic b cilli, including P. eruginos n d Hemophilus influenz , but re less ctive g inst erobic Gr m-positive cocci th n the first-gener tion gents. The popul r p renter l gents in this group r e cefot xime (Cl for n), ceftri xone (Rocephin), nd ceft zidime (Fort z). The l tter gent is not ble for its ctivity g inst P. eruginos . The fourth-gener tion ceph losporins re just beginning to emerge. One of the gents in this grou p is cefepime (M xipime), which h s the Gr m-neg tive ntib cteri l spectrum of ceft zidime (i.e., it covers P. eruginos ), but is lso ctive g inst erobic Gr m-positive cocci (eg. streptococci nd methicillin-sensitive st phylococci) ( 24). DOSING The doses for some p renter l ceph losporins re shown in T ble 35.5. Most gent s re given in doses of 1 to 2 g every 6 to 8 hours. P.570 Ceftri xone is the ex ception, nd c n be given every 12 to 24 hours. The dose djustments in ren l f ilure re lso included in T ble 33.5 (22). Note th t the dose is djusted by ex tending the dosing interv l r ther th n decre sing the mount of drug given with e ch dose. This is done to preserve concentr tion-dependent b cteri l killing. Ceftri xone requires no dose djustment in ren l f ilure. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 17)03-M y-05 08:21:34

Ovid: ICU Book TABLE 35.5. PARENTERAL CEPHALOSPORIN DOSING TOXICITY Adverse re ctions to ceph losporins re uncommon nd nonspecific (e.g., n use , r sh, nd di rrhe ). There is 5 to 15% incidence of cross- ntigenicity with pe nicillin (23), nd ceph losporins should be voided in p tients with prior n phyl ctic re ction to penicillin. COMMENT The ceph losporins th t see most ction in the ICU setting re the third-gener t ion gents (cefot xime, ceftri xone, nd ceft zidime), which re used to tre t i nfections c used by Gr m-neg tive p thogens (T ble 35.1). Ceft zidime is p rticu l rly popul r bec use of its ctivity g inst P. eruginos . At the N tion l C n cer Institute, monother py with ceft zidime is one of the preferred regimens for the empiric ther py of febrile neutropeni (25). The fourth-gener tion gent ce fepime h s lso been used in febrile neutropeni (24), nd this gent m y repl c e ceft zidime bec use of its extended ctivity g inst Gr m-positive p thogens. IMIPENEM Imipenem h s the bro dest spectrum of ntib cteri l ctivity of ny ntibiotic c urrently v il ble (26). Bec use of its wide spectrum of ctivity, imipenem is u seful for tre ting polymicrobi l infections (e.g., bowel sepsis) nd for empiric ther py of febrile neutropenic p tients. P.571 ANTIBACTERIAL SPECTRUM As is evident in T ble 35.6, the ntib cteri l spectrum of imipenem includes mos t of the p thogens involved in nosocomi l infections. The most not ble exception is methicillin-resist nt st phylococci. In ddition to its ctivity g inst the Enterob cteri ce e, imipenem covers ll n erobic p thogens, including B. fr gi lis nd Enterococcus f ec lis. Some str ins of Pseudomon s (e.g., P. cep ci ) r e resist nt to imipenem, nd cquired resist nce in P. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 17)03-M y-05 08:21:34

Ovid: ICU Book eruginos is reported (26).

TABLE 35.6. THE ANTIBACTERIAL ACTIVITY OF IMIPENEM DOSING Imipenem is in ctiv ted by enzymes in the ren l tubules, so it is impossible to chieve high levels of the drug in urine. To overcome this problem, the commerci l prep r tion of imipenem cont ins n enzyme inhibitor, cil st tin. The combin tion imipenemcil st tin prep r tion is v il ble s Prim xin. The dose recommend tions for imipenemcil st tin represent the dose of imipenem. The usu l intr venou s dose in dults is 500 mg every 6 hours. In suspected Pseudomon s infections, t he dose is doubled to 1 g every 6 hours. In ren l f ilure, the dose should be re duced by 50 to 75% (22). TOXICITY The m jor effect of toxicity ssoci ted with imipenem is gener lized seizures, w hich occur in 1 to 3% of p tients receiving the drug (26). Most p tients h ve history of seizure disorder, n intr cr ni l m ss, or ren l f ilure. Although this is n uncommon occurrence, m ximum d ily dose of 2 g or 25 mg/kg h s been recommended (26). COMMENT The ultim te ntibiotic would be effective g inst ll p thogens nd produce no dverse re ctions. Imipenem comes closer to this ide l P.572 th n ny ntibiotic currently v il ble. Monother py with imipenem is useful regimen for the empi ric ther py of suspected sepsis, nd it is one of the preferred regimens for the empiric ther py of febrile neutropeni (25,27). However, bec use of emerging re sist nce in P. eruginos , monother py with imipenem is not recommended for susp ected Pseudomon s infections. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 17)03-M y-05 08:21:34

Ovid: ICU Book THE PENICILLINS Almost 30 penicillins re v il ble for clinic l use. The few th t re useful in the ICU re listed in T ble 35.7. TABLE 35.7. USES FOR PARENTERAL PENICILLINS IN THE ICU NATURAL PENICILLINS The penicillin discovered by Alex nder Fleming in 1929 is benzylpenicillin, or p enicillin G. This subst nce is ctive g inst erobic streptococci (S. pneumoni e, S. pyogenes) nd n erobic mouth flor . Penicillin G w s ctive g inst st ph ylococci when first discovered, but these org nisms h ve cquired resist nce t o penicillin by virtue of n enzyme (c lled -lactamase or penicillinase) that dis rupts the penicillin molecule. The limited spectrum of penicillin G limits its u se in the ICU. The major use for this agent is the occasional patient with pneum ococcal pneumonia or disseminated pneumococcal sepsis. PENICILLINASE-RESISTANT PENICILLINS To regain the original activity against staphylococci, the penicillin G molecule was modified to make it resistant to the penicillinase enzyme. This resulted in the introduction of methicillin, oxacillin, and nafcillin to treat staphylococc al infections. However, staphylococci have demonstrated an increasing resistance to these agents over the last 15 to 20 years. These resistant organisms, known as methicillin-resistant staphylococci, have ecome prominent pathogens in hospi talized patients. The treatment of choice for these organisms is vancomycin. P.5 73 EXTENDED-SPECTRUM PENICILLINS The penicillins in this category have an extended anti acterial spectrum that co vers Gram-negative aero ic acilli. This category includes the aminopenicillins (ampicillin and amoxicillin), the car oxypenicillins (car enicillin and ticarcil lin), and the ureidopenicillins (azlocillin, mezlocillin, and piperacillin). All groups are active against Gram-negative enteric pathogens, ut the latter two g roups are active against P. aeruginosa. These agents are also known as antipseud omonal penicillins. In addition to treating serious infections caused y Pseudom onas http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 17)03-May-05 08:21:34

Ovid: ICU Book organisms, these agents are also used for the empiric therapy of fever in neutro penic patients (25). Dosing The recommended doses for parenteral penicillins are shown in Ta le 35.7. Penici llin G has a short half-life (30 minutes) and is given y continuous intravenous infusion. The modified penicillins are usually given in a dose of 1 to 2 g ever y 4 to 6 hours. The penicillins are excreted y the kidneys, and dose reduction is advised in renal insufficiency. In renal failure, the dose interval should e extended to 8 to 12 hours (22). Toxicity The most common adverse reactions to the penicillins are allergic-type reactions . A rash develops in up to 4% of patients, ut anaphylaxis occurs in only 0.05% of patients (28). A hypersensitivity nephritis has een linked to methicillin, ut not to the other penicillins. A reversi le neurotoxic syndrome with delirium and seizures has een linked to high-dose therapy with penicillin G, often in pa tients with renal insufficiency (28). THE QUINOLONES The quinolone anti iotics were introduced in the mid-1980s for the treatment of complicated urinary tract infections in outpatients. Two parenteral agents are c urrently availa le for clinical use: ciprofloxacin and ofloxacin. ANTIBACTERIAL SPECTRUM The quinolones are active against (methicillin-sensitive) staphylococci and most of the Entero acteriaceae, including P. aeruginosa. They are less active agains t streptococci, and have no activity against anaero ic organisms. Emerging resis tance in Pseudomonas isolates is a concern (29). P.574 DOSING The dose of intravenous ciprofloxacin is 400 mg every 12 hours for serious infec tions (30). The drug is irritating to veins and should e either given via a cen tral vein or infused slowly over 1 hour. In renal failure, the daily dose should e reduced y 25 to 50% (22). TOXICITY The quinolones are safe in most patients. There are rare reports of a hypersensi tivity interstitial nephritis from ciprofloxacin (31), and one case of exacer at ion of myasthenia gravis (32). The quinolones also interfere with the hepatic me ta olism of theophylline and warfarin and can potentiate the actions of oth of these drugs (29,33). Ciprofloxacin causes a 25% increase in serum theophylline l evels, and com ined therapy has resulted in symptomatic theophylline toxicity (3 4). Although no dose adjustments are necessary, serum theophylline levels and pr othrom in times should e monitored carefully when ciprofloxacin is given in com ination with these two agents. COMMENT http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 17)03-May-05 08:21:34

Ovid: ICU Book Despite their popularity as oral agents, the quinolones have not found a niche i n the treatment of infections in the ICU. Because of the rapid emergence of resi stance in Pseudomonas organisms, the use of quinolones should e limited (29,34) . They can e used to treat the occasional urinary tract infection caused y org anisms that are not suscepti le to other conventional agents. VANCOMYCIN Vancomycin may e the single most popular anti iotic in critical care. However, it is overused, which is creating pro lems with acquired resistance. ANTIBACTERIAL SPECTRUM Vancomycin is active against all Gram-positive cocci, including anaero ic strept ococci (E. faecalis), coagulasenegative staphylococci (S. epidermidis), and meth icillin-resistant strains of S. aureus (35). It is also one of the most active a gents against Clostridium difficile, the pathogen responsi le for anti iotic-ass ociated pseudomem ranous colitis. In recent years, nosocomial strains of Enteroc occus have demonstrated an emerging resistance to vancomycin, and in a survey of U.S. hospitals conducted in 1993, 8% of all enterococcal isolates were vancomyc in-resistant (36). P.575 DOSING The usual intravenous dose of vancomycin is 500 mg every 6 hours. The drug must e infused slowly (no faster than 10 mg/minute) to minimize the risk of infusion -related reactions. Dose reduction is necessary in renal insufficiency. In patie nts with renal failure, the drug is given once every 4 days, and no supplemental doses are given after hemodialysis (22). Serum drug levels are often monitored to limit toxicity and maintain efficacy. Peak levels should e elow 40 mg/ L to reduce the risk of ototoxicity, and trough levels should e a ove 5 mg/L to mai ntain anti acterial activity (37).

TOXICITY Infusion-related Toxicity Rapid administration of vancomycin is associated with vasodilation, flushing, an d hypotension (38,39). The mechanism is vancomycin-induced histamine release fro m mast cells (38). Slow infusions (less than 10 mg/ minute) reduce the risk of t his reaction. Ototoxicity Vancomycin can cause reversi le hearing loss for high-frequency sounds when seru m drug levels exceed 40 mg/ L (37). Irreversi le deafness has een reported when serum levels exceed 80 mg/L (37). Both complications are uncommon, possi ly ec ause serum drug levels are monitored frequently. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 17)03-May-05 08:21:34

Ovid: ICU Book Nephrotoxicity Reversi le renal insufficiency is reported in 5% of patients receiving vancomyci n (37). There is no relationship to vancomycin dose, ut the incidence is higher during com ined therapy with aminoglycosides. Many of the patients are seriousl y ill and have other reasons for renal impairment, so the nephrotoxic potential of vancomycin may e overstated. COMMENT Vancomycin is a valua le drug in the ICU ecause of its activity for methicillin -resistant and coagulase-negative staphylococci. However, it is clearly overused , and the emergence of vancomycin-resistant enterococci P.576 should prompt clin icians to curtail the use of this agent in hospitalized patients. REFERENCES AMINOGLYCOSIDES 1. Edson RS, Terrell CL. The aminoglycosides. Mayo Clin Proc 1991;66:11581164. 2. Walting DM, Dasta JF. Aminoglycoside dosing considerations in intensive care unit patients. Ann Pharmacother 1993;27:351357. 3. Pancoast SJ. Aminoglycoside anti iotics in clinical use. Med Clin North Am 19 88;72:581612. 4. Triginer C, Izquierdo I, Fernandez R, et al. Gentamicin volume of distri utio n in critically ill septic patients. Intensive Care Med 1990;16:303306. 5. Rotschafer JC, Za inski RA, Walker KJ. Pharmacodynamic factors in anti iotic efficacy. Pharmacotherapy 1992;12:64S70S. 6. Schumock GT, Ra er SR, Crawford SY, et al. National survey of once-daily dosi ng of aminoglycoside anti iotics. Pharmacotherapy 1995;15:201209. 7. Wilson SE. Aminoglycosides: assessing the potential for nephrotoxicity. Surg Gynecol O stet 1986;171 (Suppl):2430. 8. Isenstein DA, Venner DS, Duggan J. Neuromuscular lockade in the intensive ca re unit. Chest 1992;102:12581266. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 17)03-May-05 08:21:34

Ovid: ICU Book 9. Drachman DB. Myasthenia gravis. N Engl J Med 1994;330:1791810. P.577 10. Lippmann M, Yang E, Au E, Lee C. Neuromuscular locking effects of to ramyci n, gentamicin, and cefazolin. Anesth Analg 1982;61:767770. ANTIFUNGAL AGENTS 11. Gallis HA, Drew RH, Pickard WW. Amphotericin B: 30 years of clinical experie nce. Rev Infect Dis 1990;12:308329. 12. Bult J, Franklin CM. Using amphotericin B in the critically ill: a new look at an old drug. J Crit Illness 1996;11:577585. 13. Carlson MA, Condon RE. Nephrotoxicity of amphotericin B. J Am Coll Surg 1994 ;179:361381. 14. Terrell CL, Hughes CE. Antifungal agents used in deep-seated mycotic infecti ons. Mayo Clin Proc 1992;67:6991. 15. Anaissie E, Bodey GP, Kantarjian H, et al. Fluconazole therapy for chronic d isseminated candidiasis in patients with leukemia and prior amphotericin therapy . Am J Med 1991;91:142150. 16. Rex JH, Bennett JE, Sugar AM, et al. A randomized trial comparing fluconazol e with amphotericin B for the treatment of candidemia in patients without neutro penia. N Engl J Med 1994;331:13251330. 17. Goodman JL, Winston DJ, Greenfield RA, et al. A controlled trial of fluconaz ole to prevent fungal infections in patients undergoing one marrow transplantat ion. N Engl J Med 1992;326:845851. 18. Crussel-Porter LL, Rindone JP, Ford MA, Jaskar DW. Low-dose fluconazole ther apy potentiates the hypoprothrom inemic effect of warfarin sodium. Arch Intern M ed 1993;153:102104. 19. Cadle RM, Zenon GJ, Rodriguez-Barradas MC, Hamill RJ. Fluconazole-induced sy mptomatic phenytoin toxicity. Ann Pharmacother 1994;28:191194. 20. Gearhart MO. Worsening of liver function with fluconazole and a review of az ole antifungal hepatotoxicity. Ann Pharmacother 1994;28:11771181. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 17)03-May-05 08:21:34

Ovid: ICU Book AZTREONAM 21. Brewer NS, Hellinger WC. The mono actams. Mayo Clin Proc 1991;66:11521157. 22. Bennett WM, Aronoff GR, Golper TA, et al., eds. Drug prescri ing in renal fa ilure. 3rd ed. Philadelphia: American College of Physicians, 1994. CEPHALOSPORINS 23. Gustafferro CA, Steckel erg JM. Cephalosporin antimicro ial agents and relat ed compounds. Mayo Clin Proc 1991;66:10641073. 24. Ramphal R, Gucalp R, Rotstein C, et al. Clinical experience with single agen t and com ination regimens in the management of infection in the fe rile neutrop enic patient. Am J Med 1996;100(Suppl 6A):83S89S. 25. Pizzo PA. Choosing empiric therapy for fe rile neutropenic patients. J Crit Illness 1995;10:165168. IMIPENEM 26. Hellinger WC, Brewer NS. Imipenem. Mayo Clin Proc 1991;66:10741081. P.578 27. Freifield A, Walsh T, Marshall D, et al. Monotherapy for fever and neutropen ia in cancer patients: a randomized comparison of ceftazidime versus imipenem. J Clin Oncol 1995;13:165176. PENICILLINS 28. McEvoy GK, ed. AHFS drug information monographs. Bethesda, MD: American Soci ety of Hospital Pharmacists, 1995;225338. QUINOLONES 29. Walker RC, Wright AJ. The fluoroquinolones. Mayo Clin Proc 1991;66:12491259. 30. Kljucar S, Heimesaat M, von Pritz uer E, et al. A comparison of intravenous ciprofloxacin dosage regimens in severe nosocomial infections. Infect Med 1992;9 (Suppl B):5872. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 17)03-May-05 08:21:34

Ovid: ICU Book 31. Allon M, Lopez EJ, Min K-W. Acute renal failure due to ciprofloxacin. Arch I ntern Med 1990;150:2187 2189. 32. Moore B, Safani M, Keesey J. Possi le exacer ation of myasthenia gravis y c iprofloxacin. Lancet 1988;1:882. 33. Ro son RA. The effects of quinolones on xanthine pharmacokinetics. Am J Med 1992;92(Suppl 4A):22S 26S. 34. Maddix DS. Do we need an intravenous fluoroquinolone? West J Med 1992;157:555 9. VANCOMYCIN 35. Wilhelm HP. Vancomycin. Mayo Clin Proc 1991;66:11701191. 36. Gin AS, Zhanel GG. Vancomycin-resistant enterococci. Ann Pharmacother 1996;3 0:615623. 37. Saunders NJ. Why monitor peak vancomycin concentrations? Lancet 1994;344:174 81750. 38. Levy JH, Kettlekamp N, Goertz P, et al. Histamine release y vancomycin: the mechanism for hypotension in man. Anesthesiology 1987;67:122125. 39. Romanelli VA, Howie MB, Myerowitz D, et al. Intraoperative and postoperative effects of vancomycin administration in cardiac surgery patients: a prospective , dou le- lind, randomized trial. Crit Care Med 1993;21:11241131. GENERAL WORKS 40. Mayo Clinic Proceedings Symposium on Antimicro ial Agents. Rochester, MN: Ma yo Clinic Proceedings, 1992. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 17)03-May-05 08:21:34

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 36 ACID-BASE INTERPRETATIONS A little learning is a dangerous thing. Drink deep, or taste not the Pyrean spri ng. --Alexander Pope In a survey conducted at a university teaching hospital (1) , 70% of the participating physicians claimed that they were well versed in the diagnosis of acid- ase disorders and that they needed no assistance in the inter pretation of arterial lood gases (ABGs). These same physicians were then given a series of ABG measurements to interpret, and they correctly interpreted only 4 0% of the test samples. A survey at another teaching hospital revealed that inco rrect acid- ase interpretations led to errors in patient management in one-third of the ABG samples analyzed (2). These surveys reveal serious deficiencies in a n area that tends to e ignored. This can cause trou le in the ICU, where 9 of e very 10 patients may have an acid ase disorder (3). This chapter presents a stru ctured approach to acid- ase interpretations ased on a set of welldefined rules that are applied to ABG analysis (4,5,6,7 and 8). This approach is taken from a computer program that interprets ABGs (8), so it should work. BASIC CONCEPTS The hydrogen ion concentration [H+] in extracellular fluid is determined y the alance etween the partial pressure of car on dioxide (PCO2) and the concentrat ion of icar onate [HCO3] in the fluid. This relationship is expressed as follow s (3): Using a normal arterial PCO2 of 40 mm Hg and a normal serum HCO3 concentration o f 24 mEq/L, the normal [H+] in arterial lood is 24 (40/24) = 40 nEq/L. P.582 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 16)03-May-05 08:08:10

Ovid: ICU Book [H+] AND PH Note that the [H+] in extracellular fluid is expressed in nanoequivalents (nEq) per liter. A nanoequivalent is one-millionth of a milliequivalent, so there are millions more sodium, chloride, and other ions measured in mEq than there are hy drogen ions. Because nanoequivalents is a cum ersome term, the [H+] is routinely expressed in pH units, which are derived y taking the negative logarithm ( ase 10) of the [H+] in nEq/L. The relationship etween pH and [H+] is shown in Figu re 36.1. A normal [H+] of 40 nEq/L corresponds to a pH of 7.40. Because the pH i s a negative logarithm of the [H+], changes in pH are inversely related to chang es in [H+] (e.g., a decrease in pH is associated with an increase in [H+]). Note that as the pH decreases from its highest value (7.60), the slope of the curve progressively increases. Thus, as the pH decreases, a progressively larger chang e occurs in [H+] associated with a given change in pH. The num ers a ove the cur ve indicate the change in [H+] associated with each change of 0.1 pH units. Over the normal pH range from 7.36 to 7.44 (indicated y the shaded area in Figure 3 6.1), the change in [H+] is less than 10 nEq/L (this shows how tightly the [H+] is controlled). At the acidotic end of the curve, the change in [H+] is more tha n threefold higher than at the alkalotic end of the curve (20 nEq/L versus 6 nEq /L per 0.1 pH unit, respectively). Therefore, the acid- ase consequences P.583 o f a given change in pH depends on the underlying acid- ase status of the patient . http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 16)03-May-05 08:08:10

Ovid: ICU Book Figure 36.1. The relationship etween hydrogen ion concentration [H+] and pH. Th e num ers a ove the curve indicate the change in [H+] associated with a change o f 0.1 pH unit. The shaded area shows the normal pH range in extracellular fluid. COMPENSATORY CHANGES According to the determinants of [H+] shown earlier, the sta ility of the extrac ellular pH is determined y the sta ility of the PCO2/HCO3 ratio. Maintaining a constant PCO2/HCO3 ratio will maintain a constant extracellular pH. This is the asis for the primary and compensatory acid- ase changes shown in Ta le 36.1. Wh en a primary acid- ase distur ance alters one component of the PCO2/HCO3 ratio, the compensatory response alters the other component in the same direction to ke ep the PCO2/HCO3 ratio constant. Thus, when the primary disorder is meta olic (i .e., a change in HCO3), the compensatory response is respiratory (i.e., a change in PCO2), and vice-versa. It is important to emphasize that compensatory respon ses limit rather than prevent changes in pH (i.e., compensation is not synonymou s with correction). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 16)03-May-05 08:08:10

Ovid: ICU Book TABLE 36.1. PRIMARY AND SECONDARY ACID-BASE DERANGEMENTS Respiratory Compensation The ventilatory control system provides the compensation for meta olic acid- ase distur ances, and the response is prompt. The changes in ventilation are mediat ed y H+sensitive chemoreceptors located in the carotid ody (at the carotid if urcation in the neck) and in the lower rainstem. A meta olic acidosis excites t he chemoreceptors and initiates a prompt increase in ventilation and a decrease in arterial PCO2. A meta olic alkalosis silences the chemoreceptors and produces a prompt decrease in ventilation and increase in arterial PCO2. When these vent ilatory responses are fully functional, the compensatory, or expected, arterial PCO2 can e defined according to the equations shown in Ta le 36.2. The compensa tory response to meta olic alkalosis has varied in different reports; however, t he equation shown in Ta le 36.2 has proven relia le, at least up to a HCO3 level of 40 mEq/L (9). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 16)03-May-05 08:08:10

Ovid: ICU Book TABLE 36.2. EXPECTED CHANGES IN ACID-BASE DISORDERS Meta olic Compensation The kidneys provide the compensation for respiratory acid- ase disorders y adju sting HCO3 rea sorption in the proximal tu ules. Respiratory acidosis stimulates HCO3 rea sorption, which increases the serum HCO3 concentration, whereas respir atory alkalosis inhi its HCO3 rea sorption, which decreases the serum HCO3 conce ntration. P.584 However, the compensatory response in the kidneys is not immedia te. It egins to appear in 6 to 12 hours and slowly increases to a steady-state response over the next few days. Because of this delay in renal compensation, re spiratory acid- ase disorders are classified as acute ( efore renal compensation egins) and chronic (after renal compensation is fully developed). The changes in pH that accompany the acute and chronic respiratory disorders can e defined using the equations shown in Ta le 36.2. THE RULES OF ACID-BASE INTERPRETATION The approach that follows is ased on three acid- ase varia les: pH, PCO2, and H CO3. The reference ranges for each of these varia les is shown elow (3). Any me asurement that falls outside of these ranges is considered to e a normal. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 16)03-May-05 08:08:10

Ovid: ICU Book PRIMARY METABOLIC DISORDERS Rule 1. A primary meta olic acid- ase disorder is present if the pH is a normal and the pH and PCO2 change in the same direction. Thus, a primary meta olic acidosis is present if the arterial pH is elow 7.36 and the arterial PCO2 (PaCO2) is decrea sed, whereas a primary meta olic alkalosis is present if the pH is a ove 7.44 an d the PaCO2 is increased. If a primary meta olic acid- ase disorder is identifie d, proceed to Rule 2. Rule 2. A superimposed respiratory acid- ase disorder is present if any of the following conditions are satisfied. The measured PCO2 is normal. The measured PCO2 is hig her than the expected PCO2 (this indicates a superimposed respiratory acidosis). The measured PCO2 is less than the expected PCO2 (this indicates a superimposed respiratory alkalosis). Thus, once a primary meta olic disorder has een identi fied, the appropriate equation in Ta le 36.2 should e used to define the PCO2 t hat is expected with full respiratory compensation. If the measured P.585 PCO2 i s a ove or elow the expected PCO2, a com ined (meta olic and respiratory) acid ase disorder is present. PRIMARY RESPIRATORY DISORDERS Rule 3. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 16)03-May-05 08:08:10

Ovid: ICU Book A primary respiratory acid- ase disorder is present if the PaCO2 is a normal and the PaCO2 and pH change in opposite directions. Thus, a primary respiratory aci dosis is present if the PaCO2 is a ove 44 mm Hg and the arterial pH is decreased , whereas a respiratory alkalosis is present if the PaCO2 is elow 36 mm Hg and the arterial pH is increased. If a primary respiratory acid- ase disorder is ide ntified, proceed to Rule 4. Rule 4. The expected change in pH (determined y using the equations in Ta le 36.2) is u sed to determine whether the respiratory disorder is acute or chronic and whethe r a superimposed meta olic acid- ase disorder is present. If the change in pH is 0.008 times the change in PaCO2, the respiratory disorder is acute (uncompensat ed). If the change in pH is 0.003 to 0.008 times the change in PaCO2, the respir atory disorder is partially compensated.. If the change in pH is 0.003 times the change in PaCO2, the respiratory disorder is chronic (fully compensated). If th e change in pH is more than 0.008 times the change in PaCO2, a superimposed meta olic acid- ase disorder is present. For example, if the PCO2 has increased to 5 0 mm Hg, then an acute respiratory acidosis would have a decrease in pH of (0.00 8 10 mm Hg) 0.08 pH units, a chronic respiratory acidosis would have a change in pH of (0.003 10 mm Hg) 0.03 pH units, and a partially compensated respiratory a cidosis would have a change in pH etween 0.03 and 0.08 pH units. If the decreas e in pH is greater than 0.08 pH units, then a superimposed meta olic acidosis is present. MIXED DISORDERS Rule 5. A mixed (acidosis and alkalosis) acid- ase disorder is present if the PaCO2 is a normal and the pH is unchanged or normal, or if the pH is a normal and the PaCO 2 is unchanged or normal. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 16)03-May-05 08:08:10

Ovid: ICU Book This rule is ased on the fact that compensatory responses to primary acid- ase disorder do not completely correct the primary a normality. Therefore, if the Pa CO2 is 50 mm Hg and the pH is 7.40, the compensatory change in pH is more than e xpected, indicating that there is a meta olic alkalosis in addition to the compe nsated respiratory acidosis. RULE-ORIENTED ACID-BASE INTERPRETATIONS This section illustrates how the 5 rules just descri ed can e applied to the in terpretation of ABGs. Each of the interpretation egins with the arterial pH. ACIDEMIA If the pH is elow 7.36, check the PCO2 and proceed as follows: A low or normal PaCO2 indicates a primary meta olic acidosis (Rule 1). P.586 The difference etw een the measured and expected PaCO2 is then used to identify a superimposed resp iratory disorder (Rule 2). A high PCO2 indicates a primary respiratory acidosis (Rule 3). The change in pH is then used to determine whether the disorder is acu te or chronic, or whether a superimposed meta olic acid- ase disorder is present (Rule 4). ALKALEMIA If the pH is a ove 7.44, check the PCO2 and proceed as follows: A normal or high PaCO2 indicates a primary meta olic alkalosis (Rule 1). A comparison of the mea sured and expected PaCO2 is then used to identify an associated respiratory diso rder (Rule 2). A low PaCO2 indicates a primary respiratory alkalosis (Rule 3). T he change in pH is then used to determine whether the disorder is acute or chron ic, or whether a superimposed meta olic disorder is present (Rule 4). NORMAL PH http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 16)03-May-05 08:08:10

Ovid: ICU Book If the arterial pH is unchanged or normal, the PCO2 should e checked: A high PC O2 indicates a mixed respiratory acidosis-meta olic alkalosis (Rule 5). A low PC O2 indicates a mixed respiratory alkalosis-meta olic acidosis (Rule 5). A normal pH com ined with a normal PaCO2 is not a solute evidence against an acid- ase d isorder ecause a meta olic acidosis co-existing with a meta olic alkalosis can e accompanied y a normal pH and PaCO2. THE ANION GAP The anion gap (AG) is an acid- ase parameter that is used to evaluate patients w ith a meta olic acidosis to determine whether the pro lem is an accumulation of hydrogen ions (e. g., lactic acidosis) or a loss of icar onate (e.g., diarrhea) . Although it does not perform well in certain situations, the AG can still prov ide valua le information when it is used judiciously (10). THE CONCEPT To achieve electrochemical alance the ionic elements in the extracellular fluid must have a net charge of zero. Thus, the concentration of negatively-charged a nions and positivelycharged cations must alance. All ions participate in this alance, oth those that are routinely measured, such as sodium (Na), chloride (C l), and icar onate (HCO3), and those that are not measured. The unmeasured cati ons (UC) and P.587 nunmeasured anions (UA) have the following relationships to t he commonly measured electrolytes: or when the terms are rearranged: The relationship (UA UC) is a measurement of the relative a undance of unmeasure d anions and is called the AG. Determinants The anions and cations that contri ute to the AG are shown in Ta le 36.3. The pl asma proteins are the major source of unmeasured anions, whereas potassium and c alcium make up the ulk of the unmeasured cations (descri ing these electrolytes as unmeasured is hardly http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 16)03-May-05 08:08:10

Ovid: ICU Book appropriate). The charge difference etween the two groups reveals an anion exce ss (anion gap) of 12 mEq/L. Much of this difference is due to the plasma protein s. For example, a 50% reduction in plasma proteins can result in a 75% reduction in the AG (from 12 to 4 mEq/ L), as indicated at the ottom of Ta le 36.3. Most of the charge donated y plasma proteins is from al umin, so hypoal uminemia ca n have a significant influence in reducing the AG (3). TABLE 36.3. DETERMINANTS OF THE ANION GAP Reference Range The normal range for the AG was originally defined as 8 ever this range was defined using automated systems for (the SMA-6 and SMA-12) that are no longer used in most sing the newer automated systems (e.g., the ASTRA), the is 3 to 11 mEq/L (12), which is lower than the original nificance of this lower reference range is unclear, and o the interpretation of the AG. to 16 mEq/L (11,12). How electrolyte measurements clinical la oratories. U normal range for the AG reference range. The sig it adds more confusion t

METABOLIC ACIDOSIS As mentioned previously, the AG has een used to identify certain types of meta olic acidosis. Specifically, it is used to differentiate etween P.588 meta olic acidoses caused y an accumulation of hydrogen ions and meta olic acidoses caus ed y a loss of icar onate ions. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 16)03-May-05 08:08:10

Ovid: ICU Book High Anion Gap When a meta olic acidosis is due to the accumulation of hydrogen ions in the ext racellular fluid (e.g., lactic acidosis), the hydrogen ions com ine with icar o nate to form car onic acid. This decreases the icar onate concentration in the extracellular fluid, which in turn increases the AG (as predicted y the relatio nship AG = Na [CL + HCO3]). Therefore, a meta olic acidosis with a high AG is mo st likely caused y organic acid accumulation (i.e., lactic acid or ketoacids) o r renal failure with impaired hydrogen ion excretion. Although a high AG can e helpful in the setting of a meta olic acidosis, it should not e used as evidenc e of a meta olic acidosis unless it exceeds 30 mEq/L (13). In fact, an elevated AG can e a sign of an underlying meta olic alkalosis. That is, alkalosis can el evate the AG, presuma ly y increasing the strength of the negative charge on al umin molecules (10,11).

Relia ility The AG has not een proven to e a sensitive marker of organic acid accumulation (i.e., lactic acidosis). Although lactic acid accumulation should e accompanie d y a high AG, numerous reports of lactic acidosis with a normal anion gap exis t (14,15). In a recent case report, the AG was only 11 mEq/L, despite a markedly elevated serum lactate of 13 mEq/L (15). Some of these cases can e explained y the presence of conditions that could prevent the AG from increasing in respon se to acid accumulation (e.g., hypoal uminemia) (16). However, this does not add to the poor performance of the AG as a screen for lactic acidosis. MIXED METABOLIC DISTURBANCES Mixed meta olic distur ances (e.g., high AG from ketoacidosis plus normal AG fro m diarrhea) can e identified using the relationship etween the increase in AG and the decrease in serum HCO3. This relationship is expressed as the AG excess/ HCO3 deficit ratio, which is sometimes called the gap-gap. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 16)03-May-05 08:08:10

Normal Anion Gap When a meta olic racellular fluid gain of chloride ease in chloride ncentration, the rmal AG is taken increased icar

acidosis is caused y the loss of icar onate ions from the ext (e.g., diarrhea), the icar onate loss is counter alanced y a ions to maintain electrical charge neutrality. Because the incr concentration is proportional to the decrease in icar onate co AG remains unchanged. Therefore, a meta olic acidosis with a no as a sign of a icar onate-wasting process, such as diarrhea or onate losses in the urine in early renal failure.

Ovid: ICU Book P.589 MIXED METABOLIC ACIDOSES When hydrogen ions accumulate in lood, the decrease in serum HCO3 is equivalent to the increase in AG and the AG excess/HCO3 deficit ratio is unity. When a hyp erchloremic acidosis is present, the ratio approaches zero. When a mixed acidosi s (high AG + normal AG) is present, the AG excess/HCO3 deficit ratio indicates t he relative contri ution of each type to the acidosis. For example, a ratio of 0 .5 indicates an equivalent contri ution from high AG and normal AG acidoses. Dia etic Ketoacidosis Dia etic ketoacidosis is expected to present as a high AG meta olic acidosis. Ho wever, after therapy with fluids and insulin egins, the high AG acidosis change s to a normal AG acidosis (17). This is partly due to the chloride load in the i ntravenous fluids. In this situation, the serum icar onate remains low (dilutio nal effect) ut the AG excess/HCO3 deficit ratio shows a steady decline. In this situation, the persistent decrease in serum HCO3 will create a false impression that the ketoacids are not eing cleared. This illustrates the value of the AG excess/HCO3 deficit ratio for the management of patients with ketoacidosis. MIXED ACIDOSISALKALOSIS When alkali is added in the presence of a high AG acidosis, the decrease in seru m icar onate is less than the decrease in AG and the AG excess/HCO3 deficit rat io is greater than unity. Meta olic alkalosis is common in the ICU ecause of th e popularity of nasogastric suction and diuretics. Therefore, mixed meta olic ac idosis and alkalosis may e more common than suspected. VENOUS BLOOD GASES ABG analysis continues to thrive, despite the fact that the acid- ase status in arterial lood is not likely to reflect the acid- ase status in the peripheral t issues. The discrepancy etween arterial and venous lood gases is shown in Figu re 36.2. (18). These data are taken from a study of 16 patients in the ICU with indwelling pulmonary artery catheters who had experienced a cardiac arrest. The measurements were o tained during cardiopulmonary resuscitation. Note that the a rterial lood appears reasona ly normal (pH is 7.41 and PCO2 is http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 16)03-May-05 08:08:10

Ovid: ICU Book 32 mm Hg) and that the venous lood reveals severe acidemia (pH of 1.5) and hype rcapnia (PCO2 of 74 mm Hg). Thus, even in the most extreme conditions, the arter ial lood is not a sensitive marker of the acid- ase conditions at the tissue le vel. Remem er this when performing cardiopulmonary resuscitation. Figure 36.2. Acid- ase parameters in arterial and venous lood during cardiopulm onary resuscitation. Height of the vertical columns indicates the mean; cross- a rs indicate standard deviation. Study group: 16 adult patients. (From Weil MH, R ackow EC, Trevino R, et al. Difference in acid- ase state etween venous and art erial lood during cardiopulmonary resuscitation. N Engl J Med 1986;315:153156.) P.590 REFERENCES SUGGESTED READINGS Arieff AI, DeFronzo RA, eds. Fluid electrolyte and acid- ase disorders. New York : http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 16)03-May-05 08:08:10

Ovid: ICU Book Churchill Livingstone, 1985. Rose BD. Clinical physiology of acid- ase and electrolyte disorders. 4th ed. New York: McGraw-Hill, 1994.

3. Gilfix BM, Bique M, Magder S. A physical chemical approach to the analysis of acid ase alance in the clinical setting. J Crit Care 1993;8:187197. P.591 COMPREHENSIVE REVIEWS 4. Narins RG, Emmett M. Simple and mixed acid- ase disorders: a practical approa ch. Medicine 1980;59:161187 (89 references). 5. Laski ME, Kurtzman NA. Acid- ase disorders in medicine. Dis Mon 1996;XLII:5712 8 (200 references). MINIREVIEWS 6. Morganroth M. An analytical approach to diagnosing acid- ase disorders. J Cri t Illness 1990;5:138150 (5 references). 7. Ha er RJ. A practical approach to acid- ase disorders. West J Med 1991;155:14 6 151 (22 references). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 16)03-May-05 08:08:10

2. Broughton JO, Kennedy TC. Interpretation of arterial lood gases Chest 1984;85:148149.

INTRODUCTION 1. Hingston DM. A computerized interpretation of arterial pH and do physicians need it? Respir Care 1982;27:809815.

lood gas data: y computer.

Ovid: ICU Book SOFTWARE 8. Krasner J, Marino PL. Respiratory expert. Philadelphia: WB Saunders, 1987. SELECTED REFERENCES 9. Javaheri S, Kazemi H. Meta olic alkalosis and hypoventilation in humans. Am R ev Respir Dis 1987;136:10111016. 10. Emmet M, Narins RG. Clinical use of the anion gap. Medicine 1977;56:3854. 11. Oster JR, Perez GO, Materson BJ. Use of the anion gap in clinical medicine. South Med J 1988;81:229237. 12. Winter SD, Pearson JR, Ga ow PA, et al. The fall of the serum anion gap. Arc h Intern Med 1990;150:311313. 13. Ga ow PA, Kaehny WD, Fennessey PV. Diagnostic importance of an increased ani on gap. N Engl J Med 1980;303:854858. 14. I erti TS, Lie owitz AB, Papadakos PJ, et al. Low sensitivity of the anion g ap as a screen to detect hyperlactatemia in critically ill patients. Crit Care M ed 1990;18:275277. 15. Schwartz-Goldstein B, Malik AR, Sarwar A, Brandtsetter RD. Lactic acidosis a ssociated with a normal anion gap. Heart Lung 1996;25:7980. 16. Ernest D, Herkes RG, Raper RF. Alterations in the anion gap following cardio pulmonary ypass. Crit Care Med 1992;20:5256. 17. Paulson WD. Anion gap- icar onate relationship in dia etic ketoacidosis. Am J Med 1986;81:9951000. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 16)03-May-05 08:08:10

Ovid: ICU Book 18. Weil MH, Rackow EC, Trevino R. Difference in acid- ase state etween venous and arterial lood during cardiopulmonary resuscitation. N Engl J Med 1986;315:1 53156. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 16)03-May-05 08:08:10

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 37 THE ORGANIC ACIDOSES This chapter focuses on the clinical disorders produced y the accumulation of o rganic acids (i. e., lactic acid and ketoacids) in the extracellular fluid. The important dictum to emphasize in these conditions is that the acid- ase distur a nce is not the primary illness, ut it is a clinical marker of an underlying pat hologic disorder. Therefore, the focus should e on identifying and correcting t he pathologic disorder rather than treating the acid- ase distur ance. LACTIC ACIDOSIS LACTATE METABOLISM Lactate is the end-product of anaero ic glycolysis and is normally produced at a rate of 1 mmol/ kg/hour (or 1920 mmol/day for a 175-l adult) (1,2,3 and 4). Th e overall reaction for anaero ic glycolysis is as follows: Note that the reaction produces lactate, a negatively-charged ion, not lactic ac id. The hydrogen ions needed to convert lactate to lactic acid must e generated y the hydrolysis of ATP (5). Therefore, lactate production is not synonymous w ith lactic acid production. Most of the lactate production occurs in skeletal mu scle, owel, rain, and circulating erythrocytes. The lactate generated in these tissues can e taken up y the liver and converted to glucose (via gluconeogene sis) or can e used as a primary oxidative fuel. P.593 Energetics The energy yield from glucose meta olism is outlined in Figure 37.1 (see also Ta le 13.5). The anaero ic meta olism of glucose generates 47 kilocalories (kcal) of energy per mole of glucose, whereas the aero ic meta olism of glucose generat es 673 kcal per mole of glucose (6). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 20)03-May-05 08:08:38

Ovid: ICU Book Therefore, the energy yield from anaero ic glycolysis is only 7% of the energy y ield from aero ic glucose meta olism. However, lactate can serve as an oxidative fuel, and the oxidative use of lactate can correct the energy deficit associate d with anaero ic glycolysis. The oxidation of lactate generates 326 kcal per mol e of lactate (6), and ecause 1 mole of glucose generates 2 moles of lactate, th e energy yield from anaero ic glycolysis will e increased y 652 (2 326) kcal p er mole of glucose if the lactate that is generated is completely oxidized. This role of lactate as an oxidative fuel (called the lactate shuttle) has een desc ri ed in exercise (7), and it may also e operative in the developing stages of clinical shock. For example, when the skeletal muscles ecome anaero ic in the e arly stages of clinical shock, the lactate that is generated could e used as a source of energy y other vital organs that are not yet anaero ic, such as the h eart and central nervous system. Figure 37.1. The salient features of glucose and lactate meta olism. Hyperlactatemia The normal concentration of lactate in lood is less than 2 mmol/L while at rest , and up to 5 mmol/ L during exercise (2,3 and 4). Therefore, a resting lood la ctate level a ove 2 mmol/L is considered a normal. However, mild elevations in lood lactate (2 to 4 mmol/L) may not P.594 e accompanied y an acidosis. Theref ore, hyperlactatemia is not synonymous with lactic acidosis. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 20)03-May-05 08:08:38

Ovid: ICU Book ETIOLOGIES The principal causes of hyperlactatemia and lactic acidosis are indicated in ita licized print in Figure 37.1. Oxygen Deprivation The single most important cause of lactic acidosis is deficient cell oxygenation in shock (e.g., hypovolemic, cardiogenic, and septic shock). Hypoxemia and anem ia, which are presumed to cause impaired cellular oxygenation, are rarely accomp anied y lactic acidosis. The lack of association etween severe hypoxemia and h yperlactatemia is shown in Ta le 24.1. The ina ility of severe anemia to produce hyperlactatemia is descri ed in Chapter 44. In patients with clinical shock, th e severity of the hyperlactatemia has prognostic value (2,3,4,8,9 and 10). This is demonstrated in Figure 37.2. Note that when the lood lactate level rises to 10 mmol/L, the chances of survival are negligi le. The prognostic value of lood lactate levels pertains only to patients with clinical shock. Figure 37.2. The predictive value of lood lactate levels in patients with circu latory shock. (From Weil MH, Afifi AA. Experimental and clinical studies on lact ate and pyruvate as indicators of the severity of acute circulatory failure [sho ck]. Circulation 1970;16:9891001.) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 20)03-May-05 08:08:38

Ovid: ICU Book P.595 Endotoxemia As descri ed in Chapter 13, endotoxemia can elevate lood lactate levels without producing cellular oxygen deprivation (see Figure 13.5). This effect is due to endotoxin-mediated inhi ition of pyruvate dehydrogenase, the enzyme that initiat es pyruvate oxidation in the mitochondria (11). This effect implies that sepsis with endotoxin-producing gram-negative organisms can e a source of hyperlactate mia in the a sence of septic shock. This may explain the reported association e tween lactic acidosis and the acquired immunodeficiency syndrome (12). Thiamine Deficiency Thiamine serves as a co-factor for the pyruvate dehydrogenase enzyme that initia tes pyruvate oxidation in the mitochondria. Therefore, thiamine deficiency can e accompanied y hyperlactatemia (13). As with endotoxemia, the hyperlactatemia in thiamine deficiency may not e accompanied y evidence of clinical shock. Bec ause thiamine deficiency may e common in critically ill patients (see Chapter 4 6), this diagnosis should e considered in all cases of unexplained hyperlactate mia in the ICU. Alkalosis Severe alkalosis (respiratory or meta olic) can raise lood lactate levels (14) ecause the increased activity in pH-dependent enzymes in the glycolytic pathway promotes lactate production. When liver function is normal, the liver clears th e extra lactate generated during alkalosis and hyperlactatemia ecomes evident o nly when the pH of the lood is 7.6 or higher. However, in patients with impaire d liver function, hyperlactatemia can e seen with less severe alkalemia. Alkalo sis-enhanced lactate formation is an undesira le consequence of alkali therapy f or lactic acidosis. Others Other possi le causes of hyperlactatemia in patients in the ICU are seizures (fr om increased lactate production), hepatic insufficiency (from reduced lactate cl earance), epinephrine infusions (from enhanced glycogenolysis with increased lac tate production), nitroprusside toxicity (from cyanide accumulation), and acute asthma (possi ly from enhanced lactate production y the respiratory muscles) (1 5,16 and 17). Hyperlactatemia associated with epinephrine infusions and hepatic insufficiency is often mild and not accompanied y lactic acidosis (15). Hyperla ctatemia that accompanies generalized seizures can e severe ut is transient (1 6). Hyperlactatemia during nitroprusside infusions is a manifestation of cyanide http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 20)03-May-05 08:08:38

Ovid: ICU Book intoxication and is an ominous sign (see Chapter 53). P.596 DIAGNOSIS Hyperlactatemia is a possi le cause of any meta olic acidosis in a critically il l patient, regardless of the anion gap. The Anion Gap As descri ed in Chapter 36, the anion gap should e elevated in lactic acidosis, ut there are numerous reports of a normal anion gap in patients with lactic ac idosis (16,18). Therefore, the anion gap should not e used as a screening test for possi le lactic acidosis. For more information on the anion gap, see Chapter 36. Blood Lactate Lactate concentrations can e measured in plasma or whole lood. A lactate-speci fic electrode that requires a lood sample of only 0.13 mL is availa le and can complete the lactate measurement in less than 2 minutes (10). This electrode is com ined with other su stratespecific electrodes in a porta le chemistry analyze r (NOVA SP7, NOVA Biomedical, Waltham, Massachusetts) that can e used at the e dside. If immediate measurements are unavaila le, the lood sample should e pla ced on ice to retard lactate production y red lood cells in the sample. A lact ate level a ove 2 mmol/L (in plasma or whole lood) is a normal, and in patients with circulatory shock, a lood lactate level a ove 4 mmol/L is evidence of sig nificant tissue oxygen deficits (see Ta le 13.1). The prognostic value of the 2 mmol/L versus 4 mmol/L threshold in patients with suspected shock is shown in Ta le 13.4.

D-LACTIC ACIDOSIS The lactate produced y mammalian tissues is a levo-isomer ( ends light to the l eft), whereas a dextro-isomer of lactate ( ends light to the right) is produced y certain strains of acteria that can populate the owel (19). D-lactate gener ated y acterial fermentation in the owel can gain access to the systemic circ ulation and produce a meta olic acidosis, often com ined with a meta olic enceph alopathy (20). Most cases of D-lactic acidosis have een reported after extensiv e small owel resection or after jejunoileal ypass for mor id o esity (19,20 an d 21). Diagnosis http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 20)03-May-05 08:08:38

Ovid: ICU Book D-lactic acidosis can produce an elevated anion gap. However, the standard la or atory assays for lood lactate measure only L-lactate and will miss D-lactic aci dosis. Most clinical la oratories will perform a specific assay for D-lactate on request. P.597 ALKALI THERAPY FOR LACTIC ACIDOSIS The primary goal of therapy in lactic acidosis is to correct the underlying meta olic a normality. Alkali therapy aimed at preventing severe acidemia (i.e., ser um pH elow 7.2) has also een popular (22) ut is of questiona le value. The al kali therapy's lack of value is evident in the most recent recommendations of th e American Heart Association to avoid icar onate therapy during cardiac arrest (see Ta le 17.5). The following issues are relevant to the use of alkali therapy in lactic acidosis. RATIONALE The principal fear in acidosis is the risk of impaired myocardial contractility (23). However, in the intact organism, acidemia is often accompanied y an incre ase in cardiac output (24). This is explained y the a ility of acidosis to stim ulate catecholamine release from the adrenals and to produce vasodilation. There fore, impaired contractility from acidosis is less of a concern in the intact or ganism. Furthermore, acidosis may have a protective role in the setting of clini cal shock. For example, extracellular acidosis has een shown to protect energydepleted cells from cell death (25). PROTOCOL Alkali therapy is carried out with sodium icar onate solutions such as the ones shown in Ta le 37.1. Bicar onate administration is guided y the estimated defi cit in the icar onate (HCO3) uffer pool: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 20)03-May-05 08:08:38

Ovid: ICU Book TABLE 37.1. BICARBONATE-CONTAINING BUFFER SOLUTIONS The ody weight in this equation is lean ody weight. The factor 0.6 represents the apparent icar onate space (a reflection of total ody uffering capacity), which is 60% of the ody weight in mild to moderate acidosis (26). In severe aci dosis (serum HCO3 less than 10 mEq/L), the apparent icar onate space is increas ed to 70% of the ody P.598 weight, so 0.7 can e used in Equation 37.2 when the measured serum HCO3 is elow 10 mEq/L (26). In the a sence of a superimposed re spiratory acid- ase disorder, a serum HCO3 of 15 mEq/L should e sufficient to k eep the serum pH in a safe range. Therefore, a desired HCO3 of 15 mEq/ L can e used to calculate the icar onate deficit. One-half of the icar onate deficit i s replaced immediately, and the remaining deficit is infused over the next 4 to 6 hours. EFFICACY The administration of sodium icar onate has limited success in raising the seru m pH in patients with lactic acidosis (27). This has een attri uted to the tend ency of icar onate to generate CO2, a volatile acid. However, a simpler explana tion is provided y examining the titration curve for the car onic acid- icar on ate uffer system, which is shown in Figure 37.3. The HCO3 uffer pool is genera ted y the dissociation of car onic acid (H2CO3): http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 20)03-May-05 08:08:38

Ovid: ICU Book Figure 37.3. The titration curve for the car onic acid- icar onate uffer system . The large, shaded area indicates the effective pH range for the icar onate u ffer system, which does not coincide with the normal pH range for extracellular fluid. (Adapted from Comroe JH. Physiology of respiration. Chicago: Year ook Med ical Pu lishers, 1974;203.) The dissociation constant (pK) for car onic acid (i.e., the pH at which the acid is 50% dissociated) is 6.1, as indicated on the titration curve. Buffers are mo st effective within 1 pH unit on either side of the pK (28), P.599 so the effect ive range of the icar onate uffer system should e an extracellular pH etween 5.1 and 7.1 pH units (indicated y the shaded area on the titration curve). The refore, icar onate is not expected to e an effective uffer in the usual pH ra nge of extracellular fluid. The limited uffer capacity for icar onate in the p hysiologic pH range, as predicted y its titration curve, deserves more emphasis in discussions of icar onate therapy for the organic acidoses. In fact, icar onate is est descri ed as a transport form for car on dioxide in lood (see Fig ure 2.3), and not as a uffer. Hemodynamics Bicar onate therapy has also shown a limited a ility to correct the hemodynamic alterations in shock-associated lactic acidosis (29). In fact, whatever hemodyna mic improvements are produced y icar onate administration may e the result of the sodium load (Ta le 37.1) and http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 20)03-May-05 08:08:38

Ovid: ICU Book not the alkali therapy (30). ADVERSE EFFECTS A num er of undesira le effects are associated with sodium icar onate therapy. One of the principal disadvantages is the a ility of sodium icar onate to gener ate CO2 and raise the PCO2 in oth intracellular and extracellular fluids (24,29 ,30 and 31). Considering the PCO2 of the sodium icar onate solution in Ta le 37 .1, this effect is not surprising. The increase in PCO2 is undesira le ecause i t creates a (volatile) acid load that must e excreted y the lungs and reduces the uffering capacity of sodium icar onate. Infusions of sodium icar onate ca n also increase lood lactate levels (29). This effect is attri uted to alkalosi s-induced augmentation of lactate production, and it hardly seems desira le in t he setting of hyperlactatemia. Finally, icar onate can ind calcium and lower t he concentration of ionized calcium in the lood. This can impair myocardial con tractility and promote hypotension (27). OTHER BUFFER SOLUTIONS Car icar Car icar is a commercially availa le uffer solution that is a 1:1 mixture of s odium icar onate and disodium car onate. As shown in Ta le 37.1, Car icar has less icar onate and a much lower PCO2 than the standard 7.5% sodium icar onate solution. As a result, Car icar does not produce the increase in PCO2 seen wit h sodium icar onate infusions. Because of this, Car icar is a more effective uffer than sodium icar onate (29,30). Tromethamine The amine uffer tromethamine (TRIS or THAM) provides intracellular and extracel lular uffering without generating CO2 (31). The P.600 uffer reaction for THAM has a pK of 7.8 at ody temperatures, so THAM provides effective uffering over the pH range of 6.8 to 8.8 (a more appropriate pH range than seen with icar ona te). THAM is availa le as a 0.3 M solution (0.3 mEq/mL) and is given according t o the ase deficit: THAM (mEq/L) = 0.3 ody weight (kg) ase deficit. THAM is ex creted in the urine and should e avoided in patients with renal failure. This uffer may prove eneficial in protecting the rain from local tissue acidosis du ring and after cardiac arrest (31). However, the clinical experience with THAM i s limited at present. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 20)03-May-05 08:08:38

Ovid: ICU Book SUMMARY Alkali therapy has a limited role in the management of patients with lactic acid osis. First, it is unclear if acidosis is detrimental. Furthermore, therapy with sodium icar onate has limited efficacy in raising the pH, and it produces an u ndesira le increase in the PCO2 of the ody fluids. Buffers that do not generate CO2 are availa le, ut their clinical value is unproven. Recommendation If severe lactic acidosis develops (pH elow 7.1) and the patient is deteriorati ng, a trial infusion of icar onate can e attempted y administering one-half o f the estimated icar onate deficit. If cardiovascular improvement occurs, icar onate therapy can e continued. If no improvement or further deterioration occu rs, further icar onate administration is not warranted.

KETOSIS In conditions of reduced nutrient intake, adipose tissue releases free fatty aci ds, which are then taken up in the liver and meta olized to form the ketones ace toacetate and -hydroxy utyrate. These ketones are released from the liver and can e used as oxidative fuels y vital organs such as the heart and central nervou s system. The oxidative meta olism of ketones yields 4 kcal/ g, which is a great er energy yield than the 3.4 kcal/g produced y car ohydrate meta olism (see Cha pter 46). The normal concentration of ketones in the lood is negligi le (less t han 0.1 mmol/L), ut after 3 days of starvation, lood ketone levels increase te nfold. Ketones are strong acids, and thus progressive ketosis eventually produce s a meta olic acidosis. The prevalence of acetoacetate (AcAc) and -hydroxy utyrat e (BOHB) in lood is determined y the following redox reaction: The alance of this reaction favors the formation of -hydroxy utyrate. P.601 In c onditions of enhanced ketone production, the BOHB:AcAc ratio ranges from 3:1 (in dia etic ketoacidosis) to as high as 8:1 (in alcoholic ketoacidosis). The conce ntration of ketones in the lood in dia etic and alcoholic ketoacidosis is shown in Figure 37.4. Note the preponderance of hydroxy utyrate in oth conditions. Be cause of this preponderance, ketoacidosis is more accurately called -hydroxy utyr ic acidosis. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 20)03-May-05 08:08:38

Ovid: ICU Book Figure 37.4. The concentrations of acetoacetate and -hydroxy utyrate in the lood in dia etic ketoacidosis (DKA) and alcoholic ketoacidosis (AKA). The horizontal hatched line represents the minimum concentration of acetoacetate required to p roduce a positive nitroprusside reaction. NITROPRUSSIDE REACTION The nitroprusside reaction is a colorimetric method for detecting acetoacetate a nd acetone in lood and urine. The test can e performed with ta lets (Acetest) or reagent strips (Ketostix, La stix, Multistix). A detecta le reaction requires a minimum acetoacetate concentration of 3 mEq/L. Because this reaction does not detect the predominant ketoacid -hydroxy utyrate (32), it is an insensitive meth od for monitoring the severity of ketoacidosis. This is illustrated in Figure 37 .4. In alcoholic ketoacidosis, the total concentration of ketoacids in lood is 13 mEq/L, which represents more than a hundredfold increase over the normal conc entration of lood ketones, yet the nitroprusside reaction will e negative eca use the acetoacetate concentration is elow 3 mEq/L. While discussing the nitrop russide reaction, it may e worth mentioning that acetylcysteine can produce a f alsely-negative nitroprusside P.602 reaction (33). The clinical significance of this interaction seems marginal; however, the antioxidant properties of acetylcy steine (see Chapter 3) may lead to more common use of this http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 20)03-May-05 08:08:38

Ovid: ICU Book agent in the ICU. DIABETIC KETOACIDOSIS Dia etic ketoacidosis (DKA) is usually seen in insulin-dependent dia etic patien ts, ut in 20% of cases, there is no previous history of dia etes mellitus (34). DKA is most often the result of inappropriate insulin dosing, ut 50% of patien ts can have a concurrent illness, most commonly an infection (34). CLINICAL FEATURES The hallmark of DKA is the com ination of hyperglycemia and a serum icar onate level elow 20 mEq/L. The lood glucose is usually a ove 250 mg/dL, ut may not e a ove 350 mg/dL. No correlation exists etween the severity of the hyperglyce mia and the severity of the ketoacidosis (35). Semiquantitative methods for dete cting ketones (acetoacetate) in lood and urine will e positive; however, as me ntioned earlier, these methods underestimate the degree of ketonemia. Anion Gap The increase in ketoacids should produce an elevated anion gap; however, this is varia le, and the anion gap can e normal in DKA (36). The renal excretion of k etones is accompanied y an increase in chloride rea sorption in the renal tu ul es, and the resulting hyperchloremia limits the increase in the anion gap. MANAGEMENT The management of DKA is summarized in Ta le 37.2. The following are some of the details. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 20)03-May-05 08:08:38

Ovid: ICU Book TABLE 37.2. MANAGEMENT OF DIABETIC KETOACIDOSIS P.603 Insulin Insulin therapy is given intravenously, starting with a olus dose of 0.1 units per kilogram ody weight (37). Because intravenous insulin has a half-life of on ly 5 minutes, it must e given y continuous infusion. The initial dose rate is 0.1 U/kg/hour. The lood glucose should fall 10% in the first hour of therapy. I f it does not, the insulin dose rate should e dou led. The lood glucose levels should e measured every 1 to 2 hours during intravenous insulin therapy. Finge rstick glucose determinations can e performed if the lood glucose is elow 500 mg/dL (34). Fluids Volume deficits average 50 to 100 mL/kg (or 4 to 8 L for a 175-l adult). If no evidence of hypovolemic shock exists, crystalloid fluids are appropriate for vol ume replacement. Fluid therapy egins with 0.9% (isotonic) saline infused at a r ate of approximately 1 L/hour for the first 2 hours. This is followed y infusio n of 0.45% (half-normal) saline at 250 to 500 mL/hour. When the lood glucose fa lls to 250 mg/dL, dextrose can e used in the intravenous fluids and the infusio n rate is dropped to 100 to 250 mL/hour. If evidence of hypovolemia shock (e.g., hypotension, reduced urine output) does exist, fluid replacement should egin w ith colloid fluids (5% al umin or 6% hetastarch). Colloid resuscitation is given at one-third the volume recommended for crystalloid fluid replacement. Potassium http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 20)03-May-05 08:08:38

Ovid: ICU Book Potassium depletion is almost universal in DKA, and the average deficit is 3 to 5 mEq/kg. However, the serum potassium is often normal (74% of patients) or elev ated (22% of patients) at presentation. The serum potassium falls during insulin therapy (transcellular shift), and this fall can e dramatic. Therefore, potass ium replacement should e started as soon as possi le (Ta le 37.2), and the seru m potassium should e monitored hourly for the first 4 to 6 hours of therapy. Phosphate Phosphorous depletion is also common in DKA and averages 1 to 1.5 mmol/kg. Howev er, phosphorous replacement seems to have little impact on the outcome in DKA, a nd therefore phosphate replacement is not recommended routinely (34). The serum phosphate level should e measured 4 hours after the start of therapy. If the le vel is severely depressed (less than 1 mg/ dL), phosphate replacement is advised (see Ta le 37.1 for the recommended replacement dose). Alkali Therapy Bicar onate therapy does not improve the outcome in DKA, and ecause of the disa dvantages of alkali therapy descri ed earlier, icar onate P.604 therapy is not recommended in DKA, regardless of the severity of the acidemia (34). MONITORING The serum icar onate level may not e a relia le parameter for following the co urse of the ketoacidosis. Fluid replacement therapy often produces a hyperchlore mic acidosis y promoting ketoacid excretion in the urine, which increases chlor ide rea sorption in the renal tu ules. This can keep the icar onate from rising despite a resolving ketoacidosis. In this situation, the pattern of the acidosi s is changing (i.e., changing from a high anion gap to a low anion gap acidosis) . Therefore, monitoring the pattern of the acidosis as therapy proceeds can e m ore informative. This is accomplished y monitoring the anion gap excess : icar onate deficit ratio, as descri ed in Chapter 36. This ratio is 1.0 in pure keto acidosis, and decreases toward zero as the ketoacidosis resolves and is replaced y the hyperchloremic acidosis. When the ketones have een cleared from the lo odstream, the ratio approaches zero. ALCOHOLIC KETOACIDOSIS Alcoholic ketoacidosis (AKA) is a complex acid- ase disorder that occurs in chro nic alcoholics and usually appears 1 to 3 days after a period of heavy inge dri nking (38). Several mechanisms seem to e involved in the ketosis, including red uced nutrient intake (which initiates enhanced http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 20)03-May-05 08:08:38

Ovid: ICU Book ketone production), hepatic oxidation of ethanol (which generates NADH and enhan ces hydroxy utyrate formation), and dehydration (which impairs ketone excretion i n the urine). CLINICAL FEATURES Patients with AKA tend to e chronically ill and have several concurrent disorde rs (e.g., pancreatitis, upper gastrointestinal leeding, hepatitis, ethanol with drawal, seizures). The presentation usually includes nausea, vomiting, and a dom inal pain (38). Electrolyte a normalities are common, particularly the hypos (e. g., hyponatremia, hypokalemia, hypophosphatemia, hypomagnesemia, hypoglycemia). Mixed acid- ase disorders are also common in AKA. More than half the patients ca n have lactic acidosis (caused y other concurrent conditions), and a hyperchlor emic acidosis is common (possi ly caused y ketone excretion in the urine). A me ta olic alkalosis is also seen in patients with protracted vomiting. DIAGNOSIS The diagnosis of AKA is suggested y the clinical setting (i.e., after a period of inge drinking), an elevated anion gap, and the P.605 presence of ketones in the lood or urine. However, the nitroprusside reaction for detecting ketones ca n e negative in AKA. The oxidation of ethanol in the liver generates NADH, and this favors the conversion of acetoacetate to -hydroxy utyrate. As a result, the acetoacetate in lood and urine are low and may e elow the threshold for keton e detection y the nitroprusside reaction, as shown in Figure 37.4. However most cases of AKA have a positive nitroprusside reaction for ketones (38). MANAGEMENT The management of AKA is nota le for its simplicity. Infusion of dextrose-contai ning saline solutions is all that is required. The glucose helps retard hepatic ketone production, while the infused volume promotes the renal clearance of keto nes. The ketoacidosis usually resolves within 24 hours. Other electrolyte defici encies are corrected as needed. Bicar onate therapy is unnecessary (38). REFERENCES SUGGESTED READINGS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 20)03-May-05 08:08:38

Ovid: ICU Book Cohn RM, Roth KS. Biochemistry and disease. Baltimore: Williams & Wilkins, 1996. Rose BD. Clinical physiology of acid- ase and electrolyte disorders. 4th ed. New York: McGraw-Hill, 1994. GENERAL REVIEWS 1. Laski ME, Kurtzman NA. Acid- ase disorders in medicine. Dis Mon 1996;XLII:5712 8 (200 references). LACTIC ACIDOSIS: REVIEWS 2. Mizock BA. Lactic acidosis. Dis Mon 1989;XXXV:235300 (322 references). 3. Mizock BA, Falk JL. Lactic acidosis in critical illness. Crit Care Med 1992;2 0:8093 (144 references).

LACTIC ACIDOSIS: SELECTED REFERENCES 5. A erti KGMM, Cuth ert C. The hydrogen ion in normal meta olism: a review. CIB A Foundation Symposium 87. Meta olic acidosis. London: Pitman Books, 1982;115. 6. Lehninger AL. Bioenergetics. New York: WA Benjamin, 1965;16. P.606 7. Brooks GA. Lactate production under fully aero ic conditions: the lactate shu ttle during rest and exercise. Fed Proc 1986;45:29242929. 8. Weil MH, Afifi AA. Experimental and clinical studies on lactate and pyruvate as indicators of the severity of acute circulatory failure (shock). Circulation 1970;16:9891001. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 20)03-May-05 08:08:38

4. Stacpoole PW. Lactic acidosis. Endocrinol Meta

Clin North Am 1993;22:221245.

Ovid: ICU Book 9. Stacpoole PW, Wright EC, Baumgartner TG, et al. Natural history of acquired l actic acidosis in adults. Am J Med 1994;97:4754. 10. Aduen J, Bernstein WK, Khastgir T, et al. The use and clinical importance of a su strate-specific electrode for rapid determinations of lood lactate concen trations. JAMA 1994;272:16781684. 11. Curtis SE, Cain SM. Regional and systemic oxygen delivery/uptake relations a nd lactate flux in hyperdynamic, endotoxin-treated dogs. Am Rev Respir Dis 1992; 145:348 354. 12. Chattha G, Arieff AI, Cummings C, Tierney LM. Lactic acidosis complicating t he acquired immunodeficiency syndrome. Ann Intern Med 1993;118:3739. 13. Camp ell CH. The severe lactic acidosis of thiamine deficiency: acute, perni cious or fulminating eri eri. Lancet 1984;1:446449. 14. Bersin RM, Arieff AI. Primary lactic alkalosis. Am J Med 1988;85:867871. 15. Kruse JA, Zaidi SAJ, Carlson RW. Significance of lood lactate levels in cri tically ill patients with liver disease. Am J Med 1987;83:7782. 16. Brivet F, Bernadin M, Cherin P, et al. Hyperchloremic acidosis during grand mal seizure acidosis. Intensive Care Med 1994;20:2731. 17. Mountain RD, Heffner JE, Brackett NC, Sahn SA. Acid- ase distur ances in acu te asthma. Chest 1990;98:651655. 18. I erti TS, Lie owitz AB, Papadakos PJ, et al. Low sensitivity of the anion g ap as a screen to detect hyperlactatemia in critically ill patients. Crit Care M ed 1990;18:275277. LACTIC ACIDOSIS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 20)03-May-05 08:08:38

Ovid: ICU Book 19. Anonymous. The colon, the rumen, and d-lactic acidosis. Lancet 1990;336:59960 0 (editorial). 20. Thurn JR, Pierpoint GL, Ludvigsen CW, Eckfeldt JH. D-lactate encephalopathy. Am J Med 1985;79:717720. 21. Bustos D, Ponse S, Pernas JC et al. Fecal lactate and the short owel syndro me. Dig Dis Sci 1994;39:23152319. ALKALI THERAPY 22. Bie uyck JF. Sodium icar onate in the treatment of su types of acute lactic acidosis: physiologic considerations. Anesthesiology 1990;72:10641076. 23. Sonnett J, Pagani FD, Baker LS, et al. Correction of intramyocardial hyperca r ic acidosis with sodium icar onate. Circ Shock 1994;42:163173. 24. Mehta PM, Kloner RA. Effects of acid- ase distur ance, septic shock, and cal cium and phosphorous a normalities on cardiovascular function. Crit Care Clin 19 87;3:747758. 25. Gores GJ, Nieminen AL, Fleischman KE, et al. Extracellular acidosis delays o nset of cell death in ATP-depleted hepatocytes. Am J Physiol 1988;255:C315C322. 26. Rose BD. Clinical physiology of acid- ase and electrolyte disorders. 4th ed. New York: McGraw-Hill, 1994;590. 27. Graf H, Arieff AI. The use of sodium icar onate in the therapy of organic a cidoses. Intensive Care Med 1986;12:286288. P.607 28. Comroe JH. Physiology of respiration. Chicago: Year ook Medical Pu lishers, 1974;203. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 20)03-May-05 08:08:38

Ovid: ICU Book 29. Rhee KY, Toro LO, McDonald GG, et al. Car icar , sodium icar onate, and sod ium chloride in hypoxic lactic acidosis. Chest 1993;104:913918. 30. Benjamin E, Oropello JM, A alos A, et al. Effects of acid- ase correction on hemodynamics, oxygen dynamics, and resuscita ility in severe canine hemorrhagic shock. Crit Care Med 1994;22:161616723. 31. Rosen erg JM, Martin GB, Paradis NA, et al. The effect of CO2 and nonCO2gener ating uffers on cere ral acidosis after cardiac arrest: a 31P NMR study. Ann Em erg Med 1989;18:341347. DIABETIC KETOACIDOSIS 32. Umpierrez GE, Kita achi AE. A rational approach to diagnosing dia etic ketoa cidosis. J Crit Illness 1996;11:428432. 33. Holcom e BJ, Messick CR. Drug:la interactions: implications for nutrition s upport. Nutr Clin Pract 1994;9:196198. 34. Fish LH. Dia etic ketoacidosis. Postgrad Med 1994;96:7596. 35. Brandt KR, Miles JM. Relationship etween severity of hyperglycemia and meta olic acidosis in dia etic ketoacidosis. Mayo Clin Proc 1988;63:10711074.

36. Gam lin GT, Ash urn RW, Kemp DG, Beuttel SC. Dia etic ketoacidosis presentin g with a normal anion gap. Am J Med 1986;80:758760. 37. Umpierrez GE, Kita achi AE. Management strategies for dia etic ketoacidosis. J Crit Illness 1996;11:437443. ALCOHOLIC KETOACIDOSIS 38. Wrenn KD, Slovis CM, Minion GE, Rutkowsli R. The syndrome of alcoholic http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 20)03-May-05 08:08:38

Ovid: ICU Book ketoacidosis. Am J Med 1991;91:119128. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 20)03-May-05 08:08:38

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 38 METABOLIC ALKALOSIS Meta olic alkalosis does not have the notoriety of meta olic acidosis, yet it is the most common acid- ase disorder in hospitalized patients (1). One of the fea tures of meta olic alkalosis that makes it so prevalent is its self-sustaining a ility after the inciting condition resolves. This is attri uted to chloride dep letion, which enhances icar onate rea sorption in the renal tu ules (to maintai n electrical neutrality in the extracellular fluids). In fact, chloride depletio n and repletion are the yin and yang of meta olic alkalosis in hospitalized pati ents.

NASOGASTRIC SUCTION Each milliequivalent of hydrogen ion (H+) that is secreted y the gastric mucosa generates 1 mEq of icar onate (HCO3). The gastric juice has a H+ concentration of 50 to 100 mEq/L, so a considera le quantity of HCO3 is generated y gastric acid secretion. However, this is not normally accompanied y an increase in plas ma HCO3 ecause the acidic gastric juice also stimulates pancreatic secretion of HCO3, which counter alances the HCO3 generated y gastric acid secretion. The l oss of HCO3 in pancreatic secretions is eliminated when gastric acid is removed y vomiting or nasogastric suction, and this allows the plasma HCO3 to rise. Add itional factors that promote meta olic alkalosis during nasogastric suction are loss of potassium and hypovolemia.

DIURETICS Diuretics promote alkalosis y increasing the urinary loss of electrolytes and f ree water. The following electrolytes are involved: P.609 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 12)03-May-05 08:09:04

COMMON ETIOLOGIES The following is a (1,2).

rief description of the common causes of meta olic alkalosis

Ovid: ICU Book q Chloride excretion is enhanced y diuretics ecause chloride follows the sodium that is lost in the urine. To maintain electrical neutrality in the extracellula r fluid, the urinary chloride loss is counter alanced y an increase in HCO3 rea sorption across the renal tu ules. This results in an increase in extracellular (plasma) HCO3 concentration. q

Potassium loss in the urine is enhanced y diuretics ecause sodium delivery to the distal tu ules is increased, and this promotes potassium secretion via the s odium potassium exchange pump in the distal tu ule. The resultant potassium deple tion can then produce an extracellular alkalosis y promoting the intracellular accumulation of H (transcellular H+K+ exchange) and y promoting the secretion of H+ in the distal renal tu ules. q + Magnesium is also lost in the urine during diuresis, and this promotes potassium depletion through unclear mechanisms. Magnesium depletion is an important facto r in diuretic-induced potassium depletion, as discussed in Chapter 42. VOLUME DEPLETION Volume depletion can promote meta olic alkalosis in two ways. First, loss of fre e water concentrates the existing HCO3 stores and raises the HCO3 concentration in extracellular fluids. Second, a decrease in circulating lood volume stimulat es the reninangiotensin aldosterone axis, and the aldosterone release promotes the loss of potassium and hydrogen ions in the distal tu ule. The importance of vol ume depletion as a cause of meta olic alkalosis is unclear ecause fluid losses are almost always accompanied y electrolyte losses and the latter condition cou ld promote the alkalosis, as descri ed previously. In fact, volume depletion can e accompanied y a decrease in the glomerular filtration rate (GFR), and this should promote a meta olic acidosis (as occurs in renal failure). ORGANIC ANIONS The administration of organic anions such as lactate (in lactated Ringer's solut ion), acetate (in parenteral nutrition solutions), and citrate (in anked lood) could produce a meta olic alkalosis. Of these, only citrate administration in lood transfusions seems to e capa le of causing a meta olic alkalosis (3). Howe ver, at least 8 units of lood must e transfused efore http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 12)03-May-05 08:09:04

Ovid: ICU Book the plasma HCO3 egins to rise (2). POSTHYPERCAPNIA The compensatory response to CO2 retention is a decrease in renal icar onate ex cretion, which produces a compensatory meta olic alkalosis. If chronic hypercapn ia is corrected acutely (e.g., y overventilation P.610 during mechanical ventil ation), the compensatory meta olic alkalosis will ecome a primary acid- ase dis order. However, this should e a transient condition ecause the decrease in art erial PCO2 promotes renal icar onate excretion. ADVERSE EFFECTS Meta olic alkalosis has several potential adverse effects. The following ones de serve mention, although their clinical significance is unclear. NEUROLOGIC MANIFESTATIONS The neurologic manifestations attri uted to alkalosis include depressed consciou sness, generalized seizures, and carpopedal spasms. However, these manifestation s are almost always associated with respiratory alkalosis, not meta olic alkalos is. This is explained y the greater tendency for respiratory alkalosis to influ ence the acid- ase status of the central nervous system. HYPOVENTILATION Meta olic alkalosis can depress ventilation, and this has een mentioned as a co ntri uting factor in failed weaning from mechanical ventilation (1). P.611 Howev er, as mentioned in Chapter 36, the ventilatory depression in meta olic alkalosi s is a varia le and often minor effect. The lack of hypoventilation in meta olic alkalosis is easily demonstrated in patients who receive chronic diuretic thera py (i.e., most of these patients have a meta olic alkalosis, ut few have CO2 re tention). The expected arterial PCO2 at any given serum HCO3 concentration is de scri ed y the equation shown in Ta le 36.2 (4). This equation is used to plot t he relationship etween serum HCO3 and arterial PCO2 shown in Figure 38.1. As sh own in the graph, hypercapnia is not expected until the serum HCO3 rises http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 12)03-May-05 08:09:04

Ovid: ICU Book a ove 30 to 35 mEq/L. Figure 38.1. The relationship etween serum icar onate (HCO3) and arterial PCO2 (PaCO2) as predicted y the equation shown at the top of the graph. Note that t he serum HCO3 must rise a ove 30 to 35 mEq/L to produce hypercapnia (i.e., PaCO2 a ove 44 mm Hg). The relative lack of ventilatory depression in meta olic alkalosis may e a refl ection of the aseline activity in the peripheral chemoreceptors, where meta oli c acid- ase disorders exert most of their influence on ventilation. That is, the peripheral chemoreceptors show minimal activity under normal conditions (5), so there is little aseline activity for alkalosis to inhi it. TISSUE OXYGENATION Meta olic alkalosis has several effects that could impair tissue oxygenation. Th ese are demonstrated in Figure 38.2. Alkalosis can decrease tissue oxygen delive ry in two ways. First, alkalosis increases the fraction of serum calcium that is ound to al umin, and the resultant decrease in ionized (free) calcium can impa ir myocardial contractility P.612 and reduce cardiac output. Second, alkalosis s hifts the oxyhemoglo in dissociation curve to the left (Bohr effect), which decr eases the tendency for hemoglo in to release oxygen in the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 12)03-May-05 08:09:04

Ovid: ICU Book tissue capillaries. In addition, intracellular alkalosis stimulates glycolysis, which can increase tissue oxygen requirements (6). Therefore, meta olic alkalosi s can decrease tissue oxygen availa ility while increasing tissue oxygen demands . Although the effects of meta olic alkalosis on tissue oxygenation are unknown ( ecause it is impossi le to directly monitor tissue oxygenation), the potential for impaired tissue oxygenation is an important consideration in patients with shock (in whom tissue oxygenation is already impaired). Figure 38.2. Effects of meta olic alkalosis on the determinants of tissue oxygen ation. EVALUATION The cause of a meta olic alkalosis is often readily apparent (e.g., nasogastric suction, diuretics). However, the urinary chloride concentration can also e use ful in identifying the source of a meta olic alkalosis. As shown in Ta le 38.1, the meta olic alkaloses can e classified as chloride-responsive or chloride-res istant ased on the urinary chloride concentration. The only exception to this r ule is in the early stages of diuretic therapy, when the urinary chloride concen tration is elevated ut the resulting meta olic alkalosis is chlorideresponsive. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 12)03-May-05 08:09:04

Ovid: ICU Book TABLE 38.1. CLASSIFICATION OF METABOLIC ALKALOSIS CHLORIDE-RESPONSIVE ALKALOSIS A chloride-responsive meta olic alkalosis is characterized y a low urinary chlo ride concentration (i.e., less than 15 mEq/L), indicating chloride depletion. Th is type of meta olic alkalosis is the result of gastric acid loss, diuretic ther apy, volume depletion, or renal compensation for hypercapnia. As indicated y th e inciting conditions, volume depletion is common in chloride-responsive meta ol ic alkalosis. The majority of cases of meta olic alkalosis in hospitalized patie nts are the chloride-responsive variety. CHLORIDE-RESISTANT ALKALOSIS A chloride-resistant meta olic alkalosis is characterized y an elevated urinary chloride concentration (i.e., a ove 25 mEq/L). Most P.613 cases of chloride-res istant alkalosis are caused y mineralocorticoid excess (e.g., hyperadrenal cond itions) or potassium depletion (these two conditions often co-exist). This type of meta olic alkalosis is characterized y volume expansion rather than volume d epletion. The disorders associated with chloride-resistant alkalosis are not com monly seen in patients in the ICU, with the possi le exception of aggressive cor ticosteroid therapy. SPOT URINE CHLORIDE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 12)03-May-05 08:09:04

Ovid: ICU Book When the cause of a meta olic alkalosis is unclear, the concentration of chlorid e in a random (spot) urine sample can help identify the possi le sources of the pro lem. One source of error occurs in the early stages of diuretic therapy, whe n the urinary chloride concentration is elevated in a chloride-responsive meta o lic alkalosis. Another enefit of measuring the spot urinary chloride concentrat ion is in selecting the appropriate therapy to correct the alkalosis. These are descri ed elow. MANAGEMENT Because most meta olic alkaloses that occur in hospitalized patients are chlorid e-responsive, chloride replacement is the mainstay of therapy for meta olic alka losis in most patients. The chloride can e replaced as sodium chloride, potassi um chloride, or hydrochloric acid (HCl). SALINE INFUSION Because volume depletion is common in chloride-responsive meta olic alkalosis, i nfusion of isotonic saline (0.9% sodium chloride) is the most common method of c hloride replacement in this condition. Method The volume of isotonic saline needed can e determined y estimating the chlorid e deficit, as shown in Ta le 38.2. For example, if a patient weighs 80 kg (175 l ) and the serum chloride concentration is 80 mEq/L, the chloride deficit will e 0.3 80 20 = 480 mEq. The chloride P.614 concentration in isotonic saline is 15 4 mEq/L, so the volume of saline needed to replace the chloride deficit will e 480/154 = 3.1 L. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 12)03-May-05 08:09:04

Ovid: ICU Book TABLE 38.2. SALINE INFUSION FOR METABOLIC ALKALOSIS POTASSIUM CHLORIDE The administration of potassium chloride is not an effective method of chloride repletion ecause the maximum rate of potassium infusion that is safe is 40 mEq/ hour (see Chapter 41). Therefore, potassium chloride administration is indicated only for patients who are hypokalemic. However, ecause hypokalemia can promote meta olic alkalosis, correcting hypokalemia is an important measure for correct ing a meta olic alkalosis. It is important to emphasize that the administration of potassium chloride will not replenish potassium stores if there is concurrent magnesium depletion (7). Therefore, it is important to identify and correct mag nesium depletion efore attempting to replace potassium deficits (see Chapter 42 for information on identifying and correcting magnesium deficiency). HYDROCHLORIC ACID Infusion of dilute solutions of HCl produces the most rapid correction of meta o lic alkalosis (1). This method can e used when rapid correction of severe alkal emia (pH greater than 7.5) is desira le, particularly when the cause of the alka losis is loss of gastric acid. Method The amount of HCl required to correct a meta olic alkalosis can e determined y estimating the hydrogen ion (H+) deficit, as shown in Ta le 38.3. A plasma HCO3 concentration of 35 mEq/L can e used as the desired plasma HCO3 level ecause immediate correction of the plasma HCO3 to normal is rarely necessary. Once the H+ deficit is calculated, the corresponding volume and infusion rate of the HCl solution can e determined. A popular solution is 0.1N HCl, which contains 100 m Eq of H+ per liter (equivalent to the H+ concentration in gastric secretions, wh ich is 50 to 100 mEq/L). Because HCl solutions are corrosive, they must e infus ed through a large, central vein (8). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 12)03-May-05 08:09:04

Ovid: ICU Book TABLE 38.3. HYDROCHLORIC ACID INFUSION FOR METABOLIC ALKALOSIS P.615 To illustrate the use of the equations in Ta le 38.3, consider a patient w ho weighs 80 kg and has a plasma HCO3 of 50 mEq/L. Using a desired plasma HCO3 o f 35 mEq/L, one can calculate the H+ deficit as follows: 0.3 80 (50 ms 35) = 360 mEq. The corresponding volume of 0.1N is 360/100 = 3.6 L. The infusion rate of the 0.1N HCl solution is (0.2 80)/100 = 0.16 L/hour, or 2.6 ml/minute. Adverse Effects The major concern with HCl infusions is the corrosive effects of the HCl solutio ns. Extravasation of HCl solutions can produce severe tissue necrosis, even when the solution is infused through a central vein (9). Solutions more concentrated than 0.1N HCl can also promote corrosion of intravascular catheters (10). To li mit the risk for corrosion, HCl solutions that are more concentrated than the 0. 1N solution should e avoided. OTHER CHLORIDE SOLUTIONS Ammonium chloride can e converted to ammonia and HCl in the liver, and this can correct a meta olic alkalosis indirectly. However, ammonium chloride infusions are discouraged ecause the ammonia that is generated can produce an encephalopa thy (with o tundation and coma), particularly in patients with hepatic or renal insufficiency (1,2). Arginine hydrochloride is also roken down to arginine and HCl in the liver. However, the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 12)03-May-05 08:09:04

Ovid: ICU Book movement of arginine (a cation) into cells is coupled with the movement of potas sium out of cells, and this can lead to severe and life-threatening hyperkalemia (1,2). HISTAMINE H2 BLOCKERS Inhi ition of gastric acid secretion with histamine H2 receptor antagonists has een used to prevent meta olic alkalosis during prolonged periods of nasogastric suctioning (1,2). However, considering the consequences of gastric acid suppres sion on colonization of the upper gastrointestinal tract (descri ed in Chapter 6 and Chapter 33), this approach should e discouraged.

CHLORIDE-RESISTANT ALKALOSIS The management of chloride-resistant meta olic alkalosis is aimed at treating th e underlying cause of the mineralocorticoid excess (e.g., hyperadrenalism, renal artery stenosis) and correcting potassium deficits. Because this type of meta o lic alkalosis usually is accompanied y extracellular volume expansion (not volu me contraction), the diuretic corrective measures descri ed elow can also e us ed. P.616 Acetazolamide Acetazolamide (Diamox) is a car onic anhydrase inhi itor that locks HCO3 rea so rption in the proximal renal tu ules and promotes urinary HCO3 excretion. Acetaz olamide has diuretic effects, and ecause chloride-resistant meta olic alkalosis usually is accompanied y a high extravascular volume, acetazolamide is well-su ited for correcting chloride-resistant meta olic alkaloses. The recommended dose is 5 to 10 mg/kg IV (or PO), and the maximum effect is seen after an average of 15 hours (11). Acetazolamide promotes potassium depletion as well as volume dep letion, and thus it should not e used in chloride-resistant meta olic alkalosis that is accompanied y hypokalemia. REFERENCES SUGGESTED READING Rose BD. Clinical physiology of acid- ase and electrolyte disorders. 4th ed. New York: McGraw-Hill, 1994. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 12)03-May-05 08:09:04

Ovid: ICU Book REVIEWS 1. Friedman BS, Lum PD. Prevention and management of meta olic alkalosis. J Int ensive Care Med 1990;5(Suppl):S22S27. 2. Rose BD. Meta olic alkalosis. In: Clinical physiology of acid- ase and electr olyte disorders. 4th ed. New York: McGraw-Hill, 1994;515539. SELECTED REFERENCES 3. Driscoll DF, Bistrian BR, Jenkins RL. Development of meta olic alkalosis afte r massive transfusion during orthotopic liver transplantation. Crit Care Med 198 7;15:905 908. 4. Javaheri S, Kazemi H. Meta olic alkalosis and hypoventilation in humans. Am R ev Respir Dis 1987;136:10111016. 5. Marino PL. Brainstem chemoreception. Ann Ar or: University of Michigan Press, 1974. 6. Rastegar HR, Woods M, Harken AH. Respiratory alkalosis increases tissue oxyge n demand. J Surg Res 1979;26:687692. 7. Whang R, Flink EB, Dyckner T, et al. Mg depletion as a cause of refractory po tassium depletion. Arch Intern Med 1985;145:16861689. 8. Brimioulle S, Vincent JL, Dufaye P, et al. Hydrochloric acid infusion for tre atment of meta olic alkalosis: effects on acid- ase alance and oxygenation. Cri t Care Med 1985;13:738742. 9. Jankauskas SJ, Gursel E, Antonenko DR. Chest wall necrosis secondary to hydro chloric acid use in the treatment of meta olic alkalosis. Crit Care Med 1989;17: 963 964. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 12)03-May-05 08:09:04

Ovid: ICU Book 10. Kopel R, Dur in CG. Pulmonary artery catheter deterioration during hydrochlo ric infusion for the treatment of meta olic alkalosis. Crit Care Med 1989;17:6886 89. 11. Marik PE, Kussman BD, Lipman J, Kraus P. Acetazolamide in the treatment of m eta olic alkalosis in critically ill patients. Heart Lung 1991;20:455458. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 12)03-May-05 08:09:04

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 39 ACUTE OLIGURIA Lack of urine output in the acutely hypovolemic patient is renal success, not re nal failure. --Ronald V. Maier, MD An acute decrease in urine output can represe nt a functional adaptation, as Dr. Maier points out, ut more often it represent s trou le. The trou le is acute renal failure, which can have a mortality of 80% in critically ill patients (1,2). Furthermore, the introduction of acute hemodi alysis has not reduced the mortality in acute renal failure (1,2), which has ser ious implications. Let's hope this information does not fall into the hands of t he outcomes analysts, who elieve that an intervention is not justified if it does not improve the outcome. OVERVIEW Oliguria (urine output less than 400 mL/day) may not e associated with a normal renal function (4). This condition (which can e caused y antidiuretic hormone [ADH] excess) differs from the oliguria descri ed in this chapter, which should e accompanied y any of the following: q An increase in serum creatinine of at least 0.5 mg/dL a ove aseline q An increase in serum creatinine of at least 50% over aseline q A reduction in calculated creatinine clearance of at least 50% q Severe renal dysfunction requiring some form of renal replacement therapy http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 15)03-May-05 07:47:23

Ovid: ICU Book P.620 The creatinine clearance can e estimated using the formulas shown in Ta l e 39.1 (5). This helps not only to determine the degree of renal functional impa irment, ut also to determine appropriate drug dosages. TABLE 39.1. QUANTITATIVE ASSESSMENT OF RENAL FUNCTION The four criteria listed a ove identify the condition known as acute oliguric re nal failure (AORF). The causes of AORF are traditionally separated into three ca tegories, as shown in Figure 39.1 (1,6). Each category is named according to the anatomic location of the pro lem responsi le for the oliguria. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 15)03-May-05 07:47:23

Ovid: ICU Book Figure 39.1. Causes of acute oliguria ased on anatomic location of the pro lem. PRERENAL DISORDERS The prerenal sources of AORF are located proximal to the kidneys and are charact erized y a decrease in renovascular flow. The disorders in this category includ e hypovolemia, severe cardiac dysfunction, loss of vascular tone, drugs that pro mote renal vasoconstriction (e.g., nonsteroidal antiinflammatory agents), and dr ugs that reduce glomerular filtration pressure (e. g., angiotensin-converting en zyme inhi itors). Prerenal disorders are responsi le for roughly half of the cas es of AORF (1,2 and 3). INTRINSIC RENAL DISORDERS The intrinsic renal disorders involve the renal parenchyma and are characterized y impaired glomerular filtration, renal tu ular dysfunction, or oth. These di sorders are traditionally descri ed as three entities: acute glomerulonephritis, acute tu ular necrosis (ATN), and acute interstitial nephritis (AIN). ATN is th e most common intrinsic renal disorder in AORF. It is most often caused y sepsi s, circulatory shock, and nephrotoxins (6). The toxins include drugs (e.g., amin oglycosides), radiocontrast dye, and pigments (e.g., myoglo in). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 15)03-May-05 07:47:23

Ovid: ICU Book P.621 Acute Tu ular Necrosis Acute renal failure is characterized y ineffective filtration across the glomer uli. However, in ATN (an unfortunate term ecause there may e no tu ular necros is), the pathologic defect does not involve the glomerulus, ut rather involves the renal tu ules and adjacent parenchyma (i.e., tu ulointerstitial injury). If the glomerulus is not involved, what shuts off the glomerular filtration rate (G FR)? This is explained in the photomicrograph in Figure 39.2. The center of this picture is a proximal tu ule that is filled with tu ular epithelial cells that have sloughed into the lumen of the tu ule. This mass of cells creates an o stru ction, and this produces a ack pressure that is exerted on the proximal to the o struction. This ack pressure then serves to reduce the net filtration pressur e across the glomerular capillaries. This tu ulo-glomerular feed ack pressure re duces the GFR. Figure 39.2. Photomicrograph of acute tu ular necrosis (ATN) showing a proximal tu ule filled with exfoliated epithelial cells (magnification 400). (From Racuse n LC. Histopathology of acute renal failure. New Horiz 1995;3:662668.) The damage to the renal tu ules in ATN is elieved to e the result of ischemia as well as inflammatory cell injury (7). The latter process can e widespread, a nd it can e accompanied y dysfunction in multiple P.622 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 15)03-May-05 07:47:23

Ovid: ICU Book organs (the multiorgan dysfunction syndrome, descri ed in Chapter 31). In this s ituation, the intrinsic renal failure is part of a more widespread systemic illn ess and is not an isolated organ injury. This is an important consideration eca use it means that the therapy for acute intrinsic renal failure should not e di rected solely at the kidneys (as it usually is). POSTRENAL DISORDERS The postrenal causes of AORF occur distal to the renal parenchyma and are charac terized y o struction of urinary flow. The disorder in this category can involv e o struction in the collecting system (papillary necrosis), the ureters (retrop eritoneal tumors), or the ladder outlet (strictures, prostatism). Postrenal o s truction is an uncommon cause of oliguria unless a solitary functional kidney is present. URINARY PARAMETERS The initial evaluation of the oliguric patient should include a microscopic and, if necessary, chemical analysis of the urine. The methods descri ed elow will help distinguish prerenal from intrinsic renal disorders. The urinary evaluation is less valua le for identifying postrenal o struction (except in papillary nec rosis, where sloughed papilla can e evident on urine microscopy). P.623 URINE MICROSCOPY Microscopic examination of the urine sediment is the easiest and least expensive diagnostic procedure. The presence of a undant tu ular epithelial cells with ep ithelial cell casts is virtually pathognomonic of ATN. In addition, the presence of white cell casts identifies an interstitial nephritis, and the presence of p igmented casts identifies myoglo inuria. Urine microscopy is not a valua le proc edure for identifying prerenal causes of oliguria. If the urine microscopy is un revealing, the urinary sodium determination can e useful. URINE SODIUM When renal perfusion is diminished, sodium rea sorption increases and urinary so dium excretion decreases. On the other hand, intrinsic renal failure is associat ed with impaired sodium rea sorption and an increase in urinary sodium excretion . Therefore, in the setting of oliguria, a urine sodium elow 20 mEq/L usually i ndicates a prerenal disorder. However, a urine sodium a ove 40 mEq/L does not ne cessarily rule out a prerenal disorder. An elevated http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 15)03-May-05 07:47:23

Ovid: ICU Book urine sodium can occur when a prerenal disorder is superimposed on intrinsic ren al dysfunction (with o ligatory sodium loss) or when there is ongoing diuretic t herapy. Elderly patients often have o ligatory sodium loss in the urine, and the y can have a high urine sodium in the face of prerenal disorders. Therefore, a u rine sodium a ove 40 mEq/L is not a solute evidence in favor of renal over prere nal disorders. FRACTIONAL EXCRETION OF SODIUM The fractional excretion of sodium the glomerulus that is excreted in learance divided y the creatinine Na is normally less than 1%; i.e., ed in the urine. In the setting of ormation: (FENa) is the fraction of sodium filtered at the urine. This is equivalent to the sodium c clearance, as shown in Ta le 39.1 (8). The FE less than 1% of the filtered sodium is excret oliguria, the FENa provides the following inf

Although exceptions to these criteria exist (i.e., the FENa can e less than 1% in an occasional case of intrinsic renal failure), the FENa is one of the most r elia le urinary parameters for distinguishing prerenal from renal causes of AORF . However, it is a cum ersome determination to perform, which limits its popular ity. BEDSIDE MANAGEMENT The principal task in the early management of the oliguric patient is to identif y and correct any prerenal disorders that might e responsi le P.624 for the oli guria. This is a simple task when invasive hemodynamic monitoring is availa le. OPTIMIZE CENTRAL HEMODYNAMICS The approach to optimizing central hemodynamics is very similar to the initial p art of the flow diagram shown in Figure 13.3. The first step is to determine the adequacy of cardiac filling. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 15)03-May-05 07:47:23

Ovid: ICU Book Figure 39.3. The technique of continuous arteriovenous hemofiltration (CAVH). Cardiac Filling Pressures The cardiac filling pressures (central venous pressure [CVP] and wedge pressures ) should e normal or unchanged (a decrease in the filling pressures in excess o f 4 mm Hg is considered a significant change). If the filling pressures are low (CVP less than 4 mm Hg, wedge pressure less than 8 mm Hg), then volume should e infused until the CVP increases to 6 to 8 mm Hg and the wedge pressure reaches 12 to 15 mm Hg. Once the filling pressures are in these ranges, the cardiac outp ut should e measured.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 15)03-May-05 07:47:23

Cardiac Output If the cardiac output is low, more volume should s 10 to 12 mm

e infused until the CVP reache

Ovid: ICU Book Hg and the wedge pressure is close to 20 mm Hg. If the cardiac output is still l ow despite filling pressures in the high-normal range, then the appropriate card iac support drug is selected using the lood pressure. If the lood pressure is normal, do utamine is given (start at 5 g/kg/ min) for inotropic support. If the lood pressure is low, dopamine is given (start at 5 g/kg/min) for pressure and f low support. The goal should e a cardiac index a ove 3 L/min/m2.

GLOMERULOTUBULAR FLOW If the oliguria persists despite adequate systemic pressure and flow, the likely source of the pro lem is intrinsic renal failure. In this situation, little els e can e done. The following two maneuvers are popular ut ineffective. LOW-DOSE DOPAMINE Despite thirty years of use as a renal vasodilator, low-dose dopamine (2 g/kg/min ) has shown no evidence of enefit when used in patients with acute oliguric ren al failure (9,10,11 and 12). A summary of the P.625 experience with dopamine in this setting is shown in Ta le 39.2. Although dopamine can augment renal lood f low and promote a diuresis, these effects are not accompanied y improved renal function or etter outcomes. The diuretic effect of dopamine is unrelated to its vascular effects, so the increase in urine output is not a reflection of an inc rease in GFR. TABLE 39.2. LOW-DOSE DOPAMINE IN AORF http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 15)03-May-05 07:47:23

Ovid: ICU Book In addition to eing ineffective, low-dose dopamine has also een identified as a potential risk ecause of the actions of dopamine as a splanchnic vasoconstric tor (which could predispose to owel ischemia) (12). Therefore, as a result of i ts unproven value and potential for harm, low-dose dopamine is no longer recomme nded for the management of acute oliguric renal failure (1,11). FUROSEMIDE The goal of diuretic administration in acute tu ular injury is to promote tu ula r flow and reduce the ack pressure that impedes glomerular filtration. However, furosemide has shown little value in promoting a diuresis in this setting. Cons idering that less than 10% of a furosemide dose reaches the lumen of the renal t u ules in renal failure (13), the lack of diuretic effect in renal failure is no t surprising. When furosemide is given y continuous intravenous infusion, the d iuretic effects are much greater than seen with olus drug administration (the d iuretic effect of furosemide is a function of the drug delivery time) (14). Ther efore, this may e the preferred method of attempted diuresis in acute renal fai lure. The optimal dose is unknown, ut the dose rate in clinical reports average s 1 to 9 mg/hour, although dose rates up to 0.75 mg/kg/ hour have een reported (14). There is no evidence that a loading dose is required efore the infusion. SPECIFIC RENAL DISORDERS DYE-INDUCED RENAL FAILURE Standard iodinated radiocontrast agents can produce a hyperosmotic endothelial i njury in the small vessels of the renal medulla, which leads to ischemic injury and an ATN-like picture. The renal P.626 injury usually ecomes apparent as a ri sing serum creatinine within 48 hours of the procedure. Oliguria is uncommon (10 % of cases), and most cases resolve within 2 weeks (15,16). This type of injury is common in dia etic patients with renal insufficiency, and dehydration is cons idered a risk factor. Saline infusions (150 to 200 mL/hour) efore, during, and after the procedure are effective in reducing the incidence of dye-induced ATN. Despite the introduction of more isotonic dyes, the pro lem persists. INTERSTITIAL NEPHRITIS AIN is caused y infections (usually viral or atypical pathogens) and hypersensi tivity drug http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 15)03-May-05 07:47:23

Ovid: ICU Book reactions. Virtually any drug can produce AIN, ut the most notorious offenders are listed in Ta le 39.3 (3). Many of these drugs are common in the ICU, so you should e aware of their nephrotoxic potential. TABLE 39.3. DRUGS THAT CAN CAUSE INTERSTITIAL NEPHRITIS AIN can e difficult to distinguish from ATN. The characteristic signs of a hype rsensitivity reaction (e.g., fever, rash, eosinophilia) may not e present. Ongo ing therapy with a high-risk drug can raise suspicion for AIN, ut the renal inj ury can occur months and even years after the onset of drug therapy (17). This c an direct attention away from the possi ility of a druginduced nephropathy. One of the aids to the diagnosis of AIN is the presence of white cell casts on urine microscopy. No specific treatment exists for AIN other than removing the offend ing drug. Therefore, in cases where the cause of acute renal failure is uncertai n, it is wise to stop use of as many medications as possi le. MYOGLOBINURIC ATN Widespread muscle reakdown (rha domyolysis) leads to the release of myoglo in i nto the loodstream, and when the myoglo in is filtered in the kidneys, it damag es the renal tu ules and produces an ATN-like illness. However, this is usually a milder form of renal failure and is nonoliguric in most cases. Most cases of r ha domyolysis are toxin-induced or traumatic. The num er of potential toxins is staggering; one review lists over 150 candidates (18). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 15)03-May-05 07:47:23

Ovid: ICU Book Diagnosis A positive urinary dipstick test for occult lood, com ined with the a sence of erythrocytes on urine microscopy, is diagnostic of myoglo inuria P.627 (except i n the rare case of severe intravascular hemolysis with hemoglo inuria). Rha domy olysis may e suspected when serum creatinine rises at a rate that is greater th an expected. As shown in Ta le 39.4, the serum creatinine will rise 1 to 2 mg/dL daily in the a sence of renal function. However, in the presence of widespread muscle reakdown, the serum creatinine will rise y more than 2 mg/dL each day. The diagnosis of rha domyolysis is secured y measuring any one of several muscl e enzymes in lood (e.g., creatine phosphokinase, lactate dehydrogenase). I pref er the aldolase enzyme, ecause it is specific for skeletal muscle. TABLE 39.4 CONSEQUENCES OF RENAL SHUTDOWN Management The management of myoglo inuric ATN includes volume infusion and careful monitor ing of potassium and phosphate levels in lood (these electrolytes are released from muscle, and their serum concentration can rise quickly to dangerous levels) . Volume infusion is very important ecause the damaging effects of myoglo in in the kidneys is markedly enhanced y volume depletion. Alkalinization of the uri ne is often recommended in myoglo inuria, ut is rarely necessary. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 15)03-May-05 07:47:23

Ovid: ICU Book HEMOFILTRATION Although hemodialysis is the standard method of lood purification in renal fail ure, a num er of alternative methods have surfaced in recent years (19). One of these is hemofiltration, which is descri ed elow. BASIC FEATURES Whereas hemodialysis removes solutes y diffusion, hemofiltration uses convectio n (solvent drag) to remove solutes. This is a less efficient method of solute re moval. However, the mem ranes used for hemofiltration are porous, and they allow large molecules (molecular weights up to 25,000 daltons) to pass through the me m ranes. This facilitates oth solute and fluid removal. CONTINUOUS HEMOFILTRATION Although hemofiltration removes solutes at a slower rate than hemodialysis, this allows hemofiltration to e used for longer periods P.628 (days). The technique of continuous hemofiltration is shown in Figure 39.3. In this case, the filter is placed etween an artery and a vein. This is called continuous arteriovenous hemofiltration (CAVH). The hemofilter can also e placed etween two veins. This method is called continuous venovenous hemofiltration (CVVH). Continuous Arteriovenous Hemofiltration No pumps are involved in CAVH. The arteriovenous pressure difference is the pres sure gradient for flow through the filter. The pressure gradient for filtration is the vertical distance etween the filter and the ultrafiltrate collection ag . The filtration pressure can thus e adjusted y adjusting the position of the collection ag relative to the filter. P.629 Also note the replacement fluid in the CAVH circuit. This is necessary ecause the concentration of solutes in the ultrafiltrate is the same as in the lood. Therefore, the concentration of solut es in lood will not decrease during hemofiltration unless a replacement fluid i s infused (dilutional effect). The use of a replacement fluid allows CAVH to red uce the concentration of toxins that accumulate in renal failure (i.e., it allow s CAVH to function as a method of dialysis). Over a 24-hour period, CAVH can rem ove 10 L of fluid (which is replaced) and 12 g of urea (20). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 15)03-May-05 07:47:23

Ovid: ICU Book Uses Hemofiltration can e used to remove solutes or fluids (e.g., in renal failure o r heart failure). Because the removal rate of fluid is slower than with hemodial ysis, hemofiltration is etter tolerated in patients who are hemodynamically com promised. The a ility to remove large molecules also has several potential appli cations. One of these is in severe sepsis, in which hemofiltration has een used to remove potentially harmful inflammatory mediators (21). This type of radiato r flush is likely to ecome a very popular practice in the near future (if it wo rks). REFERENCES SUGGESTED READINGS Bellomo R, ed. Acute renal failure. New horizons. Vol. 3, No. 4. Baltimore: Will iams & Wilkins, 1995. REVIEWS 1. Thadani R, Pascual M, Bonventre JV. Acute renal failure. N Engl J Med 1996;33 4:14481460 (153 references). OVERVIEW 2. Anderson RA. Prevention and management of acute renal failure. Hosp Pract 199 3;28:6175. 3. Garella S. Drug-induced renal disease. Hosp Pract 1993;28:129140. 4. Zaloga GP, Hughes SS. Oliguria in patients with normal renal function. Anesth esiology 1990;72:598602. 5. Cockroft DW, Gault MN. Prediction of creatinine clearance from serum creatini ne. Nephron 1976;16:3141. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 15)03-May-05 07:47:23

Ovid: ICU Book 6. Brivet FG, Kleinknecht DJ, Loirat P, et al. Acute renal failure in intensive care units: causes, outcome, and prognostic factors of hospital mortality: a pro spective multicenter study. Crit Care Med 1996;24:192198. 7. Johnson JP, Rockaw MD. Sepsis or ischemia in experimental acute renal failure : what have we learned? New Horiz 1995;3:608614. P.630 URINARY PARAMETERS 8. Steiner RW. Interpreting the fractional excretion of sodium. Am J Med 1984;77 :699 702. LOW-DOSE DOPAMINE 9. Bersten AD, Holt AW. Vasoactive drugs and the importance of renal perfusion p ressure. New Horiz 1995;3:650661. 10. Marik PE. Low-dose dopamine in critically ill oliguric patients. Heart Lung 1993;22:171175. 11. Chertow GM, Sayegh MH, Allgren RL, Lazarus JM. Is the administration of dopa mine associated with adverse or favora le outcomes in acute renal failure? Am J Med 1996;1:4953. 12. Thompson BT, Cockrill BA. Renal-dose dopamine: a Siren song? Lancet 1994;344 :7 8. FUROSEMIDE 13. Brater DC, Anderson SA, Brown-Cartwright D. Response to furosemide in chroni c http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 15)03-May-05 07:47:23

Ovid: ICU Book renal insufficiency: rationale for limited doses. Clin Pharmacol Ther 1986;40:13 4139. 14. Martin SJ, Danzinger LH. Continuous infusion of loop diuretics in the critic ally ill: a review of the literature. Crit Care Med 1994;22:13231329. RENAL DISORDERS 15. Wish JB, Moritz CE. Preventing radiocontrast-induced acute renal failure. J Crit Illness 1990;5:1631. 16. Hock R, Anderson RJ. Prevention of drug-induced nephrotoxicity in the intens ive care unit. J Crit Care 1995;10:3343. 17. Ten RM, Torres VE, Millner DS, et al. Acute interstitial nephritis. Mayo Cli n Proc 1988;3:921930. 18. Curry SC, Chang D, Connor D. Drug and toxin-induced rha domyolysis. Ann Emer g Med 1989;18:10681084. HEMOFILTRATION 19. Ronco C, Bar acini S, Digito A, Zoccali G. Achievements and new directions i n continuous renal replacement therapy. New Horiz 1995;3:708716. 20. Merrill RH. Techniques of continuous arteriovenous hemofiltration and hemodi alysis. J Crit Illness 1991;6:381387. 21. Vincent J-L, Tielemans C. Continuous hemofiltration in severe sepsis: is it eneficial? J Crit Care 1995;10:2732. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 15)03-May-05 07:47:23

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 40 HYPERTONIC AND HYPOTONIC SYNDROMES This chapter descri es the clinical disorders that are characterized y an a nor mal distri ution of total ody water (TBW) in the intracellular and extracellula r fluid compartments. These disorders are characterized y a change in the effec tive osmolarity of the extracellular fluid, and many are associated with an a no rmal plasma sodium concentration. BASIC CONCEPTS The following is a description of the forces that determine the movement of wate r etween the intracellular and extracellular fluid compartments (1,2,3 and 4). OSMOTIC ACTIVITY The activity (concentration) of solute particles in a solution is inversely rela ted to the activity (concentration) of water molecules in the solution. The solu te activity in a solution is also called the osmotic activity and is expressed i n osmoles (osm). The total osmotic activity in a solution is the sum of the indi vidual osmotic activities of all the solute particles in the solution. For monov alent ions, the osmotic activity in milliosmoles (mOsm) per unit volume is equiv alent to the concentration of the ions in milliequivalents (mEq) per unit volume . Thus, the osmotic activity in isotonic saline (0.9% sodium chloride) is as fol lows: P.632 Osmolarity is the osmotic activity per volume of solution (solutes plus wa ter) and is expressed as mOsm/L. Osmolality is the osmotic activity per volume o f water and is expressed as mOsm/kg H2O. The osmotic activity of ody fluids usu ally is expressed in relation to the volume of water (i.e., osmolality). However , the volume of water in ody fluids is far greater than the volume of solutes, so there is little difference etween the osmolality and osmolarity of ody flui ds. Thus, the terms osmolality and osmolarity can e used interchangea ly to des cri e the osmotic activity in ody fluids. TONICITY When two solutions are separated y a mem rane that allows the passage of water ut not solutes, the water passes from the solution with the lower osmotic activ ity to the solution with the higher osmotic activity. The http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 16)03-May-05 07:48:01

Ovid: ICU Book relative osmotic activity in the two solutions is called the effective osmolalit y, or tonicity. The solution with the higher osmolality is descri ed as hyperton ic, and the solution with the lower osmolality is descri ed as hypotonic. Thus, the tendency for water to move into and out of cells is determined y the relati ve tonicity of the intracellular and extracellular fluids. When the mem rane sep arating two fluids is permea le to oth solutes and water and a solute is added to one of the fluids, the solute equili rates fully across the mem rane. In this situation, the solute increases the osmolality of oth fluids ut does not chan ge the relative tonicity of either fluid. Therefore, the water will not move fro m one fluid compartment to the other. An example of a solute that ehaves in thi s manner is urea. Urea is freely permea le across cell mem ranes, so an increase in the urea concentration in extracellular fluid (i.e., an increase in the plas ma lood urea nitrogen [BUN] increases the osmolality of the extracellular fluid s ut does not increase the tonicity of the extracellular fluids or cause a net movement of water out of cells. Thus, azotemia (an increase in BUN) is a hyperos motic condition, ut not a hypertonic condition. PLASMA OSMOLALITY The osmolality of the extracellular fluids can e measured in the clinical la or atory using the freezing point of plasma (a solution containing 1 osm/L will fre eze at 1.86 C). This is the freezing point depression method for measuring osmolal ity. The osmolality of the extracellular fluids can also e calculated using the concentrations of sodium, chloride, glucose, and urea in plasma (these are the major solutes in extracellular fluid). The calculation elow uses a plasma [Na] of 140 mEq/L, a plasma [glucose] of 90 mg/dL, and a plasma [BUN] of 14 mg/dL. P.633 The sodium concentration is dou led to include the osmotic contri ution of chloride. The serum glucose and urea are measured in milligrams per deciliter, and the factors 18 and 2.8 (the atomic weights divided y 10) are used to conver t mg/dL to mOsm/kg H2O. OSMOLAL GAP Because solutes other than sodium, chloride, glucose, and urea are present in th e extracellular fluid, the measured plasma osmolality will e greater than the c alculated plasma osmolality. This osmolar gap (i.e., the difference etween the measured and calculated plasma osmolality) is normally as much as 10 mOsm/kg H2O (4,5). An increase in the osmolar gap occurs when certain toxins (e.g., ethanol , methanol, ethylene glycol, or the unidentified toxins that accumulate in renal failure) are in the extracellular fluid. Therefore, the osmolar gap has een pr oposed as a screening test for identifying the presence of toxins in the extrace llular fluid. In the case of renal failure, the osmolar gap has een recommended as a relia le test for distinguishing acute from chronic renal failure: the osm olar gap is expected to e normal in acute renal failure and elevated in chronic renal failure (4). In reality, the osmolar gap is used infrequently. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 16)03-May-05 07:48:01

Ovid: ICU Book PLASMA TONICITY Because urea passes freely across cell mem ranes, the effective osmolality or to nicity of the extracellular fluid can e calculated y eliminating urea (BUN) fr om the plasma osmolality equation. Because the concentration of urea contri utes little to the total solute concent ration in extracellular fluids, there is little difference etween the osmolalit y and tonicity of the extracellular fluid. This equation P.634 esta lishes the p lasma sodium concentration as the principal determinant of the effective osmolal ity of extracellular fluid. Because the effective osmolality determines the tend ency for water to move into and out of cells, the plasma sodium concentration is the principal determinant of the relative volumes of the intracellular and extr acellular fluids. HYPERNATREMIA The normal plasma (serum) sodium concentration is 135 to 145 mEq/L. Therefore, h ypernatremia (i.e., a serum sodium concentration a ove 145 mEq/L) can e the res ult of loss of fluid that has a sodium concentration elow 135 mEq/L (hypotonic fluid loss) or gain of fluid that has a sodium concentration a ove 145 mEq/L (hy pertonic fluid gain). Each of these conditions can e identified y assessing th e state of the extracellular volume, as shown in Ta le 40.1. TABLE 40.1. CHANGES IN TOTAL BODY SODIUM AND WATER IN HYPERNATREMIA AND HYPONATR EMIA http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 16)03-May-05 07:48:01

Ovid: ICU Book EXTRACELLULAR VOLUME If invasive hemodynamic monitoring is availa le, the intravascular volume (IVV) can e evaluated y noting the cardiac filling pressures and the cardiac output (as descri ed in Chapter 10, Chapter 11, Chapter 14, and Chapter 16). In the a s ence of hypoproteinemia (which shifts fluids from the intravascular to extravasc ular space), the state of the IVV can e used as a reflection of the state of th e extracellular volume (ECV). If invasive hemodynamic monitoring is unavaila le, the state of the ECV can e inferred from a few clinical varia les. The first i s the sudden loss of weight (i.e., in a nonedematous person, a sudden loss of we ight over a few days is an indication of a decreased ECV). The second is the pre sence of peripheral edema (i.e., in the a sence of hypoproteinemia, the presence of peripheral edema is an indication of an increased ECV). The third is the con centration of sodium in a random (spot) urine sample (i.e., a urine sodium conce ntration less than 10 mEq/L is an indication of a decreased ECV). Finally, the c linical P.635 manifestations shown in Ta le 14.2 can e used as evidence for a d ecreased ECV. Once the state of the ECV is determined, the strategies shown in F igure 40.1 can e applied. Figure 40.1. Management strategies for hypernatremia ased on the extracellular volume. (From Marino PL, Krasner J, O'Moore P. Fluid and electrolyte expert. Phi ladelphia: WB Saunders, 1987.) Low ECV indicates loss of hypotonic fluids. Common causes are excessive diuresis , vomiting, and diarrhea. The management strategy is to replace the sodium defic it quickly (to maintain IVV) and to replace the free water deficit slowly (to pr event intracellular overhydration). Normal ECV indicates a net loss of free wate r. This can e seen in dia etes insipidus, or when loss of hypotonic fluids (e.g ., diuresis) is treated y replacement with isotonic saline in a 1:1 volume-to-v olume ratio. The management strategy is to replace the free water deficit slowly (to prevent intracellular overhydration). High ECV indicates a gain of hyperton ic fluids. This is seen with aggressive use of hypertonic saline or sodium icar onate solutions. The management strategy is to induce sodium loss in the urine with diuresis http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 16)03-May-05 07:48:01

Ovid: ICU Book and to replace the urine volume loss with fluids that are hypotonic to the urine . Each of these conditions is descri ed in more detail in the following sections . HYPOVOLEMIC HYPERNATREMIA The most common cause of hypernatremia is loss of hypotonic ody fluids. The con centration of sodium in the ody fluids that are commonly P.636 lost is shown in Ta le 40.2. With the exception of small owel secretions and pancreatic secreti ons, loss of any of these ody fluids will result in hypernatremia. TABLE 40.2. SODIUM CONCENTRATION IN BODY FLUIDS CONSEQUENCES All of the ody fluids listed in Ta le 40.2 contain sodium, so the loss of these fluids will e accompanied y deficits in total ody sodium as well as TBW. The sodium deficits predispose to hypovolemia, whereas the free water deficits pred ispose to hypertonicity in the extracellular fluids. Therefore, the two conseque nces of hypotonic fluid loss are hypovolemia and hypertonicity. Hypovolemia The most immediate threat with hypotonic fluid loss is hypovolemia, which predis poses to hypoperfusion of the vital organs. Fortunately, hypovolemia is not as p rominent when hypotonic fluids are lost as when whole lood is lost. This is ec ause the resultant hypertonicity draws water out of cells, and this helps mainta in the volume of the extracellular (intravascular) fluid compartment. Hypertonicity http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 16)03-May-05 07:48:01

Ovid: ICU Book The hypertonicity of the extracellular fluids predisposes to cellular dehydratio n. The most serious consequence of hypertonic hypernatremia is a meta olic encep halopathy (6). Clinical findings include depressed consciousness that can progre ss to frank coma, generalized seizures, and focal neurologic deficits. Hypernatr emic encephalopathy has an associated mortality of up to 50% (6), ut management should proceed slowly. VOLUME REPLACEMENT The most immediate concern in hypovolemic hypernatremia is to replace volume def icits and to maintain the cardiac output. Volume P.637 replacement can e guided y the cardiac filling pressures and the cardiac output, or y the clinical var ia les shown in Ta le 14.2. When solute losses are severe and hemodynamic compro mise is present, colloid fluid replacement with 5% al umin or 6% hetastarch can rapidly restore the IVV, as descri ed in Chapter 15. When crystalloid fluids are given for volume replacement, isotonic saline should always e used and less co ncentrated fluids (e.g., half-normal saline), which predispose to cellular overh ydration, should e avoided. FREE WATER REPLACEMENT When hypovolemia has een corrected, the next step is to calculate and replace t he free water deficit. The calculation of free water deficit is ased on the ass umption that the product of TBW and plasma sodium concentration (PNa) is always constant. Using a normal plasma sodium concentration of 140 mEq/L and rearranging terms yi elds the following relationship: The normal TBW (in liters) usually is 60% of lean ody weight (in kg) in men and 50% of lean ody weight in women. However, in hypernatremia associated with fre e water deficits, the normal TBW should e approximately 10% less than usual (3) . Thus in men, the normal TBW is 0.5 ody weight (kg), and in women, the normal TBW is 0.4 ody weight (in kg). Once the current TBW is calculated, the water de ficit is taken as the difference etween the normal TBW and the current TBW. Example Calculation Assume that an adult man with a lean ody weight of 70 kg has a plasma sodium of 160 mEq/L. The normal TBW will e 0.5 70 = 35 L. The current TBW will e 35 140 /160 = 30.5 L. The TBW deficit will e 35 30.5 = 4.5 L. Replacement Volume The volume of the replacement fluid needed to correct the water deficit is deter mined y the concentration of sodium in the replacement fluid. The replacement v olume can e determined as follows (1): http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 16)03-May-05 07:48:01

Ovid: ICU Book P.638 where X is the ratio of the sodium concentration in the replacement fluid to the sodium concentration in isotonic saline (X = replacement fluid Na/154). I f the water deficit is 4.5 L and the replacement fluid is halfnormal saline (Na = 75 mEq/L), the replacement volume will e 4.5 (1/0.5) = 9 L. Cere ral Edema The rain cells initially shrink in response to a hypertonic extracellular fluid , ut cell volume is restored within hours. This restoration of cell volume is a ttri uted to the generation of osmotically active su stances called idiogenetic osmoles (6). Once the rain cell volume is restored to normal, the aggressive re placement of free water can predispose to cere ral edema. To limit the risk of c ere ral edema, free water deficits should e replaced slowly, over 48 to 72 hour s (6). DIABETES INSIPIDUS The most noted cause for hypernatremia without apparent volume deficits is dia e tes insipidus (DI), which is a condition of impaired renal water conservation (7 ). This condition results in excessive loss of urine that is almost pure water ( devoid of solute). The underlying pro lem in DI is related to antidiuretic hormo ne (ADH), a hormone secreted y the posterior pituitary gland that promotes wate r rea sorption in the distal tu ule. Two defects related to ADH can occur in DI: Central DI is caused y inhi ition of ADH release from the posterior pituitary. Common causes of central DI in critically ill patients include closed head inju ry, anoxic encephalopathy, and meningitis (5,7). The onset is heralded y polyur ia that usually is evident within 24 hours of the inciting event. Nephrogenic DI is caused y defective end-organ responsiveness to ADH. Possi le causes of neph rogenic DI in critically ill patients includes hypokalemia, aminoglycosides, amp hotericin, radiocontrast dyes, and the polyuric phase of ATN. The defect in urin e concentrating a ility in nephrogenic DI is not as severe as it is in central D I. DIAGNOSIS The hallmark of DI is a dilute urine in the face of plasma hypertonicity. In cen tral DI, the urine osmolarity is often elow 200 mOsm/L, whereas in nephrogenic DI, the urine osmolarity is usually etween 200 and 500 mOsm/L (5). The diagnosi s of DI is confirmed y noting the urinary response to fluid restriction. Failur e of the urine osmolarity to increase more than 30 mOsm/L in the first hours of complete fluid restriction is diagnostic of DI. The fluid losses can e excessiv e during fluid restriction in DI (particularly central DI), and thus fluid restr iction must e monitored carefully. Once the diagnosis of DI is confirmed, P.639 the response to vasopressin (5 units intravenously) will differentiate central from nephrogenic DI. In central DI, the urine osmolality increases y at least 5 0% almost immediately after vasopressin administration, whereas the urine osmola lity remains unchanged after vasopressin in nephrogenic DI. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 16)03-May-05 07:48:01

Ovid: ICU Book MANAGEMENT The fluid loss in DI is almost pure water, so the replacement strategy is aimed at replacing free water deficits only. The water deficit is calculated as descri ed previously, and the free water deficit is corrected slowly (over 2 to 3 days ) to limit the risk of cere ral edema. In central DI, vasopressin administration is also required to prevent ongoing free water losses. The usual dose is 5 to 1 0 units of aqueous vasopressin su cutaneously every 6 to 8 hours (5,7). The seru m sodium must e monitored carefully during vasopressin therapy ecause water in toxication and hyponatremia can occur if the central DI egins to resolve.

HYPERVOLEMIC HYPERNATREMIA Hypernatremia from hypertonic fluid gain is uncommon. Possi le causes are hypert onic saline resuscitation, sodium icar onate infusions for meta olic acidosis ( see Ta le 37.1), and ingestion of excessive amounts of ta le salt (8). MANAGEMENT In patients with normal renal function, excess sodium and water are excreted rap idly. When renal sodium excretion is impaired, it might e necessary to increase renal sodium excretion with a diuretic (e.g., furosemide). Because the sodium c oncentration in urine during furosemide diuresis is approximately 75 mEq/ L, exc essive urine output will aggravate the hypernatremia ( ecause the urine is hypot onic to plasma). Therefore, urine volume losses must e partially replaced with a fluid that is hypotonic to the urine. HYPERGLYCEMIA The formula for plasma tonicity presented earlier predicts that hyperglycemia wi ll e accompanied y a hypertonic extracellular fluid. When progressive hypergly cemia does not result in ketosis, the major clinical consequence is a hypertonic encephalopathy similar to the one descri ed for hypernatremia (6). The syndrome of nonketotic hyperglycemia (NKH) usually is seen in patients who have enough e ndogenous insulin to prevent ketosis. The condition usually is precipitated y a physiological stress (e.g., infection, trauma), and the patients may or may not have a prior history of dia etes mellitus (9). The plasma P.640 glucose is ofte n 1000 mg/dL or higher (9) (whereas in ketoacidosis, the plasma glucose is usual ly elow 800 mg/dL). The persistent loss of glucose in the urine produces an osm otic diuresis that can lead to profound volume losses. CLINICAL MANIFESTATIONS Patients with NKH usually have an altered mental status and may show signs of hy povolemia. The altered mental status can progress to frank coma when the plasma tonicity rises a ove 330 mOsm/kg H2O (9). Advanced cases of encephalopathy can e accompanied y generalized seizures and focal neurologic deficits, as descri e d for hypernatremic encephalopathy. MANAGEMENT http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 16)03-May-05 07:48:01

Ovid: ICU Book The fluid management of NKH is similar to that descri ed for hypovolemic hyperna tremia. Volume deficits tend to e more profound in NKH than in simple hypovolem ic hypernatremia ecause of the osmotic diuresis that accompanies the glycosuria . Therefore, rapid correction of the IVV (i.e., with 5% al umin or isotonic sali ne) may e necessary. Free Water Deficit Once the IVV is restored, free water deficits are estimated and replaced slowly. However when calculating the free water deficit that accompanies hyperglycemia, it is necessary to correct the plasma sodium for the increase in plasma glucose . This is ecause the hyperglycemia draws water from the intracellular space, an d this creates a dilutional effect on the plasma sodium concentration. The decre ase in plasma sodium in hypernatremia can vary according to the state of the ECV . In general, for every 100 mg/dL increment in the plasma glucose, the plasma so dium should fall y 1.6 to 2 mEq/L (10). Therefore, for a patient with a plasma glucose of 1000 mg/dL and a measured plasma sodium of 145 mEq/L, the actual or c orrected plasma sodium will average 145 + (900/100 1.8) = 161 mEq/L (the factor 1.8 is taken as the average value etween 1.6 and 2 mEq/L). The restoration of rain cell volume can occur rapidly in hypertonic states due to hyperglycemia (9) . Therefore, the free water replacement should e particularly judicious in NKH. Insulin Therapy Because insulin drives oth glucose and water into cells, insulin therapy can ag gravate hypovolemia. Therefore, in patients who are hypovolemic, insulin should e withheld until the IVV is restored. Once this is accomplished, insulin therap y can e given as advised for P.641 dia etic ketoacidosis (see Ta le 37.2). The insulin requirement will diminish as the hypertonic condition is corrected, so p lasma glucose concentrations should e monitored hourly during intravenous insul in therapy in NKH. HYPONATREMIA Hyponatremia (serum sodium less than 135 mEq/L) is found in up to 4.5% of hospit alized elderly patients (11) and in 1% of postoperative patients (12). It is par ticularly prevalent in patients with the acquired immunodeficiency syndrome (AID S) and can e seen in up to 40% of hospitalized patients with AIDS (13). The mor tality in hyponatremic patients is as much as dou le the mortality in patients w ith a normal plasma sodium concentration (11,13). This increase in mortality can e a reflection of the treatment as well as the consequences of hyponatremia. PSEUDOHYPONATREMIA Extreme elevations in plasma lipids or proteins increase the plasma volume and c an reduce the measured plasma sodium concentration. The increase in plasma volum e in these situations is in the nonaqueous phase of plasma, and ecause the sodi um is contained in the aqueous phase of plasma, the hyponatremia in this http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 16)03-May-05 07:48:01

Ovid: ICU Book situation does not represent a decrease in extracellular sodium relative to extr acellular water (i.e., true or hypotonic hyponatremia). This condition is theref ore called pseudohyponatremia. Because the nonaqueous phase of plasma represents only 7% of the total plasma volume, large increases in plasma lipids and plasma proteins are needed to produce significant decreases in the measured plasma sod ium concentration. The correction factors for hyperlipidemia and hyperproteinemi a are as follows (1): q Plasma triglycerides (g/L) 0.002 = mEq/L decrease in plasma Na. q Plasma protein level minus 8 (g/dL) 0.025 = mEq/L decrease in plasma Na. Ion-Specific Electrodes The conventional method for measuring plasma sodium (flame emission spectrophoto metry) includes oth the aqueous and nonaqueous phases of plasma. However, the n ewer ion-specific sodium electrodes measure the sodium concentration only in the aqueous phase of plasma. Therefore, pseudohyponatremia will not occur when ionspecific electrodes are used to measure the plasma sodium concentration (14). P. 642 HYPOTONIC HYPONATREMIA True or hypotonic hyponatremia represents an increase in free water relative to sodium in the extracellular fluids. It does not necessarily represent an increas e in the volume of extracellular fluids. As shown in Ta le 40.1, the ECV can e low, normal, or high in patients with hyponatremia. The diagnostic approach to h yponatremia can egin with an assessment of the ECV, as shown in Figure 40.2 (14 ,15). (The assessment of ECV is descri ed earlier, for the assessment of hyperna tremia.) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 16)03-May-05 07:48:01

Ovid: ICU Book Figure 40.2. Diagnostic approach to hyponatremia. SIADH = syndrome of inappropri ate antidiuretic hormone. HYPOVOLEMIC HYPONATREMIA This condition is characterized y fluid losses com ined with volume replacement using a fluid that is hypotonic to the lost fluid (e.g., diuresis replaced y d rinking tap water). The result is a net loss of sodium relative to free water, w hich decreases oth the ECV and the extracellular sodium concentration. The conc entration of sodium in a random (spot) urine sample can sometimes help determine if the sodium loss is renal or extrarenal in origin. Site of Sodium Loss Renal Extrarenal Urine Sodium >20 mEq/L <10 mEq/L http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 16)03-May-05 07:48:01

Ovid: ICU Book Renal sodium losses would e seen in diuretic overuse and in adrenal P.643 insuf ficiency, whereas extrarenal sodium losses can occur with diarrhea and persisten t vomiting. ISOVOLEMIC HYPONATREMIA Isovolemic hyponatremia is characterized y a small gain in free water, ut not enough to e clinically detected (approximately 5 L of excess water is necessary to produce detecta le peripheral edema in the average-size adult). In this situ ation, the major disorders to consider are inappropriate (nonosmotic) release of ADH and acute water intoxication (psychogenic polydipsia). The urine sodium and urine osmolality will help distinguish etween these two disorders. Clinical Disorder Inappropriate ADH Water intoxication Urine Sodium >20 mEq/L <10 mEq/L Urine Osmolality >100 mOsm/kg H2O <100 mOsm/kg H2O The inappropriate (nonosmotic) release of ADH is characterized y an inappropria tely concentrated urine (urine osmolality a ove 100 mOsm/kg H2O) in the face of a hypotonic plasma (plasma tonicity elow 290 mOsm/kg H2O). This condition can e seen in certain groups of stressed patients, such as patients who have undergone recent surgery. It can also e produced y a variety of tumors and infections. This latter condition is known as the syndrome of inappropriate ADH (SIADH), and it can e accompanied y severe hyponatremia (plasma sodium elow 120 mEq/L). HYPERVOLEMIC HYPONATREMIA Hypervolemic hyponatremia represents an excess of sodium and water, with the wat er gain exceeding the sodium gain. In this situation, the urine sodium can somet imes help identify the source of the pro lem. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 16)03-May-05 07:48:01

Ovid: ICU Book Common Causes Heart failure Renal failure Hepatic failure Urine Sodium <20 mEq/L >20 mEq/L <20 mEq/L The urine sodium can e misleading if the patient is also receiving diuretics (w hich are commonly used in these conditions). The clinical picture is usually hel pful, although these conditions can co-exist in critically ill patients. SYMPTOMATIC HYPONATREMIA The major complication of hyponatremia is a meta olic encephalopathy, which can e oth irreversi le and fatal (12,16,17). This condition is due to cere ral ede ma and increased intracranial pressure (17). In P.644 addition to having the sam e manifestations as the encephalopathy associated with the hypertonic syndromes (i.e., depressed level of consciousness, seizures, and focal neurologic findings ), this encephalopathy can e accompanied y the acute respiratory distress synd rome (18). In addition to hyponatremic encephalopathy another distinct encephalo pathy is associated with the therapy of hyponatremia, particularly when the hypo natremia is corrected too rapidly (17). This encephalopathy is characterized y diffuse demyelinating lesions and can e accompanied y pituitary damage and ocu lomotor nerve palsies. A specific type of demyelinating disorder known as centra l pontine myelinolysis has also een attri uted to rapid correction of hyponatre mia (19). As descri ed in the following section, the risk of this second encepha lopathy has led to specific recommendations for limiting the maximum rate and en d-point of corrective therapy. MANAGEMENT STRATEGIES The management of hyponatremia is determined y the state of the ECV (i.e., low, normal, or high) and y the presence or a sence of neurologic symptoms. Symptom atic hyponatremia requires more aggressive corrective therapy than asymptomatic hyponatremia. However, to limit the risk of a demyelinating encephalopathy, the rate of rise in plasma sodium should not exceed 0.5 mEq/L/hour and the final pla sma sodium concentration should not exceed 130 mEq/L (17). The general managemen t strategies ased on the ECV are as follows: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 16)03-May-05 07:48:01

Ovid: ICU Book Low ECV: Infuse hypertonic saline (3% NaCl) in symptomatic patients, and isotonic saline in asymptomatic patients. Normal ECV: Com ine furosemide diuresis with infusion of hypertonic saline in symptomatic pa tients, or isotonic saline in asymptomatic patients. High ECV: Use furosemide-induced diuresis in asymptomatic patients. In symptomatic patient s, com ine furosemide diuresis with judicious use of hypertonic saline. SODIUM REPLACEMENT When corrective therapy requires the infusion of isotonic saline or hypertonic s aline, the replacement therapy can e guided y the calculated sodium deficit. T his is determined as follows (using a plasma sodium of 130 mEq/L as the desired end-point of replacement therapy): P.645 The normal TBW (in liters) is 60% of the lean ody weight (in kg) in men, and 50% of the lean ody weight in women. Thus, for a 60 kg woman with a plasma sodium of 120 mEq/L, the sodium deficit will e 0.5 60 (130 120) = 300 mEq. Beca use 3% sodium chloride contains 513 mEq of sodium per liter, the volume of hyper tonic saline needed to correct a sodium deficit of 300 mEq will e 300/513 = 585 mL. Using a maximum rate of rise of 0.5 mEq/L/ hour for the plasma sodium (to l imit the risk of a demyelinating encephalopathy), the sodium concentration defic it of 10 mEq/L in the previous example should e corrected over at least 20 hour s. Thus, the maximum rate of hypertonic fluid administration will e 585/20 = 29 mL/hour. If isotonic saline is used for sodium replacement, the replacement vol ume will e 3.3 times the replacement volume of the hypertonic 3% saline solutio n. REFERENCES SUGGESTED READING Rose BD. Clinical physiology of acid- ase and electrolyte disorders. 4th ed. New York: McGraw-Hill, 1994. REVIEWS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 16)03-May-05 07:48:01

Ovid: ICU Book 1. Marino PL, Krasner J, O'Moore P. Fluid and electrolyte expert. Philadelphia: WB Saunders, 1987 (software). 2. Oh MS, Carroll HJ. Disorders of sodium meta olism: hypernatremia and hyponatr emia. Crit Care Med 1992;20:94103 (34 references). BASIC CONCEPTS 3. Rose BD. The total ody water and the plasma sodium concentration. In: Clinic al physiology of acid ase and electrolyte disorders. 4th ed. New York: McGraw-Hi ll, 1994;219234. 4. Sklar AK, Linas SL. The osmolal gap in renal failure. Ann Intern Med 1983;98: 481482. HYPERNATREMIA 5. Gehe M. Clinical approach to the hyperosmolar patient. Crit Care Clin 1987;5 :797815. 6. Arieff AI, Ayus JC. Strategies for diagnosing and managing hypernatremic ence phalopathy. J Crit Illness 1996;11:720727. 7. Blevins LS, Wand GS. Dia etes insipidus. Crit Care Med 1992;20:6979. 8. Moder KG, Hurley DL. Fatal hypernatremia from exogenous salt intake: report o f a case and review of the literature. Mayo Clin Proc 1990;65:15871594. P.646

10. Moran SM, Jamison RL. The varia le hyponatremic response to hyperglycemia. W est J Med 1985;142:4953. HYPONATREMIA http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 16)03-May-05 07:48:01

HYPERGLYCEMIA 9. Rose BD. Hyperosmolal states: hyperglycemia. In Clinical physiology of acidase and electrolyte disorders. 4th ed. New York: McGraw-Hill, 1994;737762.

Ovid: ICU Book 11. Terzian C, Frye EB, Piotrowski ZH. Admission hyponatremia in the elderly. J Gen Intern Med 1994;9:8991. 12. Ayus JC, Wheeler JM, Arieff AI. Postoperative hyponatremic encephalopathy in menstruant women. Ann Intern Med 1992;117:891897. 13. Tang WW, Kaptien EM, Feinstein EI, Massry SG. Hyponatremia in hospitalized p atients with the acquired immunodeficiency syndrome (AIDS) and the AIDS-related complex. Am J Med 1993;94:169 174. 14. Weis erg LS. Pseudohyponatremia: a reappraisal. Am J Med 1988;86:315318. 15. Schrier RW, Briner VA. The differential diagnosis of hyponatremia. Hosp Prac t 1990;25:2937. 16. Ayus JC, Arieff AI. Symptomatic hyponatremia: making the diagnosis rapidly. J Crit Illness 1990;5:846856. 17. Arieff AI, Ayus JC. Pathogenesis of hyponatremic encephalopathy. Chest 1993; 103:607610. 18. Ayus JC, Arieff AI. Pulmonary complications of hyponatremic encephalopathy: noncardiogenic pulmonary edema and hypercapnic respiratory failure. Chest 1995;1 07:517521. 19. Bruner JE, Redmond JM, Haggar AM, et al. Central pontine myelinolysis and po ntine lesions after rapid correction of hyponatremia: a prospective magnetic res onance imaging study. Ann Neurol 1990;27:6166. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 16)03-May-05 07:48:01

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 41 POTASSIUM Early sea-living organisms exhi ited a preference for intracellular potassium an d a disdain for intracellular sodium, which eventually changed the composition o f the oceans from a potassium salt solution to a sodium salt solution. This eha vior is also found in mammalian organisms, in whom potassium is the major intrac ellular cation and sodium is the major extracellular cation. This pattern is the result of the sodiumpotassium exchange pump on cell mem ranes, which sequesters potassium and extrudes sodium. In humans, only 2% of the total ody potassium st ores are found outside cells. This lack of extracellular representation limits t he value of the plasma (extracellular) potassium concentration as an index of to tal ody potassium stores. POTASSIUM DISTRIBUTION The marked discrepancy etween the intracellular and extracellular content of po tassium is illustrated in Figure 41.1. The total ody potassium content in healt hy adults is approximately 50 mEq/kg (1), so a 70-kg adult will have 3500 mEq of total ody potassium. However, only 70 mEq (2% of the total amount) is found in the extracellular fluids. Because the plasma accounts for approximately 20% of the extracellular fluid volume, the potassium content of plasma will e a out 15 mEq, which is 0.004% of the total amount of potassium in the ody. This suggest s that the plasma potassium will e an insensitive marker of changes in total o dy potassium stores. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 17)03-May-05 07:48:39

Ovid: ICU Book Figure 41.1. The intracellular and extracellular potassium content in a 70-kg ad ult with a total ody potassium of 50 mEq/L. SERUM POTASSIUM The relationship etween changes in total ody potassium and changes in serum po tassium is curvilinear, as shown in Figure 41.2 (2,3). The slope of the curve de creases on the deficit side of the graph, indicating that the change in serum pota ssium is much smaller P.648 when potassium is depleted than when potassium accum ulates. In an averaged-size adult with a normal serum potassium concentration (i .e., 3.5 to 5.5 mEq/L), a total ody potassium deficit of 200 to 400 mEq is requ ired to produce a 1 mEq/L decrease in serum potassium, whereas a total ody pota ssium excess of 100 to 200 mEq is required to produce a 1 mEq/L rise in serum po tassium (3). In other words, potassium depletion must e twice as great as potas sium accumulation P.649 to produce a significant (1 mEq/L) change in the serum p otassium concentration. This difference is due to the large pool of intracellula r potassium, which can replenish extracellular stores when potassium is lost. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 17)03-May-05 07:48:39

Ovid: ICU Book Figure 41.2. The relationship etween the serum potassium concentration and chan ges in total ody potassium content. (Redrawn from Brown RS. Extrarenal potassiu m homeostasis. Kidney Int 1986;30:116127.) HYPOKALEMIA Hypokalemia is a serum potassium concentration elow 3.5 mEq/L. The causes of hy pokalemia can e classified according to whether an intracellular shift of potas sium (transcellular shift) occurred or whether a decrease in total ody potassiu m content (potassium depletion) occurred (4). The following are some of the poss i le causes of hypokalemia that are likely to e encountered in the ICU. TRANSCELLULAR SHIFT Potassium movement into cells is facilitated y stimulation of 2-adrenergic recep tors on muscle cell mem ranes. Inhaled -agonist ronchodilators (e.g., al uterol) are well known for their a ility to reduce the serum potassium concentration, ut this effect is mild (0.5 mEq/L or less) in the usual therapeutic doses (5). A more significant effect is seen when inhaled agonists are given in com ination w ith diuretics (6). Other factors that promote the transcellular shift of potassi um into cells include alkalosis (respiratory or meta olic), hypothermia (acciden tal or induced), and insulin. Alkalosis has a varia le and unpredicta le effect on the serum potassium (7). Hypothermia causes a transient drop in serum potassi um that resolves during rewarming. Lethal cases of hypothermia can e accompanie d y hyperkalemia ecause http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 17)03-May-05 07:48:39

Ovid: ICU Book of widespread cell death (8). POTASSIUM DEPLETION Potassium depletion can e the result of either renal or extrarenal potassium lo sses. The site of potassium loss can often e identified y using a com ination of urinary potassium and chloride concentrations, as shown in Figure 41.3. Figure 41.3. Diagnostic approach to hypokalemia. Renal Potassium Loss The leading cause of renal potassium wasting is diuretic therapy. Other causes l ikely to e seen in the ICU include nasogastric drainage, alkalosis, and magnesi um depletion. The urinary chloride is low (less than 15 mEq/L) when nasogastric drainage or alkalosis is involved, and it is high (greater than 25 mEq/L) when m agnesium depletion or diuretics are responsi le. Magnesium depletion impairs pot assium rea sorption across the renal tu ules and may play http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 17)03-May-05 07:48:39

Ovid: ICU Book a very important P.650 role in promoting and sustaining potassium depletion in c ritically ill patients, particularly those receiving diuretics (9). Extrarenal Potassium Loss The major cause of extrarenal potassium ation in stool is 75 mEq/L, ut ecause ess each day, little potassium is lost. stool can e as high as 10 L, and thus in significant potassium depletion. loss is diarrhea. The potassium concentr the stool volume is normally 200 mL or l In diarrheal states, the daily volume of severe or prolonged diarrhea can result

CLINICAL MANIFESTATIONS Severe hypokalemia (serum K+ elow 2.5 mEq/L) can e accompanied y diffuse musc le weakness and mental status changes. Milder degrees of hypokalemia (serum K+ 2 .5 to 3.5 mEq/L) are often asymptomatic. A normalities in the ECG, including pro minent U waves (more than 1 mm in height), flattening and inversion of T waves, and prolongation of the QT interval, can e present in more than half of the cas es (10). None of these changes are specific for hypokalemia. The T wave changes and U waves can e seen with digitalis or left ventricular hypertrophy, and QT p rolongation can e seen with hypocalcemia and hypomagnesemia. P.651 Arrhythmias There is a misconception a out the a ility of hypokalemia to promote cardiac arr hythmias. Hypokalemia alone does not produce serious cardiac arrhythmias (10). H ypokalemia is often com ined with other conditions that can promote arrhythmias (e.g., magnesium depletion, digitalis), and the hypokalemia may enhance the proa rrhythmic effects of these other conditions. Hypokalemia is well known for its a ility to promote digitalis-induced arrhythmias (see Chapter 53). MANAGEMENT OF HYPOKALEMIA The first concern in hypokalemia is to eliminate or treat any condition that pro motes transcellular potassium shifts (e.g., alkalosis). If the hypokalemia is du e to potassium depletion, proceed as descri ed in the following section. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 17)03-May-05 07:48:39

Ovid: ICU Book POTASSIUM DEFICIT If the hypokalemia is due to potassium depletion, a potassium deficit of 10% of the total ody potassium stores is expected for every 1 mEq/L decrease in the se rum potassium (11). The correlation etween potassium deficits and the severity of hypokalemia is shown in Ta le 41.1. These estimates do not consider any contr i ution from transcellular potassium shifts, and thus they are meant only as rou gh guidelines for gauging the severity of potassium depletion. TABLE 41.1. POTASSIUM DEFICITS IN HYPOKALEMIA* POTASSIUM REPLACEMENT Solutions The usual replacement fluid is potassium chloride, which is availa le as a conce ntrated solution (1.5 and 2 mEq/mL) in ampules containing 10, 20, 30, and 40 mEq of potassium. These solutions are extremely hyperosmotic (the 2 mEq/L solution has an osmolality of 4000 P.652 mOsm/kg H2O) and must e diluted (12). A potassi um phosphate solution is also availa le (contains 4.5 mEq potassium and 3 mM pho sphate per mL) and is preferred y some for potassium replacement in dia etic ke toacidosis ( ecause of the phosphate depletion that accompanies ketoacidosis). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 17)03-May-05 07:48:39

Ovid: ICU Book Infusion Rate The standard method of intravenous potassium replacement is to add 20 mEq of pot assium to 100 mL of isotonic saline and infuse this mixture over 1 hour (13). Th e maximum rate of intravenous potassium replacement is usually set at 20 mEq/hou r (13), ut dose rates up to 40 mEq/hour occasionally may e necessary (e.g., wi th serum K+ elow 1.5 mEq/L or serious arrhythmias), and dose rates as high as 8 0 mEq/hour have een used safely (4). A large central vein should e used for in fusion ecause of the irritating properties of the hyperosmotic potassium soluti ons. However, if the desired replacement rate is greater than 20 mEq/hour, the i nfusion should not e given through a central venous catheter ecause of the the oretical risk of transient hyperkalemia in the right heart cham ers, which can p redispose to cardiac standstill. In this situation, the potassium dose can e sp lit and administered via two peripheral veins. Response The serum potassium may e slow to rise at first, ecause of the position on the flat part of the curve in Figure 41.2. Full replacement usually takes a few day s, particularly if potassium losses are ongoing. If the hypokalemia seems refrac tory to replacement therapy, the serum magnesium level should e checked. Magnes ium depletion promotes urinary potassium losses and can cause refractory hypokal emia (14). The management of hypomagnesemia is presented in Chapter 42. HYPERKALEMIA While hypokalemia is often well tolerated, hyperkalemia (serum K+ greater than 5 .5 mEq/L) can e a serious and life-threatening condition (15). PSEUDOHYPERKALEMIA Potassium release from traumatic hemolysis during the venipuncture can produce a spurious elevation in serum potassium. This is more common than suspected, and has een reported in 20% of lood samples with an elevated serum potassium (16). Potassium release from muscles distal to a tourniquet can also e a source of s puriously high serum potassium levels (17). Because of the risk of spurious hype rkalemia, an unexpected finding of hyperkalemia in an asymptomatic P.653 patient should always prompt a repeat measurement efore any diagnostic or therapeutic measures are initiated. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 17)03-May-05 07:48:39

Ovid: ICU Book Potassium release from cells during clot formation in the specimen tu e can also produce pseudohyperkalemia when severe leukocytosis (white lood cell count gre ater than 50,000/ mm3) or throm ocytosis (platelet count greater than 1 million/ mm3) is present. When this condition is suspected, the serum potassium should e measured in an unclotted lood sample. URINE POTASSIUM Hyperkalemia can e caused y potassium release from cells (transcellular shift) or y impaired renal potassium excretion. If the source of the hyperkalemia is unclear, the urinary potassium concentration can e helpful. A high urine potass ium (greater than 30 mEq/L) suggests a transcellular shift, and a low urine pota ssium (less than 30 mEq/L) indicates impaired renal excretion. TRANSCELLULAR SHIFT Acidosis traditionally has een listed as a cause of hyperkalemia ecause of the tendency for acidosis to oth enhance potassium release from cells and reduce r enal potassium excretion. However, hyperkalemia does not always accompany respir atory acidosis (18), and no clear evidence exists that organic acidoses (i.e., l actic acidosis and ketoacidosis) can produce hyperkalemia (7,18,19). Although hy perkalemia can accompany acidoses associated with renal failure and renal tu ula r acidosis, hyperkalemia in these instances may e caused y impaired renal pota ssium excretion. Myonecrosis can release large amounts of potassium into the ext racellular fluid, ut if renal function is normal, the extra potassium is prompt ly cleared y the kidneys. For example, severe exercise can raise the serum pota ssium to 8 mEq/L, ut the hyperkalemia resolves with a half-time of 25 seconds ( 20). Drugs that can promote hyperkalemia via transcellular potassium shifts incl ude -receptor antagonists and digitalis (Ta le 41.2). Serious P.654 hyperkalemia (i.e., serum potassium a ove 7 mEq/L) is possi le only with digitalis toxicity. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 17)03-May-05 07:48:39

Ovid: ICU Book TABLE 41.2. DRUGS THAT CAN CAUSE HYPERKALEMIA IMPAIRED RENAL EXCRETION Renal insufficiency can produce hyperkalemia when the glomerular filtration rate falls elow 10 mL/minute or the urine output falls elow 1 L/day (21). Exceptio ns are interstitial nephritis and hyporeninemic hypoaldosteronism (21). The latt er condition is seen in elderly dia etic patients who have defective renin relea se in response to reduced renal lood flow. Adrenal insufficiency is a well know n cause of hyperkalemia from impaired renal potassium excretion, ut is not a co mmon cause of hyperkalemia in the ICU. Drugs that impair renal potassium excreti on are considered one of the leading causes of hyperkalemia. A list of common of fenders is shown in Ta le 41.2. The drugs most commonly implicated are angiotens in-converting enzyme inhi itors, potassium sparing diuretics, and nonsteroidal a ntiinflammatory drugs (15,16). Other potential offenders in the ICU are heparin, trimethoprimsulfamethoxazole, and pentamidine (22,23 and 24). All of these agent s promote hyperkalemia y inhi iting or locking the reninangiotensinaldosterone s ystem, and all promote hyperkalemia particularly when given with potassium suppl ements. BLOOD TRANSFUSIONS Massive lood transfusions (i.e., when the transfusion volume exceeds the estima ted lood volume) can promote hyperkalemia when given to patients with circulato ry shock. Potassium leakage from erythrocytes results in a steady rise in plasma potassium levels in stored lood. In whole lood, the plasma potassium rises an average of 1 mEq/L/day. However, ecause http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 17)03-May-05 07:48:39

Ovid: ICU Book one unit of whole lood contains 250 mL of plasma, this represents an increase o f only 0.25 mEq/day in the plasma potassium content per unit of whole lood. Aft er 14 days of storage, the plasma potassium load is 4.4 mEq per unit of whole l ood and 3.1 mEq per unit of packed red cells (25). The potassium load in lood t ransfusions normally is cleared y the kidneys, and thus no sustained rise in pl asma potassium occurs. However, in patients with circulatory shock, the extra po tassium from lood transfusions can accumulate and produce hyperkalemia. Further more, when the volume of distri ution for potassium is curtailed y widespread h ypoperfusion, the potassium accumulation can e rapid and life-threatening. CLINICAL MANIFESTATIONS The most serious consequence of hyperkalemia is the tion in the heart. The ECG can egin to change when aches 6.0 mEq/L, and it is always a normal when the mEq/L (21). Figure 41.4 illustrates the ECG changes hyperkalemia. slowing of electrical conduc P.655 the serum potassium re serum potassium reaches 8.0 associated with progressive

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 17)03-May-05 07:48:39

Ovid: ICU Book Figure 41.4. The ECG manifestations of progressive hyperkalemia. (Adapted from B urch GE, Winsor T. A primer of electrocardiography. Philadelphia: Lea & Fe iger, 1966;143.) The earliest change in the ECG is a tall, tapering (tented) T wave that is most evident in precordial leads V2 and V3. Similar peaked T waves have een o served i n meta olic acidosis (26). As the hyperkalemia progresses, the P wave amplitude decreases and the PR interval lengthens. The P waves eventually disappear and th e QRS duration ecomes prolonged. The final event is ventricular asystole. MANAGEMENT OF HYPERKALEMIA The acute management of hyperkalemia is guided y the serum potassium level and the ECG. The therapeutic maneuvers are outlined in Ta le 41.3. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 17)03-May-05 07:48:39

Ovid: ICU Book TABLE 41.3. ACUTE MANAGEMENT OF HYPERKALEMIA MEMBRANE ANTAGONISM Calcium directly antagonizes the mem rane actions of potassium. When hyperkalemi a is severe (i.e., a ove 7 mEq/L) or accompanied y advanced ECG changes (i.e., loss of P waves and prolonged QRS duration), calcium gluconate is administered i n the dose shown in Ta le 41.3. If there is no response to calcium within a few minutes, a second dose can e given. A third dose will not e effective if there was no response to the second dose of calcium. The response to calcium lasts on ly 20 or 30 minutes, so other therapies should e initiated to enhance potassium clearance. Calcium must e given cautiously to patients on digitalis ecause hy percalcemia can potentiate digitalis cardiotoxicity. For patients receiving digi talis, the calcium gluconate should e added to 100 mL of isotonic saline and in fused over 20 to 30 minutes. If the hyperkalemia is a manifestation of digitalis toxicity, calcium is contraindicated. When hyperkalemia is accompanied y evide nce of circulatory compromise, calcium chloride http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 17)03-May-05 07:48:39

Ovid: ICU Book is preferred to calcium gluconate. One ampule (10 mL) of 10% calcium chloride co ntains three times more elemental calcium than one ampule of 10% calcium glucona te (see Ta le 43.3), and the extra calcium in calcium chloride may prove enefic ial in promoting cardiac contraction and maintaining peripheral vascular tone. TRANSCELLULAR SHIFT InsulinDextrose Com ined therapy with insulin and dextrose will drive potassium into muscle cell s and decrease the serum potassium y an average of 1 mEq/L. However, this is a temporary effect, and other maneuvers aimed at enhancing potassium clearance are also required. P.656 P.657 Sodium Bicar onate The administration of sodium icar onate (44 to 88 mEq) can also shift potassium into cells. However, the most common acidotic condition associated with hyperka lemia is renal failure, and in this condition, insulindextrose is much more effec tive in lowering the serum potassium than icar onate (27). Furthermore, icar o nate inds calcium and should not e given after calcium is administered. For th ese reasons, there is little value in using icar onate to treat hyperkalemia. ENHANCED CLEARANCE Measures aimed at enhancing the removal of potassium from the ody can e used a lone in mild cases of hyperkalemia (i.e., serum K less than 7 mEq/L) without adv anced ECG changes or can serve as a follow-up to calcium and insulindextrose ther apy. Exchange Resin Polystyrene sulfonate (Kayexalate) is a cation exchange resin that can enhance p otassium clearance across the gastrointestinal mucosa (gastrointestinal dialysis ). This resin can e given orally or y retention enema, and it is mixed with 20 % sor itol to prevent concretion. For P.658 each mEq of potassium removed, 2 to 3 mEq of sodium are added. If there is concern a out http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 17)03-May-05 07:48:39

Ovid: ICU Book

Loop Diuretics The loop diuretics furosemide and ethacrynic acid enhance urinary potassium excr etion. These agents can e used as a follow-up measure to calcium and insulindext rose. This approach is ineffective in renal failure. Hemodialysis Hemodialysis is the most effective method of lowering the serum potassium in pat ients with renal failure (27). REFERENCES SUGGESTED READING Androgue HJ, Wesson DE. Potassium. Boston: Blackwell Scientific, 1994. POTASSIUM DISTRIBUTION 1. Rose BD. Potassium homeostasis. In: Clinical physiology of acid- ase and elec trolyte disorders. 4th ed. New York: McGraw-Hill, 1994;346376. 2. Brown RS. Extrarenal potassium homeostasis. Kidney Int 1986; 30:116-127. 3. Sterns RH, Cox M, Feig PU, Singer I. Internal potassium alance and the contr ol of the plasma potassium alance. Medicine 1981;60:339351. HYPOKALEMIA 4. Freedman BI, Burkhart JM. Hypokalemia. Crit Care Clin 1991;7:143153. 5. Bodenhammer J, Bergstrom R, Brown D, et al. Frequently ne ulized eta-agonist s for http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 17)03-May-05 07:48:39

the added sodium, one or two doses of furosemide can sis.

e used to enhance natriure

Ovid: ICU Book asthma: effects on serum electrolytes. Ann Emerg Med 1992;21:13371342. 6. Lipworth BJ, McDevitt DG, Struthers AD. Prior treatment with diuretic augment s the hypokalemic and electrocardiographic effects of inhaled al uterol. Am J Me d 1989;86:653657. 7. Androgue HJ, Madias NE. Changes in plasma potassium concentration during acut e acid- ase changes. Am J Med 1981;71:456467. 8. Schaller MD, Fischer AP, Perret CH. Hyperkalemia: a prognostic factor during acute, severe hypothermia. JAMA 1990;264:18421845. 9. Salem M, Munoz R, Chernow B. Hypomagnesemia in critical illness. Crit Care Cl in 1991;7:225252. 10. Flake G, Villarread D, Chapman D. Is hypokalemia a cause of ventricular arr hythmias? J Crit Illness 1986;1:6674. 11. Stanaszek WF, Romankiewicz JA. Current approaches to management of potassium deficiency. Drug Intell Clin Pharmacol 1985;19:176184. P.659 12. Trissel LA. Hand ook on injecta le drugs. 8th ed. Bethesda, MD: American Soc iety for Hospital Pharmacists, 1994;886902. 13. Kruse JA, Carlson RW. Rapid correction of hypokalemia using concentrated int ravenous potassium chloride infusions. Arch Intern Med 1990;150:613617. 14. Whang R, Flink EB, Dyckner T, et al. Mg depletion as a cause of refractory p otassium depletion. Arch Intern Med 1985;145:16861689. HYPERKALEMIA http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 17)03-May-05 07:48:39

Ovid: ICU Book 15. Williams ME. Hyperkalemia. Crit Care Clin 1991;7:155174. 16. Rimmer JM, Horn JF, Gennari FJ. Hyperkalemia as a complication of drug thera py. Arch Intern Med 1987;147:867869.

18. Burger GA, Howard R. Acidosis and [K+]. Anesth Analg 1993;76:680. 19. Orringer CE, Eustace JC, Wunsch CD, Gardner LB. Natural history of lactic ac idosis after grand-mal seizure. N Engl J Med 1977;297:796799. 20. Med o JL, Sejersted OM. Plasma potassium changes with high intensity exercis e. J Physiol 1990;42:105122. 21. Williams ME, Rosa RM. Hyperkalemia: disorders of internal and external potas sium alance. J Intensive Care Med 1988;3:5264. 22. Oster JR, Singer I, Fishman LM. Heparin-induced aldosterone suppression and hyperkalemia. Am J Med 1995;98:575586. 23. Green erg S, Reiser IW, Chou SY, Porush JG. Trimethoprimsulfamethoxazole indu ces reversi le hyperkalemia. Ann Intern Med 1993;119:291295. 24. Peltz S, Hashmi S. Pentamidine-induced severe hyperkalemia. Am J Med 1989;87 :698699. 25. Michael JM, Dorner I, Burns D, et al. Potassium load in CPD-preserved whole lood and two types of packed red cells. Transfusion 1975;15:144149.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 17)03-May-05 07:48:39

17. Don BR, Se astian A, Cheitlin M, et al. Pseudohyperkalemia caused enching during phle otomy. N Engl J Med 1990;322:12901293.

y fist cl

Ovid: ICU Book 26. Dreyfuss D, Jondeau G, Couturier R, et al. Tall T waves during meta olic aci dosis without hyperkalemia: a prospective study. Crit Care Med 1989;17:404408. 27. Blum erg A, Weidmann P, Gnadinger M. Effect of various therapeutic approache s on plasma potassium and major regulating factors in terminal renal failure. Am J Med 1988;85:507512. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 17)03-May-05 07:48:39

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 42 MAGNESIUM Magnesium is the second most a undant intracellular cation in the human ody (po tassium eing the first), where it serves as a cofactor for more than 3000 enzym e reactions that involve adenosine triphosphate (1). One of the magnesium-depend ent enzyme systems is the mem rane pump that generates the electrical gradient a cross cell mem ranes. As a result, magnesium plays an important role in the acti vity of electrically excita le tissues (1,2,3 and 4). Magnesium also regulates t he movement of calcium into smooth muscle cells, which gives it a pivotal role i n the maintenance of cardiac contractile strength and peripheral vascular tone ( 2). MAGNESIUM BALANCE The content and distri ution of magnesium in the human ody is shown in Ta le 42 .1 (1). The average-size adult contains approximately 24 g (1 mole, or 2000 mEq) of magnesium; a little over half is located in one, whereas less than 1% is lo cated in plasma. This lack of representation in the plasma limits the value of t he plasma magnesium concentration as an index of total ody magnesium stores. Th is is particularly true in patients with magnesium deficiency, in whom serum mag nesium levels can e normal in the face of total ody magnesium depletion (5,6). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 19)03-May-05 07:50:15

Ovid: ICU Book TABLE 42.1. MAGNESIUM DISTRIBUTION IN ADULTS SERUM MAGNESIUM Serum is favored over plasma for magnesium assays ecause the anticoagulant used for plasma samples can e contaminated with citrate or other anions that ind m agnesium (6). The normal range for serum magnesium depends on the daily magnesiu m intake, which P.661 varies according to geographic region. The normal range fo r healthy adults residing in the United States is shown in Ta le 42.2 (7). TABLE 42.2. REFERENCE RANGES FOR MAGNESIUM* Ionized Magnesium Only 55% of the magnesium in plasma is in the ionized (active) form, and the rem aining 45% is either ound to plasma proteins (33% of the total) or chelated wit h divalent anions such as phosphate and sulfate (12% of the total) (1,3,4). The standard assay for magnesium (i.e., spectrophotometry) measures all three fracti ons of magnesium. Therefore, when the serum magnesium is a normally low, it is i mpossi le to determine whether the pro lem is a decrease in the ionized (active) fraction or a decrease in the ound fractions (e.g., hypoproteinemia) (8). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 19)03-May-05 07:50:15

Ovid: ICU Book The level of ionized magnesium can e measured with an ion-specific electrode (9 ) or y ultrafiltration of plasma (10), ut these techniques are not routinely a vaila le for clinical use. However, ecause the total amount of magnesium in pla sma is small, the difference etween the ionized and ound magnesium content may not e large enough to e clinically relevant. URINARY MAGNESIUM The normal range for urinary magnesium excretion is shown in Ta le 42.2. Under n ormal circumstances, only small quantities of magnesium P.662 are excreted in th e urine. When magnesium intake is deficient, the kidneys conserve magnesium and urinary magnesium excretion falls to negligi le levels. This is shown in Figure 42.1 (11). After the start of a magnesium deficient diet, the urinary magnesium excretion promptly falls to negligi le levels and the serum magnesium remains in the normal range. This illustrates the relative value of urinary magnesium over serum magnesium levels in the detection of magnesium deficiency. This is discus sed again later in this chapter. Figure 42.1. Urinary and plasma magnesium levels in a healthy volunteer placed o n a magnesium-free diet. Solid ars on the vertical axes indicate the normal ran ge for urine and plasma magnesium. (Adapted from Shils ME. Experimental human ma gnesium deficiency. Medicine 1969;48:6182.) MAGNESIUM DEFICIENCY http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 19)03-May-05 07:50:15

Ovid: ICU Book Magnesium deficiency is common in hospitalized patients (12,13,14,15,16,17,18,19 and 20). Hypomagnesemia is reported in 10 to 20% of patients on general medical wards (15,16 and 17) and in 60 to 65% of patients in ICUs (18,19). Because magn esium depletion may not e accompanied y hypomagnesemia, the incidence of magne sium depletion is even higher than indicated y the surveys of hypomagnesemia. I n fact, magnesium depletion has een descri ed as the most underdiagnosed electro lyte a normality in current medical practice (20). PREDISPOSING CONDITIONS Because serum magnesium levels have a limited a ility to detect magnesium deplet ion, recognizing the conditions that predispose to magnesium depletion may e th e only clue of an underlying electrolyte P.663 im alance. The common predisposin g conditions for magnesium depletion are listed in Ta le 42.3. TABLE 42.3. MARKERS OF POSSIBLE MAGNESIUM DEPLETION Diuretic Therapy Diuretics are the leading cause of magnesium deficiency. Drug-induced inhi ition of sodium rea sorption also interferes with magnesium rea sorption, and the res ultant urinary magnesium losses can parallel urinary sodium losses. Urinary magn esium excretion is most http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 19)03-May-05 07:50:15

Ovid: ICU Book pronounced with the loop diuretics (furosemide and ethacrynic acid). Magnesium d eficiency has een reported in 50% of patients receiving chronic therapy with fu rosemide (21). The thiazide diuretics show a similar tendency for magnesium depl etion, ut only in elderly patients (22). Magnesium depletion is not a complicat ion of therapy with potassium-sparing diuretics such as triamterene (23). Anti iotic Therapy The anti iotics that promote magnesium depletion are the aminoglycosides, amphot ericin and pentamidine (24,25). The aminoglycosides lock magnesium rea sorption in the ascending loop of Henle, and hypomagnesemia has een reported in 30% of patients receiving aminoglycoside therapy (25). The other risk associated with a nti iotic use occurs with anti iotic-associated diarrhea, which can e accompani ed y significant magnesium losses in the stool. Other Drugs A variety of other drugs have een associated with magnesium depletion, includin g digitalis (26), adrenergic agents (27), and the chemotherapeutic agents cispla tin (28) and cyclosporine (29). The first two agents shift magnesium into cells, whereas the latter two promote renal magnesium excretion. P.664 Alcohol-Related Illness Hypomagnesemia is reported in 30% of hospital admissions for alcohol a use, and in 85% of admissions for delirium tremens (30,31). The magnesium depletion in th ese conditions is due to a num er of factors, including generalized malnutrition and chronic diarrhea. In addition, there is an association etween magnesium de ficiency and thiamine deficiency (32). Magnesium is required for the transformat ion of thiamine into thiamine pyrophosphate, so magnesium deficiency can promote thiamine deficiency in the face of adequate thiamine intake. For this reason, t he magnesium status should e monitored periodically in patients receiving daily thiamine supplements. Secretory Diarrhea A high concentration of magnesium (10 to 14 mEq/L) is present in secretions from the lower gastrointestinal tract (33), and thus secretory diarrhea can e accom panied y profound magnesium depletion (31). Upper tract secretions are not rich in magnesium (1 to 2 mEq/L), so http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 19)03-May-05 07:50:15

Ovid: ICU Book vomiting does not pose a risk for magnesium depletion. Dia etes Mellitus Magnesium depletion is common in insulin-dependent dia etic patients, pro a ly a s a result of urinary magnesium losses that accompany glycosuria (34). Hypomagne semia is reported in only 7% of admissions for dia etic ketoacidosis, ut the in cidence increases to 50% over the first 12 hours after admission (35), pro a ly as a result of insulin-induced movement of magnesium into cells. Acute Myocardial Infarction As many as 80% of patients with acute myocardial infarction (MI) can have hypoma gnesemia in the first 48 hours after the event (36). The mechanism is unclear u t may e due to an intracellular shift of magnesium caused y endogenous catecho lamines excess. The importance of magnesium in patients with acute MI is discuss ed in Chapter 19. CLINICAL MANIFESTATIONS Although no clinical manifestations are specific for magnesium deficiency, the f ollowing clinical findings are suggestive of an underlying magnesium deficiency (Ta le 42.3). Associated Electrolyte A normalities Magnesium depletion is often accompanied y depletion of other electrolytes, suc h as potassium, phosphate, calcium, and phosphate (See Ta le 42.3) (37). As ment ioned in Chapter 41, the hypokalemia P.665 that accompanies magnesium depletion can e refractory to potassium replacement therapy (see Reference 14 in Chapter 41), and magnesium repletion is often necessary efore potassium repletion is po ssi le. The hypocalcemia that accompanies magnesium depletion is due to impaired parathormone release (38) com ined with an impaired end-organ response to parat hormone (39). In addition, magnesium deficiency may act on one directly to redu ce calcium release, independent of parathyroid hormone (40). As with the hypokal emia, the hypocalcemia from magnesium depletion is difficult to correct unless m agnesium deficits are corrected. Hypophosphatemia is a cause rather than effect of magnesium depletion. The mechanism is enhanced renal magnesium excretion (41) . Therefore, when hypophosphatemia accompanies http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 19)03-May-05 07:50:15

Ovid: ICU Book hypomagnesemia, the phosphate stores should e replenished to ensure adequate re pletion of magnesium stores. Arrhythmias Because magnesium is required for proper function of the mem rane pump on cardia c cell mem ranes, magnesium depletion will depolarize cardiac cells and promote tachyarrhythmias. Because oth digitalis and magnesium deficiency act to inhi it the mem rane pump, magnesium deficiency will magnify the digitalis effect and p romote digitalis cardiotoxicity. Intravenous magnesium is effective in suppressi ng digitalis-toxic arrhythmias, even when serum magnesium levels are normal (42, 43). Intravenous magnesium can also e effective in a olishing refractory arrhyt hmias (i.e., unresponsive to traditional antiarrhythmic agents) in the a sence o f hypomagnesemia (44). This effect may e due to a mem rane-sta ilizing effect o f magnesium that is unrelated to magnesium repletion. One of the most serious ar rhythmias associated with magnesium depletion is torsades de pointes (polymorpho us ventricular tachycardia). The role of magnesium in this arrhythmia is discuss ed in Chapter 20. Neurologic Findings The neurologic manifestations of magnesium deficiency include altered mentation, generalized seizures, tremors, and hyperreflexia. All are uncommon, nonspecific , and have little diagnostic value. A neurologic syndrome descri ed recently tha t can a ate with magnesium therapy deserves mention. The clinical presentation i s characterized y ataxia, slurred speech, meta olic acidosis, excessive salivat ion, diffuse muscle spasms, generalized seizures, and progressive o tundation (4 5). The clinical features are often rought out y loud noises or odily contact , and thus the term reactive central nervous system magnesium deficiency has ee n used to descri e this disorder. This syndrome is associated with reduced magne sium levels in cere rospinal fluid, and it resolves with magnesium infusion. The prevalence of this disorder is unknown at present. P.666 DIAGNOSIS As mentioned several times, the serum magnesium level is an insensitive marker o f magnesium depletion. When magnesium depletion is due to nonrenal factors (e.g. , diarrhea), the urinary magnesium excretion is a more sensitive test for magnes ium depletion (46). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 19)03-May-05 07:50:15

Ovid: ICU Book However, ecause most cases of magnesium depletion are due to enhanced renal mag nesium excretion, the diagnostic value of urinary magnesium excretion may e lim ited. Magnesium Retention Test In the a sence of renal magnesium wasting, the urinary excretion of magnesium in response to a magnesium load may e the most sensitive index of total ody magn esium stores (47,48). This method is outlined in Ta le 42.4. The normal rate of magnesium rea sorption is close to the maximum tu ular rea sorption rate (Tmax), so most of the infused magnesium will e excreted in the urine when magnesium s tores are normal. However, when magnesium stores are deficient, the magnesium re a sorption rate is much lower than the Tmax, so more of the infused magnesium wi ll e rea sor ed and less will e excreted in the urine. When less than 50% of t he infused magnesium is recovered in the urine, magnesium deficiency is likely, and when more than 80% of the infused magnesium is excreted in the urine, magnes ium deficiency is unlikely. This test can e particularly valua le in determinin g the end-point of magnesium replacement therapy (i.e., magnesium replacement is continued until urinary magnesium excretion is at least 80% of the infused magn esium load). It is important to P.667 emphasize that this test will e unrelia l e in patients with impaired renal function or when there is ongoing renal magnes ium wasting. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 19)03-May-05 07:50:15

Ovid: ICU Book TABLE 42.4. RENAL MAGNESIUM RETENTION TEST MAGNESIUM REPLACEMENT THERAPY PREPARATIONS The magnesium preparations availa le for oral and parenteral use are listed in T a le 42.5 (49,50). The oral preparations can e used for daily maintenance thera py (5 mg/kg in normal su jects) and for correcting mild, asymptomatic magnesium deficiency. However, ecause intestinal a sorption of oral magnesium is erratic, parenteral magnesium is preferred for treating symptomatic or severe magnesium deficiency. TABLE 42.5. ORAL AND PARENTERAL MAGNESIUM PREPARATIONS Magnesium Sulfate The standard intravenous preparation is magnesium sulfate (MgSO4). Each gram of magnesium sulfate has 8 mEq (4 mmol) of elemental magnesium (3). A 50% magnesium sulfate solution (500 mg/mL) has an osmolarity of 4000 mOsm/L (50), so it must e diluted to a 10% (100 mg/mL) or 20% (200 mg/mL) solution for intravenous use. Saline solutions should e used as the diluent for magnesium sulfate. Ringer's solutions should not e used ecause the calcium in Ringer's solutions will coun teract the actions of the infused magnesium. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 19)03-May-05 07:50:15

Ovid: ICU Book REPLACEMENT PROTOCOLS The following magnesium replacement protocols are recommended for patients with normal renal function (51). Mild, Asymptomatic Hypomagnesemia The following guidelines can e used for patients with mild hypomagnesemia and n o apparent complications (51): P.668 q Assume a total magnesium deficit of 1 to 2 mEq/kg. q Because 50% of the infused magnesium can e lost in the urine, assume that the t otal magnesium requirement is twice the magnesium deficit. q Replace 1 mEq/kg for the first 24 hours, and 0.5 mEq/kg daily for the next 3 to 5 days. q

Moderate Hypomagnesemia The following therapy is intended for patients with a serum magnesium level less than 1 mEq/L or when hypomagnesemia is accompanied y other electrolyte a norma lities: q Add 6 g MgSO4 (48 mEq Mg) to 250 or 500 mL isotonic saline and infuse over 3 hou rs. q Follow with 5 g MgSO4 (40 mEq Mg) in 250 or 500 mL isotonic saline infused over the next 6 hours. q Continue with 5 g MgSO4 every 12 hours ( y continuous infusion) for the next 5 d ays. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 19)03-May-05 07:50:15

If the serum magnesium is greater than 1 mEq/L, oral magnesium can eplacement therapy.

e used for r

Ovid: ICU Book Life-Threatening Hypomagnesemia When hypomagnesemia is associated with serious cardiac arrhythmias or generalize d seizures, do the following: q Infuse 2 g MgSO4 (16 mEq Mg) intravenously over 2 minutes. q Follow with 5 g MgSO4 (40 mEq Mg) in 250 or 500 mL isotonic saline infused over the next 6 hours. q Continue with 5 g MgSO4 every 12 hours ( y continuous infusion) for the next 5 d ays. Serum magnesium levels will rise after the initial magnesium olus ut will egi n to fall after 15 minutes. Therefore, it is important to follow the olus dose with a continuous magnesium infusion. Serum magnesium levels may normalize after 1 to 2 days, ut it will take several days to replenish the total ody magnesiu m stores. Hypomagnesemia and Renal Insufficiency Hypomagnesemia is not common in renal insufficiency ut can occur when severe or chronic diarrhea is present and the creatinine clearance is greater than 30 mL/ minute. When magnesium is replaced in the setting of renal insufficiency, no mor e than 50% of the magnesium in the standard replacement protocols should e admi nistered, (51) and the serum magnesium should e monitored carefully. MAGNESIUM ACCUMULATION Magnesium accumulation occurs almost exclusively in patients with impaired renal function. In one survey of hospitalized patients, hypermagnesemia P.669 (i.e., a serum magnesium greater than 2 mEq/L) was reported in 5% of patients (15). PREDISPOSING CONDITIONS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 19)03-May-05 07:50:15

Ovid: ICU Book Hemolysis The magnesium concentration in erythrocytes is approximately three times greater than that in serum (1), so hemolysis can increase the plasma magnesium. This ca n occur either in vivo from a hemolytic anemia or in vitro from traumatic disrup tion of erythrocytes during phle otomy. In hemolytic anemia, the serum magnesium is expected to rise y 0.1 mEq/L for every 250 mL of erythrocytes that lyse com pletely (1), so hypermagnesemia is expected only with massive hemolysis. Renal Insufficiency The renal excretion of magnesium ecomes impaired when the creatinine clearance falls elow 30 mL/minute (52). However, hypermagnesemia is not a prominent featu re of renal insufficiency unless magnesium intake is increased. Others Other conditions that can predispose to mild hypermagnesemia are dia etic ketoac idosis (transient), adrenal insufficiency, hyperparathyroidism, and lithium into xication (52). CLINICAL FEATURES The clinical consequences of progressive hypermagnesemia are listed elow (52). Serum Magnesium Clinical Finding 4.0 mEq/L >5.0 mEq/L >10 mEq/L >13 mEq/L Hyporeflexia Prolonged atrioventricular conduction Complete heart lock Cardiac arrest http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 19)03-May-05 07:50:15

Ovid: ICU Book Magnesium has een descri ed as nature's physiologic calcium locker (53), and m ost of the serious consequences of hypermagnesemia are due to calcium antagonism in the cardiovascular system. Most of the cardiovascular depression is the resu lt of cardiac conduction delays. Depressed contractility and vasodilation are no t prominent. MANAGEMENT Hemodialysis is the treatment of choice for severe hypermagnesemia. Intravenous calcium gluconate (1 g IV over 2 to 3 minutes) can P.670 e used to antagonize t he cardiovascular effects of hypermagnesemia temporarily, until dialysis is star ted (52). If fluids are permissi le and some renal function is preserved, aggres sive volume infusion com ined with furosemide may e effective in reducing the s erum magnesium levels in less advanced cases of hypermagnesemia. REFERENCES GENERAL REVIEWS 1. Elin RJ. Magnesium meta olism in health and disease. Dis Mon 1988;34:161219 (2 38 references). 2. White RE, Hartzell HO. Magnesium ions in cardiac function. Biochem Pharmacol 1989;38:859867 (109 references). 3. McLean RM. Magnesium and its therapeutic uses: a review. Am J Med 1994;96:6376 (130 references). 4. Marino PL. Calcium and magnesium in critical illness: a practical approach. I n: Sivak ED, Higgins TL, Seiver A, eds. The high risk patient: management of the critically ill. Baltimore: Williams & Wilkins, 1995;11831195 (87 references). MAGNESIUM BALANCE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 19)03-May-05 07:50:15

Ovid: ICU Book 5. Reinhart RA. Magnesium meta olism: a review with special reference to the rel ationship etween intracellular content and serum levels. Arch Intern Med 1988;1 48:24152420. 6. Elin RJ. Assessment of magnesium status. Clin Chem 1987;33:19651970. 7. Lowenstein FW, Stanton MF. Serum magnesium levels in the United States 1971 19 74. J Am Coll Nutr 1986;5:399414. 8. Kroll MH, Elin RJ. Relationship etween magnesium and protein concentrations in serum. Clin Chem 1985;31:244246. 9. Alvarez-Lefmans FJ, Giraldez F, Gamino SM. Intracellular free magnesium in ex cita le cells: its measurement and its iologic significance. Can J Physiol Phar macol 1987;65:915925.

11. Shils ME. Experimental human magnesium depletion. Medicine 1969;48:6182. MAGNESIUM DEPLETION Reviews 12. Salem M, Munoz R, Chernow B. Hypomagnesemia in critical illness. Crit Care C lin 1991;7:225252 (154 references. 13. Reinhart RA. Magnesium deficiency: recognition and treatment in the emergenc y medicine setting. Am J Emerg Med 1992;10:7883. 14. Whang RW, Hampton EM, Whang DD. Magnesium homeostasis and clinical disorders of magnesium deficiency. Ann Pharmacother 1994;28:220225 (66 references). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 19)03-May-05 07:50:15

10. Munoz R, Khilnani P, Salem M, et al. Ionized hypomagnesemia: a frequent pro lem in critically ill neonates. Crit Care Med 1991;19:S48.

Ovid: ICU Book P.671 Prevalence 15. Whang R, Ryder KW. Frequency of hypomagnesemia and hypermagnesemia: requeste d vs routine. JAMA 1990;263:30633064. 16. Martin BJ, Black J, McLelland AS. Hypomagnesemia in elderly hospital admissi ons: a study of clinical significance. Q J Med 1991;78:177184. 17. Ru eiz GJ, Thill-Baharozian M, Hardie D, Carlson RW. Association of hypomagn esemia with mortality in acutely ill medical patients. Crit Care Med 1993;21:2032 09. 18. Ryzen E, Wagers PW, Singer FR, Rude RK. Magnesium deficiency in a medical IC U population. Crit Care Med 1985;13:1921. 19. Chernow B, Bam erger S, Stoiko M, et al. Hypomagnesemia in patients in posto perative intensive care. Chest 1989;95:391397. 20. Whang R. Magnesium deficiency: pathogenesis, prevalence, and clinical implic ations. Am J Med 1987;82(3A):2429. Predisposing Factors 21. Dykner T, Wester PO. Potassium/magnesium depletion in patients with cardiova scular disease. Am J Med 1987;82(Suppl 3A):1117. 22. Hollifield JW. Thiazide treatment of systemic hypertension: effects on serum magnesium and ventricular ectopic activity. Am J Med 1989;63:22G25G. 23. Ryan MP: Diuretics and potassium/magnesium depletion. Am J Med 1987;82(Suppl http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 19)03-May-05 07:50:15

Ovid: ICU Book 3A):3847. 24. Shah GM, Kirschen aum MA. Renal magnesium wasting associated with therapeuti c agents. Miner Electrolyte Meta 1991;17:5864. 25. Zaloga G, Chernow B, Pock A, et al. Hypomagnesemia is a common complication of aminoglycoside therapy. Surg Gynecol O stet 1984;158:561564. 26. Whang R, Oci TO, Watawa e A. Frequency of hypomagnesemia in hospitalized pat ients receiving digitalis. Arch Int Med 1985;145:655656. 27. Whyte K, Addis GJ, Whitesmith R, Reid JL. Adrenergic control of plasma magne sium in man. Clin Sci 1987;72:135138. 28. Ashraf M, Scotchel PL, Krall JM, et al. Cis-platinum-induced hypomagnesemia and peripheral neuropathy. Gynecol Oncol 1983;16:309318. 29. Thompson CB, June CH, Sullied KM, Themes ED. Association etween cyclosporin neurotoxicity and hypomagnesemia. Lancet 1984;ii:11161120. 30. Balesteri FJ. Magnesium meta olism in the critically ill. Crit Care Clin 198 5;5:217 226. 31. Martin HE. Clinical magnesium deficiency. Ann N Y Acad Sci 1969;162:891903.

33. Kassirer JP, Hrick DE, Cohen JJ. Repairing ody fluids: principles and pract ice. Philadelphia: WB Saunders, 1989;118129. 34. Sjogren A, Floren CH, Nilsson A. Magnesium deficiency in IDDM related to lev el of http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 19)03-May-05 07:50:15

32. Dyckner T, Ek B, Nyhlin H, Wester PO. Aggravation of thiamine deficiency magnesium depletion. Acta Med Scand 1985;218:129131.

Ovid: ICU Book glycosylated hemoglo in. Dia etes 1986;35:459463. 35. Lau K. Magnesium meta olism: normal and a normal. In: Arieff AI, DeFronzo RA , eds. Fluids, electrolytes, and acid- ase disorders. New York: Churchill Living stone, 1985;575623. 36. A raham AS, Rosenmann D, Kramer M, et al. Magnesium in the prevention of let hal arrhythmias in acute myocardial infarction. Arch Intern Med 1987;147:753755. P.672 Clinical Findings 37. Whang R, Oei TO, Aikawa JK, et al. Predictors of clinical hypomagnesemia. Ar ch Intern Med 1984;144:17941796. 38. Anast CS, Winnacker JL, Forte LR. Impaired release of parathyroid hormone in magnesium deficiency. J Clin Endocrinol Meta 1976;42:707717. 39. Rude RK, Oldham SB, Singer FR. Functional hypoparathyroidism and parathyroid hormone end-organ resistance in human magnesium deficiency. Clin Endocrinol 197 6;5:209224. 40. Gra er ML, Schulman G. Hypomagnesemic hypocalcemia independent of parathyroi d hormone. Ann Intern Med 1986;104:804805. 41. Dominiquez JH, Gray RW, Lemann J Jr. Dietary phosphate deprivation in women and men: effects on mineral and acid alances, parathyroid hormone and meta olis m of 25-OH-vitamin D. J Clin Endocrinol Meta 1976;43:10561068. 42. Cohen L, Kitzes R. Magnesium sulfate and digitalis-toxic arrhythmias. JAMA 1 983;249:28082810. 43. French JH, Thomas RG, Siskind AP, et al. Magnesium therapy in massive digoxi n http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 19)03-May-05 07:50:15

Ovid: ICU Book intoxication. Ann Emerg Med 1984;13:562566. 44. Tsivoni DT, Keren A. Suppression of ventricular arrhythmias y magnesium. Am J Cardiol 1990;65:13971399. 45. Langley WF, Mann D. Central nervous system magnesium deficiency. Arch Intern Med 1991;151:593596. Diagnosis 46. Fleming CR, George L, Stoner GL, et al. The importance of urinary magnesium levels in patients with gut failure. Mayo Clin Proc 1996;71:2124. 47. Rasmussen HS, McNair P, Goransson L, et al. Magnesium deficiency in patients with ischemic heart disease with and without acute myocardial infarction uncove red y an intravenous loading test. Arch Intern Med 1988;148:329332. 48. Clague JE, Edwards RHT, Jackson MJ. Intravenous magnesium loading in chronic fatigue syndrome. Lancet 1992;340:124125. Magnesium Replacement Therapy 49. Dipalma JR. Magnesium replacement therapy. Am Fam Physician 1990;42:173176. 50. Trissel LA. Hand ook on injecta le drugs. 8th ed. Bethesda, MD: American Soc iety of Hospital Pharmacists, 1994;633639. 51. Oster JR, Epstein M. Management of magnesium depletion. Am J Nephrol 1988;8: 349354. Magnesium Acculmulation http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 19)03-May-05 07:50:15

Ovid: ICU Book 52. Van Hook JW. Hypermagnesemia. Crit Care Clin 1991;7:215223.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 19)03-May-05 07:50:15

53. Iseri LT, French JH. Magnesium: nature's physiologic calcium rt J 1984;108:188193.

locker. Am Hea

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 43 CALCIUM AND PHOSPHORUS Calcium and phosphorus are responsi le for much of the structural integrity of t he ony skeleton. Although neither is found in a undance in the soft tissues, o th play an important role in vital cell functions. Phosphorus participates in ae ro ic energy production, whereas calcium participates in several diverse process es, such as lood coagulation, neuromuscular transmission, and smooth muscle con traction. Considering the important functions of these electrolytes, it is surpr ising that a normalities in calcium and phosphorus alance are so well tolerated . CALCIUM Calcium is the most a undant electrolyte in the human ody (the average adult ha s more than half a kilogram of calcium), ut 99% is in one (1). In the soft tis sues, calcium is 10,000 times more concentrated in the extracellular fluids (2). This preference for the extracellular fluid seems odd ecause calcium seems mos t involved with smooth muscle contraction (3), which is an intracellular process . PLASMA CALCIUM The calcium in plasma is present in three forms, as depicted in Figure 43.1. App roximately 50% of the calcium is ound to plasma proteins, and al umin is respon si le for 80% of the protein inding. An additional 5 to 10% is chelated to plas ma anions such as sulfates and phosphates. The remainder is present as free or u nattached calcium ions. This ionized fraction is the physiologically active frac tion of calcium in plasma. The concentration of total and ionized calcium in pla sma P.674 is shown in Ta le 43.1. These values may vary slightly in different cl inical la oratories. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 21)03-May-05 07:49:38

Ovid: ICU Book Figure 43.1. The three fractions of calcium in plasma and the contri ution of ea ch to the total plasma calcium concentration. The column on the right shows how a decrease in plasma al umin can reduce the total plasma calcium without affecti ng the ionized calcium. TABLE 43.1. NORMAL RANGES FOR CALCIUM AND PHOSPHOROUS IN BLOOD Total versus Ionized Calcium The calcium assay used y most clinical la oratories measures all three fraction s of calcium, which can e misleading. The column on the right in Figure 43.1 de monstrates the effects of a http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 21)03-May-05 07:49:38

Ovid: ICU Book decrease in the concentration of al umin in plasma. Because al umin is responsi le for 80% of the protein- ound calcium in plasma, a decrease in al umin decreas es the amount of calcium in the protein- ound fraction. The total calcium in pla sma decreases y the same amount, ut the ionized calcium remains unchanged. Bec ause the ionized calcium is the physiologically active fraction, the hypocalcemi a caused y hypoal uminemia is not physiologically significant. The hypocalcemia that is physiologically significant is ionized hypocalcemia. Various correction factors have een proposed for determining the P.675 effects of hypoal uminemia on the plasma calcium concentration. However, none of these correction factors have proven relia le (4,5), and the only method of identifying true (ionized) hy pocalcemia in the face of hypoal uminemia is to measure the ionized fraction of calcium in plasma. IONIZED CALCIUM MEASUREMENT Ionized calcium can e measured in whole lood, plasma, or serum with ion-specif ic electrodes that are now availa le in most clinical la oratories (4,5 and 6). The normal concentration of ionized calcium in plasma is shown in Ta le 43.1. Blood Collection Several conditions can alter the level of ionized calcium in lood samples (5). Acidosis decreases the inding of calcium to al umin and increases the ionized c alcium, whereas alkalosis has the opposite effect. Loss of car on dioxide from a lood sample could falsely lower the ionized calcium, so it is important to avo id gas u les in the lood sample. Anticoagulants (e.g., heparin, citrate, and EDTA) can ind calcium, so lood samples should not e placed in collection tu e s that contain these anticoagulants. Tu es with red stoppers (red top tu es) conta in silicone and are adequate for measuring ionized calcium in serum samples. Hep arinized syringes can e used for measuring ionized calcium in whole lood. Alth ough heparin also inds calcium, the effect is minimal if the heparin level is l ess than 15 U/ mL of lood (5).

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 21)03-May-05 07:49:38

IONIZED HYPOCALCEMIA Ionized hypocalcemia has 7). The common disorders ICU are listed in Ta le emia in outpatients, ut

een reported in 50 to 65% of admissions to the ICU (6, associated with ionized hypocalcemia in patients in the 43.2. Hypoparathyroidism is a leading cause of hypocalc it is not

Ovid: ICU Book

TABLE 43.2. CAUSES OF IONIZED HYPOCALCEMIA IN THE ICU PREDISPOSING CONDITIONS Magnesium Depletion Magnesium depletion promotes hypocalcemia y inhi iting parathormone secretion a nd reducing end-organ responsiveness to parathormone (see Chapter 42). Hypocalce mia from magnesium depletion is refractory to calcium replacement therapy, and m agnesium repletion often corrects the hypocalcemia without calcium replacement. Sepsis Sepsis is a common cause of hypocalcemia in the ICU (6,7,8 and 9). The mechanism is unclear, ut it may involve an increase in calcium inding to al umin caused y elevated levels of circulating free fatty acids. Hypocalcemia is independent of the vasodilation that accompanies sepsis (9), and thus the clinical signific ance of the hypocalcemia in sepsis is questioned. Alkalosis http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 21)03-May-05 07:49:38

P.676 a consideration in the ICU unless neck surgery has .

een performed recently

Ovid: ICU Book As mentioned earlier, alkalosis promotes the inding of calcium to al umin and c an reduce the fraction of ionized calcium in lood. Symptomatic hypocalcemia is more common with respiratory alkalosis than with meta olic alkalosis. Infusions of sodium icar onate can also e accompanied y ionized hypocalcemia ecause ca lcium directly inds to the infused icar onate. Blood Transfusions Ionized hypocalcemia has een reported in 15% of patients receiving lood transf usions (6). The mechanism is calcium inding y the citrate preservative in ank ed lood. Hypocalcemia from lood transfusions usually is transient, and resolve s when the infused citrate is meta olized y the liver and kidneys (1). In patie nts with renal or hepatic failure, a more prolonged hypocalcemia can result. Alt hough hypocalcemia from lood transfusions could impede lood coagulation, this is not considered to e a significant effect and calcium infusions are no longer recommended in massive lood transfusions. Drugs A num er of drugs can ind calcium and promote ionized hypocalcemia (1,6,7). The ones most often used in the ICU are aminoglycosides, cimetidine, heparin, and t heophylline. Renal Failure Ionized hypocalcemia can accompany renal failure as a result of phosphate retent ion and impaired conversion of vitamin D to its active P.677 form in the kidneys . The treatment in this setting is aimed at lowering the phosphate levels in lo od with antacids that lock phosphorus a sorption in the small owel. However, t he value of this practice is unproven. The acidosis in renal failure can decreas e the inding of calcium to al umin, so hypocalcemia in renal failure does not i mply ionized hypocalcemia. Pancreatitis Severe pancreatitis can produce ionized hypocalcemia through several mechanisms. The prognosis is adversely affected y the appearance of hypocalcemia (10), alt hough a causal relationship has not een proven. CLINICAL MANIFESTATIONS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 21)03-May-05 07:49:38

Ovid: ICU Book The clinical manifestations of hypocalcemia are related to enhanced cardiac and neuromuscular excita ility and reduced contractile force in cardiac muscle and v ascular smooth muscle. Neuromuscular Excita ility Hypocalcemia can e accompanied y hyperreflexia, generalized seizures, and teta ny. Chvostek's and Trousseau's signs are often listed as manifestations of hypoc alcemia. However, Chvostek's sign is nonspecific (it is present in 25% of normal adults), and Trousseau's sign is insensitive (it can e a sent in 30% of patien ts with hypocalcemia) (1). Cardiovascular Effects The cardiovascular complications of hypocalcemia include hypotension, decreased cardiac output, and ventricular ectopic activity. These complications are rarely seen in mild cases of ionized hypocalcemia (i.e., ionized calcium 0.8 to 1.0 mm ol/L). However, advanced stages of ionized hypocalcemia (i.e., ionized calcium l ess than 0.65 mmol/L) can e associated with ventricular tachycardia and refract ory hypotension (1). CALCIUM REPLACEMENT THERAPY The treatment of ionized hypocalcemia should e directed at the underlying cause of the pro lem. However, symptomatic hypocalcemia is considered a medical emerg ency (1), and the treatment of choice is intravenous calcium. The calcium soluti ons and dosage recommendations P.678 for intravenous calcium replacement are sho wn in Ta le 43.3. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 21)03-May-05 07:49:38

Ovid: ICU Book TABLE 43.3. INTRAVENOUS CALCIUM REPLACEMENT THERAPY Calcium Salt Solutions The two most popular calcium solutions for intravenous use are 10% calcium chlor ide and 10% calcium gluconate. Both solutions have the same concentration of cal cium salt (i.e., 100 mg/ mL), ut calcium chloride contains three times more ele mental calcium than calcium gluconate. One 10-mL ampule of 10% calcium chloride contains 272 mg (13.6 mEq) of elemental calcium, whereas one 10-mL ampule of 10% calcium gluconate contains only 90 mg (4.6 mEq) of elemental calcium (11). Dosage Recommendations The intravenous calcium solutions are hyperosmolar and should e given through a large central vein if possi le. If a peripheral vein is used, calcium gluconate is the preferred solution ecause of its lower osmolarity (Ta le 43.3). A olus dose of 200 mg elemental calcium (diluted in 100 mL isotonic saline and given o ver 10 minutes) should raise the total serum calcium y 1 mg/dL (1), ut levels will egin to fall after 30 minutes. Therefore, the olus dose of calcium should e followed y a continuous infusion at a dose rate of 1 to 2 mg (elemental cal cium) per kilogram of ody weight per hour. This should e continued for at leas t 6 hours. Su sequent doses of calcium should e guided y the level of ionized calcium in lood. Maintenance Therapy The daily maintenance dose of calcium is 2 to 4 g in adults. This can e adminis tered orally http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 21)03-May-05 07:49:38

Ovid: ICU Book using calcium car onate (e.g., Oscal) or calcium gluconate ta lets (500 mg calci um per ta let). Caution Intravenous calcium replacement can e risky in select patient populations. Calc ium infusions can promote vasoconstriction and ischemia P.679 in any of the vita l organs (12). The risk of calcium-induced ischemia should e particularly high in patients with low cardiac output who are already vasoconstricted. In addition , aggressive calcium replacement can promote intracellular calcium overload, whi ch can produce a lethal cell injury (13), particularly in patients with circulat ory shock. Because of these risks, calcium infusions should e used judiciously. Intravenous calcium is indicated only for patients with symptomatic hypocalcemi a or an ionized calcium level elow 0.65 mmol/L (1). HYPERCALCEMIA Hypercalcemia is not nearly as common as hypocalcemia: it is reported in less th an 1% of hospitalized patients (14). In 90% of cases, the underlying cause is hy perparathyroidism or malignancy (15). Less common causes include prolonged immo ilization, thyrotoxicosis, and drugs (lithium, thiazide diuretics). Malignancy i s the most common cause of severe hypercalcemia (i.e., total serum calcium a ove 14 mg/dL or ionized calcium a ove 3.5 mmol/L) (16).

q Gastrointestinal (GI): nausea, vomiting, constipation, ileus, and pancreatitis q Cardiovascular: hypovolemia, hypotension, and shortened QT interval q Renal: polyuria and nephrocalcinosis q Neurologic: confusion and depressed consciousness, including coma http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 21)03-May-05 07:49:38

CLINICAL MANIFESTATIONS The manifestations of hypercalcemia usually are nonspecific and can ed as follows:

e categoriz

Ovid: ICU Book These manifestations can ecome evident when the total serum calcium rises a ove 12 mg/dL (or the ionized calcium rises a ove 3.0 mmol/L), and they are almost a lways present when the serum calcium is greater than 14 mg/dL (or the ionized ca lcium is a ove 3.5 mmol/L) (16). MANAGEMENT Treatment is indicated when the hypercalcemia is associated with adverse effects , or when the serum calcium is greater than 14 mg/dL (ionized calcium a ove 3.5 mmol/L). The management of hypercalcemia is summarized in Ta le 43.4 (15,16). TABLE 43.4. MANAGEMENT OF SEVERE HYPERCALCEMIA Saline Infusion Hypercalcemia usually is accompanied y hypercalciuria, which produces an osmoti c diuresis. This eventually leads to hypovolemia, P.680 which reduces calcium ex cretion in the urine and precipitates a rapid rise in the serum calcium. Therefo re, volume infusion to correct hypovolemia and promote renal calcium excretion i s the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 21)03-May-05 07:49:38

Ovid: ICU Book first goal of management for hypercalcemia. Isotonic saline is recommended for t he volume infusion ecause natriuresis itself promotes renal calcium excretion. Furosemide Saline infusion will not return the calcium to normal levels. To do this require s the addition of furosemide (40 to 80 mg IV every 2 hours) to further promote u rinary calcium excretion. The goal is an hourly urine output of 100 to 200 mL/mi nute. The hourly urine output must e replaced with isotonic saline. Failure to replace urinary volume losses will favor a return to hypovolemia, which is count erproductive. Calcitonin Although saline and furosemide will correct the hypercalcemia acutely, this appr oach does not treat the underlying cause of the pro lem, which (in malignancy) i s enhanced one resorption. Calcitonin is a naturally occurring hormone that inh i its one resorption. It is availa le as salmon calcitonin, which is given su c utaneously or intramuscularly in a dose of 4 U/kg every 12 hours. The response i s rapid (onset within a few hours), ut the effect is mild (the maximum drop in serum calcium is 0.5 mmol/L). Hydrocortisone Corticosteroids can reduce the serum calcium y impeding the growth of lymphoid neoplastic tissue and enhancing the actions of P.681 vitamin D. Steroids are usu ally com ined with calcitonin and can e particularly useful in the hypercalcemi a associated with multiple myeloma or renal failure (15,16). The standard regime n uses hydrocortisone, 200 mg IV daily in 2 or 3 divided doses. Pamidronate Calcitonin can e used for rapid reduction of serum calcium, ut the mild respon se will not keep the calcium in the normal range. A group of compounds known as iphosphonates (pyrophosphate derivatives) are more potent inhi itors of one re sorption and maintain a normal serum calcium. However, their onset of action is delayed, and thus they are not useful when rapid control of serum calcium is des ired. Pamidronate is currently the iphosphonates of choice for the management o f severe hypercalcemia (15). The dose is 90 mg, which is given y continuous int ravenous infusion over http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 21)03-May-05 07:49:38

Ovid: ICU Book

Plicamycin Plicamycin (formerly mithramycin) is an antineoplastic agent that inhi its one resorption. It is similar to pamidronate in that it is more potent than calciton in ut has a delayed onset of action. The dose is 25 g/kg (intravenously over 4 h ours), which can e repeated in 24 to 48 hours if necessary. Because of the pote ntial for serious side effects (e.g., one marrow suppression), plicamycin has l argely een replaced y pamidronate. Dialysis Dialysis is effective in removing calcium in patients with renal failure. Either hemodialysis or peritoneal dialysis can e used (15). PHOSPHORUS The average adult has 500 to 800 g of phosphorus (17). Most phosphorus is contai ned in organic molecules such as phospholipids and phosphoproteins, and 85% is l ocated in the ony skeleton. The remaining 15% in soft tissues is present as fre e, inorganic phosphorus. Unlike calcium, inorganic phosphorus is predominantly i ntracellular in location, where it participates in glycolysis and high energy ph osphate production. The normal concentration of inorganic phosphorus in plasma i s shown in Ta le 43.1. This reference range pertains only to adults; the normal range in children is higher (17). HYPOPHOSPHATEMIA Hypophosphatemia (serum PO4 less than 2.5 mg/dL or 0.8 mmol/L) can e the result of an intracellular shift of phosphorus, an increase P.682 in the renal excreti on of phosphorus, or a decrease in phosphorus a sorption from the GI tract. Most cases of hypophosphatemia are due to movement of PO4 into cells. PREDISPOSING CONDITIONS Glucose Loading http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 21)03-May-05 07:49:38

24 hours. The peak effect is seen in 4 to 5 days, and the dose can t that time if necessary.

e repeated a

Ovid: ICU Book The movement of glucose into cells is accompanied y a similar movement of PO4 i nto cells, and if the extracellular content of PO4 is marginal, this intracellul ar PO4 shift can result in hypophosphatemia. Glucose loading is the most common cause of hypophosphatemia in hospitalized patients (18,19), usually seen during refeeding in alcoholic, malnourished, or de ilitated patients. It can occur with oral feedings, enteral tu e feedings, or with total parenteral nutrition. The i nfluence of parenteral nutrition on serum PO4 levels is shown in Figure 43.2. No te the gradual decline in the serum PO4 and the severe degree of hypophosphatemi a (serum PO4 elow 1 mg/dL) seen after 7 days of intravenous nutrition. This tre nd in serum PO4 is one of the reasons why parenteral nutrition regimens are adva nced gradually for the first few days after eing started. Figure 43.2. The cumulative effect of total parenteral nutrition (TPN) on the se rum phosphate level. (Data from Knochel JP. The pathophysiology and clinical cha racteristics of severe hypophosphatemia. Arch Intern Med 1977;137:203220.) As mentioned, oral and enteral feedings create a similar risk of hypophosphatemi a, particularly in de ilitated or malnourished patients. In fact, hypophosphatem ia may e responsi le for the progressive weakness and inanition that characteri zes the refeeding syndrome in malnourished patients (20). Respiratory Alkalosis Respiratory alkalosis can increase intracellular pH, and this accelerates glycol ysis. The increase in glucose use then increases glucose and http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 21)03-May-05 07:49:38

Ovid: ICU Book P.683 phosphorus movement into cells. This may e an important source of hypopho sphatemia in ventilator-dependent patients ecause overventilation and respirato ry alkalosis is common in these patients. -Receptor Agonists Stimulation of -adrenergic receptors can move PO4 into cells and promote hypophos phatemia. This effect is evident in patients treated with -agonist ronchodilator s. In one study of patients with acute asthma who were treated aggressively with ne ulized al uterol (2.5 mg every 30 minutes), the serum PO4 decreased y 1.25 mg/dL (0.4 mmol/L) 3 hours after the onset of therapy (21). However, the signifi cance of this effect is unclear. Sepsis There is a common association etween septicemia and hypophosphatemia in some st udies (19). A causal relationship is unproven, ut sepsis could promote a transc ellular shift of PO4 as a result of elevated levels of endogenous catecholamines . Phosphorus-Binding Agents Aluminum can form insolu le complexes with inorganic phosphorus. As a result, al uminumcontaining compounds, such as sucralfate (Carafate), or antacids that cont ain aluminum hydroxide (e.g., Amphojel) can impede the a sorption of phosphate i n the upper GI tract and promote phosphate depletion (18). The major concern is sucralfate, which is eing used with increasing frequency for the prophylaxis of stress ulcer leeding in critically ill patients (see Chapter 6). Sucralfate ad ministration is associated with an increased incidence of hypophosphatemia in pa tients in the ICU (21). However a direct cause-and-effect relationship etween s ucralfate and phosphate depletion remains to e esta lished. Dia etic Ketoacidosis The osmotic diuresis from glycosuria promotes the urinary loss of PO4, and patie nts with prolonged or severe hyperglycemia are often phosphate depleted. As ment ioned in Chapter 37, phosphate depletion is almost universal in patients who hav e dia etic ketoacidosis, ut it does not ecome evident until insulin therapy dr ives PO4 into cells. Because phosphate supplementation does not alter the outcom e in dia etic ketoacidosis (see Chapter 37), the significance of the phosphate d epletion in this disorder is unclear. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 21)03-May-05 07:49:38

Ovid: ICU Book CLINICAL MANIFESTATIONS Hypophosphatemia is often clinically silent, even when the serum PO4 falls to ex tremely low levels. In one study of patients with severe hypophosphatemia (i.e., serum PO4 less than 1.0 mg/dL), none of the P.684 patients showed evidence of h armful effects (23). Despite the apparent lack of harm, phosphate depletion crea tes a risk for impaired energy production in all aero ic cells. Aero ic Energy Production Phosphate depletion has several effects that could impair cellular energy produc tion. These are summarized in Figure 43.3. To egin with, each of the following determinants of systemic oxygen delivery can e adversely affected y phosphate depletion. Figure 43.3. The effects of phosphate depletion that can impair cellular energy production. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 21)03-May-05 07:49:38

Ovid: ICU Book q Cardiac output: Phosphate depletion can impair myocardial contractility and redu ce cardiac output. Hypophosphatemic patients with heart failure have shown impro ved cardiac performance after phosphate supplementation (24). q Hemoglo in: Reduction of high energy phosphate production from glycolysis in ery throcytes can reduce the deforma ility of red cells. This may explain why severe hypophosphatemia can e accompanied y a hemolytic anemia (18). q Oxyhemoglo in dissociation: Phosphate depletion is accompanied y depletion of 2 ,3-diphosphoglycerate, and this shifts the oxyhemoglo in dissociation curve to t he left. When this occurs, hemoglo in is less likely to release oxygen to the ti ssues. P.685 In addition to the adverse effects on tissue oxygen availa ility, phosphat e depletion can directly impede cellular energy production y reducing the avail a ility of inorganic phosphorus for high-energy phosphate production and decreas ing the activity of the glycolytic pathway. Muscle Weakness One of the possi le consequences of impaired energy production from phosphate de pletion is skeletal muscle weakness. Biochemical evidence of skeletal muscle dis ruption (e.g., elevated creatine kinase levels in lood) is common in patients w ith hypophosphatemia, ut overt muscle weakness is usually a sent (25). There is one report of respiratory muscle weakness and failure to wean from mechanical v entilation in patients with severe hypophosphatemia (26). However, other studies show that respiratory muscle weakness is common in hyposphaphatemia ut is not clinically significant in most patients (27). At present, the evidence linking p hosphate depletion with clinically significant skeletal muscle weakness is scant . PHOSPHORUS REPLACEMENT THERAPY Intravenous phosphorus replacement is recommended for all patients with severe h ypophosphatemia (i.e., serum PO4 elow 1.0 mg/dL or 0.3 mmol/L) and for patients with hypophosphatemia of any degree who also have cardiac dysfunction, respirat ory failure, muscle weakness, or impaired tissue oxygenation. The phosphate solu tions and dosage recommendations are shown in Ta le 43.5 (21,22). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 21)03-May-05 07:49:38

Ovid: ICU Book TABLE 43.5. INTRAVENOUS PHOSPHATE REPLACEMENT THERAPY Once the serum PO4 rises a ove 2 mg/dL, phosphate replacement can e continued u sing oral phosphate preparations like Neutra-Phos or K-Phos. The oral replacemen t dosage is 1200 to 1500 mg phosphorus daily. Remem er that sucralfate and phosp hate- inding antacids need to e discontinued when oral phosphate preparations a re used. The P.686 tendency for oral phosphate preparations to promote diarrhea limits the use of high-dosage oral PO4 replacement therapy. Maintenance Therapy The normal daily maintenance dosage of phosphate is 1200 mg if given orally (1). As shown in Ta le 43.6, the content of enteral feeding formulas varies widely, and thus thus enteral nutrition may not provide the daily phosphorus requirement s without additional phosphate supplementation. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 21)03-May-05 07:49:38

Ovid: ICU Book TABLE 43.6. THE PHOSPHORUS CONTENT OF ENTERAL TUBE FEEDINGS In patients who cannot tolerate enteral nutrition, daily phosphate requirements are provided intravenously. The IV maintenance dosage of PO4 is approximately 80 0 mg/day (this dosage is lower than the oral maintenance dosage ecause only 70% of orally administered phosphate is a sor ed from the GI tract). HYPERPHOSPHATEMIA Most cases of hyperphosphatemia in the ICU are the result of impaired PO4 excret ion from renal insufficiency or PO4 release from cells ecause of widespread cel l necrosis (e.g., rha domyolysis or tumor lysis). Hyperphosphatemia can also e seen in dia etic ketoacidosis, ut as descri ed earlier, this disorder is almost always accompanied y phosphate depletion, which ecomes evident after the onse t of insulin therapy. CLINICAL MANIFESTATIONS The clinical manifestations of hyperphosphatemia are not well documented. The mo st noted concern is the formation of insolu le calciumphosphate P.687 complexes t hat are deposited in soft tissues and promote tissue damage. However, little inf ormation exists on this su ject. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 21)03-May-05 07:49:38

Ovid: ICU Book MANAGEMENT There are two approaches to hyperphosphatemia. The first is to promote PO4 indi ng in the upper GI tract, which can lower the serum PO4 even in the a sence of a ny oral intake of phosphate (i.e., GI dialysis). Sucralfate or aluminum-containi ng antacids can e used for this purpose. In patients with significant hypocalce mia, calcium acetate ta lets (PhosLo, Braintree La s) can help raise the serum c alcium while lowering the serum PO4. Each calcium acetate ta let (667 mg) contai ns 8.45 mEq elemental calcium. The recommended dosage is 2 ta lets three times a day. The other approach to hyperphosphatemia is to enhance PO4 clearance with h emodialysis. This is reserved for patients with renal failure, and is rarely nec essary. REFERENCES GENERAL REFERENCES 1. Zaloga GP, Chernow B. Divalent cations: calcium, magnesium and phosphorus. In : Chernow B, ed. The pharmacologic approach to the critically ill patient. 3rd e d. Baltimore: Williams & Wilkins, 1994;777804 (190 references). CALCIUM REVIEWS 2. Marino PL. Calcium and magnesium in serious illness: a practical approach. In : Sivak ED, Higgins TL, Seiver A, eds. The high risk patient: management of the critically ill. Baltimore: Williams & Wilkins, 1995;11831195 (87 references). 3. Smith JB. Calcium homeostasis in smooth muscle. New Horiz 1996;4:218 (137 refe rences). PLASMA CALCIUM 4. Weaver CA. Assessing calcium status and meta olism. J Nutr 1990;120(Suppl 11) :14701473. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 21)03-May-05 07:49:38

Ovid: ICU Book 5. Foreman DT, Lorenzo L. Ionized calcium: it's significance and clinical useful ness. Ann Clin La Sci 1991;21:297304. 6. Cagir B, Walsh CB, Mahoney WD, Herz BL. Hypocalcemia in surgical critical car e patients: measurements of ionized calcium. Contemp Surg 1994;45:7178. HYPOCALCEMIA 7. Zaloga GP. Hypocalcemia in critically ill patients. Crit Care Med 1992;20:2512 62. 8. Desai TK, Carlson RW, Gehe MA. Prevalence and clinical implications of hypoc alcemia in acutely ill patients in a medical intensive care setting. Am J Med 19 88;84:209214. P.688 9. Burchard KW, Simms H, Ro inson A, et al. Hypocalcemia during sepsis. Arch Sur g 1992;127:265272. 10. Stein erg W, Tenner S. Acute pancreatitis. N Engl J Med 1994;330:11981210. 11. Trissel LA. Hand ook of injecta le drugs. 8th ed. Bethesda, MD: American Soc iety of Hospital Pharmacists, 1994;134148. 12. Shapiro MJ, Mistry M. Calcium regulation and nonprotective properties of cal cium in surgical ischemia. New Horiz 1995;4:134138. 13. Trump BF, Berezesky IK. Calcium-mediated cell injury and cell death. New Hor iz 1995;4:139150. HYPERCALCEMIA http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 21)03-May-05 07:49:38

Ovid: ICU Book 14. Shek CC, Natkunam A, Tsang V, et al. Incidence, Causes and mechanism of hype rcalcemia in a hospital population in Hong Kong. Q J Med 1990;77:12771285. 15. Mundy GR. Evaluation and treatment of hypercalcemia. Hosp Pract 1994;29:7986. 16. Bilezikian JP. Management Of Acute Hypercalcemia. N Engl J Med 1992;326:1196 1203. PHOSPHORUS REVIEWS 17. Peppers M, Gehe M, Desai T. Hypophosphatemia and hyperphosphatemia. Crit Ca re Clin 1991;7:201214 (64 references). HYPOPHOSPHATEMIA 18. Brown GR, Greenwood JK. Drug- and nutrition-induced hypophosphatemia: mechan isms and relevance in the critically ill. Ann Pharmacother 1994;28:626632. 19. Halevy J, Bulvik S. Severe hypophosphatemia in hospitalized patients. Arch I ntern Med 1988;148:153155. 20. Solomon SM, Kir y DF. The refeeding syndrome: a review. J Parenter Enteral N utr 1990;14:9097. 21. Bodenhamer J, Berstrom R, Brown D, et al. Frequently ne ulized eta-agonists for asthma: effects on serum electrolytes. Ann Emerg Med 1992;21:13371342. 22. Miller SJ, Simpson J. Medication-nutrient interactions: hypophosphatemia ass ociated with sucralfate in the intensive care unit. Nutr Clin Pract 1991;6:199201 . 23. King AL, Sica DA, Miller G, Pierpaoli S. Severe hypophosphatemia in a genera l hospital population. South Med J 1987;80:831835. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 21)03-May-05 07:49:38

Ovid: ICU Book 24. Davis SV, Olichwier KK, Chakko SC. Reversi le depression of myocardial perfo rmance in hypophosphatemia. Am J Med Sci 1988;295:183187. 25. Singhal PC, Kumar A, Desroches L, et al. Prevalence and predictors of rha do myolysis in patients with hypophosphatemia. Am J Med 1992;92:458464. 26. Agusti AG, Torres A, Estopa R, Agusti-Vidal A. Hypophosphatemia as a cause o f failed weaning: the importance of meta olic factors. Crit Care Med 1984;12:1421 43. 27. Gravelyn TR, Brophy N, Siegert C, Peters-Golden M. Hypophosphatemia-associat ed respiratory muscle weakness in a general inpatient population. Am J Med 1988; 84:870 875. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 21)03-May-05 07:49:38

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 44 ERYTHROCYTE TRANSFUSIONS One of the important discoveries, I elieve is the realization that anemia is wel l tolerated providing lood volume is maintained. --Daniel J. Ullyott. J Thoracic Cardiovasc Surg 1992;103:1007. The transfusion of erythrocyte products to corre ct anemia is one of the least scientific practices to e encountered in the ICU. One of the pro lems, as highlighted in Figure 44.1, is an apparent confusion a out the relative merits of lood cells versus lood volume in maintaining tissue via ility. As indicated in the introductory quote, the human organism suffers l ittle from deficits of red lood cells (anemia), as long as deficits of lood vo lume are prevented. The relative importance of lood volume over lood cells is shown y the fact that hypovolemia is a well-known cause of shock (impaired tiss ue oxygenation), whereas anemic shock is a nonentity. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 24)03-May-05 08:05:23

Ovid: ICU Book Figure 44.1. The statement on the left is a popular claim of the American Red Cr oss. However, the statement on the right is more accurate and would help foster a more judicious use of erythrocyte transfusions. This chapter egins y examining some of the asic assumptions for the use of er ythrocyte transfusions to correct anemia. Following this is a more practical dis cussion of the indications, methods, and complications of erythrocyte transfusio ns (1,2,3,4 and 5). ASSUMPTIONS The transfusion of erythrocyte products to correct anemia is rooted in the follo wing three assumptions. q It is possi le to accurately monitor anemia. q It is possi le to identify when an anemia impairs tissue oxygenation. q Erythrocyte transfusions improve tissue oxygenation. This section riefly evaluates the validity of these assumptions. MONITORING ANEMIA Anemia is a condition that is characterized y a decrease in the oxygen carrying P.692 capacity of lood (5). Because the oxygen carrying capacity of lood is d etermined y the mass of circulating red lood cells, anemia can e defined as a decrease in the red cell mass. The red cell mass is measured as a volume using chromium-tagged autologous erythrocytes; the normal values in adults are shown i n Ta le 44.1. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 24)03-May-05 08:05:23

Ovid: ICU Book TABLE 44.1. NORMAL VALUES FOR RED CELL PARAMETERS IN ADULTS Because red cell volume is not easily measured in the clinical setting, the clin ical definition of anemia is ased on the hematocrit and hemoglo in concentratio n in whole lood. Anemia is defined as any hematocrit or hemoglo in concentratio n that falls elow the normal ranges shown in Ta le 44.1. Relia ility The pro lem with the clinical definition of anemia is the influence of the plasm a volume on the hematocrit and hemoglo in concentration. For example, an increas e in the plasma volume (e. g., from hydration) will decrease the hematocrit and hemoglo in concentration and give the impression of a developing or worsening an emia, even though the oxygen carrying capacity of the lood (i.e., the red cell volume) remains unchanged. Therefore, changes in the hemoglo in and hematocrit c an e misleading ecause they may not reflect a change in the oxygen carrying ca pacity of the lood. This is revealed in clinical studies P.693 that show a poor correlation etween changes in the hematocrit and hemoglo in and changes in the red cell volume (6,7). Thus the clinical markers of anemia are unrelia le ecau se they are relative measurements (i. e., relative to plasma volume), whereas th e true marker of anemia is an a solute measurement (i.e., a deficit in red cell volume). This means that the first assumption stated at the eginning of the sec tion (i.e., that it is possi le to accurately monitor an anemia and determine it s severity) is invalid. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 24)03-May-05 08:05:23

Ovid: ICU Book CONSEQUENCES OF ANEMIA Cardiac Output A decrease in the num er of circulating erythrocytes will decrease the viscosity of lood (see Ta le 1.2), and according to the HagenPoiseuille equation shown in Equation 44.1 (where is the sym ol for viscosity), this will favor an increase in the rate of lood flow. (See Chapter 1 for a description of the HagenPoiseuill e equation.) Anemia will therefore e accompanied y an increase in cardiac output, and this helps maintain the rate of oxygen delivery to the peripheral tissues. The severi ty of anemia required to elicit an increase in cardiac output varies in differen t reports. Early o servations showed that cardiac output egins to rise when the hemoglo in falls elow 7 g/dL (8). However, su sequent studies have shown that hemoglo in levels as low as 4.5 g/dL ( oth acute and chronic) may not e accompa nied y an increase in cardiac output (9,10). Systemic Oxygenation Another compensatory response to anemia that helps maintain tissue oxygenation i s an increase in oxygen extraction from the systemic capillaries. This is demons trated in Figure 44.2. As the hematocrit decreases elow normal, there is a decr ease in systemic oxygen delivery (DO2), ut the oxygen extraction ratio (O2ER in creases, which helps maintain a constant oxygen uptake (VO2) into the tissues. H owever, when the hematocrit falls elow 10%, the increase in oxygen extraction i s no longer sufficient to maintain a constant VO2 and tissue oxygenation egins to fall. Further decreases in hematocrit eyond this point result in impaired ti ssue oxygenation, as shown y the increase in lood lactate levels (lower graph) . http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 24)03-May-05 08:05:23

Ovid: ICU Book Figure 44.2. The effects of progressive isovolemic anemia on oxygen delivery (DO 2), oxygen extraction ratio (O2ER), oxygen uptake (VO2), and lood lactate level s in anesthetized primates reathing 100% oxygen. (From Wilkerson DK et al. Oxyg en extraction ratio: a valid indicator of myocardial meta olism in anemia. J Sur g Res 1987;42:629634.) Figure 44.2 shows a threshold hematocrit of 10% (which corresponds to a hemoglo in of 3 g/ dL) for impaired tissue oxygenation. These results are taken from a s tudy of anesthetized primates reathing 100% P.694 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 24)03-May-05 08:05:23

Ovid: ICU Book oxygen (14). Similar results have een reported in anesthetized humans ( oth chi ldren and adults) reathing 100% oxygen (11,12) and in nonanesthetized primates reathing room air (13). However, impaired tissue oxygenation is expected at hig her levels of hematocrit and hemoglo in in awake patients who are reathing less than 100% oxygen and who have either cardiac dysfunction or hypermeta olism. Be cause these latter conditions are common in patients in the ICU, the level of P. 695 anemia that compromises tissue oxygenation is expected to vary in individual patients. TRANSFUSION TRIGGER As just indicated, a single hematocrit or hemoglo in level cannot e used as the transfusion trigger for all patients in the ICU. This is stated in the Clinical Practice Guideline on Elective Red Blood Cell Transfusion from the American Col lege of Physicians, which cautions to avoid an empiric, automatic transfusion th reshold, such as a hemoglo in less than 10 g/dL (2). Physiologic Parameters The physiologic markers of impaired tissue oxygenation (see Ta le 13.1) should e superior to the hemoglo in and hematocrit as transfusion triggers in individua l patients in the ICU (1,13,14,15 and 16). The oxygen extraction ratio has een specifically recommended for this purpose (14,15). As demonstrated in Figure 44. 2, the point at which the compensatory increase in oxygen extraction egins to f ail corresponds to an O2ER of 0.5 (50%). Therefore, an O2ER of 0.5 can e used a s a transfusion trigger (14,15). Oxygen extraction can e monitored continuously in the ICU y using pulse oximetry (for arterial O2 saturation) com ined with m ixed venous oximetry (for venous O2 saturation), as descri ed in Chapter 22. The refore, the second assumption stated at the eginning of the section (i.e., that it is possi le to identify when an anemia impairs tissue oxygenation) is not va lid if the hemoglo in and hematocrit are monitored, ut can e valid if the phys iologic parameters of tissue oxygenation are monitored. EFFICACY OF RED CELL TRANSFUSIONS Cardiac Output Just as anemia can increase the cardiac output y reducing lood viscosity, corr ecting anemia can reduce the cardiac output y increasing lood viscosity (16). This tendency to reduce http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 24)03-May-05 08:05:23

Ovid: ICU Book cardiac output limits the a ility of erythrocyte transfusions to improve systemi c oxygen transport. Systemic Oxygenation Erythrocyte transfusions have a varia le effect on systemic oxygenation in indiv idual patients (17,18 and 19). This is demonstrated in Figure 44.3. Each line in this graph shows the effects of erythrocyte transfusions on the hemoglo in conc entration and systemic oxygen uptake (VO2) in individual postoperative patients with isovolemic anemia. P.696 The num er of units of packed cells transfused in each patient is indicated y the num ers intersecting each line in the graph. Of the six patients who were transfused, three patients showed an increase in VO2 and three patients showed a decrease in VO2. Note that the direction of change i n the VO2 after the erythrocyte transfusions is unrelated to either the aseline level of anemia (the aseline hemoglo in is less than 7 g/dL in all patients) o r the magnitude of increase in the serum hemoglo in concentration. This graph de monstrates that an increase in hemoglo in concentration after erythrocyte transf usions does not necessarily indicate that tissue oxygenation has also improved. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 24)03-May-05 08:05:23

Ovid: ICU Book Figure 44.3. The effects of erythrocyte transfusions on hemoglo in concentration and systemic oxygen uptake (VO2) in individual postoperative patients with isov olemic anemia. Each line on the graph represents the changes recorded in an indi vidual patient. The num er of packed cells transfused in each case is indicated y the num ers that transect each line. (Data from personal o servations.)

Therefore, the third assumption stated at the eginning of this section (i.e., t hat erythrocyte transfusions improve tissue oxygenation) is valid for some ut n ot all patients. Furthermore, some parameter of systemic oxygenation, and not ju st the hematocrit and hemoglo in concentration, should e monitored to identify which patients enefit from lood transfusions. INDICATIONS FOR ERYTHROCYTE TRANSFUSIONS The following statements are relevant regarding the need for erythrocyte concent rates (packed red cells) in individual patients with normovolemic anemia. These statements represent a summary of the recommendations in References 1,2,3,4 and 5. q Erythrocyte transfusions are not indicated for the following: r Enhancing a sense of well- eing P.697 r Promoting wound healing r Expanding the intravascular volume r Correcting a hemoglo in elow 10 g/dL in a patient who has no evidence of ongoin g tissue ischemia (e.g., angina, ischemic stroke, hyperlactatemia) r Correcting anemia of any degree in patients who do not have cardiac dysfunction, coronary artery disease, or cere rovascular disease r Erythrocyte transfusions are indicated for the following: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 24)03-May-05 08:05:23

Ovid: ICU Book Evidence of impaired tissue oxygenation (e.g., VO2 less than 100 mL/ min/m2 or h yperlactatemia) or ongoing coronary or cere rovascular ischemia in patients with an adequate lood volume s An O2ER a ove 0.5 in patients with an adequate cardiac output s Correction of hemoglo in elow 7 g/dL in patients with a history of active coron ary artery disease, cere rovascular insufficiency, or significant cardiac dysfun ction When erythrocyte products are given to improve systemic oxygen transport in asym ptomatic patients, 1 or 2 units of packed cells should e transfused and the eff ects on oxygen transport varia les should e measured approximately 15 to 30 min utes after the transfusion is completed. If a significant rise in hemoglo in occ urs ut no improvement in systemic oxygen transport is seen, no further transfus ions should e given at that time. ERYTHROCYTE PRODUCTS All lood products containing erythrocytes are stored at 4 C using a liquid antic oagulant preservative that contains citrate, phosphate, and dextrose (CPD). The citrate inds ionized calcium and acts as an anticoagulant. The phosphate helps retard the reakdown of 2,3diphosphoglycerate, and the dextrose serves as a fuel source for the erythrocytes. Erythrocytes stored in CPD at 4 C are via le for at least 21 days (20). WHOLE BLOOD A unit of whole lood contains an average of 510 mL ( lood plus CPD solution) (2 1). Most lood anks will store whole lood only on request. Otherwise, the loo d is fractionated into erythrocyte and plasma fractions within a few hours of co llection. The separation of whole lood into its component fractions allows more efficient use of lood products to achieve specific transfusion goals. PACKED RED CELLS Erythrocyte concentrates, or packed red cells, are prepared y centrifuging whol e lood and removing 250 mL of the plasma supernatant (20). P.698 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 24)03-May-05 08:05:23

Ovid: ICU Book Each unit of packed cells contains approximately 200 mL of cells (mostly erythro cytes) and 50 to 100 mL of plasma and CPD solution. The hematocrit is usually e tween 60 and 80%, and the hemoglo in concentration is etween 23 and 27 g/dL (20 ). LEUKOCYTE-POOR RED CELLS Removal of the leukocytes in packed red cells is recommended when transfusing pa tients with a history of fe rile, nonhemolytic transfusion reactions (which are caused y anti odies to leukocytes in donor lood) (21). The leukocytes can e s eparated y centrifugation or filters, ut separation is never complete and up t o 30% of the leukocytes remain in the sample. WASHED RED CELLS Packed cells can e washed with isotonic saline to remove leukocytes and residua l plasma. The removal of plasma helps prevent allergic reactions caused y prior sensitization to plasma proteins in donor lood. Washed red cells are therefore used for transfusing patients with a history of hypersensitivity transfusion re actions. INFUSING ERYTHROCYTE PRODUCTS A standard infusion system for transfusing erythrocyte products is shown in Figu re 44.4. Each of the num ered components in the system is descri ed riefly in t he following paragraphs. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 24)03-May-05 08:05:23

Ovid: ICU Book Figure 44.4. An infusion system for transfusing erythrocyte products. Each of th e num ered components is descri ed in the text. PRESSURIZED INFUSIONS The gravity-driven flow of whole lood and packed cells is shown in Ta le 44.2 ( 22). Because these flow rates do not approach the 250 mL/min or greater flow rat es needed for the resuscitation of trauma victims, pressure-generating devices a re used to speed infusion rates. The most common device used for this purpose is a standard lood pressure cuff that is wrapped around the collapsi le plastic lood containers. When the cuff is inflated to a pressure of 200 mm Hg, the infus ion rate of whole lood and packed cells increases approximately threefold, as s hown in Ta le 44.2. Manual hand pumps are also availa le for increasing infusion rates. These hand pumps are not as effective as the lood pressure cuff (at 200 mm http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 24)03-May-05 08:05:23

Ovid: ICU Book Hg) for infusing whole lood ut are equivalent to the pressure cuffs for infusi ng packed cells (Ta le 44.2). TABLE 44.2. PRESSURIZED INFUSION OF ERYTHROCYTE PRODUCTS SALINE DILUTION As descri ed in Chapter 14 (see Figure 14.4), the infusion rate of lood product s is inversely related to the density of erythrocytes in P.699 P.700 the fluid ( a viscosity effect). As a result, packed cells infuse at approximately one-third the rate of whole lood. Special Y-configured tu ing in lood infusion sets all ows packed cells to e diluted with an equal volume of isotonic saline. When thi s is done, the infusion rate of packed cells is compara le to that of whole loo d. Only isotonic saline should e used as a diluent for packed red cells. Ringer 's solutions are not advised for diluting erythrocyte products ecause the calci um in Ringer's solutions can promote clotting in the lood sample (23). BLOOD FILTERS Erythrocyte products are infused through filters that trap small clots and other cellular de ris (e. g., decomposed platelets and fi rin-coated leukocytes). The se filters can ecome an impediment to flow as they collect trapped de ris, and thus they should e replaced periodically (e.g., after every 4 units of lood). The standard filters have a pore size of 170 to http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 24)03-May-05 08:05:23

Ovid: ICU Book 260 microns (23), which allows small fi rin microaggregates to pass freely. Thes e microaggregates can ecome lodged in the pulmonary capillaries and create a no rmalities in gas exchange. Smaller microaggregate filters are availa le, ut the ir value in preventing pulmonary complications has not een proven (24). BLOOD WARMERS Warming reduces the viscosity of refrigerated lood and can increase infusion ra tes y 30 to 50% (see footnote in Ta le 44.2). However, the major value of warmi ng lood is considered to e the prevention of hypothermia from rapid transfusio ns (when 1 unit of lood is transfused every 5 to 10 minutes) (25). The recommen ded temperature for infused lood is 33 to 35 C (25). Temperatures of 37 C or highe r can promote hemolysis. A simple method for warming lood is to immerse the lo od storage ags in hot water efore transfusion. However, this rewarming method can take up to 30 minutes, and can produce hemolysis from overheating. Controlle d warming devices that can warm lood to the desired infusion temperatures at fl ow rates slightly in excess of 100 mL/minute are availa le (26). However, at the infusion rates often used to resuscitate trauma victims (i.e., greater than 250 mL/minute), lood warming devices are often una le to warm lood to the desired temperature (26). ADVERSE REACTIONS The complications of homologous lood transfusion (i.e., lood donor from the sa me species) that are most noted are listed in Ta le 44.3 (27,28,29,30,31,32 and 33). P.701 The acute transfusion reactions are descri ed riefly in the followin g paragraphs. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 24)03-May-05 08:05:23

Ovid: ICU Book TABLE 44.3. RISKS OF HOMOLOGOUS BLOOD TRANSFUSION ACUTE HEMOLYTIC REACTIONS Acute hemolytic transfusion reactions are uncommon and are rarely severe enough to e lifethreatening (27,28 and 29). These reactions are produced y anti odies in the recipient that ind to ABO surface antigens on donor erythrocytes. These anti odies fix complement and can produce rapid lysis. One unit of packed red c ells can e completely lysed in less than 1 hour (27). Lysis of red cells someho w provokes a severe systemic inflammatory response that can lead to hypotension and progressive multiorgan dysfunction. This type of transfusion reaction is usu ally the result of identification errors leading to transfusion of ABO mismatche d lood (28). Clinical Manifestations Severe reactions require as little as 10 mL of donor lood (27) and are usually evident within a few minutes after the transfusion is started. Fever, dyspnea, c hest pain, and low ack pain are common (27,28 and 29). Hypotension can develop suddenly and may e the only sign in comatose patients. Severe reactions are acc ompanied y a consumptive coagulopathy and progressive multiorgan dysfunction. A cute renal failure is prominent in 5 to 10% of cases (29). Bedside Strategy The following approach is recommended for any patient who develops a fever soon after the onset of a homologous lood transfusion. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 24)03-May-05 08:05:23

Ovid: ICU Book q Stop the transfusion immediately. This is imperative ecause the P.702 mor idity and mortality in hemolytic reactions is a function of the volume of incompati l e lood transfused (27). q Check the lood pressure immediately. If the pressure is dropping, do the follow ing: r Infuse volume (colloids may e preferred ecause of their a ility to rapidly exp and the vascular volume). r Start dopamine (at 5 g/kg/minute). This agent has een preferred for its renal va sodilating effects ecause renal failure is a poor prognostic sign in hemolytic transfusion reactions (29). However, the relative enefits of dopamine over othe r pressor agents is unproven. q Once the patient is sta ilized, do the following: q

O tain a freshly voided urine specimen and perform a dipstick test for lood. q Send a lood sample for a direct Coom 's test. A positive test confirms a hemoly tic transfusion reaction. However, a negative test is possi le if most of the do nor erythrocytes have already lysed. FEBRILE NONHEMOLYTIC REACTIONS Fever that is not related to hemolysis is the most common acute transfusion reac tion, appearing in approximately 1% of transfusions. This reaction is the result of anti odies in the recipient to leukocytes in donor lood. The antileukocyte anti odies are produced in response to prior transfusions or prior pregnancies, so this type of response is usually seen in multiparous women or in patients wit h a history of prior transfusions. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 24)03-May-05 08:05:23

O tain a lood sample and inspect the plasma for the pink-to-red hue of hemoglo in. q

Ovid: ICU Book Clinical Manifestations The fever usually appears 1 to 6 hours after starting the transfusion (later tha n the onset of fever associated with hemolytic transfusion reactions), and it is not usually accompanied y signs of systemic illness. However, severe reactions can occur, and patients can have a toxic appearance. Bedside Strategy The edside approach is the same as outlined for hemolytic transfusion reactions . The diagnosis of a nonhemolytic fe rile reaction is confirmed y excluding the presence of hemolysis with the tests descri ed previously. Bacterial contaminat ion of lood products is a potential cause of fever, ut it is uncommon (32). Ho wever, some recommend routine culture of donor and recipient lood for any fe ri le transfusion P.703 reaction accompanied y other signs of systemic illness (e. g., rigors, dyspnea). Future Transfusions More than 50% of patients who develop a fe rile nonhemolytic transfusion reactio n will not experience a similar reaction during su sequent lood transfusions. T herefore, no special precautions are necessary for future transfusions. If a sec ond fe rile reaction develops, leukocyte-poor red cell preparations are advised for all further erythrocyte transfusions. ALLERGIC REACTIONS Hypersensitivity reactions (rash, anaphylaxis) are the result of sensitization t o plasma proteins in prior transfusions. Patients with IgA deficiency are partic ularly prone to hypersensitivity reactions to lood transfusions, and can do so without prior exposure to plasma products (33). Clinical Manifestations The usual manifestation is mild urticaria that appears during the transfusion an d may e accompanied y fever. Anaphylactic reactions are rare (Ta le 44.3). Bedside Strategy Mild urticaria without fever does not require interruption of the transfusion. H owever, the common practice is to stop the transfusion temporarily and administe r antihistamines http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 24)03-May-05 08:05:23

Ovid: ICU Book (diphenhydramine, 25 to 50 mg orally or intramuscularly every 6 hours). Antihist amines may help alleviate the pruritus associated with urticaria ut otherwise a re without enefit. The rare instance of transfusion-associated anaphylaxis shou ld e managed as descri ed in Chapter 31 (see Ta le 31.4). Patients who develop anaphylaxis should e tested for an underlying IgA deficiency. Future Transfusions Future transfusions should e avoided if possi le for all cases of transfusion-a ssociated anaphylaxis. For less severe allergic reactions, washed red cell prepa rations (plasma removed) are advised for su sequent transfusions. Antihistamine premedication is a popular ut unproven practice. P.704 ACUTE LUNG INJURY Acute respiratory failure is a much talked a out ut uncommon transfusion reacti on, with an estimated incidence of 1 per 5000 transfusions (34). It has een o s erved after a single transfusion of either whole lood or packed red cells (15). The prevailing theory is that antileukocyte anti odies in donor lood ind to c irculating granulocytes in the recipient and promote leukocyte sequestration in the pulmonary microcirculation. This then leads to granulocyte-mediated lung inj ury, which presents as acute respiratory distress syndrome (ARDS). However, unli ke most cases of ARDS, this lung injury is rarely fatal. Clinical Manifestations Signs of respiratory compromise (dyspnea, hypoxemia) usually develop within a fe w hours after the transfusion egins. Fever is common, and hypotension has een reported (34). The chest x-ray film eventually shows diffuse pulmonary infiltrat es. Although the acute syndrome can e severe, the process usually resolves with in a week. As mentioned, fatalities are rare. Bedside Strategy The transfusion should e stopped (if still running) at the first signs of respi ratory compromise. The remainder of the management is similar to that descri ed for ARDS in Chapter 23. Future Transfusions http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 24)03-May-05 08:05:23

Ovid: ICU Book There are no firm recommendations regarding future transfusions in patients who develop transfusion-associated ARDS. Some advocate the use of washed red cells ( to remove any anti ody-containing plasma in the donor sample), while others caut ion against any future transfusions unless a solutely necessary (34). AUTOLOGOUS TRANSFUSION Reinfusion of the patient's own lood (autologous transfusion) not only eliminat es many of the risks associated with homologous transfusions, ut also reduces t he strain on the nation's lood supply. The following is a rief description of the different approaches to autologous lood transfusion (35). PREOPERATIVE DEPOSIT Patients undergoing elective surgical procedures with minimal lood losses (e.g. , orthopedic procedures) can donate their own lood P.705 efore the procedure. The patients must have a aseline hemoglo in of 11 g/dL or higher and e free of any conditions that pose a risk for phle otomy (e.g., frequent angina, aortic s tenosis). One unit of lood can e removed every 4 days (usually weekly) for 2 o r 3 weeks efore the procedure. Any lood transfusions required during or after surgery are then performed with the patient's own lood. INTRAOPERATIVE SALVAGE Reinfusion of lost lood is a popular practice for loody surgical procedures such as openheart surgery and liver transplantation. It is also allowed y some Jeho vah's Witnesses (35). Blood is aspirated with a modified suction wand and fed to a specialized cell-saver instrument that centrifuges the lood to concentrate the erythrocytes and washes the specimen with saline. High-speed instruments that c an process a unit of packed cells in less than 5 minutes are now availa le (35).

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 24)03-May-05 08:05:23

Complications Surprisingly few risks are associated with intraoperative nal coagulopathy is pro a ly the most common complication, rger volume reinfusions. Consumptive coagulopathy and ARDS complications (salvaged lood syndrome) and may e caused ated leukocytes (36).

lood salvage. Dilutio and it occurs with la are the most serious y reinfusion of activ

Ovid: ICU Book POSTOPERATIVE SALVAGE The most common method of postoperative lood salvage is the reinfusion of shed mediastinal lood after median sternotomy procedures (37). Chest tu e drainage f rom the mediastinum is directed to a rigid cannister that contains two collectio n ags. The first ag contains a 170micron filter to trap large clots and cellul ar de ris. Because the lood has undergone endogenous defi rination in the chest , no anticoagulants need to e added. The lood passes through the filter and dr ains into the second collection ag, which has an 800 mL capacity. When this ag is filled, it is detached and hung on an IV pole for reinfusion. This method di ffers from intraoperative salvage in that there is no centrifugation or saline w ashing of the sample. Traumatic disruption of cells is common, and the hematocri t of the reinfused lood is only 15 to 25%. This low hematocrit may explain why reinfusion of shed mediastinal lood does not always reduce homologous transfusi on needs (38). Complications are considered uncommon. However, the supernatant i n shed mediastinal lood contains high concentrations of leukocyte elastase (39) , and this could produce a clinical syndrome P.706 similar to the salvaged lood syndrome seen with intraoperative lood salvage. REFERENCES CONSENSUS STATEMENTS 1. Consensus Conference on Perioperative Red Blood Cell Transfusion. JAMA 1988;2 60:27002702. 2. Practice strategies for elective red lood cell transfusion. A Clinical Pract ice Guideline from the American College of Physicians. Ann Intern Med 1992;116:4 03406. REVIEWS 3. Ro ertie PG, Gravlee GP. Safe limits of isovolemic hemodilution and recommend ations for erythrocyte transfusions. Int Anesthesiol Clin 1990;28:197204 (54 refe rences). 4. Welch HG, Meehan KR, Goognough LT. Prudent strategies for elective red lood cell http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 24)03-May-05 08:05:23

Ovid: ICU Book transfusion. Ann Intern Med 1992;116:393402 (120 references). 5. Spence RK, Cernaianu AC, Carson J, DelRossi AJ. Transfusion and surgery. Curr Pro l Surg 1993;30:11011192 (212 references). MONITORING ANEMIA 6. Jones JG, Holland BM, Wardrop CAJ. Total circulating red cells versus hematoc rit as a primary descriptor of oxygen transport y the lood. Br J Hematol 1990; 76:228232. 7. Cordts PR, LaMorte WW, Fisher JB, et al. Poor predictive value of hematocrit and hemodynamic parameters for erythrocyte deficits after extensive elective vas cular operations. Surg Gynecol O stet 1992;175:243248. CONSEQUENCES OF ANEMIA 8. Brannon ES, Merrill AJ, Warren JV, Stead EA Jr. Cardiac output in patients wi th chronic anemia as measured y the technique of right heart catheterization. J Clin Invest 1945;24:332337. 9. Duke M, A elmann WH. The hemodynamic response to chronic anemia. Circulation 1969;39:503515. 10. Rosen AL, Gould S, Sehgal LR, et al. Cardiac output response to extreme hemo dilution with hemoglo in solutions of various P50 values. Crit Care Med 1979;7:3 80 382. 11. Fontana JL, Wel orn L, Mongan PD, et al. Oxygen consumption and cardiovascul ar function in children during profound intraoperative normovolemic hemodilution . Anesth Analg 1995;80:219225. 12. Messmer K. Therapeutic threshold values for acute alterations in hemoglo in concentration. In: Zander R, Mertzlufft F, eds. The oxygen status of arterial l ood. Basel, http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 24)03-May-05 08:05:23

Ovid: ICU Book Switzerland: Karger Pu lishers, 1991;167173. 13. Levine E, Rosen A, Sehgal L, et al. Physiologic effects of acute anemia: imp lications for a reduced transfusion trigger. Transfusion 1990;30:1114. P.707 TRANSFUSION TRIGGER 14. Wilkerson DK, Rosen AL, Gould SA, et al. Oxygen extraction ratio: a valid in dicator of myocardial meta olism in anemia. J Surg Res 1987;42:629634. 15. Levy PS, Chavez RP, Crystal GJ, et al. Oxygen extraction ratio: a valid indi cator of transfusion need in limited coronary vascular reserve? J Trauma 1992;32 :769774. EFFICACY OF RED CELL TRANSFUSIONS 16. Shah DM, Gottlie M, Rahm R, et al. Failure of red cell transfusions to incr ease oxygen transport or mixed venous PO2 in injured patients. J Trauma 1982;22: 741746. 17. Silverman H, Tuma P. Gastric tonometry in patients with sepsis: effects of d o utamine and packed red cell transfusions. Chest 1992;102:184188. 18. Marik PE, Si ald W. Effect of stored- lood transfusion on oxygen delivery i n patients with sepsis. JAMA 1993;269:30243029. 19. Ro ins JM, Keating K, Orlando R, Yeston N. Effects of lood transfusion on oxygen delivery and consumption in critically ill surgical patients. Contemp Sur g 1993;43:281 285. ERYTHROCYTE PRODUCTS 20. American Association of Blood Banks Technical Manual. 10th ed. Arlington, VA : http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 24)03-May-05 08:05:23

Ovid: ICU Book American Association of Blood Banks, 1990;3758, 635637. 21. Davies SC. Transfusion of red cells. In: Contreras M, ed. ABC of transfusion . London: British Medical Journal, 1990;913. INFUSING ERYTHROCYTE PRODUCTS 22. Dula DJ, Muller A, Donovan SW. Flow rate variance of commonly used IV infusi on techniques. J Trauma 1981;21:480482. 23. Blood Transfusion Practice. In: American Association of Blood Banks Technica l Manual. 10th ed. Arlington, VA: American Association of Blood Banks, 1990;34137 5. 24. Kruskall MS, Bergen JJ, Klein HG, et al. Transfusion therapy in emergency me dicine. Ann Emerg Med 1988;17:327335. 25. Iserson KV, Huestis DW. Blood warming: current applications and techniques. Transfusion 1991;31:558571.

RISKS OF HOMOLOGOUS TRANSFUSION 27. Seyfried H, Walewska I. Immune hemolytic transfusion reactions. World J Surg 1987;11:2529. 28. Gloe D. Common reactions to transfusions. Heart Lung 1991;20:506512. 29. Nicholls MD. Transfusion: mor idity and mortality. Anesth Intensive Care 199 3;21:15 19. P.708 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 24)03-May-05 08:05:23

26. Uhl L, Pacini D, Kruskall MS. A comparative study of e. Anesthesiol 1992;77:10221028.

lood warmer performanc

Ovid: ICU Book 30. Schrei er GB, Busch MP, Akinman S, et al. The risk of transfusion-transmitte d viral infections. N Engl J Med 1996;334:16851690. 31. Klein HG. New insights into the management of anemia in the surgical patient . Am J Med 1996;101(Suppl 2A):12S15S. 32. Gottlie T. Hazards of acterial contamination of lood products. Anesth Int ensive Care 1993;21:2023. 33. Is ister JP. Adverse reactions to plasma and plasma components. Anesth Inten sive Care 1993;21:3138. 34. Gans ROB, Duurkens VAM, van Zundert AA, et al. Transfusion-related acute lun g injury. Intensive Care Med 1988;14:654657. AUTOLOGOUS TRANSFUSIONS 35. Ereth MH, Oliver WC Jr, Santrach PJ. Perioperative interventions to decrease transfusion of allogenic lood products. Mayo Clin Proc 1994;69:575586. 36. Bull BS, Bull MH. The salvaged lood syndrome: a sequel to mechanochemical a ctivation of platelets and leukocytes? Blood Cells 1990;16:520. 37. Ward HB, Smith RR, Landis KP, et al. Prospective, randomized trial of autotr ansfusion after routine cardiac operations. Ann Thorac Surg 1993;56:137141. 38. Ro erts SR, Early G, Brown B, et al. Autotransfusion of unwashed mediastinal shed lood fails to decrease anked lood requirements in patients undergoing a ortocoronary ypass surgery. Am J Surg 1991;162:477480. 39. Har ison S, Chung S, Kucick B, Marino PL. Leukocyte disruption and elastase release in autotransfused lood. Crit Care Med 1989;17:S42. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (23 of 24)03-May-05 08:05:23

Ovid: ICU Book SUGGESTED READINGS Anderson KC, Ness P. Scientific asis of transfusion medicine: implications for clinical practice. Philadelphia: WB Saunders, 1994. Blood: Bearer of life and death: New ways to fight diseases caused y faults in the loodstream. A report from the Howard Hughes Medical Institute. Chevy Chase, MD: Howard Hughes Medical Institute, 1993. Contreras M, ed. ABC of transfusion. London: British Medical Journal, 1990. Gravlee GP, ed. Blood conservation. Int Anesthesiol Clin 1990;28:183243. Hillman RS, Finch CA. Red cell manual. 6th ed. Philadelphia: FA Davis, 1992. Spe nce RK, ed. New insights into the management of anemia in the surgical patient. Am J Med 1996;101(Suppl 2A):1S44S. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (24 of 24)03-May-05 08:05:23

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 45 PLATELET DISORDERS AND REPLACEMENT Platelet disorders fall into two categories: those characterized y a normal num ers of circulating platelets and those characterized y a normal platelet funct ion. This chapter introduces the common causes of oth types of platelet disorde rs in the ICU population. This is followed y a rief description of the indicat ions, methods, and complications of platelet transfusion therapy (1,2,3 and 4). PLATELETS AND HEMOSTASIS The normal adult has an average of 250 illion platelets per liter of lood. Ass uming a normal lood volume of 5.5 L, the total num er of platelets in the lood stream will e slightly in excess of 1 trillion. To keep this platelet count con stant, as many as 45 illion new platelets must e added to each liter of lood daily (5). For a lood volume of 5.5 L, this corresponds to a total production o f 248 illion platelets each day. These num ers are astounding, and indicate the level of marrow activity needed to maintain normal platelet homeostasis. THROMBOSIS Platelets are not really cells ecause they contain no nucleus and cannot synthe size proteins. They are a portion of megakaryocyte cytoplasm that has een pinch ed off and completely surrounded y a cell mem rane. This cytoplasm is composed of dense granules that are rich in calcium. When the vascular endothelium is den uded, platelets adhere to the exposed su endothelium ( y virtue of mem rane glyc oproteins called integrins that act as adhesion receptors) and release the conte nts of their granules. The calcium that is released helps initiate the coagulati on cascade that ends in the production of fi rin strands. P.710 The strands of f i rin then form an interlacing meshwork with the platelets to produce a throm us . THROMBOCYTOPENIA http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 14)03-May-05 08:06:14

Ovid: ICU Book Throm ocytopenia is defined as a platelet count elow 150,000/mm3 (or 150 109/L in SI Units) (6). However, the a ility to form a hemostatic plug is retained unt il the platelet count falls elow 100,000/mm3. Therefore, clinically significant throm ocytopenia can e defined as a platelet count elow 100,000/mm3. The lee ding tendency in throm ocytopenia is determined primarily y the presence or a s ence of a structural lesion that is prone to leeding and is not a function of t he platelet count. Although the traditional teaching has een that spontaneous h emorrhage in the a sence of a structural lesion can occur when the platelet coun t falls elow 20,000/mm3 (1), this is no longer considered valid (7). In the a s ence of risk factors for leeding, platelet counts elow 5000/mm3 can e tolerat ed without trou lesome leeding (8). PLATELET ADHESION When the a ility of platelets to adhere to the su endothelium is diminished (e.g ., as in uremia), the risk of hemorrhage can e increased despite platelet count s exceeding 100,000/mm3. Defects in platelet adhesion can e detected y a prolo nged leeding time. However, no correlation exists etween the leeding time and the tendency for leeding (9), and thus the leeding time is not useful for det ecting clinically significant a normalities in platelet adhesion. The recognitio n of platelet adhesion a normalities in individual patients is accomplished y i dentifying the conditions that alter platelet adhesiveness. THROMBOCYTOPENIA The only survey of throm ocytopenia in patients in the ICU is from a medical ICU , where 23% of the patients had platelet counts elow 100,000/mm3 at some time d uring their ICU stay (10). The causes of throm ocytopenia that are most likely t o e encountered in the ICU are listed in Ta le 45.1. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 14)03-May-05 08:06:14

Ovid: ICU Book TABLE 45.1. CAUSES OF PLATELET DISORDERS IN THE ICU P.711 HEPARIN Heparin can com ine with a heparin- inding protein (platelet factor 4) in platel ets to form an antigenic complex that induces the formation of IgG. This immunog lo ulin can then ind to platelets and promote platelet clumping. If severe enou gh, this process can result in a consumptive throm ocytopenia and clinically app arent throm osis (11). Clinical Features Throm ocytopenia is reported in 1 to 3% of patients receiving heparin (12). It u sually appears within 14 days after heparin is started and is independent of the heparin dose. Even the small doses of heparin used for heparin flushes and the small amounts on heparin-coated pulmonary artery catheters can induce throm ocyt openia (13). Throm ocytopenia is less common with lowmolecular-weight heparin tha n with unfractionated heparin (14). The major complication of heparin-induced th rom ocytopenia is throm osis, not leeding. In a 14-year survey of patients with heparin-induced throm ocytopenia, venous throm oem olism was identified in 70% of the patients and arterial throm osis was identified in 15% of the patients (1 2). In approximately half the patients surveyed, the disorder was first recogniz ed after a throm otic episode had occurred. In the other patients who had isolat ed throm ocytopenia, 50% developed a throm osis within the next 30 days despite the discontinuation of heparin therapy (12). Diagnosis There are two criteria for the diagnosis of heparin-induced throm ocytopenia (14 ). The first is throm ocytopenia that develops more than 5 days after the initia l exposure to heparin. The second is a positive assay for heparin-induced IgG an ti odies. The anti ody is detected y the release of 14C-la eled serotonin from platelets that are added to a sample of the patient's serum. This assay is avail a le in most clinical la oratories and is costly [$222 for the assay performed y the Smith Kline Clinical La oratories (Philadelphia, PA) in 1996]. Management http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 14)03-May-05 08:06:14

Ovid: ICU Book In patients with isolated throm ocytopenia (without throm osis), heparin adminis tration should e discontinued. Remem er to discontinue heparin flushes and remo ve any heparincoated intravascular catheters. If anticoagulation is necessary, C oumadin can e started. Dextran (500 mL of Dextran-40 daily) can e used as a te mporary antithrom otic measure until the Coumadin achieves therapeutic levels of anticoagulation. The management of heparin-induced throm osis is difficult. Cou madin is usually ineffective (15), and alternative anticoagulants (e.g., P.712 d anaproid and hirudin) are currently not availa le in the United States. Throm ol ytic therapy may prove useful here (16). INFECTIONS Throm ocytopenia is reported in more than 50% of patients with systemic sepsis ( 4), and in 50% of patients with acquired immunodeficiency syndrome (AIDS) (17). In oth conditions platelet destruction is enhanced; ut in AIDS it is immune-me diated, and in sepsis it is not. Bleeding is not common in either condition unle ss the throm ocytopenia is accompanied y other coagulation a normalities. DISSEMINATED INTRAVASCULAR COAGULATION Widespread endothelial damage, as can occur with septicemia or multiple trauma, releases a protein known as tissue factor that activates the endogenous coagulat ion cascade and the fi rinolytic system. This can result in a severe coagulopath y characterized y widespread microvascular throm osis accompanied y depletion of circulating platelets and procoagulant proteins (18). This condition is calle d disseminated intravascular coagulation (DIC). Clinical Features The microvascular throm osis in DIC produces multiorgan dysfunction. The lungs a re commonly involved, and the clinical picture is similar to the acute respirato ry distress syndrome, which is descri ed in Chapter 23. Advanced cases are accom panied y acute oliguric renal failure and progressive hepatocellular injury. De pletion of platelets and coagulation factors can e accompanied y leeding from multiple sites, particularly the gastrointestinal tract. Diagnosis http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 14)03-May-05 08:06:14

Ovid: ICU Book The diagnosis of DIC is ased on the clinical setting (i.e., severe sepsis or mu ltiple trauma) com ined with la oratory evidence of widespread coagulation defec ts. Throm ocytopenia can e severe and is accompanied y prolongation of the pro throm in time and activated partial throm oplastin time (from depleted coagulant proteins) and elevated fi rin degradation products (from fi rinolysis). Fi rino gen levels can e misleading ecause fi rinogen is an acute phase reactant, and this can result in normal fi rinogen levels in the face of increased fi rinogen use (19). Management Acute, fulminant DIC often has a fatal outcome. Heparin usually is ineffective i n retarding the microvascular throm osis, pro a ly ecause of depletion of antit hrom in-III (20). Antithrom inII concentrates can e given with heparin (the ATIII dosage is 90 to 120 Units P.713 as a load, then 90 to 120 Units daily for 4 days), ut the value of this practice remains unproven (21). Bleeding is particu larly trou lesome to manage ecause the administration of coagulation factors an d platelets can aggravate the microvascular throm osis. THROMBOTIC THROMBOCYTOPENIA PURPURA Throm otic throm ocytopenia purpura (TTP) is a rare ut life-threatening conditi on caused y immune-mediated platelet aggregation with widespread microvascular throm osis (similar to DIC). It usually is seen in young adults, particularly wo men, and often follows a nonspecific illness. Clinical Features A com ination of five clinical features (pentad) is characteristic of TTP: fever , neurologic changes, acute renal failure, throm ocytopenia, and microangiopathi c hemolytic anemia. Patients usually experience fever and depressed consciousnes s that can progress rapidly to coma and generalized seizures. Throm ocytopenia i s not associated with other coagulation a normalities, which distinguishes TTP f rom DIC. The diagnosis is confirmed y the presence of schistocytes in the lood smear (indicating a microangiopathic hemolytic anemia). Management The treatment of choice for TTP is plasma exchange (22,23). This can e performe d with plasmapheresis equipment that removes lood and separates the plasma from the red cells. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 14)03-May-05 08:06:14

Ovid: ICU Book The plasma is discarded and the red cells are reinfused with fresh frozen plasma . This is continued until 1.5 times the plasma volume has een exchanged (the no rmal plasma volume is 40 mL/kg in adult men and 36 mL/kg in adult women). This i s repeated daily for approximately 1 week. If plasmapheresis equipment is unavai la le, a poor man's plasma exchange can e performed y inserting an arterial cath eter into the femoral artery and withdrawing 500 mL aliquots of lood (equivalen t to 1 Unit of whole lood) into a lood collection ag. This is then sent to th e lood ank for centrifugation to separate the plasma from the lood cells. The cells are then returned and reinfused with a unit of fresh frozen plasma. This is continued until at least one plasma volume has een exchanged. Acute, fulmina nt TTP is almost always fatal if untreated. With prompt use of plasma exchange, as many as 90% of patients can survive the acute episode (22,23). Platelet trans fusions can aggravate the underlying throm osis in TTP, and thus are contraindic ated. BLOOD TRANSFUSIONS The via ility of platelets in whole lood and erythrocyte concentrates (packed c ells) is almost completely lost after 24 hours of storage. P.714 Therefore, larg e-volume transfusions can produce dilutional throm ocytopenia. This effect ecom es prominent when the transfusion volume exceeds 1.5 times the lood volume (24) . A rare type of posttransfusion throm ocytopenia appears approximately 1 week a fter transfusion, usually in multiparous women, and is caused y antiplatelet an ti odies. This condition is called posttransfusion purpura, and throm ocytopenia is often severe and prolonged (25). Platelet counts can fall to 10,000/mm3 or l ower for as long as 40 days (18). If hemorrhage ensues, plasma exchange is the t reatment of choice. ABNORMAL PLATELET FUNCTION The common causes of impaired platelet aggregation in patients in the ICU are li sted in Ta le 45.1 (26). RENAL INSUFFICIENCY Impaired platelet adhesion is seen in oth acute and chronic renal insufficiency . No correlation exists etween the severity of the renal dysfunction and the se verity of the platelet function a normality. The pro lem is corrected y hemodia lysis or peritoneal dialysis. It is unclear if this http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 14)03-May-05 08:06:14

Ovid: ICU Book platelet function a normality is clinically significant (26). CARDIOPULMONARY BYPASS Platelet adhesiveness is impaired y unknown mechanisms when lood passes throug h the oxygenator apparatus used during cardiopulmonary ypass. The severity of t he platelet function defect is directly related to the duration of ypass (26). In most cases, the a normality resolves within a few hours after ypass is compl eted. However, defects in platelet adhesion may contri ute to trou lesome medias tinal leeding in the immediate postoperative period. Aprotinin Aprotinin (Trasylol) is a proteinase inhi itor that can prevent the platelet dys function associated with cardiopulmonary ypass (27). When administered intraope ratively at the dosage shown elow, aprotinin can reduced the requirement for pl atelets and erythrocyte products (28). Dose: 280 mg IV as a loading dose, followed y a continuous infusion of 70 mg/hour. Al so add 280 mg to the oxygenator prime solution. The drug is infused for the dura tion of the operative procedure. Although there is a concern that aprotinin migh t promote throm osis in the ypass grafts, no evidence exists to support this co ncern (28). P.715 ASPIRIN Aspirin inhi its platelet aggregation y locking prostaglandin-mediated platele t degranulation. The effect is irreversi le and lasts for the entire life span o f the platelets (10 days). One aspirin ta let (325 mg) is enough to produce the platelet defect. Emergency Surgery The effects of aspirin on platelet adhesion ecomes an important consideration i n patients taking aspirin daily (e.g., as a preventive measure for coronary thro m osis) who require emergency surgery. This is particularly true for cardiopulmo nary ypass surgery ecause aspirin can add to the risk of leeding from ypassinduced platelet dysfunction. Most studies of emergency ypass surgery show incr eased leeding associated with recent aspirin use http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 14)03-May-05 08:06:15

Ovid: ICU Book (26). This aspirin effect is less pronounced when aprotinin is given during the ypass procedure (29).

INDICATIONS FOR PLATELET TRANSFUSIONS The following guidelines for platelet transfusions are summarized from the first four references listed at the end of this chapter. ACTIVE BLEEDING The following statements apply to all instances of active leeding other than ec chymotic and petechial hemorrhage. Platelet transfusions are indicated in the fo llowing situations: q The platelet count is elow 50,000/mm3 and the throm ocytopenia is not the resul t of immune mechanisms. q The platelet count is a ove 50,000/mm3 and a condition that significantly impair s platelet adhesion (e.g., cardiopulmonary ypass) is present. Platelet transfusions are not indicated in the following situations: q The platelet count is a ove 50,0000/mm3 and platelet function is normal. q Throm ocytopenia is caused y antiplatelet anti odies. P.716 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 14)03-May-05 08:06:15

OTHER DRUGS A variety of other drugs can impair platelet adhesion, including -lactam anti iot ics, nonsteroidal anti-inflammatory agents, and dextrans (26). However, the clin ical significance of these drug-induced platelet effects is unclear.

Ovid: ICU Book MASSIVE TRANSFUSION The traditional practice of giving platelet transfusions after transfusion of 8 to 10 Units of whole lood or packed cells is no longer recommended. Platelet co unts should e monitored and platelets transfused if the platelet count falls to elow 50,000/mm3 and evidence of continued leeding is present. As mentioned pr eviously, significant dilutional throm ocytopenia is not expected until the tran sfused lood volume exceeds 1.5 times the lood volume. PROPHYLAXIS The following statements apply to patients with no evidence of active leeding o ther than ecchymotic or petechial hemorrhage. Platelet transfusions are indicate d in the following situations: q The platelet count is elow 5,000/mm3. q The platelet is elow 20,000/mm3 and a high-risk condition for leeding (e.g., p eptic ulcer disease, prior hemorrhage from diverticulosis or arteriovenous malfo rmations) is present. q The platelet count is elow 50,000/mm3, and the following procedures are planned : Endoscopic iopsy um ar puncture Major surgery Platelet transfusions are not indicated in the following situations: q The platelet count is 5000 to 20,000/mm3 ut no risk factors for leeding are pr esent. q Throm ocytopenia is caused y antiplatelet anti odies. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 14)03-May-05 08:06:15

Ovid: ICU Book PLATELET TRANSFUSIONS Platelet concentrates are prepared y centrifuging fresh whole lood and suspend ing the platelet pellet in a small volume of plasma. Each platelet concentrate ( from 1 Unit of whole lood) contains 50 to 100 illion platelets in 50 mL of pla sma (30). Platelets can e stored for up to 7 days, ut via ility egins to decl ine after 3 days. Platelet transfusions are usually given as multiples of 6 to 1 0 individual platelet concentrates that are pooled together. EFFICACY In the average-size adult, each platelet concentrate should raise the circulatin g platelet count y 5000 to 10,000/mm3 (30) and the effect should last approxima tely 8 days. Smaller increments for a shorter duration are expected when the rat e of platelet destruction increases, either in the platelet concentrate or in th e patient. P.717 COMPLICATIONS Because one platelet transfusion involves platelet concentrates from 6 to 8 indi vidual donors, the infectious risks associated with homologous lood transfusion s (see Ta le 44.3) are increased sixfold to eightfold with platelet transfusions . Fe rile nonhemolytic reactions are also more common with platelet transfusions , and fever has een reported in up to 30% of platelet transfusion recipients (3 1). Hypersensitivity reactions triggered y proteins in the plasma fraction of p latelet concentrates can also occur. Although platelet mem ranes have ABO antige ns, transfusion reactions as a result of ABO incompati ility do not occur with p latelet transfusions. However, recipients of multiple platelet transfusions can develop antiplatelet anti odies that limit the effectiveness of platelet transfu sions. If this ecomes a pro lem, platelets can e harvested y apheresis from s ingle, HLA-matched donors. REFERENCES CLINICAL PRACTICE GUIDELINE 1. College of American Pathologists, Administration Practice Guidelines Developm ent Task Force. Fresh frozen plasma, cryoprecipitate, and platelets. JAMA 1994;2 71:777 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 14)03-May-05 08:06:15

Ovid: ICU Book 781. REVIEWS 2. Rintels PB, Kenney RM, Cowley JP. Therapeutic support of the patient with thr om ocytopenia. Hematol Oncol Clin North Am 1994;8:11311157.

4. Chang JC. Postoperative throm ocytopenia with etiologic, diagnostic, and ther apeutic considerations. Am J Med Sci 1996;311:96105. PLATELETS AND HEMOSTASIS 5. Tomer A, Harker LA. Megakaryocytopoiesis and platelet kinetics. In: Rossi EC, Simon TL, Moss GS, eds. Principles of transfusion therapy. Baltimore: Williams & Wilkins, 1991;167179. 6. American Association of Blood Banks Technical Manual. 10th ed. Arlington, VA: American Association of Blood Banks, 1990;649. 7. Beutler E. Platelet transfusions: the 20,000 per microliter trigger. Blood 19 93;81:1411 1413. 8. Beutler E. Commentary. A str Clin Care Guidelines 1994;6:4. 9. Rodgers RP, Levin J. A critical reappraisal of the leeding time. Semin Throm Hemost 1990;16:120.

THROMBOCYTOPENIA http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 14)03-May-05 08:06:15

3. Machin SJ, Kelsey H, Seghatchian MJ. Platelet transfusion. Throm 5;74:246252.

Haemost 199

Ovid: ICU Book 10. Baughman RP, Lower EF, Flessa HC, Tollerud DJ. Throm ocytopenia in the inten sive care unit. Chest 1993;104:12431247. P.718 11. Aster RH. Heparin-induced throm ocytopenia and throm osis. N Engl J Med 1995 ;332:13741376. 12. Warkentin TE, Kelton JG. A 14-year study of heparin-induced throm ocytopenia . Am J Med 1996;101:502507. 13. Laster J, Silver D. Heparin-coated catheters and heparin-induced throm ocyto penia. J Vasc Surg 1988;7:667672. 14. Warkentin TE, Levine MN, Hirsh J, et al. Heparin-induced throm ocytopenia in patients treated with low-molecular-weight heparin or unfractionated heparin. N Engl J Med 1995;332:13301335. 15. Warkentin TE, Hirsh J, Kelton JG. Heparin-induced throm ocytopenia. N Engl J Med 1995;333:1007. 16. Dieck JA, Rizo-Patron C, Unisa A, et al. A new manifestation and treatment a lternative for heparin-induced throm osis. Chest 1990;98:15241526. 17. Doweiko JP, Groopman JE. Hematologic consequences of HIV infection. In: Brod er S, Merigan TC, Bolognesi D, eds. Text ook of AIDS medicine. Baltimore: Willia ms & Wilkins, 1994;617628. 18. Bell WR. The pathophysiology of disseminated intravascular coagulation. Semi n Hematol 1994;31(Suppl 1):1924. 19. Bovill EG. La oratory diagnosis of disseminated intravascular coagulation. S emin Hematol 1994;31(Suppl 1):3539. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 14)03-May-05 08:06:15

Ovid: ICU Book 20. Clark J, Ru in RN. A practical approach to managing disseminated intravascul ar coagulation. J Crit Illness 1994;9:265280. 21. Fourrier F, Chopin C, Huart J-J, et al. Dou le- lind, place o-controlled tri al of antithrom in II concentrates in septic shock with disseminated intravascul ar coagulation. Chest 1993;104:882888. 22. Rock GA, Shumack KH, Buskard NA, et al. Comparison of plasma exchange with p lasma infusion in the treatment of throm otic throm ocytopenia purpura. N Engl J Med 1991;325:393397. 23. Hayward CP, Sutton DMC, Carter WH Jr, et al. Treatment outcomes in patients with adult throm otic throm ocytopenic purpura-hemolytic uremic syndrome. Arch I ntern Med 1994;154:982987. 24. Reiner A, Kickler TS, Bell W. How to administer massive transfusions effecti vely. J Crit Illness 1987;2:1524. 25. Heffner JE. What caused post-op throm ocytopenia in this 82 year-old man? J Crit Illness 1996;11:666671. ABNORMAL PLATELET FUNCTION 26. George JN, Shattil SJ. The clinical importance of acquired a normalities of platelet function. N Engl J Med 1991;324:2739.

28. Lemmer JH Jr, Stanford W, Bonney SL, et al. Aprotinin for coronary ypass op erations: efficacy, safety, and influence on early saphenous vein graft patency. J Thorac Cardiovasc Surg 1994;107:543553. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 14)03-May-05 08:06:15

27. Mohr R, Goor DA, Lusky A, Lavee J. Aprotinin prevents cardiopulmonary induced platelet dysfunction. Circulation 1992;86(Suppl II):405409.

ypass

Ovid: ICU Book 29. Murkin JM, Lux J, Shannon NA, et al. Aprotinin significantly decreases leed ing and transfusion requirements in patients receiving aspirin and undergoing ca rdiac operations. J Thorac Cardiovasc Surg 1994;107:554561. PLATELET TRANSFUSIONS 30. Simon TL. Platelet transfusion therapy. In: Rossi EC, Simon TL, Moss GS, eds . Principles of transfusion medicine. Baltimore: Williams & Wilkins, 1991;219222. 31. Heddle NM, Klama L, Singer J, et al. The role of the plasma from platelet co ncentrates in transfusion reactions. N Engl J Med 1994;331:625628. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 14)03-May-05 08:06:15

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 46 NUTRIENT AND ENERGY REQUIREMENTS The fundamental goal of nutritional support is to provide individual patients wi th their daily nutritional requirements. This chapter explains how to determine the nutrient and energy needs of each patient in the ICU (1,2 and 3). OXIDATIVE ENERGY CONVERSION OXIDATIVE COMBUSTION According to the Laws of Thermodynamics, energy can neither e produced nor dest royed. Therefore, the only way to o tain energy is to transfer it from an energy source in nature. Natural su stances that are rich in stored energy are called fuels, and the device that performs the energy transfer is called an engine. The process of energy transfer y two types of engines is illustrated in Figure 46. 1. The automo ile has a mechanical engine that mixes oxygen with a fossil fuel ( e.g., gasoline) at high temperatures, and this releases the energy from the fuel that is then used to power the automo ile. Likewise, the human ody has a ioch emical engine (meta olism) that mixes oxygen with an organic fuel (e.g., car ohy drates) at high temperatures, and this releases energy from the fuel that is the n used to power the human ody. The process that allows the energy to e release d from a fuel is called oxidation, or the chemical reaction etween oxygen and a fuel (see Chapter 3). If the oxidation reaction is conducted at high temperatur es, the energy release from the fuel is more rapid. Such hightemperature oxidati on reactions are called com ustion reactions. Thus, oth the automo ile engine a nd oxidative meta olism are internal com ustion engines that capture the energy stored in natural fuels. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 21)03-May-05 08:01:43

Ovid: ICU Book Figure 46.1. Energy conversion y two internal com ustion engines. One engine is mechanical, and the other is iochemical. P.722 ORGANIC FUELS The three organic (car on- ased) fuels used y the human ody are car ohydrates, proteins, and lipids. The energy yield from the com ustion of these fuels is me asured as heat production in kilocalories (kcal) per gram of su strate. The ener gy yield from the com ustion of each of the organic fuels is shown in Ta le 46.1 . The information in this ta le can e stated as follows: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 21)03-May-05 08:01:43

Ovid: ICU Book TABLE 46.1. THE OXIDATIVE METABOLISM OF ORGANIC FUELS P.723 The summed meta olism of all three organic su strates determines the total - ody O2 consumption (VO2), CO2 production (VCO2), and energy expenditure (EE) f or any given period. The 24-hour EE then determines the daily calorie requiremen ts that must e provided y nutrition support. DAILY ENERGY EXPENDITURE The daily energy expenditure of each individual patient can e estimated or meas ured. PREDICTIVE EQUATIONS In the early part of the twentieth century, the daily energy expenditure of a gr oup of healthy adults (136 men and 103 women) was measured (4). The results of t his study were expressed as regression equations for daily energy expenditure a sed on sex, ody weight (in kilograms), and height (in inches). These equations are known as the HarrisBenedict equations (named after the principal investigator s in the study), and they are shown in Ta le 46.2. The daily energy expenditure is expressed as the asal energy expenditure (BEE), which is the heat production of asal meta olism in the resting and fasted state. Because the ody weight in the HarrisBenedict equations does not allow for changes in ody weight caused y o esity or edema fluid, the ideal ody weight should e used in these predictiv e equations. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 21)03-May-05 08:01:43

Ovid: ICU Book TABLE 46.2. METHODS FOR DETERMINING DAILY ENERGY REQUIREMENTS Another more simplified predictive equation for the BEE is as follows: P.724 This relationship has proven to e equivalent to the more complicated Harr isBenedict equations (5). Although it has not een tested rigorously, this simple relationship provides a allpark estimate of BEE for determining nutritional need s. Adjustments in BEE To allow for the thermal effect of food intake, the BEE is multiplied y 1.2 to derive the resting energy expenditure (REE), which is the energy expenditure of asal meta olism in the resting ut not fasted state. Other adjustments in the B EE that allow for enhanced energy expenditure in hypermeta olic conditions are s hown elow: Fever: BEE 1.1 (for each C a ove the normal ody temperature) Mild st ress: BEE 1.2 Moderate stress: BEE 1.4 Severe stress: BEE 1.6 The actual adjustm ents for severe illness can vary widely in individual patients (6). Studies comp aring predicted and actual energy expenditure in critically ill patients have sh own that the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 21)03-May-05 08:01:43

Ovid: ICU Book predictive equations (with adjustments for degree of stress) overestimate daily energy needs y 20 to 60% (6,7 and 8). For this reason, measurements of energy e xpenditure are more accurate than predictive equations in patients in the ICU. INDIRECT CALORIMETRY Because it is impossi le to measure meta olic heat production in clinical practi ce, the meta olic energy expenditure is measured indirectly y measuring the who le- ody VO2 and VCO2. This technique is called indirect calorimetry (2). The REE can e derived from the whole- ody VO2 and VCO2 y using the equation shown in Ta le 46.2 (9). The original REE equation, which incorporated a measurement of t he daily urinary nitrogen excretion, was proposed y the Scottish physiologist J . B. de V. Weir in 1949 (10). A num er of adaptations of the original Weir equat ion have een proposed (11,12), ut the REE equations used in the clinical setti ng do not include the urinary nitrogen excretion. Method Indirect calorimetry is performed with specialized instruments called meta olic carts that measure the exchange of O2 and CO2 across the lungs. These instrument s can e placed at the edside, and gas exchange measurements are o tained over 15 to 30 minutes. The VO2 and VCO2 are then extrapolated to a 24-hour period, an d the 24-hour REE is calculated y using an equation similar to the one shown in Ta le 46.2. P.725 Total Energy Expenditure The REE o tained from indirect calorimetry is usually measured for 15 to 30 minu tes, and then extrapolated to a 24-hour period. The total energy expenditure (TE E), measured over 24 hours, is equivalent to the extrapolated REE in patients wh o are not hypermeta olic (13), ut the TEE can e as much as 40% higher than the extrapolated REE in hypermeta olic septic patients (13). Therefore, the REE mea sured over limited periods is not necessarily equivalent to the total daily ener gy expenditure in hypermeta olic patients in the ICU. Limitations Indirect calorimetry is the most accurate method for determining the daily energ y requirements of individual patients in the ICU. However, several factors limit the popularity of indirect calorimetry in the clinical setting. First and forem ost, the technique requires relatively http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 21)03-May-05 08:01:43

Ovid: ICU Book expensive equipment and specially trained personnel, and it is not universally a vaila le. In addition, the oxygen sensor in most meta olic carts is not relia le at inspired oxygen levels a ove 50%, so indirect calorimetry can e unrelia le in patients with respiratory failure who require inhaled oxygen concentrations a ove 50% (2). Because of all these limitations, daily caloric needs are often es timated using predictive formulas such as the HarrisBenedict equations, whereas i ndirect calorimetry (if availa le) is reserved for selected patients who require careful titration of daily energy intake (e.g., ventilator-dependent patients). NONPROTEIN CALORIES The daily energy requirement should e provided y calories derived from car ohy drates and lipids, and protein intake should e used to maintain the stores of e ssential enzymatic and structural proteins. The proportion of daily calories tha t is provided y lipids and car ohydrates is a matter of some de ate, ut no cle ar evidence shows one su strate to e superior to the other as a source of calor ies (2). CARBOHYDRATES Car ohydrates supply approximately 70% of the nonprotein calories in the average American diet. Because the human ody has limited car ohydrate stores (Ta le 46 .3), daily intake of car ohydrates is necessary P.726 to ensure proper functioni ng of the central nervous system, which relies heavily on glucose as its princip al fuel source. However, excessive intake of car ohydrates can prove detrimental for the following reasons. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 21)03-May-05 08:01:43

Ovid: ICU Book TABLE 46.3. ENDOGENOUS FUEL STORES IN HEALTHY ADULTS q Car ohydrates stimulate the release of insulin, and insulin inhi its the mo iliz ation of free fatty acids from adipose tissue. Because adipose tissue fat is the major source of endogenous calories (Ta le 46.3), excessive car ohydrate intake impairs the a ility of the ody to rely on endogenous fat stores during periods of inadequate nutrition. q The oxidative meta olism of glucose produces an a undance of CO2 relative to the oxygen consumed, as indicated y the respiratory quotients listed in Ta le 46.1 . Furthermore, ingestion of excessive car ohydrates leads to de novo lipogenesis , which has a respiratory quotient of 8.0. Therefore, the ingestion of excessive car ohydrates can e accompanied y an exaggerated production of CO2 (14), and this could promote hypercapnia in patients with compromised lung function. In fa ct, excessive calories from any nutrient source can e accompanied y excessive CO2 production (15). LIPIDS Dietary lipids have the highest energy yield of the three organic fuels (Ta le 4 6.1), and lipid stores in adipose tissues represent the major endogenous fuel so urce in healthy adults (Ta le 46.3). Most nutritional regimens use exogenous lip ids to provide approximately 30% of the daily energy needs. Linoleic Acid Dietary lipids are triglycerides, which are composed of a glycerol molecule link ed to three fatty acids. The only dietary fatty acid that is considered essentia l (i.e., must e provided in the diet) is linoleic acid, a long chain, polyunsat urated fatty acid with 18 car on atoms (16). A deficient intake of this essentia l fatty acid produces a clinical disorder characterized y a scaly dermopathy, c ardiac dysfunction, and increased suscepti ility to infections (16). This disord er is prevented y providing 0.5% of the dietary fatty acids as linoleic acid. S afflower oil is used as the source of linoleic acid in most nutritional support regimens. PROTEIN REQUIREMENTS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 21)03-May-05 08:01:43

Ovid: ICU Book The goal of protein intake is to match the rate of protein cata olism in the ind ividual patient. Protein intake can e estimated y using P.727 the following ge neralized predictions for normal and hypercata olic patients (17): Condition Normal meta olism Hypercata olism Daily Protein Intake 0.8 to 1.0 g/kg 1.2 to 1.6 g/kg The estimated protein intake in hypercata olic patients is limited y the ina il ity to determine the severity of protein cata olism. A more accurate assessment of daily protein requirements requires some measure of protein cata olism. This measure is the urinary excretion of nitrogen, as descri ed elow. NITROGEN BALANCE Two-thirds of the nitrogen derived from protein reakdown is excreted in the uri ne (17). Because protein is 16% nitrogen, each gram of urinary nitrogen (UN) rep resents 6.25 g of degraded protein. The total- ody nitrogen (N) alance can ther efore e determined as follows (18): where UUN is the urinary urea nitrogen excretion (in grams) in 24 hours, and the factor 4 represents the daily nitrogen loss (in grams) other than UUN. If the U UN is greater than 30 (g/24 hours), a factor of 6 is more appropriate for the da ily nitrogen losses other than UUN (19). The goal of the nitrogen alance is to maintain a positive alance of 4 to 6 grams. Total versus Urea Nitrogen http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 21)03-May-05 08:01:43

Ovid: ICU Book Under normal circumstances, approximately 85% of the nitrogen in the urine is co ntained in urea and the remainder is contained in ammonia and creatinine. Howeve r, in certain groups of patients in the ICU (e.g., postoperative patients), urea may contain less than 50% of the total nitrogen in the urine (20). Therefore, t he UUN can underestimate urinary nitrogen losses in patients in the ICU. Measuri ng the urinary ammonia excretion in addition to the UUN will provide a more accu rate assessment of the TUN in these patients (21). However, the clinical signifi cance of this added measurement is unknown at present. NITROGEN BALANCE AND CALORIC INTAKE The first step in achieving a positive nitrogen alance is to provide enough non protein calories to spare proteins from eing degraded P.728 to provide energy. This is demonstrated in Figure 46.2, which shows the relationship etween the in take of nonprotein calories and the nitrogen alance. When the daily protein int ake is constant, the nitrogen alance ecomes positive only when the intake of n onprotein calories is sufficient to meet the daily energy needs (i.e., the REE). If the nonprotein calorie intake is insufficient, some of the protein provided in the diet will e roken down to provide calories, which will produce a negati ve nitrogen alance. Therefore, when the daily intake of nonprotein calories is insufficient, increasing the protein intake ecomes an inefficient method of ach ieving a positive nitrogen alance. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 21)03-May-05 08:01:43

Ovid: ICU Book Figure 46.2. The relationship etween daily intake of nonprotein calories (in re lation to the resting energy expenditure) and the nitrogen alance at a constant protein intake. REE = Resting energy expenditure. VITAMIN REQUIREMENTS Twelve vitamins are considered an essential part of the daily diet. The recommen ded daily dose of individual vitamins in enteral and parenteral nutritional regi mens is shown in Ta le 46.4 (2). It is important to emphasize that the daily vit amin requirements may e much higher than indicated in this ta le in seriously i ll, hypermeta olic patients. In fact, deficiencies in several vitamins have een documented in hospitalized patients, despite the daily provision of vitamins in nutritional support regimens (22,23). The normal vitamin levels in lood are in cluded in the section on Reference Ranges in the Appendix at the end of this tex t. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 21)03-May-05 08:01:43

Ovid: ICU Book TABLE 46.4. RECOMMENDED DAILY REQUIREMENTS FOR VITAMINS Although it is impossi le to comment on the importance of each of P.729 the vita mins in ICU patients, the following comments on thiamine and the antioxidant vit amins are deserved. THIAMINE Thiamine (vitamin B1) is a component of thiamine pyrophosphate, an essential cof actor in car ohydrate meta olism. Thiamine deficiency is likely to e common in patients in the ICU for the following reasons. First, the normal ody content of thiamine is only approximately 30 mg (24), so assuming a daily thiamine require ment of 3 mg in patients in the ICU (Ta le 46.4), lack of thiamine intake could result in depletion of endogenous thiamine stores after just 10 days. Second, th e use of thiamine is increased eyond expected levels in hypercata olic conditio ns (25) and may also e increased in patients receiving nutritional support with glucose-rich formulas. Third, urinary thiamine excretion is increased y furose mide (26), which is a commonly used diuretic in the ICU. Finally, magnesium is n ecessary for the conversion of thiamine into thiamine pyrophosphate, so magnesiu m depletion (which is common in patients in the ICU) causes a functional form of t hiamine deficiency (27). Clinical Features http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 21)03-May-05 08:01:43

Ovid: ICU Book Four clinical disorders are associated with thiamine deficiency (24,28,29 and 30 ): (a) cardiac dysfunction ( eri eri heart disease), ( ) a meta olic (Wernicke's ) encephalopathy, (c) lactic acidosis (see Fig. 37.1), and (d) a peripheral neur opathy. Similar disorders, such as cardiac dysfunction and meta olic encephalopa thy, are common in patients P.730 in the ICU, and thiamine deficiency should e considered in each case in which one of these disorders is unexplained. Diagnosis The la oratory evaluation of thiamine status is shown in Ta le 46.5. Although pl asma levels of thiamine can e useful in detecting thiamine depletion, the most relia le assay of functional intracellular thiamine stores is the erythrocyte tr ansketolase assay (31). This assay measures the activity of a thiamine pyrophosp hatedependent (transketolase) enzyme in the patient's red lood cells in response to the addition of thiamine pyrophosphate (TPP). An increase in enzyme activity of greater than 25% after the addition of TPP indicates a functional thiamine d eficiency. I use the plasma thiamine levels to screen for thiamine depletion and reserve the transketolase assay for determining the end-point of thiamine reple tion in patients with documented thiamine deficiency. TABLE 46.5. LABORATORY EVALUATION OF THIAMINE STATUS ANTIOXIDANT VITAMINS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 21)03-May-05 08:01:43

Ovid: ICU Book Two vitamins serve as important endogenous antioxidants: vitamin C and vitamin E . Both of these antioxidant vitamins are descri ed in Chapter 3 (see Ta le 3.1), and they will not e descri ed further here. Considering the important role tha t oxidation-induced cell injury may have in multiorgan damage in serious illness es (see Ta le 3.2), it is wise to maintain adequate ody stores of the antioxida nt vitamins in critically ill patients. The increased rates of iological oxidat ion that are common in critical illness are likely to increase the daily require ments for vitamin C and vitamin E far a ove those listed in Ta le 46.4. Therefor e, it is important to monitor vitamin C and vitamin E status carefully in seriou sly ill patients in the ICU (see the Appendix for the normal plasma levels of vi tamins C and E). P.731 ESSENTIAL TRACE ELEMENTS A trace element is a su stance that is present in the ody in amounts less than 50 g per gram of ody tissue (32). Seven trace elements are considered essential in humans (i.e., associated with a deficiency syndrome), and these are listed in Ta le 46.6 along with their recommended daily maintenance doses (2). As with th e vitamin requirements, the trace element requirements in Ta le 46.6 are for hea lthy adults; the requirements in hypermeta olic patients in the ICU may e far g reater. TABLE 46.6. DAILY REQUIREMENTS FOR ESSENTIAL TRACE ELEMENTS The following trace elements are mentioned ecause of their relevance to oxidati on-induced cell injury. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 21)03-May-05 08:01:43

Ovid: ICU Book IRON One of the interesting features of iron in the human ody is how little is allow ed to remain as free, un ound iron. The normal adult has approximately 4.5 g of iron, yet there is virtually no free iron in plasma (33). Most of the iron is o und to hemoglo in, and the remainder is ound to ferritin in tissues and transfe rrin in plasma. Furthermore, the transferrin in plasma is only approximately 30% saturated with iron, so any increase in plasma iron will e quickly ound y tr ansferrin, thus preventing any rise in plasma free iron. Iron and Oxidation Injury One reason why the ody may e so concerned with inding iron is the a ility of free iron to promote oxidation-induced cell injury (see Reference 27 and Referen ce 28 in Chapter 3). As descri ed in Chapter 3, iron in the reduced state (Fe-II ) promotes the formation of hydroxyl radicals (see Figure 3.1), and hydroxyl rad icals are considered the most reactive oxidants known in iochemistry. In this c ontext, the a ility to ind and sequester iron has een called the major antioxi dant function of P.732 plasma (see Reference 28, Chapter 3). This might explain why hypoferremia is a common occurrence in patients who have conditions associat ed with hypermeta olism (34) ( ecause this would limit the destructive effects o f hypermeta olism.) In light of this description of iron, a reduced serum iron l evel in a critically ill patient should not prompt iron replacement therapy unle ss there is evidence of total- ody iron deficiency. This latter condition can e detected with a plasma ferritin level; that is, a plasma ferritin elow 18 g/L i ndicates pro a le iron deficiency, whereas a plasma ferritin a ove 100 g/L means that iron deficiency is unlikely (35). SELENIUM Selenium is an endogenous antioxidant y virtue of its role as a co-factor for g lutathione peroxidase, one of important endogenous antioxidant enzymes (see Ta l e 3.1). Selenium use is increased in acute illness, and plasma selenium levels c an fall to su normal levels within 1 week after the onset of acute illness (36). Since selenium supplementation is not routinely included in parenteral nutritio n support regimens, prolonged parenteral nutrition is accompanied y selenium de ficiency (37). The com ination of increased selenium use and lack of daily selen ium supplementation may make selenium deficiency common in patients in the ICU. Such a condition will promote oxidant cell injury. The assessment of selenium st atus is descri ed in Chapter 3. The minimum daily requirement http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 21)03-May-05 08:01:43

Ovid: ICU Book for selenium is 55 g for women and 70 g for men (38). This requirement is likely t o e much higher in hypermeta olic patients in the ICU. The maximum daily dose o f selenium that is considered safe is 200 g, and this dose is pro a ly more appro priate for ICU patients. FOOD FOR THOUGHT Before leaving this chapter, it is important to point out that there is a fundam ental pro lem with the practice of promoting nutrient intake in critically ill p atients. This pro lem relates to the fate of administered nutrients in seriously ill patients. A normal Nutrient Processing The goal of nutrient intake in the malnourished patient is to correct the malnou rished state. However, the malnutrition P.733 that accompanies critical illness is different from the malnutrition that accompanies starvation, and this differe nce has important implications for the value of nutrient intake in each situatio n. The malnutrition from starvation is due to deficits in the ody stores of ess ential nutrients, and thus nutrient intake will correct this type of malnutritio n. However, the malutrition that accompanies serious illnesses is due to disease -induced a normal nutrient processing, and thus nutrient intake will not correct this type of malnutrition as long as the underlying disease is active. In other words, the important factor in correcting the malnutrition from critical illnes s is not the intake of nutrients, ut a decrease in the activity of the underlyi ng disease process (39). In fact, in the setting of a normal nutrient processing , nutrient intake can e used to generate meta olic toxins. This is demonstrated elow. Nutrient Toxicity In healthy su jects, less than 5% of exogenously administered glucose is meta ol ized to form lactate. However, in acutely ill patients, up to 85% of an exogenou s glucose load can e recovered as lactate (40). The graph in Figure 46.3 demons trates the a ility of exogenous glucose to generate lactic acid in acutely stres sed patients undergoing major surgery (41). In this case, patients undergoing a dominal aneurysm surgery were given intraoperative fluid therapy with either Rin ger's solutions or 5% dextrose solutions. In the patients who received the 5% de xtrose solution (total amount of dextrose infused averaged P.734 200 g), the lo od lactate increased y 3 mmol/L, whereas in the patients who received an equiva lent volume of the glucose-free (Ringer's) solution, the lood lactate level inc reased only http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 21)03-May-05 08:01:43

Ovid: ICU Book 1 mmol/L. Thus an organic nutrient (car ohydrate) can e used to generate a meta olic toxin (lactic acid) when nutrient processing is a normal (during the stres s of a dominal aneurysm surgery).

Figure 46.3. Effect of car ohydrate infusion on arterial lactate levels during a dominal aortic surgery. Each point represents a mean lactate level for 10 patie nts receiving Ringer's solution (closed squares) and 10 patients receiving 5% de xtrose solution (open squares) for intraoperative fluid management. Total volume infused is equivalent with oth fluids. (Data from Reference 41.)

The study in Figure 46.3 illustrates that nutrient intake can have very differen t consequences in different su jects, and that nutrients can ecome toxins in th e diseased host. Lucretius realized this more than 2000 years ago when he stated , What is food to one man may e fierce poison to others. For this reason, one sho uld not immediately jump on the andwagon for aggressive nutritional support in critically ill patients. REFERENCES REVIEWS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 21)03-May-05 08:01:43

Ovid: ICU Book 1. Mandt JM, Teasley-Straus erg KM, Shronts EP. Nutritional requirements. In: Te asleyStraus erg KM, ed. Nutrition support hand ook. Cincinnati, OH: Harvey-Whitn ey Books, 1992;1936 (61 references). 2. McClave SA, Snider HL. Use of indirect calorimetry in clinical nutrition. Nut r Clin Pract 1992;7:207221 (83 references). 3. Dark DS, Pingleton SK. Nutrition and nutritional support in critically ill pa tients. J Intensive Care Med 1993;8:1633 (163 references). DAILY ENERGY EXPENDITURE 4. Harris JA, Benedict FG. A iometric study of asal meta olism in man. Washing ton, DC: Carnegie Institute of Washington, Pu lication 279, 1919. 5. Paauw JD, McCamish MA, Dean RE, Ouelette TR. Assessment of caloric needs in s tressed patients. J Am Coll Nutr 1984;3:5159. 6. Mann S, Westenskow DR, Houtchens BA. Measured and predicted caloric expenditu re in the acutely ill. Crit Care Med 1985;13:173177. 7. Weissman C, Kemper M, Askanazi J, et al. Resting meta olic rate of the critic ally ill patient: measured versus predicted. Anesthesiology 1986;64:673679. P.735 8. Makk LJK, McClave SA, Creech PW. Clinical application of the meta olic cart t o the delivery of total parenteral nutrition. Crit Care Med 1990;18:13201327. 9. Burzstein S, Saphar P, Singer P, Elwyn DH. A mathematical analysis of indirec t calorimetry measurements in critically ill patients. Am J Clin Nutr 1989;50:22 7230. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 21)03-May-05 08:01:43

Ovid: ICU Book 10. Weir JB de V. New methods for calculating meta olic rate with special refere nce to protein meta olism. J Physiol 1949;109:19. 11. Westenskow DR, Schipke CA, Raymond JL, et al. Calculation of meta olic expen diture and su strate utilization from gas exchange measurements. J Parent Ent Nu tr 1988;12:2024. 12. Cunningham JJ. Calculation of energy expenditure from indirect calorimetry: assessment of the Weir equation. Nutrition 1990;6:222223. 13. Koea JB, Wolfe RR, Shaw JHF. Total energy expenditure during total parentera l nutrition: am ulatory patients at home versus patients with sepsis in surgical intensive care. Surgery 1995;118:5462. NONPROTEIN CALORIES 14. Rodriguez JL, Askanazi J, Weismann C, et al. Ventilatory and meta olic effec ts of glucose infusions. Chest 1985;88:512518. 15. Talpers SS, Rom erger DJ, Bunce SB, Pingleton SK. Nutritionally associated i ncreased CO2 production. Chest 1992;102:551555. 16. Linscheer WG, Vergroesen AJ. Lipids. In: Shils ME, Olson JA, Shike M, eds. M odern nutrition in health and disease. 8th ed. Philadelphia: Lea & Fe iger, 1994 ;4788. PROTEIN REQUIREMENTS 17. Crim MC, Munro HN. Proteins and amino acids. In: Shils ME, Olson JA, Shike M , eds. Modern nutrition in health and disease. 8th ed. Philadelphia: Lea & Fe ig er, 1994;3 35. 18. Black urn G, Bistrian B, Maini B, et al. Nutritional and meta olic assessmen t of the hospitalized patient. J Parent Ent Nutr 1977;1:1122. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 21)03-May-05 08:01:43

Ovid: ICU Book 19. Velasco N, Long CL, Otto DA, et al. Comparison of three methods for the esti mation of total nitrogen losses in hospitalized patients. J Parent Ent Nutr 1990 ;14:517522. 20. Konstantinides F, Konstantinides N, Li J, et al. Urinary urea nitrogen: too insensitive for calculating nitrogen alance studies in surgical clinical nutrit ion. J Parent Ent Nutr 1991;15:189193. 21. Burge JC, Cho an P, McKnight T, et al. Urinary ammonia excretion as an estim ate of total urinary nitrogen in patients receiving parenteral nutritional suppo rt. J Parent Ent Nutr 1993;17:529531. VITAMIN REQUIREMENTS 22. Dempsey DT, Mullen JL, Rom eau JL, et al. Treatment effects of parenteral vi tamins in total parenteral nutrition patients. J Parent Ent Nutr 1987;11:229237. 23. Beard M, Hatipov C, Hamer J. Acute onset of folate deficiency in patients un der intensive care. Crit Care Med 1980;8:500503. 24. Tanphaichitr V. Thiamine. In: Shils MER, Olson JA, Sike M, eds. Modern nutri tion in health and disease. 8th ed. Philadelphia: Lea & Fe iger, 1994;359365. 25. McConachie I, Haskew A. Thiamine status after major trauma. Intensive Care M ed 1988;14:628631. 26. Seligmann H, Halkin H, Rauchfleisch S, et al. Thiamine deficiency in patient s with congestive heart failure receiving long-term furosemide therapy: a pilot study. Am J Med 1991;91:151155.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 21)03-May-05 08:01:43

27. Dyckner T, Ek B, Nyhlin H, Wester PO. Aggravation of thiamine deficiency magnesium depletion. Acta Med Scand 1985;218:129131.

Ovid: ICU Book 28. Tan GH, Farnell GF, Hemsrud DD, Litin SC. Acute Wernicke's encephalopathy at tri uta le to pure dietary thiamine deficiency. Mayo Clin Proc 1994;69:849850. P. 736 29. Oriot D, Wood C, Gottesman R, Huault G. Severe lactic acidosis related to ac ute thiamine deficiency. J Parent Ent Nutr 1991;15:105109. 30. Skelton WP, Skelton NK. Thiamine deficiency neuropathy: it's still common to day. Postgrad Med 1989;85:301306. 31. Boni L, Kieckens L, Hendrikx A. An evaluation of modified erythrocyte transk etolase assay for assessing thiamine nutritional adequacy. J Nutr Sci Vitaminol 1980;26:507514. ESSENTIAL TRACE ELEMENTS 32. Fleming CR. Trace element meta olism in adult patients requiring total paren teral nutrition. Am J Clin Nutr 1989;49:573579.

34. Shan ogue LKR, Paterson N. Effect of sepsis and surgery on trace minerals. J Parent Ent Nutr 1990;14:287289. 35. Guyatt GH, Patterson C, Ali M, et al. Diagnosis of iron deficiency anemia in the elderly. Am J Med 1990;88:205209. 36. Hawker FH, Stewart PM, Switch PJ. Effects of acute illness on selenium homeo stasis. Crit Care Med 1990;18:442446. 37. Sando K, Hoki M, Nezu R, et al. Platelet glutathione peroxidase activity in long-term total parenteral nutrition with and without selenium supplementation. J Parent Ent Nutr 1992;16:5468. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 21)03-May-05 08:01:43

33. Halliwell B, Gutteridge JMC. Free radicals in Oxford: Clarendon Press, 1989;3438.

iology and medicine. 2nd ed.

Ovid: ICU Book 38. Food and Nutrition Board, National Research Council. Recommended dietary all owances. 10th ed. Washington, DC: National Academy Press, 1989;217223. FOOD FOR THOUGHT 39. Marino PL, Finnegan MJ. Nutrition support is not eneficial, and can e harm ful, in critically ill patients. Crit Care Clin 1996;12:667676. 40. Gunther B, Jauch K-W, Hartl W, et al. Low-dose glucose infusion in patients who have undergone surgery. Arch Surg 1987;122:765771. 41. Degoute C-S, Ray M-J, Manchon M, et al. Intraoperative glucose infusion and lood lactate: endocrine and meta olic relationships during a dominal aortic sur gery. Anesthesiology 1989;71:355361.

SUGGESTED READINGS Burzstein S, Elwyn DH, Askanazi J, Kinney JM, eds. Energy meta olism, indirect c alorimetry, and nutrition. Baltimore: Williams & Wilkins, 1989. Food and Nutrition Board, National Research Council. Recommended dietary allowan ces. 10th ed. Washington, DC: National Academy Press, 1989. Shils ME, Olson JA, Shike M, eds. Modern nutrition in health and disease. 8th ed . Philadelphia: Lea & Fe iger, 1994. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 21)03-May-05 08:01:43

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 47 ENTERAL NUTRITION One of the important features of the gastrointestinal (GI) tract (as descri ed i n Chapter 6 and Chapter 33) is the role of the intestinal epithelium as a arrie r to invasion y pathogenic microorganisms. As discussed in this chapter, the a rrier function of the owel mucosa is maintained y the intake and processing of ulk nutrients along the digestive tract. Therefore, providing nutrients via th e enteral route not only provides nutritional support for the vital organs, ut also supports host defenses against invasive infection (1,2 and 3). TROPHIC EFFECT OF ENTERAL NUTRIENTS Complete owel rest is accompanied y progressive atrophy and disruption of the intestinal mucosa. This effect ecomes evident after just a few days and is not prevented y parenteral (intravenous) nutrition. The influence of luminal nutrie nts on the morphology of the intestinal mucosa is shown in Figure 47.1 (4). The photomicrograph at the top shows the normal appearance of the small owel mucosa . Note the fingerlike projections (microvilli), which serve to increase the surf ace area for nutrient a sorption. The photomicrograph at the ottom shows the mu cosal changes after 1 week of a protein-deficient diet. Note the shortening of t he microvillus on the left of the picture and the generalized disruption of the surface architecture. This demonstrates that depletion of nutrients in the owel lumen is accompanied y degenerative changes in the owel mucosa. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 22)03-May-05 08:03:30

Ovid: ICU Book Figure 47.1. Photomicrographs showing the normal appearance of the small owel m ucosa (upper), and the mucosal changes after 1 week of a protein-deficient diet (lower). (Reprinted with permission from Deitch EA et al. Ann Surg 1987;205:68169 0.) O servations like those in Figure 47.1 indicate that the owel mucosa relies on nutrients in the owel lumen to provide its nutritional needs. One of the nutrie nts that may play an important role in this process is the amino acid glutamine, which is considered the principal meta olic fuel for intestinal epithelial cell s (5). The use of glutamine in enteral feedings is discussed later in this chapt er. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 22)03-May-05 08:03:30

Ovid: ICU Book P.738 P.739 TRANSLOCATION The process of translocation, where enteric pathogens move across the owel muco sa and into the systemic circulation, is descri ed in Chapter 6 (see Figure 6.1) . Translocation has een documented during periods of owel rest in critically i ll patients (6), and this has een attri uted to mucosal disruption from lack of luminal nutrients. This means that enteral nutrition could help prevent translo cation and su sequent sepsis y maintaining the functional integrity of the owe l mucosa. The potential for enteral nutrition to prevent sepsis of owel origin is one of the major reasons why enteral nutrition has ecome favored over parent eral (intravenous) nutrition in critically ill patients.

CONTRAINDICATIONS Enteral feedings in any amount are contraindicated in patients with circulatory shock, intestinal ischemia, complete mechanical owel o struction, or ileus. Tot al enteral nutrition is not advised in patients with the following conditions: p artial mechanical owel o struction, severe or unrelenting diarrhea, pancreatiti s, or high-volume (more than 500 mL daily) enterocutaneous fistulas. Partial (lo w volume) enteral support is, however, possi le in these conditions (1). In the case of pancreatitis, enteral feedings can e delivered into the jejunum (see Jej unostomy Feedings). FEEDING TUBES Standard nasogastric tu es (14 to 16 French) are no longer favored for enteral t u e feedings ecause of patient discomfort. Although there has een concern a ou t gastroesophageal reflux with these tu es, clinical studies do not support this concern (7). The feeding tu es that are currently favored are narrower (8 to 10 French) and more flexi le than standard nasogastric tu es (8). Because these tu es are so flexi le, a rigid stylet is also provided to facilitate insertion. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 22)03-May-05 08:03:30

PATIENT SELECTION In the a sence of contraindications, enteral tu trient intake has een inadequate for more than isk of acterial translocation across the owel ngs should e started as soon as possi le after intake (1,2).

e feedings are indicated when nu 5 days. In patients who are at r (e.g., urn victims), tu e feedi the onset of inadequate nutrient

Ovid: ICU Book INSERTION Feeding tu es are inserted through the nares and advanced into the stomach or du odenum. The distance that the tu e must e advanced P.740 to reach the stomach c an e estimated y measuring the distance from the tip of the nose to the earlo e and then to the xiphoid process (9). Proper placement in the stomach is someti mes possi le to determine y measuring the pH (with litmus paper) of a specimen aspirated from the tip of the feeding tu e (10). If the specimen has a pH less t han 5, the tip of the tu e is likely to e in the stomach. Feeding tu es that ar e equipped with a pH sensor (GrapHpro e, FT Zinetics Corp., Salt Lake City, Utah ) are also availa le. Remem er that pH testing will e unrelia le in patients re ceiving histamine H2 lockers. TRACHEAL INTUBATION The principal complication of feeding tu e placement is accidental tracheal intu ation (8). Because feeding tu es are narrow, they readily pass through the lary nx and around the inflated cuffs on tracheal tu es. Accidental intu ation of the trachea is often asymptomatic (pro a ly ecause of sedation, depressed consciou sness, or an a normal cough reflex), and in the a sence of symptoms, tu es can e advanced into the distal airways. This is illustrated in Figure 47.2, which sh ows the tip of a small- ore feeding tu e in the lower lo e of the right lung. If feeding tu es are advanced too far into the lungs, the rigid stylet makes it ea sy to puncture the visceral pleura and produce a pneumothorax (8). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 22)03-May-05 08:03:30

Ovid: ICU Book Figure 47.2. Routine postinsertion chest x-ray film showing a small- ore feeding tu e in the lower lo e of the right lung. P.741 Because of the risk of asymptomatic intu ation of the lungs, a postinserti on chest x-ray study is often required to evaluate tu e placement (unless pH tes ting confirms gastric placement). Auscultation of the upper a domen while insuff lating air through the tu e is not a relia le method for excluding tu e malposit ion in the lungs ecause sounds emanating from a tu e in the lower airways can e transmitted into the upper a domen (8). DUODENAL PLACEMENT For those who prefer tu e feedings placed in the duodenum instead of the stomach , gastric tu es must e advanced past the pylorus and into the duodenum. This ca n sometimes e accomplished y specialized maneuvers at the edside (see Referen ce 11) or may require fluoroscopic guidance. Tu e passage into the duodenum can e confirmed y an increase in the pH of feeding tu e aspirates to a ove 6.0 (7) , or y radiographic localization. Feeding Site The proposed advantage of duodenal feedings is a reduced risk of reflux of feedi ng solution into the esophagus and su sequent pulmonary aspiration. However, cli nical studies show that http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 22)03-May-05 08:03:30

Ovid: ICU Book the risk of aspiration in duodenal feedings is the same as that in gastric feedi ngs (3,12). Therefore, the time and effort devoted to advancing gastric tu es in to the duodenum is not justified. FEEDING FORMULAS More than 80 liquid feeding formulas are availa le for enteral nutrition (13). T he formulas that are most commonly used are descri ed in Ta le 47.1, Ta le 47.2, Ta le 47.3, Ta le 47.4 and Ta le 47.5. TABLE 47.1. CHARACTERISTICS OF SELECTED ENTERAL FEEDING FORMULAS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 22)03-May-05 08:03:30

Ovid: ICU Book TABLE 47.2. ENTERAL FORMULAS WITH A HIGH CALORIC DENSITY TABLE 47.3. FEEDING FORMULAS WITH AN ALTERED LIPID COMPOSITION TABLE 47.4. GLUTAMINE-ENRICHED FEEDING FORMULAS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 22)03-May-05 08:03:30

Ovid: ICU Book TABLE 47.5. FIBER-ENRICHED ENTERAL FEEDING FORMULAS

CALORIC DENSITY The caloric density of feeding formulas is determined primarily y the car ohydr ate content. Most formulas provide 1 to 2 kilocalories per liter of solution. Th e formulas that provide 1 to 1.5 kcal/L (standard caloric density) are listed in Ta le 47.1, and the formulas that provide 1.5 to 2 kcal/L (high caloric density ) are listed in Ta le 47.2. The energy-rich formulas in Ta le 47.2 are well-suit ed for patients with excessive daily energy needs and for patients who are volum erestricted. OSMOLALITY The osmolality of liquid feeding formulas varies from 280 to 1100 mOsm/kg H2O. T he major determinant of osmolality is the car ohydrate content ( ecause this is the most a undant nutrient in most feeding formulas). Because car ohydrates also determine caloric density, osmolality and caloric density are directly related. Formulas with the lowest caloric density (1 kcal/L) have the lowest osmolalitie s (approximately 300 mOsm/kg H2O) and are usually isotonic to the ody fluids. F ormulas with the highest caloric density (2 kcal/L) have the highest osmolalitie s (1000 mOsm/kg H2O) and are markedly hypertonic to the ody fluids. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 22)03-May-05 08:03:30

P.742 P.743 The following is a ral feeding formulas.

rief description of some of the features of ente

Ovid: ICU Book Hypertonic formulas should e infused into the stomach to take advantage of the dilutional effects of the gastric secretions. PROTEIN Liquid feeding formulas provide 35 to 40 grams of protein per liter. Although so me formulas are designated as eing protein-rich (these formulas often have the suffix HN to indicate high nitrogen), they provide only 20% more protein than the standard feeding formulas. P.744 Protein Complexity Most enteral formulas provide intact proteins that are roken down into amino ac ids in the upper GI tract. Because small peptides are a sor ed more rapidly than amino acids, some feeding formulas contain small peptides instead of intact pro tein to facilitate a sorption. Peptide ased formulas such as Peptamen (Clintec) and Vital HN (Ross) can e used in patients with impaired intestinal a sorption (e.g., from inflammatory owel disease). These formulas also promote water rea s orption from the owel (13), and thus they could prove eneficial in patients wi th severe or unrelenting diarrhea. LIPIDS The lipid emulsions used in feeding formulas are rich in long-chain triglyceride s derived from vegeta le oils. These lipids represent a concentrated source of c alories, with an energy yield (9 kcal/g) that is almost three times that of car ohydrates (3.4 kcal/g). Because excessive fat ingestion is not well tolerated (i .e., it promotes diarrhea), the lipid content of most feeding formulas is limite d to 30% of the total calories. Some formulas with an altered lipid composition are descri ed in the following sections. These formulas are summarized in Ta le 47.3. Lipid-Rich Formula One liquid feeding formula with a high fat content is Pulmocare (Ross), which us es lipids to provide 55% of the total calories. This formula is intended for pat ients with respiratory failure. The proposed enefit is ased on the low rate of CO2 production relative to O2 consumption associated with lipid meta olism (see Ta le 46.1). Thus when lipids replace car ohydrates as the principal nutrient s u strate, meta olic CO2 production will decline and there will e less of a tend ency for CO2 retention in patients with compromised lung function (14). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 22)03-May-05 08:03:30

Ovid: ICU Book Alternative Lipids The two feeding formulas descri ed in Ta le 47.3 contain dietary fat from source s other than vegeta le oils. Polyunsaturated fatty acids from vegeta le oils can serve as precursors for inflammatory mediators (eicosanoids) that are capa le o f producing widespread cell injury (15). Fatty acids derived from alternative so urces such as marine oils (e.g., fish oil) do not promote the production of harm ful inflammatory mediators, and thus they might e preferred to the standard die tary fats to limit the risk of inflammatory-mediated tissue injury. Two feeding formulas contain fatty acids from marine oils, and oth are included in Ta le 47 .3. Impact (Sandoz Nutrition) also contains arginine, which promotes immunocompe tence (16), and Perative (Ross La s) P.745 has eta-carotene, a vitamin A analog ue that can have antioxidant activity. Both formulas are intended for patients w ith systemic inflammation who are at risk for inflammatory-mediated tissue injur y. Impact is also supposed to improve immune function. ADDITIVES GLUTAMINE As mentioned earlier, glutamine is the principal fuel for the owel mucosa (5). Therefore, daily supplementation with glutamine seems a reasona le measure for m aintaining the functional integrity of the owel mucosa. Although glutamine is n ot an essential amino acid ( ecause it is produced in skeletal muscle), tissue g lutamine stores decline precipitously in acute, hypercata olic states. Therefore , glutamine can ecome an essential nutrient in the hypermeta olic, stressed pat ient (17). Glutamine Enriched Formulas Because glutamine is a natural constituent of proteins, all feeding formulas tha t contain intact protein will also contain glutamine (18). However, little of th is glutamine is in the free or un ound form. The formulas that contain glutamine as a free amino acid are listed in Ta le 47.4. With the exception of AlitraQ (R oss La oratories), the glutamine content of enteral feeding formulas is low and may e insufficient to provide a enefit (19,20). In one study of glutamine admi nistration in healthy adults (20), the average glutamine dosage (oral and intrav enous) was 0.35 g/kg/day, or 24.5 g/day for a 70-kg su ject. Assuming a daily ca loric intake of 2000 kcal, the only feeding formula in Ta le 47.4 that will prov ide a glutamine dosage of 0.35 g/kg/day is AlitraQ. In the setting of hypercata olism, the glutamine provided y most enteral formulas will e even more inadequ ate. Therefore, although the use of glutaminehttp://gateway.ut.ovid.com/gw1/ovid we .cgi (10 of 22)03-May-05 08:03:30

Ovid: ICU Book

DIETARY FIBER The term fi er refers to a group of plant products that are not degrada le y hu man digestive enzymes. These products are classified y their fermentative prope rties. Fermenta le fi er (cellulose, pectin, gums) is degraded y intestinal ac teria to form shortchain fatty acids (e.g., acetate), which are used as an energ y su strate y the large owel mucosa. This type of fi er can slow gastric empty ing and ind ile salts, and oth of these actions can help alleviate diarrhea. Nonfermenta le fi er (lignins) is not degraded y intestinal acteria, P.746 ut it can create an osmotic force that adsor s water from the owel lumen. This ty pe of fi er can therefore reduce the tendency for watery diarrhea. Thus fi er ha s several actions that can reduce the tendency for diarrhea during enteral feedi ngs. Furthermore, fermenta le fi er can serve as a source of meta olic support f or the mucosa of the large owel (21). This latter effect could play an importan t role in limiting the tendency for translocation across a disrupted large owel mucosa. Several feeding formulas contain fi er, and they are shown in Ta le 47. 5. The added fi er in all cases is soy polysaccharide, which is a fermenta le fi er. Thus there is little difference etween the fi er enriched formulas, either in type or in content of fi er. Fi er enhanced feedings can also e achieved y adding Metamucil (nonfermenta le fi er) or Kaopectate (fermenta le fi er) to th e feeding regimen. Performance The effects of fi er enriched feedings on the incidence of diarrhea have een in consistent, with some studies showing a reduced incidence of diarrhea (22), and others showing no effect (23). However, relying on fi er to prevent diarrhea neg lects the source of the diarrhea; the focus of prevention should e to eliminate or treat the process responsi le for the diarrhea.

MISCELLANEOUS Branched Chain Amino Acids http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 22)03-May-05 08:03:30

fortified enteral formulas seems reasona le, the amount of glutamine provided most formulas may e inadequate.

Ovid: ICU Book The ranched chain amino acids (BCAAs) isoleucine, leucine, and valine are avail a le in feeding formulas intended for trauma victims and patients with hepatic e ncephalopathy. In trauma victims, the BCAAs can e used as a fuel source in skel etal muscle, there y sparing the degradation of other muscle proteins to provide energy. In hepatic encephalopathy, the BCAAs can antagonize the uptake of aroma tic amino acids (e.g., tryptophan) into the central nervous system, and this hel ps prevent the su sequent reakdown of the aromatic amino acids to form false ne urotransmitters, which have een implicated in the pathogenesis of hepatic encep halopathy (24). Examples of feeding formulas enriched with BCAAs include Traum-A id (McGaw) and Stresstein (Sandoz) for trauma victims, and Hepatic-Aid II (McGaw ) and Travasor Hepatic (Clintec) for hepatic encephalopathy. The enefits of th ese formulas are unproven. Carnitine Carnitine is necessary for the transport of fatty acids into mitochondria for fa tty acid oxidation (25). Deficiency of carnitine can occur in prolonged states o f hypercata olism or when carnitine intake is P.747 eliminated. The clinical con sequences of carnitine deficiency include cardiomyopathy, skeletal muscle myopat hy, and hypoglycemia. The recommended daily dose of carnitine is 1 to 3 g in adu lts (13). Enteral formulas that are supplemented with carnitine include Glucerna (Ross), Isocal HN (Mead Johnson), Jevity (Ross), and Peptamen (Clintec). FEEDING REGIMEN Tu e feedings are usually infused for 12 to 16 hours in each 24-hour period. Con tinuous infusion without a period of owel rest is an unrelenting stress to the owel mucosa and promotes mala sorption and diarrhea. Intermittent olus feeding s more closely approximate the normal condition, ut the volumes required are of ten too large to e given safely.

GASTRIC RETENTION Before gastric feedings are started, it is necessary to determine how much volum e will e retained in the stomach over a 1-hour period ecause this will determi ne how fast the feedings can e administered. A volume of water that is equivale nt to the desired hourly feeding volume should e infused over 1 hour. After the infusion is complete, the feeding tu e should e clamped for 30 minutes. The tu e should then e unclamped, and any residual volume should

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 22)03-May-05 08:03:30

Ovid: ICU Book e aspirated from the stomach. If the residual lume infused, gastric feeding can proceed (9). vely high, duodenal or jejunal feedings may e c residual volume is measured, it is important olus ecause this will produce acute gastric tion of the residual volume. volume is less than 50% of the vo If the residual volume is excessi more appropriate. When the gastri not to administer the volume as a distension and lead to overestima

STARTER REGIMENS The traditional approach to initiating tu e feedings is to egin with dilute for mulas and a slow infusion rate and gradually advance the formula concentration a nd infusion rate over the next few days until the desired nutrient intake is ach ieved. This presuma ly allows the atrophic owel mucosa time to regenerate after a period of owel rest. The draw ack with starter regimens is the fact that nut rient intake is inadequate for the time required to advance to full nutritional support. In the malnourished patient, this added period of inadequate nutrition adds to the malnutrition. Studies involving intragastric feedings show that full feedings can e delivered immediately without trou lesome vomiting or diarrhea (26,27). This is presuma ly due to the a ility of gastric secretions to P.748 di lute the feeding formula and reduce the osmotic load associated with the feeding s. Therefore, starter regimens are unnecessary for gastric feedings. Because of the limited reservoir function of the small owel, starter regimens are usually required for duodenal and jejunal feedings. COMPLICATIONS The complications associated with enteral feedings include occlusion of the feed ing tu e, reflux of gastric contents into the airways, and diarrhea. TUBE OCCLUSION Narrow- ore feeding tu es can ecome occluded y accumulation of residue from th e feeding formulas. One important mechanism is the precipitation of proteins in the feeding solution y acidic gastric juice that refluxes up the feeding tu es (28). Standard preventive measures include flushing the feeding tu es with 30 mL of water every 4 hours, and using a 10-mL water flush after medications are ins tilled (29). Relieving the O struction http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 22)03-May-05 08:03:30

Ovid: ICU Book If there is still some flow through the tu e, warm water should e injected into the tu e and agitated with a syringe. This can relieve the o struction in 30% o f cases (30). If this is ineffective, pancreatic enzyme can e used as follows ( 30): Dissolve 1 ta let of Viokase and 1 ta let of sodium car onate (324 mg) in 5 mL of water. Inject this mixture into the feeding tu e and clamp for 5 minutes. Follow with a warm water flush. This should relieve the o struction in approxim ately 75% of cases (30). If the tu e is completely occluded and it is impossi le to introduce warm water or pancreatic enzyme, an attempt should e made to inse rt a flexi le wire or a drum cartridge catheter to clear the o struction. ASPIRATION Retrograde regurgitation of feeding formula is reported in as many as 80% of pat ients receiving gastric or duodenal feedings (8). As stated earlier, the risk of reflux in gastric feedings is the same as that in duodenal feedings (3,12). Ele vating the head of the ed to 45 degrees can reducealthough not eliminatethe risk of reflux (8). Glucose Reagent Strips Aspiration of feeding formulas into the airways can e detected y testing trach eal aspirates with glucose oxidase reagent strips. The P.749 results are measure d with an automated glucose meter (AccuCheck, Boehringer Mannheim). A glucose co ncentration greater than 20 mg/dL in tracheal aspirates is evidence of aspiratio n (31). Coloring the feeding formulas with food coloring and inspecting the colo r of the tracheal secretions is an insensitive method for detecting aspiration ( 31). DIARRHEA Diarrhea occurs in approximately 30% of patients receiving enteral tu e feedings (32). Although the hypertonicity of enteral feeding formulas can induce an osmo tic diarrhea, in most cases of diarrhea associated with enteral feedings, the fe eding formula is not responsi le for the diarrhea (32,33). The cause of the diar rhea in many cases is a medicinal elixir that contains sor itol (an osmotic agen t) to improve palata ility (32,34). Some of the sor itolcontaining liquid drug p reparations are shown in Ta le 47.6 (34). Also shown is the daily dosage of sor itol that would accompany each drug when given in the usual therapeutic dosages (34). In most cases, the daily dosage of sor itol can e enough to induce an osm otic diarrhea. Therefore, a search for sor itol-containing medicinal elixirs sho uld e the first concern in the evaluation of diarrhea during enteral tu e feedi ngs. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 22)03-May-05 08:03:30

Ovid: ICU Book TABLE 47.6. SORBITOL-CONTAINING LIQUID DRUG PREPARATIONS Stool Osmolal Gap Clostridium difficile enterocolitis is also a possi le cause of diarrhea during enteral feedings. One method of differentiating the secretory diarrhea caused y C. difficile enterocolitis from the osmotic diarrhea caused y hypertonic feedi ngs or medicinal elixirs is to calculate the stool osmolal gap as follows: P.750 A stool osmolal gap greater than 160 mOsm/kg H2O suggests an osmotic diarr hea secondary to hypertonic tu e feedings or medicinal elixirs, whereas a smalle r (or negative) osmolal gap suggests a secretory diarrhea caused y C. difficile enterocolitis. (For more information on the diagnosis and treatment of C. diffi cile enterocolitis, see Chapter 33.) JEJUNOSTOMY FEEDINGS Although a dominal surgery usually is accompanied y 24 to 48 hours of gastric h ypomotility, the motility of the small owel is often unimpaired (35). Infusion of liquid feeding formulas into the jejunum takes advantage of the continued sma ll owel motility after a dominal surgery and allows immediate postoperative nut rition. Jejunal feedings can also e performed for nutritional support of patien ts with pancreatitis. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 22)03-May-05 08:03:30

Ovid: ICU Book NEEDLE CATHETER JEJUNOSTOMY A feeding jejunostomy can e performed as a complimentary procedure during laparot omy. A needle catheter jejunostomy is shown in Figure 47.3 (36). A loop of jejun um (15 to 20 cm distal to the ligament P.751 of Treitz) is mo ilized to the ante rior a dominal wall, and a 16-gauge catheter is tunneled through the su mucosa o f the jejunum for a distance of 30 to 45 cm and then advanced into the owel lum en. The jejunum is then secured to the peritoneum on the underside of the a domi nal wall, and the catheter is secured to the skin. Figure 47.3. A needle catheter jejunostomy. Feeding Method As mentioned, the small owel does not have the reservoir capacity of the stomac h, so starter regimens are recommended for jejunal feedings. Feedings are usuall y initiated at a rate of 15 to 25 mL/hour, and the infusion rate is gradually in creased over the next few days until full nutritional support is achieved (37). Catheters are flushed with 10 mL of isotonic saline every 6 hours to promote cat heter patency. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 22)03-May-05 08:03:30

Ovid: ICU Book Complications The principal complications of needle catheter jejunostomies are diarrhea and oc clusion of the narrow feeding catheters (37). Because of the latter complication , needle catheter jejunostomies are used only for temporary nutritional support (approximately 1 week). If more prolonged jejunal feedings are desired, a needle catheter jejunostomy can e converted to a standard jejunostomy (which uses a 1 2 French feeding tu e) using a technique descri ed in Reference 38. REFERENCES GUIDELINES 1. A.S.P.E.N. Board of Directors. Guidelines for the use of parenteral and enter al nutrition in adult and pediatric patients. J Parent Ent Nutr 1993;17(Suppl):1 SA51SA. 2. American Gastroenterological Association Medical Position Statement: Guidelin es for the use of enteral nutrition. Gastroenterology 1995;108:12801281. REVIEWS 3. Benya R, Mo arhan S. Enteral alimentation: administration and complications. J Am Coll Nutr 1991;10:209219 (92 references).

5. Herskowitz A, Sou a WW. Intestinal glutamine meta olism during critical illne ss: a surgical perspective. Nutrition 1990;6:199206. 6. Mainous MR, Deitch EA. The gut arrier. In: Zaloga GP, ed. Nutrition in criti cal care. St. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 22)03-May-05 08:03:30

TROPHISM 4. Deitch EA, Wintertron J, Li MA, Berg R. The gut as a portal of entry for eremia. Ann Surg 1987;205:681690.

act

Ovid: ICU Book Louis: Mos y, 1994;557568. P.752 FEEDING TUBES 7. Dotson RG, Ro inson RG, Pingleton SK. Gastroesophageal reflux with nasogastri c tu es. Am J Respir Crit Care Med 1994;149:16591662. 8. Metheny N. Minimizing respiratory complications of nasoenteric tu e feedings: state of the science. Heart Lung 1993;22:213223. 9. Rom eau JL, Caldwell MD, Forlaw L, Guenter PA, eds. Atlas of nutritional supp ort techniques. Boston: Little, Brown, 1989;77106. 10. Metheny NA, Clouse RE, Clark JM, et al. pH testing of feeding-tu e aspirates to determine placement. Nutr Clin Pract 1994;9:185190. 11. Zaloga GP. Bedside method for placing small owel feeding tu es in criticall y ill patients. Chest 1991;100:16431646. 12. Strong RM, Condon SC, Solinger MR, et al. Equal aspiration rates from post-p ylorus and intragastric-placed small- ore nasoenteric feeding tu es: a randomize d, prospective study. J Parent Ent Nutr 1992;16:5963. FEEDING FORMULAS 13. Shronts EP, Havala T. Enteral formulas. In: Teasley-Straus urg KM, Cerra F, Lehmann S, Shronts EP, eds. Nutrition support hand ook. Cincinnati, OH: Harvey W hitney Books, 1992;147186. 14. Al-Saady NM, Blackmore CM, Bennett ED. High fat, low car ohydrate, enteral f eeding lowers PaCO2 and reduces the period of ventilation in artificially ventil ated patients. Intensive Care Med 1989;15:290295. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 22)03-May-05 08:03:30

Ovid: ICU Book 15. Bagley JS, Wan JMF, Georgieff M, et al. Cellular nutrition in support of mul tiorgan failure. Chest 1991;100(Suppl):182S188S. 16. Daly JM, Lie erman MD, Goldfine J, et al. Enteral nutrition with supplementa l arginine, RNA, and omega-3 fatty acids in patients after operation: immunologi c, meta olic, and clinical outcome. Surgery 1992;112:5667. ADDITIVES 17. Lacey JM, Wilmore DW. Is glutamine a conditionally essential amino acid? Nut r Rev 1990;48:297310. 18. Swails WS, Bell SJ, Borlase BC, et al. Glutamine content of whole proteins: implications for enteral formulas. Nutr Clin Pract 1992;7:7780. 19. AlitraQ. Manufacturer's product description. Ross La oratories, 1991. 20. Ziegler TR, Benfell K, Smith RJ, et al. Safety and meta olic effects of L-gl utamine administration in humans. J Parent Ent Nutr 1990;14(Suppl):137S146S. 21. Palacio JC, Rom eau JL. Dietary fi er: a rief review and potential applicat ion to enteral nutrition. Nutr Clin Pract 1990;5:99106. 22. Homan H-H, Kemen M, Fuessenich C, et al. Reduction in diarrhea incidence y solu le fi er in patients receiving total or supplemental enteral nutrition. J P arent Ent Nutr 1994;18:486490. 23. Frankenfield DC, Beyer PL. Soy-polysaccharide fi er: effect on diarrhea in t u e-fed, head-injured patients. Am J Clin Nutr 1989;50:533538. 24. Alexander WF, Spindel E, Harty RF, Cerda JJ. The usefulness of ranched chai n amino acids in patients with acute or chronic hepatic encephalopathy. Am J Gas troenterol http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 22)03-May-05 08:03:30

Ovid: ICU Book 1989;84:9196. 25. Vockley J. The changing face of disorders of fatty acid oxidation. Mayo Clin Proc 1994;69:249257. P.753 FEEDING REGIMEN 26. Rees RGP, Keohane PP, Grim le GK, et al. Elemental diet administered nasogas trically without starter regimens to patients with inflammatory owel disease. J Parent Ent Nutr 1986;10:258262. 27. Mizock BA. Avoiding common errors in nutritional management. J Crit Illness 1993;10:11161127. COMPLICATIONS 28. Marcuard SP, Perkins AM. Clogging of feeding tu es. J Parent Ent Nutr 1988;1 2:403 405. 29. Benson DW, Griggs BA, Hamilton F, et al. Clogging of feeding tu es: a random ized trial of a newly designed tu e. Nutr Clin Pract 1990;5:107110. 30. Marcuard SP, Stegall KS. Unclogging feeding tu es with pancreatic enzyme. J Parent Ent Nutr 1990;14:198200. 31. Potts RG, Zaroukian MH, Guerrero PA, Baker CD. Comparison of lue dye visual ization and glucose oxidase test strip methods for detecting pulmonary aspiratio n of enteral feedings in intu ated adults. Chest 1993;103:117121. 32. Edes TE, Walk BE, Austin JL. Diarrhea in tu e-fed patients: feeding formula not necessarily the cause. Am J Med 1990;88:9193. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 22)03-May-05 08:03:30

Ovid: ICU Book 33. Eisen erg PG. Causes of diarrhea in tu e-fed patients: a comprehensive appro ach to diagnosis and management. Nutr Clin Pract 1993;8:119123. 34. Cheng EY, Hennen CR, Nimphius N. Unsuspected source of diarrhea in an ICU pa tient. Clin Intensive Care 1992;3:3336. JEJUNOSTOMY FEEDINGS 35. Sagar PM, Kreuger G, Macfie J. Nasogastric intu ation and elective a dominal surgery. Br J Surg 1992;79:11271131. 36. Nance ML, Gorman RC, Morris JB, Mullen JL. Techniques for long-term jejunal access. Contemp Surg 1995;46:2125. 37. Collier P, Kudsk KA, Glezer J, Brown RO. Fi er-containing formula and needle catheter jejunostomies: a clinical evaluation. Nutr Clin Pract 1994;9:101103. 38. Antinori CH, Andrew C, Villanueva DT, et al. A technique for converting a ne edle catheter jejunostomy into a standard jejunostomy. Am J Surg 1992;164:6869. SUGGESTED READINGS Lipman TO, ed. A i liography for specialized nutrition support. 4th ed. Silver Springs, MD: American Society of Parenteral and Enteral Nutrition, 1994. Rom eau JL, Caldwell MD, eds. Clinical nutrition: enteral and tu e feeding. 2nd ed. Philadelphia: WB Saunders, 1990. Teasley-Straus urg KM, Cerra F, Lehmann S, Shronts EP, eds. Nutrition support ha nd ook. Cincinnati, OH: Harvey Whitney Books, 1992. Zaloga GP, ed. Nutrition in critical care. St. Louis: Mos y, 1994. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 22)03-May-05 08:03:30

Ovid: ICU Book http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 22)03-May-05 08:03:30

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 48 PARENTERAL NUTRITION When full nutritional support is not possi le with enteral tu e feedings, the in travenous delivery of nutrients can e used to supplement or replace enteral nut rition. This chapter introduces the asic features of intravenous nutritional su pport (1,2) and explains how to create a total parenteral nutrition (TPN) regime n to meet the needs of individual patients. INTRAVENOUS NUTRIENT SOLUTIONS DEXTROSE SOLUTIONS As mentioned in Chapter 46 and Chapter 47, the standard nutritional support regi men uses car ohydrates to supply approximately 70% of the daily (nonprotein) cal orie requirement. These are provided y dextrose (glucose) solutions, which are availa le in the strengths shown in Ta le 48.1. Because dextrose is not a potent meta olic fuel (see Ta le 46.1), the dextrose solutions must e concentrated to provide enough calories to satisfy daily requirements. As a result, the dextros e solutions used for TPN are hyperosmolar and should e infused through large ce ntral veins. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 15)03-May-05 08:04:52

Ovid: ICU Book TABLE 48.1. INTRAVENOUS DEXTROSE SOLUTIONS P.755 AMINO ACID SOLUTIONS Amino acid solutions are mixed together with the dextrose solutions to provide t he daily protein requirements. A variety of amino acid solutions are availa le f or specific clinical settings, as demonstrated in Ta le 48.2. The standard amino acid solutions contain approximately 50% essential amino acids (N = 9) and 50% nonessential (N = 10) plus semiessential (N = 4) amino acids (3). The nitrogen i n essential amino acids is partially recycled for the production of nonessential amino acids, so the meta olism of essential amino acids produces less of a rise in the lood urea nitrogen concentration than meta olism of nonessential amino acids. For this reason, amino acid solutions designed for use in renal failure a re rich in essential amino acids (see Aminosyn RF in Ta le 48.2). Finally, for r easons stated in Chapter 47, nutritional formulas for hypercata olic conditions (e.g., trauma) and hepatic failure can e supplemented with ranched chain amino acids (isoleucine, leucine, and valine), and two specialty amino acid solutions for each of these conditions are included in Ta le 48.2. It is important to emp hasize that none of these specialized nutrient formulas have improved the outcom es in the disorders for which they are designed (4). TABLE 48.2. STANDARD AND SPECIALTY AMINO ACID SOLUTIONS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 15)03-May-05 08:04:52

Ovid: ICU Book Glutamine As mentioned in Chapter 47, glutamine is the principal meta olic fuel used y in testinal epithelial cells, and lack of glutamine may e at least partly responsi le for the atrophy of the owel mucosa that accompanies prolonged periods of o wel rest (5). Glutamine-enriched TPN has een shown to reduce the atrophic chang es in the owel mucosa during periods of owel rest (5). Therefore, glutamine-su pplemented TPN may play an important role in maintaining the functional P.756 in tegrity of the owel mucosa and preventing acterial translocation (6). Although glutamine is not an essential amino acid ( ecause it is produced in skeletal mu scle), glutamine levels in lood and tissues drop precipitously in acute, hyperc ata olic conditions (e. g., trauma), so glutamine may e a conditionally essentia l amino acid (5). The amino acids that contain glutamic acid are shown in Ta le 4 8.3. Glutamine is formed when glutamic acid com ines with ammonia in the presenc e of the enzyme glutamine synthetase, so glutamic acid can e an exogenous sourc e of glutamine. The value of these glutamate-containing amino acid solutions is unknown at present. TABLE 48.3. AMINO ACID SOLUTIONS WITH GLUTAMATE LIPID EMULSIONS Intravenous lipid emulsions are long-chain triglycerides from vegeta le oils (sa fflower or soy ean oils) and are rich in linoleic acid, an essential, polyunsatu rated fatty acid that is not http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 15)03-May-05 08:04:52

Ovid: ICU Book produced y the human ody (7). As shown in Ta le 48.4, lipid emulsions are avai la le in 10% and 20% strengths (the percentage refers to grams of triglyceride p er 100 mL of solution). The 10% emulsions provide approximately 1 kcal/mL, and t he 20% emulsions provide 2 kcal/mL. Unlike the hypertonic dextrose solutions, li pid emulsions are roughly isotonic to plasma and can e infused through peripher al veins. The lipid emulsions are availa le in unit volumes of 50 to 500 mL and can e infused P.757 separately (at a maximum rate of 50 mL/hour) or added to th e dextroseamino acid mixtures. The triglycerides introduced into the loodstream are not cleared for 8 to 10 hours, and lipid infusions often produce a transient , lipemic-appearing (whitish) plasma. TABLE 48.4. INTRAVENOUS LIPID EMULSIONS FOR CLINICAL USE Lipid Restriction Lipids are used to provide up to 30% of the daily (nonprotein) calorie requireme nts. However, ecause dietary lipids are oxidation-prone and can promote oxidant -induced cell injury (8), restricting the use of lipids in critically ill patien ts (who often have high oxidation rates) seems wise. Although lipid infusion is necessary to prevent essential fatty acid deficiency (cardiomyopathy, skeletal m uscle myopathy), this can e accomplished with minimal amounts of lipid (see foo tnote in Ta le 48.3). ADDITIVES http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 15)03-May-05 08:04:52

Ovid: ICU Book Commercially availa le mixtures of electrolytes, vitamins, and trace elements ar e added directly to the dextroseamino acid mixtures. ELECTROLYTES Most electrolyte mixtures contain sodium, chloride, potassium, and magnesium; th ey also may contain calcium and phosphorous. The daily requirement for potassium or any specific electrolyte can e specified in the TPN orders. If no electroly te requirements are specified, the electrolytes are added to replace normal dail y electrolyte losses. VITAMINS Aqueous multivitamin preparations are added to the dextroseamino acid mixtures. O ne unit vial of a standard multivitamin preparation will provide the normal dail y requirements for most vitamins (see Ta le 46.4), with the exception of vitamin K (9). Enhanced vitamin requirements in hypermeta olic patients in the ICU may not e satisfied. Furthermore, some vitamins are degraded efore they are delive red. Two examples are vitamin A (which is degraded y light) and thiamine (which is degraded y sulfite ions used as preservatives for amino acid solutions) (10 ). TRACE ELEMENTS A variety of trace element additives are availa le, and two commercial mixtures are shown in Ta le 48.5. Most trace element mixtures contain chromium, copper, m anganese, and zinc, ut they do not contain iron and iodine. Some mixtures conta in selenium, and others do P.758 not. Considering the importance of selenium in endogenous antioxidant protection (see Chapter 3), it seems wise to select a tra ce element additive that contains selenium. Routine administration of iron is no t recommended in critically ill patients ecause of the prooxidant actions of ir on (see Chapter 3 and Chapter 46). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 15)03-May-05 08:04:52

Ovid: ICU Book TABLE 48.5. TRACE ELEMENT PREPARATIONS AND DAILY REQUIREMENTS CREATING A TPN REGIMEN The following stepwise approach shows how to create a TPN regimen for an individ ual patient. The patient in this example is a 70-kg adult who is not nutritional ly depleted and has no volume restrictions. Step 1 The first step is to estimate the daily protein and calorie requirements as desc ri ed in Chapter 46. For this example, the daily calorie requirement will e 25 kcal/kg, and the daily protein requirement will 1.4 g/kg. Therefore, for the 70kg patient, the protein and calorie requirements are as follows: Step 2 The next step is to take a standard mixture of 10% amino acids (500 mL) and 50% dextrose (500 mL) and determine the volume of this mixture that is needed to del iver the estimated daily protein requirement. Although the dextroseamino acid mix ture is referred to as A10D50, the final mixture actually represents 5% amino aci ds (50 grams of protein per liter) and 25% dextrose (250 grams dextrose per lite r). Therefore, the volume of the A10D50 mixture needed http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 15)03-May-05 08:04:52

Ovid: ICU Book to provide the daily protein requirement is as follows: P.759 If this mixture is infused continuously over 24 hours, the infusion rate w ill e 1900 mL/24 hours = 81 mL/hour (or 81 microdrops/minute). Step 3 Using the total daily volume of the dextroseamino acid mixture determined in Step 2, the next step is to determine the total calories that will e provided y th e dextrose in the mixture. Using an energy yield of 3.4 kcal/g for dextrose, the total dextrose calories can e determined as follows: Because the estimated requirement for calories is 1750 kcal/day, the dextrose wi ll provide all ut 135 kcal/day. These remaining calories can e provided y an intravenous lipid emulsion. Step 4 If a 10% lipid emulsion (1 kcal/mL) is used to provide 135 kcal/day, the daily v olume of the lipid emulsion will e 135 mL/day. Because the lipid emulsion is av aila le in unit volumes of 50 mL, the volume can e adjusted to 150 mL/day to av oid wastage. Thus volume can e infused at half the maximum recommended rate (50 mL/hour) to minimize the tendency to develop lipemic serum during the infusion.

q Provide standard TPN with A10D50 to run at 80 mL/hour. q Add standard electrolytes, multivitamins, and trace elements. q Give 10% Intralipid: 150 mL to infuse over 6 hours. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 15)03-May-05 08:04:52

Step 5 The daily TPN orders for the previous example can then

e written as follows:

Ovid: ICU Book TPN orders are rewritten each day. Specific electrolyte, vitamin, and trace elem ent requirements are added to the daily orders as needed. The example just prese nted applies to the separate administration of dextroseamino acid mixtures and li pid emulsions. Another practice that is gaining popularity is to add the nutrien t solutions and additives together to form a total nutrient admixture (TNA). Alt hough this simplifies nutrient administration and reduces cost, there are linger ing concerns regarding compati ility (e.g., multivitamin preparations may not e compati le with lipid emulsions). COMPLICATIONS A multitude of complications are associated with parenteral nutrition (2,11). So me of the more prominent ones are mentioned in the following paragraphs. P.760 CATHETER-RELATED COMPLICATIONS Because the dextrose and amino acid solutions are hyperosmolar (Ta le 48.1 and T a le 48.2), TPN is administered through large, central veins. The complications associated with central venous catheters are descri ed in Chapter 5 and Chapter 6. One complication that can e particularly frustrating is the misdirected cath eter, like the one shown in Figure 48.1. In this case, a catheter inserted into the right su clavian vein has entered the internal jugular vein and advanced in a retrograde direction up into the neck. When this occurs, the catheter can e r epositioned over a guidewire as follows (13). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 15)03-May-05 08:04:52

Ovid: ICU Book Figure 48.1. Porta le chest x-ray film showing a central venous catheter that ha s een advanced up into the neck. Catheter Repositioning When a catheter has een misdirected up into the neck, the patient should e pla ced in a semirecum ent or upright position if possi le and the catheter should e withdrawn until only a few centimeters of the catheter tip remains inserted. A flexi le guidewire is then inserted through the catheter and advanced 10 cm. Th e catheter is removed over the guidewire, and a new catheter is inserted and adv anced 15 cm. The guidewire is removed and a Doppler pro e (the one used y nurse s to detect pedal pulses) is placed over the internal jugular vein P.761 in the neck. A olus of saline is then injected through the catheter. If the catheter h as een rethreaded up into the neck, the olus injection will produce an audi le noise from the Doppler pro e. If this occurs, a new catheter will need to e in serted into the internal jugular vein on the same side. If no sound is detected, a repeat chest x-ray study should e o tained to determine whether the catheter has een repositioned in the superior vena cava. CARBOHYDRATE INFUSIONS Hyperglycemia http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 15)03-May-05 08:04:52

Ovid: ICU Book Glucose intolerance is one of the most common complications of TPN. Even though this pro lem can e reduced y providing fewer nonprotein calories as glucose (a nd more as lipids), persistent hyperglycemia usually requires the addition of in sulin to the TPN solutions. It is important to emphasize that insulin adsor s to all plastics and glass used in intravenous infusion sets. The amount lost to ad sorption varies with the amount of insulin added, ut an average loss of 20 to 3 0% should e expected (13). Al umin has een used to reduce insulin inding to i ntravenous infusion sets (13), ut this is a costly and unrelia le measure. Inst ead, the insulin dosage is adjusted to achieve the desired glycemic control. Whe n TPN is discontinued, the insulin requirement will e less than that needed dur ing TPN. Hypophosphatemia The effects of TPN on the serum phosphate level is shown in Figure 43.2. This ef fect is due to enhanced uptake of phosphate into cells associated with glucose e ntry into cells. The phosphate is then used to form thiamine pyrophosphate, an i mportant cofactor in car ohydrate meta olism. Fatty Liver When glucose calories exceed the daily calorie requirements, there is lipogenesi s in the liver and this can progress to fatty infiltration of the liver and elev ated levels of transaminase enzymes in the lood (11,14). It is unclear whether this process has any pathologic consequences or whether it merely serves as a ma rker of excess car ohydrate calories. Hypercapnia Excess car ohydrates promote CO2 retention in patients with respiratory insuffic iency. Although this has een attri uted to the high respiratory quotient associ ated with car ohydrate meta olism (see Ta le 46.1), P.762 this may e a reflecti on of overfeeding in general and not specific overfeeding with car ohydrates (15 ). LIPID INFUSIONS Oxidant Injury One of the major (and often overlooked) toxicities associated with lipid infusio ns is an http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 15)03-May-05 08:04:52

Ovid: ICU Book increased risk of oxidation-induced cell injury (8). The damaging effects associ ated with oxidation of mem rane lipids is descri ed in Chapter 3. The same oxida tion-prone lipids in cell mem ranes (i.e., polyunsaturated fatty acids) are also a undant in intravenous lipid emulsions. Because excessive and unprotected iol ogical oxidation is likely to e common in patients in the ICU (see Ta le 3.2), infusion of oxidation-prone exogenous lipids could promote widespread oxidant da mage in these patients. This consideration deserves much more attention that it currently receives. Impaired Oxygenation Lipid infusions are associated with impaired oxygenation are well known for their a ility to damage the pulmonary em olism syndrome), and infusions of oleic acid (a fatty mulsions) are used to produce acute respiratory distress rimental animals. This effect of lipids may e caused y to the pulmonary capillary endothelium. (16). Free fatty acids capillaries (e.g., fat acid present in lipid e syndrome (ARDS) in expe oxidant-induced injury

Mucosal Atrophy The a sence of ulk nutrients in the owel produces atrophy and disruption of th e owel mucosa. This is descri ed in Chapter 47 and is illustrated in Figure 47. 1. These changes can predispose to translocation of enteric pathogens across the owel mucosa and su sequent septicemia. Because TPN is usually accompanied y owel rest, one of the indirect complications of TPN is acterial translocation a nd sepsis of owel origin (17). As mentioned earlier, glutamine-supplemented TPN may help reduce the risk of this complication. Acalculous Cholecystitis The a sence of lipids in the proximal small owel prevents cholecystokinin-media ted contraction of the gall ladder and the ile stasis that P.763 results may pr omote acalculous cholecystitis (2). This disorder is descri ed in Chapter 33. PERIPHERAL PARENTERAL NUTRITION Parenteral nutrition can occasionally e delivered via peripheral veins for shor t periods. The http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 15)03-May-05 08:04:52

GASTROINTESTINAL COMPLICATIONS Two indirect complications of TPN are related to the a sence of n the owel.

ulk nutrients i

Ovid: ICU Book goal of peripheral parenteral nutrition (PPN) is to provide just enough nonprote in calories to spare the reakdown of proteins to provide energy (i.e., proteinsparing nutritional support). PPN does not create enough of a positive nitrogen alance to uild up protein stores, and thus it is not intended for patients who are protein depleted or for patients who are hypercata olic and at risk of eco ming protein depleted. The osmolarity of peripheral vein infusates should e kep t elow 900 mOsm/L to prevent osmotic damage to the vessels (18). Therefore, PPN must e delivered with dilute amino acid and dextrose solutions. Because lipid emulsions are isotonic to plasma, lipids can e used to provide a significant pr oportion of the nonprotein calories in PPN.

METHOD A common admixture used in PPN is a mixture of 3% amino acids and 20% dextrose. This mixture produces a final concentration of 1.5% amino acids (15 grams of pro tein per liter) and 10% dextrose (100 grams of dextrose per liter), with an osmo larity of approximately 500 mOsm/ L. The dextrose will provide 340 kcal/L, so 2. 5 L of the mixture will provide 850 kcal. If 250 mL of 20% Intralipid is added t o the regimen (adding 500 kcal), the total nonprotein calories will increase to 1350 kcal/day. This should e close to the nonprotein calorie requirement of an average-size adult at rest (25 kcal/kg/day). In hypermeta olic patients, large v olumes of PPN are required to satisfy daily energy requirements. In summary, per ipheral intravenous nutrition can e used as a temporary measure to prevent or l imit protein reakdown in patients who are not already protein depleted and are expected to egin oral feedings within a few days. Postoperative patients seem est suited for this form of nutritional support. REFERENCES GUIDELINES 1. A.S.P.E.N. Board of Directors. Guidelines for the use of parenteral and enter al nutrition in adult and pediatric patients. J Parent Ent Nutr 1993;17(Suppl):1 SA52SA. P.764 REVIEWS http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 15)03-May-05 08:04:52

Ovid: ICU Book 2. Phelps SJ, Brown RO, Helms RA, et al. Toxicities of parenteral nutrition in t he critically ill patient. Crit Care Clin 1991;7:725753 (212 references). INTRAVENOUS NUTRIENT SOLUTIONS 3. Teasley-Staus urg KM. Amino acid solutions. In Teasley-Straus urg KM, Cerra F B, Lehmann S, Shronts EP, eds. Nutrition support hand ook. Cincinnati, OH: Harve y Whitney Books, 1992:4772.

5. Sou a WW, Klim erg VS, Plumley DA, et al. The role of glutamine in maintainin g a healthy gut and supporting the meta olic response to injury and infection. J Surg Res 1990;48:383391. 6. Grant J. Use of L-glutamine in total parenteral nutrition. J Surg Res 1988;44 :506510. 7. Warshawsky KY. Intravenous fat emulsions in clinical practice. Nutr Clin Prac t 1992;7:187196. 8. Hardin TC. Cytokine mediators of malnutrition: clinical implications. Nutr Cl in Pract 1993;8:5559. 9. Manufacturer's product description for M.V.I.-12. West orough, MA: Astra USA, 1995. 10. LaFrance RJ, Miyagawa CI. Pharmaceutical considerations in total parenteral nutrition. In: Fischer JE, ed. Total parenteral nutrition. 2nd ed. Boston: Littl e, Brown, 1991;5798. COMPLICATIONS 11. Perry DA, Markin RS, Rose SG, Schenken JR. Changes in la oratory values in http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 15)03-May-05 08:04:52

4. Andris DA, Krzywda EA. Nutrition support in specific diseases: ack to . Nutr Clin Pract 1994;9:2832.

asics

Ovid: ICU Book

12. Benotti PN, Bistrian BR. Practical aspects and complications of total parent eral nutrition. Crit Care Clin 1987;3:115131. 13. Trissel LA. Hand ook on injecta le drugs. 8th ed. Bethesda, MD: American Soc iety of Hospital Pharmacists, 1994;585590. 14. Freund HR. A normalities of liver function and hepatic damage associated wit h total parenteral nutrition. Nutrition 1991;7:16. 15. Talpers SS, Rom erger DJ, Bunce SB, Pingleton SK. Nutritionally associated i ncreased car on dioxide production. Chest 1992;102:551555. 16. Sleie B, Askanazi J, Rothkopf M, et al. Intravenous fat emulsions and lung f unction: a review. Crit Care Med 1988;16:183193.

PERIPHERAL PARENTERAL NUTRITION 18. Teasley-Straus urg KM. Indications for parenteral and enteral nutrition. In: TeasleyStraus urg KM, Cerra FB, Lehmann S, Shronts EP, eds. Nutrition support h and ook. Cincinnati, OH: Harvey Whitney Books, 1992;3746. P.765 SUGGESTED READINGS Fischer JE. Total parenteral nutrition. 2nd ed. Boston: Little, Brown, 1991. Grant JP. Hand ook of total parenteral nutrition. 2nd ed. Philadelphia: WB Saund ers, http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 15)03-May-05 08:04:52

17. Alverdy JC, Aoys E, Moss GS. Total parenteral nutrition promotes ranslocation from the gut. Surgery 1988;104:185190.

patients receiving total parenteral nutrition. La

Med 1990;21:97102.

acterial t

Ovid: ICU Book 1992. Lipman TO, ed. A i liography for specialized nutrition support. 4th ed. Silver Springs, MD: American Society of Parenteral and Enteral Nutrition, 1994. Teasley-Straus urg KM, Cerra F, Lehmann S, Shronts EP, eds. Nutrition support ha nd ook. Cincinnati, OH: Harvey Whitney Books, 1992. Zaloga GP, ed. Nutrition in critical care. St. Louis: Mos y, 1994. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 15)03-May-05 08:04:52

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 49 ADRENAL AND THYROID DYSFUNCTION Endocrine disorders that involve the adrenal and thyroid glands are noted for th eir a ility to act as catalysts for serious, life-threatening conditions while e scaping notice themselves. This chapter explains how to unmask an underlying or occult disorder of adrenal or thyroid function and how to treat each disorder ap propriately (1,2,3,4,5,6,7 and 8). ADRENAL INSUFFICIENCY The adrenal gland plays a major role in the adaptive response to stress. The adr enal cortex releases glucocorticoids and mineralocorticoids that promote glucose availa ility and maintain extracellular volume, while the adrenal medulla relea ses catecholamines that support the circulation. Attenuation or loss of this adr enal response (i.e., adrenal insufficiency) leads to hemodynamic insta ility, vo lume depletion, and defective energy meta olism. The important feature of adrena l insufficiency is its a ility to remain silent until the adrenal gland is calle d on to respond to a physiologic stress. When this occurs, adrenal insufficiency ecomes an occult catalyst that speeds the progression of acute, life-threateni ng conditions (1,2 and 3). Adrenal insufficiency can e a primary or secondary d isorder (the latter eing due to hypothalamicpituitary dysfunction). The descript ion that follows pertains to primary adrenal insufficiency. PREDISPOSING CONDITIONS Several conditions that are common in patients in the ICU can predispose to prim ary adrenal insufficiency. These include major surgery, circulatory failure, sep tic shock, severe coagulopathy, and human immunodeficiency virus (HIV) infection s (1,2). In some of these conditions, P.767 the adrenal insufficiency is caused y pathologic destruction of the adrenal gland (e.g., coagulopathy with adrenal hemorrhage). In others, the pro lem is diminished adrenal responsiveness (e.g., septic shock). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 14)03-May-05 08:05:54

Ovid: ICU Book INCIDENCE In surveys of random ICU patients, the incidence of adrenal insufficiency has va ried from 0 to 30% in different reports (1,2). In patients with septic shock, th e incidence is higher at 25 to 40% (9,10 and 11). In many of these cases, adrena l insufficiency was not evident clinically ut was uncovered y iochemical evid ence of a normal adrenal responsiveness. In patients with la oratory evidence of adrenal insufficiency, the mortality is more than dou le that of patients with normal adrenal responsiveness (9,10 and 11). CLINICAL MANIFESTATIONS In critically ill patients, the most prominent manifestation of adrenal insuffic iency is hypotension that is refractory to vasopressors (1,2,9,10 and 11). Other features of adrenal insufficiency, such as electrolyte a normalities (hyponatre mia, hyperkalemia), weakness, and hyperpigmentation, are either uncommon or not specific enough to suggest the diagnosis in ICU patients. Hemodynamics In mild or chronic cases of adrenal insufficiency, the hemodynamic changes are o ften a reflection of hypovolemia (low filling pressures, low cardiac output, hig h systemic vascular resistance). In acute adrenal failure, the hemodynamic chang es are similar to those of hyperdynamic shock (high cardiac output, low systemic vascular resistance) (10,12). Because many cases of adrenal insufficiency are u ncovered in patients with septic shock, where the hemodynamic changes are simila r to those of acute adrenal failure (i.e., hyperdynamic shock), it is often impo ssi le to identify adrenal failure ased on hemodynamic profiles in critically i ll patients. ACTH STIMULATION TEST Adrenal insufficiency should e suspected in any patient in the ICU who develops sudden hypotension of unclear etiology or has hypotension that is refractory to vasopressors. In critically ill patients, the diagnostic test of choice for pri mary adrenal insufficiency is the rapid adrenocorticotropic hormone (ACTH) stimu lation test (1,2 and 3). This test evaluates the acute adrenal response to a ol us injection of synthetic ACTH (Cortrosyn). P.768 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 14)03-May-05 08:05:54

Ovid: ICU Book METHOD The rapid ACTH stimulation test can e performed at any time and is not influenc ed y diurnal variations in cortisol secretion (which are often a sent in critic ally ill patients anyway). An initial lood sample is o tained for a plasma cort isol level, and synthetic ACTH (0.25 mg) is injected intravenously. A post-ACTH plasma cortisol level is then o tained 1 hour after the ACTH injection. Results The graph in Figure 49.1 shows the plasma cortisol level efore and after ACTH i njection in three different situations. Under normal conditions, there is more t han a twofold increase in the plasma cortisol level 1 hour after the ACTH inject ion. In the presence of a physiological stress (which pertains to most ICU patie nts), the aseline plasma cortisol level is higher than normal, ut the response to ACTH is lunted. In the setting of adrenal insufficiency, the aseline corti sol is elow normal, and the response to ACTH is markedly diminished. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 14)03-May-05 08:05:54

Ovid: ICU Book Figure 49.1. Plasma cortisol levels efore and after intravenous injection of sy nthetic ACTH (Cortrosyn). (Redrawn from Chernow B. Hormonal and meta olic consid erations in critical care medicine. In: Thompson WL, Shoemaker W, eds. Critical care: state of the art. Volume 3. Fullerton, CA: Society of Critical Care Medici ne, 1982.) Interpretation The interpretation of the ACTH stimulation test is outlined in Ta le 49.1. In th e presence of a physiologic stress, a aseline plasma cortisol level that is a o ve 15 g/dL (414 nmol/L) indicates normal adrenal function. An increment in plasma cortisol (at 1 hour after ACTH injection) of less than 7 g/dL (193 nmol/L) is ev idence of a limited adrenal reserve (as would e seen when the adrenal gland is maximally stimulated). A aseline plasma cortisol of less than 15 g/dL and a post stimulation increment in plasma cortisol of less than 7 g/dL is evidence of prima ry adrenal insufficiency. TABLE 49.1. INTERPRETATION OF THE ACTH STIMULATION TEST STEROID THERAPY In patients with suspected adrenal insufficiency who have severe or refractory h ypotension, steroids can e started immediately, efore the ACTH stimulation tes t is performed. Steroid administration can proceed as follows: http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 14)03-May-05 08:05:54

Ovid: ICU Book Dexamethasone (Decadron) will not interfere with plasma cortisol assay (13) and thus can e given efore or during the ACTH stimulation test. The initial dose s hould e 10 mg (as an intravenous olus), which is equivalent to 270 mg of hydro cortisone (cortisol). q Methylprednisolone (Solu-Medrol) will not interfere with the cortisol assay if a radioimmunoassay (RIA) is performed (13). The initial dose of methylprednisolon e is 60 mg (as an intravenous olus), which is equivalent to 300 mg of hydrocort isone. q After the ACTH stimulation test is completed, empiric therapy P.769 P.770 can e gin with hydrocortisone (Solu-Cortef). The initial dose is 250 mg (intravenous olus), followed y 100 mg given intravenously (IV) every 6 hours until the test results are availa le. q If the ACTH stimulation test is normal, the hydrocortisone can e a ruptly disco ntinued, without a taper. If the test reveals primary adrenal insufficiency, the hydrocortisone should e continued at the dosage of 100 mg IV every 6 hours unt il the patient is no longer in a stressed condition. When this occurs, the daily dose of hydrocortisone should e reduced to 20 mg (which is equivalent to the a mount of cortisol secreted daily y the adrenal glands). EVALUATION OF THYROID FUNCTION La oratory tests of thyroid function can e a normal in 75% of hospitalized pati ents (14) and 90% of critically ill patients (15). In most cases, the a normalit y represents an adaptive response to nonthyroidal illness and is not a sign of p athologic thyroid disease (14,15 and 16). This section descri es the la oratory evaluation of thyroid function (17) and explains how to determine whether a la o ratory a normality represents a disorder of thyroid function or a condition know n as euthyroid sick. A summary of the pertinent thyroid function tests is shown in Ta le 49.2. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 14)03-May-05 08:05:54

Ovid: ICU Book TABLE 49.2. THE LABORATORY EVALUATION OF THYROID FUNCTION SERUM THYROXINE Thyroxine (T4) is the principal hormone secreted y the thyroid gland, ut the a ctive form of the hormone is triiodothyronine (T3), which is formed y the deiod ination of thyroxine in extrathyroidal tissues. Despite eing the active form of thyroid hormone, T3 levels in plasma are not a relia le index of thyroid functi on. Serum T3 levels can e normal in 30% of patients with hypothyroidism, and su normal P.771 in 70% of patients who are euthyroid sick (17). Therefore, measure ments of T3 levels in the lood are not advised for the evaluation of thyroid fu nction (17). T4 plasma is partly ound to carrier proteins such as thyroid indi ng glo ulin. However, the free (un ound) T4 is the active form of the hormone. T he assay for T4 in plasma is an RIA that measures oth the protein- ound (inacti ve) and free (active) hormone levels. Thus, an a normal total T4 assay can represe nt an a normality in protein inding and not an a normality in thyroid function. The total serum T4 is su normal in 50% of patients in the ICU (17), even though most of these patients do not have a normal thyroid function. Therefore, the ro utine assay for (total) T4 levels in plasma is not recommended for the evaluatio n of thyroid function in patients in the ICU (17). FREE THYROXINE INDEX http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 14)03-May-05 08:05:54

Ovid: ICU Book The free thyroxine index (FTI) is an indirect assessment of free (un ound) T4 le vels in plasma. The FTI is the product of the total serum T4 level and the T3 re sin uptake (T3RU). The latter test measures the inding capacity of carrier prot eins for T3. An increase in the T3RU indicates a decrease in the inding capacit y of plasma proteins (which could occur when the amount of protein- ound T4 has increased or when protein concentration has decreased). A decrease in the T3RU i ndicates an increase in protein inding capacity (which could occur when there i s a decrease in the amount of protein- ound T4 or when there is an increase in t he protein concentration). When the total T4 and T3RU change in opposite directi ons (FTI unchanged), the change in the total T4 is due to altered inding y car rier proteins (euthyroid state). When the total T4 and T3RU change in the same d irection (FTI changed), the change in total T4 is due to a normal thyroid functi on. The FTI is increased in hyperthyroidism and decreased in hypothyroidism. FREE THYROXINE The amount of free T4 in plasma can e measured y allowing the plasma to pass t hrough a semipermea le mem rane that allows only the free (un ound) T4 to pass. This test is the most relia le la oratory measure of thyroid function, ut it is a complex assay and is not performed routinely in many clinical la oratories. THYROID-STIMULATING HORMONE When a normal thyroid function is detected y the FTI or free T4 assay, the leve l of thyroidstimulating hormone (TSH) in plasma can help determine whether the p ro lem is a primary thyroid disorder or the result of hypothalamicpituitary dysfu nction. For example, in P.772 hypothyroidism caused y pathologic injury of the thyroid gland (primary hypothyroidism), the negative feed ack effect of T4 on TS H secretion will e lost and the plasma level of TSH will e elevated. On the ot her hand, in hypothyroidism caused y hypothalamicpituitary dysfunction, the leve l of TSH in plasma will e su normal. The normal concentration of TSH in plasma is 0.5 to 3.5 4 mU/L (6). Because TSH is normally present in a low concentration , only an elevated TSH level is interpreta le. In primary hypothyroidism, TSH le vels usually rise a ove 20 mU/L. Both dopamine and high-dosage glucocorticoids c an suppress the increase in TSH in primary hypothyroidism, so a normal TSH in pa tients receiving dopamine infusions or high-dosage steroids does not rule out th e possi ility of primary hypothyroidism. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 14)03-May-05 08:05:54

Ovid: ICU Book HYPERTHYROIDISM Most cases of hyperthyroidism are due to primary thyroidal disorders (e.g., Grav e's disease, autoimmune thyroiditis). Chronic therapy with amiodarone, an iodine -containing antiarrhythmic agent, can also cause hyperthyroidism (18). CLINICAL MANIFESTATIONS Some of the common or characteristic manifestations of hyperthyroidism are liste d in Ta le 49.3. It is important to note that elderly patients with hyperthyroid ism may e lethargic rather than agitated (apathetic thyrotoxicosis). The com in ation of lethargy and unexplained atrial fi rillation is characteristic of apath etic thyrotoxicosis in the elderly (7,19). TABLE 49.3. MANIFESTATIONS OF THYROID DYSFUNCTION Thyroid Storm An uncommon ut severe form of hyperthyroidism known as thyroid storm can e pre cipitated y acute illness or surgery. This condition, P.773 characterized y fe ver, severe agitation, and high-output heart failure, can progress to hypotensio n and coma (20) and is uniformly fatal if overlooked and left untreated. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 14)03-May-05 08:05:54

Ovid: ICU Book DIAGNOSIS As mentioned in the previous section, hyperthyroidism will e accompanied y an elevated FTI and an increased free T4 level in plasma. Because hyperthyroidism i s usually caused y primary thyroidal disorders, the TSH is not a valua le test in hyperthyroidism. MANAGEMENT -Receptor Antagonists Immediate management of trou lesome tachyarrhythmias can e achieved y administ ering intravenous propanolol (1 mg every 5 minutes until the desired effect is a chieved). Oral maintenance therapy (20 to 120 mg every 6 hours) can e used unti l antithyroid drug therapy is effective. Antithyroid Drugs The two drugs used to suppress thyroxine production are methimazole and propylth iouracil (PTU). Methimazole is usually preferred to PTU ecause it causes a more rapid decline in serum thyroxine levels and has a lower incidence of serious si de effects (agranulocytosis) (8). Both antithyroid drugs are given orally. The i nitial dose of methimazole is 10 to 20 mg once a day, and the initial dose of PT U is 75 to 100 mg TID (8). The dose of oth drugs is reduced y 50% after 4 to 6 weeks of therapy. Iodide In severe cases of hyperthyroidism, iodide (which locks thyroxine release from the thyroid gland) can e added to therapy with PTU. Iodide can e given orally as Lugol's solution (4 drops every 12 hours) or intravenously as sodium iodide ( 500 to 1000 mg every 12 hours). If the patient has an iodide allergy, lithium (3 00 mg orally every 8 hours) can e used as a su stitute. Thyroid Storm In addition to the a ove measures, the treatment of thyroid storm often requires aggressive volume infusion (to replace fluid losses as a result of vomiting, di arrhea, and heightened insensi le fluid loss). Thyroid storm can accelerate gluc ocorticoid meta olism and create a relative adrenal insufficiency. Thus in cases of thyroid storm associated P.774 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 14)03-May-05 08:05:54

Ovid: ICU Book with severe or refractory hypotension, hydrocortisone (300 mg IV as a loading do se, followed y 100 mg IV every 8 hours) may help correct the hypotension. Succe ssful management of thyroid storm also requires treatment of the precipitating e vent (20). HYPOTHYROIDISM Hypothyroidism is uncommon in hospitalized patients. When present, most cases re present primary hypothyroidism (16). CLINICAL MANIFESTATIONS Some of the more common or characteristic manifestations of hypothyroidism are l isted in Ta le 49.3. The most common cardiovascular manifestation is pericardial effusion (21), which develops in approximately 30% of cases, and is the most co mmon cause of an enlarged cardiac silhouette in patients with hypothyroidism (21 ). These effusions usually accumulate slowly and do not cause cardiac compromise . Pleural effusions are also common in hypothyroidism. The pleural and pericardi al effusions are due to an increase in capillary permea ility and are exudative in quality. Hypothyroidism can also e associated with hyponatremia and a skelet al muscle myopathy, with elevations in muscle enzymes (creatine phosphokinase, a ldolase, lactate dehydrogenase). Enhanced release of creatinine from skeletal mu scles can also raise the serum creatinine in the a sence of renal dysfunction (2 2). Myxedema Coma Advanced cases of hypothyroidism are accompanied y hypothermia and depressed co nsciousness. Although this condition is called myxedema coma, frank coma is unco mmon (5). The edematous appearance in myxedema is due to intradermal accumulatio n of proteins (5) and does not represent accumulation of interstitial edema flui d. DIAGNOSIS As mentioned earlier, hypothyroidism should e accompanied y a decrease in the FTI and a decrease in the free T4 levels in plasma. Furthermore, primary hypothy roidism will e accompanied y an elevated serum TSH level (usually greater than 20 mU/L). A normal total serum T4 level will virtually exclude the diagnosis of hypothyroidism. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 14)03-May-05 08:05:54

Ovid: ICU Book THRYROID REPLACEMENT THERAPY The treatment for mild to moderate hypothyroidism is levothyroxine, which is giv en orally in a single daily dose of 50 to 200 g(6). P.775 The initial dose is usu ally 50 g/day, and this is increased in 50 g/day increments every 3 to 4 weeks. Th e optimal replacement dose of levothyroxine is determined y monitoring the seru m TSH level. The optimal dose is the lowest dose of levothyroxine that returns t he TSH to within the normal range (0.5 to 3.5 mU/L). In 90% of cases, this occur s with a levothyroxine dose of 100 to 200 g/day (6). Oral thyroxine therapy can a lso e effective in severe hypothyroidism, ut intravenous therapy is often reco mmended (at least initially) ecause of the risk of impaired gastrointestinal mo tility in severe hypothyroidism. One recommended regimen includes an initial int ravenous thyroxine dose of 250 g, followed on the next day y a dose of 100 g, and followed thereafter y a daily dose of 50 g (5). T3 Replacement Therapy Because the conversion of T4 to T3 (the active form of thyroid hormone) can e d epressed in critically ill patients (5), oral therapy with T3 can e used to sup plement thyroxine replacement therapy. In patients with depressed consciousness, oral T3 can e given in a dose of 25 g every 12 hours until the patient awakens (23). However, the enefits of T3 supplementation are unproven. REFERENCES REVIEWS: ADRENAL DYSFUNCTION 1. Knowlton AI. Adrenal insufficiency in the intensive care setting. J Intensive Care Med 1988;4:3545 (111 references). 2. Chin R. Adrenal crisis. Crit Care Clin 1991;7:2342 (58 references). 3. Oelkers W. Adrenal insufficiency. N Engl J Med 1996;335:12061212 (43 reference s). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 14)03-May-05 08:05:54

Ovid: ICU Book REVIEWS: THYROID DYSFUNCTION 4. Brent GA. The molecular asis of thyroid hormone action. N Engl J Med 1994;33 1:847853 (59 references). 5. Myers L, Hays J. Myxedema coma. Crit Care Clin 1991;7:4356 (48 references). 6. Toft AD. Thyroxine therapy. N Engl J Med 1994;331:174180 (74 references). 7. Reasner CA 2nd, Isley WL. Thyrotoxicosis in the critically ill. Crit Care Cli n 1991;7:57 74 (117 references). 8. Franklyn JA. The management of hyperthyroidism. N Engl J Med 1994;330:1731 173 8 (89 references). ADRENAL INSUFFICIENCY 9. Rothwell PM, Udwadia ZF, Lawler PG. Cortisol response to corticotrophin and s urvival in septic shock. Lancet 1991;337:582583. P.776 10. Moran JL, Chapman MJ, O'Fathartaigh MS, et al. Hypocortisolemia and adrenoco rtical responsiveness at onset of septic shock. Intensive Care Med 1994;20:489495 . 11. Soni A, Pepper GM, Wyrwinski PM, et al. Adrenal insufficiency occurring duri ng septic shock: incidence, outcome, and relationship to peripheral cytokine lev els. Am J Med 1995;98:266271. 12. Dorin RI, Kearns PJ. High output circulatory failure in acute adrenal insuff iciency. Crit Care Med 1988;16:296297. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 14)03-May-05 08:05:54

Ovid: ICU Book 13. Passmore JM Jr. Adrenal cortex. In: Geelhoed SW, Chernow B, eds. Endocrine a spects of acute illness. Clinics in critical care medicine. Volume 5. New York: Churchill Livingstone, 1985;97134. EVALUATION OF THYROID FUNCTION 14. Simons RJ, Simon JM, Demers LM, Santen RJ. Thyroid dysfunction in elderly ho spitalized patients. Arch Intern Med 1990;150:12491253. 15. Sumita S, Ujike Y, Namika A, et al. Suppression of the thyrotropin response to thyrotropin-releasing hormone and its association with severity of critical i llness. Crit Care Med 1994;22:16031609. 16. Isley WL. Thyroid dysfunction in the severely ill and elderly. Postgrad Med 1993;94:111128. 17. Surks MI, Chopra IJ, Mariash CN, et al. American Thyroid Association guideli nes for use of la oratory tests in thyroid disorders. JAMA 1990;263:15291532. HYPERTHYROIDISM 18. Trip MD, Wiersinga W, Plomp TA. Incidence, predicta ility, and pathogenesis of amiodarone-induced thyrotoxicosis and hypothyroidism. Am J Med 1991;91:507511. 19. Klein I. Thyroid hormone and the cardiovascular system. Am J Med 1990;88:6316 37. 20. Ehrmann DA, Sarne DH. Early identification of thyroid storm and myxedema com a. Crit Illness 1988;3:111118. HYPOTHYROIDISM 21. Ladenson PW. Recognition and management of cardiovascular disease related to thyroid dysfunction. Am J Med 1990;88:638641. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 14)03-May-05 08:05:54

Ovid: ICU Book 22. Lafayette RA, Costa ME, King AJ. Increased serum creatinine in the a sence o f renal failure in profound hypothyroidism. Am J Med 1994;96:298299. 23. McCulloch W, Price P, Hinds CJ, Wass JAH. Effects of low dose triiodothyroni ne in myxedema coma. Intensive Care Med 1985;11:259262. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 14)03-May-05 08:05:54

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Chapter 50 DISORDERS OF MENTATION A normal mental function is one of the most recogniza le signs of serious illnes s. In patients in the ICU, a normal mental function is associated with a higher mortality rate, a prolongation of mechanical ventilation, and a longer stay in t he ICU (1,2). This chapter focuses on two disorders of mental function that are common in critically ill patients: depressed consciousness and delirium. The fin al section of this chapter descri es the most severe disorder of mental function that will ever e encountered: rain death. MENTAL FUNCTION The mental aspects of nervous system function are responsi le for the manner in wh ich individuals interact with their environment. Mental function can e consider ed normal when all the following mental processes are intact: q Awareness of self and surroundings q Accurate perception of what is experienced (orientation) q A ility to process input data to generate more meaningful information (judgment and reasoning) q A ility to store and retrieve information (memory) The first mental process is known as consciousness, and the latter three mental processes make up what is known as cognition. The disorders of mental function c an therefore e classified as disorders of consciousness and disorders of cognit ion. DEPRESSED CONSCIOUSNESS Consciousness has two components: arousal (or wakefulness) and awareness (or res ponsiveness). Because awareness is not possi le without arousal, the tendency fo r arousal is the most important determinant P.780 of the level or degree of cons ciousness. The following are some definitions of level of consciousness ased on the tendency for arousal. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 17)03-May-05 07:56:34

Ovid: ICU Book LEVELS OF CONSCIOUSNESS Awakearoused and aware Somnolenteasily aroused and aware Stuporousaroused with diff iculty, impaired awareness Comatoseunarousa le and unaware Vegetative statearoused ut unaware Because coma is characterized y the a sence of arousal, it must al so e characterized y the a sence of awareness (responsiveness). Therefore, com a can e defined as a state of unarousa le unresponsiveness. Vegetative State A vegetative state is produced y diffuse cere ral injury without involvement of the rainstem reticular activating system (which determines arousal). In this c ondition, the patient is awake (eyes open) ut is unresponsive to ver al or noxi ous stimuli. A condition that can mimic the vegetative state is called the locke din syndrome. The patient with this condition is awake and aware ut una le to g enerate a motor response to ver al and noxious stimuli. Locked-in syndrome can e produced y destruction of the motor pathways in the ventral portion of the lo wer rainstem or y the induction of neuromuscular paralysis without adequate se dation. ETIOLOGIES The common causes of a depressed level of consciousness in patients who have not sustained a head injury are listed in Figure 50.1. Most of the conditions liste d in this figure can e classified as types of encephalopathies (i.e., infectiou s, ischemic, drug-related, or meta olic encephalopathies). These conditions are included in the simplified pneumonic SMASHED (3): http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 17)03-May-05 07:56:34

Ovid: ICU Book Figure 50.1. Possi le causes of a normal mental function (i.e., depressed consci ousness and delirium) in patients who have not sustained a head injury. SSu strate deficiencies (e.g., glucose, thiamine) MMeningoencephalitis or Mental i llness (e.g., malingering, psychogenic coma) AAlcohol or Accident (i.e., cere rov ascular accident) SSeizures HHypers (e.g., hypercapnia, hyperglycemia, hyperthyroi d, hyperthermia) or Hypos (e.g., hypoxia, hypotension, hypothyroid, hypothermia) EElectrolyte a normalities (e.g., hypernatremia, hyponatremia, hypercalcemia) an d Encephalopathies (e.g., hepatic, septic, uremic) P.781 DDrugs (e.g., opioid ana lgesics, sedativehypnotic agents). In one survey of neurologic complications in a medical ICU (1), ischemic stroke was the most common cause of depressed conscio usness in patients admitted to the ICU and septic encephalopathy was the most co mmon cause of depressed consciousness that developed after admission to the ICU. Septic Encephalopathy The condition known as septic encephalopathy applies to all infections other tha n primary infections of the central nervous system (4). Encephalopathy is report ed in 70% of patients in the ICU with sepsis and can appear as an early sign of sepsis (4). The underlying mechanism is unclear. Multiple, small microa scesses in oth cere ral hemispheres is the most consistent finding in fatal cases of se ptic encephalopathy (4), ut the incidence of these lesions in less advanced cas es is unknown. The amino acid alterations that are http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 17)03-May-05 07:56:34

Ovid: ICU Book characteristic of hepatic encephalopathy (i.e., increased aromatic amino acids a nd decreased ranched chain amino acids in plasma) are also found in septic ence phalopathy (5), ut the significance of this a normality is unclear. Septic ence phalopathy may e one feature of a more widespread multiorgan injury associated with the systemic inflammatory response (see Chapter 33). If this is the case, o xidant-induced cell injury may play a role in the pathogenesis of septic encepha lopathy. P.782 BEDSIDE EVALUATION The following features of the edside evaluation of depressed consciousness dese rve mention. Pupils The conditions that affect pupillary size and light reactivity are shown in Ta l e 50.1 (6,7,8 and 9). Dilated and unreactive pupils are usually signs of suprate ntorial (cere ral) injury. If the injury is due to diffuse cere ral ischemia, th e pupillary a normality is ilateral. If the injury is due to a mass lesion or c ere ral edema with herniation, the pupillary a normality will e unilateral (9). When atropine is given in the usual doses during cardiopulmonary resuscitation, the pupils will dilate ut they usually remain reactive (6). If the pupils rema in unreactive for longer than 6 hours after resuscitation from cardiac arrest, t he prognosis for neurologic recovery is very poor (8). TABLE 50.1. CONDITIONS THAT AFFECT PUPILLARY SIZE AND REACTIVITY Ocular Motility Spontaneous eye movements (conjugate or dysconjugate) are a nonspecific sign in comatose patients. However, a fixed gaze preference involving one or oth eyes i s highly suggestive of a mass lesion or seizures (9). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 17)03-May-05 07:56:34

Ovid: ICU Book Ocular Reflexes The ocular reflexes can e used to evaluate the functional integrity of the lowe r rainstem (9). These reflexes are illustrated in Figure 50.2. Figure 50.2. The ocular reflexes in the evaluation of coma. The oculocephalic re flex is demonstrated in the panels on the left, and the oculovesti ular reflex i s demonstrated in the panels on the right. The oculocephalic reflex is elicited y rotating the patient's head from side to side (slowly at first, then riskly). When the cere ral hemispheres are impaire d ut the lower rainstem is intact, the eyes will deviate away from the directi on of rotation (see the upper left panel in Figure 50.2). Because this eye movem ent resem les the fixed P.783 forward gaze of a doll's eyes, it is popularly kno wn as the doll's-eyes movement (9). When the lower rainstem is damaged (or when the patient is awake), the eyes will follow the direction of head rotation (see the lower left panel in Figure 50.2). The oculocephalic reflex should not e at tempted in patients with cervical arthritis or a suspected injury involving the cervical spine. The oculovesti ular reflex is elicited y injecting 30 mL of col d saline or tap water in the external auditory canal (using a 50-mL syringe and a 2-inch plastic catheter). When the rainstem is functionally intact, oth eyes will deviate toward the irrigated ear. This conjugate eye movement is lost when the lower rainstem is http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 17)03-May-05 07:56:34

Ovid: ICU Book damaged. The Extremities Clonic movements elicited y flexing the patient's hands or feet (asterixis) is a sign of a diffuse meta olic encephalopathy (4). A focal motor or sensory defec t in the extremities (e.g., hemiparesis or asymmetric reflexes) can also e asso ciated with a diffuse meta olic encephalopathy. However, the presence of focal n eurologic defects should always prompt further investigation (with computed tomo graphy) for a structural rain lesion. P.784 THE GLASGOW COMA SCALE The severity of a depressed level of consciousness is often evaluated with the G lasgow Coma Scale, which is shown in Ta le 50.2. Although this scale was origina lly proposed for patients with head injuries (10), it has gained widespread acce ptance for use in the evaluation of nontraumatic coma as well (11,12,13,14 and 1 5). In the initial evaluation of nontraumatic coma, patients with a Glasgow Coma score of 6 or higher (out of a maximum score of 15) are seven times more likely to awaken than patients with a score of 5 or lower (11). TABLE 50.2. THE GLASGOW COMA SCALE Limitations The Glasgow Coma Scale is ased on three indicators of cere ral function: eye op ening, ver al http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 17)03-May-05 07:56:34

Ovid: ICU Book communication, and motor responses to ver al and noxious stimuli. The ver al com munication component is a pro lem in patients in the ICU who are intu ated ecau se these patients are una le to communicate ver ally (13,14). The predictive val ue of the other components of the Glasgow Coma Scale are shown in Ta le 50.3. In this case, the predictive value refers to the chances of a satisfactory neurolo gic recovery in patients who remain comatose after resuscitation from cardiac ar rest (15). In patients who show no response to ver al or noxious stimuli 1 hour after the cardiac arrest, 70 to 80 % will not have a satisfactory neurologic rec overy. If these deficits persist for more than 3 days after the cardiac arrest, the chances of a satisfactory neurologic recovery are practically nil. Therefore , the P.785 Glasgow Coma Scale can still provide prognostic information in patie nts who are intu ated and una le to ver alize. TABLE 50.3. THE CLINICAL FEATURES OF DELIRIUM The data in Ta le 50.3 also demonstrates that the Glasgow Coma Scale does not re ach its full predictive power in the first few hours after cardiac arrest. There fore, the Glasgow Coma Scale should not e used to predict the chances for neuro logic recovery in the first few hours after cardiac arrest. DELIRIUM Delirium is the most common mental disorder in hospitalized elderly patients (16 ,17,18 and 19), and it is the most common postoperative complication in the elde rly (20). Surveys of hospitalized patients have found delirium in up to 50% of t he elderly patients (19). Unfortunately, as many as three-fourths of the cases o f delirium will e missed y physicians and nurses caring for the patients (16). CLINICAL FEATURES The clinical features of delirium are summarized in Ta le 50.4 (20). Delirium is a cognitive disorder that is characterized y attention deficits, disordered th inking, and a fluctuating symptomatology. The hallmark of delirium (and the feat ures that distinguish delirium from dementia) is its acute onset and fluctuating clinical http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 17)03-May-05 07:56:34

Ovid: ICU Book course. TABLE 50.4. THE CLINICAL FEATURES OF DELIRIUM Hypoactive Delirium There is a tendency to consider delirium as a state of agitation (as in the deli rium tremens syndrome). However, as shown in Ta le 50.4, there is also a hypoact ive form of delirium that is characterized y lethargy rather than agitation. In fact, hypoactive delirium (with lethargy) is the most common form of delirium i n the elderly (20), and this is certainly a source of many missed diagnoses of d elirium in elderly patients. P.786 Delirium versus Dementia Delirium and dementia are distinct mental disorders that are easily confused ec ause they have overlapping clinical features (e.g., attention deficits and disor dered thinking) (21). As mentioned earlier, the principal features of delirium t hat distinguish it from dementia are its acute onset and fluctuating course. How ever, as many as three-fourths of cases of delirium are superimposed on an under lying dementia (20), so the diagnosis of delirium in any individual patient does not exclude the presence of dementia. ETIOLOGIES The possi le causes of delirium are listed in Figure 50.1. Any type encephalopat hy (i.e., infectious, ischemic, drug-related, or meta olic) can cause a state of delirium. Drugs are implicated as causative or contri utory factors in up to 40 % of cases of delirium in the elderly (17,18 and 19). The drugs that are most li kely to e responsi le for delirium in patients in the ICU are listed in Ta le 5 0.5. The principal offenders in this list are alcohol (withdrawal), long-acting enzodiazepines (diazepam and flurazepam), and opiates (particularly http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 17)03-May-05 07:56:34

Ovid: ICU Book meperidine) (17,18 and 19). TABLE 50.5. DRUGS THAT CAN CAUSE DELIRIUM* MANAGEMENT The management of delirium should focus on identifying and treating the underlyi ng cause of the pro lem. If agitation and disruptive ehavior ecome a pro lem, the administration of sedatives is advised. P.787 The choice of sedative is dete rmined y the cause of the delirium (22,23,24 and 25), as indicated in the follo wing. ICU and Postoperative Delirium For delirium that is ICU-related or develops postoperatively, the sedative of ch oice is haloperidol (22). (See Ta le 8.6 for haloperidol dosing recommendations. ) The only draw ack with intravenous haloperidol is its delayed onset of action (approximately 10 minutes). If rapid sedation is desired, a enzodiazepine (e.g. , midazolam) can e used in the initial stages of management. However, enzodiaz epines should e avoided for continued management ecause they can aggravate ICU -related and postoperative delirium (22). Delirium Tremens The management of delirium that accompanies alcohol withdrawal is the opposite o f the management for ICUrelated delirium. In this situation, the preferred sedat ives are enzodiazepines, whereas haloperidol can aggravate the delirium and low er the seizure threshold (24,25). (See Ta le 8.3 for enzodiazepine dosing recom mendations.) If trou lesome hypertension is a feature of delirium tremens, adjun ctive therapy with clonidine (a centrallyacting 2- gonist) c n help both decre se the blood pressure nd dd to the sed tion produced by benzodi zepines (24,25). The dose of clonidine is 0.1 mg or lly every 2 hours until the pressure is cont rolled http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 17)03-M y-05 07:56:34

Ovid: ICU Book or cumul tive dose of 0.5 mg h s been given (26). Clonidine is lso v il ble in tr nsderm l p tch, but this is not recommended bec use of the del yed onset of ction. P.788 Coc ine-Induced Delirium The m n gement of coc ine-induced delirium is essenti lly the s me s h s been d escribed for delirium tremens: benzodi zepines re the preferred sed tives, nd h loperidol is not dvised (23). BRAIN DEATH Br in de th is condition ch r cterized by irreversible cess tion of life-suppo rting function in the centr l nervous system (27,28). This condition is most oft en the result of severe intr cr ni l injury or m ssive intr cerebr l hemorrh ge or inf rction. It is not common consequence of the conditions listed in Figure 50.1. DIAGNOSIS A checklist for the di gnosis of br in de th is shown in T ble 50.6. E ch of the criteri on this checklist should be s tisfied on two sep r te occ sions. The p rincip l fe ture of br in de th is the irreversible cess tion of ll purposeful electric l ctivity in the br in, including ctivity in the lower br instem resp ir tory centers. This cess tion of ctivity should h ve no contribution from sed ting drugs or hypothermi . http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 17)03-M y-05 07:56:34

Ovid: ICU Book TABLE 50.6. CHECKLIST FOR THE DIAGNOSIS OF BRAIN DEATH Apne Testing The h llm rk of the br in de th determin tion is the demonstr tion of persistent pne despite the presence of hyperc pnic stimulus to ventil tion. To perform this pne test, the p tient is sep r ted from the ventil tor to llow the rte ri l PCO2 to rise t le st 10 mm Hg bove the norm l or b seline levels (the b s eline rteri l PCO2 should not be below norm l before this test). To prevent lif e-thre tening hypoxemi during the pne test, the test is preceded by 10 to 1 5 minute period of ventil tion with 100% oxygen (29,30). In ddition, 100% oxyge n c n be insuffl ted into the tr che t r te of 15 L/minute during the period of pne (30). After the p tient h s been sep r ted from the ventil tor for p redetermined period (usu lly 10 minutes), n rteri l blood g s me surement is o bt ined to document the desired incre se in rteri l PCO2. If pne persists des pite the desired incre se in rteri l PCO2, the test is confirm tory for the http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 17)03-M y-05 07:56:34

Ovid: ICU Book di gnosis of br in de th. ORGAN DONATION If p tient with suspected or documented br in de th is c ndid te for org n d on tion, the following me sure c n be used to preserve org n vi bility during n d fter the br in de th determin tion (31). P.789 A S fer Apne Test To prevent d m ging hypoxemi during pne testing, n ltern tive method is v il ble in which the rteri l PCO2 is llowed to rise while the p tient continues to receive ssisted ventil tion (32). This ltern tive method involves mech nic l ventil tion with 100% oxygen using intermittent m nd tory ventil tion (IMV). Hypoventil tion is then induced by decre sing the number of m chine bre ths deli vered by the ventil tor, which llows the rteri l PCO2 to rise. This method is demonstr ted in Figure 50.3. The tot l ventil tion delivered by the ventil tor i s bruptly decre sed from 11 to 1 L/minute, nd the rise in rteri l PCO2 is mon itored noninv sively with n end-tid l PCO2 monitor. Note th t the rteri l oxyg en s tur tion (S O2), which P.790 is monitored continuously with pulse oximete r, is m int ined bove 90% during the entire period of progressive hyperc pni ( bec use the p tient continues to receive some ssisted ventil tion). When the r teri l PCO2 rises to the desired level, the p tient c n then be sep r ted from t he ventil tor to observe for persistent pne .

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 17)03-M y-05 07:56:34

Ovid: ICU Book Figure 50.3. A s fer pne test th t produces progressive hyperc pni while the p tient continues to receive ssisted ventil tion. The step decre se in minute v entil tion (from 11 to 1 L/minute) is produced by reducing the number of m chine bre ths delivered during intermittent m nd tory ventil tion (IMV). The rise in rteri l PCO2 is monitored noninv sively using n end-tid l CO2 monitor (PETCO2) , nd rteri l oxygen tion (S O2) is monitored continuously using pulse oximet er. (From Gutm nn DH, M rino PL. An ltern tive pne test for the ev lu tion of br in de th. Ann Neurol 1991;30:852853.)

Hemodyn mics Hypotension or n imp ired c rdi c output in potenti l org n donors should be co rrected by infusing dobut mine (for positive inotropic effects) or dop mine (for positive inotropic nd v soconstrictor effects) s described in Ch pter 18. If pulmon ry rtery c theter is in pl ce, the go l of hemodyn mic m n gement shou ld be systemic oxygen upt ke bove 100 mL/min/m2 nd serum l ct te concentr tion below 4 mmol/ L (see Ch pter 13). P.791

Pituit ry F ilure More th n h lf of p tients with br in de th will develop pituit ry f ilure with di betes insipidus nd second ry dren l insufficiency (31). Both conditions c n le d to profound hypovolemi ( nd reduced org n perfusion) nd hypertonic hyper n tremi ( nd cell dehydr tion). If there is evidence of centr l di betes insi pidus (i.e., spont neous diuresis with urine osmol lity below 200 mOsm/L), tre tment with desmopressin ( form of v sopressin th t does not c use v soconstric tion) is dvised (31). If there is evidence of dren l insufficiency

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 17)03-M y-05 07:56:34

Ovid: ICU Book (i.e., hypotension refr ctory to v sopressors), tre tment with hydrocortisone (s ee dos ges in Ch pter 49) is dvised. REFERENCES INTRODUCTION 1. Bleck TP, Smith MC, Pierre-Louis J-C, et l. Neurologic complic tions of crit ic l medic l illnesses. Crit C re Med 1993;21:98103. 2. Kelly BJ, M tth y MA. Prev lence nd severity of neurologic dysfunction in cr itic lly ill p tients. Chest 1993;104:18181824. DEPRESSED CONSCIOUSNESS 3. Roberts JR, W son S, Soegel E. Pneumonic for di gnosis of cute ment l st tus ch nges. Ann Emerg Med 1990;19:221222. 4. M hler J, Young GB. Septic enceph lop thy. Intensive C re Med 1993;38:177187.

6. Goetting MG, Contrer s E. Systemic tropine dministr tion during c rdi c rr est does not c use fixed nd dil ted pupils. Ann Emerg Med 1991;20:5557. 7. Ong GL, Bruning HA. Dil ted fixed pupils due to dministr tion of high doses of dop mine hydrochloride. Crit C re Med 1981;9:658660. 8. Steen-H nsen JE, H nsen NN, V genes P, Schreiner B. Pupil size nd light re ctivity during c rdiopulmon ry resuscit tion: clinic l study. Crit C re Med 19 88;16:6970.

GLASGOW COMA SCALE http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 17)03-M y-05 07:56:34

9. Weisberg LA. Differenti l di gnosis of com : Illness 1989;4:97108.

step-by-step str tegy. J Crit

5. Sprung CL, Cerr FB, Freund HR, et l. Amino cid lter tions nd enceph lop thy in the sepsis syndrome. Crit C re Med 1991;19:753757.

Ovid: ICU Book

11. S cco RL, V nGool R, Mohr JP, H user WA. Nontr um tic com : Gl sgow Com Sco re nd com etiology s predictors of 2-week outcome. Arch Neurol 1990;47:1181118 4.

13. Seg tore M, W y C. The Gl sgow Com Sc le: time for ch nge. He rt Lung 1992; 21:548557.

15. Edgren E, Hedstr nd U, Kelsey S, et l. Assessment of neurologic prognosis i n com tose survivors of c rdi c rrest. L ncet 1994;343:10551059. DELIRIUM 16. Inouye SK. The dilemm of delirium: clinic l nd rese rch controversies reg rding di gnosis nd ev lu tion of delirium in hospit lized elderly medic l p tie nts. Ann Intern Med 1994;97:278288. 17. Rumm ns TA, Ev ns JM, Kr hn LE, Fleming KC. Delirium in elderly p tients: ev lu tion nd m n gement. M yo Clin Proc 1995;70:989998.

18. Fr ncis J, M rtin D, K poor WN. A prospective study of delirium in the hospi t lized elderly. JAMA 1990;263:10971101. 19. M rc ntonio ER, Ju rez G, Goldm n L, et l. The rel tionship of postoper tiv e delirium with psycho ctive medic tions. JAMA 1994;272:15181522. 20. Inouye SK, v n Dyck CH, Alessi CA, et l. Cl rifying confusion: the confusio n ssessment method: new method for detection of delirium. Ann Intern Med 1990 ;113:941948. 21. Fleming KC, Ad ms AC, Petersen RC. Dementi : di gnosis nd ev lu tion. M yo Clin Proc 1995;70:10931107. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 17)03-M y-05 07:56:34

14. M rion D. The Gl sgow Com Sc le score: contempor ry C re World 1994;11:101102.

pplic tion. Intensive

12. H mel MB, Goldm n L, Teno J, et l. Identific tion of com tose p tients igh risk for de th or severe dis bility. JAMA 1995;273:18421848. P.792

10. Te sd le G, Jennett B. Assessment of com 1974;2:8184.

nd imp ired consciousness. L ncet

t h

Ovid: ICU Book 22. Weissm n C. Str tegies for m n ging delirium in critic lly ill p tients. J C rit Illness 1996;11:295307. 23. Hoffm n RS. An effective str tegy for m n ging coc ine-induced git ted deli rium. J Crit Illness 1994;9:139149.

25. Crippen DW. Str tegies for m n ging delirium tremens in the ICU. J Crit Illn ess 1997;12:140149. 26. G les MA. Or l ntihypertensives for hypertensive urgencies. Ann Ph rm cothe r 1994;28:352357. BRAIN DEATH 27. Curry PD, Bion JF. The di gnosis nd m n gement of br in de th. Curr An esth Crit C re 1994;5:3640. 28. Wijdicks EFM. Di gnosis nd m n gement of br in de th in the intensive c re unit. In: Neurology of critic l illness. Phil delphi : FA D vis, 1995;323337.

30. M rks SJ, Zisfein J. Apneic oxygen tion in pne tests for br in de th. Arch Neurol 1990;47:1066 1068.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 17)03-M y-05 07:56:34

32. Gutm nn D, M rino PL. An ltern tive pne de th. Ann Neurol 1991;30:852853. P.793

test for the ev lu tion of br in

31. Detterbeck FC, Mill MR, Willi ms W, Eg n TM. Org n don tion nt of the multiple org n donor. Contemp Surg 1993;42:281285.

nd the m n geme

29. Benzel EC, Gross CD, H dden TA, et l. The of br in de th. J Neurosurg 1989;71:191194.

pne test for the determin tion

DELIRIUM TREMENS 24. Lohr RH. Tre tment of Proc 1995;70:777782.

lcohol withdr w l in hospit lized p tients. M yo Clin

Ovid: ICU Book SUGGESTED READINGS Grott JC, ed. M n gement of the cutely ill neurologic p tient. New York: Churc hill Livingstone, 1993. Plum F, Posner JB. The di gnosis of stupor nd com . 3d ed. Phil delphi : FA D v is, 1982. Wijdicks EFM. Neurology of critic l illness. Phil delphi : FA D vis, 1995. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (17 of 17)03-M y-05 07:56:34

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 51 DISORDERS OF MOVEMENT This ch pter focuses on three gener l types of movement disorders encountered in critic lly ill p tients: involunt ry movements (seizures), we k or ineffective movements (neuromuscul r we kness), nd no movements (neuromuscul r block de). SEIZURES Seizures re second only to met bolic enceph lop thies s the most common neurol ogic complic tion encountered in p tients in the ICU (1). DEFINITIONS The following definitions will prove helpful in the description, ev lu tion, nd tre tment of seizures (2,3). Types of Movement Seizures c n be ccomp nied by ny of the following p tterns of muscle contr cti on: tonic contr ctions (sust ined muscle contr ction), tonic contr ctions ( bse nce of muscle contr ction), clonic contr ctions (periodic muscle contr ctions wi th regul r mplitude nd frequency), or myoclonus (periodic muscle contr ction s with n irregul r mplitude nd frequency). Seizures m y lso be ccomp nied b y f mili r movements c lled utom tisms (e. g., lip sm cking or chewing). Gener lized Seizures Gener lized seizures rise from symmetric nd synchronous electric l disch rges involving the entire cerebr l cortex. These seizures m y or m y not be ccomp ni ed by muscle contr ctions. Gener lized P.795 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 15)03-M y-05 07:56:59

Ovid: ICU Book

P rti l Seizures P rti l seizures rise from electric l disch rges th t re confined to foc l o r restricted p rt of the cerebr l cortex. They re subdivided into simple p rti l seizures (do not imp ir consciousness) nd complex p rti l seizures (c use imp ired consciousness). Two types of p rti l complex seizures th t deserve mention re tempor l lobe seizures (ch r cterized by motionless st re nd utom tisms ) nd epilepsi p rti lis continu (ch r cterized by persistent tonic-clonic mov ements of the f ci l muscles nd limb muscles on one side of the body).

St tus Epilepticus St tus epilepticus is defined s more th n 30 minutes of continuous seizure cti vity or two or more sequenti l seizures without n intervening period of conscio usness. ETIOLOGIES New-onset seizures c n be the result of drug intoxic tion (e.g., theophylline) , drug withdr w l (e.g., eth nol), infections (e.g., meningoenceph litis, bsces s), ischemic injury (e. g., foc l or diffuse), sp ce-occupying lesions (e.g., tu mor or hemorrh ge), or met bolic der ngements (e.g., hep tic or uremic enceph lo p thy, hypoglycemi , hypon tremi , hypoc lcemi ). In one survey of new-onset sei zures in p tients in the ICU, the most common c uses were drug intoxic tions nd drug withdr w l (4). The drugs most likely to be implic ted s c use of seizu res in p tients in the ICU re listed in T ble 51.1 (5). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 15)03-M y-05 07:56:59

seizures th t re not ccomp nied by prominent muscle contr ctions sence seizures (formerly known s petit-m l seizures).

re c lled b

Ovid: ICU Book TABLE 51.1. DRUG-RELATED SEIZURES IN THE ICU P.796 St tus Epilepticus Approxim tely 50% of c ses of st tus epilepticus occur s new-onset seizures. Th e most common c use is cute cerebrov scul r injury (ischemi , hemorrh ge). Hypo glycemi nd centr l nervous system infections re uncommon offenders. Ev lu tion The ev lu tion of new-onset seizures should focus on the etiologies previously m entioned. In the bsence of n obvious met bolic or drug-rel ted c use or when t he physic l ex min tion reve ls foc l bnorm lity, further ev lu tion (with ne uroim ging studies nd lumb r puncture) is dvised. MANAGEMENT Most seizures encountered in the ICU re gener lized or p rti l seizures ssoci ted with tonicclonic movements (i.e., convulsive seizures). The cute m n gement of convulsive seizures is summ rized in T ble 51.2, which uses the recommend ti ons for st tus epilepticus (6). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 15)03-M y-05 07:56:59

Ovid: ICU Book TABLE 51.2. DRUG TREATMENT OF CONVULSIVE STATUS EPILEPTICUS Benzod zepines Intr venous benzodi zepines re the initi l drugs of choice for suppression of c onvulsive seizures. Intr venous di zep m (V lium) in dose of 0.2 mg/kg will te rmin te seizure ctivity within 5 minutes in pproxim tely 80% of c ses, but the effect l sts only 30 minutes (6). P.797 Intr venous lor zep m (Ativ n) in dos e of 0.1 mg/kg is s effective s di zep m nd h s longer dur tion of ction ( 12 to 24 hours). Phenytoin Bec use of its short dur tion of ction, intr venous di zep m should lw ys be f ollowed by phenytoin to prevent recurrence of seizures. The st nd rd intr venous dose of phenytoin is 20 mg/kg in dults, but sm ller dose of 15 mg/kg is reco mmended for elderly p tients (6). A m ximum infusion r te of 50 mg/minute is dv ised to reduce the risk of c rdiov scul r depression from phenytoin (which m y b e rel ted to the propylene glycol solvent used in intr venous phenytoin prep r t ions) (7). If the initi l dose of phenytoin is unsuccessful, ddition l doses of 5 mg/kg c n be given to tot l cumul tive dose of 30 mg/kg (6). The ther peuti c serum level for phenytoin is 10 to 20 mg/L. In p tients with history of nti convuls nt hypersensitivity syndrome (fever, r sh, nd lymph denop thy in respon se to phenytoin dministr tion), st tus epilepticus c n be m n ged http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 15)03-M y-05 07:56:59

Ovid: ICU Book with continuous intr venous infusion of di zep m using the infusion r te shown in T ble 51.2 (8). Phenob rbit l The combin tion of benzodi zepines nd phenytoin will control seizures in 90% of c ses of st tus epilepticus. In the rem ining 10% of c ses, intr venous phenob rbit l will chieve control in pproxim tely h lf the c ses (6). Phenob rbit l i s given in dose of 100 mg/minute until the seizures re controlled or m ximu m dose of 20 mg/kg is chieved. The ther peutic serum level for phenob rbit l is 10 to 40 mg/L. The higher dos ge r nge will produce profound nd prolonged sed tion, so the p tient m y not w ken for sever l hours fter the seizures re sup pressed. Approxim tely 5% of p tients with st tus epilepticus will be refr ctory to benzodi zepines, phenytoin, nd phenob rbit l. These refr ctory c ses m y re quire inh l tion l nesthesi nd neuromuscul r p r lysis (9). At this st ge, neurologic consult tion should be obt ined immedi tely. NEUROMUSCULAR WEAKNESS The following is brief description of three neuromuscul r disorders th t c n p roduce severe nd life-thre tening neuromuscul r we kness. Some of the comp r ti ve fe tures of these disorders re included in T ble 51.3. TABLE 51.3. COMPARATIVE FEATURES OF THREE NEUROMUSCULAR DISORDERS http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 15)03-M y-05 07:56:59

Ovid: ICU Book MYASTHENIA GRAVIS My stheni gr vis is n utoimmune dise se ch r cterized by ntibody-medi ted de struction of postsyn ptic cetylcholine receptors t P.798 neuromuscul r junctio ns. The illness is uncommon nd ffects pproxim tely 1 in every 100,000 dults (10). Clinic l Fe tures The we kness in my stheni gr vis ch r cteristic lly becomes worse with repe ted ctivity. Signs of we kness re usu lly first evident in the eyelids nd extr o cul r muscles, nd gener lized we kness follows in 85% of c ses (10). The proxim l limb muscles re often ffected, nd we kness c n involve the di phr gm nd t hor cic muscul ture. R pid progression to respir tory f ilure nd ventil tor dep endence, c lled my sthenic crisis, c n occur. The deficit in my stheni is purel y motor, with no sensory involvement. Deep tendon reflexes re usu lly preserved . In ddition to concurrent illness nd surgery, sever l drugs c n precipit te o r ggr v te the my sthenic syndrome. The princip l offenders re ntibiotics (e. g., minoglycosides), c rdi c drugs (e.g., lidoc ine, proc in mide, quinidine), nd m gnesium (11). Aminoglycosides c n precipit te my sthenic syndrome th t r esolves when the drugs re discontinued. M gnesium blocks the presyn ptic rele s e of cetylcholine nd c n be p rticul rly detriment l in my sthenic p tients. Di gnosis The di gnosis of my stheni gr vis is b sed on the ch r cteristic p ttern of mus cle we kness (e.g., eyelid or extr ocul r muscle we kness, worse with ctivity) nd the finding of enh nced muscle strength fter the dministr tion of n ntic holinester se inhibitor such s edrophonium (Tensilon). A r dioimmuno ss y for cetylcholine receptor ntibodies in the blood is positive in 85% of c ses (10) nd confirms the di gnosis. Once the di gnosis is confirmed, se rch for ssoci ted conditions such s thymic tumors (10% of c ses) nd hyperthyroidism (5% of c ses) is dvised. P.799 Tre tment http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 15)03-M y-05 07:56:59

Ovid: ICU Book Anticholinester se gents re the first line of ther py in my stheni gr vis. Py ridostigmine (Mestinon) is st rted t dos ge of 60 mg or lly (or 2 mg intr mus cul rly) every 8 hours nd c n be incre sed if necess ry to m ximum dos ge of 120 mg or lly (or 4 mg intr muscul rly) every 8 hours. Immunother py is dded if needed with prednisone (20 to 60 mg/d y), z thioprine (2 to 3 mg/kg/d y), or c yclosporine (5 mg/kg/d y) (10). In dv nced c ses requiring mech nic l ventil ti on, pl sm pheresis (to remove cetylcholine ntibodies) is often effective in pr oducing short-term improvement. Surgic l thymectomy is often dvised in p tients under 60 ye rs of ge to reduce the need for immunosuppressive ther py (10). GUILLAIN-BARR SYNDROME Guill in-B rr syndrome is n cute infl mm tory demyelin ting polyneurop thy th t is preceded by n cute infectious illness in two-thirds of c ses (12,13). The disorder is considered to be immune in origin, but the precise mech nism rem ins uncle r. Clinic l Fe tures This condition presents with p resthesi s nd symmetric limb we kness th t evolv es over period of few d ys to few weeks. Symptoms of preceding infection usu lly subside before the we kness becomes pp rent. Approxim tely 20% of p ti ents progress to respir tory f ilure th t requires mech nic l ventil tion (12,13 ). Adv nced c ses c n lso be ch r cterized by utonomic inst bility nd bulb r p r lysis. The condition resolves spont neously in pproxim tely 85% of c ses, b ut residu l neurologic deficits re common (12). Di gnosis The di gnosis is b sed on the clinic l picture (i.e., progressive, symmetric lim b we kness following n cute infectious illness). Axon l degener tion c n be do cumented by nerve conduction studies if necess ry. Tre tment The tre tment of Guill in-B rr syndrome mostly involves supportive c re. In sever e c ses requiring mech nic l ventil tion, both pl sm pheresis nd intr venous im mune globulin (0.4 g/ kg/d y) c n produce tr nsient improvement (14), but the tr e tment effects re much less m rked th n in my stheni gr vis. Bec use this dis order resolves spont neously in most c ses, c reful ttention to complic tions ( e.g., respir tory f ilure, thromboembolism) is m nd tory for s tisf ctory P.80 0 http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 15)03-M y-05 07:56:59

Ovid: ICU Book outcome. The respir tory m n gement of this disorder is described l ter in the c h pter. CRITICAL ILLNESS POLYNEUROPATHY In p tients with systemic sepsis nd multiorg n dysfunction (see Ch pter 31), on e of the org ns th t c n be ffected is the peripher l nervous system. The resul t is diffuse peripher l neurop thy c lled critic l illness polyneurop thy (15) . Clinic l Fe tures As m ny s 70% of p tients with systemic sepsis nd multiorg n dysfunction will h ve evidence of combined motor nd sensory polyneurop thy on nerve conduction studies (16). This disorder is clinic lly silent in most c ses, but it c n prod uce severe limb nd trunc l we kness on occ sion. Most c ses occur in p tients w ho h ve been ventil tor-dependent for t le st 1 week (17). The we kness usu lly becomes evident when the sepsis begins to resolve, nd it c n del y we ning fro m mech nic l ventil tion (17). M ny of the p tients h ve received neuromuscul r blocking gents, nd the condition presents s persistent we kness fter neuromu scul r p r lysis is discontinued. Autonomic dysfunction nd bulb r we kness re uncommon, which helps distinguish this condition from Guill in-B rr syndrome (17) . The di gnosis of critic l illness polyneurop thy is suggested by the clinic l setting (i.e., limb nd trunc l we kness ssoci ted with severe sepsis nd multi org n dysfunction) nd by the exclusion of other neuromuscul r disorders. Nerve conduction studies or electromyogr phy c n performed to document xon l degener tion (17), but these studies should not be necess ry in most c ses. No specific tre tment exists, but spont neous recovery occurs in m ny c ses. PULMONARY COMPLICATIONS The pulmon ry consequences of progressive neuromuscul r we kness re summ rized in T ble 51.4. Respir tory muscle strength must P.801 decre se consider bly befo re pulmon ry complic tions ppe r. The e rliest complic tion is depressed coug h with difficulty cle ring secretions. This c n le d to ret ined secretions with infection nd irw ys obstruction. When p tients re un ble to cle r secretions dequ tely, tr che l intub tion is indic ted to help cle r secretions. As the n euromuscul r we kness progresses, telect sis nd hypoxemi become prominent, fo llowed by lveol r hypoventil tion nd progressive CO2 retention. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 15)03-M y-05 07:56:59

Ovid: ICU Book TABLE 51.4. Respir tory Consequences of Neuromuscul r We kness One of the import nt fe tures to note in T ble 51.4 is the f ct th t tr che l in tub tion is often necess ry before the ppe r nce of progressive hypoxemi nd h yperc pni . In other words, in neuromuscul r disorders th t involve the respir t ory muscles, e rly intub tion before the onset of respir tory f ilure is often n ecess ry (13,18) The most sensitive me sure of respir tory muscle strength is th e m ximum inspir tory pressure (PIm x) (19), which is described in Ch pter 29. A PIm x th t f lls below 30 cm H2O is evidence of severe respir tory muscle we kn ess nd is n indic tion for tr che l intub tion nd ssisted ventil tion. NEUROMUSCULAR BLOCKADE Drug-induced neuromuscul r block de is common pr ctice in ICUs for m n ging ve ntil tordependent p tients who re git ted nd difficult to ventil te. However, this pr ctice h s serious dr wb cks (20), s described l ter. MECHANISMS Neuromuscul r blocking gents ct by binding to cetylcholine receptors on the p ostsyn ptic side of the neuromuscul r junction. Once bound, there re two differ ent modes of ction. The depol rizing gents ct like cetylcholine nd depol ri ze the postsyn ptic membr ne. However, they produce sust ined depol riz tion, which blocks the prop g tion of electric l impulses long the muscle. The nondep ol rizing gents ct by blocking cetylcholine-induced ctiv tion of the postsyn ptic membr ne. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 15)03-M y-05 07:56:59

Ovid: ICU Book NEUROMUSCULAR BLOCKERS The neuromuscul r blocking drugs th t re used most commonly in the ICU re show n in T ble 51.5 (21). All three gents in this t ble re nondepol rizing blocker s. TABLE 51.5. COMPARATIVE FEATURES OF NEUROMUSCULAR BLOCKERS P ncuronium (P vulon) is rel tively long- cting neuromuscul r blocker (dur tio n of ction is 1 to 2 hours) th t w s introduced for clinic l use in 1972. The i niti l popul rity of this gent h s dwindled bec use of its long dur tion of ct ion (c n ccumul te with repe ted use) nd v golytic effects (which c n produce t chyc rdi nd hypertension). P.802 Although p ncuronium c n be given by contin uous infusion, it is usu lly given s intermittent bolus doses bec use of the ri sk of drug ccumul tion. P ncuronium is excreted by the kidneys, nd dos ge redu ctions re often necess ry in ren l f ilure. Vecuronium (Norcuron) is chemic l n log of p ncuronium but is shorter cting nd h s no v golytic effects. The d osing of vecuronium is simil r to th t for p ncuronium. Bec use of its short dur tion of ction (30 minutes), it is often given by continuous infusion. Atr curi um (Tr crium) is lso shorter cting th n p ncuronium nd c n be given by contin uous infusion. This gent c n rele se hist mine from m st cells nd produce hypo tension when given too r pidly. For this re son, bolus injections should be give n over 2 to 4 minutes. Succinylcholine (Anectine) is depol rizing gent th t i s not included in T ble 51.5 bec use it is used rel tively infrequently. This g ent is ultr -short cting nd is most often used to http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 15)03-M y-05 07:56:59

Ovid: ICU Book f cilit te tr che l intub tion. An intr venous dose of 1 to 2 mg/kg produces p r lysis within 2 minutes, nd the effect dis ppe rs within 10 minutes. The prolon ged depol riz tion produced by this gent is ccomp nied by l rge efflux of po t ssium out of muscle cells, nd this results in tr nsient rise in serum pot s sium of 0.5 to 1.0 mEq/L (21). Life-thre tening incre ses in serum pot ssium h v e been recorded in p tients with cute neurologic injury nd burns.

MONITORING The st nd rd method of monitoring the dequ cy of neuromuscul r block de involve s stimul ting the uln r nerve t the wrist to produce dduction of the thumb. Th e stimul ting electric l current is sm ll (50 to 80 milli mps) nd it is imperce ptible to the p tient. Bec use nondepol rizing neuromuscul r block de is ex gger ted by repe ted nerve stimul tion, series of repetitive stimuli c n be pplie d to the uln r nerve nd the movement of the thumb c n be observed for evidence of progressive decrement in response. The st nd rd method is to pply series of 4 stimuli in succession nd comp re the thumb P.803 movement in response to the first nd fourth stimuli. This method is c lled the tr in-of-four stimul tio n technique (23). Tot l bsence of thumb movement is evidence for excessive neur omuscul r block de, where s bsence of decrement in thumb contr ctions with re petitive nerve stimul tion is evidence of in dequ te block de.

DISADVANTAGES In dequ te Sed tion Neuromuscul r blocking drugs do not produce sed tion, nd bec use p r lysis is n extremely frightening nd even p inful experience, he vy sed tion is m nd tory during neuromuscul r block de. Unfortun tely, it is impossible to determine whe ther sed tion is dequ te in p tient who is p r lyzed, so generous doses of se d tive gents is dvised. Bec use immobiliz tion is often p inful experience, opi tes m y be preferred for sed tion. Prolonged P r lysis A number of reports h ve documented prolonged neuromuscul r p r lysis fter disc ontinuing ther py with neuromuscul r blocking gents (24). This phenomenon often occurs fter prolonged neuromuscul r block de (longer th n 1 week), but it c n lso occur fter just few doses of neuromuscul r blockers. There is no evidenc e th t this problem is rel ted to the drug th t is used. Shorter periods nd int ermittent cess tion of neuromuscul r block de h ve been http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 15)03-M y-05 07:56:59

Ovid: ICU Book recommended to limit the risk of prolonged p r lysis (24). However, the best pre ventive me sure is to elimin te the use of neuromuscul r block de whenever possi ble.

Venous Thromboembolism Loss of the milking ction of muscle contr ction on the venous return from the l egs predisposes to venous thrombosis during neuromuscul r p r lysis. Therefore, prophyl xis for venous thrombosis ( s described in Ch pter 7) is m nd tory durin g neuromuscul r p r lysis. In summ ry, the risks of neuromuscul r p r lysis re consider ble enough to consider voiding the use of neuromuscul r blocking gent s whenever possible. P.804 With the ggressive use of sed tion, voiding these gents will be e sier th n nticip ted. REFERENCES SEIZURES 1. Bleck TP, Smith MC, Pierre-Louis SJ-C, et l. Neurologic complic tions of cri tic l medic l illnesses. Crit C re Med 1993;21:98103.

3. Se mens CM, Slovis CM, Kr mer D, G vin LJ. Seizures: current clinic l guideli nes for ev lu tion nd emergency m n gement. Emerg Med Rep 1995;16:2330. 4. Wijdicks EFM, Sh rbrough FW. New-onset seizures in critic lly ill p tients. N eurology 1993;43:10421044. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 15)03-M y-05 07:56:59

2. Mosewich RK, So EL. A clinic l ppro ch to the cl ssific tion of seizures epileptic syndromes. M yo Clin Proc 1996;71:405414.

nd

Hypost tic Pneumoni Cle r nce of respir tory secretions is in dequ . Endotr che l suction c theters re un ble to the bsence of cough to help cle r secretions, ns in dependent lung regions. This c n le d to

te during neuromuscul r p r lysis re ch the dist l irw ys, nd in there will be pooling of secretio hypost tic pneumoni .

Ovid: ICU Book 5. Wijdicks EFM. Seizures in the intensive c re unit. In: Neurology of critic l illness. Phil delphi : FA D vis, 1995;1833. 6. Epilepsy Found tion of Americ 's Working Group on St tus Epilepticus. Tre tme nt of Convulsive St tus Epilepticus. JAMA 1993;270:854859. 7. Louis S, Kott H, McDonell F. The c rdiocircul tory ch nges c used by intr ven ous Dil ntin nd its solvent. Am He rt J 1967;74:523529. 8. Bertz RJ, Howrie DL. Di zep m by continuous intr venous infusion for st tus e pilepticus in nticonvuls nt hypersensitivity syndrome. Ann Ph rm cother 1993;27 :298 301.

MYASTHENIA GRAVIS 10. Dr chm n DB. My stheni gr vis. N Engl J Med 1994;330:17971810. 11. London SF, Ringel SP. Neuromuscul r emergencies. In: Weiner WJ, ed. Emergent nd urgent neurology. Phil delphi : JB Lippincott, 1992;5978. GUILLAIN-BARR SYNDROME 12. Ropper AH. Guill in-B rr syndrome. N Engl J Med 1992;326:11301136.

14. v n der Meche FGA, Schmitz PIM, the Dutch Guill in-B rr Study Group. A http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 15)03-M y-05 07:56:59

13. Hund EF, Borel CO, Cornbl th DR, et l. Intensive m n gement f Guill in-B rr syndrome. Crit C re Med 1993;21:433446.

9. J god A, Riggio S. Refr ctory st tus epilepticus in 93;22:13371348.

dults. Ann Emerg Med 19

nd tre tment o

Ovid: ICU Book r ndomized tri l comp ring intr venous immune globulin nd pl sm exch nge in Gu ill in-B rr syndrome. N Engl J Med 1992;326:11231129. CRITICAL ILLNESS POLYNEUROPATHY 15. Bolton CF. Polyneurop thy in critic lly ill p tients. J Neurol Neurosurg Psy chi try 1984;47:12231231. 16. Witt NJ, Zochodne DW, Bolton CF, et l. Peripher l nerve function in sepsis nd multiple org n f ilure. Chest 1991;99:176184. P.805

17. Leijten FSS, H rinck-de Weerd JE, Poortvliet DCJ, de Weerd AW. The role of p olyneurop thy in motor conv lescence fter prolonged mech nic l ventil tion. JAM A 1995;274:12211225. PULMONARY COMPLICATIONS 18. Newton-John H. Prevention of pulmon ry complic tions in severe Guill in-B rr syndrome by e rly ssisted ventil tion. Med J Aust 1985;142:444. 19. Mier-Jedrzejowicz AK, Brophy C, Green M. Respir tory muscle function in my s theni gr vis. Am Rev Respir Dis 1988;138:867873. NEUROMUSCULAR BLOCKADE 20. H nsen-Fl schen J, Cowen J, R ps E. Neuromuscul r block de in the intensive c re unit. More th n we b rg ined for. Am Rev Respir Dis 1993;147:234236. 21. Prielipp RC, Coursin DB. Applied ph rm cology of common neuromuscul r blocki ng gents in critic l c re. New Horiz 1994;2:3447. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 15)03-M y-05 07:56:59

Ovid: ICU Book 22. Armstrong DK, Crisp CB. Ph rm coeconomic issues of sed tion, n lgesi , nd neuromuscul r block de in critic l c re. New Horiz 1994;2:8593. 23. D vidson JE. Neuromuscul r block de: indic tions, peripher l nerve stimul ti on, nd other concurrent interventions. New Horiz 1994;2:7584. 24. W tling SM, D st JF. Prolonged p r lysis in intensive c re unit p tients f ter the use of neuromuscul r blocking gents: review of the liter ture. Crit C re Med 1994;22:884893.

SUGGESTED READINGS Grott JC, ed. M n gement of the cutely ill neurologic p tient. New York: Churc hill Livingstone, 1993. Wijdicks EFM. Neurology of critic l illness. Phil delphi : FA D vis, 1995. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 15)03-M y-05 07:56:59

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 52 STROKE AND RELATED DISORDERS Stroke ( poorly descriptive term for cute br in injury of v scul r origin) is the third le ding c use of de th in the United St tes nd is responsible for pp roxim tely one-fourth of ll de ths in the dult popul tion. Considering these c redenti ls, stroke should occupy top position in the hier rchy of life-thre te ning conditions. However, stroke h s tr dition lly received little ttention fro m critic l c re speci lists. This h s ch nged in recent ye rs. The current view of cute cerebr l inf rction emph sizes the simil rities to cute myoc rdi l inf rction nd stresses the v lue of ppro ching br in tt ck with the s me ggre ssive me sures used in the ppro ch to he rt tt ck (1,2). DEFINITIONS The clinic l disorders described in this ch pter re cerebrov scul r disorders. The following re some definitions nd cl ssific tions for these disorders propo sed by the N tion l Institute of Neurologic Disorders nd Stroke (3).

q An cute neurologic disorder q Produced by nontr um tic injury in the centr l nervous system th t is v scul r i n origin q Accomp nied by foc l r ther th n glob l neurologic dysfunction q Persists for longer th n 24 hours or results in de th within the first 24 hours http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (1 of 16)03-M y-05 07:58:44

STROKE Stroke is

clinic l condition with ll the following fe tures (3,4):

Ovid: ICU Book P.807 Cl ssific tions Stroke c n be cl ssified s ischemic or hemorrh gic b sed on the type of p tholo gic injury. Approxim tely 80% of strokes re ischemic nd 15% re hemorrh gic (1 0% c used by intr cerebr l hemorrh ge, nd 5% c used by sub r chnoid hemorrh ge) (5). Ischemic strokes c n be further cl ssified s thrombotic or embolic in ori gin. Thrombotic strokes origin te in the s me f shion s described for cute myo c rdi l inf rction (MI) (see Ch pter 19). Embolic strokes ccount for 20% of isc hemic strokes (6). Most emboli origin te from thrombi in the left trium (from tri l fibrill tion) nd left ventricle (from cute MI), but occ sion lly they c n rise from deep vein thrombosis in the legs th t embolized through p tent fo r men ov le (7). Stroke c n lso be cl ssified ccording to the r pidity of neur ologic recovery. A minor stroke, lso c lled reversible ischemic neurologic de ficit (RIND), is ch r cterized by complete recovery of neurologic function withi n 3 weeks fter the cute event (3). A m jor stroke is ch r cterized by neurolog ic deficits th t persist for longer th n 3 weeks fter the event. TRANSIENT ISCHEMIC ATTACK A tr nsient ischemic tt ck (TIA) is n episode of foc l loss of br in function ( s result of ischemi ) th t l sts less th n 24 hours (3). The m jor distincti on between TIA nd stroke is the underlying p thology; i.e., ischemi in TIA ver sus inf rction or hemorrh ge in stroke. This, in turn, determines the dur tion o f the neurologic deficits: less th n 24 hours in TIA nd longer th n 24 hours in stroke. BEDSIDE EVALUATION The p tient with suspected stroke will h ve new-onset foc l neurologic deficit s th t re not tr um tic in origin. If the deficits h ve been present for less t h n 24 hours, it is often impossible to distinguish TIA from stroke. The followi ng fe tures of the clinic l present tion c n be useful in the ev lu tion of susp ected stroke (8). These fe tures re included in T ble 52.1. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (2 of 16)03-M y-05 07:58:44

Ovid: ICU Book TABLE 52.1. THE BEDSIDE EVALUATION OF SUSPECTED STROKE Seizures Gener lized convulsive seizures nd convulsive st tus epilepticus re uncommon i n TIA nd stroke. Seizures develop in pproxim tely 10% of c ses of stroke (4). They usu lly ppe r in the first 24 hours nd re foc l r ther th n gener lized in most c ses. Fever Fever is uncommon in TIA but c n be present in pproxim tely 50% of p tients wit h stroke (9). P.808 In most c ses of fever ssoci ted with stroke, the fever is due to process other th n the stroke (e.g., infection or thromboembolism). Consciousness The reticul r ctiv ting system in the br instem is responsible for rous l or w kefulness (consciousness). Bec use most c ses of stroke re the result of cereb r l inf rction, loss of consciousness is not common finding in uncomplic ted s troke (1,4). When foc l neurologic deficits re ccomp nied by com , the most li kely di gnoses re intr cerebr l hemorrh ge, m ssive cerebr l inf rction with ce rebr l edem , br instem inf rction, or seizures (nonconvulsive seizures or posti ct l st te). http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (3 of 16)03-M y-05 07:58:44

Ovid: ICU Book Aph si The left cerebr l hemisphere is the domin nt hemisphere for speech in 90% of sub jects. D m ge involving the left cerebr l hemisphere produces condition known s ph si , which is defined s disturb nce in the comprehension nd formul ti on of l ngu ge (10). P tients with ph si c n h ve difficulty underst nding ver b l rem rks (receptive ph si ), difficulty in verb l expression (expressive ph si ), or both (glob l ph si ). Most p tients with ph si will h ve cerebr l inf rction in the distribution of the left middle cerebr l rtery (10). Other c uses of ph si re tumors, he d injury, nd Alzheimer's dementi .

We kness The h llm rk of ischemi or hemorrh gic injury involving the cerebr l hemisphere s is we kness in the contr l ter l limbs. Limb we kness is present if the p tien t is un ble to hold the rm in 90 degrees of bduction for 10 seconds or un ble to hold the leg 30 degrees bove P.809 the horizont l pl ne for 5 seconds (11). The presence of hemip resis supports the di gnosis of TIA or stroke. However, hemip resis h s lso been described in met bolic enceph lop thy c used by ren l f ilure (12) nd sepsis (13). DIAGNOSTIC EVALUATION The di gnostic ev lu tion of suspected stroke h s tr dition lly proceeded t s low p ce. However, s mentioned in the introduction to this ch pter, the current view of ischemic stroke stresses the simil rities with cute MI, nd the need t o ppro ch cute stroke with the s me l crity used in the ppro ch to cute MI. According to the U.S. N tion l Stroke Associ tion, the ev lu tion of suspected stroke should be completed within 6 hours fter the onset of symptoms (14). Or s t ted more succinctly, time is br in (15). ROUTINE STUDIES The ev lu tion of suspected stroke should include blood chemistries to se rch fo r hypoglycemi , hypon tremi , hypern tremi , nd ren l f ilure. Addition l routi ne studies should include n INR if the p tient is being tre ted with coum din, n electroc rdiogr m if tri l fibrill tion is suspected, nd chest x-r y if t he p tient h s fever.

http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (4 of 16)03-M y-05 07:58:44

COMPUTED TOMOGRAPHY Computed tomogr phy (CT) of the br in c n identify ischemic inf rction rh ge nd

nd hemor

Ovid: ICU Book c n distinguish between the two (16). The sensitivity of CT sc ns is 70% for cer ebr l inf rction (17) nd over 90% for intr cerebr l hemorrh ge (1). However, th e sensitivity of CT sc ns is influenced by the time p ssed from the onset of str oke to the time the sc ns re performed. Timing The influence of timing on the di gnostic yield from CT sc ns is illustr ted in Figure 52.1. For cerebr l inf rctions, the di gnostic yield from CT sc ns is 50% lower if the sc ns re performed within 24 hours fter the inf rction (16). The refore, n unreve ling CT sc n performed within 24 hours fter the onset of susp ected stroke does not rule out the possibility of cerebr l inf rction. Figure 52.1. The influence of timing on the yield from CT sc ns. Both CT sc ns re from the s me p tient with suspected stroke. The sc n on the left w s obt ine d within 24 hours fter the onset of symptoms nd is unreve ling. The sc n on th e right w s obt ined 3 d ys l ter nd shows l rge hypodense re (inf rction) with m ss effect in the left cerebr l hemisphere. (Reproduced with permission fr om Reference 16.) Indic tions Two m jor benefits re derived from CT sc ns in suspected stroke. First, CT sc n s c n distinguish inf rction from hemorrh ge, which is http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (5 of 16)03-M y-05 07:58:44

Ovid: ICU Book

Cost As m y be expected, CT sc ns re costly. At Presbyteri n Medic l Center (Univers ity of Pennsylv ni ), the ch rge to the p tient for n unenh nced CT sc n of the he d is $1342 (ch rge for the fisc l ye r 1996). MAGNETIC RESONANCE IMAGING M gnetic reson nce im ging (MRI) h s higher di gnostic yield th n CT sc ns for bl nd inf rctions (p rticul rly those involving the cerebellum nd br instem) ( 18). The v lue of MRI in suspected stroke P.811 is illustr ted by the c se in Fi gure 52.2. The MRI sc n in this figure is from previously he lthy 39-ye r-old wom n who experienced cute onset of right rm nd leg we kness. The initi l CT sc n w s unreve ling, but the MRI sc n reve ls multiple (hyperdense) inf rctions long the course of the left middle cerebr l rtery (where s inf rctions re hy podense on CT sc ns, they re hyperdense on MRI sc ns). This prompted cerebr l ngiogr m, which reve led prob ble v sculitis s c use of the cerebr l inf rc tion. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (6 of 16)03-M y-05 07:58:44

P.810 import nt for selecting the ppropri te ntify the occ sion l c se of suspected stroke (tumor or bscess). For these re sons, CT sc gnostic procedure in p tients with suspected

ther py. Second, CT sc ns will ide c used by sp ce-occupying lesion ns re recommended s routine di stroke (1).

Ovid: ICU Book

Indic tions MRI is reserved for the occ sion l c se of suspected stroke in which CT sc ns r e unreve ling. However, bec use of the expense of MRI P.812 (see below), it shou ld be reserved for c ses in which the results of MRI will le d to improved ther py nd better ch nce for recovery. Contr indic tions Bec use MRI uses m gnetic pulses, it is contr indic ted in p tients with impl nt ed http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (7 of 16)03-M y-05 07:58:44

Figure 52.2. A rightsided we of inf rction istory nd MRI

T2-weighted MRI sc n from 39-ye r-old wom n with cute onset of kness nd norm l CT sc n. The rrows point to hyperdense re s long the distribution of the left middle cerebr l rtery. (C se h sc n courtesy of Dr. S mi Khell , M.D.)

Ovid: ICU Book p cem kers, cerebr l neurysm clips, intr ocul r met l, nd cochle r impl nts (1 8). Other met l impl nts nd ven c v filters re rel tive contr indic tions to MRI (18). Cost At Presbyteri n Medic l Center (University of Pennsylv ni ), the p tient ch rge for n MRI of the br in is $1761 (for fisc l ye r 1996).

q Lumb r puncture is not indic ted in most p tients with suspected stroke. It c n be useful in the occ sion l c se when CT sc n reve ls equivoc l evidence of su b r chnoid hemorrh ge or when n bscess is in close proximity to the sub r chno id sp ce. q Echoc rdiogr phy is indic ted when stroke is ssoci ted with tri l fibrill tion , cute MI, or left-sided endoc rditis. It m y lso be indic ted in stroke of un determined etiology to identify p tent for men ov le ( nd possible p r doxic l cerebr l embolism). q

HYPERTENSION Hypertension is common in the e rly period fter cute stroke, but ntihypertens ive ther py is not dvised s routine pr ctice for two re sons. First, cerebr l utoregul tion m y be lost in the region of the br in th t is d m ged, nd thu s cute lowering of the blood pressure could enh nce the extent of injury. This is verified by clinic l study showing th t cute lowering of blood pressure in hypertensive stroke victims is often ccomp nied by worsening of the http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (8 of 16)03-M y-05 07:58:44

EARLY MANAGEMENT The following discussion refers to the m n gement in the first 24 hours fter ute stroke.

Electroenceph logr phy is indic ted in c ses in which seizures re suspected the c use of the neurologic deficits.

OTHER DIAGNOSTIC TESTS The following tests re

ppropri te for the indic tions cited.

Ovid: ICU Book neurologic deficits (19). This is pplied to p tients with severe hypertension ( di stolic blood pressure bove 120 mm Hg) s well. The second re son to void P. 813 ntihypertensive ther py is the tendency for the hypertension to resolve spo nt neously in the d ys following n cute stroke (20). The Stroke Council of the Americ n He rt Associ tion recommends ntihypertensive tre tment when the systo lic pressure is bove 220 mm Hg or when the me n blood pressure is bove 130 mm Hg (1). However, this recommend tion l cks experiment l v lid tion. If ntihyper tensive ther py is used, the go l must be gr du l r ther th n prompt reduction o f blood pressure. Both nitroglycerin nd nitroprusside should be voided bec use these cerebr l v sodil tors c n incre se intr cr ni l pressure (1). Nic rdipine ( c lcium ch nnel blocker th t preserves cerebr l blood flow) or ngiotensin-c onverting enzyme inhibitors (which h ve little effect on cerebr l vessels) m y b e the most ppropri te gents for lowering blood pressure in cute stroke. ANTICOAGULATION Approxim tely 20% of p tients with cute ischemic stroke develop progressive neu rologic deficits over the ensuring 4 d ys (21,22). Ther peutic ntico gul tion w ith hep rin h s been the tr dition l pr ctice for p tients with progressive isch emic stroke (22). Although e rly studies showed possible benefit from this pr ctice, these studies were not well designed. More recent studies reve l little o r no benefit from full ntico gul tion in progressive ischemic stroke (22). THROMBOLYTIC THERAPY Considering the simil rities between cute thrombotic inf rction of the br in n d cute (thrombotic) MI nd the success with thrombolytic ther py in cute MI (s ee Ch pter 19), it follows th t thrombolytic ther py should be ev lu ted in cut e ischemic stroke. The studies of thrombolytic ther py in ischemic stroke h ve n ot been encour ging. However, in the most recent study, which w s sponsored by t he N tion l Institute of Neurologic Disorders nd Stroke (NINDS) (23), p tients given tissue pl sminogen ctiv tor (0.9 mg/kg over 1 hour) within 3 hours fter the onset of ischemic stroke showed signific ntly fewer neurologic deficits 3 mo nths l ter. (The clinic l course in the e rly period fter cute ischemic stroke w s not f vor bly influenced by thrombolytic ther py in this study.) B sed on t he NINDS study just described, the Food nd Drug Administr tion h s pproved the use of tissue pl sminogen ctiv tor in the first 3 hours fter the onset of cu te ischemic stroke (24). However, bec use CT sc n must be obt ined to rule out hemorrh ge http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (9 of 16)03-M y-05 07:58:44

Ovid: ICU Book before thrombolytic ther py is st rted, this me ns th t p tients with ischemic s troke must seek tre tment in the emergency dep rtment nd h ve CT sc n complet ed within 3 hours to be c ndid tes for thrombolytic ther py. In light of these s trict requirements, it is unlikely P.814 th t thrombolytic ther py will become common tre tment mod lity in cute ischemic stroke. INCREASED INTRACRANIAL PRESSURE Incre sed intr cr ni l pressure c n be the result of intr cerebr l hemorrh ge or m ssive inf rction with cerebr l edem . In either c se, it c rries poor progn osis. Sever l me sures re used to lower intr cr ni l pressure following he d in jury (25). However, m ny of these re imed t reducing cerebr l blood flow, whi ch c n ggr v te n ischemic stroke. The use of selected methods to lower intr c r ni l pressure in cute ischemic stroke c n be summ rized s follows: q Intr cr ni l pressure monitoring is of unproven v lue for m n ging intr cr ni l hypertension in ischemic stroke (1) nd is not recommended s routine me sure. q Elev ting the he d of the bed to 30 degrees will reduce intr cr ni l pressure by promoting venous return from the he d (25), nd this me sure is recommended for ll p tients with intr cr ni l hypertension (1). q Endotr che l suctioning c n incre se intr cr ni l pressure, even when hypoxemi is prevented by preceding period of 100% oxygen inh l tion (26). Therefore, en dotr che l suctioning should be reduced in frequency nd dur tion (if possible) in p tients with intr cr ni l hypertension (26). q Hyperventil tion to induce hypoc pni nd reduce cerebr l blood flow does not im prove outcome in p tients with he d injury (25). This l ck of documented benef it, together with the risk of ex gger ted ischemi during hyperventil tion, m ke s hyperventil tion n undesir ble intervention is ischemic stroke. q High-dos ge corticosteroids do not improve outcome in ischemic stroke with cereb r l edem nd c n incre se the risk of infection (1). Therefore, steroids should be voided in ll c ses of intr cr ni l hypertension. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (10 of 16)03-M y-05 07:58:44

Ovid: ICU Book M nnitol lowers intr cr ni l pressure by dr wing w ter out of cerebr l tissues ( 27). Although of unproven v lue, m nnitol c n be given in c ses of severe or pro gressive cerebr l edem from cute stroke. The dose is 0.25 to 0.5 g/kg IV over 20 minutes. Hypertonic fluids like m nnitol c n incre se the perme bility of the blood-br in b rrier (25), which f vors the entry of m nnitol into cerebr l tiss ues. Bec use of this risk, m nnitol should not be given in repe ted doses to con trol intr cr ni l pressure (25). SUBARACHNOID HEMORRHAGE Sub r chnoid hemorrh ge (SAH) is usu lly the result of neurysm l rupture or ble eding from n rteriovenous m lform tion. Predisposing f ctors for SAH include c oc ine buse nd bleeding disorders (28). P.815 Although cl ssified s type of stroke (5), SAH c n differ from the other types of stroke in both present tion nd m n gement.

CLINICAL PRESENTATION The h llm rk of the clinic l present tion of SAH is he d che. The full-blown syn drome m y be preceded by severe but self-limited he d che c lled sentinel he d che (29), which is presum bly the result of neurysm l dil tion or sm ll he morrh gic le k. The he d che of SAH is usu lly brupt in onset, persistent nd p rogressive, nd worse with exertion. Severe he d che th t is worse with exertion is more ch r cteristic of SAH th n the myri d other c uses of he d che (29). Al though the he d che of SAH tends to be centered t the b se of the skull or in t he cervic l region, this fe ture is not specific for SAH. Other m nifest tions, such s n use nd vomiting, ment l st tus ch nges, nd stiff neck, m y or m y n ot be present. DIAGNOSTIC EVALUATION As mentioned e rlier, CT sc ns of the he d (unenh nced) h ve 90% sensitivity f or the detection of hemorrh ge, including sub r chnoid hemorrh ge, nd thus re the initi l di gnostic test of choice for suspected SAH. However, CT sc ns c n m iss SAH in the posterior foss (where the br instem nd cerebellum re loc ted). The im ge in Figure 52.3 is n MRI sc n from 30-ye r-old wom n with severe n d persistent he d che who h d norm l CT sc n of the he d. The MRI sc n shows hyperdense re (indic ted by the rrows) just ventr l to the pons, which repre sents SAH. Thus, even though CT sc ns h ve high sensitivity for SAH, neg t ive CT sc n does not elimin te the possibility for SAH. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (11 of 16)03-M y-05 07:58:44

Ovid: ICU Book Figure 52.3. An MRI sc n from 30-ye r-old wom n with severe, persistent he d c he nd norm l CT sc n of the he d. Note the rrows pointing to hyperdense r e ventr l to the br instem. This represents prepontine sub r chnoid hemorrh g e. Lumb r puncture confirmed the presence of blood in the sub r chnoid sp ce. (C se history nd MRI sc n courtesy of Dr. S mi Khell , M.D.)

MANAGEMENT The morbidity nd mort lity in SAH is rel ted to two processes: recurrence of th e hemorrh ge nd cerebr l v sosp sm. Recurrent SAH In most c ses of SAH, the bleeding h s subsided t the time of di gnosis. To pre vent recurrence of the SAH, cerebr l ngiogr phy is performed to identify the responsible v scul r bnorm lity for surgic l correction. However, ngiogr phy i s usu lly del yed until the p tient is w ke nd clinic lly recovered. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (12 of 16)03-M y-05 07:58:44

Ovid: ICU Book Cerebr l V sosp sm The neurologic deficits in SAH re c used by v sosp sm of cerebr l vessels with result nt cerebr l ischemi . The v sosp sm is produced P.816 by blood in the sub r chnoid sp ce, lthough the ex ct mech nism is uncle r. This v scul r response is ttenu ted by nimodipine, c lcium ch nnel blocker th t h s preferenti l v sodil ting effect on cerebr l vessels. Nimodipine in dos ge of 0.35 mg/kg or lly every 4 hours h s proven effective in reducing v sosp sm (30) nd improving neurologic function (31) in p tients with SAH. As result, this gent is used routinely in SAH. REFERENCES REVIEWS 1. Stroke Council of the Americ n He rt Associ tion. Guidelines for the m n geme nt of p tients with cute ischemic stroke. Circul tion 1994;90:15881601 (179 refe rences). 2. N r dz y JFX, G sch WR. Acute stroke. Emerg Med Clin North Am 1996;14:197 216 (123 references). DEFINITIONS 3. Speci l Report from the N tion l Institute of Neurologic Disorders nd Stroke . Cl ssific tion of cerebrov scul r dise ses III. Stroke 1990;21:637676. P.817 4. B mford J. Clinic l ex min tion in di gnosis nd subcl ssific tion of stroke. L ncet 1992;339:400405. 5. D'Cost DF. Subcl ssific tion of strokes. L ncet 1992;339:1541. 6. H rt RG. C rdioembolic stroke. Am J Med 1996;100:465474. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (13 of 16)03-M y-05 07:58:44

Ovid: ICU Book

BEDSIDE EVALUATION 8. Goldstein LB, M tch r DB. Clinic l ssessment of stroke. JAMA 1994;271:1114112 0. 9. Oppenheimer S, H chinski V. Complic tions of cute stroke. L ncet 1992;339:72 1 724. 10. D m sio A. Aph si . N Engl J Med 1992;326:531559. 11. Brott T, Ad ms HP, Olinger CP, et l. Me surements of cute cerebr l inf rct ions: clinic l ex min tion sc le. Stroke 1989;20:864870. 12. Bolton CF, Young GB. Neurologic complic tions of ren l f ilure. Toronto: But terworths, 1990. 13. M her J, Young GB. Septic enceph lop thy. Intensive C re Med 1993;8:177187. DIAGNOSTIC EVALUATION 14. N tion l Stroke Associ tion Consensus St tement. Stroke: the first 6 hours. Stroke, Clinic l Upd tes. Vol. IV. Englewood, CO: N tion l Stroke Associ tion, 1 993. 15. McC rthy M. Time is br in. L ncet 1993;341:13391340. 16. Gr ves VB, P rtington VB. Im ging ev lu tion of cute neurologic dise se. In : Goodm n LR, Putm n CE, eds. Critic l c re im ging. 3rd ed. Phil delphi : WB S unders, 1992;391409. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (14 of 16)03-M y-05 07:58:44

7. Di Tullio M, S cco RL, Gop l A, et l. P tent for men ov le s for cryptogenic stroke. Ann Intern Med 1992;117:461465.

risk f ctor

Ovid: ICU Book 17. Bry n RN, Levy LM, Whitlow WD, et l. Di gnosis of cute cerebr l inf rction : comp rison of CT nd MR im ging. Am J Neuror diol 1991;12:611620. 18. Kent DL, H ynor DR, Longstreth WT, L rson EB. The clinic l effic cy of m gne tic reson nce im ging in neuroim ging. Ann Intern Med 1994;120:856874. MANAGEMENT 19. Phillips SJ. P thophysiology nd m n gement of hypertension in cute, ischem ic stroke. Hypertension 1994;23:131136. 20. O'Connell JE, Gr y CS. Tre ting hypertension fter stroke. BMJ 1994;308:1523 1524.

22. Rothrock JF, H rt RG. Antithrombotic ther py in cerebrov scul r dise se. Ann Intern Med 1991;115:885895. 23. The N tion l Institute of Neurologic Disorders nd Stroke rt-PA Stroke Study Group. Tissue pl sminogen ctiv tor for cute ischemic stroke. N Engl J Med 199 5;333:1581 1587. 24. Bleck TP. Thrombolysis for cute ischemic stroke: how, when, nd why. J Crit Illness 1996;11:645657. P.818 25. Cold GE, Holdg rd HO. Tre tment of intr cr ni l hypertension in cute he d injury with soci l reference to the role of hyperventil tion nd sed tion with b rbitur tes: review. Intensive C re World 1992;9:172178. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (15 of 16)03-M y-05 07:58:44

21. Britton M, Roden A. Progression of stroke fter 1985;16:629633.

rriv l t hospit l. Stroke

Ovid: ICU Book

27. N th F, G lbr ith S. The effect of m nnitol on cerebr l white m tter w ter c ontent. J Neurosurg 1986;65:4143. SUBARACHNOID HEMORRHAGE 28. Ozer MN, M terson RS, C pl n LR. M n gement of persons with stroke. St. Loui s: Mosby, 1994;8384. 29. S per JR. He d che: urgent consider tions in di gnosis nd tre tment. In: We iner WJ. Emergent nd urgent neurology. Phil delphi : JB Lippincott, 1992;509531. 30. Allen GS, Ahn HS, Preziosi TJ, et l. Cerebr l rteri l sp sm controlled tri l of nimodipine in p tients with sub r chnoid hemorrh ge. N Engl J Med 1983;308 :619624.

31. Petruk KC, West M, Mohr G, et l. Nimodipine tre tment in poor-gr de neurys m p tients: results of multicenter double-blind pl cebo-controlled tri l. J Ne urosurg 1988;68:505517. SUGGESTED READINGS Ozer MN, M terson RS, C pl n LR. M n gement of persons with stroke. St. Louis: M osby, 1994. Wijdicks EFM. Neurology of critic l illness. Phil delphi : FA D vis, 1995. http://g tew y.ut.ovid.com/gw1/ovidweb.cgi (16 of 16)03-M y-05 07:58:44

26. Rudy EB, Turner BS, B un M, et l. Endotr che l suctioning in d injury. He rt Lung 1991;20:667674.

dults with he

Ovid: ICU Book Copyright 1998 Lippincott Willi ms & Wilkins M rino, P ul L. ICU Book, 2nd Editio n Ch pter 53 PHARMACEUTICAL TOXINS AND ANTIDOTES Poisons nd medicine re oftentimes the s me subst nce given with different inte nts. --Peter L th m (1865) There is little doubt th t ph rm ceutic l mish ps re source of consider ble morbidity nd even mort lity. Adverse drug re ctions re responsible for t le st 10% of hospit l dmissions nd s m ny s 20% of ICU dmissions (1,2). Once in the hospit l, s m ny s 30% of p tients experience n dverse drug re ction before disch rge (3). The ver ge p tient in the ICU rec eives 6 to 9 different medic tions d ily nd 8 to 12 different medic tions durin g the ICU st y (1). Therefore, the ICU is fertile environment for ph rm ceutic l mis dventures. This ch pter describes the clinic l toxicity ssoci ted with t he ph rm ceutic l gents listed below, nd the tre tment of e ch with specific ntidotes (shown in p rentheses). Included re toxic drug ingestions th t prompt dmission to the ICU nd toxic drug re ctions th t surf ce fter dmission to th e ICU. (For list of the most common toxic ingestions in the United St tes, see the Appendix.) Acet minophen ( cetylcysteine) Benzodi zepines (flum zenil) -Bloc kers (glucagon) Calcium antagonists (calcium, glucagon) Digitalis (anti ody frag ments) Nitroprusside (nitrites, thiosulfate) Opioids (naloxone) Tricyclic antide pressants ( icar onate) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 36)03-May-05 07:54:17

Ovid: ICU Book ACETAMINOPHEN Acetaminophen gained widespread popularity in the 1960s as a less toxic analgesi c antipyretic agent than aspirin. Ironically, acetaminophen P.822 is now the seco nd leading cause of toxic drug ingestions in the United States (4). Although it is safe when administered in the usual therapeutic dosages (less than 4 g daily) , acetaminophen can e a lethal hepatotoxin when ingested in large doses. TOXIC MECHANISM The toxicity of acetaminophen is related to its meta olism, which is shown in Fi gure 53.1. The ulk of acetaminophen meta olism involves the formation of sulfat e and glucuronide conjugates in the liver, which are then excreted in the urine. Approximately 5% of the meta olism involves oxidation y cytochrome P-450depende nt enzymes to form a highly reactive intermediate that can promote oxidant injur y in hepatic parenchymal cells. This toxic meta olite is normally removed y con jugation with glutathione (an intracellular antioxidant descri ed in Chapter 3). Large doses of acetaminophen prompt increased use of the glutathione pathway, a nd this can lead to glutathione depletion. When hepatic glutathione stores fall to 30% of normal levels, the toxic meta olite can accumulate and promote widespr ead hepatocellular damage (5). Figure 53.1. The hepatic meta olism of acetaminophen and the mechanism of action of Nacetylcysteine. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 36)03-May-05 07:54:17

Ovid: ICU Book Predisposing Conditions According to the meta olic sequence in Figure 53.1, the following conditions pre dispose to acetaminophen hepatotoxicity (5,6,7 and 8): Therapy with drugs that e nhance hepatic microsomal oxidation, such as ar iturates, phenytoin, and rifamp in Conditions associated with glutathione depletion, such as chronic P.823 ethan ol ingestion, malnutrition, and infections with the human immunodeficiency virus (HIV) These conditions are important to consider when interpreting serum acetam inophen levels, as discussed later in this chapter. CLINICAL PRESENTATION The period following acetaminophen overdose is divided into three stages (5). In the initial stage (the first 24 hours), symptoms are either a sent or nonspecif ic (e.g., nausea) and no la oratory evidence of hepatic injury exists. In patien ts who develop hepatotoxicity, a second stage (24 to 72 hours after drug ingesti on) occurs, during which clinical manifestations continue to e minimal or a sen t, ut la oratory evidence of hepatic injury (i.e., elevated serum transaminases ) egins to appear. In advanced cases of hepatic injury, a third stage follows ( after 72 hours after ingestion) that is characterized y clinical and la oratory evidence of progressive hepatic injury and hepatic insufficiency (e.g., encepha lopathy and coagulopathy), occasionally com ined with renal insufficiency. DIAGNOSIS In most cases of acetaminophen overdose, the initial presentation occurs within 24 hours after drug ingestion, when no manifestations of hepatic injury are pres ent. The principal task at this time is to identify the patients who are likely to develop hepatotoxicity in later stages. Two varia les can have prognostic val ue in this regard. Ingested Dose The toxic dose of acetaminophen can vary in individual patients. However, when t he ingested dose of acetaminophen is known, the following criteria can e used t o predict the likelihood of hepatotoxicity. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 36)03-May-05 07:54:17

Ovid: ICU Book Ingested Dose 250 mg/kg >140 mg/kg > 70 mg/kg Hepatotoxicity Pro a le Possi le Possi le (high-risk conditions) Serum Drug Levels The serum acetaminophen level is the most relia le predictor of toxic risk in th e early period after drug ingestion (5,6). Figure 53.2 shows a widely used nomog ram that relates toxic risk to the serum acetaminophen level (in micrograms/mill iliter) in the first 24 hours after drug ingestion. The first 4 hours after drug ingestion is not included in the nomogram ecause serum drug levels may not rea ch P.824 steady-state levels during this time. The dotted line (lower threshold) is used for patients with a condition that predisposes to acetaminophen hepatot oxicity (9); the solid line is used for all other patients. If a serum acetamino phen level falls in the high-risk region of the nomogram, the risk of developing hepatotoxicity is 60% or higher (5), and this is sufficient to warrant the trea tment descri ed here. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 36)03-May-05 07:54:17

Ovid: ICU Book Figure 53.2. Nomogram for predicting the risk of hepatotoxicity from the serum a cetaminophen level in the first 24 hours after drug ingestion. Any serum level t hat falls on or a ove the threshold lines is an indication to egin therapy with N-acetylcysteine. The normal threshold (solid line) is from Reference 5, and th e threshold for predisposing conditions (dashed line) is from Reference 6. N-ACETYLCYSTEINE The goal of therapy in acetaminophen overdose is to limit or prevent accumulatio n of the toxic meta olite. Because glutathione does not readily cross cell mem r anes, the administration of exogenous glutathione is not well suited for this ta sk. N-acetylcysteine (NAC) is a glutathione analog that can cross cell mem ranes and act as an intracellular glutathione surrogate (10). As shown in Figure 53.1 , NAC contains a sulfhydryl group that allows it to act as a reducing agent and inactivate the toxic acetaminophen meta olite. This a ility to act as a reducing agent also allows N-acetylcysteine to function as an antioxidant agent, as desc ri ed in Chapter 3. In fact, the toxic effect of the acetaminophen meta olite is an oxidant-induced injury, similar to the cell injury produced P.825 y oxygen meta olites, and N-acetylcystein protects against acetaminophen toxicity ecause of its antioxidant actions. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 36)03-May-05 07:54:17

Ovid: ICU Book Timing Therapy with NAC provides protection against acetaminophen hepatotoxicity when t reatment is initiated within 24 hours after drug ingestion (5,6,11). Although so me evidence exists that NAC can e protective when given 24 to 36 hours after dr ug ingestion (12), the consensus is that N-acetylcysteine is indicated only when therapy can e started within 24 hours after acetaminophen overdose. Protection is greatest when therapy starts within 8 hours after drug ingestion (11), and t hereafter the protective effects of NAC decline steadily with time. Therefore, a voiding treatment delays is the single most important measure for ensuring optim al therapy with N-acetylcysteine. Therapeutic Regimens NAC can e given orally or intravenously using the dosing regimens shown in Ta l e 53.1. The intravenous route is preferred in most countries except the United S tates, where orally administered NAC is the only regimen approved for clinical u se. Despite the preference for oral therapy in the United States, oth routes of therapy are considered equally effective. TABLE 53.1. TREATMENT OF ACETAMINOPHEN OVERDOSE WITH NACETYLCYSTEINE (NAC) Adverse Reactions The sulfur content of NAC gives the liquid drug preparation a very disagreea le taste (often http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 36)03-May-05 07:54:17

Ovid: ICU Book descri ed as rotten eggs). As a result, oral administration of NAC often incites vomiting. NAC can e diluted in grapefruit juice to improve tolerance, ut in m any instances, a nasogastric tu e must e inserted to administer therapy. The or al regimen of P.826 NAC is also associated with a dose-dependent diarrhea that a ppears in approximately half the patients who complete the 72-hour regimen (13). Fortunately, this diarrhea resolves with continued therapy in over 90% of cases . Intravenous NAC can cause trou lesome anaphylactoid reactions (14), ut these reactions are uncommon. ACTIVATED CHARCOAL Acetaminophen is completely a sor ed from the gastrointestinal tract in the firs t few hours after drug ingestion (5). Therefore, activated charcoal (1 g/kg ody weight) is recommended only in the first 4 hours after acetaminophen overdose ( 5,15). Although charcoal can also adsor N-acetylcysteine, this interaction is p ro a ly not significant (5). BENZODIAZEPINES Benzodiazepines are the most commonly overdosed prescription drugs in the United States (4). However, admissions for drug overdose are not the only source of e nzodiazepine toxicity in the ICU. Surveys reveal that roughly 50% of patients in the ICU receive enzodiazepines for sedation (16,17), and adverse reactions to enzodiazepine sedation are likely to e a significant source of clinical toxici ty. The use of enzodiazepines for sedation in the ICU is descri ed in Chapter 8 .

CLINICAL TOXICITY Benzodiazepines produce a dose-dependent depression in the level of consciousnes s, ut this effect usually is not accompanied y respiratory depression or cardi ovascular depression. The lack of significant respiratory or cardiovascular depr ession explains why enzodiazepine overdose is rarely fatal (18). However, sever al factors in the ICU can promote respiratory and cardiovascular depression with enzodiazepines. These include advanced age, com ined therapy with opioid analg esics, and prolonged drug therapy. Drug Accumulation As mentioned in Chapter 8, the enzodiazepines are lipophilic agents, and thus h ave a tendency to accumulate in lipid-rich tissues such as the central nervous s ystem. This can result in oversedation when enzodiazepines are given for prolon ged periods (19,20,21 and http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 36)03-May-05 07:54:17

Ovid: ICU Book 22). This tendency for drug accumulation is underappreciated, particularly with midazolam, which is considered a rapidly acting agent. There is a general miscon ception that the short elimination half-life of midazolam (90 minutes) indicates that the drug is rapidly cleared from the ody. However, the elimination of mid azolam from the circulation is caused largely y uptake into the central nervous system and other P.827 lipid-laden tissues. Therefore, the short elimination ha lf-life of midazolam does not represent rapid elimination from the ody. As a re sult, prolonged infusions of midazolam can e associated with significant drug a ccumulation in the central nervous system, leading to oversedation and delays in weaning from mechanical ventilation (21,22). FLUMAZENIL Flumazenil is a enzodiazepine antagonist that inds to enzodiazepine receptors in the central nervous system ut does not exert any agonist actions (23). It i s most effective in reversing the sedative effects of the enzodiazepines ut is inconsistent in reversing enzodiazepine-induced respiratory depression (24). F lumazenil can also improve the sensorium in ethanol intoxication (25), ut the d oses required are large (5 mg) and potentially hazardous. Drug Administration Flumazenil is given as an intravenous olus. The initial dose is 0.2 mg, and thi s can e repeated at 1-minute intervals if necessary to a cumulative dose of 1.0 mg. The response is rapid in onset (1 to 2 minutes), and lasts for approximatel y 1 hour (19,21,23). Because flumazenil has a shorter duration of action than th e enzodiazepines, resedation is common after 30 to 60 minutes. Because of the r isk of resedation, the initial olus dose of flumazenil is often followed y a c ontinuous infusion at 0.3 to 0.4 mg/hour (19). Adverse Reactions Flumazenil produces few undesira le side effects (23,26,27). It can precipitate a enzodiazepine withdrawal syndrome in patients with a long-standing history of enzodiazepine use (26). Flumazenil can also precipitate seizures in patients r eceiving enzodiazepines for seizure control and in mixed overdoses involving tr icyclic antidepressants (28). Clinical Uses http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 36)03-May-05 07:54:17

Ovid: ICU Book Because of the enign nature of enzodiazepine toxicity, flumazenil is a treatme nt in search of an illness. The principal use of flumazenil is in patients with known or suspected enzodiazepine overdose, ut only when a mixed overdose invol ving tricyclic antidepressants is not suspected, and only in patients who are no t taking enzodiazepines for seizure control. Flumazenil has even fewer uses in the ICU. Although flumazenil can reverse oversedation with enzodiazepines in ve ntilator-dependent patients (19,20,21 and 22), the hazards of oversedation are m inimal in patients receiving ventilatory support. Flumazenil has een reported t o hasten weaning from mechanical ventilation (21,22 and 23), ut P.828 this appl ication seems limited y the lack of respiratory depression in enzodiazepine se dation. -RECEPTOR ANTAGONISTS Toxic ingestion of -receptor antagonists is increasing yearly, and more than 5000 cases of locker overdose were reported in 1992 (4). Intentional overdose is not the only source of locker toxicity in the ICU. Blockers are used to manage seve ral conditions in the ICU, including hypertension, narrow-complex tachycardias, unsta le angina, and acute myocardial infarction, and these uses create an addit ional source of - locker toxicity. The -receptor antagonists used most often in th e ICU are shown in Ta le 53.2. TABLE 53.2. COMPARISON OF INTRAVENOUS -RECEPTOR ANTAGONISTS CLINICAL TOXICITY http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 36)03-May-05 07:54:17

Ovid: ICU Book The manifestations of -receptor lockade arise primarily from the cardiovascular system and the central nervous system (29,30). Cardiovascular Toxicity The most common manifestations of - locker toxicity are radycardia and hypotensi on (29,30). The radycardia is usually sinus in origin and is well tolerated. Th e hypotension can e caused y peripheral vasodilation (renin lockade) or a dec rease in cardiac output (-1 receptor lockade). Hypotension that is sudden in ons et is usually a reflection of a decrease in cardiac output and is an ominous sig n. The more lipophilic - lockers (e.g., propranolol) have a mem rane-sta ilizing or quinidine-like action that can prolong atrioventricular (AV) conduction. This effect, which is not the result of lockade, can produce various degrees of hear t lock. P.829 Neurotoxicity Most - lockers are lipid solu le to some degree and thus have a tendency to accum ulate in lipid-rich organs such as the central nervous system. As a result, - loc ker overdose is often accompanied y lethargy, depressed consciousness, and gene ralized seizures. The latter manifestation is more prevalent than suspected and has een reported in 60% of overdoses with highly lipophilic agents such as prop ranolol (30). Like prolonged AV conduction, the neurologic manifestations are no t the result of -receptor lockade. GLUCAGON The cardiovascular depression from -receptor lockade is resistant to conventiona l therapy with atropine (1 mg intravenously), isoproterenol (0.1 to 0.2 mg/minut e), and transvenous pacing (29). The regulatory hormone glucagon is the agent of choice for reversing the cardiovascular depression in -receptor lockade. Mechanism of Action The diagram in Figure 53.3 shows the chain of events responsi le for the positiv e inotropic actions of -1 receptor activation in the heart. The -receptor is funct ionally linked (via specialized G proteins) to the adenyl cyclase enzyme on the inner surface of the cell mem rane. Activation of the receptorenzyme complex resu lts in the hydrolysis P.830 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 36)03-May-05 07:54:17

Ovid: ICU Book of adenosine triphosphate (ATP) to form cyclic adenosine monophosphate (cAMP). T he cAMP then activates a protein kinase that promotes the inward movement of cal cium through the cell mem rane. The influx of calcium promotes interactions etw een contractile proteins and there y augments the strength of cardiac contractio n. Figure 53.3. The mechanism of action of drugs that affect the strength of cardia c contraction. ATP = adenosine triphosphate, cAMP = cyclic adenosine monophospha te, PDE = phosphodiesterase, AMP = adenosine monophosphate. The diagram in Figure 53.3 shows that glucagon can activate adenyl cyclase throu gh a mem rane receptor that is distinct from the -receptor. This allows glucagon to mimic the positive inotropic effects of -receptor activation when the -receptor s are quiescent. Indications Glucagon is indicated for the treatment of hypotension and symptomatic radycard ia associated with toxic exposure to - lockers (Ta le 53.3). When used at the app ropriate dosages, glucagon elicits a favora le response in 90% of patients (29). Glucagon is not indicated for reversing the prolonged AV conduction or neurolog ic a normalities in - locker overdose ecause these effects are not mediated y -r eceptor lockade. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 36)03-May-05 07:54:17

Ovid: ICU Book TABLE 53.3. ANTIDOTE THERAPY WITH GLUCAGON Preparation Glucagon is provided as a powder that must e reconstituted. The diluent provide d y the manufacturer usually includes phenol (car olic acid) as a preservative. Phenol can cause a variety of toxic side effects, including agitation, seizures , and hypotension. The threshold tolerance for phenol in adults is 50 mg (31), a nd this threshold will e reached after 5 hours with a glucagon infusion at the standard dosage (i.e., glucagon at 5 mg/hour delivers phenol at 10 mg/hour). The refore, the commercial phenol-containing diluent should e avoided and glucagon should e reconstituted in saline or sterile water. P.831 Dosing Recommendations The effective dose of glucagon can vary in individual patients, ut a olus dose of 3 to 5 mg intravenously should e effective in most adults (29,30). The init ial dose is 3 mg (or 0.05 mg/kg), and this can e followed y a second dose of 5 mg (or 0.07 mg/kg) when necessary. The response to glucagon is most pronounced when the plasma ionized calcium is normal (32). The effects of glucagon can e s hort-lived (5 minutes), so a favora le response should e followed y a continuo us infusion (5 mg/hour). Adverse Effects http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 36)03-May-05 07:54:17

Ovid: ICU Book Nausea and vomiting are common at glucagon doses a ove 5 mg/hour. Mild hyperglyc emia is common, and it is caused y glucagon-induced glycogenolysis and gluconeo genesis. The insulin response to the hyperglycemia can drive potassium into cell s and promote hypokalemia. Finally, glucagon stimulates catecholamine release fr om the adrenal medulla, and this can raise lood pressure in hypertensive patien ts. This hypertensive response is exaggerated in pheochromocytoma, so glucagon i s contraindicated in patients with pheochromocytoma. CALCIUM ANTAGONISTS The calcium antagonists are responsi le for more than 6000 cases of drug overdos e each year and are the fourth leading cause of death from toxic drug ingestions in the United States (4). The three original calcium antagonists (verapamil, ni fedipine, and diltiazem) are responsi le for most of the clinical experience wit h calcium antagonist toxicity. MECHANISMS Calcium has a profound influence on the electrical and mechanical performance of smooth muscle. Its role in the contraction of cardiac smooth muscle is shown in Figure 53.3. The inward movement of calcium across the cell mem rane (triggered y depolarization of the cell mem rane or y activation of the cAMP pathway) pr omotes the interaction etween contractile proteins that ultimately determines t he strength of muscle contraction. Although not shown in the figure, calcium inf lux also triggers calcium release from the sarcoplasmic reticulum, which is anot her source of calcium for muscle contraction. This process from mem rane depolar ization to muscle contraction is called excitationcontraction coupling (33). Calc ium also participates in the propagation of electrical impulses in smooth muscle . The depolarization-triggered inward movement of calcium P.832 facilitates the propagation of electrical impulses in cardiac muscle, and speeds conduction thro ugh the AV node. The calcium antagonists lock the inward movement of calcium ac ross smooth muscle mem ranes, ut not across the sarcoplasmic reticulum. This l ockade can result in negative inotropic and chronotropic effects, prolonged atri oventricular conduction (negative dromotropic effect), decreased arrhythmogenici ty, vascular dilation, and ronchial dilation. The individual calcium antagonist s differ in their a ility to elicit these responses. CLINICAL TOXICITY http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 36)03-May-05 07:54:17

Ovid: ICU Book The toxic manifestations of the three most popular calcium antagonists (verapami l, nifedipine, and diltiazem) are shown in Ta le 53.4 (34). Verapamil is most li kely to produce hypotension and prolonged AV conduction. Verapamil is only a wea k vasodilator, and the hypotension is caused y a decrease in cardiac output (ne gative inotropic effect) without compensatory vasoconstriction. Nifedipine is pr edominantly a vasodilator (hence the high incidence of sinus tachycardia) and ha s little influence on AV conduction. Diltiazem is similar to verapamil in its a ility to prolong AV conduction, ut it causes less cardiac depression than verap amil and less vasodilation than nifedipine. TABLE 53.4. CLINICAL FEATURES OF OVERDOSES WITH DIFFERENT CALCIUM ANTAGONISTS Noncardiovascular manifestations of calcium locker toxicity include lethargy an d depressed consciousness (most common), generalized seizures, and hyperglycemia (caused y inhi ition of insulin release, which is calcium-dependent) (34,35). TREATMENT There are two approaches to calcium channel lockade (29). The first involves th e administration of calcium to antagonize the lockade on the outer surface of t he cell mem rane. The second involves the use of drugs that activate the cAMP pa thway, which antagonizes the lockade from the inner surface of the cell mem ran e. P.833 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 36)03-May-05 07:54:17

Ovid: ICU Book Intravenous Calcium Intravenous calcium is the traditional first-line therapy for reversing calcium channel lockade and elicits favora le responses in 35 to 75% of cases (34,35 an d 36). As descri ed in Chapter 43, there are two calcium salts for intravenous u se (calcium chloride and calcium gluconate), and equivalent amounts of each salt do not contain equivalent amounts of elemental calcium. This is shown in Ta le 53.5; 1 g (1 ampule) of 10% calcium chloride contains roughly three times as muc h elemental calcium as 1 g (1 ampule) of 10% calcium gluconate. Therefore, the c ommon practice of ordering calcium y ampules, without identifying the calcium s alt, is inappropriate. TABLE 53.5. ANTIDOTE THERAPY WITH INTRAVENOUS CALCIUM Although the calcium dosage varies widely in clinical reports, calcium is most e ffective when given in dosages that increase the serum calcium level (35). The d osage of calcium for reversing calcium channel lockade shown in Ta le 53.5 shou ld e sufficient to raise the serum calcium level. The response to calcium may l ast only 10 to 15 minutes, so the initial response to calcium should e followed y a continuous infusion at 0.3 to 0.7 mEq/kg/hour (29). Calcium infusions are not recommended for patients eing treated with digitalis. Glucagon Glucagon is emerging as a promising antidote for calcium channel lockade (37). However, the experience with glucagon is limited to individual case reports (38, 39 and 40). There is some evidence that the effective dosage of glucagon may e higher for calcium channel lockade than for -receptor lockade (38). However, un til further studies are availa le, the http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 36)03-May-05 07:54:17

Ovid: ICU Book recommended dosage of glucagon for - lockade (Ta le 53.3) is a reasona le estimat e of the effective dosage for calcium channel lockade. Catecholamines A variety of catecholamines have een used to antagonize calcium channel lockad e, ut dopamine (starting at 5 g/kg/minute) seems to produce the most favora le r esponses (29). Cases that are refractory to dopamine can e treated with high-do se epinephrine (starting at 1 g/kg/minute) (29). P.834 Phophodiesterase Inhi itors The phosphodiesterase inhi itor amrinone can add to the effects of glucagon and catecholamines in reversing the cardiovascular toxicity of calcium antagonists ( 40,41). However, the experience with amrinone is limited to case reports at pres ent. PREVENTIVE THERAPY Hypotension is a common complication of therapy with intravenous verapamil for s upraventricular tachyarrhythmias. The initial recommendation for eliminating thi s risk was to pretreat with 1 ampule of 10% calcium chloride (13.6 mEq calcium). However, this dose of calcium can produce transient sinus radycardia and atrio ventricular lockade (42). More recent experience indicates that pretreatment wi th one-third of an ampule of calcium chloride or 1 ampule of 10% calcium glucona te (4.6 mEq calcium) is effective in preventing verapamilinduced hypotension in most cases (Ta le 53.5) (36). DIGITALIS Digoxin is the fifth most commonly prescri ed medication in the United States (4 3), and toxic reactions are reported in up to 25% of patients receiving the drug (44). It is the second leading cause of adverse drug reactions in hospitalized patients (45) and the seventh leading cause of death from pharmaceutical exposur es (4). TOXIC MANIFESTATIONS Neurotoxicity http://gateway.ut.ovid.com/gw1/ovidwe .cgi (16 of 36)03-May-05 07:54:17

Ovid: ICU Book The neurologic manifestations of digitalis toxicity include lethargy and depress ed consciousness, delirium, psychosis with hallucinations, and seizures (46). Vi sual distur ances, including decreased acuity, scotomata, and the appearance of colored halos around lights, are reported in over 90% of cases (44). Anorexia, n ausea, and vomiting, which originate from chemoreceptor trigger zones in the low er rainstem, are each reported in over 40% of cases of digitalis toxicity (44). Cardiotoxicity Digitalis toxicity can produce a variety of cardiac rhythm distur ances. The spe ctrum includes sinus arrest, sinoatrial lock, Mo itz type I AV lock, ectopic a trial tachycardia with AV lock, atrial fi rillation with AV junctional escape, AV junctional tachycardia (more than 80 eats/ minute), AV junctional exit lock , AV dissociation, ventricular ectopic depolarizations, ventricular tachycardia, and ventricular fi rillation (47). None of these rhythm distur ances are specif ic for P.835 digitalis toxicity. The one that is most predictive of digitalis to xicity is atrial fi rillation with a junctional rhythm (i.e., atrial fi rillatio n with a regular RR interval). Others A dominal pain and diarrhea are reported in 65% and 40% of cases of digitalis to xicity, respectively (44). Severe hyperkalemia is an uncommon ut life-threateni ng manifestation of digitalis toxicity, and can e resistant to conventional the rapy (48). CONVENTIONAL MANAGEMENT Activated Charcoal As much as 30% of a digoxin dose (oral or intravenous) is excreted in the ile, and repeated doses of activated charcoal (1 g/kg every 4 hours) enhances clearan ce of digoxin from the loodstream (49), and may e particularly valua le in aid ing digoxin clearance in patients with renal insufficiency. However, digoxin has a large volume of distri ution (6 to 7 L/kg in adults) and attempts to enhance drug elimination y any means (e.g., hemodialysis and gastrointestinal dialysis) have limited success (44). Magnesium http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 36)03-May-05 07:54:17

Ovid: ICU Book Digitalis cardiotoxicity is elieved to e the result of inhi ition of the magne sium-dependent mem rane pump for sodium and potassium. Therefore, depletion of m agnesium, which is common in patients receiving digitalis, can promote digitalis cardiotoxicity. Magnesium deficiency also promotes potassium depletion, which i s another predisposing factor for digitalis cardiotoxicity. For these reasons, i ntravenous magnesium is recommended for digitalis cardiotoxicity (except radyca rdia and prolonged AV conduction), particularly when accompanied y hypokalemia. Dosing Regimen. Give 2 g magnesium intravenously over 20 minutes and follow with a continuous in fusion of 1 to 2 g/hour to maintain a serum magnesium of 4 to 5 mEq/L. Magnesium infusion is not recommended for patients with renal insufficiency (for more on magnesium, see Chapter 42). Antiarrhythmic Agents Dilantin and lidocaine are favored for suppressing digitalis-induced ventricular ectopic rhythms ecause they do not prolong AV conduction. Dilantin can actuall y enhance AV conduction, ut lidocaine is often preferred ecause it is given y continuous infusion and is easily titrated. P.836 Lidocaine. Load with 1 to 2 mg/kg intravenously and follow with a continuous infusion at 1 to 4 mg/minute. Dilantin. Infuse at 25 to 50 mg/minute until arrhythmia is suppressed or the total dose is 15 mg/kg. Follow with oral therapy at 400 to 600 mg/day. Quinidine and procaina mide are contraindicated ecause they depress AV conduction (quinidine also elev ates serum digoxin levels), and retylium is contraindicated ecause it can e p roarrhythmic (44). Atropine (0.5 to 1 mg as an intravenous olus, repeated every 5 to 10 minutes as needed) can e used to reverse mild forms of digitalis-induc ed radyarrhythmias (e.g., sinus radycardia and low-grade AV lock). Cardiac pa cing is indicated for complete heart lock. Temporary external pacing (pending i mmunotherapy) may e preferred to transvenous pacing ecause of http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 36)03-May-05 07:54:17

Ovid: ICU Book the myocardial irrita ility that accompanies digitalis toxicity (44). IMMUNOTHERAPY Digoxin-specific anti ody fragments (Digi ind, Burroughs Wellcome) were introduc ed for clinical use in 1986 and have proven effective in life-threatening cases of digoxin toxicity (48,49). Indications The indications for digoxin-specific anti ody fragments (Fa ) are listed in Ta l e 53.6. All of the conditions in this ta le are life-threatening, and many are r esistant to therapy with conventional modalities. The largest clinical experienc e with anti ody therapy has een in cases of hyperkalemia, high-grade AV lock, and malignant ventricular arrhythmias. In these conditions, immunotherapy has el icited favora le responses in 90% of cases (50). TABLE 53.6. INDICATIONS FOR IMMUNOTHERAPY OF DIGITALIS TOXICITY P.837 Dosage Recommendations The dosage of anti ody fragments is determined according to the relationships sh own in Ta le 53.7. Each vial of Digi ind contains 40 mg of anti ody fragments, a nd inds 0.6 mg of digoxin. The total ody content of digoxin (estimated using t he digoxin dose or the serum digoxin level) http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 36)03-May-05 07:54:17

Ovid: ICU Book is used to determine the appropriate dose of anti ody fragments. If neither the dose nor the serum digoxin level is known, anti ody fragments can e given in th e empiric dose shown in Ta le 53.7. TABLE 53.7. DOSING RECOMMENDATIONS FOR DIGOXIN-SPECIFIC ANTIBODY FRAGMENTS (FAB) The dose of anti ody fragments is usually infused over 30 minutes ut can e giv en y rapid intravenous injection (48). Because the response is not immediate, olus injection may e preferred when a rapid response is desira le. Most respons es are evident within 1 hour after drug administration, and peak responses requi re another 3 hours to develop (50). Plasma Digoxin Levels An increase in plasma digoxin levels occurs in all patients who receive digoxinspecific anti ody fragments. The anti ody fragments have such a high affinity fo r digoxin that they displace digoxin from tissue inding sites, and this raises the amount of Fa - ound digoxin in extracellular fluid. Within minutes after ant i ody administration, plasma levels of digoxin (most ound to Fa ) can e 20 tim es higher than pretreatment levels (51). These levels gradually decline over the next few days, as Fa fragments are cleared y the kidneys. The clinical la ora tory uses a radioimmunoassay that measures oth free and ound forms of digoxin, so monitoring plasma digoxin levels is misleading in the first few days followi ng anti ody therapy. Financial Burden http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 36)03-May-05 07:54:17

Ovid: ICU Book Therapy with anti ody fragments is extremely costly, as indicated in Ta le 53.7. As a result, immunotherapy should e limited to cases in which it is clearly in dicated. P.838 NITROPRUSSIDE Nitroprusside is a rapidly acting vasodilator whose therapeutic uses and potenti al for toxicity are descri ed in Chapter 18. The toxic potential of nitroprussid e is a reflection of its cyanide content, which accounts for 44% of the molecula r weight of nitroprusside (52,53 and 54). Nitroprusside-induced cyanide intoxica tion may e responsi le for over 1000 deaths each year (52), and this has led to the recommendation that nitroprusside e a andoned for clinical use (52). CYANIDE METABOLISM The nitroprusside molecule is a ferricyanide complex that contains five cyanide atoms and a nitrosyl group ound to iron in the oxidized (Fe-III) state. The mol ecule is disrupted in circulating lood, releasing the nitrosyl group as nitric oxide, which mediates the vasodilator actions of nitroprusside. The fate of the free cyanide that is released is shown in Figure 53.4 (53,54 and 55). When cyani de com ines with the oxidized iron in cytochrome oxidase, it locks the use of o xygen and inhi its the production of high-energy phosphate compounds. Two chemic al reactions prevent cyanide from reacting with cytochrome oxidase. One involves the inding of cyanide to the oxidized iron in methemoglo in. The other reactio n involves the transfer of sulfur from a donor molecule (thiosulfate) to cyanide to form a thiocyanate compound, which is then cleared y the kidneys. The latte r transulfuration reaction is the principal mechanism for removing cyanide from the human ody. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 36)03-May-05 07:54:17

Ovid: ICU Book

Capacity for Cyanide Removal Healthy adults of average size have enough methemoglo in to ind the cyanide in 18 mg of nitroprusside, and enough thiosulfate to ind the cyanide in 50 mg of n itroprusside (53). When taken together, the human ody has the capacity to detox ify 68 mg of nitroprusside. At a nitroprusside infusion of 2 g/kg/minute (therape utic range) in an 80-kg adult, this 68 mg capacity is reached 500 minutes (8.3 h ours) after the onset of infusion. This illustrates the limited capacity of the human ody to detoxify nitroprusside, even under ideal conditions. The removal o f cyanide is further limited y thiosulfate depletion, which is common in smoker s and postoperative patients (53,54). http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 36)03-May-05 07:54:17

Figure 53.4. The fate of free cyanide released planation.

y nitroprusside. See text for ex

Ovid: ICU Book CYANIDE INTOXICATION The limited capacity of the human ody to remove cyanide was not appreciated whe n nitroprusside was introduced for clinical use in 1976 (53). As a result, cyani de accumulation is common during nitroprus-side P.839 infusions, even in low the rapeutic doses (e.g., 1 g/kg/minute) (53,54). The clinical manifestations of cyan ide intoxication are listed in Ta le 53.8. One of the early signs of cyanide acc umulation is nitroprusside tachyphylaxis (53). Signs of impaired oxygen use (i.e ., a decrease in oxygen extraction ratio and lactic acidosis) often do not appea r until the late stages of cyanide intoxication (55). Lactic acidosis P.840 is n ot a sensitive marker of cyanide intoxication (53,54), and the a sence of lactic acidosis during nitroprusside infusion does not exclude the possi ility of cyan ide accumulation. TABLE 53.8. CYANIDE INTOXICATION: DIAGNOSIS AND THERAPY Diagnosis Whole lood cyanide levels can e used to document cyanide intoxication, as show n in Ta le http://gateway.ut.ovid.com/gw1/ovidwe .cgi (23 of 36)03-May-05 07:54:17

Ovid: ICU Book 53.8. However, the results of cyanide assays are not immediately availa le (a ST AT specimen usually requires 3 to 4 hours to e processed) (54), so immediate de cisions regarding cyanide intoxication are often ased on the clinical presentat ion. Nitroprusside tachyphylaxis is an important early marker of cyanide accumul ation. Prevention Because cyanide intoxication is common and often overlooked, measures aimed at p reventing cyanide accumulation during nitroprusside therapy are particularly val ua le. The most effective preventive measure is to avoid the use of nitroprussid e. Another effective measure is adding thiosulfate to the nitroprusside infusate , using the doses presented in Chapter 18 (54). This latter measure should e a mandatory practice for nitroprusside infusions. Therapy Treatment of cyanide intoxication should egin with the inhalation of 100% oxyge n. The Cyanide Antidote Kit (Eli Lilly & Co.) can e used as descri ed in Ta le 53.8 (56). Nitrites promote the oxidation P.841 of hemoglo in to methemoglo in a nd promote cyanide inding to methemoglo in. The cyanide ound to methemoglo in must eventually e eliminated from the ody y transulfuration, so thiosulfate s hould always e given with nitrites. Because nitrite therapy creates methemoglo inemia, alternative methods of cyanide inding have een explored. The affinity of cyanide for co alt has led to the use of hydroxoco alamin (100 mL of a 5% sol ution infused over 15 minutes) (57), which com ines with cyanide to form cyanoco alamin (vitamin B12), which is then excreted in the urine. This strategy is pop ular in Europe, ut the lack of a suita le hydroxoco alamin preparation has hamp ered its use in the United States. THIOCYANATE INTOXICATION The most important mechanism for cyanide removal involves the formation of thioc yanate, which is slowly excreted in the urine. When renal function is impaired, thiocyanate can accumulate and produce a toxic syndrome that is distinct from cy anide intoxication (53,54). The clinical features include anxiety, confusion, pu pillary constriction, tinnitus, hallucinations, and generalized seizures (53,54) . Thiocyanate can also produce hypothyroidism y locking the thyroid uptake of iodine (54). The diagnosis of thiocyanate toxicity is esta lished y the serum t hiocyanate level. Normal http://gateway.ut.ovid.com/gw1/ovidwe .cgi (24 of 36)03-May-05 07:54:17

Ovid: ICU Book levels are elow 10 mg/L, and clinical toxicity is usually accompanied y levels a ove 100 mg/ L (54). Thiocyanate intoxication can e treated y hemodialysis o r peritoneal dialysis. OPIOIDS The opioids are common offenders in overdoses involving illicit street d the opioid analgesic morphine is the most common cause of toxic drug involving hospitalized patients (45). The adverse side effects of the algesics are descri ed in Chapter 8. The following description focuses eatment of opioid intoxication with the opioid antagonist naloxone.

NALOXONE Naloxone is a pure opioid antagonist that inds to endogenous opioid receptors ut does not elicit any agonist responses. It is most effective in locking -recep tors (analgesia, respiratory depression) and -receptor ( edation, pupillary con triction), and lea t effective in bloc ing receptor (delirium, hallucination ). Route of Admini tration Naloxone (0.4 mg/mL or 1 mg/mL) i u ually given intravenou ly (58), but it can al o be given endotracheally (59), by intralingual injection, P.842 or by ubmen tal injection (60,61). The ubmental injection of naloxone i ea ily performed w ith a 23-gauge, 1.5-in. needle in erted in the midline about halfway between the mandible and thyroid cartilage. The needle i directed uperiorly and advanced approximately 1 in. to ma e the injection. The drug i introduced into the mu cl e at the ba e of the tongue and i ab orbed into the ublingual circulation. Th i approach i afe and can elicit a re pon e in le than 2 minute (61).

Depre ed mental tate. For patient with a depre ed en orium but no re piratory depre ion, the initi al do e of naloxone hould be 0.4 mg intravenou pu h. Thi can be repeated in 2 minute , if nece ary. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (25 of 36)03-May-05 07:54:17

Do ing Recommendation In opioid overdo e, rever al of the edation u ually require aloxone than rever al of the re piratory depre ion.

maller do e of n

drugs, an reactions opioid an on the tr

A total do e of 0.8 mg hould be effective if the mental tatu change are cau ed by an opioid derivative (62). In patient with nown opioid dependency, the b olu do e of naloxone hould be reduced to 0.1 or 0.2 mg (62). Re piratory depre ion. For patient who have evidence of re piratory depre ion the initial do e of nal oxone hould be 2 mg intravenou pu h. Thi do e i repeated every 2 minute if nece ary, to a total do e of 10 mg (58,62). The effect of naloxone la t approx imately 60 to 90 minute , which i le than the duration of action of mo t opio id . Therefore, a favorable re pon e to naloxone hould be followed by repeat do e at 1-hour interval or by a continuou infu ion. For a continuou naloxone i nfu ion, the hourly do e of naloxone hould be two-third of the effective bolu do e (a 6hour do e can be added to 250 or 500 mL of aline and the olution inf u ed over 6 hour ) (63). To achieve teady- tate drug level in the early infu i on period, a econd bolu of naloxone (at one-half the original bolu do e) i g iven 30 minute after the infu ion i tarted. The duration of the infu ion vari e (according to the drug and the do e inge ted), but the average duration i 10 hour (62). Empiric Therapy Patient with a depre ed mental tate of undetermined etiology are often given naloxone (0.8 to 2 mg intravenou pu h) a empiric therapy. Thi practice ha be en que tioned becau e it elicit a favorable re pon e in fewer than 5% of ca e (64). An alternative approach ha been propo ed where empiric naloxone i indica ted only for patient with pinpoint pupil who have circum tantial evidence of o pioid abu e (e.g., needle trac ) (62,64). When naloxone i u ed in thi manner, a favorable re pon e i expected in approximately 90% of patient (64). Adver e Reaction Naloxone ha few ide effect . The mo t common adver e reaction i the opioid wi thdrawal yndrome (anxiety, abdominal cramp , vomiting, and piloerection). There are ca e report of acute pulmonary P.843 edema (mo t in the early po toperativ e period) and generalized eizure following naloxone admini tration (62), but t he e complication are rare. TRICYCLIC ANTIDEPRESSANTS http://gateway.ut.ovid.com/gw1/ovidweb.cgi (26 of 36)03-May-05 07:54:17

Ovid: ICU Boo

Overdo e of tricyclic antidepre ant (e.g., amitriptyline, de ipramine, doxepin , and imipramine) i the leading cau e of death from pharmaceutical drug overdo e in the United State (4). TOXIC MANIFESTATIONS The toxic effect of the tricyclic antidepre ant re ult primarily from their a nticholinergic action and their ability to bloc the reupta e of neurotran mitt er uch a norepinephrine (65). Seriou toxicity u ually emanate from the cent ral nervou y tem or cardiova cular y tem. Central Nervou Sy tem Early manife tation include fever, agitation, and pupillary dilatation. Advance d ca e are characterized by delirium, coma, and generalized eizure . The neuro toxic manife tation are attributed to anticholinergic effect . Cardiova cular Early ign include tachycardia and hyperten ion, which are attributed to bloc a de of norepinephrine reupta e. Depletion of norepinephrine in nerve terminal le ad to po tural and upine hypoten ion. Cardiac conduction abnormalitie (prolon ged QRS interval), arrhythmia (ventricular ectopic , wide-complex tachycardia ) , and a decrea ed cardiac output are een in advanced ca e . CLINCAL EVALUATION Admi ion to the ICU i warranted for patient with life-threatening complicatio n , uch a delirium, coma, eizure , cardiac arrhythmia , or hypoten ion. Toxic manife tation often require 6 to 12 hour to become apparent, and in the early period after drug inge tion, the principal ta i to identify patient who are li ely to develop eriou toxicity. Pla ma drug level can be u eful in confirm ing the diagno i of tricyclic overdo e but not in predicting progre ion to er iou toxicity (65). The electrocardiogram ha the greate t predictive value in a ymptomatic patient . Prolongation of the QRS interval can be a progno tic ign for progre ion to eizure (QRS above 0.10 econd) and eriou arrhythmia (QRS above 0.16 econd) (66). Therefore, a ymptomatic patient with a prolonged QRS interval u ually are admitted for ob ervation (admi ion to a telemetry area i appropriate). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (27 of 36)03-May-05 07:54:17

Ovid: ICU Boo

Seizure Benzodiazepine u ually are effective in treating generalized eizure in tricyc lic overdo e. Diazepam (5 to 10 mg) and lorazepam (1 to 2 mg) are equally effect ive. Seizure refractory to the e agent have re ponded to midazolam (10 mg intr avenou ly, then 6 mg/hour) (67). Barbiturate can al o be u ed for refractory e izure , but phenytoin ha limited efficacy (65). Arrhythmia The treatment of choice for eriou arrhythmia in tricyclic overdo e i al alin ization of the blood with odium bicarbonate (68). Al alinization rever e the c ardiotoxic action of the tricyclic by an un nown mechani m. The goal of al ali nization i a erum pH of 7.50 to 7.55, which can be achieved by intermittent bo lu do e of odium bicarbonate at a do age of 1 to 2 mEq/ g. Al alinization ha al o been accompanied by an improved mental tatu and re olution of hypoten io n in tricyclic overdo e (68). Hypoten ion Hypoten ion in tricyclic overdo e can be the re ult of cardiac dy function, peri pheral va odilation, or both. Therefore, hypoten ion that doe not re pond to vo lume infu ion hould prompt in ertion of a pulmonary artery catheter to guide ma nagement. If the problem i a cardiac dy function, dobutamine hould be u ed; if the problem i a low y temic va cular re i tance, norepinephrine hould be u e d. Dopamine can promote arrhythmia in tricyclic overdo e and hould be avoided (65).

http://gateway.ut.ovid.com/gw1/ovidweb.cgi (28 of 36)03-May-05 07:54:17

REFERENCES INTRODUCTION 1. Da ta JF. Drug pre cribing i ue in the inten ive care unit: finding an wer to common que tion . Crit Care Med 1994;22:909912.

MANAGEMENT Management of tricyclic overdo e in the ICU u ually involve rrhythmia , or hypoten ion. P.844

eizure , cardiac a

Ovid: ICU Boo

3. Jan el CA, Speedie SM. Detecting drug interaction : a review of the literatur e. DICP Ann Pharmacother 1990;24:982989. ACETAMINOPHEN 4. Litovitz TL, Holm KC, Clancy C, et al. 1992 Annual Report of the American A ociation of Poi on Control Center Toxic Expo ure Surveillance Sy tem. Am J Emer g Med 1993;11:494555.

6. Jane J, Routledge PA. Recent development in the management of paracetamol ( acetaminophen) poi oning. Drug Safety 1992;7:170177 7. Whitcomb DC, Bloc GD. A ociation of acetaminophen hepatotoxicity with fa ti ng and ethanol u e. JAMA 1994;272:18451850. P.845 8. Dequay B, Malinverni R, Lauterburg BH. Glutathione depletion in HIV-infected patient : role of cy teine deficiency and effect of oral N-acetylcy teine. AIDS 1992;6:815 820. 9. Briti h National Formulary No. 27. London: Briti h Medical A ociation, 1994; 2021. 10. Holdine MR. Clinical pharmaco inetic of N-acetylcy teine. Clin Pharmaco i net 1991;20:123134. 11. Smil tein MJ, Knapp GL, Kulig KW, Rumac BH. Efficacy of oral N-acetylcy te ine in http://gateway.ut.ovid.com/gw1/ovidweb.cgi (29 of 36)03-May-05 07:54:17

5. An er AL, Smil tein MJ. Acetaminophen: concept Clin North Am 1994;12:335350.

and controver ie . Emerg Med

2. Trunet P, Borda IT, Rouget AV, et al. The role of drug-induced illne i ion to an inten ive care unit. Inten ive Care Med 1986;12:4346.

in adm

Ovid: ICU Boo

the treatment of acetaminophen overdo e. Analy i of the National Multicenter St udy. N Engl J Med 1988;319:15571562. 12. Harri on PM, Keay R, Bray GP, et al. Improved outcome of paracetamol-induce d fulminant hepatic failure by late admini tration of N-acetylcy teine. Lancet 1 990;335:15721573.

15. Spiller HA, Krenzelo EP, Grande GA, et al. A pro pective evaluation of the effect of activated charcoal before oral N-acetylcy teine in acetaminophen overd o e. Ann Emerg Med 1994;23:519523. BENZODIAZEPINES 16. Da ta JF, Fuhrman TM, McCandle C. Pattern of pre cribing and admini tering drug for agitation and pain in a urgical inten ive care unit. Crit Care Med 1 994;22:974 980. 17. Han en-Fla chen JH, Brazin y S, Ba ile C, Lan en PN. U e of edating drug and neuromu cular bloc ing agent in patient requiring mechanical ventilation f or re piratory failure. JAMA 1991;266:28702875. 18. Gaudreault P, Guay J, Thivierge RL, Verdy I. Benzodiazepine poi oning. Drug Safety 1991;6:247265. 19. Bodenham A. Rever al of prolonged edation u ing flumazenil in critically il l patient . Ane the ia 1989;44:603605. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (30 of 36)03-May-05 07:54:17

14. Sunman W, Hughe A, Sever P. Anaphylactoid re pon e to intravenou acetylcy teine. Lancet 1992;339:12311232.

13. Miller LF, Rumac BH. Clinical afety of high do e Oncol 1983;10(Suppl 1):7685.

of acetylcy teine. Semin

Ovid: ICU Boo

20. Ritz R, El a er S, Schwander J. Controlled edation in ventilated inten ive care patient . Re u citation 1988;16(Suppl):S83S89. 21. Pepperman ML. Double-blind tudy of the rever al of midazolam-induced edati on in the inten ive care unit with flumazenil (RO 15-1788): effect on weaning fr om ventilator. Ane th Inten ive Care 1990;18:3844. 22. Breheny FX. Rever al of midazolam edation with flumazenil. Crit Care Med 19 92;20:736739. 23. Howland MA. Flumazenil. In: Goldfran LR, Flomenbaum NE, Lewin NA, et al., e d . Goldfran ' toxicologic emergencie . Norwal , CT: Appleton & Lange, 1994;8058 10. 24. Shalan y SJ, Naumann TL, Englander FA. Therapy update: effect of flumazenil on benzodiazepine-induced re piratory depre ion. Clin Pharmacol 1993;12:483487. 25. Marten F, Koppel C, Ibe K, et al. Clinical experience with the benzodiazepi ne antagoni t flumazenil in u pected benzodiazepine or ethanol poi oning. J Tox icol Clin Toxicol 1990;28:341356.

27. Chern TL, Hu SC, Lee CH, Deng JF. Diagno tic and therapeutic utility of flum azenil in comato e patient with drug overdo e. Am J Emerg Med 1993;11:122124. 28. Haver o GP, DiSalvo RP, Imhoff TE. Fatal eizure after flumazenil admini t ration in a patient with mixed overdo e. Ann Pharmacother 1994;28:13471349. P.846 BETA RECEPTOR ANTAGONISTS 29. Kern W II, Kline J, Ford MD. Beta bloc er and calcium channel bloc er toxic ity. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (31 of 36)03-May-05 07:54:17

26. Doyon S, Robert 1994;12:301316.

JR. Reapprai al of the coma coc tail. Emerg Med Clin North Am

Ovid: ICU Boo

Emerg Med Clin North Am 1994;12:365390. 30. Wein tein RS. Recognition and management of poi oning with beta-adrenergic b loc ing drug . Ann Emerg Med 1984;13:11231131. 31. Brancato DJ. Recognizing potential toxicity of phenol. Vet Hum Toxicol 1982; 24:29 30. 32. Chernow B, Zaloga G, Malcolm D, et al. Glucagon' chronotropic action i cal ciumdependent. J Pharmacol Exp Ther 1987;241:833837. 33. Lucche i BR. The role of calcium in excitationcontraction coupling in cardiac and va cular mooth mu cle. Circulation 1989;80:IV1IV10. CALCIUM ANTAGONISTS 34. Ramo a EA, Spiller HA, Winter M, Bory D. A one-year evaluation of calcium channel bloc er overdo e : toxicity and treatment. Ann Emerg Med 1993;22:196200. 35. Ramo a EA, Spiller HA, Myer A. Calcium channel bloc er toxicity. Ann Emerg Med 1990;19:649653. 36. Jame on SJ, Hargarten SW. Calcium pretreatment to prevent verapamil-induced hypoten ion in patient with SVT. Ann Emerg Med 1992;21:84 (editorial). 37. Zarit y A, Morowicz M, Chernow B. Glucagon antagoni m of calcium bloc erind uced myocardial dy function. Crit Care Med 1988;16:246251. 38. Doyon S, Robert J. The u e of glucagon in a ca e of calcium channel bloc er overdo e. Ann Emerg Med 1993;22:12291233.

http://gateway.ut.ovid.com/gw1/ovidweb.cgi (32 of 36)03-May-05 07:54:17

39. Walter FG, Frye G, Mullen JT, et al. Amelioration of nifedipine poi oning a ociated

Ovid: ICU Boo

with glucagon therapy. Ann Emerg Med 1993;22:12341237. 40. Wolf LR, Spadafora MP, Otten EJ. U e of amrinone and glucagon in a ca e of c alcium channel bloc er overdo e. Ann Emerg Med 1993;22:12251228. 41. Goenen M, Col J, Compere A, et al. Treatment of evere verapamil poi oning w ith combined amrinone-i oproterenol therapy. Am J Cardiol 1986;58:11421143. 42. Kuhn M. Severe bradyarrhythmia following calcium pretreatment. Am Heart J 1 991;121:18131814. DIGITALIS 43. Red Boo . 100th ed. Montvale, NJ: Medical Economic , 1996;98. 44. Kri anda TJ. Digitali toxicity. Po tgrad Med 1992;91:273284. 45. Evan PS, Pe totni SL, Cla en DC, et al. Preventing adver e drug event in ho pitalized patient . Ann Pharmacother 1994;28:523527. 46. McDonnell Coo e D. The u e of central nervou y tem manife tation in the e arly detection of digitali toxicity. Heart Lung 1993;22:477481.

48. Martiny SS, Phelp SJ, Ma ey KL. Treatment of digitali intoxication with d igoxin pecific antibody fragment : a clinical review. Crit Care Med 1988;16:62963 5. 49. Lalonde RL, De hpande R, Hamilton PP, et al. Acceleration of digoxin clearan ce by activated charcoal. Clin Pharmacol Ther 1985;37:367371. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (33 of 36)03-May-05 07:54:17

47. Moorman JR, Pritchett ELC. The arrhythmia ntern Med 1986;145:12891292.

of digitali intoxication. Arch I

Ovid: ICU Boo

50. Antmann EM, Wenger TL, Butler VP, et al. Treatment of 150 ca e of life-thre atening digitali intoxication with digoxin- pecific Fab antibody fragment . Cir culation 1990;81:17441752. P.847 51. Ujhelyi MR, Robert S, Cumming DM, et al. Influence of digoxin immune Fab th erapy and renal dy function on the di po ition of total and free digoxin. Ann In tern Med 1993;119:273277. NITROPRUSSIDE 52. Robin ED, McCauley R. Nitropru ide-related cyanide poi oning. Time (long pa t due) for urgent, effective intervention . Che t 1992;102:18421845. 53. Curry SC, Arnold-Capell P. Nitropru ide, nitroglycerin, and angioten in-con verting enzyme inhibitor . In: Blumer JL, Bond GR, ed . Toxic effect of drug u ed in the ICU. Critical care clinic . Philadelphia:WB Saunder , 1991;555582. 54. Hall VA, Gue t JM. Sodium nitropru ide-induced cyanide intoxication and pre vention with odium thio ulfate prophylaxi . Am J Crit Care 1992;2:1927. 55. Arieff AI. I mea urement of venou oxygen aturation u eful in the diagno i of cyanide poi oning? Am J Med 1992;93:582583. 56. Kir MA, Gerace R, Kulig KW. Cyanide and methemoglobin inetic in mo e inh alation victim treated with the Cyanide Antidote Kit. Ann Emerg Med 1993;22:141 3 1418. 57. Curry SC, Connor DA, Ra ch e RA. Effect of the cyanide antidote hydroxocobal amin on commonly ordered erum chemi try tudie . Ann Emerg Med 1994;24:6567. OPIOIDS http://gateway.ut.ovid.com/gw1/ovidweb.cgi (34 of 36)03-May-05 07:54:17

Ovid: ICU Boo

58. Naloxone hydrochloride. AHFS Drug Information 95. Bethe da, MD: American Soc iety of Ho pital Sy tem Pharmaci t , 1995;14181420. 59. Tandleberg D, Abercrombie D. Treatment of heroin overdo e with endotracheal naloxone. Ann Emerg Med 1982;11:443445. 60. Maio RF, Gau el B, Freeman B. Intralingual naloxone injection for narcotic-i nduced re piratory depre ion. Ann Emerg Med 1987;16:572573. 61. Salvucci AA Jr, Ec tein M, I covich AI. Submental injection of naloxone. An n Emerg Med 1995;25:719720.

64. Hoffman JR, Schriger DL, Luo JS. The empiric u e of naloxone in patient wit h altered mental tatu : a reapprai al. Ann Emerg Med 1991;20:246252. TRICYCLIC ANTIDEPRESSANTS 65. Wei man R, Howland MA, Hoffman RS, Cohen H. et al. Cyclic antidepre ant . I n: Goldfran LR, Flomenbaum NE, Lewin NA, et al., ed . Goldfran ' toxicologic e mergencie . 5th ed. Norwal , CT: Appleton & Lange, 1994;725734. 66. Boehnert M, Lovejoy FH Jr. Value of the QRS duration ver u the erum drug l evel in predicting eizure and ventricular arrhythmia after acute overdo e of tricyclic antidepre ant . N Engl J Med 1985;313:474479.

http://gateway.ut.ovid.com/gw1/ovidweb.cgi (35 of 36)03-May-05 07:54:17

67. Kumar A, Blec TP. Intravenou midazolam for the treatment of refractory tu epilepticu . Crit Care Med 1992;20:483488.

63. Goldfran L, Wei man RS, Erric JK, et al. A do ing nomogram for continuou infu ion intravenou naloxone. Ann Emerg Med 1986;15:566569.

62. Doyon S, Robert 01316.

J. Reapprai al of the coma coc tail. Emerg Med Clin 1994;12:3

ta

Ovid: ICU Boo

68. Hoffman JR, Votey SR, Bayer M, Silver L. Effect of hypertonic odium bicarbo nate in the treatment of moderate to evere cyclic antidepre ant overdo e. Am J Emerg Med 1993;11:336341.

http://gateway.ut.ovid.com/gw1/ovidweb.cgi (36 of 36)03-May-05 07:54:17

SUGGESTED READING Goldfran LR, Flomenbaum NE, Lewin NA, et al., ed . Goldfran ' gencie . 5th ed. Norwal , CT: Appleton & Lange, 1994.

toxicologic emer

Ovid: ICU Boo

Copyright 1998 Lippincott William & Wil in Marino, Paul L. ICU Boo , 2nd Editio n Chapter 54 DRUG DOSING ADJUSTMENTS IN THE ICU Thi final chapter i a compilation of ome of the important drug do ing adju tm ent mentioned throughout the boo . The information i pre ented in tabular form , with each drug or group of drug li ted in alphabetical order. Included for ea ch drug i the normal or u ual do age, along with pertinent do age adju tment f or drug interaction , renal in ufficiency, hepatic in ufficiency, and advanced a ge. The ource for the recommended do age alteration are li ted in the bibliog raphy at the end of thi chapter. The information in thi chapter i by no mean exhau tive (a thi would require a eparate text of it own), but it doe prov ide a quic reference of do age adju tment for drug that are commonly u ed in the ICU. Mo t of the recommendation in thi chapter are for parenteral drug the rapy. GLOMERULAR FILTRATION RATE Many of the do e alteration in thi chapter are ba ed on the glomerular filtrat ion rate (GFR). Thi can be e timated in adult u ing the equation hown below ( ee Coc roft and Gault, 1976).

Note the additional factor of 0.85 for women, indicating that the GFR in women i 85% of the GFR in men. The body weight in thi calculation hould be the ideal or lean body weight ( ee the Reference Table in the Appendix for a li t of ide al body weight in adult men and women). However, if a patient ha a body weight that i below the ideal weight, the actual body weight hould be u ed in the ca lculation ( ee Robert et al, 1993). E timate of GFR ba ed on the Coc croftGault P.849 equation hown here are reported to be more accurate than e timate of GFR ba ed on urinary creatinine clearance in critically ill patient ( ee Robert e t al, 1993). http://gateway.ut.ovid.com/gw1/ovidweb.cgi (1 of 15)03-May-05 07:51:42

Ovid: ICU Boo

The Elderly Note that the Coc roftGault equation predict that the GFR in adult will decline with advancing age. After the age of 40, the GFR decrea e approximately 1 mL/m inute each year ( ee Zawada and Boice, 1993). Thi mean that a 70-year-old pati ent will have a GFR that i 30 mL/minute le than that of a 40-year-old patient . Thi age-related decline in GFR can influence the do ing guideline for drug that are eliminated by the idney . Therefore, the GFR i an important con idera tion for do ing adju tment in elderly patient a well a in patient with rena l di ea e. Drug Normal Do e

Recommendation Avoid acetaminophen. Increa ed ri of hepatotoxicity from acetam inophen. Avoid acetaminophen if po ible, or u e with extreme caution. Acetaminophen 650 mg q6hr Hepatic in ufficiency Patient with: HIV infection Malnutrition Chronic ETOH GFR <30 mL/min Acetazolamide Adeno ine 250 mg q6hr 612 mg IV Avoid acetazolamide. Reduce do age by 50%. Reduce do age by 50%. Avoid combined treatment. Admini tration through central venou catheter Combined treatment with: Blockers Calcium lockers Dipyridamole Theophylline ( locks adenosine receptors) Aminoglycosides Amikacin Gentamicin To ramycin Amphotericin 23 mg/kg q8hr 11.5 mg/kg q8hr 11.5 mg/kg q8hr GRF 1050 mL/min GFR <10 mL/min

0.51 mg/kg/day GFR <10 mL/min TPN solutions Extend dosage interval to 48 or 72 hr. Do not add to TPN solutions (incompati le ). Ampicillin Drug Amrinone Aztreonam 13 g q6hr Normal Dose 510 g/kg/min GFR <15 mL/min Conditions

Reduce single doses y 30 to 70% and give q1218hr. Reduce single doses give q2448hr.

Condition

y 70% and

Ovid: ICU Boo

Extend dosage interval to 24 hr. Recommendation Reduce dosage rate 25 to 50%. Re duce dosage y 70%. Reduce dosage y 25%. GFRI <10 mL/min GFR <10 mL/min Hepatic insufficiency 12 g q812 hr http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 15)03-May-05 07:51:42

Ovid: ICU Book Benzodiazepines Diazepam Midazolam See Ta le 8.3 and Ta le 8.4. Com ined treatment with: Cimetidine Erythromycin Isoniazid Ketoconazole Metoprol ol Propranolol Valproic acid Com ined treatment with theophylline Com ined treat ment with rifampin O esity Impaired meta olism of diazepam and midazolam may lead to increased dose require ments. Avoid erythromycin midazolam com ination. Otherwise, adjust enzodiazepine dosage only as needed. Avoid theophylline (antagonizes enzodiazepine sedation) if sedation an importan t goal. Enhanced meta olism of midazolam and diazepam may lead to increased dosa ge requirements. Use ideal ody weight for midazolam dosing. Calcium 200 mg IV over 10 minutes Digitalis treatment Digitalis toxicity

Drug Normal Dose 1 g q8hr 2 g q8hr 2 g q8hr Conditions Recommendation Extend dosage interval to 2448 hr. Extend dosage interval to 24 hr . Extend dosage interval to 24 hr. Extend dosage interval to 48 hr. GFR 2050 mL/m in GFR <20 m/min Reduce dosage rate y 25% and do not exceed rate of 67 mg/hr. R educe dosage rate y 75% and do not exceed rate of 33 mg/hr. Cephalosporins Cefazolin Cefotaxime Ceftazidime GFR <10 mL/min GFR <10 mL/min GFR 1050 mL/min GFR <10 mL/min Cimetidine Continuous infusion: Start at 37.5 mg/ hr. Increase in 12.5 mg/hr increments if needed to a maximum of 100 mg/hr. Note:Cimetidine locks renal ine without a change in GFR. floxacin 400 mg IV q12hr GFR ylline Reduce dosage y 50%. creatinine secretion and can increase serum creatin This can cause a falsely low estimate of GFR. Cipro <10 mg/min Com ined treatment with: Coumadin Theoph See recommendations for coumadin and theophylline.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 15)03-May-05 07:51:42

Avoid if possi le, or reduce dosage cium infusions.

y 50% and infuse over 30 minutes. Avoid cal

Ovid: ICU Book Coumadin 510 mg PO daily Com ined treatment with: Amiodarone Ciprofloxacin Cimetidine Erythromycin Flucon azole Isoniazid Metronidazole Quinidine Salicylates Trimothoprim sulfamethoxazole Com ined treatment with: Corticosteroids Estrogen Nafcillin Pheno ar ital Rifam pin Sucralfate Possi le enhanced anticoagulant effect. Avoid com ined treatment with amiodarone , metronidazole, and salicylates. Otherwise, monitor INR. Possi le reduced anticoagulant effect. Monitor INR. Drug Digoxin Normal Dose 0.1250.5 mg qd Conditions Recommendation Possi le enhanced digoxin effect. Reduce digoxin dosage y 50% fo r com ined treatment with amiodarone, quinidine, and verapamil. Otherwise, monit or serum digoxin levels. Com ined treatment with: Amiodarone Quinidine Verapamil Captopril Diltiazem GFR <10 mL/min

Avoid haloperidol. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 15)03-May-05 07:51:42

Note: In renal failure, there may e a spuriously high serum digoxin level y ra dioimmunoassay. Fluconazole 200 to 400 mg/day GFR 2050 mL/min GFR <20 mL/min Com ined treatment with: Phenytoin Warfarin Haloperidol See Ta le 8.6. Prolonged QT Interval Reduce dosage y 50%. Reduce dosage y 75%. See recommendations for phe nytoin and warfarin.

Decrease single dose

y 75% and administer every 48 hr.

Ovid: ICU Book Heparin See Ta le 7.4. Mor id o esity ( ody weight >130 kg) Highdosage nitroglycerin (>350 g/min) Heparin dose may e much smaller than recommended in the standard weight- ased d osing nomogram. May e increased heparin requirement. Monitor PTT. Drug Imipenem Normal Dose 0.51 g q6hr Conditions Recommendation Extend dosage interval to 8 hr. Extend dosage interval to 12 hr. Do not exceed a daily dose of 2 g or 25 mg/kg. Possi le decrease in insulin requ irement. Monitor lood glucose. Expect 20 to 30% loss of insulin dose as a resul t of adsorption to plastic and glass in IV infusion sets. Reduce dosage y 50%. Reduce dosage y 50% ecause of impaired lidocaine meta olism. GFR 2050 mL/min GFR <20 mL/min Seizure disorder Insulin Varia le Treatment with ACE Inhi itors IV fluid containers and infusion sets Ketorolac Lidocaine 15 to 30 mg q6hr 14 mg/min GFR <50 mL/min Hepatic insufficiency Low cardiac output Com ined treatment with: Cimetidine -Blockers Elderly patients Limit lidocaine infusion to less than 12 hr. Recommendation Reduce dosage y 50% . See recommendations for phenytoin. Drug Normal Dose Conditions Metronidazole 7.5 mg/kg q6hr GFR <10 mL/min Com ined treatment with phenytoin Nitroglycerin 10200 g/min Polyvinychloride tu ing Increased intracranial pressure Nitroglycerin a sor s to PVC. Use Nitrostat infusion sets (nonadsor ent material

). Avoid nitroglycerin. Nitroprusside Opioids Meperidine Morphine Fentanyl Pentamidine 0.210 g/kg/min 50100 mg q4hr 16 mg/hr 30100 g/hr 4 mg/kg/day GFR <10 mL/min GFR <50 mL/min GFR <10 mL/min Hepatic insufficiency GFR <10 mL/mi n Avoid nitroprusside. Avoid (neurotoxic meta olite accumulates when renal functio n is impaired). Reduce dosage y 50%. Avoid if possi le. Reduce dosage y 50%. GFR <10 mL/min Extend dosage interval to 48 hr http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 15)03-May-05 07:51:42

Ovid: ICU Book Phenytoin 300400 mg/day Com ined treatment with: Amiodarone Fluconazole Isoniazid Metronidazole Hypoal u minemia Com ined treatment with: Coumadin Diazepam Glucocorticoids Pheno ar ital Theophylline Valproic acid Glucose-containing intravenous fluids Possi le enhanced phenytoin toxicity. Monitor serum levels. Same as a ove. Possi le reduction in phenytoin efficacy. Monitor serum levels. Do not add phenytoin to glucose-containing fluids (incompati le). Drug Normal Dose 26 mg/min Conditions Recommendation Reduce dosage y 25%. Reduce dosage y 50%. Possi le increase in serum procainamide levels. Monitor serum levels. Avoid if possi le. Procainamide Mild heart failure Elderly patients with renal dysfunction Cimetidine treatment GFR <10 mL/min Propofol Ranitidine Theophylline See Ta le 8.4. 50 mg IV q810hr O esity Dose is determined y ideal ody weight. Extend dosage interval to 24 hr. Reduce dosage y 50% and follow serum drug levels. Reduce dosge y 30 to 50% and follo w serum theophylline levels. GFR <50 mL/min Low cardiac output Hepatic insufficiency Com ined treatment with: Cimetidine Ciprofloxacin Erythromycin Enoxacin Propranolol Verapamil Com ined t reatment with: Phenytoin Pheno ar ital O esity See Ta le 25.4. Normal dosing may e su therapeutic. Follow serum theophylline levels.

http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 15)03-May-05 07:51:42

Dosing

y ideal ody weight may e su therapeutic. Follow serum drug levels.

Ovid: ICU Book Thiamine 3100 mg/day Hypomagnesemia Parenteral nutrition Thiamine will e ineffective until magnesium depletion is corrected. Do not add thiamine to amino acid solutions (may e degraded y sulfites in the solutions). Drug Normal Dose Conditions Recommendation GFR 15 to 30 mL/ min GFR <15 mL/min Normal dose for 48 hr, then r educe daily dose y 50% and extend dose interval to 12 hr. Reduce daily dose y 50%, and extend dosage interval to 12 hr. Trimethoprim sulfamethoxazole (TMPSMX) For Pneumocystis carinii pneumonia (PCP): 20 mg TMP/kg/day and 100 mg SMX/kg/day in 3 to 4 divided doses Vancomycin 500 mg IV q6hr GFR 1030 mL/min GFR <10 mL/min Heparin infusion Extend dosage interval to 48 hr. Extend dosage interval to 4 days. Heparin degra des vancomycin. Use a separate IV line. Possi le enhanced negative inotropic and vasodilator effects of verapamil. Avoid com ined treatment if possi le. Verapamil 510 mg IV Com ined treatment with: ACE inhi itors - lockers Quinidine Digoxin treatment See recommendations for digoxin. P.855 REFERENCES GENERAL DRUG INFORMATION SOURCES 1. Young DS. Effects of drugs on clinical la oratory tests. 3d ed. Washington, D C: American Association for Clinical Chemistry, 1990. 2. Rizock MA, Hillman CDM, eds. The Medical Letter hand ook of adverse drug inte ractions. New Rochelle, NY: The Medical Letter, 1991. 3. Schrier RW, Gam ertoglio JG, eds. Hand ook of drug therapy in liver and kidne y disease. Boston: Little, Brown, 1991. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 15)03-May-05 07:51:42

Ovid: ICU Book 4. Trissel LA. Hand ook on injecta le drugs. 8th ed. Bethesda, MD: American Soci ety of Hospital Pharmacists, 1994. 5. Bennett WM, Aronoff GR, Golper TA, et al. Drug prescri ing in renal failure: dosing guidelines for adults. 3d ed. Philadelphia: American College of Physician s, 1994. 6. Chernow B, ed. The pharmacologic approach to the critically ill patient. 3d e d. Baltimore: Williams & Wilkins, 1994. 7. McEvoy GK, ed. AHFS Drug Information, 1995. 38th ed. Bethesda, MD: American S ociety of Health-System Pharmacists, 1995. 8. Semla TP, Beizer JL, Hig ee MD. Geriatric dosage hand ook. 2nd ed. Hudson, OH : Lexi-Comp Inc., 1995. REVIEWS 1. Bass NM, Williams RL. Guide to drug dosage in hepatic disease. Clin Pharmacok inet 1988;15:396420. 2. Zawada ET, Boice JL. Clinical pharmacology in aged intensive care unit patien ts. J Intensive Care Med 1993;8:289297 (38 references). 3. Ku isty CA, Arns PA, Wedlund PJ, Branch RA. Adjustment of medications in live r failure. In: Chernow B, ed. The pharmacologic approach to the critically ill p atient. 3d ed. Baltimore: Williams & Wilkins, 1994;95113 (186 references). 4. Kroh UF. Drug administration in critically ill patients with acute renal fail ure. New Horiz 1995;3:748759 (42 references). 5. Preston L, Briceland LL, Lomaestro BM, et al. Dosing adjustment of 10 antimic ro ials for patients with renal impairment. Ann Pharmacother 1995;29:12021207. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 15)03-May-05 07:51:42

Ovid: ICU Book GLOMERULAR FILTRATION RATE 1. Cockroft DW, Gault MN. Prediction of creatinine clearance from serum creatini ne. Nephron 1976;16:3141. 2. Ro ert S, Zarowitz BJ, Peterson EL, et al. Predicta ility of creatinine clear ance estimates in critically ill patients. Crit Care Med 1993;21:14871495. ADENOSINE 1. McCollam PL, U er W, Van Bakel AB. Adenosine-related ventricular asystole. An n Intern Med 1993;118:315316. 2. Rankin AC, Brooks R, Ruskin JM, McGovern BA. Adenosine and the treatment of s upraventricular tachycardia. Am J Med 1992;92:655664. P.856 3. Chronister C. Clinical management of supraventricular tachycardia with adenos ine. Am J Crit Care 1993;2:4147. AMINOGLYCOSIDES 1. See Bennett WM et al, pp. 18-19. AMPHOTERICIN 1. See Bennett WM et al, p. 31. AMRINONE http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 15)03-May-05 07:51:42

Ovid: ICU Book 1. Bottorff MB, Rutledge DR, Pieper JA. Evaluation of intravenous amrinone: the first of a new class of positive inotropic agents with vasodilator properties. P harmacotherapy 1984;5:227236. AZTREONAM 1. See Bennett WM et al, p. 25. 2. McLeod CM, Bartley EA, Payne JA, et al. Effects of cirrhosis on kinetics of a ztreonam. Antimicro Agents Chemother 1984;26:493497. BENZODIAZEPINES 1. Levine RL. Pharmacology of intravenous sedatives and opioids in critically il l patients. Crit Care Clin 1994;10:709731.

CEPHALOSPORINS 1. Gustaffero CA, Steckel erg JM. Cephalosporin antimicro ial agents and related compounds. Mayo Clin Proc 1991;66:10641073. CIMETIDINE 1. Ben-Menachem T, Fogel R, Patel RV, et al. Prophylaxis for stress-related gast ric hemorrhage in the medical intensive care unit. Ann Intern Med 1994;121:568575 . P.857 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 15)03-May-05 07:51:42

3. Olkkola KT, Aranko K, Luurila H, et al. A potentially hazardous interaction etween erythromycin and midazolam. Clin Pharmacol Ther 1993;53:298305.

2. Hoegholm A, Steptoe P, Fogh B, et al. Benzodiazepine antagonism ine. Acta Anesth Scand 1989;33:164166.

y aminophyll

Ovid: ICU Book COUMADIN 1. Crussel-Porter LL, Rindone JP, Ford MA, Jaskar DW. Low-dose fluconazole thera py potentiates the hypoprothrom inemic response to warfarin. Arch Intern Med 199 3;153:102104. 2. See McEvoy GK, ed., p. 928.

FLUCONAZOLE 1. Terrell CL, Hughes CE. Antifungal agents used for deep-seated mycotic infecti ons. Mayo Clin Proc 1992;67:6991. HEPARIN 1. Holliday DM, Watling SM, Yanos J. Heparin dosing in a mor idly o ese patient. Ann Pharmacother 1994;28:11101111. 2. Jaffrani NA, Ehrenpreis S, Laddu A, Som urg J. Therapeutic approach to unsta le angina: nitroglycerin, heparin, and com ined therapy. Am Heart J 1993;126:123 91242. IMIPENEM 1. Hellinger WC, Brewer NS. Imipenem. Mayo Clin Proc 1991;66:10741081. 2. Vos MC, Vincent HH, Yzerman EPF. Clearance of imipenem/cilastatin in acute re nal http://gateway.ut.ovid.com/gw1/ovidwe .cgi (11 of 15)03-May-05 07:51:42

DIGOXIN 1. Marcus MI. Pharmacokinetic interactions Coll Cardiol 1985;5:82A90A.

etween digoxin and other drugs. J Am

Ovid: ICU Book

INSULIN 1. See Trissel LA, pp. 585590. LIDOCAINE 1. Marcus FI, Opie LH. Antiarrhythmic drugs. In: Opie LH, ed. Drugs for the hear t. 4th ed. Philadelphia: WB Saunders, 1995;221222. NITROGLYCERIN 1. See Trissell LA, pp. 777780. NITROPRUSSIDE 1. Curry SC, Arnold-Cappell P. Nitroprusside, nitroglycerin, and angiotensin-con verting enzyme inhi itors. Crit Care Clin 1991;7:555582. P.858 2. FDA Medical Bulletin 21:34, March, 1991. OPIATES 1. Shochet RB, Murray GB. Neuropsychiatric toxicity of meperidine. J Intensive C are Med 1988;3:246252. 2. Chauvin M, Sandouk P, Scherrmann JM, et al. Morphine pharmacokinetics in rena l failure. Anesthesiology 1987;66:327331. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (12 of 15)03-May-05 07:51:42

failure patients treated Med 1992;18:282285.

y continuous hemodiafiltration (CAVHD). Intensive Care

Ovid: ICU Book PHENYTOIN 1. Cadle RM, Zenon GJ, Rodriguez-Barradas MC, Hamill RJ. Fluconazole-induced sym ptomatic phenytoin toxicity. Ann Pharmacother 1994;28:191195. 2. Smart HL, Somerville KW, Williams J, et al. The effects of sucralfate upon ph enytoin a sorption in man. Br J Pharmacol 1985;20:238240. 3. Lindow J, Wijdicks EFM. Phenytoin toxicity associated with hypoal uminemia in critically ill patients. Chest 1994;105:602604. 4. Wijdicks EFM. Seizures in the intensive care unit. In Neurology of critical i llness. Philadelphia: FA Davis, 1995;1933. 5. Epilepsy Foundation of America's Working Group on Status Epilepticus. Treatme nt of convulsive status epilepticus. JAMA 1993;270:854859. PROCAINAMIDE 1. Marcus FI, Opie LH. Antiarrhythmic drugs. In: Opie LH, ed. Drugs for the hear t. Philadelphia: WB Saunders, 1995;216217. PROPOFOL 1. Barr J. Propofol: A new drug for sedation in the intensive care unit. Int Ane sthesiol Clin 1993;31:131154. RANITIDINE 1. McEvoy GK. AHFS Drug Information 1995. Bethesda, MD: American Society of Heal th System Pharmacists, 1995;20572062. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (13 of 15)03-May-05 07:51:42

Ovid: ICU Book THEOPHYLLINE 1. Sessler CN, Brady W. Theophylline toxicity: How to minimize the potential ris k. J Crit Illness 1991;6:10451054. 2. Joeng CS, Huang SC, Jones DW. Theophylline disposition in Korean patients wit h congestive heart failure. Ann Pharmacother 1994;28:396401. 3. Spivey JM, Laughlin PH, Goss TF. Theophylline toxicity secondary to ciproflox acin administration. Ann Emerg Med 1991;20:11311134. 4. Rizzo A, Mira ella A, Bonanno A. Effect of ody weight on the volume of distr i ution of theophylline. Lung 1988;166:269276. P.859 THIAMINE 1. Dyckner T, Nyhlin H, Wester PO. Aggravation of thiamine deficiency y magnesi um depletion. Acta Med Scand 1985;218:129131. 2. LaFrance RJ, Miyagawa CI. Pharmaceutical considerations in total parenteral n utrition. In: Fischer JE, ed. Total parenteral nutrition. 2nd ed. Boston: Little , Brown, 1991;5798. TRIMETHOPRIMSULFAMETHOXAZOLE 1. Paap CM, Nahata MC. Trimethoprim/sulfamethoxazole dosing during renal dysfunc tion. Ann Pharmacother 1995;29:1300. VANCOMYCIN 1. Brown DL, Manro LS. Vancomycin dosing chart for use in patients with renal http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 15)03-May-05 07:51:42

Ovid: ICU Book impairment. Am J Kidney Dis 1988;11:1519. VERAPAMIL 1. Piepho RW, Cul ertson VL, Rhodes RS. Drug interactions with the calcium entry lockers. Circulation 1987;75:V181V194. 2. See McEvoy GK, ed., pp. 11531154. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 15)03-May-05 07:51:42

Ovid: ICU Book Copyright 1998 Lippincott Williams & Wilkins Marino, Paul L. ICU Book, 2nd Editio n Section XV APPENDICES Thus in dealing with any su ject matter, find out what entities are undenia ly i nvolved, and state everything in terms of these entities. Bertrand Russell P.862 P.863 Appendix 1 UNITS AND CONVERSIONS The units of measurement in the medical sciences are taken from the metric syste m (centimeter, gram, second) and the Anglo-Saxon system (foot, pound, second). T he metric units were introduced during the French Revolution and were revised in 1960. The revised units are called Systme International (SI) units and are curre ntly the worldwide standard. The United States initially refrained from adopting the SI units, ut this position has softened in recent years. P.864 UNITS OF MEASUREMENT IN THE SYSTME INTERNATIONAL (SI) Parameter Length Area Volum e L L2 L3 Dimensions Meter (m) Square meter (m2) Cu ic meter (m3) Basic SI Unit (Sym ol) 1 inch = 2.54 cm 1 square centimeter (cm2) = 104 m2 1 liter (L) = 0.001 m3 1 milliliter (ml) = 1 cu ic centimeter (cm3) Mass Density M M/L3 Kilogram (k g) Kilogram per cu ic meter (kg/m3) 1 pound (l ) = 453.5 g 1 kg = 2.2 l s 1 kg/m 3 = 0.001 kg/dm3 Density of water = 1.0 kg/dm3 Density of mercury = 13.6 kg/dm3 Velocity Acceleration L/T L/T2 M (L/T2) Meters per second (m/sec) Meters per sec ond squared (m/ sec2) Force Newton (N) = kg (m/sec2) 1 dyne = 105 N 1 mile per ho ur (mph) = 0.4 m/sec 1 ft/sec2 = 0.03 m/sec2 Equivalences http://gateway.ut.ovid.com/gw1/ovidwe .cgi (1 of 22)03-May-05 18:15:41

Ovid: ICU Book Pressure Pascal (Pa) = N/m2 1 kPa = 7.5 mm Hg = 10.2 cm H2O 1 mm Hg = 1.00000014 torr (See conversion ta le for kPa and mmHg) Heat Temperature Viscosity M (L/T2) L None M, 1/L, 1/T Joule (J) = N m Kelvin (K) Newton second per square meter (N sec/m2) 1 kilocalorie (kcal) = 4184 J 0 C = 273 K (See conversion ta le for C and F) Centipo ise (cP) = 103N sec/m2 Equivalent (Eq) = mol valence Ionic strength = mol/kg Amount of a su stance Concentration N N/L3 N/M Mole (mol) = molecular weight in grams mol/m3 = Molarity mol/kg = Molality TEMPERATURE CONVERSIONS (C) 100 41 40 39 38 37 36 35 34 33 32 31 30 0 F = (9/5 C) + 32 C = 5/9 (F 32) (F) 212 105.8 104 102.2 100.4 98.6 96.8 95 93.2 91.4 89.6 87.8 8 6 32 P.865 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (2 of 22)03-May-05 18:15:41

Ovid: ICU Book APOTHECARY AND HOUSEHOLD CONVERSIONS Apothecary 1 grain = 60 mg 1 ounce = 30 g 1 fluid ounce = 30 mL 1 pint = 500 mL 1 quart = 1000 mL Household 1 teaspoonful = 15 mL 1 dessertspoonful = 10 mL 1 ta lespoonful = 15 mL 1 wineglassful = 60 mL 1 teacupful = 120 mL 1 tum lerful = 240 mL 1 petroleum arrel = 42 gal PRESSURE CONVERSIONS mmHg 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 kPa 5.45 5.59 5.72 5.85 5.99 6.12 6.25 6.38 6.52 6.65 6.78 6.92 7.05 7.18 7.32 7.45 7.58 7.71 7.85 7.98 mmHg 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 kPa 8.11 8.25 8.38 8.51 8.65 8.78 8.91 9.04 9.18 9.31 9.44 9.58 9 .71 9.84 9.98 10.11 10.24 10.37 10.51 10.64 mmHg 81 82 83 84 85 86 87 88 89 90 9 1 92 93 94 95 96 97 98 99 100 kPa 10.77 10.91 11.04 11.17 11.31 11.44 11.57 11.7 0 11.84 11.97 12.10 12.24 12.37 12.50 12.64 12.77 12.90 13.03 13.17 13.90 Kilopascal (kPa) = 0.133 mmHg mmHg = 7.50 kPa P.866 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (3 of 22)03-May-05 18:15:41

Ovid: ICU Book pH AND HYDROGEN ION CONCENTRATION pH 6.8 6.9 7.0 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 [H+] (nEq/L) 160 125 100 80 63 50 40 32 26 20 16 SIZES OF PLASTIC TUBE DEVICES French Size 1 4 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 * 0.01 0.05 0.10 0.13 0.16 0.18 0.21 0.23 0.26 0.28 0.31 0.34 0.36 0.39 0.41 0.44 0.47 0.50 2.6 3.3 4.0 4.6 5.3 6.0 6.6 7.3 8.0 8.6 9.3 10.0 10.6 1 1.3 12.0 12.6 Chest tu es Outside Diameter* inches 0.3 1.3 mm Vascular catheters Device Small- ore feeding tu es Nasogastric tu es

P.867 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (4 of 22)03-May-05 18:15:41

Diameters can vary with manufacturers. However, a useful rule of thum ) 3 = French size.

is OD (mm

Ovid: ICU Book SIZES OF INTRAVASCULAR CATHETERS Outside Diameter* Gauge 26 25 24 23 22 21 20 19 18 16 14 12 10 inches 0.018 0.020 0.022 0.024 0.028 0.032 0.036 0.040 0.048 0.0 64 0.080 0.104 0.128 0.71 0.81 0.91 1.02 1.22 1.62 2.03 2.64 3.25 Central venous catheters mm 0.45 0.50 0.56 0.61 Peripheral vascular catheters Type of Catheter Butterfly devices Introducer catheters * Diameters can vary with manufacturers. P.868 P.869

Appendix 2 SELECTED REFERENCE RANGES REFERENCE RANGES FOR SELECTED CLINICAL LABORATORY TESTS Su stance Acetoacetate A lanine aminotransferase (SGPT) Al umin Aldolase Alkaline phosphatase Ammonia Amy lase Aspartate aminotransferase (SGOT) -Hydroxy utyrate Bicar onate Biliru in: To tal Conjugated Fluid* P, S S S CSF S S P S S S S S S Traditional Units 0.33.0 mg/ dL 035 U/L 46 g/dL 1148 mg/dL 06 U/L (F)30100 U/L (M)45115 U/L 1080 g/dL 0130 U/L 0 <1.0 mg/dL 2226 mEq/L 0.11.0 mg/dL 0.2 mg/dL k 97.95 0.016 10 0.01 16.6 0.016 0.587 0.016 0.016 96.05 1 17.1 = 330 mol/L 00.58 kat/L 4060 g/L 0.110.48 g/L 0100 nkat/L 0. .67 kat/L 0.751.92 kat/L 550 mol/L 02.17 kat/L 00.58 kat/L <100 mol/L 2226 mmol/L ol/L SI Units http://gateway.ut.ovid.com/gw1/ovidwe .cgi (5 of 22)03-May-05 18:15:41

Ovid: ICU Book Blood urea nitrogen (BUN) Calcium: Total Ionized Chloride P, S S P P, S CSF U

818 mg/dL 8.510.5 mg/dL 2.22.3 mEq/L 95105 mEq/L 120130 mEq/L 10200 mEq/L 0.61.5 mg/dL 1525 mg/kg/24 hr <5 g/dL >30 g/dL 150350 mg/dL <10 g/mL 70100 mg/dL 5080 mg/dL <2.0 mE /L <4.0 mEq/L 50150 U/L 0160 U/L 1.83.0 mg/dL 1,52.4 mEq/L 280296 mOsm/kg 2.55.0 mg/dL 3,55.0 mEq/L 6.08.0 g/dL <40 mg/dL <150 mg/24 hr 135147 mEq/L 411 g/dL 0.82.8 ng/dL 7 5220 ng/dL 0.367 1 1 3.06.5 mmol/L 2.22.6 mmol/L 1.101.15 mmol/L 95105 mmol/L 120130 mmol/L 10200 mmol/L Creatinine Cyanide: Nontoxic Lethal Fi rinogen Fi rin split products Glucose (fa sting) Lactate: Resting Exercise Lactate dehydrogenase (LDH) Lipase Magnesium Os molality Phosphate Potassium Total protein S U WB P S P CSF P S S P, S S S P, S P, S CSF U P, S S S 0.09 0.009 3.8 0.01 1 0.06 1 0.017 0.017 0.41 0.5 1 0.32 1 10 0.01 0.01 1 12.9 0 .015

0.050.13 mmol/L 0.130.22 mmol/kg/24 h <19 mol/L >144 mol/L 1.53.5 g/L <10 mg/L 3.96.1 mmol/L 2.84.4 mmol/L <2 mmol/L <4 mmol/L 0.822.66 kat/L 02.66 kat/L 0.81.2 mmol/L 0.81 2 mmol/L 280296 mmol/kg 0.801.60 mmol/L 3.55.0 mmol/L 6080 g/L <0.40 g/L <1.5 g/24 h r 135147 mmol/L 51142 nmol/L 1036 pmol/L 1.23.4 nmol/L Sodium Thyroxine: Total Free Triiodothyronine (T3) *

Adapted from the New England Journal of Medicine SI Unit Conversion Guide. Walth am, MA: Massachusetts Medical Society, 1992. P.870 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (6 of 22)03-May-05 18:15:41

P = Plasma, S = serum, U = urine, WB = whole BC = red lood cell.

lood, CSF = cere rospinal fluid, R

Ovid: ICU Book

REFERENCE RANGES FOR VITAMINS AND TRACE ELEMENTS Su stance Chromium Copper Folat e Iron Ferritin Manganese Pyridoxine Ri oflavin Selenium Thiamine (total) Vitami n A Vitamin B2 Vitamin C Vitamin D Vitamin E Zinc * Fluid* S S RBC S P, S WB P S WB P P, S S S S P, S S Traditional Units 0.140.15 ng/mL 70140 g/dL 140960 ng/mL (M) 80180 g/dL (F)60160 g/dL (M)20250 ng/mL (F)10120 ng/mL 0.42.0 g/dL 2090 ng/mL 2.63. 8234 g/dL 3.44.8 g/dL 1050 g/dL 2001000 pg/mL 0.62 mg/dL 2440 ng/mL 0.781.25 mg/dL L 1 0.018 5.98 26.57 0.012 0.003 0.349 0.737 56.78 2.599 23.22 0.153 17.85 0.16 2.26 0.18 k = SI Units 2.52.7 nmol/L 1122 mol/L 3172169 nmol/L (M)1432 mol/L (F)1129 /L (M)20250 g/L (F)10120 g/L 0.73.6 mol/L 120540 nmol/L 70100 nmol/L 0.72.5 mol/L 9 l/L 0.351.75 mol/L 150750 pmol/L 30100 mol/L 60105 nmol/L 1829 mol/L 11.518.5 mol/L P = plasma, S = serum, WB = whole lood, RBC = red lood cell. Adapted from the New England Journal of Medicine SI Unit Conversion Guide, Walth am, MA: Massachusetts Medical Society, 1992. LABORATORY MEASUREMENTS THAT ARE INFLUENCED BY BODY POSITION % Decrease When Upr ight Measurement Hemoglo in Hematocrit Serum calcium Total protein Serum al umin Cholesterol Alkaline phosphatase Alanine aminotransferase Average 5 6 4 9 9 9 9 7 Range 37 49 26 710 614 515 511 414 From Ravel R. Clinical la oratory medicine, Chicago: Year ook Medical Pu lishing , 1989;4. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (7 of 22)03-May-05 18:15:41

Ovid: ICU Book VOLUME USED BY AUTOMATED ANALYZERS IN THE CLINICAL LABORATORY La oratory Test Ar terial lood gases Electrolyte panel Complete lood cell count Glucose For serum assays, the volume of whole lood to e withdrawn is: Instrument Volume (mL) 1. 0 0.15 0.125 0.04 From Mayo Clin Proc 1993;68:255. P.871

DESIRABLE WEIGHTS FOR ADULTS* Height Feet 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 2 3 4 5 6 7 8 9 10 11 0 1 2 3 4 Inches 128134 130136 132138 134140 136142 138145 140148 14215 44154 146157 149160 152164 155168 158172 162176 Females 4 4 5 5 5 5 5 10 11 0 1 2 3 4 02111 103113 104115 106118 108121 111124 114127 109121 111123 113126 115129 11813 38 112131 120134 122137 125140 128143 131147 134151 Small Frame 131141 133143 13514 139151 142154 145157 148160 151163 154166 157170 160174 164178 167182 171187 Male Frame Large Frame 138150 140153 142156 144160 146164 149168 152172 155176 158180 161 4188 168192 172197 172202 181207 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (8 of 22)03-May-05 18:15:41

Ovid: ICU Book 5 5 5 5 5 5 5 6 * 5 6 7 8 9 10 11 0 117130 120133 123136 126139 129142 132145 135148 138151 127141 130144 133147 136150 139153 142156 145159 148162 137155 140159 143163 146167 149170 152173 155176 158179 Unclothed weights associated with the longest life expectancies. From the statis tics ureau of the Metropolitan Life Insurance Company, 1983. P.872 BASAL METABOLIC RATES Body Weight (kg) 40 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 1340 1485 1505 1555 1580 1600 1630 1660 1690 1725 1765 1785 1815 1845 1870 1900 Male kcal/24 hours Female 1241 1399 1429 1458 1487 1516 1544 1572 1599 1626 1653 1679 1705 1731 1756 1781 1805 From Tal ot FB. Am J Dis Child 1938;5:455459. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (9 of 22)03-May-05 18:15:41

Ovid: ICU Book DETERMINATIONS OF BODY SIZE Ideal Body Weight* Body Mass Index Body Surface Area Du ois Formula Jaco son Formula * Devine BJ. Drug Intell Clin Pharm 1974;8:650. Matz R. Ann Intern Med 1993;118:23 2. Du ois EF. Basal meta olism in health and disease. Philadelphia: Lea & Fe ige r, 1936. Jaco son B. Medicine and clinical engineering. Englewood Cliffs, NJ: Pr entice-Hall, 1977. P.873 BODY BUILD AND BLOOD VOLUME IN ADULTS Average Blood Volume (mL/kg) Body Build Th in Normal Muscular O ese Males 65 70 75 60 Females 60 65 70 55 From Documenta Geigy Scientific Ta les. 7th ed. Basel, Switzerland: JR Geigy, SA , 1970;528. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (10 of 22)03-May-05 18:15:41

Ovid: ICU Book

BLOOD VOLUMES IN THE ELDERLY Volume Whole lood Plasma Erythrocytes Elderly Men (3809 BSA) 2362 (1,9995 BSA) 667 (1761 BSA) 1609 Elderly Women (1591 BSA) + 889 (925 BSA) + 802 (716 BSA) + 14 From Cordtes PR et al. Surg Gynecol O stet 1992;175:243248. BODY FLUID DISTRIBUTION IN HEALTHY ADULTS Parameter Total ody water Interstitia l fluid Blood volume (BV) Erythrocyte volume (EV) Plasma volume (PV) Hematocrit (Hct) Derivation 0.55 ody wt (kg) 0.16 ody wt (kg) 0.065 ody wt (kg) EV = BV Hct PV = BV EV EV/BV 100 Male 600 mL/kg 160 mL/kg 70 mL/kg 33 mL/kg 37 mL/kg 47% (mean) 4054% (range) Female 500 mL/kg 160 mL/kg 65 mL/kg 27 mL/kg 38 mL/kg 42% ( mean) 3747% (range) From Documenta Geigy Scientific Ta les. 7th ed. Basel, Switzerland: JR Geigy SA, 1970. P.874 PEAK EXPIRATORY FLOW RATES FOR HEALTHY MALES Average Peak Flow (L/min) Age (yr) Ht: 60 65 70 75 http://gateway.ut.ovid.com/gw1/ovidweb.cgi (11 of 22)03-May-05 18:15:41

Ovid: ICU Book 20 25 30 35 40 45 50 55 60 65 70 602 590 577 565 552 540 527 515 502 490 477 649 636 622 609 596 583 569 556 542 529 515 693 679 664 651 636 622 607 593 578 564 550 740 725 710 695 680 665 649 634 618 603 587 Peak Flow (L/min) = [3.95 (0.0151 Age)] Ht (cm) Regression equation from Leiner GC et al. Am Rev Respir is 1963;88:646. PEAK EXPIRATORY FLOW RATES FOR HEALTHY ) 20 25 30 35 40 45 50 55 60 65 70 Ht: 40 55 423 418 413 408 402 397 391 386 18 412 406 400 65 70 496 490 483 476

Peak Flow (L/min) = [2.93 (0.0072 Age)] Ht (cm) Regression equation from Leiner GC et al. Am Rev Respir is 1963;88:647. P.875 Appendix 3 CLINICAL SCORING SYSTEMS APACHE II SCORING SYSTEM The APACHE (Acute Physiology and Chronic Health Evaluation) scoring system was d eveloped to provide an objective assessment of severity of illness in patients i n the ICU. The scoring system is not meant for burn patients or postcardiopulmona ry bypass patients. Although there are limitations in predicting mortality http://gateway.ut.ovid.com/gw1/ovidweb.cgi (12 of 22)03-May-05 18:15:41

FEMALES 309 385 380 375 470 464

Average 380 375 369 60 457 451

Peak Flow (L/min) Age (yr 370 365 360 355 350 345 3 460 454 448 442 436 430 424 4 445 439 432

Ovid: ICU Book in individual patients in the ICU, the APACHE scoring system is widely used in c linical studies to provide some measure of disease severity in the study patient s. The following pages demonstrate how to generate an APACHE II score (1). Altho ugh there is an APACHE III scoring system (2), the APACHE II score is more widel y used. The APACHE II score is made up of three components: q Acute Physiology Score (APS). The largest component of the APACHE II score is de rived from 12 clinical measurements that are obtained within 24 hours after admi ssion to the ICU. The most abnormal measurement is selected to generate the APS component of the APACHE II score. If a variable has not been measured, it is ass igned zero points. q Age Adjustment. From one to six points is added for patients older than 44 years of age. q Chronic Health Evaluation. An additional adjustment is made for patients with se vere and chronic organ failure involving the heart, lungs, kidneys, liver, and i mmune system. P.876 ACUTE PHYSIOLOGY SCORE Points: Temperature (C) Mean arterial pressure Heart rate Respiratory rate 1 2 +4 41 160 180 50 500 +3 3940.9 130159 140179 3549 350499 +2 +1 38.538.9 0 3638.4 70109 70109 +1 3435.9 +2 3233.9 5069 5569 +3 3031.9 +4 29.9 49 110129 110139 2534 200349 4054 39 5 1224 <200 >70 1011 69

A-aPO2 PAO2 6170 7.257.32 5560 7.157.24 <55 <7.15 Arterial pH 3 7.7 52 7.67.69 7.57.59 7.337.49 Serum 4151.9 3240.9 2331.9 1821.9 1517.9 <15 bicarbonate (mEq/ L) Serum sodium (mEq/L) Serum potassium (mEq/L) Serum creatini ne (mg/dL) Hematocrit WBC count 180 7 3.5 60 40 66.9 23.4 160179 155159 5.55.9 150154 3.55.4 0.61.4 130149 33.4 120129 2.52.9 <0.6 2029.9 12.9 111119 110 <2.5 1.51.9 5059.9 2039.9 4649.9 1519.9 3045.9 314.9 <20 <1 http://gateway.ut.ovid.com/gw1/ovidweb.cgi (13 of 22)03-May-05 18:15:41

Ovid: ICU Book 4 15 (Glasgow Coma Score) = 1. If FiO2 >50. 2. If FiO2 <50%. 3. Use only if no ABGs. Scoring Method 1. Selec t the most abnormal measurement for each parameter in the first 24 hours after I CU admission. 2. If a parameter has not been measured, assign it zero points. 3. Add the corresponding points for all 12 parameters to obtain the Acute Physiolo gy Score. 4. Glasgow Coma Scale follows. THE GLASGOW COMA SCALE* Eye Opening: Spontaneous To Speech To Pain None Verbal C ommunication: Oriented Confused conversation Inappropriate words Incomprehensi le sounds None Motor Response: O eys commands Localizes to pain Withdraws to pain A normal flexion A normal extension None Total Points 6 5 4 3 2 1 Points 5 4 3 2 1 Points Points 4 3 2 1 Points * Adapted from Teasdale G, Jennet B. Assessment of coma and impaired consciousne ss. A practical approach. Lancet 1974;2:8186. For intu ated patients, assign a sc ore of 1 for ver al communication Best score is 15 points; worst score is 3 poin ts. P.877 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (14 of 22)03-May-05 18:15:41

Ovid: ICU Book AGE ADJUSTMENT Age (yr) <44 4554 5564 6574 >75 Points 0 2 3 5 6 CHRONIC HEALTH ADJUSTMENT For any of the following: 1. Biopsy proven cirrhosis. 2. Heart failure: NYHA Class IV 3. Severe COPD (hypercapnia, home oxygen) 4. Chr onic dialysis 5. Immunocompromised Add 2 points for elective surgery or neurosur gery, 5 points for emergency surgery. TOTAL APACHE II SCORE APS score Age adjustment Chronic health adjustment Total A PACHE II score _____________________________________________________ ___________ __________________________________________ _____________________________________ ________________ _____________________________________________________ http://gateway.ut.ovid.com/gw1/ovidwe .cgi (15 of 22)03-May-05 18:15:41

Ovid: ICU Book

APACHE II SCORE AND MORTALITY IN 5815 ICU PATIENTS Hospital Mortality (%) APACHE II Score 04 59 1014 1519 2024 2529 3034 35 Nonoperative 4 6 12 22 40 51 71 82 Postop tive 1 3 6 11 29 37 71 87 ata from Knaus WA et al. Crit Care Med 1985;13:818829.

Limitations The following limitations of the APACHE II score deserve mention. q The APS score has no adjustments for measurements obtained in the presence of in terventions such as hemodynamic support drugs, mechanical ventilation, or antipy retic therapy. P.878 q There is an exaggerated penalty for old age. For example, age greater than 65 ye ars adds more points than an A-a PO2 gradient above 500 mm Hg (6 points versus 4 points, respectively). q There is no consideration for malnutrition or cachexia in the chronic health eva luation. MULTIPLE ORGAN YSFUNCTION SCORE The multiple organ dysfunction score was developed because of the direct relatio nship between the number of major organ systems failed and mortality in ICU pati ents. Unlike the APACHE II scoring system, which can only be used on admission t o the ICU, the multiple organ dysfunction score can be determined daily. This al lows identification of changing mortality risks in individual patients during th e ICU stay. http://gateway.ut.ovid.com/gw1/ovidweb.cgi (16 of 22)03-May-05 18:15:41

Ovid: ICU Book

Parameter PAO2/FiO2 Serum creatinine ( mol/L) Serum bilirubin ( mol/L) Pulse-adjuste d heart rate* Platelet count (per L) Glascow Coma Score 0 1 >300 100 20 10 >120 15 2 226300 101200 2160 10.115 81120 1314 151225 201350 61120 15.120 5180 1012 3 76150 351500 121240 20.130 2150 79 4 75 500 >240 >30 20 6 * PAR = HR (RAP/MAP); HR = heart rate, RAP = right atrial pressure, MAP = mean a rterial pressure. The est estimate in the a sence of sedation. Scoring Method 1. Select the most a normal result for each parameter over a 24-h our period. 2. If a parameter has not een measured, assign it zero points. 3. A dd the corresponding points for all 6 parameters to o tain the final score. 4. A new score can e o tained for any su sequent 24-hour period. From Marshall JC e t al. Multiple organ dysfunction score: a relia le predictor of a complex clinic al outcome. Crit Care Med 1995;23:16381652. P.879 MULTIPLE ORGAN DYSFUNCTION SCORE AND ICU MORTALITY* Score 0 14 58 912 Mortality (%) 0 1 3 25 Score 1316 1720 >20 Mortality (%) 50 75 100 * Mortality recorded in 692 patients in a Surgical ICU. From Marshall JC et al. Crit Care Med 1995;23:16381652. REFERENCES 1. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: A severity of diseas e classification system. Crit Care Med 1985;13:818828. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (17 of 22)03-May-05 18:15:41

THE MULTIPLE ORGAN

YSFUNCTION SCORE Points

Ovid: ICU Book 2. Knaus WA, Wagner DP, Draper EA, et al. The APACHE III Prognostic system. Ches t 1991;100:16191636. P.880 Appendix 4 HEALTH CARE AND VITAL STATISTICS CHARGES TO THE PATIENT FOR SELECTED HOSPITAL CARE SERVICES IN 1996* Inpatient La oratory Tests Sodium Chloride Glucose BUN Creatinine Na, K, Cl, Creat Calcium P hosphate Magnesium Heme Panel WBC differential Arterial lood gas Urinalysis CPK Stat charge per test $52.75 $51.75 $24.50 $40.50 $129.00 $170.75 $29.75 $29.75 $70.00 $35.25 $44.00 $173.00 $20.50 $258.25 $25.00 Miscellaneous Echo, 2D Averag e Daily Room Charge Regular room Intermediate care unit Intensive care unit * $9 93.00 $1253.00 ECG Chest physical therapy Oxygen, hourly $1149.00 $222.00 $149.0 0 $16.00 $1759.00 CT, head (unenhanced) CT, head (enhanced) MRI, head $1342.00 $ 1437.00 $1761.00 Ultrasound, a domen CT, a domen (enhanced) MRI, a domen $800.00 $1201.00 $1761.00 Radiographic Imaging Chest, single view Chest, two views A do men, one view A domen, complete Porta le charge $126.00 $167.00 $158.00 $215.00 $100.00

P.881 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (18 of 22)03-May-05 18:15:41

Patient charges are those of the University of Pennsylvania Health System, Pres yterian Medical Center, for the 1996 fiscal year.

Ovid: ICU Book ANNUAL COST OF HEALTH CARE IN THE UNITED STATES (Billions of dollars) 1960 1970 1980 1990 1994 All health-related services Hospital care Physician services Medications Healthrelated research $26.9 $9.3 $5.3 $4.2 $0.7 $73.2 $28.0 $13.6 $8.8 $2.0 $247.2 $102.7 $45.2 $21.6 $5.6 $697.5 $256.5 $140.5 $61.2 $8.5 $949.4 $338.5 $189.4 $78.6 $15.9 From Office of National Health Statistics, Health Care Financing Administration. LIFE EXPECTANCY IN THE UNITED STATES Years Expected At Birth White Black Year 1950 1960 1970 1980 1990 1995 Female 72.2 74.1 75.6 78.1 79.4 79.6 Male 60.8 67.4 68.0 70.7 72.9 73.4 Female 62.9 66.3 69.4 73.6 75.2 74.0 Male 59.1 61.1 61.3 65.3 67.0 65.4 From National Center for Health Statistics, U.S. Department of Health and Human Services. ANNUAL MORTALITY IN THE UNITED STATES Year Total Deaths* Deaths per 1000 Populat ion 1960 1970 1980 1990 1991 1992 1993 1994 1,711,982 1,921,031 1,986,000 2,162,000 2,169,518 2,175,613 2,268,000 2,286,000 9.5 9.5 8.7 8.6 8.6 8.5 8.8 8.8 http://gateway.ut.ovid.com/gw1/ovidwe .cgi (19 of 22)03-May-05 18:15:41

Ovid: ICU Book * Excludes fetal deaths and deaths outside the United States. From National Center for Health Statistics, US Department of Health and Human Services. P.882 THE ELDERLY POPULATION IN THE UNITED STATES Total U.S. Population (in millions) 179.3 248.7 262.8 297.7 393.9 16.0 (8.9%) 31.2 (12.6%) 33.5 (12.8%) 39.4 (13.2%) 78.8 (20.0%) Elderly Population (in millions) >65 yr >85 yr 0.9 (<0.01%) 3.1 (1 .2%) 3.6 (1.4%) 5.6 (1.9%) 18.2 (4.6%) Year 1960 1990 1995 2010 2050 From Bureau of the Census, U.S. Department of Commerce. WORLD POPULATION GROWTH RATE World Population (in illions) 1 2 3 4 5 Time for E ach Growth Increment (yr) 2,000,000 100 30 15 7 From Rifkin J. Entropy and the greenhouse world. New York: Bantam Books, 1989;11 8. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (20 of 22)03-May-05 18:15:41

Ovid: ICU Book RELATIVE SIZE OF THE MEDICAL SPECIALTIES IN 1995 Most Popular Specialty Internal medicine Family practice Pediatrics Psychiatry General surgery O Gyn Anesthesiol ogy Orthopedics Radiology Emergency medicine Physicians 88,241 59,110 43,609 38, 089 37,570 33,519 32,853 22,037 19,808 19,112 Specialty Forensic pathology Aeros pace medicine Colorectal surgery Preventive medicine Nuclear medicine Pu lic hea lth Thoracic surgery Occupational medicine Radiation oncology Allergy/immunology Least Popular Physicians 466 575 990 1269 1435 1760 2310 3031 3360 4040 From 1997 World Almanac and Book of Facts. Mahwah NJ: World Almanac Books, 1997; 969. P.883 MOST COMMONLY DISPENSED PRESCRIPTION DRUGS IN THE UNITED STATES IN 1995 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Premarin (estrogen) Synthroid (levothyroxine) Zantac (ran itidine) Trimox (amox/TMP/SMX) Lanoxin (digoxin) Amoxil (amoxicillin) Procardia (nifedipine) Vasotec (enalapril) Proventil (al uterol) Prozac (fluoxetine) 11. 1 2. 13. 14. 15. 16. 17. 18. 19. 20. Cardizem (diltiazem) Coumadin (warfarin) Zolo ft (sertraline HCL) Zestril (lisonopril) Augmentin (amox/clav) Biaxin (clarithro mycin) Triamterene/HCTZ Ventolin (al uterol) Cipro (ciprofloxacin) Hydrocodone w /APAP From 1996 Red Book. Montvale NJ: Medical Economics Co., 1996. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (21 of 22)03-May-05 18:15:41

Ovid: ICU Book MOST COMMONLY REPORTED TOXIC INGESTIONS Cases Reported Drugs Acetaminophen (adul t) Aspirin (adult) Benzodiazepines -Blockers Calcium antagonists Digitalis Cocain e Opiates Theophylline Tricyclic antidepressants No. Rank 25,742 5,411 33,516 5, 308 6,683 2,310 3,713 4,474 5,735 20,619 #2 #6 #1 #7 #4 #10 #9 #8 #5 #3 No. 18 8 54 16 38 19 52 33 35 146 Deaths Rank #8 #10 #2 #9 #4 #7 #3 #6 #5 #1 From Annual Report of the American Association of Poison Control Centers Toxic E xposure Surveillance System. Am J Emerg Med 1993;11:494. http://gateway.ut.ovid.com/gw1/ovidwe .cgi (22 of 22)03-May-05 18:15:41