Annals of Cardiac Anaesthesia 2007; 10: 121126

Original Article Chakravarthy et al. Cardiac Output Measurement by 4 Techniques 121

Comparison of Simultaneous Estimation of Cardiac Output by Four Techniques in Patients Undergoing Off-pump Coronary Artery Bypass Surgery- A Prospective Observational Study
Murali Chakravarthy, MD, DA, DNB; TA Patil, DA; K Jayaprakash, MD, DNB; Praveen Kalligudd, MD; Dattatreya Prabhakumar, MD; Vivek Jawali, MS, MCh. Wockhardt Heart Institute, Bangalore, Karnataka, India

We prospectively compared four techniques of cardiac output measurement: bolus thermodilution cardiac output (TDCO), continuous cardiac output (CCO), pulse contour cardiac output (PiCCO), and Flowtrac (FCCO), simultaneously in fifteen patients undergoing off-pump coronary artery bypass grafting (OPCAB). All the patients received pulmonary artery catheter (capable of measuring both bolus thermodilution cardiac output and CCO), PiCCO arterial cannula in the left and FCCO in the right femoral artery. Cardiac indices (CI) were obtained every fifteen minutes by using all the four techniques. TDCO was treated as control and the rest were treated as test values. Interchangeability of techniques with TDCO was assessed by Bland and Altman plotting and mountain plot. Four hundred and thirty eight sets of data were obtained from fifteen patients. The values of cardiac output varied between 1 to 6.9 L/min. We found that the values of all the techniques were interchangeable. At certain times, the values of CI measured by both PiCCO and FCCO appeared erratic. The values of CI measured simultaneously appeared in the following descending order of accuracy; TDCO>CCO>FCCO>PiCCO ( the % times TDCO correlated with CCO, FCCO, PiCCO was 93, 86 and 80 respectively). The bias and precision (in L/ min) for CCO were 0.03, 0.06, PiCCO 0.13, 0.1 and flowtrac 0.15, 0.04 respectively suggesting interchangeability. We conclude that the cardiac output measured by CCO technique and the pulse contour as measured by PiCCO and FCCO were interchangeable with TDCO more than 80% of the times. (Annals of Cardiac Anaesthesia 2007; 10: 121126) Key words: Cardiac output, Pulse contour, Thermodilution, Off-pump coronary artery bypass

Introduction

easurement of the cardiac output (CO) and other derived variables helps anaesthesiologists to treat haemodynamic disturbances in the perioperative period. The need for continuous measurement of these values is relevant during off-pump coronary artery bypass (OPCAB) surgery. Currently more than five techniques of continuous CO (CCO) measurement are available. Despite limitations, CO by thermodilution technique (TDCO) is still

considered the gold standard, and any new method is compared with it. 1-4 Manufacturers of the equipments such as CCO, Lithium dilution cardiac output (LIDCO), pulse contour cardiac output (PiCCO) and the recently introduced Flotrac, claim that each of their equipment can continuously and accurately measure CO on a beat-to-beat basis. The aim of the preent study was to compare these techniques with TDCO. Methods Fifteen normotensive patients undergoing OPCAB under general anaesthesia were included in the study after obtaining ethical committee approval. A written informed consent for surgery and participation in the study was obtained from

Address for Correspondence: Dr. Murali Chakravarthy, Chief Consultant Anesthesiologist, Wockhardt Heart Institute, Bangalore 560052, Karnataka, India Telephone Office: 91 80 51994444, 91 80 22261037, Fax: 91 80 22281149 E-mail: chakravarthy@vsnl.com, whhi@vsnl.com

ACA-06-110- OA.p65

121

6/28/2007, 9:47 AM

122 Chakravarthy et al. Cardiac Output Measurement by 4 Techniques

Annals of Cardiac Anaesthesia 2007; 10: 121126

each patient. Presence of preoperative inotropic therapy, intra-aortic balloon pump, ventilator therapy, peripheral vascular disease, disease of the organ systems (renal, hepatic and metabolic disorders), intracardiac shunts, regurgitation of pulmonary and/ or tricuspid valves and emergency surgery were taken as exclusion criteria. All patients were premedicated with fentanyl 2 g/kg and midazolam 0.05 mg/kg, thirty minutes prior to surgery. Upon arrival in the operation theatre, both femoral arteries were cannulated; right with 18 gauge secalon T cannula and the left with the custom made PiCCO cannula. Right internal jugular vein was cannulated with 8.5 French introducer sheath and a continuous thermodilution CO catheter was inserted in to the pulmonary artery using pressure waveform as the guide to identify cardiac chambers (Swan Ganz CCO/VIP, Edward Lifesciences, LLC, Irvine, CA USA). The vigilance monitor was used to measure CO and base line CO by CCO technique was noted down. Whenever difficulty was encountered in achieving a wedge position, an image intensifier was used to guide the pulmonary artery catheter to an acceptable position. CO was measured by bolus thermodilution technique and this value was noted as basal TDCO. The right femoral arterial cannula was connected to Flowtracmonitor (Edward Lifesciences, LLC, Irvine, CA USA) and the slave connection from the Flowtrac monitor was connected to the haemodynamic monitor for monitoring invasive arterial pressure. The custom made arterial cannula for PiCCO monitor was inserted in the left femoral artery and connected to PiCCO plus monitor (Pulsion Medical Systems AG, Stahlgruberring, Munich, Germany). An initial thermodilution curve was obtained on the PiCCO monitor by injecting 15 ml of saline through the proximal lumen of the Swan Ganz catheter at temperature of 40 C. Arterial pressure waveform was obtained from the left femoral artery; the monitor displayed the CO on a beat-tobeat basis. Baseline values of CO by TDCO, CCO, Flowtrac and PiCCO were obtained prior to induction of anaesthesia. Device characteristics: PiCCO provides continuous estimate of stroke volume and the CO is derived from the arterial

pressure waveform. Stroke volume is calculated using a proprietary algorithm using beat duration, ejection duration, mean arterial pressure and modulus and phase of the first waveform harmonic. In order to measure CO on a beat-tobeat basis, one needs to obtain an initial calibration against an independent measure of CO usually against thermodilution technique. Flowtrac device uses arterial pressure waveform analysis to determine the CO from an arterial pressure based algorithm that utilises the relationship between pulse pressure and stroke volume. It consists of a special transducer that attaches to an existing arterial cannula and then connects to a processing and display unit (Vigileo). The arterial waveform is assessed at 100 Hz and the standard deviation of the pulse pressure is determined over a 20 s window. The algorithm takes two additional factors; vessel compliance and peripheral resistance. Intermittent bolus or continuous thermodilution CO was measured using Swan-Ganz pulmonary artery catheter. Bolus CO measurement was performed using 10 ml of normal saline at room temperature, average of 3 values were considered. The continuous CO was viewed in the stat mode, and the most recent data was taken as the value for comparison. The values from PiCCO and flowtrac were recorded at the time of third thermodilution injection. General anaesthesia was induced with midazolam 0.5 mg/kg, fentanyl 2 g/kg administered intravenously and inhalation of 1 to 3 % of sevoflurane or isoflurane. Intubation was facilitated with 1 mg/kg rocuronium administered intravenously. Oral endotracheal tube was placed in the trachea and controlled ventilation was commenced. After stabilizing haemodynamic disturbances that may have occurred following induction of general anaesthesia, CO was measured every 15 minutes by using all the 4 techniquesduring surgery. Heparinization was achieved with 200 units/kg of heparin and an activated clotting time of >250 sec was considered adequate for surgery. Haemodynamic interventions were carried out based on the haemodynamic parameters obtained from the TDCO. We

ACA-06-110- OA.p65

122

6/28/2007, 9:47 AM

Annals of Cardiac Anaesthesia 2007; 10: 121126

Chakravarthy et al. Cardiac Output Measurement by 4 Techniques

123

empirically divided our data in to three groups; CO of < 4.0 L/min , between 4 and 6 L/min and in excess of 6 L/min. Intravenous infusion of inotropic agent and/ or intravenous fluid boluses (2-3 ml/kg) and/or appropriate tilt of the operating table were implemented to restore haemodynamic values to acceptable levels. Intravenous infusion of dopamine in the dose of 5 to 7.5 g/kg/min was the inotrope of choice. Other inotropic and vasodilatory agents were used depending on the values of heart rate, mean arterial pressure, central venous pressure, pulmonary artery wedge pressure and systemic vascular resistance index. Surgical access was obtained through midsternotomy. Left internal mammary artery was harvested and other suitable conduits such as saphenous vein and radial artery were harvested as required. All the patients underwent OPCAB surgery as planned. Left internal mammary artery was grafted first to left anterior descending artery. Other conduits were grafted thereafter. After the completion of anastomosis, protamine sulfate in the dose of 2 mg/kg was used to neutralize residual heparin. Return of activated clotting time in the range of 120 to 130 seconds was considered acceptable. A blinded observer recorded the COfrom the monitor at intervals of 15 minutes throughout the surgery. Statistical methods: Interchangeability of techniques was assessed by plotting values obtained by each of the techniques against TDCO values. Bias and precision between the techniques was assessed by using Bland and Altman plotting.5 To assess the interchangeability of the techniques, mountain plot tool (Krouwer et al 6) was used as complementary to Bland and Altman plotting. A P value < 0.05 was considered significant. Results There were 9 males, 8 diabetics, 5 hypertensives, 2 asthmatics and one had a past history of cerebrovascular accident. After excluding invalid recordings due to diathermy disturbance, physiotherapy and routine nursing care, 438 sets

of data were obtained from fifteen patients. All patients required 5-10 g/kg/min of dopamine infusion during the intraoperative period. The variation in the CO in our study was from 1 to 6.9 L/ min. Table 1 shows the bias and precision of other techniques compared to TDCO. Low bias and precision suggests interchangeability of techniques. Correlation of CO measured by various techniques was performed with TDCO (Table 2). The correlation co-efficient (r) was 0.6, 0.49 and 0.4
Table1. showing the bias (mean of differences) and precision (2 standard deviation n) with the different techniques Techniques TDCO/CCO TDCO/PiCCO TDCO/FCCO Bias L/ min 0.03 0.13 0.15 precision L/min 0.06 0.10 0.04

TDCO : thermodilution cardiac output, CCO : continuous cardiac output, PiCCO : pulse contour cardiac output, FCCO : Flowtrac cardiac output

for CCO, FCCO and PiCCO respectively. CCO measurements correlated with TDCO 91.5% times, underestimated 4.1% times and overestimated 4.3% of times. The errors occurred when TDCO values were <4 L/min. FCCO measurements correlated with TDCO 86.3% times, underestimated 8.2 % of times and overestimated 6.4 % times. Underestimation by FCCO occurred more often than overestimation (34 times vs 28). PiCCO measurements correlated with TDCO 80.8% times, while both underestimation and overestimation occurred almost at the same frequency (8.9 and 8.6%). In contrast to the measurements by FCCO, PiCCO measurements tended to underestimate TDCO <4 L/min (23 underestimation when TDCO <4 L/min compared to 15 underestimations when TDCO >4 L/min) and overestimate TDCO>6 L/ min (24 overestimation when TDCO >6 L/min, compared to 15 when TDCO <6 L/min). Bland and Altman plot of CCO versus TDCO shows mean difference (bias) of 0.03 L/min. The limits of agreement (2 SD) were +1.4 and -1.4. Figure 1 shows distribution of the values mostly within the limits of agreement suggesting interchangeability. Non-interchangeability of a few values may be because of time delay between the two techniques and is depicted in the figure with

ACA-06-110- OA.p65

123

6/28/2007, 9:47 AM

124 Chakravarthy et al. Cardiac Output Measurement by 4 Techniques Table 2. Frequency of correlation of cardiac output measured by TDCO compared with the other techniques
TDCO < 4 L/min CCO Correct correlation (93.8%) False low (2.5%) False high (3.8%) FCCO Correct correlation (86.3%) False low (8.2%) False high (6.4%) PiCCO Correct correlation (80.8%) False low (8.6%) False high (8.9%) 18 3 7 TDCO 4-6 L/min 261 7 4 TDCO > 6 L/min 131 1 6

Annals of Cardiac Anaesthesia 2007; 10: 121126

PiCCO Vs TDCO

6 11 8

239 8 10

136 14 6

2 27 7

245 7 8

105 8 29

TDCO : thermodilution cardiac output, CCO : continuous cardiac output, PiCCO : pulse contour cardiac output, FCCO : Flowtrac cardiac output

Fig. 2. Bland and Atman plot of data for thermodilution cardiac output (TDCO) and pulse contour cardiac output (PiCCO). The X axis is the mean of PiCCO and TDCO. The mean difference (bias) was 0.13 l/min and is depicted as black line. The limits of agreement (2 SD) were +2.24 and -2.24 are shown in dashed lines above and below zero. TDCO: thermodilution cardiac output, PiCCO: pulse contour cardiac output FCCO Vs TDCO

values of high bias (>5). Figures 2 and 3 show the Bland and Altman plots of measurements obtained by PiCCO and FCCO respectively. Plotting of both PiCCO and FCCO as compared to TDCO are depicted well within the limits of agreement. Values of FCCO appear to be distributed in a narrower range of distribution (2.24 for PiCCO as compared to 0.66 for FCCO).
CCO Vs TDCO

FCCO+TDCO/2

Fig. 3. Bland and Altman plot of Flowtrac cardiac output (FCCO) and thermodilution cardiac output (TDCO)- The X axis is the mean of FCCO and TDCO. The mean difference (bias) was 0.15 L/min and is depicted as black line. The limits of agreement (2 SD) were +0.66 and -0.66 are shown in dashed lines above and below zero. TDCO: thermodilution cardiac output, FCCO: Flowtrac cardiac output

CCO+TDCO/2 Fig. 1. Bland and Altman plot of data for theromodilution cardiac output (TDCO) and continuous cardiac output (CCO). The X axis is the mean of CCO and TDCO. The mean difference (bias) was 0.31 L/min and is depicted as black line. The limits of agreement (2 SD) were +1.4 and -1.4 are shown in dashed lines above and below zero. TDCO : thermodilution cardiac output, CCO : continuous cardiac output

interchangeability, which occurred with the techniques is seen as either the base or the apex of the mountain not aligning exactly with each other.

CCO PiCCO FCCO

Mountain plot, which is generally used complementary to Bland and Altman plots also confirmed the interchangeability of four techniques (Fig. 4). Observations are made based on comparison of shape of one mountain with the other. The median between TDCO and CCO, FCCO and PiCCO as assessed by mountain plots were 0.0, -0.40 and -0.10 respectively. After comparing the plots, the techniques could be classified in the decreasing order of accuracy as follows: TDCO>CCO>FCCO>PiCCO. Our observation made from mountain plot was similar to analysis of Bland and Altman plotting. Absence of

Fig. 4. Mountain plot comparing the interchangeability of continuous cardiac output, flowtrac cardiac output and pulse contour cardiac output with thermodiultion cardiac output. TDCO: thermodilution cardiac output, CCO: continuous cardiac output, PiCCO: pulse contour cardiac output, FCCO: Flowtrac cardiac output

ACA-06-110- OA.p65

124

6/28/2007, 9:47 AM

Annals of Cardiac Anaesthesia 2007; 10: 121126

Chakravarthy et al. Cardiac Output Measurement by 4 Techniques

125

Discussion In clinical practice, CO measurement is conventionally carried out by using bolus thermodilution technique. For the sake of convenience, CCO measurement technique has been introduced. The continuous technique measured by using Swan Ganz catheter equipped with a heating coil, is not truly continuous. A delay of 140 to 240 seconds occurs during measurement. In order to avoid these fallacies, CCO measurements in real time have been introduced in the market. PiCCO and FCCO are two such techniques. PiCCO technology employs patented algorithm to calculate beat-to-beat CO after initial validation by a termodilution technique through pulse contour analysis. Although this technique has been described as semi-invasive, clinicians need to insert a catheter in the central vein to obtain a thermodilution curve against which the subsequent arterial waveforms are compared. (thermodilution curve that is required from initial calibration is produced by injecting a known quantity of saline in the right atrium and a wash out curve is generated at the pulmonary artery). It is not necessary to introduce pulmonary artery catheter in order to measure CO by PiCCO. Our study suffers from the potential weakness that all the measurements made by these techniques may not have been in real time with respect to each other, because of potential delay in the measurement and display of CO by CCO technique. However, TDCO also has similar setback, because it is usually averaged over 3 values. In order to obviate the need for any form of central venous cannulation and calibration, Edward Life Sciences have recently introduced a device called Flowtrac , which can measure the continuous CO on a beat-to-beat basis. The advantage of this technique is that there is no need to calibrate with an initial thermodilution curve. Various authors 1-4, 7-10 have compared the accuracy of one measurement of CO to other. We chose to compare the above mentioned techniques in patients undergoing OPCAB because, sudden fluctuations in the mean arterial pressure, and CO are known to occur during OCPAB. We considered

TDCO as the control and others (PiCCO, CCO, and FCCO) as tests. PiCCO and FCCO depend on the arterial waveform for appropriate measurements, therefore patients with peripheral arterial disease were excluded from the study. Presence of inotropic agents and intra-aortic balloon pump in the preoperative period were also excluded from the study, because these may cause variations in the arterial waveforms. As has been previously observed, we have noted good degree of interchangeability between CCO and TDCO. Singh et al 7 observed similar excellent correlation, accuracy, and precision among the 3 methods of measuring CO in patients undergoing minimally invasive direct coronary artery bypass (MIDCAB) surgery. We observed that changes occurring in the haemodynamic parameters during lifting of the heart, positioning of the heart to expose the coronary arteries were immediately displayed by PiCCO and FCCO, while the display on TDCO and CCO were not immediate. The authors are of the opinion that this aspect of PiCCO and FCCO makes them suitable for surgeries where blood pressure changes can occur abruptly. If the arterial trace gets dampened, the estimation of CO by PiCCO and FCCO may get erratic because of the abnormality in the arterial waveform. It is necessary to have heparin flush on an automatic flush device in order to have accurate arterial waveforms all the time when these two techniques are used. Buhre et al 10 concluded that compared with TDCO, pulse contour analysis enables accurate measurement of continuous CO in patients undergoing MIDCAB but, Yamashila and could not demonstrate coworkers 11 interchangeability of PiCCO with TDCO in patients undergoing OPCAB. The interchangeability of FCCO with TDCO was described by Manecke. 12 Although, our observations are similar to Buhre and Manecke, we observed that at times, the values displayed by PiCCO and FCCO were non-interchangeable as compared to TDCO. These inaccuracies may be due to improper arterial waveforms caused by kinking of arterial cannula, positioning of the heart for grafting and external pressure on the course of the artery. Non-interchangeability of a few values of PiCCO and FCCO are seen on the Bland and

ACA-06-110- OA.p65

125

6/28/2007, 9:47 AM

126 Chakravarthy et al. Cardiac Output Measurement by 4 Techniques

Annals of Cardiac Anaesthesia 2007; 10: 121126

Altman plotting (Fig. 2 and 3) in patients undergoing OPCAB. It is true that the gold standard TDCO is an invasive technique, but other non-invasive or semi-invasive techniques need further validation. Both FCCO and PiCCO do not require pulmonary artery catheter insertion, but PiCCO requires a central venous catheter. Additionally, FCCO does not need any calibration against another dye dilution technique and is capable of measuring CO based on the arterial pressure waveforms. We conclude that the CO measurement by TDCO, CCO PiCCO and FCCO have good agreement and can be interchanged in patients undergoing OPCAB.

Acknowledgements
1. Kuralayanapalya Suresh, Scientist (Statistics), National Institute of Animal Nutrition and Physiology, Bangalore-560030, Karnataka, India, for statistical assistance 2. Professor Gopinath Ramachandran, Professor and head of the department, anaesthesia and critical care, Nizams institute of medical sciences, Hyderabad, India, for guidance. 3. Dr Dilip Kulkarni, Additional professor of anaesthesia, Nizams institute of medical sciences, Hyderabad, India, for statistical assistance in creating mountain plots.

References
1. Tsutsui M, Mori T, Aramaki Y, Fukuda I, Kazama T. A comparison of two methods for continuous cardiac output measurement: Pulse CO Vs CCO. Masui 2004; 53: 929-33 2. Schulz K, Abel HH, Werning P. Comparison between continuous and intermittent thermodilution measurement of cardiac output during coronary artery bypass operation. Anaesthesiol Intensivmed Notfallmed Schmerzther 1997; 32: 226-33 3. Jonas MM, Tanser SJ. Lithium dilution measurement of cardiac output and arterial pulse waveform analysis: an indicator dilution calibrated beat-by-beat system for continuous estimation of cardiac output. Curr Opin Crit Care 2002; 8: 257-61 4. Su NY, Huang CJ, Tsai P, Hsu YW, Hung YC, Cheng CR. Cardiac output measurement during cardiac surgery: esophageal Doppler versus pulmonary artery catheter. Acta Anaesthesiol Sin 2002; 40: 127-33 5. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307-10 6. Krouwer JS, Monti KL. A simple, graphical method to evaluate laboratory assays. Eur J Clin Chem Clin Biochem 1995; 33: 525-27 7. Singh A, Juneja R, Mehta Y, Trehan N. Comparison of continuous, stat, and intermittent cardiac output measurements in patients undergoing minimally invasive direct coronary artery bypass surgery. J Cardiothorac Vasc Anesth 2002; 16: 186-90 Hogue CW Jr, Rosenbloom M, McCawley C, Lappas DG. Comparison of cardiac output measurement by continuous thermodilution with electromagnetometry in adult cardiac surgical patients. J Cardiothorac Vasc Anesth 1994; 8: 631-35 Della Rocca G, Costa MG, Pompei L, Coccia C, Pietropaoli P. Continuous and intermittent cardiac output measurement: pulmonary artery catheter versus aortic transpulmonary technique. Br J Anaesth 2002; 88: 350-56 Buhre W, Weyland A, Kazmaier S, et al. Comparison of cardiac output assessed by pulse-contour analysis and thermodilution in patients undergoing minimally invasive direct coronary artery bypass grafting. J Cardiothorac Vasc Anesth 1999; 13: 437-40 Yamashita K, Nishiyama T, Yokoyama T, Abe H, Manabe M. Cardiac output by Pulse CO is not interchangeable with thermodilution in patients undergoing OPCAB. Can J Anesth 2005; 52: 530-34 Manecke GR. Edwards FloTrac sensor and Vigileo monitor: easy, accurate, reliable cardiac output assessment using the arterial pulse wave. Expert Rev Med Devices 2005; 2: 523-27

8.

9.

10.

11.

12.

ACA-06-110- OA.p65

126

6/28/2007, 9:47 AM