Pharmacology of Crystalloids and Colloids

INTRODUCTION
Blood volume is a critical factor in maintaining hemodynamic equilibrium and
tissue oxygenation. Intravascular volume is regulated very closely by means of
several complex mechanisms, for which the onset of action varies widely. In
some situations, such as acute bleeding, sepsis or with the use of certain drugs,
the body must withstand absolute or relative changes in blood volume that
cannot be immediately compensated for by the regulatory mechanisms. In these
situations, the main goal of volume therapy is to temporarily increase plasma
volume until the bodys own mechanisms can correct the hypovolemia.Treatment
of hypovolemia has changed significantly in recent years. In the past, fresh
frozen plasma (FFP) or its equivalent was long the volume expander most
commonly used. Now, indications for FFP are limited to the correction of some
hemostatic disorders. A by-product of this legitimate change in practice was
increased use of human albumin. Because of the financial consequences of this
strategy, several consensus conferences have issued recommendations on the
best indications for the use of various plasma volume expanders.Despite these
recommendations, the choice of the appropriate agent in the treatment of
hypovolemia has not yet been settled. The debate on crystalloids versus colloids
continues, besides a debate on the choice of colloid.

HYPOVOLEMIC SHOCK

Absolute hypovolemia may be defined as a reduction in the normal blood

volume. Initially, the body responds by activating neurohormonal compensatory
mechanisms. These mechanisms are designed to counteract the drop in arterial
pressure and maintain blood flow to the vital organs at the expense of the
muscular, cutaneous and splanchnic tissue beds which are then under-perfused.
Subsequently, the true volume-regulating mechanisms take effect. Depending on
the extent of the volume loss, hypovolemia may be compensated and changes in
hemodynamic status minimal or undetectable, or it may be uncompensated and
produce objective hemodynamic signs.Relative hypovolemia occurs when
increased arterial and venous compliance leads to a reduction in venous return.
Relative hypovolemia is involved in the pathophysiology of septic shock and
anaphylactic shock. In shock associated with infection and anaphylaxis and in
the late phase of hemorrhagic shock, capillary permeability also increases to
varying degrees. Reduced blood volume leads to reduced venous return and thus
to reduced cardiac output. The hemodynamic and neuro-hormonal responses to
acute hypovolemia have two phases which occur progressively in conscious
humans <|[1]|>. During the first phase, mechanisms regulating blood pressure
intervene rapidly to maintain blood pressure. Baroreceptor-mediated responses
produce peripheral adrenergic vasoconstriction that compensates for the fall in
cardiac output and keeps arterial pressure near normal. Stimulation of the renin-
angiotensin system augments the sympathetic system effect, although
vasopressin and the adrenal catecholamines have little direct involvement at this
point. Sympathetic vasoconstriction is accompanied by a complex vascular
redistribution which initially favors cerebral, coronary and renal circulation and
subsequently only cerebral and coronary circulation. Vasoconstriction affects
mainly the muscular, cutaneous and splanchnic tissue beds. If a critical level of
hypovolemia is reached and remains untreated, the second phase of the
physiologic response occurs abruptly. It is characterized by absolute or relative
bradycardia and a profound fall in arterial pressure with a sudden drop in
systemic resistance. This phase occurs when blood volume has dropped by more
than 30% to 50%. The drop in blood pressure is independent of bradycardia since
atropine will counteract the bradycardia but does not restore blood pressure <|
[2]|>. Several phenomena are associated with this phase <|[1]|>. The most
important appears to be a central inhibition of the sympathetic activation seen
during the first phase. During this hypotensive phase, medulloadrenal secretion
of adrenalin becomes significant. The renin-angiotensin system becomes highly
active, releasing angiotensin II and significant amounts of vasopressin. It appears
that receptors in the cardiopulmonary system are responsible for the
development of this sympathetic-inhibiting phase. The bradycardia is the result
of a vagovagal reflex loop due to stimulation of intracardiac mechanoreceptors.
Activation of these receptors is also thought to cause vasodilatation by inducing
central inhibition of the sympathetic system. It has been suggested that these
mecha-noreceptors are stimulated by mechanical distortion of the left ventricle,
which has a practically non-existant volume at end of systole. Bradycardia could
allow for better diastolic ventricular filling in conditions of extreme hypovolemia.
Central serotoninergic and opioid mechanisms have also been implicated in this
phase of hemorrhagic shock . Major fluid shifts occur during hypovolemic shock
as the body attempts to restore plasma volume. Hormonal responses (the renin-
angiotensin system, vasopressin) are actively involved in this phase, but are not
effective in the short term. Sympathetic vasoconstriction reduces capillary
hydrostatic pressure and thus facilitates movement of interstitial fluid into the
plasma compartment. Secondarily, however, changes in cell membranes result in
a capillary plasma leak into the interstitial space with an accumulation of non-
exchangeable water in some parts of the interstitial space (third-space losses)
and the intracellular environment. When hypovolemic shock is prolonged, a
number of abnormalities develop that may become irreversible and promote the
secondary development of a multiorgan failure syndrome. Severe trauma is
responsible for a major inflammatory reaction triggered by reperfusion of
ischemic tissue and tissue attrition <|[3]|>. Ischemia-reperfusion injury occurs as
oxygen metabolites activate macrophages and neutrophils. Tissue attrition
occurs as tissue factors and the complement system become involved. As well,
various mechanisms lead to an irreversible reduction in the microcirculatory flow.
The edema of endothelial cells reduces the patency of capillaries, and
neutrophils adhere to the vascular endothelium. The gastrointestinal tract
appears to play a central role in ischemia-reperfusion phenomena and their
consequences at distant locations. Splanchnic circulation is very rapidly
sacrificed during shock, and the consequences of this under-perfusion may be
catastrophic, as myocardial depressant factors are released, bacteria are
translocated and endotoxemia is facilitated. Translocation of bacteria to portal
blood and mesenteric lymph nodes is rare in humans, but may be seen at an
advanced stage of shock <|[4, 5]|>. Impaired perfusion of the gastrointestinal
tract may persist following severe hypovolemia despite an apparent
normalization of systemic hemodynamics <|[6]|>. Increased peripheral oxygen
demand results from this severe reactive inflammatory syndrome and often is
accompanied by peripheral vasoplegia <|[7]|>. Inability to adapt to these
secondary impairments appears to indicate a poor prognosis. This inability is
reflected in lower cardiac output and arterial oxygen transport in the patients
who do not survive <|[8]|> and higher mortality in patients whose initial oxygen
consumption is low <|[9]|>.

FLUID COMPARTMENTS AND FLUID SHIFTS

Water accounts for a total of 60-70% of body weight and is contained in

three compartments: intracellular, interstitial and intravascular. Intracellular
water accounts for about 40% of body weight and 70% of total body water.
Extracellular water accounts for 30% of body weight and includes the interstitial
space (with about 23% of total body water) and the intravascular space (with 7%
of total body water). Despite significant differences in their ionic composition,
intracellular and extracellular fluids have the same osmolarity. Intracellular
osmolarity remains quite stable, whereas changes in extracellular osmolarity
result in the movement of water between these two sectors. Infusion of a
hypotonic solution will reduce extracellular osmolarity and thus result in the
movement of water into the intracellular compartment in proportion to the
electrolyte level gradient. Conversely, infusion of a hypertonic solution results in
an increase in extracellular osmolarity that leads to the movement of water out
of the intracellular space into the extracellular sector. The ionic composition of
interstitial and intravascular fluids is the same. Proteins, however, are kept with
the vascular compartment, at least to a degree, by the vascular wall, which is a
semi-permeable membrane. The concentration of proteins is much higher in the
plasma sector than in the interstitial sector. About 70% of the oncotic pressure of
plasma is created by albumin. Starling defined the factors that determine flows
of fluid between the intravascular and interstitial spaces. The overall result is
that the oncotic pressure gradient opposes the hydrostatic pressure gradient. As
a result, there is a net flow of water into the interstitial space that supplies the
lymphatic circulation. However, the capillary membrane is not totally
impermeable to proteins. Movement of protein depends on the membranes
coefficient of osmotic reflection, which in turn depends on the size, shape and
charge of the molecule.

ISOTONIC CRYSTALLOIDS

An isotonic solution does not alter extracellular osmolarity, has a volume

of distribution limited and equal to the extracellular sector, and increases plasma
volume by about 200 mL for every 1,000 mL given <|[10, 11]|>. Recent research
using a pharmacokinetic model adapted for fluid spaces suggests that the
volume of distribution is more limited, i.e. about twice the plasma volume, and
especially that elimination is significantly slower when hypovolemia is present
prior to administration <|[12, 13]|>. These findings could explain why
crystalloids have been found quite effective in the treatment of hypovolemia.
Isotonic solutions (physiological saline, Ringers lactate) are widely used as
volume expanders. Their use inevitably results in an increase in the volume of
interstitial fluid. The subsequent increase in lymphatic return is responsible for
interstitial albumin being drawn into the plasma sector. Because of these effects,
along with a lack of adverse effects, crystalloid solutions are the first choice for
intravenous infusion solutions. The efficacy of crystalloids, however, is limited in
some situations. Cervera et al. showed that the amount of crystalloid infusion
required to maintain normovolemia when compensating for blood loss during
bleeding that gradually resulted in a loss of 15% to 40% of blood volume was
equal to 5 times the volume of blood lost <|[14]|>. For greater losses, the
amount of crystalloid required to maintain normovolemia increases exponentially
<|(Figure 1)|>. These large amounts are required because of the accumulation of
fluid in part of the interstitial space that is not completely drained by the
lymphatic circulation, the rate of which reaches its limit. Furthermore, even
massive use of crystalloid is not effective in restoring normal microcirculatory
perfusion. In an animal study of normovolemic hemodilution, infusion of an
amount of Ringers lactate equal to 4 times the amount of blood withdrawn did
maintain arterial pressure and filling pressures <|[15]|>. However, tissue
perfusion as determined by counting the number of capillaries perfused was
significantly reduced by 62% and tissue oxygen partial pressure was reduced
from 19 to 8 mm Hg <|(Figure 2)|>. In a group of animals treated by exchange
with an equal volume of dextran, capillary counts and tissue oxygenation did not
change. Wang et al. studied the quality of reperfusion using laser Doppler in a
hemorrhagic shock model <|[16]|>. Microvascular blood flow was significantly
reduced during hemorrhagic shock and remained depressed in all organs
examined for the two hours following resuscitation using an amount of Ringers
lactate equal to 4 times the volume of blood withdrawn, which doubled central
venous pressures compared to baseline, pre-shock values <|(Figure 3)|>. These
authors also studied hepatic clearance in similar experimental conditions and
showed that this function did not return to normal despite volume expansion with
Ringers lactate <|[17]|>. They concluded that the persistence of such
microvascular and hepatic impairment could be the basis of organ failure
following hemorrhagic shock. Controversy persists, however, with some authors
recommending colloids and others crystalloids, using arguments essentially
based on measurements of hemodynamic parameters. However, as has been
shown, these data are not adequate to judge the quality of resuscitation. There
has been no study with sufficient statistical power to compare the various fluids
in terms of morbidity or mortality. Several meta-analyses have been undertaken
with the aim of assessing the impact of choice of volume expander on mortality.
The first such study was published by Velanovitch in 1989 <|[18]|>. This analysis
combined the mortality data from 8 earlier published studies in which the
efficacy of colloid was compared to crystalloid in resuscitation from traumatic or
septic shock. Overall, the analysis showed a 5.7% difference in mortality rate in
favor of crystalloids. When only studies on shock in trauma patients were
analyzed, there was a 12.3% difference in mortality rate in favor of crystalloids.
Conversely, when only the studies on resuscitation of non-trauma patients were
included, there was a 7.8% difference in mortality rate in favor of colloids. The
authors found that colloids did not appear to have any advantages in
resuscitating trauma patients, but in patients with capillary permeability
disorders, colloids could be more effective. A second meta-analysis was
published recently by Schierhout et al. <|[19]|>. It focused on intensive care
patients. Of the 37 studies selected for inclusion, 26 compared crystalloids to
colloids (n = 1,622) and 11 studied hypertonic solutions (n = 1,460).
Resuscitation with colloids was associated with an increased absolute risk of
mortality of 4% (95% confidence interval: 0% to 8%). The authors did not find
any differences regardless of the initial pathology, notably in trauma patients.
Overall, these meta-analyses consistently find crystalloids to be preferable.
However, they do have important limitations that deserve mention: the studies
included for analysis were often very old, and the treatment used then is no
longer the standard of care; the colloids used were previous generation products,
not modern colloids; and, these meta-analyses included a very diverse array of
pathologies. In short, the issues raised by the meta-analyses are relevant, but
the answers they provide are not satisfactory given the inadequacies of the
original studies. Crystalloids remain the volume replacement solutions of first
choice, and their use provides essential benefit in the compensation of blood
losses by correcting water and sodium deficits in the interstitial compartment. In
the context of major hypovolemia, however, their use is not appropriate since
the amounts required do not allow for adequate treatment of shock and
especially of microcirculatory disturbances, a key factor in limiting the effects of
shock on tissues.

HYPERTONIC CRYSTALLOIDS

Hypertonic crystalloids (2500 m0sm/L), often in combination with colloids,

have been studied botexperimentally and clinically. They produce increased
osmolarity in extracellular fluid which results in movement of water from the
intracellular compartment to the interstitial and vascular compartments <|[20]|
>. Plasma volume expansion, however, does not suffice to explain the efficacy of
hypertonic solutions, particularly as this effect is transitory. Their sustained
duration of action is due to reflex mechanisms responsible for venous
constriction and increased cardiac contractility <|[21]|>. The pathways and
centers of these reflex mechanisms continue to be discussed. The presence of
intact vagus nerves and the orthosympathetic system appears essential. The
increase in cardiac contractility could also be a reflex or it could result from a
direct action of hyperosmolarity on myocardial fibers <|[22]|>. Hyperosmolarity
vasodilates the splanchnic and renal tissue beds <|[23]|>, an effect that is
particularly useful in the setting of hemorrhagic and septic shock. As well, the
shrinking of erythrocytes and of swollen vascular epithelium could facilitate
tissue oxygenation <|[20, 24]|>. Furthermore, hypertonic saline solutions
decrease intracranial pressure <|[25]|>. Hypertonic solutions should be
considered to be not only volume expanders, but also real drugs with systemic
and regional cardiovascular effects. Several trials have compared initial
resuscitation of trauma patients with hypertonic saline combined with dextran to
resuscitation with isotonic crystalloid alone, but no significant difference in
mortality has been found. A meta-analysis was published recently <|[26]|>,
which included 14 different randomized studies, 8 trials of hypertonic saline with
dextran and 6 trials of hypertonic saline. In each case, control patients were
given isotonic crystalloid. Overall, the use of hypertonic saline alone for
resuscitation from hemorrhagic shock was not associated with improvement in
mortality rates. However, a trend to improvement was found with the
combination of hypertonic saline and dextran. The difference in the survival rate
was 3.5% and the odds ratio estimated to be 1.20 in favor of the combination
(95% confidence interval: 0.94 to 1.57). Ultimately, the authors found that
treatment of hemorrhagic shock with hypertonic saline and dextran was
equivalent or superior to standard treatment with crystalloid. However,
hypertonic solutions have one major drawback: they cannot be administered
repeatedly because of the hypernatremia that results after a single dose. In
cardiovascular surgical intensive care, several trials have compared the use of
hypertonic saline combined with hydroxyethyl starch to standard treatment with
colloids <|[27-29]|>. The use of the combination solutions prior to cardiovascular
bypass appears to have several advantages. The volume needed to maintain or
increase filling pressures was clearly less with hypertonic solutions <|[28, 29]|>.
The total volume of fluids administered during the procedure was significantly
less in the groups treated with the hypertonic saline-starch solution. A negative
fluid balance compared to the use of standard colloids could explain the better
pulmonary gas exchange found with this therapy <|[29]|>. Higher cardiac output
after coming off bypass was also observed <|[28]|>. This effect may be related
to increased venous return <|[30]|>, arterial vasodilatation which facilitates
cardiac ejection <|[31]|>, improvement of myocardial contractility <|[22]|>, and
increased activity of the orthosympathetic system <|[32]|>. Tolerance of
hypertonic-hyperoncotic solutions in cardiovascular surgical trials has generally
been good. Some episodes of arterial hypotension or rhythm disturbances
related to overly rapid administration have been reported however <|[27, 33]|>.
Hypertonic saline solutions combined with dextran or hydroxyethylstarch are
available in several European countries and are currently in the process of being
registered in France.

HUMAN ALBUMIN

Human albumin is a protein composed of a simple chain of amino acids

with a quaternary helix-like structure. The center of the molecule is made up of
hydrophobic radicals which are binding sites for many ligands. The outer part of
the molecute is composed of hydrophilic radicals. Albumin is a relatively small
molecule in terms of space, but its size is sufficient to prevent it from crossing
the capillary membrane. At physiological concentrations (40-45 g/L), albumin
accounts for 70% of plasmas oncotic pressure, or about 18 to 22 mm Hg of which
5 to 9 mm Hg are related to the Donnan effect (the molecules electrically neutral
charge). Albumins binding capability affects many different areas. For instance,
albumin can bind the products of degradation processes such as bilirubin or free
fatty acids. Through binding, it can regulate the ionized fraction of some cations
such as calcium and magnesium, trap free radicals and serve as a transport
protein carrying amino acids and drugs. Body stores of albumin are about 4.5 to
5.0 g/kg and are divided between the intravascular and interstitial
compartments. About a third of the store of exchangeable albumin is located in
the intravascular compartment, and the rest is in the interstitial compartment.
Whereas distribution of albumin is uniform within the intravascular compartment,
its distribution within the interstitial space is uneven depending on the tissue
because of large polysaccharides, glycosaminoglycans. Albumin distribution and
metabolism have been studied using various models <|[34-37]|>. Following an
intravenous injection, plasma activity falls gradually with a distribution half-life of
4 h, which corresponds to the time necessary for equilibration with the interstitial
sector, and an elimination half-life of 17 to 18 days which corresponds to the
time for catabolism by the reticuloendothelial system. The normal rate of
transcapillary escape is 5%/h, meaning that every hour, 5% of the intravascular
albumin is transferred to the interstitial sector. The rate of transcapillary
exchange can be higher in many disease states such as hypertension, heart
failure or sepsis. About 10% of albumin is catabolized, meaning that about 14 g
of albumin are broken down daily. To maintain serum albumin levels, mobilization
of extravascular albumin by the lymphatic circulation and albumin synthesis
must compensate for losses by transcapillary passage and catabolism. Every
day, about 90% of the extravascular albumin returns to the intravascular
compartment via the lymphatic circulation. Under physiological conditions, the
liver does not store albumin; instead, synthesis compensates for metabolic
losses. Synthesis of albumin occurs exclusively in the liver and takes 1 to 2
minutes. Excretion of albumin takes about 30 minutes. Regulation of albumin
synthesis is complex and involves many different hormones along with the
plasma colloid osmotic pressure as an essential regulatory factor. Abnormalities
of water and albumin distribution in the postoperative period have been
documented <|[35, 36, 38-40]|>. These abnormalities are particularly
pronounced after major abdominal, cardiac or vascular surgery. During the
postoperative period, the reduction in plasma volume is greater than that of red
cell volume, which indicates a plasma leak into the interstitial space <|[39]|>.
Actually, the reduction in plasma volume is associated with a reduction in
exchangeable interstitial water because of the development of edema which is
not drained by the lymphatic circulation. The reduction in exchangeable
interstitial water diminishes lymphatic output, thus limiting the mobilization of
extravascular albumin. Reduced lymphatic output is a first factor leading to
postoperative hypoalbuminemia. Increased capillary permeability, however, is
the main mechanism responsible for hypoalbuminemia since it tends to
equilibrate albumin concentrations of the the vascular and interstitial
compartments <|[41]|>. In some inflammatory or septic contexts, increased
catabolism of albumin may be a third cause of hypoalbuminemia. A number of
studies have been performed to measure the volume expansion properties of
albumin <|[34, 37]|>. It is generally accepted that the theoretical volume
expanding power of albumin is 18 to 20 mL per g, or 400 mL for a 500 mL
infusion of a 4% albumin solution. Actually, volume expansion differs depending
on the population studied, and the patients hydration status and body stores of
albumin. In patients undergoing minor surgery <|[34]|>, infusion of 50 g of
albumin resulted in an increase of plasma volume of about 500 mL. The same
degree of volume expansion was observed regardless of the concentration of the
albumin solution used, i.e. 1,000 mL at 5%, 250 mL at 20% or 200 mL at 25%.
Schwartzkopff et al. measured volume expansion in surgical patients and healthy
volunteers <|[37]|>. Based on their findings, the authors determined that 5 to 8
g of albumin are lost from the intravascular space per hour. Albumin leak is
greater in normovolemic, hypoprotidemic subjects. Duration of volume expansion
is extremely complex to study and depends on blood volume, protein levels and
capillary permeability, among other factors. Albumin is often used to treat
hypoalbuminemia in intensive care patients <|[42]|>. The administration of
albumin, however, does not improve morbidity or mortality <|[43, 44]|>. Golub
et al. performed a prospective, randomized study of 219 surgical intensive care
patients comparing albumin supplementation to non-supplementation <|[45]|>.
The complication rate was 44% in the albumin group, compared to 36.9% in the
control group; the mortality rate was 10.3% in the albumin group vs 5.8% in the
control group (p = 0.22). No significant difference was found in the length of time
on ventilatory support or length of stay in intensive care. The authors found that
albumin replacement therapy offered no benefit in this context and should be
abandoned. A recent British meta-analysis <|[46]|> found very negative results
with the use of albumin as a volume expander. This analysis included 30
randomized studies and a total of 1,419 patients treated for hypovolemia, burns
or hypoprotidemia with either albumin or similar products (plasma protein
fractions), compared to crystalloids. The pooled results showed an excess risk of
death of 6% (95% confidence interval: 3% to 9%) with a relative risk of 2.40
(95% confidence interval: 1.11 to 5.19). The authors found that there was no
evidence of the efficacy of albumin over crystalloid; instead, albumin appeared
to carry additional risk. As a result, the authors called for rigorously conducted
randomized trials to determine the appropriate indications for human albumin.
This study was widely criticized in letters to the editor published in the British
Medical Journal. Criticism mainly focused on the poor selection of studies, i.e. the
inclusion of studies of not only albumin but also fractionation products of varying
quality, and the use of mortality as a main study outcome measure. While these
criticisms are valid, doubts remain and merely serve to confirm the position
taken by the consensus conferences that use of albumin should be restricted to a
few well-defined situations and albumin used as a second-choice infusion
solution when other products are not indicated, are contraindicated or have been
used up to their maximum dose. Albumin can now be extracted only from plasma
collected by phlebotomy. In the near future, it will probably be obtained from
transgenic animals. Most techniques currently used are based on the Cohn
principle and involve sequential fractionation of plasma proteins. Concentrated
preparations of 4%, 5%, 20% or 25% albumin solutions are pasteurized. The
frequency of anaphylactoid reactions is significantly lower than that seen with
the other colloids, except for hydroxyethyl starch <|[47]|>. The incidence of
reactions, which are of the chills-and-fever type, appears to be largely related to
the degree of product purity. Reports of other side effects, such as a negative
inotropic effect, aluminum overload or hemostatic effects, are anecdoctal or
controversial.

DEXTRANS
 Of all the colloids used in clinical practice, dextrans are those with which
we have the most experience. The physical, chemical and pharmacological
properties of dextrans are particularly well known and among the most studied of
all plasma substitutes. The use of dextrans has been declining noticeably in
Europe because of their side effects. Dextran is a single-chain polysaccharide of
bacterial origin. The average molecular weight of these of variably dispersed
solutions is an important product characteristic. The average molecular weight in
weight (Mw), which is the arithmetic mean of molecular weights of the
constituent particles, is different from the molecular weight in number (Mn),
which is the average molecular weight of the particles with colloid oncotic power.
The main types of dextran solutions are designated according to their Mw:
70,000 D (dextran 70), 60,000 D (dextran 60) and 40,000 D (dextran 40).Dextran
70 and 60 are generally prepared as 6% solutions, while dextran 40 is available
in a 10% concentration. The colloid oncotic power of the various dextran
solutions is very high: 1 g of dextran 40 retains 30 mL of water and 1 g of
dextran 70, 20 to 25 mL of water. Following intravenous administration, dextran
is eliminated by three routes. Most is eliminated by the kidneys. A smaller
fraction enters the interstitial space and returns to the bloodstream via lymphatic
drainage or is metabolized by certain organs. A third, even smaller fraction is
eliminated via the gastrointestinal tract. The variety of routes of elimination and
the influence of Mw on most of these routes mean that the pharmcokinetics of
dextrans are very complex. Briefly, following intravenous administration of
dextran 40, half of the dose given is eliminated within 2 h and 80% within 6 h.
Following intravenous administration of dextran 70, 50% of the dose infused is
eliminated within 24 h. The rheologic effect of dextran 40 solutions is especially
pronounced since these solutions reduce viscosity of whole blood more for the
same degree of hemodilution compared to other plasma substitutes <|[48]|>.
Following dilution with various colloid plasma substitutes, low shear-rate viscosity
is reduced with dextran 40, whereas it is increased with dextran 70 and dextran
60. Dextran 40 solutions also have a beneficial effect on red cell rouleau
formation since they increase the time of red cell aggregation, unlike dextran 60
et 70. A new finding regarding beneficial rheologic effects of dextrans involves
leukocyte adherence, an effect that may be useful in the setting of ischemia-
reperfusion injury <|[49]|>. Longer bleeding time is generally seen after
administration of more than 1.5 g/kg of dextran <|[50]|>. This effect is greater
with high molecular weight dextrans. Reduction of platelet adhesiveness is
related to a reduction of factor VIII which acts as a cofactor for ristocetine
aggregation activity, likened to the von Willebrand factor. Hemostatic
abnormalities induced by dextrans are similar to those seen in type I von
Willebrand syndrome. This explains why this side effect can be reversed by the
administration of desmopressin. Dextran also impairs the polymerization of
fibrin. Onset of oliguric or anuric kidney failure is exceptional and has been
reported exclusively with the use of 10% dextran 40 solutions. Analysis of most
published case reports has revealed contributing factors such as age, repeated
infusion of large quantities and arteritis <|[51]|>. Experimental models have
enabled us to understand the pathophysiology of this type of kidney failure,
which is an acute hyperoncotic syndrome. Anaphylactoid reactions appear to be
frequent with dextrans <|[47]|>. The mechanisms of allergic reactions have been
thoroughly studied, particularly by Hedin and Richter <|[52]|> who showed the
role of dextran antibodies. These pathophysiological findings are the basis for
preventing allergic reactions to dextrans by means of hapten inhibition <|[53]|>.
Several studies have shown that prior injection of 20 mL of dextran 1,000 D
(Promit) a few minutes before a dextran infusion will considerably reduce the
incidence of reactions, especially severe reactions. This step should always be
taken before infusion of any type of dextran. Furthermore, signs of acute fetal
distress due to uterine hypertonia have been described when dextrans have
been used during delivery <|[54]|>. Dextrans are therefore absolutely
contraindicated in this setting. Use of dextrans is declining in most European
countries because of dextrans side effects.

GELATINS

The use of gelatins as volume expanders was described in the treatment

of hypovolemic shock as early as 1915. It was not until the 1950s, however, that
gelatins became available for clinical use, including the current products,
modified fluid gelatins and urea-bridge gelatins. Gelatins are a product of bovine
origin. Consequently, there is the problem of the potential risk related to the
pathogen involved in spongiform encephalopathy. The French national agency for
develoment of medical evaluation (ANDEM) has made the following comments,
in accord with the viral safety group of the government drug agency and an
opinion issued by the European Medicine Evaluation Agency. Three factors
combine to contribute to the safety of gelatins used in pharmaceuticals:
manufacturers must not use raw material from the United Kingdom; the tissues
used as raw material are classified as not having any detectable level of
infectiousness; the method of preparation which includes extended acid and
alcaline processing and filtration is sufficient to eliminate any risk. Nevertheless,
although gelatin has never been implicated in the transmission of known non-
conventional disease agents, the biological risk can never be zero. The residual
risk, however small it may be, must therefore be taken into account in a
risk/benefit analysis. The pharmacokinetics of gelatins is not fully understood.
Data on all products are fragmentary. Gelatins are cleared essentially by
glomerular filtration <|[55]|>. About 20 to 30% of the dose given travels from
the intravascular compartment into the interstitial space. From there, it can
return to the bloodstream via the lymphatic circulation. Gelatins may also be
broken down by proteases into small peptides and amino acids in the
reticuloendothelial system <|[56]|>. Gelatins do not accumulate in the body <|
[57]|>. These data indicate that the plasma half-life of both urea-bridge gelatins
and modified fluid gelatins is 2-3 hours <|[58]|>. The various gelatin solutions
have comparable volume-expanding power. Using old data <|[59]|>, it was
estimated that in the most favorable case, which is a hypovolemic subject
without any disturbance of capillary permeability, a 500 mL infusion of gelatin
would increase plasma volume by a volume equal to the amount infused. Four
hours later, volume expansion would be only half that much. Other studies show
more limited volume expansion, often much less than the volume infused, due to
rapid passage of the gelatin into the interstitial space. In a comparative study in
healthy volunteers, Kolher et al. found less volume expansion with gelatin than
with dextran 40 or low molecular weight starches <|[60]|>. Loss of volume
expansion was also found to be faster than with the other expanders. A
randomized study compared gelatin to albumin in intensive care patients <|[61,
62]|>. Despite the difference in albumin levels, no significant difference in
pulmonary edema, kidney failure or mortality was found between the 2 groups.
Plasma albumin levels were significantly lower in survivors than in patients who
died. It was long thought that gelatins had no side effects on hemostasis
regardless of the daily dose. However, a recent study performed in the setting of
progressive hemodilution showed that compared to physiological saline, gelatin,
like hydroxyethyl starch and albumin, induced significant changes in the
thromboelastrogram that compromised coagulation <|[63]|>. Another study
showed that urea-bridge gelatins (Haemaccel) result in greater reduction of
platelet aggregation than do modified fluid gelatins <|[64]|>. The incidence of
allergic reactions with gelatins is higher than that seen with hydroxyethyl
starches and comparable to that with dextrans <|[47]|>. The mechanism
involved in these reactions is not as well understood as that associated with
dextrans. In conclusion, clinical use of gelatins is declining because of their
modest effectiveness as volume expanders and because their theoretical
advantages are gradually being outweighed by their effects on coagulation.

HYDROXYETHYL STARCHES

Hydroxyethyl starches are modified natural polysaccharides. Solutions of natural

starch are unstable and are rapidly hydrolyzed by -amylase. Hydroxylation or
etherification are used to stablize the solution and slow hydrolysis <|[65]|>, and
increase the molecules hydrophilia considerably and expand its conformation.
The extent of hydroxyethylation may be measured by two features: the degree
of substitution and molar substitution ratio. This second characteristic takes into
account the di- and tri-substitutions that occur with some molecules of glucose
and better reflects the starchs resistance to hydrolysis by -amylase. The site of
hydoxyethylation on the glucose molecule is preferentially C2, but etherification
at C3 or C6 is also possible. Hydroxyethylation at C2 gives the most resistance to
-amylase. The ratio of C2/C6 reflects the types of hydroxyethylation. An
important characteristic of these products is also molecular weight in weight
(Mw) and molecular weight in number (Mn). However, molecular weight is not
the parameter that determines the starchs pharmacokinetics, which depend
mainly on the degree and type of hydroxyethylation. However, molecular weight
is a major determinant of the solutions side effects. The first hydroxyethyl starch
was marketed in Germany and the United States and had a high Mw (450 kD).
However, this starch had side effects on hemostasis that led to its being
withdrawn from the market. Other starches with a lower molecular weight have
now been developed. In France, the main products are Elohes , Lomol , Heafusine
and Haes-Steril. These products have similar although differing characteristics.
Elohes is a 6% solution, has a Mw of 200 kD and a molar substitution rate of
0.62. Lomol is a 10% solution, has a Mw of 250 kD and a molar substitution rate
of 0.45. Haes-Steril and Heafusine have similar although not identical
characteristics to Lomol and are 6% solutions. Unlike the dextrans, the
pharmacokinetics of hydroxyethyl starches are not influenced mainly by Mw, but
depend mainly on the degree of hydroxyethylation <|[66, 67]|> <|(Figure 4)|>.
The main route of elimination of hydroxyethyl starches is urinary. A fraction is
taken up by the reticuloendothelial system where the starch is slowly broken
down. The rate of uriniary excretion in the 24 h following administration of
hydroxyethyl starch depends mainly on the degree of hydroxylation. For Elohes,
the elimination half-life of the medium-size molecules is 7 h and 5 days for the
large molecules. For Lomol, the elimination half-life of the medium-size
molecules is 3 h and 2 days for the large molecules. Actually, the usual data are
poorly suited to describing the pharmacokinetics of these variably dispersed
solutions, since they describe the average kinetics of the solution rather than of
its various fractions. It has been shown that after an infusion of hydoxyethyl
starch, the dispersion of the Mw molecules changes, first because the smaller
molecules are rapidly eliminated and then because the large molecules are
partially hydrolyzed to become medium-sized molecules. This partial hydrolysis
tends to increase or stabilize plasma volume expansion over time. This
phenomenon is reported to predominate for 2 to 4 h following infusion.
Intravascular hydrolysis by -amylase is more limited with some hydroxyethyl
starches because of their high degree of hydroxyethylation <|[68]|>. Tissue
distribution of these starch solutions has been studied in an animal model. The
reticuloendothelial system, including the spleen, accumulates hydroxyethyl
starch for a long time and catabolizes it gradually by means of maltases and the
sucrase-isomaltase complex. Metabolism of hydroxyethyl starches has been
studied by monitoring blood glucose and urine glucose, and no change has been
seen <|[69]|>. In animals, the molecular weight of molecules excreted in urine is
low, but it is much higher than the weight of glucose molecules. In vitro, the
addition of hydroxyethyl starch to serum or solution containing amylase does not
result in any increase in the glucose level. Together, these data indicate that
metabolism of hydroxyethyl starches occurs by means of the production of
increasingly smaller molecules down to a weight of about 40,000 to 50,000 D, at
which point the molecules are small enough to be excreted in urine, without
metabolism continuing to the point of formation of glucose or hydroxyethyl
glucose. Based on these data, it may be concluded that hydroxyethyl starch
infusions do not change blood glucose levels. However, cases of hyperglycemia
and even glycosuria have been reported in non-insulin-dependent diabetics. Of
course, volume expanders are used in patients under conditions (shock, surgery)
in which other factors that disturb glucose metabolism are already present, such
that hydroxyethyl starch is not necessarily responsible for disturbing glucose
metabolism. Given the number of hydroxyethyl starches on the market and the
variety of their physical and chemical characteristics, comparison of different
products is difficult. A classification by in vitro Mw, i.e. high Mw (450 kD),
medium Mw (200 kD) and low Mw (70 kD) does not take into consideration the
degree of hydroxyethyl substitution or the C2/C6 ratio. It would make more sense
to compare hydroxyethyl starches according to their in vivo Mw after partial
hydrolysis of the original solution <|[66]|>. The in vivo Mw depends on the
original Mw, the extent of hydroxyethylation and the C2/C6 ratio. The higher the
values for all three of these characteristics, the higher the in vivo Mw. This
approach would allow for easy comparisons between products since a single
feature could be used to differentiate one hydroxyethyl starch from another.
Furthermore, in vivo Mw is the key parameter for evaluating colloid osmotic
power, pharmacokinetics, accumulation in plasma and tissue and side effects on
coagulation and renal function. Colloid osmotic power depends on the number of
molecules present, a value that can be determined by dividing the mass
concentration by the average in vivo Mw. If we look at two hydroxyethyl starches,
one of which has an in vivo Mw that is half that of the others, this means that for
the same concentration, the solution with the smaller Mw has twice the colloid
osmotic power of the other. In other words, for the solution with the smaller Mw,
half of the concentration would suffice to produce an equivalent effect <|[70]|>.
As well, a lower Mw means that the solution will be cleared more rapidly, and
less will accumulate in the plasma and the reticuloendothelial system. Side
effects on coagulation and perhaps renal function also depend on the in vivo Mw
and plasma concentration. The lower the in vivo Mw, the less starch accumulates
in plasma in the event of repeated administration and the fewer the disturbances
of coagulation <|[71, 72]|>. It would seem that the best hydroxyethyl starch is
the one will the lowest in vivo Mw above the threshold of renal eliminiation,
which is 50-60 kD. The in vivo Mw of Elohes is 140-150 kD <|[68, 71]|>, higher
than that of Haes-Steril, at 110-120 kD <|[71]|>. With Elohes, initial volume
expansion greater than the volume infused has been documented by several
studies on postoperative volume expansion or normovolemic hemodilution <|[65,
68, 73]|>. These studies confirm that the effectiveness of this solution equals or
surpasses that of human albumin in terms of volume expansion or cardiovascular
efficacy <|[68, 73]|>. The duration of action is about 24 h. Degrmont et al. ont
found that infusion of 500 mLof Elohes initially expanded volume postoperatively
by 693 mL. Volume expansion lasted for 24 hours, although the concentration of
Elohes gradually fell to 35% of peak concentration at 24 hours. In vivo Mw fell
within the hour following administration and then remained stable for 24 hours.
The albumin concentration initially fell following the Elohes infusion, and then
gradually rose until 24 hours. These findings indicate that although the products
duration of action is certainly related to its intravascular persistance, other
mechanisms are involved and help explain the long duration of action of Elohes,
i.e. intravascular hydrolysis that reduces the in vivo Mw and, even more
importantly, mobilization of interstitial albumin and renal adaptations.
Comparative studies that measure intravascular volume of other hydroxyethyl
starches are not available. As a result, only indirect comparisons may be made.
With products such as Lomol ou Haes-Steril, volume expansion equal to or
greater than the amount infused has been reported <|[74]|>. The duration of
action is not as long as that of Elohes, being about 6 hours. Kolher et al. <|[75]|>
compared a 200/0.5 hydroxyethyl starch in 6% and 10% solutions to a dextran
40 in 10% solution and a polygelatin in 5.5% solution <|(Figure 5)|>. This study
showed volume expansion of at least 6 hoursduration with this hydroxye-thyl
starch, longer than that of the gelatin which was eliminated in 3 hours. The
effectiveness of hydroxyethyl starch 200 has been shown in several clinical
settings: normovolemic hemodilution, intraoperative blood loss replacement,
cardiac surgery, and sepsis <|[24, 76, 77]|>. In intensive care, the concept of
impaired capillary permeability has often been used as an argument against
colloid use. The reasoning is that colloids do not remain in the intravascular
compartment; instead, they increase the interstitial oncotic pressure and
promote the development of edema. However, Rackow et al. studied 26 septic
patients and found a lower incidence of pulmonary edema in the group treated
with hydroxyethyl starch compared to the group treated with crystalloids <|[78]|
>. However, these results were obtained with hetastarch which has a particularly
high in vivo Mw. Using an experimental model of ischemia and reperfusion, Zikria
showed a reduction in the infarct and less myocardial edema in the group treated
with hydroxyethyl starch <|[79]|>. This study was performed using a particular
hydroxyethyl starch called pentastarch because the product mainly contains a
select category of Mw molecules. Various publications have confirmed the ability
of this particular product to reduce edema in experimental models of burns <|
[80]|>, ischemia-reperfusion injury <|[79, 81-84]|> or sepsis <|[85, 86]|>. Only
one published study describes the use of pentastarch in patients with sepsis, but
this study compared the product to albumin and did not look at pulmonary
edema. Regardless of experimental evidence, it cannot be concluded that
pentastarch has demonstrated ability to reduce edema clinically. As well, this
product is not yet on the market. Interestingly, however, the experimental
studies suggest that the anti-edema properties of pentastarch are not related to
the products colloid osmotic power, but to other properties that are not yet fully
understood. Other research with currently available medium Mw hydroxyethyl
starches, not pentastarch, provides evidence for this. For instance, using cultured
endothelial cells, Collis et al. showed that hydroxyethyl starch could inhibit
endothelial cell activation and limit adverse change in capillary permeability,
compared to albumin <|[87]|>. Hydroxyethyl starches might also reduce
adherence of leukocytes <|[76]|> which play an important role in ischemic-
reperfusion events. Schmand et al. found that hydroxyethyl starch had no
negative effects on cell-mediated immune functions and macrophage function
following resuscitation from hemorrhagic shock <|[88]|>. Eastlund et al. found
that cytokine release, chemotaxis and monocyte migration were not affected by
resuscitation with hydroxyethyl starch following hemorrhagic shock <|[89]|>. A
body of experimental evidence thus suggests that hydroxyethyl starches would
in fact have a beneficial effect on the inflammatory processes associated with
hypovolemic shock. A clinical study conducted on septic patients appeared to
support this indirectly by showing better splanchnic oxygenation as measured by
the gastric intramucosal pH in patients treated with hydroxyethyl starch, as
compared to those treated with albumin <|[77]|> <|(Figure 6)|>. Overall, the
evidence supports the use of hydroxyethyl starches in intensive care patients.
Their beneficial effects seem to be more related to their action on inflammatory
processes than to their colloid osmotic power. In a risk/benefit analysis of
hydroxyethyl starches, however, their side effects should also be weighed. The
effects of hydroxyethyl starches on hemostasis have been studied by many
authors <|[63, 67, 90, 91]|>. Hydroxyethyl starches with a high in vivo Mw
should be distinguished from other hydroxyethyl starches. Numerous cases of
abnormal bleeding have been reported with use of one such product, Hetastarch
<|[66]|>. Lengthened APTT and decreased levels of factor VIII and von
Willebrand factor have been documented <|[92-94]|>. These findings and case
reports of side effects would justify no longer using this product. Case reports
involving Elohes are more complex since occasional use does not result in
hematologic changes even at high doses (33 mL/kg) for 24 h <|[95]|>. Repeated
use for 10 days, however, has been associated with clearly abnormal laboratory
findings, including reduced levels of factor VIII and von Willebrand factor <|[66]|
>. Furthermore, one drug safety monitoring report contains about a dozen cases
of bleeding events related to repeated use of Elohes over several days, most
often in a neurosurgical setting. Most of these cases were characterized by
hematologic disturbances, particularly reduced levels of von Willebrand factor.
Following this report, the marketing authorization and labeling information were
amended to indicate that the use of Elohes should be limited to a maximum of 3
consecutive days. Treib et al. investigated the effects of various hydroxyethyl
starches on hemostasis extensively and showed the influence of in vivo Mw on
the type and extent of coagulation disorders. Coagulation abnormalities are seen
after repeated administration as part of 10-day hemodilution therapy with
products that have a high in vivo Mw compared to initial (in vitro) Mw, high
degree of hydroxyethyl substitution or high C2/C6 ratio <|[66, 71, 96]|>. The
typical product in this category is Elohes. Such products were found to
accumulate in the body with a gradual increase in plasma concentrations.
Adverse effects on coagulation parameters were proportional to plasma
concentration. Shortened thrombin time and decreased fibrinogen levels are
probably the result of accelerated polymerization of fibrinogen. Prolonged partial
prothrombin time is mainly the result of reduced factor VIII and von Willebrand
factor levels. The most likely mechanism for this effect is accelerated clearance
of factor VIII-von Willebrand complex after binding by hydroxyethyl starch
molecules. Decreased levels of factors XI and XII are seen only with very high in
vivo Mw hydroxyethyl starches. These coagulation abnormalities are particularly
pronounced after repeated administration of Elohes over 10 days and are minor
or non-existant with products of the Haes-Steril or Heafusine type <|[66]|>. In
summary, the data show that hydroxyethyl starches with a low in vivo Mw (Haes-
Steril or Heafusine) have little or no effects on hemostasis, even when given
repeatedly for 10 days. The same is not true for Elohes; its use in this way is to
be avoided. The effects of hydroxyethyl starches on kidney function have been
discussed in recent publications. Two types of situations should be distinguished:
the perioperative setting and the special case of kidney transplantation. The
acute hyperoncotic kidney failure syndrom was first reported with dextran use <|
[97]|>. This syndrome occurs when colloid osmotic pressure rises to a level
where it offsets the hydraulic pressure of glomerular filtration and thereby
suppresses urine output. Such a situation occurs when a high plasma level of
colloid is reached, generally after repeated administration. Anuria occurs more
readily in any situation in which renal perfusion pressure may decrease, such as
shock, arteriopathy, or renal artery stenosis <|[98]|>. This syndrome has now
been reported with nearly all colloids: gelatins <|[99]|>, dextrans <|[97]|>,
starches <|[100]|>, and concentrated albumin <|[101]|>. In the case of
starches, the development of this syndrome may theoretically be promoted by
the repeated administration of a hydroxyethyl starch with a high in vivo Mw
leading to a gradual increase in plasma levels. However, these products do not
appear to increase the risk of postoperative renal failure even when used in large
amounts intraoperatively in thoracic aortic or thora-coabdominal aortic surgery
<|[102]|> or orthopedic surgery <|[103]|>. In the setting of kidney
transplantation, osmotic-nephrosis-like damage has been seen on biopsy of
transplanted kidneys when the donor had been resuscitated with hydroxyethyl
starch. These lesions appear to be more frequent than in historical series in
which donors were not treated with hydroxyethyl starches <|[104]|>. A
prospective, randomized study comparing a gelatin to Elohes showed poorer
recovery of kidney function after transplantation and a greater number of
patients requiring hemodialysis in the groups with kidneys exposed to
hydroxyethyl starch <|[105]|>. The mechanism for impaired renal function might
be an accumulation of hydroxyethyl starch in the tubular cells. The presence of
hydroxyethyl starch in the osmotic-nephrosis-like lesions has not been shown,
however, and many other drugs, especially cyclosporin, can produce the same
type of damage. Furthermore, another study showed that the incidence of
osmotic-nephrosis-like lesions was apparently not influenced by the use of
hydroxyethyl starch <|[106]|>. As well, a multicenter retrospective study did not
confirm any adverse effects of hydroxyethyl starch on graft function after kidney
transplantation <|[107]|>. The immunological tolerance of hydroxyethyl starches
appears to be excellent <|[47]|>. The frequency of allergic reactions is lower
than with dextrans and gelatins. Severe reactions are particularly rare.

CONCLUSION

Crystalloids remain the intravenous solutions of first choice, and their use
provides essential benefit in the replacement of blood losses by correcting water
and sodium deficits in the interstitial compartment. In major hypovolemia, the
use of crystalloids is inappropriate since the amounts needed do not sufficiently
treat the shock, especially the microcirculatory disturbances, a key factor in
limiting the consequences of shock on tissues. Evidence is building that albumin
should be restricted to a few well-defined situations and be considered as a
infusion solution of second choice, for use when other products are not indicated,
are contraindicated or have been used up to their maximum dose. In particular,
it seems that the use of albumin to correct hypoalbuminemia should be
abandoned. Hydroxyethyl starches are the synthetic colloids with the
pharmcological properties that are the closest to natural colloids. In addition,
there is evidence to support the use of hydroxyethyl starches in intensive care
patients. Their beneficial effects appear to be related more to their action on
inflammatory processes than to their colloid osmotic power. In a risk/benefit
analysis of hydroxyethyl starches, their side effects should also be weighed. Side
effects are limited when hydroxyethyl starches with a low in vivo Mw are used.