Crystalloids versus Colloids: The Controversy

Crystalloids as well as colloids are commonly used in clinical practice to substitute fluid derangements in connection with surgical procedures or in the treatment of critically ill patients. Salt solutions will freely cross capillary membranes and equilibrate within the whole extracellular compartment. Within 20-30 min 75-80% of the infused volume will lodge mainly in the interstitial fluid space. Infusion of large volume of crystalloid is consequently required for correction of plasma volume deficits. Such a therapy includes a risk of tissue oedema formation, which at excessive fluid resuscitation, may impair vital organ function. The greater capacity of colloids to remain within the intravascular space results in a more efficient expansion of the intravascular volume, better haemodynamic stability and better haemorheological effects on microvascular blood flow and oxygen transport.



Meta-analytic assessments of randomised, controlled studies comparing crystalloid and colloid resuscitation indicate that colloid (especially albumin) resuscitation may increase mortality in critically ill patients with increased capillary permability. Non-trauma surgical patients, on the other hand, seem to benefit from colloid based resuscitation fluids, since colloids in general are more effective than crystalloides for optimising physiological variables related to flow in critically ill patients. The clinical relevance of data obtained from assessment of historical studies for present evidence-based practice of medicine may be questioned. Therefore, the colloid versus crystalloid debate will continue until well-controlled randomised clinical trials have been presented. In the meanwhile clinical fluid therapy should be based on the specific needs of each individual patient. Colloids are in most situations to be preferred when the main indication is to increase intravascular volume, while crystalloids are needed for correction of extravasacular fluid derangements.



INTRODUCTION



Fluid therapy with water and salts was first given in the 1830s for treatment of dying cholera patients and dramatic responses to the treatment were noted <|[1]|>. In spite of remarkable immediate effects of the fluid resuscitation on vital signs, the effects on outcome were less impressive at that time due to problems with electrolyte composition, tonicity, and sterility of the solutions. The use of saline fluid resuscitation in shock was initially described by the turn of the century while colloids were introduced much later <|[1]|>. Gelatin, the first artificial plasma substitute to be used clinically for shock treatment, was introduced in 1915 <|[2]|> and it was used rather extensively during World War I <|[3]|>. Gelatin was in the 1940s and 1950s followed by dextrans and later also hydroxyethyl starch preparations.



Increasing knowledge about disturbances in fluid homeostasis induced by trauma and blood losses was gained during the First and Second World Wars. It was, however, not until the 1940s and 1950s when, by the use of standardised animal models, the pathophysiology of hypovolaemia and the physiological importance of fluid resuscitation were understood more in detail <|[4]|>. At that time it also became obvious that pronounced acute internal fluid fluxes between the different compartments of the body constituted an important endogenous physiological response to trauma and hypovolaemia <|[5-11]|>. It was also recognised that these internal fluid derangements had to be compensated for in order to re-establish fluid homeostasis in the post-traumatic period. Since extravascular fluids were primarily involved, infusion of crystalloids rather than of colloids was suggested by Shires and co-workers <|[12, 13]|>.



A superior relative effectiveness of colloids in comparison with crystalloids for support of plasma volume and thereby for normalisation of the haemodynamics in emergency resuscitation has, however, been acknowledged for years <|[14-16]|>. Therefore, a colloid based fluid regimen has often been suggested a better alternative than infusion of crystalloids in many clinical situations <|[14-16]|>. The optimal fluid regimen has, however, remained a matter of controversy and a crystalloid versus colloid debate has been going on for years and is still not satisfactorily settled.



It is obvious that intravenous infusion of an electrolyte solution results in a rather poor plasma volume supporting effect due to rapid redistribution of the solution throughout the whole extracellular fluid space <|[17]|>. This implies that for restitution of a plasma volume deficit there is a volume requirement of crystalloid far in excess of the actual intravascular volume deficit. At the fluid resuscitation a major part of the infused crystalloid is deposited in extravascular tissues whereby interstitial hydration is markedly increased <|[15, 16]|>. When more extensive volume deficits are substituted with crystalloid this interstitial deposition of fluid may include a risk of oedema formation. Therefore, it is not surprising that the rationale for choosing crystalloid in the clinical treatment of more extensive plasma volume deficits or perioperative blood losses may be questioned.



Some years ago An end of the crystalloid era was suggested by Twigley and Hillman <|[18]|>. Recently, however, rather alarming reports, based on meta-anlyses of published randomised studies, have instead been published <|[19-21]|>. These meta-analyses indicate an increased risk of mortality for critically ill patients resuscitated with colloids. Although albumin has been suggested the possible main cause of negative colloid associated influences on outcome, at least in critically ill patients <|[22]|>, the fact remains that at present time there is a major push for increased use of crystalloids, both in the perioperative period and in the intensive care surroundings.



It should be obvious, however, that the type of fluid required to correct fluid deficits and derangements will to a considerable extent depend on which fluid compartment is depleted and if more than one compartment is affected <|[1]|>. Since the optimal fluid regimen in different clinical conditions has remained a matter of controversy, this survey will summarise the basic physiological concepts of crystalloid and colloid resuscitation, including the present status of the ongoing crystalloid versus colloid debate.



FLUID SPACES AND INTERNAL FLUID FLUXES IN RESPONSE TO TRAUMA



Fluid spaces



The fluid spaces of the body are schematically presented in <|Figure 1|>. Total body water in the adult usually ranges from about 50 to 65% of body weight, somewhat lower in females than in males, and there is a general decrease in body water with increasing age in the elderly. About 2/3 of the body water is found within the intracellular space, i.e. approximately 28 L in a 70 kg adult male. Most of the remaining body fluid (about 14 L) is distributed within the interstitial and intravascular spaces. The relationship between the fluid content of the intravascular and interstitial spaces is in the range of 1/4 to 1/5. The capillary membrane is freely permeable to water. Therefore, the fluid movements between the intravascular and interstitial spaces are primarily influenced by the osmotic gradients of solutes. At fluid homeostasis the osmolality of both these extracellular compartments is approximately equivalent. The total fluid fluxes between the intra- and extravascular compartments are influenced by hydrostatic and colloid osmotic gradients as predicted by the Starling transcapillary fluid equilibrium equation. A near-equilibrium situation usually exists between several forces tending to move fluid out through the capillary membrane (mean capillary hydrostatic pressure, negative interstitial free fluid pressure, interstitial fluid colloid osmotic pressure), and one major force tending to move fluid back into the capillary bed (plasma colloid osmotic pressure). At near-equilibrium there is a slight imbalance in favour of outward filtration which is balanced by return of the fluid via the lymphatics. Alterations in any of these variables will change the overall balance of the fluid exchange process across the capillary membrane.



Fluid fluxes in response to surgery and trauma



The general response of the body to trauma and blood loss is a pronounced neuro-endocrine activation whereby different major compensatory defence mechanism are set inte action in order to secure the perfusion and substrate availability of central vital organs <|[23]|>. In this process major internal changes of the fluid homeostasis between the different fluid spaces of the body are induced <|[10, 11]|>. In response to the neuro-endocrine activation induced by trauma and/or haemorrhage about 1.0 L of fluid can be transferred in the adult individual from the intracellular and interstitial spaces into the intravascular compartment <|(Figure 1)|>. The main components of this endogenous plasma volume supporting defence mechanism are:



A glucose-osmotic transcapillary refill process.<|[10, 11]|>



In response to the hyperglycaemia induced by the trauma, plasma osmolality will increase and about 2 to 3 liters of fluid can be mobilised along the osmotic gradient from the intracellular compartment into the intersititial fluid space. Of this fluid about 0.5 L will reach the intravascular compartment and support blood volume. Trauma-induced insulin resistance will contribute to the maintenance of the hyperglycaemia and thereby increase the efficacy and duration of this glucose-osmotic transcapillary refill process.



A resetting of the pre- to postcapillary resistance ratio.<|[10, 11]|>



The capillary hydrostatic pressure is reduced due to precapillary vascular constriction. Thereby, the equilibrium of the Starling transcapillary exchange process is changed so that fluid reabsorption from extravascular sources is favoured. In the adult individual bout 0.5 L of fluid can be mobilised by this compensatory mechanism from the interstitial fluid space into the intravascular compartment.



In addition to direct losses of blood and plasma in connection with surgical procedures or following trauma, there is an increased overall transcapillary fluid loss resultant from increased trauma-induced activation of the cascade systems evoking a systemic inflammatory response syndrome (SIRS) influencing endothelial cell barrier function and thereby capillary permeability <|[24]|>. This more generalised increase of capillary permeability in response to trauma will further enhance the hypovolaemia and jeopardise tissue perfusion <|[25]|>.



Shires and co-workers <|[5, 6]|> have suggested a reduction of the extracellular fluid volume during major surgery and following trauma due to internal redistribution of fluid into traumatised tissues (wound oedema) and into organs, the function of which is disturbed (e.g. paralytic intestine). Such fluid movements were considered to constitute so-called third space losses. The occurrence and clinical significance of possible internal third space losses has remained unclear <|[1]|>. Trauma is often considered associated with a relative increase rather than a reduction in extracellular fluid content. A relative increase in the interstitial fluid content is explained by stress induced salt and water retention, as well as the above discussed mobilisation of intracellular fluid into the extracellular fluid space. It seems obvious, however, that in order to achieve normovolemia and hemodynamic stability and re-establish fluid homeostasis in surgical patients or trauma victims, it is necessary at the fluid resuscitation not only to consider direct blood losses but also all these internal compensatory fluid fluxes <|[11,15,16]|>.



DISTRIBUTION OF INFUSED RESUSCITATION FLUIDS



The relative distribution of crystalloids and colloids between the different fluid spaces of the body and the advantages and disadvantages of crystalloid versus colloid based fluid resuscitation regimes are summarized in <|Figure 1|>. As indicated in the figure, the plasma volume support achieved at the infusion of a colloid is usually good while the more prolonged plasma volume supporting efficacy of crystalloids is poor <|[15, 16]|>.



Crystalloids



General aspects



Crystalloid resuscitation fluids usually have a balanced electrolyte composition since infusion of a rather large volume is needed for restoration of haemodynamic stability in hypovolaemic patients <|(Figure 1)|>. Usually fluids with a "buffering capacity", i.e. fluids containing either lactate or acetate are preferred <|[16]|>. When the lactate or acetate ions are metabolised by tissue cells, bicarbonate ions will be produced and a buffer effect is achieved. Acetate containing Ringers solutions seem more advantageous than lactate containing ones since the capacity of the body to metabolise lactate may be reduced in case of disturbed organ perfusion, as seen in the connection with shock and trauma <|[16]|>. A lactate containing solution may even aggravate an already existing lactic acidosis since the metabolic capacity of the two main lactate-clearing organs, i.e. the liver and the kidney, is disturbed in severe shock. Acetate, on the other hand, can be metabolised by most tissue cells of the body.



Among the advantages of Ringers type of crystalloids are: absence of adverse anaphylactoid reactions, minimal influences on haemostasis other than those caused by the haemodilution per se, and diuresis promoting effects (Figure 1). The low cost of crystalloids as compared to colloids is often also considered advantageous.



Distribution between the different fluid spaces



Balanced salt solutions will freely cross capillary membranes and equilibrate within the whole extracellular fluid space. The intravascular retention of a crystalloid is consequently poor and it is usually considered that in connection with blood losses a large volume, i.e. 4 to 5 times the actual intravascular volume deficit, has to be infused in order to achieve normovolaemia <|[15-17]|>.



It has recently been claimed by Riddez and co-workers <|[26]|> that for optimal volume substitution following acute hypovolaemia, infusion of a much smaller volume of Ringers solution may be sufficient. In their study young volunteers were subjected to acute moderate venesection, i.e. withdrawal of 900 mL of blood. It was noted that these young individuals were able to restore up to 50% of the blood loss by internal shift of fluid from extravascular sources into the intravascular compartment, as expected on the basis of the above considered transcapillary refill process. It was also claimed that infusion of Ringers solution corresponding to about twice the lost volume was sufficient for restitution of central and regional haemodynamics. Young individuals subjected only to withdrawal of blood and not exposed to any tissue injury may not be representative for the average surgical or trauma patient population seen in the clinics. In surgical patients it has been shown by Lamke and Liljedahl <|[17]|> that infusion of 1 L of crystalloid in the postoperative period will result in a remaining plasma volume expansion of only about 0.2 L after an equilibration period of 90 min. Cervera and Moss <|[27]|> have also pointed out that crystalloid resuscitation of haemorrhage for reexpansion of the lost plasma volume may require much larger volumes than appreciated by most clinicians. At the resuscitation of major intravascular volume derangements with large volume of crystalloid it is consequently of importance to warm the infused fluid in order to prevent hypothermia and temperature associated disturbances of coagulation.



Crystalloid associated disadvantages/risks



Since 75-80% of the infused volume of a crystalloid will lodge in the extravascular compartments <|[15, 16, 28]|>, it is obvious that crystalloid resuscitation includes a risk of increased tissue hydration and oedema formation <|(Figure 1)|>. Experimental studies of the effects of a massive intravenous isotonic fluid load on the extravascular fluid content of different tissues have shown that fluid will accumulate mainly in tissues with a high compliance such as skin and connective tissue <|[29]|>. Peripheral oedema, resulting from excessive crystalloid resuscitation, has been considered mainly of cosmetic and not of functional importance. However, fluid overload with crystalloids will also increase the fluid content of vital organs, e.g. the lungs <|[29]|>. Therefore, it seems likely that the functional respiratory capacity could also be disturbed by infusion of a large volume of crystalloid resuscitation fluid.



Extravascular lung water: In the lungs a reduction of the colloid osmotic pressure (COP) in response to crystalloid resuscitation will influence the threshold hydrostatic capillary pressure at which pulmonary fluid overload will occur <|[15, 30]|>. The hydrostatic gradient (difference between the pulmonary capillary and the tissue fluid pressures) tends to push fluid out of the vasculature. However, it is counteracted by the high COP gradient between plasma and tissue fluid favouring retention of fluid within the vascular compartment. This oncotic gradient, however, is dependent on the ability of the capillary membrane to reflect substances that are colloid-osmotically active, i.e. to prevent leakage of colloidal substances out of the vascular compartment into the interstitial fluid space.



Major causes for increased extravascular lung water are increases in the hydrostatic gradient across the pulmonary capillary during fluid resuscitation and increases in the protein permeability of the capillary membrane due to systemic activation of the cascade systems influencing endothelial cell barrier function. The maintenance of a gradient of 7-9 mm Hg between plasma COP and cardiac filling pressure has been suggested of importance for prevention of pulmonary oedema <|[31]|>. However, there are several important anti-oedema safety factors, which will prevent moderate changes in capillary hydrostatic pressure from influencing the extravascular fluid content of the lung <|[32]|>. One component of this physiological defence mechanism is the increase in interstitial pressure that is initially created by an increased fluid filtration. Thereby the hydrostatic pressure gradient across the pulmonary capillary is shifted back towards the normal level and further fluid movement is prevented. Increased fluid filtration will also influence the COP gradient by diluting the protein content of the extravascular fluid. Such a reduction of the interstitial COP will oppose fluid flux by widening the colloid osmotic gradient across the pulmonary capillary bed.



Although the above considered anti-oedema safety factors are of importance in the early phase of increased pulmonary capillary hydrostatic pressure, the most important oedema preventive factor is the capacity of the lung to increase its lymph flow. From basal levels lung lymph flow can increase severalfold, up to 7-10 times <|[33]|>. Therefore, as long as the lymph drainage of the lung can clear the amount of fluid that is filtered across the pulmonary capillary, no manifest clinical pulmonary oedema will occur. This may explain why crystalloid treatment of hypovolaemia in many situations does not seem to increase extravascular lung water to any significant extent <|[34-38]|>. However, resuscitation with crystalloids may predispose for pulmonary dysfunction in case of later major challeges of the homeostatic balance <|[39]|>.



Peripheral tissue oedema: Acrystalloid based treatment regime may initially restore cardiac output during the fluid resuscitation but, due to rapid leakage of fluid out into the extravascular tissues, it may be insufficient for the maintenance of an adequate intravascular volume support <|[40]|>. Laser doppler flowmetric studies have demonstrated a persistent depression of the microvascular blood flow in spite of infusion of Ringers lactate up to four times the volume of maximal bleeding <|[41]|>. Crystalloid resuscitation consequently includes a risk of oedema formation in peripheral tissues. By compression of capillaries tissue oedema may jeopardise microvascular blood flow and thereby also tissue oxygenation <|[42-45]|>. Generalised oedema may consequently disturb the transport of oxygen and nutrients to tissue cells and thereby contribute to the development of multiple organ failure.



Iatrogenic tissue oedema caused by crystalloid resuscitation is reflected by a significant weight gain and has been considered to result in prolonged need of respirator treatment, impaired wound healing, and prolonged ICU-stay <|[44, 45]|>. Especially in elderly patients with reduced functional capacity of vital organs, including the cardiovascular and respiratory systems, such fluid overload may disturb the recovery process after surgery and trauma. Late adverse effects after an initial resuscitation with large quantities of crystalloid, seen as a "third-day" transient circulatory overload, may in addition occur due to the subsequent redistribution of the tissue oedema <|[46]|>. The problem of extravascular fluid accumulation after crystalloid volume loading seems even more critical for trauma patients suffering head injuries <|[47]|>.



Most of the extravasating crystalloid will distribute within the interstitial space but some will probably leak into hypoxic or traumatised cells in connection with surgery or trauma, i.e. into cells with a reduced functional capacity to regulate their membrane electrolyte balance and hence their volume <|[25]|>. Therefore, it is not surprising that isotonic crystalloid resuscitation has been reported not only to expand peripheral interstitial tissue but also to cause cellular oedema in vital central organs. Oedema formation may occur even in the heart following crystalloid resuscitation <|[48]|>.



Colloids



General aspects



The advantages and disadvantages of colloid-based fluid resuscitation are summarised in <|Figure 1|>. It is wellknown that addition of colloid, even in low concentrations, will markedly reduce the overall fluid volume requirements for achievement of haemodynamic stability in hypovolaemic patients <|[49-51]|>. Therefore, infusion of only a rather moderate volume is usually needed and the risk of fluid overload of extravascular tissues, as seen in connection with crystalloid fluid resuscitation, is thereby reduced. Colloid solutions contain, to varying degrees, larger, oncotically active molecules, which do not easily cross the capillary membrane. The greater capacity of colloids to remain within the intravascular space includes not only a more efficient expansion of the intravascular plasma volume but also a better maintenance of plasma COP.



Plasma volume supporting efficacy



The overall plasma volume-expanding efficacy of a colloid solution is dependent both of the size and concentration of its colloid-osmotically active molecules and their intravascular persistence. The COP of a colloid is the major factor influencing its initial plasma volume expanding capacity. Intravenous infusion of colloids with a COP lower than or equal to that of plasma will even at high infusion rates result in a mainly isovolaemic initial plasma volume expansion. Colloids with a high COP, on the other hand, will mobilise fluid from extravascular sources into the vascular compartment along the created COPgradient. Rapid infusion of colloids with a high COP may consequently include a risk of acute intravascular volume overload <|[51]|>.



The more prolonged plasma volume supporting capacity of a colloid is determined by the numbers, sizes, and configurations of the molecules in the suspension and the breakdown and elimination characteristics for the substance <|[51]|>. In has repeatedly been claimed that infusion of colloid will improve the oxygen transport to tissues and thereby enhance tissue oxygen metabolism more than crystalloid fluid resuscitation <|[45, 51, 52]|>. This may be explained by favourable haemorheological effects of colloids.



Haemorrhagic effects



The rheological behaviour of blood is altered in many clinical situations and haemorheologic aspects are important to consider at the choice of resuscitation fluid. The haemorheologic effectiveness of a colloid is determined by its haemodilutional capacity in combination with its inherent specific pharmacological effects on red cell aggregation, platelet function, plasma viscosity, and blood corpuscle-endothelial cell interactions <|[51]|>.



The reduction of the haematocrit level of the blood at the haemodilution seems to be the most important determinant for the rheologic effects and thereby for enhancement of tissue perfusion and oxygenation. Vascular resistance is reduced by lowered whole blood viscosity, which will enhance venous return and increase cardiac output <|[53-55]|>. In comparison to the dominating role played by the haemodilution per se changes in plasma viscosity seem of minor importance for alterations of tissue perfusion <|[56, 57]|>. Colloid associated reduction of red blood cell aggregation and leukocyte-endothelium interactions may be additional factors of considerable importance for the enhancement of blood flow in the microvasculature <|[51]|>. Thereby the potential activation of a systemic inflammatory response syndrome (SIRS) by surgery or trauma will be moderated and the risk of multiple organ dysfunction syndrome may be reduced <|(Figure 1)|>.



Colloid associated disadvantages/risks



In addition to the risk of volume overload at extensive fluid resuscitation with colloids, several of the available artificial colloids will also influence the haemostatic competence of the body <|[58]|>. The effects on haemostasis are partly due to haemodilution, resulting in decreased concentrations of coagulation factors, and partly to colloid specific effects on the clotting of blood. Dextrans seem to reduce the haemostatic competence more than hydroxyethyl starch preparations and gelatins <|[51]|>. The maximal dose recommendations for the different colloids are important to consider in clinical practice. Factors such as tissue accumulation, potential adverse effects on renal function, and risk of anaphylactoid reactions are also of importance, as is the cost factor <|(Figure 1)|>.



THERAPEUTIC GOALS OF CLINICAL FLUID RESUSCITATION



It is obvious that colloids and crystalloids, due to their specific characteristics, are mainly distributed to different body fluid spaces. Since colloids are distributed mainly to the intravascular space and crystalloids to the interstitial space, the decision to use colloid or crystalloid solution should depend to a major extent on whether the intravascular or the interstitial space is depleted <|[1]|>. The intravascular space is the more accessible one to clinical estimation of its fluid status. Arterial blood pressure, pulse rate, central venous pressure (CVP), peripheral perfusion, urine output, and pulmonary artery wedge pressure (PAWP) are all indicators of the intravascular blood volume. The response of these variables may also be considered of value in conjunction with a fluid challenge. A challenge with colloid, mainly staying in the circulation, is much easier to interpret than a challenge with crystalloid, which to a considerable extent relatively rapidly is redistributed, to the extravascular space <|[1]|>. Appropriate goals for the fluid resuscitation in connection with surgical procedures or in the treatment of critically ill patients can be summarised as follows <|[16, 51]|>:



maintenance or achievement of normovolaemia and haemodynamic stability


restitution of the fluid balance between the different fluid compartments


maintenance of an adequate plasma colloid osmotic pressure (COP)


enhancement of microvascular blood flow


prevention of cascade system activation and increased blood coagulability


normalisation of oxygen delivery to tissue cells and cellular metabolism and in some clinical situations


prevention of reperfusion type of injury.



In order to achieve all these goals the clinical use of both crystalloids and colloids seems indicated <|[51]|>. In spite of such evidence, however, a crystalloid versus colloid controversy has continued and there is presently no generally accepted consensus on the optimal fluid therapy in different clinical situations.



CHOICE OF FLUID REGIMEN AND CLINICAL OUTCOME



In most situations about 80% of infused crystalloids will leave the vasculature before the end of the infusion, i.e. within 20 to 40 min. The correction of the intravascular hypovolaemia with a moderate volume of crystalloid may therefore be insufficient. Even minor remaining degrees of hypovolaemia can predispose to insufficient organ perfusion. The splanchnic circulation may be critically affected and a decrease in splanchnic blood flow may occur even after a brief episode of normotensive hypovolaemia <|[59, 60]|>. Infusion of too large volumes of crystalloid solution, on the other hand, especially in critically ill patients, may due to interstitial sequestration induce overt hypoxia and increasing degree of pulmonary oedema <|[61-65]|>. The distribution of crystalloid solutions mainly to the interstitial space and the subsequent decrease in plasma COP are both factors contributing to a disturbance of the Starling equilibrium across the endothelial barrier. By comparison, colloid solutions, when adequately titrated against intravascular volume, cause little or no hypoxia <|[61-64]|>.



In spite of such arguments in favour of colloid fluid resuscitation regimens, it has been suggested on the basis of systematic reviews (meta-analyses) of randomised controlled studies that colloid administration, at least to critical ill patients, may rather increase than reduce mortality <|[19-22]|>. In <|Table 1|> the main observations of three different meta-analytic assessments of outcome in patients resuscitated with either colloids or crystalloides are summarised. Velanovich <|[19]|> included 8 clinical studies, 6 dealing with trauma patients. The meta-analysis revealed an overall treatment effect with a 5.7% relative difference (reduced mortality) in favour of crystalloid fluid therapy. For trauma patients a difference of 12.3% in mortality in favour of crystalloids was observed. However, for surgical non-trauma patients a 7.8% difference in mortality in favour of colloid treatment was found.



The more recent meta-analysis published in 1998 by Schierhout and Roberts <|[20]|> was based on systematic review of 26 published randomised studies comparing mortality (of all reasons) in critically ill patients receiving fluid therapy with either colloids or crystalloids. Of the reviewed studies 7 were dealing with trauma patients, 12 with surgical patients, 4 with burn patients, and 3 with patients with increased capillary permeability (septic/toxic states). The review indicated that the relative risk of death for trauma patients treated with colloid as compared to crystalloid was about 1.30 <|(Table 1)|>. It was thus noted that the resuscitation with colloid was associated with an absolute increase in the risk of mortality of 4% (4 extra deaths for every 100 patients resuscitated). Therefore, it was commented that as colloids are not associated with improved survival and are considerably more expensive than crystalloids, it is hard to see how their continued use outside randomised controlled trials in subsets of patients of particular concern can be justified <|[20]|>.



It should be noted, however, that in 14 out of the 26 studies the colloids infused were albumin or plasma protein fraction and in 3 of the trauma studies hypertonic (7.5%) saline rather than conventional crystalloids was used as fluid treatment regimen. Considering the reported association by the Cochrane Injuries Group Albumin Reviewers <|[22]|> between human albumin administration in critically ill patients and increased mortality, albumin rather than colloids per se could influence the outcome of critically ill patients.



In a systematic review by Choi and co-workers published in 1999 <|[21]|>, 105 articles dealing with comparisons of crystalloid versus colloid fluid resuscitation could be identified <|(Table 1)|>. However, only 17 studies, including 814 patients, were found relevant for comparative evaluation of pulmonary oedema, mortality, and the length of stay of adult patients. No overall differences between patients resuscitated with colloid or crystalloid were observed but subgroup analyses revealed a statistically significant difference in mortality in trauma in favour of crystalloid resuscitation <|(Table 1)|>. It was concluded, however, that methodological limitations preclude any evidence-based clinical recommendations and that larger well-designed randomised trials are needed to achieve sufficient power to detect potentially small differences in treatment effects, if they truly exist <|[21]|>.



CRYSTALLOIDS OR COLLOIDS IN CLINICAL PRACTICE?



The findings of the different published meta-analytic comparisons <|[19-22]|> suggest that crystalloids may be superior in the treatment of critically ill patients with increased capillary permeability while colloids are superior in the perioperative treatment of non-trauma surgical patients <|(Table 1)|>. However, colloids are in general more effective than crystalloides for optimising physiological variables related to flow in critically ill patients and for maintaining the delivery of oxygen to tissues <|[65]|>. Therefore, it is obvious that an important question related to meta-analyses remains unanswered, i.e. the clinical relevance of data obtained from assessment of historical studies for the present practice of medicine. The original publications included in the different meta-analysis <|[19-22]|> cover a time period of about 25 years. During this long time period many basic clinical therapeutic procedures, in addition to the choice of fluid regimen, have changed considerably and do not really reflect present practice. Therefore, the lack of contemporary trials may certainly be considered a problem and it makes it difficult to draw any practical overall conclusions, which may have major impact on present treatment regimens.



It should also be remembered that in many clinical situations intravenous fluid therapy is used only for short periods, e.g. in the perioperative period and the patient is usually soon able to resume oral intake. Most clinical studies are consequently designed to compare endpoints other than mortality, mostly variables indicative of effects on cardiovascular performance (haemodynamic stability, cardiac output, haemorheology, tissue perfusion, microvascular blood flow, etc.), respiratory function (extravascular lung water, shunt fraction, gas exchange, etc.), or total fluid balance (fluid sequestration between different fluid compartments, renal function, diuresis). The timing of different measurements in relation to fluid administration is another important factor that will influence the obtained results. Mortality, however, is in most clinical studies of fluid resuscitation rather an unexpected complication, dependent on patient as well as trauma or critical illness associated factors, including the adequacy of all other therapies given. Therfore, the presently available meta-analytic data can not be considered a solid evidence based platform for clinical fluid resuscitation recommendations.



As pointed out by Hillman and co-workers <|[1]|>, the crystalloid-colloid debate is flawed by an erroneous basic assumption, that both crystalloids and colloids serve the same purpose, i.e. to restore the intravascular volume. Colloids specifically have that function, while crystalloids mainly restore the volume of the interstitial compartment. The choice then, of which is the better fluid to use for resuscitation, is also misleading. Most of the colloid will remain in the circulating space, while most of the crystalloid will be distributed out of the vascular compartment. If enough crystalloid solution is administered, the circulating volume will be restored, but at the expense of expanding extravascular compartment and tissue oedema formation. Present evidence may indicate, however, that in case of increased capillary permeability, albumin may, due to its molecular characteristics, not be the optimal colloid for plasma volume expansion in critically ill patients and trauma patients <|[22, 67]|>.



CONCLUSIONS



Fluid therapy should be based on the specific needs of each individual patient.


Colloids are in most situations to be preferred when the main indication is to increase intravascular volume.


Albumin may be contraindicated in case of suspected increased capillary permeability due to SIRS.


Crystalloids are needed for correction of extravasacular fluid derangements.


Infusion of large volume of crystalloids for correction of major intravascular volume deficit includes a risk of tissue oedema and organ dysfunction.



References



1. Hillman K, Bishop G, Bristow P. The crystalloid versus colloid controversy: Present status. Baillires Clin Anaestesiol, 1997; 11: 1.


2. Hogan JJ. The intravenous use of colloidal (gelatin) solutions in shock. JAMA, 1915; 64: 721.


3. Richter W, Hedin H. Solutions and emulsions used for intravenous infusions. Handbook Exp Pharmacol, 1983; 63: 581.


4. Wiggers CJ. Physiology of Shock. The Commonwealth Fund, New York, 1950.


5. Shires GT, Brown FT, Canizaro PC, Sommerville N. Distributional changes in extracellular fluid during acute hemorrhagic shock. Surg Forum, 1960; 11: 115.


6. Shires GT, Williams J, Brown F. Acute changes in extracellular fluids associated with major surgical procedures. Ann Surg, 1961; 154: 803.


7. Gutelius JR, Shizgal HM, Lopez G. The effect of trauma on extracellular water volume. Arch Surg, 1968; 97: 206.


8. Reid DJ. Intracellular and extracellular fluid volume during surgery. Br J Surg, 1968;55: 594.


9. Jrhult J. Osmotic fluid transfer from tissue to blood during hemorrhagic hypotension. Acta Physiol Scand, 1973; 89: 213.


10. Drucker WR, Chadwick CDJ, Gann DS. Transcapillary refill in hemorrhage and shock. Arch Surg, 1981: 116: 1344.


11. Haljame H. Interstitial fluid response. Clin Surg Internat, 1984; 9: 44.


12. Shires GT, Canizaro PC. Fluid resuscitation in the severly injured. Surg Clin N Am, 1973; 53: 1341.


13. Shires GT. Pathophysiology and fluid replacement in hypovolaemic shock. Ann Clin Res, 1977; 9: 144.


14. ShoemakerWC, Schluchter M, Hopkins JA, Appel PL, Schwartz S, Chang PC. Comparison of the relative effectiveness of colloids and crystalloids in emergency resuscitation. Am J Surg, 1981; 142: 73.


15. Haljame H. Rational for the use of colloids in the treatment of shock and hypovolemia. Acta Physiol Scand, 1985; 29 (suppl. 82): 48.


16. Haljame H. Use of fluids in trauma. Int J Intensive Care, 1999; 6: 20.


17. Lamke LO, Liljedahl SO. Plasma volume changes after infusion of various plasma expanders. Resuscitation, 1976; 5: 93.


18. Twigley AJ, Hillman KM. The end of the crystalloid era? A new approach to perioperative fluid administration. Anaesthesia, 1985: 40: 860.


19. Velanovich V. Crystalloid versus colloid fluid resuscitation: a meta-analysis of mortality. Surgery, 1989; 105: 65.


20. Schierhout G, Roberts I. Fluid resuscitation with colloid or crystalloid solutions in critically ill patients: a systematic review of randomised trials. BMJ, 1998; 316: 961.


21. Choi PT, Yip G, Quinonez LG, Cook DJ. Crystalloids vs. colloids in fluid resuscitation: a systematic review. Crit Care Med, 1999; 27: 200.


22. Cochrane Injuries Group Albumin Reviewers. Human albumin administration in critically ill patients: systemic review of randomised controlled trials. BMJ, 1998; 317: 235.


23. Haljame H. Metabolic consequences of trauma. In: Aasen AO, Risberg B, eds. Surgical pathophysiology. Chur: Harwood Academic Publishers, 1990; 47-67.


24. Baue AE. Multiple organ failure, multiple organ dysfunction syndrome, and the systemic inflammatory response syndrome - where do we stand? Shock, 1994; 6: 385.


25. Haljame H. Cellular metabolic consequences of altered perfusion. In: Gutierrez G, Vincent JL, eds. Tissue oxygen utilization. Berlin Heidelberg: Springer-Verlag, 1991; 71-86.


26. Riddez L, Hahn RG, Brismar B et al. Central and regional hemodynamics during acute hypovolemia and volume substitution in volunteers. Crit Care Med, 1997; 25: 635.


27. Cervera AL, Moss G. Crystalloid requirements and distribution when resuscitating with RBCs and noncolloid solutions during hemorrhage. Circ Shock, 1978; 5: 357.


28. ShoemakerWC. Comparison of the relative effectiveness of whole blood transfusions and various types of fluid therapy in resuscitation. Crit Care Med, 1976; 4: 71.


29. Larsson M, Ware J. Effects of isotonic fluid load on plasma water and extracellular fluid volumes in the rat. Eur Surg Res, 1983; 15: 262.


30. Guyton AC, Lindsey AW. Effect of left atrial pressure and decreased plasma protein concentration on the development of pulmonary edema. Circ Res, 1959; 7: 649.


31. da Luz Pl, Shubin H, Weil MH et al. Pulmonary edema related to changes in colloid osmotic pressure and pulmonary artery wedge pressure in patients after myocardial infarction. Circulation, 1975; 51: 350.


32. Haljame H. Lung water in the ICU. Refresher course lecture. 10 th World Congress of Anaesthesiologists, Hague, 1992; B 304.


33. Zarins CK, Rice CL, Peters RM, Virgilio RW. Lymph and pulmonary response to isobaric reduction in plasma oncotic pressure. Circ Res, 1978: 43: 925.


34. Virgilio RW, Rice Cl, Smith DE et al. Crystalloid vs. colloid resuscitation: Is one better? Surgery, 1979; 85: 129.


35. Shires III GT, Peitzman AB, Albert SA et al. Response of extravascular lung water to intraoperative fluids. Ann Surg, 1983; 197: 515.


36. Boldt J, Bormann BV, Kling D, Scheld H, Hempelmann G. Influence of acute normovolemic hemodilution on extravascular lung water. Crit Care Med, 1988; 16: 336.


37. Zadrobilek E, Hackl W, Sporn P, Steinbereithner K. Effect of large volume replacement with balanced electrolyte solutions on extravascular lung water in surgical patients with sepsis syndrome. Intens Care Med, 1989; 15: 505.


38. Bickell WH, Barrett SM, Romine-Jenkins M, Hull Jr SS, Kinasewitz GT. Resuscitation of canine hemorrhagic hypotension with large-volume isotonic crystalloid: Impact on lung water, venous admixture, and systemic arterial oxygen tension. Am J Emerg Med, 1984; 12: 36.


39. Demling RH, ManoharM, Will JA. Response of the pulmonary microcirculation to fluid loading after hemorrhagic shock and resuscitation. Surgery, 1980; 87: 552.


40. Wang P, Chaudry IH. Crystalloid resuscitation restores but does not maintain cardiac output following severe hemorrhage. J Surg Res, 1991; 50: 163.


41. Wang P, Hauptman JG, Chaudry IH. Hemorrhage produces depression in microvascular blood flow which persists despite fluid resuscitation. Circ Shock, 1990; 32; 307.


42. Schtt U, Lindbom LO, Sjstrand U. Hemodynamic effects of colloid concentration in experimental hemorrhage: A comparison of Ringers acetate, 3% dextran-60, and 6% dextran 70. Crit Care Med, 1988; 16: 346.


43. Amundson B, Jennische E, Haljame H. Skeletal muscle microcirculatory and cellular metabolic effects of whole blood, Ringers acetate, and dextran 70 infusions in hemorrhagic shock. Circ Shock, 1980; 7: 111 .


44. Heughan C, Ninikoski J, Hunt TK. Effect of excessive infusion of saline solution on tissue oxygen transport. Surg Gynecol Obstet, 1972; 135:257.


45. Hauser CJ, Shoemaker WC, Turpin I, Goldberg SJ. Oxygen transport responses to colloids and crystalloids in critically ill surgical patients. Surg Gynecol Obstet, 1980; 150: 811.


46. Lucas CE, Ledgerwood AM, Shier RM et al. The renal factor in the post-traumatic fluid overload syndrome. Trauma, 1977; 17: 667.


47. Fein IA, Rackow EC, Sprung CLet al. Relation of COP to arterial hypoxemia and cerebral edema during crystalloid volume loading of patients. Ann Intern Med, 1982; 96: 570.


48. Moon PF, Hollyfield MA, Myers TL, Kramer GC. Effects of isotonic crystalloid resuscitation on fluid compartments in hemorrhaged rats. Shock, 1994; 2: 355.


49. Dawidson I, Ottoson J, Eriksson B, Hedman L, Sderberg R. Relation of colloid concentration to volume expansion for resuscitation of rats subjected to intestinal ischemic shock. Crit Care Med, 1982; 10: 597.


50. Dawidson I, Eriksson B. Statistical evaluation of plasma substitutes based on 10 variables. Crit Care Med, 1982; 10: 653.


51. Haljame H, Dahlqvist M, Walentin F. Artificial colloids in clinical practice: pros and cons. Baillires Clin Anaesthesiol, 1997; 11: 49.


52. Shoemaker WC. Hemodynamic and oxygen transport effects of crystalloids and colloids in critically ill patients. Curr Stud Hematol Blood Transf, 1986; 53: 155.


53. Brckner UB, Messmer K. Organ perfusion and tissue oxygenation after moderate isovolemic hemodilution with HES 200/0.62 and dextran-70. Anaesthetist, 1991; 40: 434.


54. Le Veen HH, Ip M, Ahmed N, Mascardo T, Guinto RB, Falk G, DOvidio N. Lowering blood viscosity to overcome vascular resistance. Surg Gynecol Obstet, 1980; 150: 139.


55. Kouraklis G, Sechas M, Skalkeas G. Effects of hemodilution on peripheral circulation. Vasc Surg, 1989; 23: 20.


56. Krieter H, Brckner UB, Kefalianakis F, Messmer K. Does colloid-induced hyperviscosity in haemodilution jeopardize perfusion and oxygenation in vital organs? A c t a Aneasthesiol Scand, 1995; 39: 236.


57. Brckner UB, MessmerK. Blood rheology and systemic oxygen transport. Biorheology, 1990; 27: 903.


58. Strauss RG. Volume replacement and coagulation: A comparative review. J Cardiothorac Anesth, 1988; 2, Suppl. 1: 24.


59. Price HL, Deutsch S, Marshal BNE, Stephen GW, Behar MG, Neufeld GR. Hemodynamic and metabolic effects of hemorrhage in man, with particular reference to the splanchnic circulation. Circ Res, 1966; 18: 469.


60. Mythen MG, Webb AR. Intra-operative gut mucosal hypoperfusion is associated with.increased post-operative complications and cost. Intens Care Med, 1994; 20: 99.


61. Rackow EC, Falk JL, Fein IA, Siegel JS, Packman MI, Haupt MT, Kaufman BS, Putnam D. Fluid resuscitation in circulatory shock: a comparison of the cardiorespiratory effects of albumin, hetastarch, and saline solutions in patients with hypovolemic and septic shock. Crit Care Med, 1983; 11: 839.


62. Brinkmeyer S, Safar P, Motoyama E, Stezoski W. Superiority of colloid over electrolyte solution for fluid resuscitation (severe normovolemic hemodilution). Crit Care Med, 1981; 9: 369.


63. Modig J. Advantages of dextran 70 over Ringer acetate solution in shock treatment and in prevention of adult respiratory distress syndrome. A randomized study in man after traumatic haemorrhagic shock. Resuscitation, 1983; 10: 219.


64. Finch JS, Reid C, Bandy K, Fickle D. Compared effects of selected colloids on extravascular lung water in dogs after oleic acid-induced lung injury and severe hemorrhage. Crit Care Med, 1983; 11: 267.


65. Dawidson I, Eriksson B. Statistical evaluation of plas ma substitutes based on 10 variables. Crit Care Med, 1982; 10: 653.


66. Hankeln KB, Beez M. Haemodynamic and oxugen transport correlates of various volume substitutes in critically ill in-patients with various aetiologies of haemodynamic instability. Int J Intens Care, 1998; 5: 8.


67. Haljame H. Albumin: To use of not to use? Contemporary alternatives? In: Baron J-F & Treib J (eds). Volume replacement. Springer-Verlag, 1998, pp 1-22.

Appendix: 



References



1. Hillman K, Bishop G, Bristow P. The crystalloid versus colloid controversy: Present status. Baillires Clin Anaestesiol, 1997; 11: 1.


2. Hogan JJ. The intravenous use of colloidal (gelatin) solutions in shock. JAMA, 1915; 64: 721.


3. Richter W, Hedin H. Solutions and emulsions used for intravenous infusions. Handbook Exp Pharmacol, 1983; 63: 581.


4. Wiggers CJ. Physiology of Shock. The Commonwealth Fund, New York, 1950.


5. Shires GT, Brown FT, Canizaro PC, Sommerville N. Distributional changes in extracellular fluid during acute hemorrhagic shock. Surg Forum, 1960; 11: 115.


6. Shires GT, Williams J, Brown F. Acute changes in extracellular fluids associated with major surgical procedures. Ann Surg, 1961; 154: 803.


7. Gutelius JR, Shizgal HM, Lopez G. The effect of trauma on extracellular water volume. Arch Surg, 1968; 97: 206.


8. Reid DJ. Intracellular and extracellular fluid volume during surgery. Br J Surg, 1968;55: 594.


9. Jrhult J. Osmotic fluid transfer from tissue to blood during hemorrhagic hypotension. Acta Physiol Scand, 1973; 89: 213.


10. Drucker WR, Chadwick CDJ, Gann DS. Transcapillary refill in hemorrhage and shock. Arch Surg, 1981: 116: 1344.


11. Haljame H. Interstitial fluid response. Clin Surg Internat, 1984; 9: 44.


12. Shires GT, Canizaro PC. Fluid resuscitation in the severly injured. Surg Clin N Am, 1973; 53: 1341.


13. Shires GT. Pathophysiology and fluid replacement in hypovolaemic shock. Ann Clin Res, 1977; 9: 144.


14. ShoemakerWC, Schluchter M, Hopkins JA, Appel PL, Schwartz S, Chang PC. Comparison of the relative effectiveness of colloids and crystalloids in emergency resuscitation. Am J Surg, 1981; 142: 73.


15. Haljame H. Rational for the use of colloids in the treatment of shock and hypovolemia. Acta Physiol Scand, 1985; 29 (suppl. 82): 48.


16. Haljame H. Use of fluids in trauma. Int J Intensive Care, 1999; 6: 20.


17. Lamke LO, Liljedahl SO. Plasma volume changes after infusion of various plasma expanders. Resuscitation, 1976; 5: 93.


18. Twigley AJ, Hillman KM. The end of the crystalloid era? A new approach to perioperative fluid administration. Anaesthesia, 1985: 40: 860.


19. Velanovich V. Crystalloid versus colloid fluid resuscitation: a meta-analysis of mortality. Surgery, 1989; 105: 65.


20. Schierhout G, Roberts I. Fluid resuscitation with colloid or crystalloid solutions in critically ill patients: a systematic review of randomised trials. BMJ, 1998; 316: 961.


21. Choi PT, Yip G, Quinonez LG, Cook DJ. Crystalloids vs. colloids in fluid resuscitation: a systematic review. Crit Care Med, 1999; 27: 200.


22. Cochrane Injuries Group Albumin Reviewers. Human albumin administration in critically ill patients: systemic review of randomised controlled trials. BMJ, 1998; 317: 235.


23. Haljame H. Metabolic consequences of trauma. In: Aasen AO, Risberg B, eds. Surgical pathophysiology. Chur: Harwood Academic Publishers, 1990; 47-67.


24. Baue AE. Multiple organ failure, multiple organ dysfunction syndrome, and the systemic inflammatory response syndrome - where do we stand? Shock, 1994; 6: 385.


25. Haljame H. Cellular metabolic consequences of altered perfusion. In: Gutierrez G, Vincent JL, eds. Tissue oxygen utilization. Berlin Heidelberg: Springer-Verlag, 1991; 71-86.


26. Riddez L, Hahn RG, Brismar B et al. Central and regional hemodynamics during acute hypovolemia and volume substitution in volunteers. Crit Care Med, 1997; 25: 635.


27. Cervera AL, Moss G. Crystalloid requirements and distribution when resuscitating with RBCs and noncolloid solutions during hemorrhage. Circ Shock, 1978; 5: 357.


28. ShoemakerWC. Comparison of the relative effectiveness of whole blood transfusions and various types of fluid therapy in resuscitation. Crit Care Med, 1976; 4: 71.


29. Larsson M, Ware J. Effects of isotonic fluid load on plasma water and extracellular fluid volumes in the rat. Eur Surg Res, 1983; 15: 262.


30. Guyton AC, Lindsey AW. Effect of left atrial pressure and decreased plasma protein concentration on the development of pulmonary edema. Circ Res, 1959; 7: 649.


31. da Luz Pl, Shubin H, Weil MH et al. Pulmonary edema related to changes in colloid osmotic pressure and pulmonary artery wedge pressure in patients after myocardial infarction. Circulation, 1975; 51: 350.


32. Haljame H. Lung water in the ICU. Refresher course lecture. 10 th World Congress of Anaesthesiologists, Hague, 1992; B 304.


33. Zarins CK, Rice CL, Peters RM, Virgilio RW. Lymph and pulmonary response to isobaric reduction in plasma oncotic pressure. Circ Res, 1978: 43: 925.


34. Virgilio RW, Rice Cl, Smith DE et al. Crystalloid vs. colloid resuscitation: Is one better? Surgery, 1979; 85: 129.


35. Shires III GT, Peitzman AB, Albert SA et al. Response of extravascular lung water to intraoperative fluids. Ann Surg, 1983; 197: 515.


36. Boldt J, Bormann BV, Kling D, Scheld H, Hempelmann G. Influence of acute normovolemic hemodilution on extravascular lung water. Crit Care Med, 1988; 16: 336.


37. Zadrobilek E, Hackl W, Sporn P, Steinbereithner K. Effect of large volume replacement with balanced electrolyte solutions on extravascular lung water in surgical patients with sepsis syndrome. Intens Care Med, 1989; 15: 505.


38. Bickell WH, Barrett SM, Romine-Jenkins M, Hull Jr SS, Kinasewitz GT. Resuscitation of canine hemorrhagic hypotension with large-volume isotonic crystalloid: Impact on lung water, venous admixture, and systemic arterial oxygen tension. Am J Emerg Med, 1984; 12: 36.


39. Demling RH, ManoharM, Will JA. Response of the pulmonary microcirculation to fluid loading after hemorrhagic shock and resuscitation. Surgery, 1980; 87: 552.


40. Wang P, Chaudry IH. Crystalloid resuscitation restores but does not maintain cardiac output following severe hemorrhage. J Surg Res, 1991; 50: 163.


41. Wang P, Hauptman JG, Chaudry IH. Hemorrhage produces depression in microvascular blood flow which persists despite fluid resuscitation. Circ Shock, 1990; 32; 307.


42. Schtt U, Lindbom LO, Sjstrand U. Hemodynamic effects of colloid concentration in experimental hemorrhage: A comparison of Ringers acetate, 3% dextran-60, and 6% dextran 70. Crit Care Med, 1988; 16: 346.


43. Amundson B, Jennische E, Haljame H. Skeletal muscle microcirculatory and cellular metabolic effects of whole blood, Ringers acetate, and dextran 70 infusions in hemorrhagic shock. Circ Shock, 1980; 7: 111 .


44. Heughan C, Ninikoski J, Hunt TK. Effect of excessive infusion of saline solution on tissue oxygen transport. Surg Gynecol Obstet, 1972; 135:257.


45. Hauser CJ, Shoemaker WC, Turpin I, Goldberg SJ. Oxygen transport responses to colloids and crystalloids in critically ill surgical patients. Surg Gynecol Obstet, 1980; 150: 811.


46. Lucas CE, Ledgerwood AM, Shier RM et al. The renal factor in the post-traumatic fluid overload syndrome. Trauma, 1977; 17: 667.


47. Fein IA, Rackow EC, Sprung CLet al. Relation of COP to arterial hypoxemia and cerebral edema during crystalloid volume loading of patients. Ann Intern Med, 1982; 96: 570.


48. Moon PF, Hollyfield MA, Myers TL, Kramer GC. Effects of isotonic crystalloid resuscitation on fluid compartments in hemorrhaged rats. Shock, 1994; 2: 355.


49. Dawidson I, Ottoson J, Eriksson B, Hedman L, Sderberg R. Relation of colloid concentration to volume expansion for resuscitation of rats subjected to intestinal ischemic shock. Crit Care Med, 1982; 10: 597.


50. Dawidson I, Eriksson B. Statistical evaluation of plasma substitutes based on 10 variables. Crit Care Med, 1982; 10: 653.


51. Haljame H, Dahlqvist M, Walentin F. Artificial colloids in clinical practice: pros and cons. Baillires Clin Anaesthesiol, 1997; 11: 49.


52. Shoemaker WC. Hemodynamic and oxygen transport effects of crystalloids and colloids in critically ill patients. Curr Stud Hematol Blood Transf, 1986; 53: 155.


53. Brckner UB, Messmer K. Organ perfusion and tissue oxygenation after moderate isovolemic hemodilution with HES 200/0.62 and dextran-70. Anaesthetist, 1991; 40: 434.


54. Le Veen HH, Ip M, Ahmed N, Mascardo T, Guinto RB, Falk G, DOvidio N. Lowering blood viscosity to overcome vascular resistance. Surg Gynecol Obstet, 1980; 150: 139.


55. Kouraklis G, Sechas M, Skalkeas G. Effects of hemodilution on peripheral circulation. Vasc Surg, 1989; 23: 20.


56. Krieter H, Brckner UB, Kefalianakis F, Messmer K. Does colloid-induced hyperviscosity in haemodilution jeopardize perfusion and oxygenation in vital organs? A c t a Aneasthesiol Scand, 1995; 39: 236.


57. Brckner UB, MessmerK. Blood rheology and systemic oxygen transport. Biorheology, 1990; 27: 903.


58. Strauss RG. Volume replacement and coagulation: A comparative review. J Cardiothorac Anesth, 1988; 2, Suppl. 1: 24.


59. Price HL, Deutsch S, Marshal BNE, Stephen GW, Behar MG, Neufeld GR. Hemodynamic and metabolic effects of hemorrhage in man, with particular reference to the splanchnic circulation. Circ Res, 1966; 18: 469.


60. Mythen MG, Webb AR. Intra-operative gut mucosal hypoperfusion is associated with.increased post-operative complications and cost. Intens Care Med, 1994; 20: 99.


61. Rackow EC, Falk JL, Fein IA, Siegel JS, Packman MI, Haupt MT, Kaufman BS, Putnam D. Fluid resuscitation in circulatory shock: a comparison of the cardiorespiratory effects of albumin, hetastarch, and saline solutions in patients with hypovolemic and septic shock. Crit Care Med, 1983; 11: 839.


62. Brinkmeyer S, Safar P, Motoyama E, Stezoski W. Superiority of colloid over electrolyte solution for fluid resuscitation (severe normovolemic hemodilution). Crit Care Med, 1981; 9: 369.


63. Modig J. Advantages of dextran 70 over Ringer acetate solution in shock treatment and in prevention of adult respiratory distress syndrome. A randomized study in man after traumatic haemorrhagic shock. Resuscitation, 1983; 10: 219.


64. Finch JS, Reid C, Bandy K, Fickle D. Compared effects of selected colloids on extravascular lung water in dogs after oleic acid-induced lung injury and severe hemorrhage. Crit Care Med, 1983; 11: 267.


65. Dawidson I, Eriksson B. Statistical evaluation of plas ma substitutes based on 10 variables. Crit Care Med, 1982; 10: 653.


66. Hankeln KB, Beez M. Haemodynamic and oxugen transport correlates of various volume substitutes in critically ill in-patients with various aetiologies of haemodynamic instability. Int J Intens Care, 1998; 5: 8.


67. Haljame H. Albumin: To use of not to use? Contemporary alternatives? In: Baron J-F & Treib J (eds). Volume replacement. Springer-Verlag, 1998, pp 1-22.