Rational or rationalized choices in fluid resuscitation?

The war between colloids and crystalloids wages on. In a large multinational survey of fluid prescribing practices in critically ill patients, we have a new and intriguing snapshot of global fluid resuscitation practices. Colloids are more often used for impaired perfusion or low cardiac output, and the choice of colloid or crystalloid varies enormously between countries. Why are some ICUs prescribing colloids more often than crystalloids when there is little convincing evidence that colloids are superior for fluid resuscitation? Are colloids advantageous in certain diseases, or in specific regional patient populations that have not yet been elucidated? Perhaps we should look inwards: the answer may not be more randomized clinical trials, but better adherence to current guidelines and treatment recommendations.


Introduction
Th e war between colloids and crystalloids wages on. Th e most recent study of fl uid prescribing practices in critically ill patients, in the previous issue of Critical Care, examined data from 5,274 patients in 391 ICUs across 25 countries [1]. In so doing, the authors have provided a snapshot of global fl uid resuscitation prescribing practices. Th e observation that colloids were more frequently prescribed than crystalloids, both on the individual patient level and according to fl uid resuscitation episodes, is novel and surprising [2]. More specifi cally, Finfer and colleagues [1] found colloids were more often used for impaired perfusion or low cardiac output, and geographically more often used in China, Great Britain and Sweden; crystalloids were more frequently used in the United States, New Zealand and Italy [1]. Th e study raises a diffi cult question: why are some ICUs prescribing colloids more often than crystalloids when there is little convincing evidence that colloids are superior for fl uid resuscitation? Asked diff erently, do colloids provide an advantage over crystalloids in certain regions or in specifi c patient populations that have not yet been elucidated?
Is there a rationale for using colloids?
Following a spate of systematic reviews with confl icting conclusions about both the safety and effi cacy of colloids, the American Th oracic Society released in 2004 a consensus statement on colloid use in the critically ill [3]. Th at guideline noted that colloids restore intravascular volume and tissue perfusion more rapidly than crystalloids in shock states. Th is is at least in part because colloids may be prescribed in volumes equivalent to whole blood loss, while crystalloids require 2 to 2.5 times greater volume infusion [4]. Does this translate into a clinically signifi cant reduction in resuscitation time?
Perhaps there are specifi c patient subgroups where colloid use may confer an important benefi t? Colloid use in cardiac surgical patients is reported to decrease pulmonary edema, pain and need for anti-emetics, with consequent faster return of bowel function due to decreased gut edema and preserved gut perfusion [5,6]. Also in cardiac surgical patients, albumin use has been associated with greater peri-operative survival [7]. Th ere is also good clinical evidence for use of colloids in dialysis-related hypotension, spontaneous bacterial peritonitis and large volume paracentesis [3]. Colloids may also be considered in combination with diuretics in patients with acute lung injury or acute respiratory distress syndrome [8,9]. Finally, albumin administration to children with malaria and to adults with sepsis may improve survival [10,11]. If these latter two populations are confi rmed to benefi t from albumin resuscitation, a strong evidence-based recommendation could be made.

Is there a rationale for using crystalloids?
Compared to crystalloids, there are substantial drug acquisition costs for colloids [12]. Given the ever-rising

Abstract
The war between colloids and crystalloids wages on. In a large multinational survey of fl uid prescribing practices in critically ill patients, we have a new and intriguing snapshot of global fl uid resuscitation practices. Colloids are more often used for impaired perfusion or low cardiac output, and the choice of colloid or crystalloid varies enormously between countries. Why are some ICUs prescribing colloids more often than crystalloids when there is little convincing evidence that colloids are superior for fl uid resuscitation? Are colloids advantageous in certain diseases, or in specifi c regional patient populations that have not yet been elucidated? Perhaps we should look inwards: the answer may not be more randomized clinical trials, but better adherence to current guidelines and treatment recommendations.
cost of healthcare internationally, it is noteworthy that among the countries with the highest colloid utilization are those with government managed or socialized healthcare systems. Clinical indications aside, eff orts to control healthcare expenditures based upon drug acquisition costs alone may lead to greater expenses and worse clinical outcomes [13].
Colloids have a combination of desirable and undesirable eff ects. Among their most common adverse reactions are general allergic responses, which are not easily predicted and may result in anaphylaxis [3]. In addition, their general anti-thrombotic properties may adversely aff ect blood coagulation [3]. Hydroxyethyl starch solutions increase the risk of acute kidney injury in sepsis and albumin may cause harm in traumatic brain injury [14][15][16].
Of course, there are detrimental eff ects of crystalloids as well. Depending on the specifi c fl uid, they may cause hyperchloremic metabolic acidosis, hypocoagulable states, reduced renal blood fl ow and urine output, and neurologic and gastrointestinal disturbances [17]. However, the clinical implications of these potentially adverse eff ects are not fully understood.

Why are we irrational?
Why do physicians behave irrationally in the face of clinical evidence that may guide appropriate decisionmaking? In this example, why do certain regions of the world utilize colloids preferentially despite the lack of evidence to prove their superiority [1,2]? Although unlikely, it is possible that genetic diff erences in patients or regional diff erences in disease result in true previously unrecognized superiority. Variations in local prescribing practices are far more likely. For example, in antibiotic prescription for bacteremia, the country of origin and infectious diseases specialist input were explanatory factors [18]. In patients with septic shock, low dose corticosteroids are used in more than half of patients in Europe but less than one-quarter in Asia [19]. As has been shown within and across healthcare systems for myriad clinical decisions, clinical practice for fl uid choices varies despite available evidence.
According to the framework developed by Cabana and colleagues [20], barriers to optimum medical care exist in three major categories: knowledge, attitude and behavior. With globally available medical information, knowledge of evidence regarding fl uid resuscitation is not the primary barrier to evidence-based practice. Physicians' practices are governed by attitudes that guideline recommendations may not produce the desired outcome, and cultural inertia restricting change in practice patterns. Attitude dictates prescribing behavior.
For all things clinical, the truth about fl uid resuscitation is inevitably in the middle. Perhaps some colloid use mixed with crystalloid use in certain patient populations is most benefi cial [21]. Previous clinical trials have led to that supposition, but we must wonder: if the ideal randomized, controlled trial defi nitively reported that as truth, would clinical practice change? Perhaps the answer is not more randomized clinical trials but better adherence to current guidelines and treatment recom mendations.

Competing interests
Emory University has received provision of albumin from Baxter Healthcare for the conduct of a randomized, clinical trial. The authors declare no other competing interests.