Critically ill patients with trauma, burns, surgery or sepsis are often given intravenous fluids to expand intravascular volume. When it comes to selecting the resuscitation fluid, one of the choices is between using a colloid or a crystalloid solution. This choice has considerable cost implications, because volume replacement with colloids is much more expensive than with crystalloids. Clinical studies have shown that colloids and crystalloids have different effects on a range of physiological parameters. Because of these differences, all-cause mortality is arguably the most clinically relevant outcome measure for randomised trials comparing the two fluid types.
This review, which was updated for the sixth time in February 2013, assessed the effects on mortality of using colloids compared to crystalloids, during fluid resuscitation in critically ill patients. A systematic search was conducted for any controlled trial in which participants were assigned to treatment on the basis of random allocation. The interventions considered were colloids (dextran 70, hydroxyethyl starches, modified gelatins, albumin or plasma protein fraction) compared to crystalloid (isotonic or hypertonic) for fluid replacement. Studies needed to include critically ill patients (excluding neonates and pregnant women) who required volume replacement. The authors excluded trials of fluids used for other purposes, such as trials of pre-loading in preparation for elective surgery, and trials in patients undergoing fluid loading before cardiopulmonary bypass. Trials of total parenteral nutrition with or without albumin were also excluded, as were randomised cross-over trials.
A total of 78 eligible trials were identified, with mortality data available from 70 of these. The analyses were divided up on the basis of the types of intervention. Considering the basic comparison of colloids versus crystalloids, the largest number of randomized patients was in the studies of albumin or plasma protein fraction, with 24 trials that reported data on mortality (9920 patients). The pooled risk ratio (RR) from these trials was 1.01 (95% confidence interval (CI) 0.93 to 1.10). 25 trials compared hydroxyethyl starch with crystalloids (9147 patients), finding a pooled RR of 1.10 (95% CI 1.02 to 1.19). In the 11 trials of modified gelatin versus crystalloid, with 506 patients, the pooled RR was 0.91 (95% CI 0.49 to 1.72) and, for the 9 trials comparing dextran with a crystalloid (834 patients), the meta-analysis produced a RR of 1.24 (95% CI 0.94 to 1.65).