Meta-analysis of colloids versus crystalloids in critically ill, trauma and surgical patients.

BACKGROUND There is uncertainty regarding the safety of different volume replacement solutions. The aim of this study was systematically to review evidence of crystalloid versus colloid solutions, and to determine whether these results are influenced by trial design or clinical setting. METHODS PubMed, Embase and the Cochrane Central Register of Controlled Trials were used to identify randomized clinical trials (RCTs) that compared crystalloids with colloids as volume replacement solutions in patients with traumatic injuries, those undergoing surgery and in critically ill patients. Adjusted odds ratios (ORs) for mortality and major morbidity including renal injury were pooled using fixed-effect and random-effects models. RESULTS Some 59 RCTs involving 16 889 patients were included in the analysis. Forty-one studies (69 per cent) were found to have selection, detection or performance bias. Colloid administration did not lead to increased mortality (32 trials, 16 647 patients; OR 0·99, 95 per cent c.i. 0·92 to 1·06), but did increase the risk of developing acute kidney injury requiring renal replacement therapy (9 trials, 11 648 patients; OR 1·35, 1·17 to 1·57). Sensitivity analyses that excluded small and low-quality studies did not substantially alter these results. Subgroup analyses by type of colloid showed that increased mortality and renal replacement therapy were associated with use of pentastarch, and increased risk of renal injury and renal replacement therapy with use of tetrastarch. Subgroup analysis indicated that the risks of mortality and renal injury attributable to colloids were observed only in critically ill patients with sepsis. CONCLUSION Current general restrictions on the use of colloid solutions are not supported by evidence.