Background:Fluid resuscitation is the cornerstone of sepsis treatment. However, whether balanced or unbalanced crystalloids or natural or synthetic colloids confer a survival advantage is unclear. Purpose:To examine the effect of different resuscitative fluids on mortality in patients with sepsis. Data Sources:MEDLINE, EMBASE, ACP Journal Club, CINAHL, HealthSTAR, the Allied and Complementary Medicine Database, and the Cochrane Central Register of Controlled Trials through March 2014. Study Selection: Randomized trials that evaluated different resuscitative fluids in adult patients with sepsis or septic shock and death. No language restrictions were applied. Data Extraction:Two reviewers extracted data on study characteristics, methods, and outcomes. Risk of bias for individual studies and quality of evidence were assessed. Data Synthesis:14 studies (18 916 patients) were included with 15 direct comparisons. Network meta-analysis at the 4-node level showed higher mortality with starches than with crystalloids (high confidence) and lower mortality with albumin than with crystalloids (moderate confidence) or starches (moderate confidence). Network meta-analysis at the 6-node level showed lower mortality with albumin than with saline (moderate confidence) and low-molecular-weight starch (low confidence) and with balanced crystalloids than with saline (low confidence) and low- and high-molecular-weight starches (moderate confidence). Limitations:These trials were heterogeneous in case mix, fluids evaluated, duration of fluid exposure, and risk of bias. Imprecise estimates for several comparisons in this network meta-analysis contribute to low confidence in most estimates of effect. Conclusion:Among patients with sepsis, resuscitation with balanced crystalloids or albumin compared with other fluids seems to be associated with reduced mortality. Primary Funding Source:The Hamilton Chapter of the Canadian Intensive Care Foundation and the Critical Care Medicine Residency Program and Critical Care Division Alternate Funding Plan at McMaster University.
Ann Intern Med. Published online 22 July 2014 doi:10.7326/M14-0178