Liberal versus restricted fluid resuscitation strategies in trauma patients: a systematic review and meta-analysis of randomized controlled trials and observational studies*.

OBJECTIVE Hemorrhage is responsible for most deaths that occur during the first few hours after trauma. Animal models of trauma have shown that restricting fluid administration can reduce the risk of death; however, studies in patients are difficult to conduct due to logistical and ethical problems. To maximize the value of the existing evidence, we performed a meta-analysis to compare liberal versus restricted fluid resuscitation strategies in trauma patients. DATA SOURCES Medline and Embase were systemically searched from inception to February 2013. STUDY SELECTION We selected randomized controlled trials and observational studies that compared different fluid administration strategies in trauma patients. There were no restrictions for language, population, or publication year. DATA EXTRACTION Four randomized controlled trials and seven observational studies were identified from 1,106 references. One of the randomized controlled trials suffered from a high protocol violation rate and was excluded from the final analysis. DATA SYNTHESIS The quantitative synthesis indicated that liberal fluid resuscitation strategies might be associated with higher mortality than restricted fluid strategies, both in randomized controlled trials (risk ratio, 1.25; 95% CI, 1.01-1.55; three trials; I(2), 0) and observational studies (odds ratio, 1.14; 95% CI, 1.01-1.28; seven studies; I(2), 21.4%). When only adjusted odds ratios were pooled for observational studies, odds for mortality with liberal fluid resuscitation strategies increased (odds ratio, 1.19; 95% CI, 1.02-1.38; six studies; I(2), 26.3%). CONCLUSIONS Current evidence indicates that initial liberal fluid resuscitation strategies may be associated with higher mortality in injured patients. However, available studies are subject to a high risk of selection bias and clinical heterogeneity. This result should be interpreted with great caution.