Controversy persists over the best choice of fluid to use for resuscitation. A number of published articles promote the use of either colloid or crystalloid fluids. Most of the arguments for use of one fluid or the other are based on cardiopulmonary data collected during and after fluid resuscitation. Although many studies report the mortality rate of patients treated with both fluids, none have critically analyzed this most important aspect of therapy. Meta-analysis is a relatively new statistical method whereby data from a number of clinical trials can be pooled to produce more reliable data. In this study meta-analysis was used to pool mortality data from reports of eight previously published, randomized, clinical trials, in which the efficacy of crystalloid and colloid fluid resuscitation was compared. The overall treatment effect when the data from all the clinical trials were pooled showed a 5.7% relative difference in mortality rate in favor of crystalloid therapy. When the data from only those studies using trauma patients were pooled, the overall treatment effect showed a 12.3% difference in mortality rate in favor of crystalloid therapy. However, when data from studies that used nontrauma patients were pooled, there was a 7.8% difference in mortality rate in favor of colloid treatment. In patients with trauma who are septic and in whom the capillary leak syndrome leads to adult respiratory distress syndrome, it may be assumed that colloid resuscitation would be no better than crystalloid resuscitation. In this study the meta-analysis of published data showed that this form of treatment is deleterious. In patients who are nonseptic or having elective surgery, however, the basement membrane is intact, and meta-analysis of data in this setting showed that treatment with colloids would be efficacious.