To assess the effect of selective decontamination of the digestive tract on respiratory tract infections and survival of patients treated in an intensive care unit, we carried out a meta-analysis of clinical studies comparing patients treated with selective decontamination with untreated controls. From eleven trials (1489 patients), differences between observed and expected respiratory tract infections and mortality were compared, and odds ratios (ORs) calculated. Analysis was done according to study design. ⋯ By contrast, the mortality benefit was less clear. Studies with historical controls and randomised trials showed that mortality was not significantly different between treatment and control patients. The evidence from these studies is at best consistent with a very limited effect of selective decontamination of the digestive tract on survival of patients in the intensive care unit, despite a clear preventive effect on the occurrence of respiratory tract infections.