BACKGROUND: We set out to evaluate early commencement of post-operative enteral nutrition versus traditional management in patients undergoing gastrointestinal surgery. METHODS: Electronic databases were searched, references lists were scanned and authors contacted for additional information. We looked for randomised controlled trials comparing early commencement of feeding (within 24 h) with no feeding in patients undergoing gastrointestinal surgery. Primary endpoints were wound infections, intra-abdominal abscesses, pneumonia, anastomotic leakage, mortality, length of hospital stay and complications of feeding. Data were combined to estimate the common relative risk of post-operative complications and associated 95% confidence intervals. RESULTS: Thirteen trials, with a total of 1,173 patients, fulfilled our inclusion criteria. Mortality was reduced with early post-operative feeding. Early post-operative feeding increased vomiting. The direction of effect is suggestive of a reduction of risk of post-surgical complications and reduced length of hospital stay. CONCLUSION: There is no obvious advantage in keeping patients 'nil by mouth' following gastrointestinal surgery. Early enteral nutrition is associated with reduced mortality, though the mechanism is not clear. This review supports the notion that early commencement of enteral feeding may be of benefit.