Patients undergoing major gastrointestinal surgery are at increased risk of developing complications. The use of immunonutrition (IN) in such patients is not widespread because the available data are heterogeneous, and some show contradictory results with regard to complications, mortality and length of hospital stay. Randomized controlled trials (RCTs) published between January 1985 and September 2009 that assessed the clinical impact of perioperative enteral IN in major gastrointestinal elective surgery were included in a meta-analysis. Twenty-one RCTs enrolling a total of 2730 patients were included in the meta-analysis. Twelve were considered as high-quality studies. The included studies showed significant heterogeneity with respect to patients, control groups, timing and duration of IN, which limited group analysis. IN significantly reduced overall complications when used before surgery (odds ratio (OR) 0·48, 95 per cent confidence interval (c.i.) 0·34 to 0·69), both before and after operation (OR 0·39, 0·28 to 0·54) or after surgery (OR 0·46, 0·25 to 0·84). For these three timings of IN administration, ORs of postoperative infection were 0·36 (0·24 to 0·56), 0·41 (0·28 to 0·58) and 0·53 (0·40 to 0·71) respectively. Use of IN led to a shorter hospital stay: mean difference -2·12 (95 per cent c.i. -2·97 to -1·26) days. Beneficial effects of IN were confirmed when low-quality trials were excluded. Perioperative IN had no influence on mortality (OR 0·90, 0·46 to 1·76). Perioperative enteral IN decreases morbidity and hospital stay but not mortality after major gastrointestinal surgery; its routine use can be recommended.
