Retrospective studies suggest that there is an increased postoperative morbidity among alcohol misusers. We have prospectively studied the risk of alcohol intake among patients undergoing surgery. We investigated 15 symptom-free subjects who required colorectal surgery and who were drinking at least 60 g of alcohol per day. These patients were matched for sex, nutrition, age, weight, cardiovascular and pulmonary disease, diagnosis, anaesthesia, and surgery to 15 control subjects who were consuming below 25 g of alcohol daily. Those drinking at least 60 g of alcohol per day developed more postoperative complications than controls (67% vs 20%, p less than 0.05) and hospital stay was prolonged (20 vs 12 days, p less than 0.05). Preoperatively, alcohol misusers had reduced left ventricular ejection fraction (median, 54% vs 68%, p less than 0.01). Delayed hypersensitivity responses were smaller in the alcohol group before (53 mm2 vs 78 mm2, p less than 0.05) and after (18 mm2 vs 55 mm2, p less than 0.01) surgery. Alcohol misusers had longer bleeding times during the first postoperative week (p less than 0.01). Surgical stress responses, as assessed by changes in plasma cortisol and catecholamines, were higher among alcoholics (p less than 0.05). Postoperative morbidity is increased in symptom-free alcohol misusers. The mechanism is probably subclinical cardiac insufficiency, immunosuppression, and decreased haemostatic function. Preoperative alcohol consumption may be a more important risk factor than previously thought.