Ever since methods of artificial nutritional support became available, attempts have been made using this form of treatment to reduce mortality and morbidity in surgical patients. Many trials have addressed this question, but very few have given a meaningful answer because of conceptual and methodological flaws. We therefore undertook a prospective randomised trial investigating the effects of at least 10 days pre-operative total parenteral nutrition (TPN) (n = 51) or total enteral nutrition (TEN) (n = 50) providing 150% basal energy expenditure (BEE) non-protein energy, to reduce major postoperative complications and mortality in a homogeneous patient group with signs of depletion. 50 patients served as a depleted control group (D) and 49 patients served as a non-depleted reference group (ND) and were operated upon without delay. Depleted control patients suffered significantly more septic complications than did patients in the non-depleted reference group (p < 0.05). There was no significant difference, however, in septic complications between either of the nutritional support groups and the non-depleted control group. In high risk patients, with weight loss >10% of body weight and over 500 ml blood loss during operation, a significant decrease in major complications was observed (p < 0.05) as a result of nutritional support. We conclude that pre-operative nutritional support, in patients with severe depletion, results in a reduction in major complications to a degree that justifies its routine use in this selected group of patients.