To ascertain the role of total enteral nutrition, compared with total parenteral nutrition, as adjunct therapy to steroids in patients with severe acute ulcerative colitis, a prospective randomized trial was conducted in 42 of such patients. Inclusion criteria were the persistence of a moderate or severe attack of the disease (Truelove's index) after 48 h on full steroid treatment (prednisone 1 mg/kg/day). Patients were randomized to receive polymeric total enteral nutrition or isocaloric, isonitrogenous total parenteral nutrition as the sole nutritional support. Remission rate and need for colectomy were similar in both groups. No significant changes in anthropometric parameters were observed in either nutritional group at the end of the study. Median increase in serum albumin was 16.7% (-0.5% to +30.4%) in the enteral feeding group, and only 4.6% (-12.0% to +13.7%) in the parenteral nutrition patients (p = 0.019). Adverse effects related to artificial nutritional support were less frequent (9% vs. 35%, p = 0.046) and milder in enterally fed patients. Postoperative infections occurred more often with parenteral nutrition (p = 0.028). These results suggest that total enteral nutrition is safe and nutritionally effective in severe attacks of ulcerative colitis. It is also cheaper and associated with fewer complications than parenteral nutrition. Total enteral nutrition should be regarded as the most suitable type of nutritional support in these patients.