Fifty-two patients had jejunoileal bypass surgery. End-to-end (Scott) or end-to-side (Payne) shunts were randomly selected for each patient; 31 standard length shunts and 21 shortened bypasses were performed. Only 22 patients had an acceptable result, whereas 30 patients had inadequate weight loss (less than 2.3 kg [5 lb] per month per year) or had gastrointestinal tract, metabolic, or surgical complications judged severe enough to render the outcome less than adequate. There was one death, and four patients required reanastomosis of the bypass. The primary determinant of success was age, ie, younger patients had clearly better results than older patients. In general, shorter shunts produced more weight loss than standard bypass procedures, but were associated with an increased complication rate. Three new complications of jejunoileal bypass are reported: acute colonic dilation with necrosis, beriberi, and lupus erythematosus.