Review of a 26-year experience with transduodenal sphincteroplasty and sphincterotomy was undertaken (1) to analyze critically the indications for and results of these procedures and (2) to determine which preoperative factors correlate with a good or poor outcome. Of 109 patients, 78 underwent sphincteroplasty, whereas 31 had a transduodenal sphincterotomy. Surgical indications included: group 1, 53 patients with common duct stones; group 2, 28 patients with dyskinesia or stenosis of the sphincter of Oddi (without choledocholithiasis or recurrent pancreatitis); and group 3, 28 patients with recurrent pancreatitis. Three elderly group 1 patients who presented with cholangitis died after surgery (a hospital mortality of 2.7%). Seventy-nine of the 103 patients (77%) in whom follow-up was obtained achieved an excellent or good result. Results were almost identical with sphincteroplasty and sphincterotomy. Abnormal preoperative liver function tests were the only predictors of a good postoperative outcome (p less than 0.05). Group 3 patients (good results in 63%) had significantly poorer (p less than 0.05) outcome than group 1 and 2 patients. Results were worst in group 3 patients who had undergone previous abdominal or pelvic surgery (p less than 0.025) and in group 2 patients who were narcotic users (p less than 0.025). The authors conclude that transduodenal sphincteroplasty or sphincterotomy can be performed relatively safely, but caution that careful selection of patients is important when the indication for surgery is either ampullary stenosis or recurrent pancreatitis.