This prospective, randomized study of selective vagotomy with antrectomy (SVA) versus proximal gastric vagotomy (PGV) for patients with duodenal ulcer compares the incidence of recurrent duodenal ulcer and postoperative morbidity during a 4- to 12-year follow-up period. In 46 patients with SVA there were no recurrent ulcers, but 26% of these individuals had serious digestive problems that were not amenable to medical treatment. The 40 patients with PGV had eight recurrent ulcers (20%), but five of these were found in the first 15 patients (33%) compared with three in the 25 patients (12%) who had operations after the need for extensive periesophageal denervation was discovered in the mid-1970s. Most recurrent ulcers were amenable to medical treatment, but 5% of the patients who had PGV had postoperative dysphagia that required periodic bougienage. The data are consistent with several interpretations, depending on the bias of the individual. However, based upon the fact that recurrent ulcers could be managed nonoperatively after PGV versus the lack of effective treatments for postgastrectomy complaints after SVA, it is reasonable to consider wider use of PGV. There are reasons to believe that the variable ulcer recurrence rates after PGV can be explained by subtle differences in operative technique, including those based upon use of the Congo red test for completeness of vagotomy. Unlike SVA, PGV remains an operative procedure in evolution that requires further clinical investigation. At this time either operation can be applied if both the surgeon and the patient have a clear understanding of the possible effects.