BACKGROUND Retention of the entire stomach during pancreatoduodenectomy should theoretically improve postoperative nutrition, but some studies have reported a high incidence of delayed gastric emptying (GE). An intact pylorus should also prevent enterogastric reflux and its attendant problems, provided that the operation does not impair sphincteric competence. METHODS We studied pyloric function in 24 patients who underwent pylorus-preserving pancreatoduodenectomy (PPP) and 12 who underwent a conventional Whipple resection including distal hemigastrectomy. Resection had been performed for benign (n = 19) or malignant (n = 17) disease. A double isotope technique was used to measure liquid and solid GE, and enterogastric reflux was measured by means of cholescintigraphy. RESULTS Only two patients who underwent PPP required nasogastric intubation beyond 8 days. At follow-up (median, 5 months; range, 2 to 42 months) one patient with persistent vomiting required remedial surgery after PPP (apart from two with recurrent carcinoma affecting the anastomosis); after Whipple resection there were three patients with intermittent vomiting and one with dumping. Median half-time for liquid GE was not significantly different (65 minutes for PPP vs 103 minutes for Whipple resection; n < 70 minutes); likewise, the median half-time for solid GE was not significantly different (140 minutes for PPP vs 180 minutes for the Whipple procedure; n < 110 minutes). Persistent enterogastric reflux was seen in three of 20 patients who underwent PPP and three of eight who underwent Whipple resection. CONCLUSIONS Isotopic test results of GE are frequently abnormal after either type of pancreatoduodenectomy, but symptoms are uncommon. The retained pylorus is competent to prevent enterogastric reflux and does not itself impair emptying of the stomach.