In order to evaluate the prognostic role of duodenal bulb deformation in the recurrence of peptic ulcer, duodenal bulb morphology and the complete healing of duodenal ulcer were endoscopically evaluated in sixty patients, who were subsequently allocated at random to either maintenance therapy with ranitidine or no treatment. Endoscopic checkups were done at regular intervals, up to the first ulcer recurrence. As expected, long-term ranitidine treatment significantly reduced the relapse rate (12 month cumulative relapse rate was 32% versus 86% in the untreated). A set of prognostic factors which might interfere with this result (sex, age, alcohol consumption, history of ulcerous relatives, duration of the disease, previous H2-blocking treatment, previous complications, smoking and morphology of the duodenal bulb) were evaluated by multivariate analysis using the Cox regression model. Only duodenal bulb morphology appeared to have any independent prognostic value. In the untreated group ulcer recurrence seemed to occur earlier (median relapse time = 2 months) in the patients with severe non-stenosing bulb deformity, and later in those with normal or mildly deformed bulb (median relapse time = 8 months); ranitidine treatment delayed relapse in deformed bulb patients (median relapse time = 14 months) and almost eliminated it in those with normal duodenal bulb morphology. No association was found between the presence of duodenal bulb deformity and the above-mentioned covariates. Our study confirms the primary importance of anti-H2 treatment and suggests that anatomical characteristics of the duodenal bulb also influence the occurrence of ulcer relapse.