The effect of cimetidine and pirenzepine on the maintenance of healing in duodenal ulceration has been compared in a multi-centre, controlled study. One hundred and sixty-six patients with endoscopically proven duodenal ulceration have been randomised to receive either cimetidine 200 mg t.i.d. and 400 mg nocte, or pirenzepine 50 mg b.i.d. for 6 weeks. Patients in each group were well matched for age, sex, weight and cigarette, alcohol and antacid consumption. After 6 weeks significantly more cimetidine-treated patients had healed ulcers: cimetidine showed 8 unhealed out of 79, and pirenzepine 19 out of 74 (p = 0.01). The follow-up phase lasted for 52 weeks or until patient withdrawal or ulcer relapse. During the follow-up phase 77.6% of cimetidine-treated and 68.8% of pirenzepine-treated patients relapsed. This difference in relapse rates is not statistically significant (p = 0.23).