Nineteen patients entered a double-blind randomized trial of the prophylactic effect of cimetidine in gastric ulcer disease. The diagnosis of gastric ulcer had been established endoscopically prior to the trial, but at the time of entry the patients were symptom-free and their ulcers had healed. The patients received either cimetidine, 400 mg twice daily or inactive tablets. Ten cimetidine-treated patients completed 12 months' treatment without suffering a recurrence of symptoms, whereas 5 of 9 placebo-treated patients suffered a symptomatic relapse after 7 to 18 weeks (P less than 0.025). In 4 of the 5 patients an ulcer was found at gastroscopy. It is concluded that maintenance treatment with cimetidine helps to prevent symptomatic relapses in gastric ulcer patients.