In a multicentre trial, 120 patients with endoscopically diagnosed duodenal ulcer were randomly allocated to treatment with either 35 micrograms enprostil b.d. or 400 mg cimetidine b.d. for up to 6 weeks on a double-blind basis. After 6 weeks, 82% (42/51) of enprostil-treated patients and 92% (44/48) of cimetidine-treated patients were healed. Corresponding healing figures on an intention-to-treat basis were 70% and 76%. No significant differences were detected between treatments with respect to healing rates or symptom control at any time. Side-effects were reported by 14 patients taking enprostil and 17 patients taking cimetidine; none were serious but they resulted in withdrawal of one and two patients respectively. Enprostil was found to be similar in efficacy and tolerance to cimetidine.