In a comparative trial of preventive medication for duodenal ulceration with 51 participating centres, 484 patients were recruited for a year's maintenance treatment with the recommended bedtime dose of ranitidine (150 mg; n = 243) or cimetidine (400 mg; n = 241). These outpatients had recently healed duodenal ulcers, confirmed by endoscopy before and after healing, and ulcer relapse was monitored by endoscopy every 4 months. The distribution of factors likely to influence ulcer recurrence was similar in the two treatment groups. A life-table method of analysis showed that the ulcer relapse rate was consistently and significantly lower on ranitidine that on cimetidine (8% v 21%, p = 0.0018 at 4 months; 14% v 34%, p less than 0.0001 at 8 months; and 23% v 37%, p = 0.004 at 12 months). Crude relapse rate calculations, which underestimate the probability of ulcer recurrence, also confirmed the significant superiority of ranitidine 150 mg over cimetidine 400 mg nightly in preventing ulcer recurrence throughout the year.