Both meal-stimulated and nocturnal acid secretions have been shown to be abnormally increased in patients with duodenal ulcer. The relative efficacy of an acid-reducing regimen aimed specifically at controlling postprandial acid secretion compared with one that controls nocturnal acid secretion is, however, not known. The endoscopic healing rates at weeks 2, 4, 6, 8, 10, and 12 of three cimetidine regimens with identical total daily dose--bedtime (1200 mg), mealtime (400 mg three times a day with meals), and reference (200 mg three times a day with meals and 600 mg at bedtime)--were compared in a randomized study on 141 patients with endoscopically proven duodenal ulcer. Evaluating endoscopists were blinded to patients' form and duration of treatment and their clinical progress; patients were unaware of the comparative design of the study. Life-table analysis for the 12 weeks of observation revealed that the mealtime regimen resulted in significantly (P less than 0.05) better healing rates than either the bedtime or the reference regimen. The differences were accounted for largely by the significantly (P less than 0.04) better healing rate at two weeks with the mealtime regimen (68%) than with either the bedtime (47%) or the reference (45%) regimen. These findings indicate that a regimen that aims at controlling meal-stimulated acid secretion achieves a faster healing rate than one that aims at controlling nocturnal acid secretion in the treatment of duodenal ulcer, and they suggest that postprandial acid secretion plays a greater role than nocturnal acid secretion in the pathophysiology of this condition.