Predictors of duodenal ulcer healing and relapse were examined in a population known to have a high healing incidence. Two double-blind prospective studies were performed in 134 patients over 4 wk and in 66 patients over 1 yr, respectively. Short-term treatment consisted either of cimetidine 1 g/day, pirenzepine 75 mg/day, or placebo. In a multiple stepwise linear regression analysis the following factors proved to increase healing incidence in decreasing order of importance: female sex, moderate alcohol consumption, abstinence from smoking, young age, and cimetidine treatment. The following factors had no influence on duodenal ulcer healing: number and total area of peptic lesions, concomitant disease, relatives with duodenal ulcer, duration of duodenal ulcer disease, and status as a migrant worker. In the long-term study, treatment consisted either of cimetidine 400 mg at bedtime, pirenzepine 30 mg at bedtime, or placebo. Cimetidine prevented ulcer relapse. Smoking favored duodenal ulcer relapse in the placebo group, but not in the cimetidine and pirenzepine group. For all the other factors no statistically significant effect was found. It is concluded that in a population with high spontaneous healing incidence (a) factors other than drug treatment such as sex, alcohol intake, smoking, and age are at least as important predictors of the outcome of short-term treatment as the drug treatment itself; (b) moderate alcohol intake might favor ulcer healing; (c) the unfavorable effect of smoking on ulcer relapse is overcome by low-dose, long-term, antisecretory treatment.