Seventy consecutive duodenal ulcer patients endoscopically followed up and treated after ulcer scarring with maintenance dosages of effective antisecretory drugs for 1 year were studied in order to identify the factors associated with ulcer relapse. Logrank test, discriminant analysis (Rao's method), and chi 2 were used in evaluating the results. Nineteen patients (27.1%) relapsed. Duodenal ulcer was found to relapse more frequently in patients with total serum pepsinogen more than 109 micrograms tyr X ml-1 X 24 h-1 and in patients with maximal acid output more than 60 mmol/h (Logrank test; both p less than 0.025). Discriminant analysis showed that blood group, age at onset of the disease, and sex were together useful in recognizing subjects relapsing within 6 months' maintenance treatment (77.1% cases correctly classified); age at onset of the disease, alcohol intake, and smoking habit were together useful in picking out subjects relapsing throughout the entire observation period (1 year) (61.76% cases correctly classified). Our data suggest that: 1) duodenal ulcer subjects with elevated serum pepsinogen or maximal acid output are significantly more prone to relapse than patients with normal or slightly increased values; 2) genetic and acquired factors (such as blood group A, early onset of the disease, male sex, alcohol intake, and cigarette consumption) may have a role in influencing ulcer relapse.