To investigate the efficacy of standard and low dosage of omeprazole in the healing of duodenal ulcer, 270 patients with endoscopically active duodenal ulcer were randomized to receive omeprazole, 10 mg or 20 mg every morning, or ranitidine, 150 mg twice daily, using the double-dummy technique. Forty-six potential prognostic factors for healing including clinical, acid-secretory, and endoscopic characteristics were prospectively obtained and healing was assessed by endoscopy at weekly intervals for up to 4 weeks. The cumulative healing rates in the 4 weeks were 43%, 77%, 94% and 95% for omeprazole, 10 mg (n = 83); 49%, 86%, 93%, and 96% for omeprazole, 20 mg (n = 87); and 29%, 63%, 83% and 93% for ranitidine (n = 84), respectively. Life-table analysis showed P less than 0.03 for omeprazole, 10 mg versus ranitidine and P less than 0.002 for omeprazole, 20 mg versus ranitidine. Life-table analysis also showed that in the omeprazole groups, healing rates were lower in smokers than in non-smokers (P less than 0.001), in late- than in early-onset patients (symptoms starting after or before the age of 30 years, respectively, P less than 0.02), in those with less than 5 months than in those with more than 5 months of remission (P less than 0.05), and in those with increased maximal acid output than in those with normal output (P less than 0.05). Patients with healed ulcer were interviewed at 2-month intervals and endoscoped at 4-month intervals or whenever symptoms recurred. The cumulative ulcer relapse rates in 1 year were not significantly different between omeprazole and ranitidine groups.(ABSTRACT TRUNCATED AT 250 WORDS)