To compare the effectiveness and tolerability of three misoprostol dosing regimens for the prevention of gastric and duodenal ulcers associated with long-term nonsteroidal anti-inflammatory drug (NSAID) therapy. A multicenter, 12-week, randomized, double-blind, placebo-controlled, parallel, four-limb study. Eligibility criteria included upper gastrointestinal symptoms during NSAID therapy and no endoscopic evidence of gastric or duodenal ulcers. A total of 1623 patients was enrolled; 1197 of these met major accession and regimen-compliance criteria and completed the trial. These 1197 patients composed the evaluable group. Patients were randomly assigned to one of four regimens: placebo four times daily; 200 micrograms of misoprostol twice daily and placebo twice daily; 200 micrograms of misoprostol three times daily and placebo once daily; and 200 micrograms of misoprostol four times daily. Upper gastrointestinal endoscopic examinations for ulcers were done after 4, 8, and 12 weeks of therapy. Tolerability and safety of the regimens were assessed by adverse-event monitoring. In the placebo group, the incidence of gastric ulcers was 15.7% and the incidence of duodenal ulcers was 7.5%. The incidence of gastric ulcers was significantly lower in the groups receiving misoprostol twice daily (8.1%; difference, 7.6% [95% CI, 2.7% to 12.5%]; P = 0.002), three times daily (3.9%; difference, 11.8% [CI, 7.4% to 16.3%]; P < 0.001), and four times daily (4%; difference, 11.7% [CI, 6.7% to 16.8%]; P < 0.001) compared with placebo. The gastric ulcer rate was significantly higher in patients receiving misoprostol twice daily compared with those receiving misoprostol three times daily (difference, 4.2% [95% CI, 0.7% to 7.7%]; P = 0.02). A significant (P = 0.02) misoprostol dose-response effect was noted in the prevention of gastric ulcers. The incidence of duodenal ulcers was significantly lower in the groups receiving misoprostol twice daily (2.6%; difference, 4.9% [CI, 1.5% to 8.2%]; P = 0.004), three times daily (3.3%; difference, 4.2% [CI, 0.6% to 7.7%]; P = 0.019), and four times daily (1.4%; difference, 6.1% [CI, 2.6% to 9.6%]; P = 0.007) compared with placebo. No significant difference was detected between patients receiving misoprostol twice daily and those receiving misoprostol three times daily, and no dose-response effect was noted with duodenal ulcers. The incidence of withdrawals for adverse events was significantly lower in the groups receiving misoprostol twice daily (12%) and three times daily (12%) than in the group receiving it four times daily (20%). The incidence of gastrointestinal adverse events was significantly higher in the group receiving misoprostol four times daily (74%) than in the placebo group (62%). Misoprostol, 200 micrograms twice or three times daily, offers substantial protection against gastric and duodenal ulcers in patients receiving long-term NSAID therapy. These dosages were better tolerated than the currently approved regimen of 200 micrograms four times daily.