OBJECTIVES: To determine the efficacy of misoprostol for the prevention of nonsteroidal anti-inflammatory drug (NSAID)-induced duodenal and gastric ulcers in arthritis patients receiving NSAID therapy. DESIGN: A randomized, double-blind, multicenter, placebo-controlled trial. SETTING: Six hundred thirty-eight private, Veterans Affairs, health maintenance, and academic practices. PATIENTS: Six hundred thirty-eight patients with chronic inflammatory or noninflammatory arthritis who were taking an NSAID but who did not have a gastric or duodenal ulcer on screening endoscopy received treatment with ibuprofen, piroxicam, naproxen, sulindac, tolmetin, indomethacin, or diclofenac daily for 3 months. Four hundred fifty-five (71%) patients completed the trial. INTERVENTIONS: Patients meeting the entry criteria were randomized to receive either misoprostol, 200 micrograms, or placebo, four times a day for 12 weeks. MAIN OUTCOME MEASURES: The endoscopy was repeated at 4, 8, and 12 weeks. The development of a duodenal or gastric ulcer (defined as a circumscribed mucosal defect > or = 0.5 cm in diameter and with perceptible depth) was regarded as prophylactic failure. RESULTS: By 12 weeks, a duodenal ulcer developed in 2 of 320 (0.6%; 95% CI, 0.2% to 3.9%) patients randomized to receive misoprostol, compared with 15 of 323 (4.6%; CI, 2.8% to 8%) patients receiving placebo (P = 0.002). A gastric ulcer developed in 6 of 320 (1.9%; (CI, 0.8% to 4.4%) patients, compared with in 25 of 323 (7.7%; CI, 5.1% to 11.4%), respectively. CONCLUSION: Misoprostol significantly lowers the frequency of both duodenal and gastric ulcer development in patients who require long-term therapy with NSAIDS.