A 6-week, multicenter, double-masked, placebo-controlled, parallel-group study compared the upper gastrointestinal (UGI) safety of Arthrotec 75 (diclofenac sodium 75 mg-misoprostol 200 microg; G.D. Searle & Co., Skokie, Illinois) administered twice daily with that of nabumetone 1500 mg administered once daily in 1203 patients with symptomatic osteoarthritis (OA) of the hip or knee. All patients had a documented clinical history of endoscopically confirmed gastric, pyloric-channel, or duodenal ulcer or > or = 10 erosions in the stomach or duodenum. UGI endoscopy was performed at baseline and again at week 6 or early withdrawal. Treatment with Arthrotec 75 resulted in a significantly lower combined incidence of endoscopically confirmed gastric and duodenal ulcers compared with nabumetone (4% vs 11%), and its rate of endoscopically confirmed ulceration was equivalent to that of placebo. The incidence of gastric ulcers alone was also significantly lower with Arthrotec 75 than with nabumetone (1% vs 9%). The incidence of duodenal ulcer with Arthrotec 75 was not significantly different from that with nabumetone (4% vs 3%). Types of adverse events were similar for all treatment groups, with GI adverse events predominating. Arthrotec 75 was well tolerated by the majority of patients. The results of this study demonstrate that Arthrotec 75 has a superior UGI safety profile, causing significantly fewer UGI ulcers, in comparison with nabumetone in patients with symptomatic OA and a documented history of ulcers or > or = 10 erosions.