Cyclooxygenase (COX)-2 selective inhibitors have been shown to have comparable efficacy to nonselective nonsteroidal anti-inflammatory drugs (NSAIDs) in the treatment of patients with osteoarthritis (OA) and rheumatoid arthritis (RA). Large outcome studies have shown that patients with OA and RA not taking low-dose aspirin have fewer symptomatic and complicated upper GI events when treated with COX-2 selective inhibitors than with nonselective NSAIDs. When used in recommended dosages, there is no convincing evidence that patients treated with COX-2 selective inhibitors have an increased incidence of cardiovascular thrombotic events, including non-fatal myocardial infarction, than patients treated with either placebo or nonselective NSAIDs other than naproxen. Co-therapy with low-dose aspirin is recommended in patients with OA and RA at increased risk for cardiovascular events; the need for gastroprotective therapy in such patients is controversial.