COX-2 selective inhibitors were developed in order to provide similar efficacy to traditional nonselective nonsteroidal anti-inflammatory drugs (NSAIDs) but with improved upper gastrointestinal safety. This paper presents an overview of randomized clinical trials demonstrating the efficacy of COX-2 selective inhibitors for the treatment of patients with arthritis, particularly osteoarthritis and rheumatoid arthritis. In osteoarthritis and rheumatoid arthritis, COX-2 selective inhibitors have been shown to be more effective than placebo and similarly effective as standard doses of nonselective NSAIDs. There are currently few randomized clinical trials comparing the efficacy of the 2 first-generation COX-2 selective inhibitors, celecoxib and rofecoxib, in osteoarthritis. Of 4 head-to-head studies comparing the 2 agents, 3 indicated similar efficacy, while the other demonstrated superiority of rofecoxib at a dose of 25 mg qd compared with celecoxib at a dose of 200 mg qd. There are no clinical trials comparing the efficacy of different agents for treatment of patients with rheumatoid arthritis. Some studies have also demonstrated efficacy for COX-2 selective inhibitors in patients with ankylosing spondylitis and gout. In aggregate, these data show that COX-2 selective inhibitors provide effective relief of pain in patients with osteoarthritis and rheumatoid arthritis, with efficacy that is similar to traditional NSAIDs. Costeffectiveness and cost-utility studies suggest, however, that their use should be limited to patients at high risk of serious upper gastrointestinal side effects, including complicated ulcers.