Nonsteroidal antiinflammatory drugs (NSAID) are effective for the relief of pain and inflammation, yet their use is tempered by the development of side effects, primarily in the gastrointestinal (GI) tract. It is now known that inhibition of the enzyme cyclooxygenase (COX) is the principal mechanism for both the efficacy and the toxicity of NSAID. Recent research has shown that COX exists as at least two isoenzymes, COX-1 and COX-2. Compelling evidence suggests that COX-1 synthesizes prostaglandins that are involved in the regulation of normal cell activity (including G1 cytoprotection), whereas COX-2 appears to produce prostaglandins mainly at sites of inflammation. These findings led to the search for compounds that would inhibit COX-2 without affecting COX-1. Several agents are under investigation in this new therapeutic category, including celecoxib (SC-58635). Celecoxib was developed as an antiinflammatory and analgesic agent, and has been studied in preclinical studies and in clinical trials. This paper focuses on the results of 5 key clinical trials of celecoxib: an efficacy trial in dental pain, a 2 week osteoarthritis (OA) efficacy trial, a 4 week rheumatoid arthritis (RA) efficacy trial, a one week endoscopic study of GI mucosal effects, and a 10 day study of effects on platelet function. The arthritis trials identified celecoxib doses that were effective in treating OA and RA and that were distinguished from placebo on standard arthritis scales. In the upper GI endoscopy study, no ulcers occurred in subjects receiving celecoxib or placebo, whereas 19% of subjects receiving naproxen developed gastric ulcers. In the platelet effects trial, no statistically significant difference from placebo was seen in the effect of celecoxib on platelet aggregation or bleeding time. In contrast, naproxen caused statistically significant reductions in platelet aggregation and a statistically significant increase in bleeding time. These preliminary trials show that celecoxib achieves analgesic and antiinflammatory efficacy in arthritis through specific COX-2 inhibition without showing evidence of two of the toxic effects of COX-1 inhibition associated with NSAID.