The purpose of this paper is to review the rationale for a new class of nonsteroidal anti-inflammatory drugs (NSAIDs) known as selective cyclooxygenase (COX)-2 inhibitors and to present preliminary clinical data on 2 COX-2 inhibitors that are approved for use in the United States. The primary mechanism of NSAIDs in the treatment of inflammation is the inhibition of COX, which exists in 2 forms. COX-I appears to regulate many normal physiologic functions, and COX-2 mediates the inflammatory response. Theoretically, an NSAID that inhibits COX-2 selectively should decrease inflammation but not influence normal physiologic functions and thus should cause fewer gastrointestinal side effects. Preliminary data suggest that celecoxib, a highly selective COX-2 inhibitor, is superior to placebo and similar to traditional NSAIDs in the short-term treatment of pain due to osteoarthritis, although it has been associated with adverse effects such as headache, change in bowel habits, abdominal discomfort, and dizziness. Celecoxib also has been shown to be as effective as traditional NSAIDs in the treatment of rheumatoid arthritis, but it may cause fewer adverse effects, including endoscopically documented ulcers. Celecoxib is metabolized in the liver by the cytochrome P-450 isozyme CYP2C9, and thus serious drug interactions are possible. In the treatment of osteoarthritis, rofecoxib has been shown to be as effective as traditional NSAIDs and may cause fewer endoscopically documented ulcers, but its complete adverse-effect profile is not known. Until the selective COX-2 inhibitors are widely used and more clinical as well as pharmacoeconomic studies are published, the exact role of COX-2 therapy cannot be determined. words: cyclooxygenase, celecoxib, rofecoxib, rheumatoid arthritis, osteoarthritis.