Arthritis and related musculoskeletal conditions occur with great frequency in the population world wide, causing significant morbidity and, in some instances, increased mortality. Affecting both the young and the old, 15% of the population in the US was estimated in 1995 to have some form of arthritis with an increase to 18% projected by the year 2020 [1]. The economic impact of arthritis and related disorders in the US alone was estimated to be 194.4 billion US dollars in 1992 and future costs are virtually certain to increase given the chronic nature of these diseases, their expanding prevalence and the considerable expense associated with newer therapies [2]. With no cure presently available, the aim of current treatment is to reduce inflammation, ameliorate symptoms and improve functional capacity. Non-steroidal anti-inflammatory drugs (NSAIDs), which suppress the formation of pro-inflammatory prostaglandins by antagonising the action of cyclooxygenase (COX), have been the mainstay of arthritis treatment for hundreds, if not, thousands of years. The clinical use of NSAIDs, however, has long been associated with significant toxicity. The recognition of two COX isoforms, cyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2), both suppressed by traditional NSAIDs, has led to an expanded hypothesis of NSAID action which consists of two postulates, namely, the efficacy of NSAIDs in the treatment of arthritis is due to the suppression of COX-2, while much of the toxicity associated with non-selective NSAIDs is the consequence of COX-1 suppression. The emergence of agents which selectively inhibit COX-2 has made it possible to clinically evaluate the validity of each of these postulates. In this report, the published experience with selective COX-2 inhibitors in the treatment of mechanical and inflammatory arthropathies is reviewed to examine the premise that isolated COX-2 suppression is comparable in efficacy to the dual COX-1/COX-2 suppression produced by non-selective NSAIDs.