Although adverse effects of non-steroidal anti-inflammatory drugs (NSAIDs) occur in only a small proportion of users, the widespread use of these drugs has resulted in a substantial overall number of affected persons who experience serious gastrointestinal complications. Dyspeptic symptoms are estimated to occur in 10–60% of NSAID users and lead to discontinuation of treatment in 5–15% of rheumatoid arthritis patients taking NSAIDs. It is now well established that the point prevalence of peptic ulcer disease in patients receiving conventional NSAID therapy ranges between 10 and 30%, representing a 10–30-fold increase over that found in the general population. One of 175 users of conventional NSAIDs in the USA will be hospitalized each year for NSAID-induced gastrointestinal damage. The mortality of hospitalized patients remains about 5–10%, with an expected annual death rate of 0.08%. The selective COX-II inhibitors (rofecoxib, celecoxib, parecoxib, etoricoxib, valdecoxib, lumiracoxib) show consistently comparable efficacy to that of conventional non-steroidal anti-inflammatory drugs (NSAIDs) in patients with rheumatoid arthritis and osteoarthritis, but have a significantly reduced propensity to cause gastrointestinal toxicity. In many cases, the gastric effects of therapeutically active doses of COX-II inhibitors are indistinguishable from placebo. The safety benefits of COX-2 inhibitors given alone appear similar to combined therapy with conventional NSAIDs and gastroprotective agents. These findings warrant the consideration of COX-II inhibitors as first-line therapy in patients requiring long-term pain control.