Toward an understanding of NSAID-related adverse events: the contribution of longitudinal data.

The ARAMIS (Arthritis, Rheumatism and Ageing Medical Information System) databanks have been used to objectify and quantify drug toxicity. The relative risk of a gastrointestinal (GI)-provoked hospitalization was more than five times greater in patients taking non-steroidal anti-inflammatory drugs (NSAIDs) than in non-NSAID-treated patients, with an excess hospitalization rate of 1.3% per annum. Additionally, there was an excess GI-related death rate of around 3% in rheumatoid arthritis (RA) patients compared with the normal population. Age, previous NSAID-related GI events, prednisone use, higher doses and greater disability predicted high-risk patients. A toxicity index showed clear differences between NSAIDs, with aspirin, salsalate and ibuprofen emerging as the least toxic, and meclofenamate and indomethacin as the most toxic. Disease modifying anti-rheumatic drugs (DMARDs) were, surprisingly, found to have similar toxicity scores to the NSAIDs. This supports the contemporary practice of employing DMARDs earlier and more aggressively in the course of RA.