The safety of the cyclo-oxygenase-2 (COX-2) selective NSAID, nimesulide, has been evaluated from information (a) in clinical trials in osteoarthritis that have been performed in Europe as well as in earlier pilot studies that were performed in patients with rheumatoid arthritis in the USA, and (b) in post-marketing studies (PMS) that have been performed by the manufacturer since the introduction of the drug in Europe and some South American countries. Upon analysis there have been clear indications of elevation of liver enzymes being related to the drug in 3/753 patients who were investigated in clinical trials in osteoarthritis. The results of the analysis of adverse drug reactions (ADRs) in patients with osteoarthritis (which is the principal indication for the drug) in PMS showed that the greatest number of these were in the skin and appendages, digestive system and organs wherein metabolic effects were manifest. There was a relatively low number of reports of gastrointestinal ulceration and haemorrhage, and these observations are in agreement with published experimental studies in humans and laboratory animals. Recent reports of ADRs in the liver in 25 patients were analysed in detail. In many of the cases there was evidence of confounding disease (e.g. cancer, prior liver damage) or prior or concurrent intake of known hepatotoxic NSAIDs (diclofenac, aspirin). Often elevations of liver enzymes above the laboratory norms are the only indication of liver injury and this in many cases is variable. The major cases of elevated liver enzymes and other liver changes have been in elderly patients.This first extensive analysis of ADRs from a COX-2 selective NSAID, nimesulide, indicates that there is a relatively low incidence of ADRs especially in the gastrointestinal tract, while those in the liver are within or below the general incidence with other NSAIDs.