A postmarketing surveillance project monitoring the adverse effect profile of the NSAID, nabumetone, was performed in 10,800 patients with osteoarthritis or rheumatoid arthritis, recruited by general practitioners in the UK. The patients were typically elderly with chronic arthritis, had many concurrent conditions, and had previously received at least 1 NSAID. Patients were followed up at 1, 3, 6 and 12 months after registration, through written information provided by the doctors. Possible serious adverse reactions were monitored by doctors directly notifying the monitoring centre, and by screening procedures at the time of computerised data entry. Only 2% of patients were lost to follow-up during the assessment period. At 12 months, 16% of patients were still receiving nabumetone, 27% of patients discontinued the drug because of lack of efficacy, and 12% because of adverse events. 52 'potentially serious' events and 96 deaths were recorded, but causality was not assessed. Evaluation of all serious events, including death, identified 11 events in 10,800 patients (0.1%) which could possibly be related to nabumetone therapy. Seven of these were gastrointestinal haemorrhage. The preliminary results of this postmarketing surveillance study thus suggest a highly acceptable tolerability profile for nabumetone.