Rheumatoid arthritis (RA) is a potentially devastating form of chronic arthritis. Methotrexate is the cornerstone of drug therapy of the disorder, and may slow or prevent joint damage. Unfortunately, this agent is not without adverse effects. Furthermore, increasing age has been been invoked as a predictor of greater toxicity and drug discontinuation by some, but not all, investigators. To assess the effect of age and other covariates on methotrexate discontinuation in a population-based sample of older patients with newly diagnosed RA. We studied the health administrative databases covering residents of the province of Québec, Canada. In these databases, we identified 246 individuals aged>or=65 years with newly diagnosed RA who had been started on methotrexate. We assessed discontinuation of methotrexate therapy using Cox proportional hazards regression models, with potential predictors of discontinuation being age, sex, co-morbidity, methotrexate dose and route (oral vs intramuscular), folic acid coadministration and disease severity. Five patients died or were lost to follow-up in the database at 6 months, and there were ten such patients at 1 year. Six months after the initial prescription of methotrexate therapy, about 80% (n=192) of remaining subjects continued to be prescribed the drug. By 1 year, 161 of 236 (68.2%) subjects continued to be prescribed the drug; by 2 years, only 108 of 217 (49.8%) subjects continued to receive the drug. Increasing age was associated with a greater risk of methotrexate discontinuation. Our population-based data indicate that increasing age is associated with a greater tendency for methotrexate discontinuation in patients with newly diagnosed RA. These results emphasize the need to ensure that older patients with RA are provided with effective therapy to minimize the effects of this chronic, potentially disabling disease.