Objectives-To determine the probability of drug continuation in a large cohort of methotrexate treated rheumatoid arthritis (RA) patients, the reasons for discontinuation of methotrexate, the overall survival of the members of this cohort, and the causes of death in these patients. Methods-Yearly follow up was conducted in methotrexate treated RA patients who formed a cohort between 1981 and 1986 at a tertiary care centre. The probability of drug continuation and the patients' survival were calculated using standard statistical procedures; standardised mortality ratios were calculated using death certificate data and USA general population and mortality tables. Results-The probability of methotrexate continuation at 10 years from the time the first members entered the cohort was 30%. Toxicity (and its severity) was the most frequent cause of discontinuing methotrexate. The cumulative probability of survival was 85% for women and 45% for men. A greater than expected number of deaths from infections was observed, but the number of deaths from cancer and cardiovascular diseases were within the range expected. Conclusions-Toxicity remains the most common cause for methotrexate discontinuation. Survival was comparable to that of other RA cohorts. Methotrexate may be implicated as an associated factor in the deaths from infections.