Rheumatoid arthritis (RA) is associated with increased mortality and an increased risk of neoplasms of the immune system (NIM). To establish whether immunosuppressive therapy alters these risks, a matched cohort study was conducted. The exposed cohort were 259 RA patients, resident in the UK, who first received immunosuppressive drugs (mainly azathioprine, cyclophosphamide and chlorambucil) between 1979 and 1982. The unexposed cohort were 259 patients matched for age, sex and disease, resident in the USA, who had never received immunosuppressives. Both cohorts had no prior reported malignancies and were followed for 10 yr. There was a small increase in mortality in the exposed compared to the unexposed cohort. Most of the excess deaths were due to malignancy. The relative risk (RR) of developing malignancy [1.5 (95% CI 0.9-2.3)] was lower than the RR of dying from malignancy [4.2 (95% CI 1.7-10.0)]. The RR of developing a NIM in the immunosuppressive-exposed group was 7.0 (95% CI 0.9-56.5). These results may be explained in part by differences in cancer registration and death rates between the UK and the USA. Nevertheless, the results suggest that exposure to immunosuppressive therapy increases the 10 yr malignancy risk in RA, but not mortality from other causes.