Cancer and rheumatoid arthritis: epidemiologic considerations.

Statistical methods used to investigate the epidemiology of cancer in rheumatoid arthritis are reviewed and their relative merits discussed. A cohort analysis of cancer morbidity was carried out on a consecutive series of 489 patients with rheumatoid arthritis seen at the Queen Elizabeth Medical Centre, Birmingham, United Kingdom between 1964 and 1978 and followed to December 31, 1983. Forty-two cancers were observed in the series in comparison with 31.13 expected (p less than 0.05) on the basis of cancer morbidity rates for the West Midlands region. The excess was due to the high relative risk of cancers of lymphatic and hematopoietic tissues (observed = 11, expected = 1.27, relative risk = 8.7, p less than 0.001). The effects of confounding factors, including duration of rheumatoid arthritis and hospital selection on the level and pattern of risk over time were examined. When two cases diagnosed soon after first attendance at hospital were excluded, lymphomas (ICD 200, 201, 8th Revision) showed a pattern of increasing relative risk with time from five years after first attendance. The increasing risk appears to be unrelated to the use of immunosuppressive or cytotoxic drugs.