ABSTRACT
OBJECTIVES: To estimate the incidence of rheumatoid arthritis (RA) in primary care and to investigate associations with consultation behaviour, risk factors, and comorbidities, using the UK General Practice Research Database (GPRD). METHODS: Subjects with a first-ever diagnosis of RA between 1 January 1996 and 31 December 1997 (n = 579) were identified from a cohort of 1 206 918 subjects aged 20-79 years without cancer. Controls from the same cohort were frequency-matched to the RA group by age, sex, and calendar year (n = 4234). Odds ratios (ORs) and 95% confidence intervals (CIs) of being diagnosed with RA in association with a range of factors were estimated using logistic regression analysis. RESULTS: RA incidence was 0.15 per 1000 person-years, was higher in women than in men, and increased with age in both sexes. Consultations and use of non-steroidal anti-inflammatory drugs (NSAIDs) prior to diagnosis were increased in subjects with RA. An increased risk of RA was observed in association with anaemia in the previous year (OR 2.63, 95% CI 1.54-4.48) and with smoking (1.33, 1.07-1.67). A decreased risk of RA was observed in association with infectious diseases (0.68, 0.50-0.94) and pregnancy in the previous year (0.22, 0.06-0.77), diabetes (0.45, 0.26-0.78), and hypertension (0.74, 0.57-0.94). We found no association with alcohol intake, obesity, or use of low-dose aspirin, oral contraceptives, or hormone replacement therapy (HRT). CONCLUSIONS: Smoking was identified as the only significant lifestyle-related risk factor for RA. Infection in the previous year was associated with a reduced likelihood of RA.
