OBJECTIVE: To evaluate whether coffee, tea, and caffeine consumption are risk factors for rheumatoid arthritis (RA) onset among older women. METHODS: These factors were evaluated in a prospective cohort study that was initiated in 1986 and that included 31,336 women ages 55-69 years without a history of RA. Risk factor data were self-reported using a mailed questionnaire. Through 1997, 158 cases of RA were identified and validated against medical records. The relative risk (RR) and 95% confidence interval (95% CI) were used as the measures of association and were adjusted for age, alcohol use, smoking history, age at menopause, marital status, and the use of hormone replacement therapy. RESULTS: Compared with those reporting no use, subjects drinking > or =4 cups/day of decaffeinated coffee were at increased risk of RA (RR 2.58, 95% CI 1.63-4.06). In contrast, women consuming >3 cups/day of tea displayed a decreased risk of RA (RR 0.39, 95% CI 0.16-0.97) compared with women who never drank tea. Caffeinated coffee and daily caffeine intake were not associated with the development of RA. Multivariable adjustment for a number of potential confounders did not alter these results. The associations of RA onset with the highest categories of decaffeinated coffee consumption (RR 3.10, 95% CI 1.75-5.48) and tea consumption (RR 0.24, 95% CI 0.06-0.98) were stronger in women with seropositive disease compared with those with seronegative disease (RR 1.54, 95% CI 0.62-3.84 and RR 0.93, 95% CI 0.27-3.20, respectively). CONCLUSION: Decaffeinated coffee intake is independently and positively associated with RA onset, while tea consumption shows an inverse association with disease onset. Further investigations of decaffeinated coffee and tea intake as arthritis risk factors are needed to verify these findings and explore their biologic basis.