Recent studies have indicated that estrogen users have a lower than expected rate of concurrent osteoarthritis. We assessed the association between estrogen replacement therapy and incident symptomatic osteoarthritis, using a nested case-control design. We identified all incident cases of hand, hip, and knee osteoarthritis in women members of the Fallon Community Health Plan, age 20-89 years, from January 1, 1990, to December 31, 1993. For each case, we selected a control woman matched by closest date of birth. We used pharmacy records to classify women as new users, past users, ongoing users (past and new users), and never-users of estrogen replacement therapy. There were 60 informative case-control pairs. After controlling for obesity and health care utilization, we found that new use of estrogen replacement therapy was a predictor of new osteoarthritis diagnosis. Past use was inversely associated with risk of osteoarthritis [adjusted odds ratio = 0.7; 95% confidence interval (CI) = 0.3-1.9]. For ongoing use of estrogen replacement therapy and osteoarthritis, the adjusted odds ratio was 1.4 (95% CI = 0.6-3.3). The associations between osteoarthritis and both new use of estrogen replacement therapy and utilization of services suggest that frequent medical care increases the likelihood of diagnosis of osteoarthritis.