A total of 14,856 women, including 921 pregnant subjects, were tested for rheumatoid factor; 4,562 were using oral contraceptives at the time of testing. The prevalence of rheumatoid factor increased directly with age. The age-adjusted prevalence of rheumatoid factor was lower in oral contraceptive users than in nonusers but this difference was not statistically significant. Rheumatoid factor remained positive in 39% of subjects undergoing retesting after an average interval of 16 months. Those women with higher titers of rheumatoid factor were more likely to remain positive (81%). Of the women having positive tests, 5.4% were identified as having rheumatoid disease.The influence of oral contraceptives (OCs) on rheumatic diseases and serologic abnormalities has been of interest from the time these hormones first came into widespread use. Because pregnancy had been observed to induce remissions among some women with rheumatoid arthritis ( l ) , and because OCs were noted to have an effect parallel to the pregnant state, initial speculation centered on the potential therapeutic benefits of contraceptive medications (2). A subsequent, challenged study (3,4) appeared to refute early claims of clinical improvement, and the use of hormones in treatment of rheumatoid disease has not gained popular acceptance.More recently, attention has turned to the adverse effects of OCs on rheumatic diseases in general (5) and on systemic lupus erythematosus (SLE) in particular (6-8). Conclusions from several studies have varied, depending upon both the scope and nature of the observations and the subjects under examination. There has been a limited number of investigations of the relationship between oral contraceptives and serologic abnormalities. Kay et a l ( 9 ) have found that women using OCs had a higher prevalence of positive tests for antinuclear antibodies, rheumatoid factor, and C-reactive protein. On the other hand, Dubois and associates (10) reported no evidence of such serologic changes in a group of 30 normal women tested for antinuclear antibodies, rheumatoid factor, and LE cell phenomenon. In another study reported by McKenna et al(1 I ) , a group of 176 women taking OCs were found to have an increased frequency of rheumatoid factor, as compared to 95 women using intrauterine devices. More recently, Tarzy et al (12) failed to demonstrate an increased prevalence of antinuclear antibodies, LE cells, or rheumatoid factor among 80 women using OCs as com-