There have been conflicting reports on, and no plausible biologic explanation for, a protective effect of oral contraceptive (OC) use on the development of rheumatoid arthritis (RA). Therefore, we investigated aspects of behavior related to OC use that could explain the preventive effect of OC on the onset of RA. In the present case—control study, past pregnancy, age at first pregnancy, and pregnancies with adverse outcome were studied as possible risk factors for RA. Interview information on reproductive variables was obtained from 135 young adult women with confirmed definite or classic RA of recent onset, and from 378 control patients with soft tissue rheumatic disorders or osteoarthritis. All patients had at least 2 years of followup to confirm the consistency of the diagnosis. We found a decreased risk of RA in women who had been pregnant. The risk of RA in women who had ever been pregnant compared with women who had never been pregnant was 0.49 (95% confidence interval 0.27–0.91). The earlier the first pregnancy, the lower the risk of RA. Pregnancy with adverse outcome (i.e., gestation <25 weeks) did not substantially change the risk of RA (relative risk 0.73, 95% confidence interval 0.50–1.07). The protective effect of pregnancy was independent of OC use, the presence of HLA–DR4, or a family history of RA. Immune modulation by female hormonal influences could be an explanation for the results of the present study.