Risk of cervical dysplasia in users of oral contraceptives, intrauterine devices or depot-medroxyprogesterone acetate. The New Zealand Contraception and Health Study Group.

Three cohorts of women aged 20-39 attending medical practitioners or family planning clinics in New Zealand for contraceptive advice were followed prospectively for five years. The three cohorts were defined by use (or at least prescription) of one of three study contraceptive methods at the time of beginning of follow-up--oral contraceptives (OC), intrauterine devices (IUD) or depot-medroxyprogesterone acetate (MPA)--and numbered 2469, 2072 and 1721 women, respectively. Follow-up was intended to be annual and included cervical smear as well as interim contraceptive and medical histories. A positive outcome (referred to here as 'dysplasia') was any degree of definite dysplasia or carcinoma of the cervix diagnosed cytologically by a central study laboratory, and confirmed by histology or analysis of DNA ploidy. In the three cohorts (OC, IUD and MPA, respectively), 12,839, 10,774 and 8,984 person-years of follow-up were accumulated and 125, 92 and 101 cases of dysplasia were confirmed. Crude annual rates of dysplasia per 1,000 person-years were approximately the same in the OC (9.6) and IUD (8.4) cohorts. Crude rates were higher in the MPA cohort (11.3 per 1,000 person-years). However, important confounding factors, principally smoking and sexual behavior, were identified with rate ratios ranging between 1 and 3 over the range of the potentially confounding variables, and multivariate analyses revealed no evidence of increased risk for the MPA cohort when these factors were taken into account. At least over the short term, there appears to be no difference in risk of cervical dysplasia between women using these three methods of contraception if differences between groups in respect to known confounding factors are taken into account.