Factors affecting the association of oral contraceptives and ovarian cancer

Abstract

We investigated the relation between epithelial ovarian cancer and the use of oral contraceptives in a case-control study of 144 white women under the age of 60 who had ovarian cancer and 139 white women under 60 who were selected from the general population. We observed a decreased risk for ovarian cancer associated with the use of oral contraceptives in subjects 40 through 59 years of age at the time of the study. The relative risk, adjusted for parity, was 0.11, with 95 per cent confidence limits of 0.04 to 0.33. In contrast to the findings in older women, a decreased risk for ovarian cancer associated with oral-contraceptive use was not found in women under 40. In this group, the adjusted relative risk associated with any use of oral contraceptives was 1.98, with 95 per cent confidence limits of 0.74 to 5.27. The lowest risk for ovarian cancer associated with the use of oral contraceptives was observed in older parous subjects and in women who had discontinued use more than 10 years previously.

PIP: Results of a case-control study of histologically verified ovarian cancer including details about oral contraceptive (OC) use are presented. 144 white Massachusetts residents under age 60 with primary epithelial ovarian cancer were matched with 139 control women randomly selected from lists of Massachusetts residents who matched the cases by residence, age within 2 years, and race. Twice as many cases as controls were single and nulliparous. 34 cases had at some time used OCs compared with 48 controls. The relative risk of ovarian cancer associated with OC use, standardized for age and parity, was .38 with 95% confidence limits of .15 to .96. A decreased risk associated with OC use in subjects 40-59 years at the time of the study was observed; the relative risk, adjusted for parity, was .11, with 95% confidence limits of .04 to .33. A decreased risk associated with OC use was not found in women under 40, whose adjusted relateive risk associated with OC use was 1.98, with 95% confidence limits of .74 to 5.27. No significant differences were observed between parous and nonparous subjects. Differences in histological characteristics of cancer in OC users and nonusers were not significant, but a larger study might have uncovered differences. The lowest risk for ovarian cancer associated with the use of OCs was observed in older parous subjects and in women who had discontinued use more than 10 years previously. Potential biases that should be considered in this study include the high refusal rate among controls and the possibility that underlying infertility might confound any association of OCs and ovarian cancer.