Oral-Contraceptive Use, Anovulatory Action, and Risk of Epithelial Ovarian Cancer

Abstract

The Authors Reply: We are grateful to Risch for giving us the opportunity to clarify further the question of oral contraceptives versus anovulation in conferring protection against epithelial ovarian cancer. We are currently undertaking an analysis concentrating on ovulatory life and its contributory factors in relation to ovarian cancer, which we hope will address in depth the issues raised by Risch. In the interim, we shall respond to Risch’s comments by pointing out that if oral contraceptive (OC) use influences ovarian cancer risk solely by suppressing ovulation, the inclusion of an estimate of lifetime ovulations (“ovulatory life”) in the analytic model should wholly nullify the OC effect. In other words the expected value of the relative risk of oral contraceptive exposure, however parameterized, should be unity in the presence of the ovulatory life variable. In our paper, we estimated that a year of OC use was associated with a 7% reduction in ovarian cancer risk (95% confidence interval 4–9%) in a model that included a carefully constructed estimate of ovulatory life. We argued further that even if errors in computing the latter value had led to imperfect control of confounding, the effect of OC use would not be seriously overestimated. We apologize for our oversight in claiming priority for our analysis including both OC use and ovulatory life, which properly belongs to Risch et al ’s 1983 paper. 1 We were aware of its existence but had overlooked their results owing to differences in terminology and analytic method. Victor Siskind Adèle Green David Purdie Christopher Bain