Cancer risk among infertile women with androgen excess or menstrual disorders (including polycystic ovary syndrome)

Abstract

Objective: To define relationships of androgen excesses to cancer risk.

Design: Retrospective cohort study.

Setting: Five large infertility practices.

Patient(s): Among 12,193 women evaluated for infertility during 1965-1988 and traced for cancer incidence through 1999, 2,560 had androgen excess or menstrual disorders; among these, 412 met established criteria for polycystic ovary syndrome.

Intervention(s): None.

Main outcome measure(s): Cancer incidence. Derivation of standardized incidence ratios (SIRs) and 95% confidence intervals (CIs) for cancer risk comparisons with the general population and rate ratios (RRs) for comparisons with other infertility patients.

Result(s): Androgen excess/menstrual disorder patients showed significant SIRs for breast (1.31; 95% CI, 1.05-1.62) and uterine (2.02; 95% CI, 1.13-3.34) cancers and melanoma (1.96; 95% CI, 1.12-3.18). Significant associations for breast and uterine cancers were restricted to primary infertility patients (respective SIRs of 1.53 and 3.48). After adjustment for other cancer predictors, the only excess risk was for uterine cancer among primary infertility patients. Compared with women with secondary infertility and no androgen excess/menstrual disorder, those with primary infertility and a disorder had an RR of 1.88 (95% CI, 0.82-4.32). Cancer risks among the women with polycystic ovary syndrome or androgen excess disorders appeared to be similar to those in the more comprehensive group.

Conclusion(s): Previous findings linking androgen excess disorders to elevated uterine cancer risks might largely reflect underlying risk profiles.

Figure 1

Field and Analytic Status of Eligible Study Subjects, Women Evaluated for Infertility, 1965-1988 **Analyses excluded 10 patients who developed cancer within the first year of follow-up. *Includes 1,151 patients who refused access to their medical record data plus 168 patients who did not reside in cancer registry states and who did not respond to our questionnaire.