Objective: To prospectively examine the association between endogenous hormones and development of ovarian cancer.
Design: Nested case-control study.
Setting: A population-based serum bank in Washington County, Maryland.
Participants: Serum samples were collected in 1974 from 20305 county residents and stored at -70 degrees C. From 1975 through 1989, a total of 31 cases of ovarian cancer were identified in women who were not taking hormones at the time of blood collection. These cases were matched to 62 controls on age, menopausal status, and, for premenopausal women, number of days from the beginning of the last menstrual period.
Main outcome measure: Prediagnostic endogenous hormone levels of cases and controls were compared.
Results: Mean follicle-stimulating hormone levels were lower among cases (43.3 IU/L) compared with controls (54.4 IU/L) (P = .04), and increasing levels were associated with significantly lower risk (P for trend = .01), particularly among postmenopausal women. Luteinizing hormone levels were 9% lower among cases than controls, but the difference was not statistically significant (P = .39). Compared with controls, cases had higher androstenedione levels (4.5 nmol/L vs 3.3 nmol; P = .03) and higher dehydroepiandrosterone (DHEA) levels (15.9 nmol/L vs 9.7 nmol/L; P = .02). The risk of ovarian cancer increased with higher levels of androstenedione and DHEA sulfate (P for trend = .008 and .11, respectively). These associations were not materially different between premenopausal and postmenopausal women.
Conclusion: The results suggest that women with low serum gonadotropin levels or high androgen levels have an increased risk of ovarian cancer. These findings do not support the hypothesis that pituitary gonadotropins increase the risk of ovarian cancer. Replication of the study in other populations is highly desirable.