Reproductive history and endometriosis among premenopausal women

Abstract

Objective: To clarify the temporal complexities of the relation between reproductive factors and endometriosis.

Methods: We analyzed 10 years of prospective data from the Nurses' Health Study II cohort. Information was obtained through questionnaires sent every 2 years to 116,678 women aged 25-42 years when enrolled in 1989. Cox proportional hazards models were used to adjust for age, calendar time, and confounding variables.

Results: During 726,205 woman-years of follow-up, 1,721 cases of laparoscopically confirmed endometriosis were reported among women with no past infertility. Greater incidence was observed among women with an earlier age at menarche (rate ratio of 1.3 comparing menarche at age < 10 to age 12 years; 95% confidence interval [CI] 1.0-1.8; P value, test for trend < .001) and shorter cycle length during late adolescence (rate ratio of 1.3 comparing < 26 days to 26-31 days; 95% CI 1.1-1.5). Time to cycle regularity was not associated with risk. Among parous women, a linear decrease in risk was observed with number of liveborn children (rate ratio of 0.5 comparing > 3 with 2 children; 95% CI 0.4-0.7; P value, test for trend < .001) and lifetime duration of lactation if time since last birth was less than 5 years (rate ratio of 0.2 comparing > 23 months with never; 95% CI 0.1-0.4; P value, test for trend < .001).

Conclusion: Hormonal and anatomical changes associated with menstruation and pregnancy may affect the rate of laparoscopically confirmed endometriosis. Within this cohort, risk was greatest among nulliparous women with earlier age at menarche and shorter menstrual cycles. Among parous women, parity and lifetime duration of lactation were associated with decreased risk.