Background: An equivocal literature exists regarding the relation between persistent organochlorine pollutants (POPs) and endometriosis in women, with differences attributed to methodologies.
Objectives: We assessed the association between POPs and the odds of an endometriosis diagnosis and the consistency of findings by biological medium and study cohort.
Methods: Using a matched cohort design, we assembled an operative cohort of women 18-44 years of age undergoing laparoscopy or laparotomy at 14 participating clinical centers from 2007 to 2009 and a population-based cohort matched on age and residence within a 50-mile catchment area of the clinical centers. Endometriosis was defined as visualized disease in the operative cohort and as diagnosed by magnetic resonance imaging in the population cohort. Logistic regression analysis was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for each POP in relation to an endometriosis diagnosis, with separate models run for each medium (omental fat in the operative cohort, serum in both cohorts) and cohort. Adjusted models included age, body mass index, breast-feeding conditional on parity, cotinine, and lipids.
Results: Concentrations were higher in omental fat than in serum for all POPs. In the operative cohort, γ-hexachlorocyclohexane (γ-HCH) was the only POP with a significant positive association with endometriosis [per 1-SD increase in log-transformed γ-HCH: adjusted OR (AOR) = 1.27; 95% CI: 1.01, 1.59]; β-HCH was the only significant predictor in the population cohort (per 1-SD increase in log-transformed β-HCH: AOR = 1.72; 95% CI: 1.09, 2.72).
Conclusions: Using a matched cohort design, we found that cohort-specific and biological-medium-specific POPs were associated with endometriosis, underscoring the importance of methodological considerations when interpreting findings.