Associations of serum persistent organic pollutant concentrations with incident diabetes in midlife women: The Study of Women’s Health Across the Nation Multi-Pollutant Study

Abstract

BackgroundOrganochlorine pesticides (OCPs), polybrominated diphenyl ethers (PBDEs), and polychlorinated biphenyls (PCBs) are persistent organic pollutants (POPs) that can negatively impact metabolic health through pathways including endocrine disruption. Few studies have evaluated diabetes risk associated with PBDEs. Little is known about the joint effect of exposure to POP mixtures on diabetes risk. ObjectivesWe investigated the relationship between POPs, individually and as mixtures, and diabetes development over 18 years (1999–2016) in midlife women. MethodsWe measured lipid-standardized serum concentrations of 34 PCBs, 19 OCPs, and 14 PBDEs in 1040 midlife women aged 45–56 years from the Study of Women’s Health Across the Nation. We tested the association between POPs measured in 1999/2000 and incident diabetes using Cox proportional hazards models. We evaluated diabetes risk associated with the overall POP mixture using Quantile-Based G-Computation (QBGC). ResultsFor most mixture components, single pollutant and mixtures analyses indicated null associations with diabetes risk, however results were inconsistent. After adjustment, hazard ratios (HRs) of developing diabetes (95% CI) associated with upper exposure tertiles (T2/T3) compared with the first tertile (T1), were 1.7 (1.0, 2.8) at T2 and 1.5 (0.84, 2.7) at T3 for hexachlorobenzene and 1.9 (1.1, 3.3) at T2 and 1.6 (0.88, 2.9) at T3 for PCB 123. A doubling of PBDE 47 was associated with 1.11 (1.00, 1.24) times the risk of T2D. QBGC identified no association for the overall joint effect of the POP mixture on diabetes (HR = 1.04 [0.53, 2.07]). ConclusionExposure to a mixture of PCBs, OCPs, and PBDEs was not associated with incident diabetes in midlife U.S. women, although some individual POPs demonstrated significant yet inconsistent associations with diabetes. Non-linear and non-monotonic dose-response dynamics deserve further exploration. More research is needed on the diabetogenic effects of PBDEs.