Prenatal Concentrations of Perfluoroalkyl Substances and Maternal Beta Cell Function at 7 to 9 Years of Follow-Up

CONTEXT: Epidemiological evidence regarding prenatal per- and polyfluoroalkyl substance (PFAS) exposure and long-term maternal metabolic health outcomes is lacking. OBJECTIVE: Quantify associations between prenatal PFAS concentrations and maternal metabolic biomarkers of glucose homeostasis 7 to 9 years later. METHODS: We measured second trimester plasma concentrations of 9 PFAS in participants enrolled in the Maternal-Infant Research on Environmental Chemicals (MIREC) study. We measured individual biomarkers of glucose homeostasis (fasting intact proinsulin, C-peptide, insulin, glucose, and hemoglobin A1C levels) in samples collected 7 to 9 years after the MIREC pregnancy (n = 258) and derived indicators of pancreatic beta cell function (proinsulin to insulin PI:INS, proinsulin to C-peptide PI:CP ratios) and insulin resistance (homeostatic model assessment for insulin resistance HOMA-IR, triglyceride-glucose index). Using multivariable linear regression models, we quantified the percent change in each outcome per doubling of individual PFAS concentrations. We used quantile g-computation and weighted quantile sum regression to evaluate the mixture of PFAS. RESULTS: Prenatal perfluorononanoic acid and perfluorodecanoic acid concentrations were associated with 13.9% (95% CI: 0.8, 28.8) and 10.5% (95% CI: -1.0, 23.4) higher HOMA-IR values as well as 11.9% (95% CI: 0.1, 25.1) and 8.9% (95% CI: -1.5, 20.3) higher fasting insulin concentrations, respectively. A doubling of perfluorooctanoic acid concentrations was associated with increases in intact proinsulin concentrations (12.8% 95% CI: -3.5, 31.8) and beta cell function ratios (PI:INS: 11.5% 95% CI: -4.4, 30.1; PI:CP: 13.5% 95% CI: -2.4, 32.0). CONCLUSION: Prenatal exposure to PFAS may impact long-term maternal insulin resistance and beta cell function, key risk factors for type 2 diabetes. These associations differ by specific PFAS.