Abstract
Per- and polyfluoroalkyl substances (PFAS) may disrupt thyroid hormones although the literature shows mixed evidence of this effect and exposure to mixtures of PFAS remains poorly understood. We used the Michigan PFAS Exposure and Health Study cohort to examine linear and nonlinear associations between serum PFAS concentrations, both alone and as a mixture, and serum thyroid hormone concentrations. Study participants included 728 adolescents and adults living in an area with past PFAS contamination of drinking water. We quantified 39 individual PFAS and four thyroid hormones in serum from participants between the years 2020 and 2021. Linear regression, weighted quantile sum (WQS) regression, supervised Principal Component Analysis (PCA), and Bayesian Kernel Machine Regression (BKMR) were used. When analyzed individually, a 1% increase in PFUnA serum concentrations was associated with a 0.023% decrease in TT3 concentration (95% CI: -0.04%, -0.01%, p < 0.05). All three mixture analyses consistently indicated an inverse relationship between PFAS mixtures and TT3 concentrations: (1) a one standard deviation increase in the WQS of the PFAS mixture was associated with a 2.0% decrease in TT3 concentration (95% CI= -4%, 0%, p < 0.05) adjusting for covariates, (2) using PCA, one standard deviation increase in a PFAS mixture was associated with a 1.2% decrease in TT3 (95% CI: -2.1%, -0.4%), and (3) BKMR similarly suggested a negative association between the PFAS mixture and TT3. We observed cross-sectional associations between a mixture of serum PFAS concentrations and thyroid hormone dysregulation, largely manifesting as decreased TT3 serum concentrations.