Mid-Childhood Plasma Concentrations of Per- and Polyfluoroalkyl Substances, Modifiable Lifestyle Factors, and Bone Mineral Density Through Late Adolescence

= 484; 50% female), we used sex-stratified linear regression and quantile g-computation mixture models to examine associations of mid-childhood (median: 7.8 years; 2007-2010) plasma PFAS concentrations with a dual-energy X-ray absorptiometry total-body aBMD Z-score in early and late adolescence (median: 12.9 and 17.6 years, respectively). We explored stratum-specific estimates by parent/self-reported physical activity and dairy intake. Using linear mixed models, we evaluated associations with aBMD accrual from mid-childhood through late adolescence. Females with higher perfluorooctanoate (PFOA) and perfluorodecanoate (PFDA) had lower early adolescent aBMD Z-score e.g., β(95%CI) per doubling PFOA: -0.19(-0.41, 0.03)]. Youth with higher PFOA and PFDA had lower late adolescent aBMD Z-score, but CIs were wide e.g., PFOA: females, -0.12(-0.40, 0.16); males, -0.10(-0.42, 0.21). Mixture models generally corroborated single PFAS results, and in linear mixed models, females with higher PFAS concentrations, and males with higher PFOA, had slower aBMD accrual. Less active males with higher PFOA, PFDA, and the PFAS mixture had lower late adolescent aBMD Z-score. Some PFAS appeared more negatively associated with the aBMD Z-score among those who consumed less dairy, but there was not consistent evidence of effect modification. Exposure to select PFAS may affect bone accrual through adolescence, with possible resilience conferred by greater physical activity and dairy intake.