PFAS endocrine disruption affecting pubertal development: exposure-response is steeper below median PFAS serum concentrations

We have recently published (van Larebeke, Environ Health 24:34, 2025) consistent and biologically relevant associations of PFAS serum concentrations with sex hormone levels in male adolescents and with a significant delay in physiological processes occurring during puberty in girls and boys. Associations with thyroid hormones differed significantly by sex with only for boys significant positive associations of FT3 with PFOA, PFHxS and PFNA. Here we report a follow-up analysis to test a work hypothesis stating that, as can be expected for receptor mediated effects, the observed associations would be stronger at lower PFAS serum concentrations (e.g. below the median) than at higher PFAS serum concentrations. To allow for a comparison between associations of biological/health effects with PFAS concentrations below or above median values, PFAS concentrations without log transformation were used in linear multiple regression with continuous effect parameters, logistic models with binary effect parameters, and Ordinal Multinomial Probit models with ordinal effect parameters. Regression coefficients and other statistical parameters were calculated with the Statistica 14.0 program. Significance of the difference between a regression coefficient observed for an association with a certain PFAS below the median serum concentration compared to above the median was assessed using a two‑tailed Wald test. The median PFAS serum values observed in our study were below present health based guidance values. Consistent with our working hypothesis, the lower exposed subpopulations (with serum concentrations below the median) had a greater change in outcome per unit increase in PFAS exposure (a stronger regression coefficient) than the higher exposed subpopulations concerning 72 of the 84 significant associations described previously and this was most pronounced for PFHxS, and to a lesser extent for PFOS(branched), PFOA and sum parameters. Continuous piecewise linear models, models with a quadratic term and, in the range of PFAS concentrations where sufficient data were present, also weighted least square graphs, yielded results supporting stronger regression coefficients below median PFAS serum concentrations. Our observations suggest that an important part of the endocrine disrupting effect of PFAS on adolescents occurs at serum levels below present health based guidance values.