Individual and mixture associations of placental per- and polyfluoroalkyl substance with neurodevelopment at 12 and 24 months of age

Abstract

Maternal serum per- and polyfluoroalkyl substances (PFAS) are linked to infant neurodevelopment vulnerabilities. However, the impact of placental PFAS exposure, a more proximal in utero exposure estimate, remains unknown. We hypothesize that elevated placental PFAS levels are associated with delayed neurodevelopment at 12 and 24 months.

Mother-infant dyads (n = 151) were enrolled in a prospective cohort in Arkansas, US. PFOA, PFOS, PFHxS, PFNA, and PFDA were detected in >65 % of placentas. The Bayley Scales of Infant and Toddler Development (BSID)-III was administered at 12 and 24 months. Individual associations of each placental PFAS on BSID-III scores were examined using linear regressions. Mixture effects were assessed using quantile g-computation and Bayesian Kernel Machine Regression.

Placental PFAS were jointly associated with lower cognitive (Ψ = -2.6; 95 % CI = -6.0, 0.9) and motor (Ψ = -2.9; 95 % CI = -6.6, 0.9) scores at 12 months and higher social-emotional scores (Ψ = 3.3; 95 % CI = -1.9, 8.5) at 24 months among males. Among females, higher PFAS mixture levels were associated with higher language (Ψ = 3.3; 95 % CI = -1.2, 7.7) and social-emotional (Ψ = 3.7; 95 % CI = -3.3, 10.6) scores at 24 months. Linear regressions showed PFHxS and PFOS were associated with lower cognitive scores at 12 months and PFDA was associated with lower motor scores at 12 months among males, while PFOS and PFNA were associated with higher social-emotional scores at 24 months among females, although confidence intervals included the null. In a matched subset with maternal serum, cord serum and placental PFAS (n = 98), we observed mostly null associations consistently across biomatrices and noted few associations, such as a negative relation of cord PFOA and PFHxS with language scores at 12 months among females.

Overall, placental PFAS levels were modestly associated with performance on neurodevelopmental assessments in early childhood and these relationships were sex- and compound-specific, and most confidence intervals cross the null.