Impact of lactational per- and polyfluoroalkyl substances exposure on infant gut microbiota and potential mediating effects on infant neurodevelopment.
Level 3 - non-randomized controlled study
Prospective observational mother-infant cohort study
PubMed 41485339 · doi:10.1016/j.jhazmat.2025.140959
What was done
A longitudinal cohort enrolled 114 mother-infant pairs to investigate whether lactational exposure to per- and polyfluoroalkyl substances (PFAS) affects infant neurodevelopment through the gut microbiome. Human milk and infant fecal samples were collected at 0.5, 1, and 3 months postpartum. Infant neurodevelopment was assessed at 6 months of age using the Ages & Stages Questionnaires, Third Edition (ASQ-3) and ASQ:Social-Emotional (ASQ:SE). Mediation analyses evaluated the role of specific gut microbial taxa in linking PFAS exposure to neurodevelopmental scores.
What was found
Higher concentrations of perfluorooctane sulfonic acid (PFOS), perfluorooctanoic acid (PFOA), 2-(N-methylperfluorooctanesulfonamido) acetic acid, and total PFAS mixture in breast milk were associated with significantly lower scores across communication, gross motor, fine motor, problem solving, and personal-social domains. PFAS exposure altered the infant gut microbiome, increasing relative abundances of Klebsiella variicola, Enterococcus faecium, Clostridium perfringens, and Veillonella atypica, while PFOS, perfluorononanoic acid (PFNA), and a chloroperfluoroether sulfonate substitute decreased Bifidobacterium breve and Streptococcus. Mediation analysis indicated that Clostridium perfringens mediated 13% of total PFAS exposure's effect on fine motor deficits, and Streptococcus mediated 11% of PFOS exposure's effect on gross motor deficits. Exact effect sizes, p-values, and confidence intervals were not reported in the abstract.
Why it matters
This study identifies early-life gut dysbiosis as a potential biological mediator connecting breast milk PFAS exposure to early developmental delays, highlighting candidate microbial pathways for future mechanistic research.
Limits
The sample size is relatively small (114 pairs), limiting statistical power for complex mediation models and multi-pollutant mixture analyses. Follow-up ended at 6 months using parent-reported screening questionnaires rather than direct standardized clinical neurodevelopmental testing. Residual confounding from unmeasured maternal dietary, environmental, or socioeconomic factors remains possible.