Investigators reported that associations between prenatal chemical exposures and child behavior were concentrated in early childhood. In the Drakenstein Child Health Study cohort of 643 mother–child pairs, three urinary PAH metabolites measured in second-trimester maternal urine—2-hydroxyfluorene, 1-naphthol, and 1-hydroxypyrene—were associated with higher externalizing scores during the preschool window (ages 2 to 5 years). The study team provided effect estimates and 95% confidence intervals for each metabolite indicating positive associations with externalizing problems. A composite PAH mixture estimate (quantified by the authors) was associated with both externalizing and internalizing CBCL outcomes in preschool. The source did not provide full metabolite concentration distributions in the abstract, but did report the direction and magnitude of associations for those metabolites and the mixture.
The investigators characterized prenatal psychosocial stress as a five-indicator latent construct derived from confirmatory factor analysis. Data indicated that higher prenatal psychosocial stress scores were linked to increased externalizing problems in both developmental windows examined: preschool (2–5 years) and school-age (6.5–10 years). The authors reported adjusted effect estimates with confidence intervals for the latent stress score on externalizing problems in both windows; unlike the chemical associations, psychosocial associations persisted into later childhood. The abstract did not list each of the five stress indicators in detail; it stated the latent construct was modeled but did not present the indicator-level parameter estimates in the summary provided.
When the team modeled joint exposure using quantile-based g-computation, the combined PAH and psychosocial stress mixture produced larger associations with behavioral outcomes than either domain alone during preschool. The joint mixture ψ estimate for externalizing problems in the preschool window exceeded the single-domain estimates, and internalizing problems also showed an elevated joint ψ. The authors noted that psychosocial factors contributed most to the joint effect estimate. Models were fit separately by developmental window, and the amplification of effect in the joint model was specific to the preschool period according to the reported results.
The trial authors conducted sex-stratified analyses and observed that preschool associations for the joint mixture were present among boys but not girls. For the joint mixture and externalizing problems in preschool, the ψ estimate was larger in boys than in girls; the interaction p-value for sex was reported by the study team and indicated statistical evidence of heterogeneity (the abstract provides the interaction p-value). The source reported no evidence that maternal psychosocial stress modified the chemical–behavior associations when formally tested.
Researchers quantified 25 urinary chemical metabolites representing nine classes (organophosphates, pyrethroids, neonicotinoid, chlorinated phenol, PAH, parabens, bisphenols, UV filters, other biocides) in second-trimester maternal urine. Prenatal psychosocial stress was captured through a latent five-indicator construct built with confirmatory factor analysis; specific indicator measures were not enumerated in the abstract. Child outcomes—internalizing and externalizing problems—were assessed repeatedly with the Child Behavior Checklist (CBCL) at 2, 3.5, 5, 6.5, 8, and 10 years. For single-exposure tests, the study team fit adjusted generalized estimating equations across repeated measures and applied false discovery rate correction for multiple comparisons. For joint-domain analyses, they applied quantile-based g-computation (QG-COMP) to assess combined effects of positively associated exposure classes. Models were stratified into two developmental windows: preschool (2–5 years) and school-age (6.5–10 years). The abstract reports point estimates, confidence intervals, and where applicable interaction p-values, but it does not list the full covariate set or model coefficients in this summary.
The authors disclosed competing interest information for one investigator and declared that other authors had no competing financial interests. Ethical approvals were obtained from relevant institutional review boards in South Africa; participant consent procedures were followed. The source indicated data-sharing commitments—the Drakenstein Child Health Study will provide de-identified data to qualified researchers on request and describes a process for collaboration on its website. The abstract did not present certain methodological specifics in full, including precise urine metabolite concentrations, detailed composition of the five psychosocial indicators in the latent construct, or a complete listing of model covariates in the text provided here. Those items would require consulting the full manuscript or supplementary materials for complete detail.
The study’s results describe distinct developmental courses for chemical versus psychosocial prenatal exposures: urinary PAH metabolites had associations restricted to early childhood, while prenatal psychosocial stress showed associations that persisted into school age. Together, PAH and psychosocial exposures produced larger joint effects on child behavioral outcomes than either exposure domain alone, with stronger joint associations observed in preschool boys compared with girls. The manuscript lists funding sources and ethical oversight, and commits to data sharing for qualified researchers; additional analytic and exposure details are available in the full paper and supplementary files referenced by the authors.