Association between prenatal trace element and toxic metal exposures and autism spectrum disorder symptoms in children: Potential mediating role of cord serum metabolites

Prenatal element exposure has been linked to autism spectrum disorder (ASD) symptoms; however, the mediating role of metabolites remains unclear. Based on the Ma'anshan Birth Cohort, 18 elements were quantified in maternal serum during the first, second, and third trimesters, and averaged concentrations of three trimesters represented the whole pregnancy exposure. Untargeted metabolomics characterized the metabolomic profiles in umbilical cord serum. ASD symptoms at ages 3, 5, and 6 years were evaluated by the Chinese version of the Clancy Autism Behavior Scale and analyzed independently across developmental stages. Analytical strategies encompassed exposure-wide associations, quantile-based g-computation, meet-in-the-middle approach, pathway enrichment, and causal mediation model. At age 3, each log10-unit increase in cadmium and magnesium concentrations was associated with 1.29-point (95%CI: 0.09, 2.50) and 4.59-point (95%CI: 0.97, 8.22) increases in ASD symptom scores, respectively, whereas selenium was associated with a 3.41-point decrease (95%CI: -6.00, -0.83). Vanadium and copper were inversely associated with ASD symptom scores at age 5; cadmium was positively associated at age 6; and mercury was inversely associated at ages 3 and 6. Overlapping exposure- and outcome-associated metabolites were identified only at age 5. Alpha-Furyl Methyl Diketone mediated an inverse indirect association between cobalt exposure and ASD symptom scores (indirect effect = -0.24, 95%CI: -0.46, -0.08), whereas 3-Methoxypropylamine showed a positive indirect association between barium exposure and ASD symptom scores (indirect effect = 0.13, 95%CI: 0.03, 0.28). Neither cobalt nor barium exhibited a significant total association; thus, these indirect associations were considered exploratory. These findings indicate element- and age-specific patterns rather than a uniform direction of association, while the indirect associations involving cord serum metabolites require confirmation in independent studies.