Antibiotic mixture exposure in meconium affects child neurodevelopment via altering early-life gut microbiome and...

Previous evidence has established associations of antibiotic exposure in early life with neurodevelopmental disorders. However, previous studies have often relied on self-reported or record-based exposure measures that may not accurately reflect internal fetal exposure, and few have integrated prenatal exposure biomarkers with infant microbiome, metabolome, and longitudinal neurodevelopmental data. This study included 300 mother-infant pairs from the Shanghai Maternal-Child Pairs Cohort. Meconium samples were used to quantify concentrations of 18 antibiotics using high-performance liquid chromatography (HPLC). Neurodevelopmental outcomes were assessed using Ages and Stages Questionnaires (ASQ) at 2, 6, 12, and 24 months of age. Multi-omics profiling of 6-month fecal samples was conducted, employing 16S rRNA gene amplicon sequencing and HPLC-based metabolomics. Linear and logistic regression models, negative binomial regression, interpretable machine learning algorithms, metabolome-wide association/pathway analyses and high-dimensional mediation modeling were used. Our study population demonstrated predominant exposure to three antibiotics: chlortetracycline, penicillin, and chloramphenicol. Each interquartile range (IQR) increase in meconium-measured antibiotic burden was associated with a lower total ASQ score at 6 months (β = -4.49, 95% CI: -7.14 to -1.86) and lower scores across all five developmental domains (βs ranging from -0.53 to -1.11; all P < 0.05). Each IQR increment of antibiotic burden was also associated with greater odds of belonging to the lower-score group of total ASQ trajectory (OR = 1.20, 95% CI: 1.04, 1.42) and the lower-score group of personal-social trajectory (OR = 1.13, 95% CI: 1.01, 1.28). Exploratory high-dimensional mediation analyses identified Lactobacillus and 2-(14,15-epoxyeicosatrienoyl) glycerol (2-14,15-EG) as candidate intermediates in the association between meconium-measured antibiotic burden and neurodevelopment. Our study suggested that higher meconium-based biomarkers of intrauterine antibiotic burden were significantly associated with lower neurodevelopmental outcomes in early childhood, with the strongest inverse associations observed at 6 months. Lactobacillus and 2-14,15-EG were identified as potential mediator biomarkers. These hypothesis-generating findings require independent replication, longitudinal temporal validation, and experimental studies before causal mechanisms or microbiota-targeted interventions can be inferred.