Research from NYU reveals early pregnancy metabolic patterns associated with behavioural differences in children, highlighting potential biological markers for developmental outcomes.
Researchers at New York University used repeated urine samples taken during pregnancy to examine whether maternal metabolic patterns were linked with later behavioural differences in children, according to a study in Nature. The work drew on the NYU Children’s Health and Environment Study, a long-running birth cohort that began recruiting pregnant people in 2016 and included families from Manhattan, Brooklyn and Bellevue hospitals. The cohort was designed to reflect the diversity of the communities served by those hospitals, and the authors say comparisons with New York City birth records supported that broader relevance.
The analysis focused on 1,070 mother-child pairs with both metabolomics data and child behaviour information. Urine was collected in early, mid and late pregnancy, then analysed with a targeted metabolomics platform from Biocrates, which the company says is widely used in clinical and epidemiological research. The method measured 188 metabolites and additional metabolite ratios and sums, with results adjusted for urine dilution using creatinine.
Children’s behaviour was assessed with the Child Behavior Checklist for preschool-aged children, a parent-reported questionnaire that captures emotional and behavioural concerns as well as traits associated with autism spectrum disorder. Rather than relying on a single pregnancy average, the researchers used mixed-effects models to account for repeated samples across gestation and negative binomial regression to handle skewed count-like outcome scores. They also examined whether associations differed by sex and ran time-point-specific analyses to look for windows of vulnerability.
The study design also accounted for several maternal and child factors, including pre-pregnancy body mass index, age, parity, ethnicity, diet quality, alcohol use and the child’s age and sex. Because prenatal tobacco use was rare in the sample, it was dropped from the final models, while some metabolic conditions were left out to avoid over-adjustment. For a subset of participants with dietary data, the team carried out sensitivity analyses using the Healthy Eating Index. The authors said the cohort showed no clear evidence of selection bias compared with the wider study population.
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
