Understanding which early-life exposures drive neurodevelopmental disorders — and whether those exposures act broadly or selectively — has enormous implications for prenatal care priorities and early intervention. A French nationwide cohort of nearly 6.8 million children, tracked for nearly a decade, offers the clearest population-level picture yet of how common biological and social factors map onto five distinct neurodevelopmental conditions.
Drawing on the EPI-MERES register linked to France's national health data system, the study identified 506,505 children — 7.5% of the cohort — with at least one neurodevelopmental disorder (NDD) by a median follow-up of 9.6 years. The disorders examined included communication disorders, specific learning disorders, ADHD, autism spectrum disorder (ASD), and intellectual disability (ID). Co-occurrence was strikingly prevalent: more than half of children with ID or ASD carried at least one additional NDD diagnosis. A cluster of prenatal exposures — male sex, preterm birth, small for gestational age (SGA) status, congenital malformations, maternal obesity, prenatal alcohol and tobacco use, and young maternal age — showed associations spanning multiple disorder categories. However, extreme prematurity emerged as particularly associated with intellectual disability, suggesting that the most severe perinatal insults carry disorder-specific developmental consequences beyond the transdiagnostic signal.
This study's scale makes it methodologically exceptional; few prior investigations have been large enough to detect disorder-specific effects while simultaneously accounting for NDD co-occurrence — a critical confounder that smaller studies routinely overlook. The findings reinforce a growing consensus that many NDDs share a partially overlapping etiological architecture rather than fully distinct causal pathways, which aligns with recent genetic research showing substantial pleiotropy across NDD diagnoses. Practically, the results elevate modifiable prenatal factors — particularly maternal obesity, alcohol exposure, and tobacco — as targets for population-level prevention that could reduce burden across the entire NDD spectrum simultaneously. Key limitations include reliance on administrative health records, which may underdiagnose milder conditions, and the observational design precluding causal inference. Still, for a domain historically fragmented by diagnosis-specific research silos, this transdiagnostic framework represents a meaningful step toward unified prevention thinking.