Gestational Age as a Risk Factor for Neurodevelopmental Disabilities: Exploring Causal Pathways.
BACKGROUND Early and late delivery are associated with increased risk of neurodevelopmental disabilities. Some experts emphasise direct causal effect, while others propose confounding by underlying pathologies. OBJECTIVES We present detailed relative-risk estimates for five major neurodevelopmental disabilities by strata of gestational age. Confounding is explored through sibling-matched analyses and the relative risks of birth defects with gestational age. METHODS Using data from Norway's Medical Birth Registry and other national health registries for 2.9 million births in 1967-2019, we conducted a prospective cohort analysis assessing gestational-age-specific relative risks for cerebral palsy (CP), intellectual impairment, epilepsy, autism spectrum disorder (ASD), and attention deficit/hyperactivity disorder (ADHD). We used an analysis of sibship-matched pairs to assess the role of unmeasured confounding by conditions that persist across successive pregnancies. We also compared these gestational-age-specific risk patterns with the corresponding risks of five birth defects present before delivery. RESULTS Multivariable-adjusted relative risks (RRs) of all five neurodevelopmental disabilities were increased with early pre-term delivery. Compared with births at 40 weeks, infants born at 23-27 weeks had RRs ranging from 2-fold for ADHD (RR 2.2, 95% confidence interval [CI] 2.0, 2.5) to 50-fold for CP (95% CI 44, 58). Relative risks for post-term delivery (≥ 43 weeks) were smaller, ranging from 1.15 to 1.45. In sibship analyses, relative risks with pre-term delivery were attenuated, suggesting the influence of unmeasured confounding. The relative risks of birth defects with preterm and post-term delivery were similar to neurodevelopmental disabilities, showing that confounding alone can produce relative risks of the magnitude seen with neurodevelopmental disabilities. CONCLUSIONS The effect of gestational age at birth on neurodevelopment is likely less-and perhaps much less-than suggested by the observed associations. Prevention of neurodevelopmental disabilities depends not just on interventions during and after delivery, but on identifying and intervening on prenatal causes of these disabilities.