Abstract
Preterm birth increases the risk of adverse neurodevelopmental outcomes, including autism spectrum disorder (ASD). Separately, polygenic risk scores (PRS) evaluate genetic liability to ASD. While both elevated PRS and preterm birth contribute to ASD risk, the extent to which gestational age interacts with genetic liability is unknown. We analyzed ASD PRS in 2387 individuals from the Center for Applied Genomics database, stratified by both ASD diagnosis (531 with ASD, 1856 without ASD) and gestational age (588 preterm, 1799 term). Term-born children with ASD had significantly higher ASD PRS compared to termborn controls (OR 1.17, 95% CI 1.03–1.32, p = 0.017), replicating prior studies. In contrast, no significant genetic difference was observed within our preterm cohort when stratifying by ASD diagnosis. We further identified a novel signal; preterm children overall had significantly lower ASD PRS compared to termborn children, even after controlling for ASD prevalence (coefficient: − 0.14, 95% CI − 0.21 to − 0.06, p = 0.00095). These findings suggest the predictive utility of ASD PRS may be contingent on gestational context. Lower genetic signal in preterm ASD could reflect stronger environmental stress, increased contributions from rare de novo variants, or undiscovered preterm-specific genetic risk loci.