Relationship of Neonatal Head Ultrasound Abnormalities to Adult Autism Spectrum Disorders in a Low Birthweight Population

Background: Preterm low birthweight (LBW) infants are at excess risk for both perinatal brain injury, as detected by neonatal head ultrasound (HUS), and for later autism spectrum disorder (ASD) diagnosis.  Although there is a well-established relationship between the presence of parenchymal lesions and/or ventricular enlargement (which are thought to reflect white matter damage)  on neonatal head ultrasounds to major motor and cognitive disability, the relationship between neonatal HUS abnormalities and ASD in preterm LBW survivors has not been well studied.

Objectives: To examine the relation of two groups of neonatal HUS abnormalities, namely parenchymal lesions and/or ventricular enlargement (PL/VE) and germinal matrix and/or intraventricular hemorrhage (GM/IVH), to adult ASD in a preterm LBW birth cohort.

Methods: A regional tricounty cohort of 1,105 infants < 2 kg born 1984-87 (the Central NJ Neonatal Brain Hemorrhage study) was systematically screened with neonatal HUS for perinatal brain injury.  ASD was assessed in two stages:screening at age 16 and diagnostic evaluation at age 21. 60% (70/117) of screen positives at age 16 years and a systematically chosen sample of 119 screen negatives were evaluated diagnostically for ASD (Total N=189) at age 21 years. 14 cases of ASD were identified. The relation of neonatal HUS abnormalities to ASD was examined with logistic regression without and with control for other pre- and perinatal risk factors and for cognitive ability and motor problems assessed at age 16.

Results: PL/VE, as detected by neonatal HUS, increased risk for later ASD (OR=7.36; 95%CI: 2.04, 26.51).  This relation withstood control for male gender and maternal hypertension during pregnancy, the only two of 32 other risk factors examined that were found to have marginal (p<.10) or significant (p < .05) relations to ASD.  GM/IVH did not increase risk for ASD.  The relation of PL/VE to ASD at age 21 also withstood control for cognitive but not motor problems. Among the 14 cases of ASD, five were in the PL/VE group; of these, the majority (N=4) had VE and only one had PL.  VE, by itself, increased risk for later ASD significantly (unadjusted OR= 8.3; 95% CI: 2.1, 32.5). 

Conclusions: In preterm LBW infants, neonatal HUS-detected PL/VE increased risk for ASD independently of cognitive ability but not motor problems.  The possibility that neonatal VE might be an important risk factor for later ASD deserves further exploration.