Risk for cognitive deficit in a population-based sample of U.S. children with autism spectrum disorders: Variation by perinatal health factors

Background: From 30% to 60% of children with an autism spectrum disorder (ASD) have an IQ measure that falls in the intellectual disability (ID) range. It is not well studied whether, for children within this ASD subgroup, there is variation in the risk for low IQ based on a child's perinatal risk factors. Objective/Hypotheses: We assessed whether preterm delivery and term small-for-gestational-age (tSGA) were associated with various measures of cognitive deficit among children with ASDs. Methods: A sample of 1129 singleton children born in 1994 and identified through school and health record review as having an ASD by age 8 years were selected from a U.S. population-based surveillance network. Mean IQ and dichotomous IQ outcomes indicating various levels of ID were examined according to whether a child was preterm (<37 weeks' gestation) or tSGA (term delivery and birth weight <10th percentile for gestational age of a U.S. referent). Results for the total sample and within race-ethnicity/maternal education strata were adjusted for child sex and ASD subtype classification. Results: Mean IQ was significantly (p < .05) lower in children delivered preterm (69.5) than term (74.5) and tSGA (69.3) than term appropriate-for gestational age (75.3). In stratified analyses, the preterm-IQ association was significant only among non-Hispanic white (NHW) children with maternal education at birth of high school or less; adjusted mean IQ was 8 points lower among those delivered preterm (65.4) than term (73.8). Term-SGA was associated with a significant 8-point deficit in adjusted mean IQ (75.5 vs. 83.8) in NHW children with maternal education greater than high school and a 6-point deficit that approached significance (68.4 vs. 74.5, p = 0.10) in NHW children with maternal education of high school or less. Non-Hispanic black children in both maternal education groups had significantly lower mean IQs than NHW children with little variation by preterm or tSGA. Conclusions: In children with ASDs, the risk for concurrent ID or IQ deficit is associated with both preterm delivery and tSGA; these associations may vary by race-ethnicity and SES. Further studies of ASD-ID co-occurrence and the effectiveness of intervention strategies should consider both perinatal and sociodemographic factors.