Behavioral Profiles of Premature Infants: Similarities and Differences to Infant Siblings Later Diagnosed with Autism

Background: The Autism Observation Scale for Infants (AOSI; Bryson et al., 2008), a direct rather than parent-report measure, reliably distinguishes autism spectrum disorder (ASD) from typical development at 12 months of age in one risk group: infant siblings of children with ASD (hereafter, ‘high-risk siblings’) compared with low-risk controls (i.e., no family history of ASD). Premature infants are another group at increased risk for ASD, although they are also at risk for other developmental disorders, which may complicate assessments of early ASD signs. Objectives: To examine early behavioral signs of ASD on the AOSI in a group of premature infants relative to groups of high-risk siblings later diagnosed with ASD and low-risk controls. Methods: Behavioral signs of ASD at 12 months of age, as indexed by AOSI total score (where higher values indicate increasing deviation) and total marker count (i.e., items with non-zero scores), were compared among premature infants (n=49; mean gestational age=32.18 weeks; tested at 12 months corrected age), high-risk siblings later diagnosed with ASD (n=53; ASD diagnoses made at 3 years of age based on the ADI-R, ADOS, and expert clinical judgment using DSM-IV, blind to prior study data), and low-risk controls (n=89) using independent-samples Kruskal-Wallis tests. We then examined the frequencies of behavioral signs exhibited on the AOSI by each group. Results: AOSI total scores and total marker counts differed among the three groups (H_{(2, N=191)}=70.35, p<.001 and H_{(2, N=191)}=66.34, p<.001, respectively). Post hoc pairwise comparisons showed that premature infants and high-risk siblings with ASD had significantly higher scores and counts than low-risk controls, although the former two groups did not differ from each other. A significant proportion of premature infants and high-risk siblings with ASD showed a lack of imitation skills (33% and 34%, respectively), variable orienting to name (27% and 40%, respectively), and decreased social behaviors (23-57% and 30-51%, respectively). Only the high-risk siblings later diagnosed with ASD had, in addition, reduced eye contact and response to a change in another person’s facial emotion, as well as more extreme reactivity, difficulty with transitions, and atypical sensory and motor behaviors (23-42% compared to 0-16% of the premature group). Conclusions: At 12 months of age, premature infants exhibit behavioral signs of ASD on the AOSI at similar overall rates to high-risk siblings later diagnosed with ASD, although AOSI scores may be higher in premature infants for reasons other than ASD. The AOSI behaviors these two groups have in common may be considered non-specific early signs, whereas only high-risk siblings later diagnosed with ASD also show specific early signs of ASD. In so far as the two groups share some but not all early ASD signs, it will be important for future research to identify the early behavioral markers that are most predictive of ASD, and to distinguish these from less or non-specific signs. We plan to continue our longitudinal investigation of high-risk siblings, premature infants, and controls in order to appraise the AOSI as an early detection and/or screening technique for use beyond empirical investigations of ASD.