A Comparison of Behavioral Markers Of ASD in a High-Risk Infant Cohort Based on Cognitive Level at 3 Years

Background: Prospective studies of high-risk infants provide opportunities to characterize the earliest signs of ASD using direct observational data. While such data can help inform early detection, possible variation in early markers related to clinical heterogeneity in ASD (e.g., with respect to cognitive functioning) has received relatively little attention.

Objectives: Using longitudinal data from a high-risk cohort to identify behavioral markers in the first year predictive of ASD, and to assess whether these markers vary between children with ASD based on the presence/absence of intellectual delays, as assessed at age 3 years.

Methods: Participants included 277 high-risk infants (HR; younger siblings of children with ASD) and 99 low-risk comparison infants (LR) followed from age 6 months to 3 years. Early markers of ASD were assessed using the Autism Observation Scale for Infants (AOSI) at 6 and 12 months. Outcomes assessed at 3 years included ASD diagnosis determined using the ADI-R, ADOS and DSM-IV-TR blind to prior assessments, and cognitive level, using the Mullen Scales of Early Learning composite score (MSEL-ELC). AOSI item scores were compared in high-risk infants diagnosed with ASD (HR-ASD), high-risk infants not diagnosed with ASD (HR-N), and LR infants using Chi-squared or Fisher exact tests. Further comparisons were completed with the HR-ASD group stratified by cognitive level (IQ <70 vs. ≥70, indexed by the MSEL-ELC). 

Results: Of 277 high-risk infants, 71 were diagnosed with ASD at 3 years, including 19 with IQ < 70 and 52 with IQ ≥ 70. HR-ASD infants as a group, compared to HR-N and LR groups, had a higher frequency at 6 months of reduced motor control, and at 12 months of atypicalities in visual tracking, orienting to name, social babbling, eye contact, social interest and affect, social referencing, response to transitions, general behavioral reactivity, and repetitive motor and sensory-oriented behaviors (all p<.05). A subset of these atypicalities were more common in the HR-ASD group with IQ<70 (vs. IQ ≥70): reduced visual tracking (26.4% vs. 9.4%, respectively; p=.029), and extreme behavioral reactivity (21.0% vs. 3.7%; p=.031). As well, reduced imitation skills, which did not distinguish HR-ASD as a group from HR-N, was more frequent among HR-ASD with IQ<70 compared to those with IQ ≥70 (38.9% vs. 5.6%; p=.002), and distinguished the former from the HR-N and LR groups. For these 3 AOSI items (visual tracking, behavioral reactivity and imitation), children in the HR-ASD group with IQ ≥70 had a similar frequency of atypicalities as the HR-N group. In contrast, IQ was not a significant modifier for other group differences identified in the initial comparison of HR-ASD with HR-N, and LR groups. 

Conclusions: Several behavioral markers observed in the first year in this high-risk cohort were associated with ASD risk and robust to IQ differences, including decreased motor control at 6 months, and atypicalities in social-communication, response to transitions and repetitive behaviors at 12 months. Atypicalities in visual tracking, general behavioral reactivity, and imitation were predictive only for children with ASD and IQ<70, thus highlighting heterogeneity in early signs related to later IQ.