What Could be Driving Phenotypic Heterogeneity? Deep Characterization of Young Children with and without Autism Spectrum Disorder from the Study to Explore Early Development (SEED)

Background: A number of studies have evaluated phenotypic profiles in samples of children with and without Autism Spectrum Disorder (ASD). Teasing out the influence of developmental function versus symptom severity on assessment measures has implications for screening and treatment recommendations. This study examines the degree that developmental differences drive phenotypic differences on social and behavioral functioning.

Objectives: To examine how developmental level differentiates performance on social, adaptive, and behavioral measures and assess the impact developmental functioning has on heterogeneity among children with ASD on symptom severity, social, and behavioral scales.

Methods: SEED is a multi-site-case-control study of children (2 to 5 years) born 2003-2006, designed to characterize ASD phenotypes and examine genetic and environmental risk factors. Children with possible ASD and developmental delay (DD) were recruited from multiple sources serving children with ASD and developmental problems. Children from the population comparison group (POP) were recruited from birth certificate records.  Children of families who expressed interest in the study were telephone screened for ASD using the Social Communication Questionnaire (SCQ) to determine initial group cohort. Children without a prior ASD diagnosis and low risk on the SCQ received a developmental assessment only (Mullen Scales of Early Learning; MSEL). Children with prior ASD diagnosis or high risk screen were evaluated with the Autism Diagnostic Interview-Revised (ADI-R), Autism Diagnostic Observation Schedule (ADOS), MSEL, Vineland Adaptive Behavior Scales – 2^nd Edition (VABS), and Child Behavior Checklist (CBCL). Children were classified with ASD in SEED based on ADI-R and ADOS results. All children were classified into three groups based on developmental function measured by the MSEL: Average/Above Average (A/AA), DD (one delay noted on MSEL domains), or Global Developmental Delay (GDD; delays on all MSEL domains). Children diagnosed with ASD were further stratified into: ASD-A/AA, ASD-DD, or ASD-GDD.

Results: 3,899 children were enrolled in SEED and 2,676 (mean age= 4.95 years; SD= 0.57), completed the clinic visit and were classified as ASD or non-ASD, and were included in these analyses. Based on developmental level, non-ASD children were characterized as:  A/AA (N=1,333; 56% male), DD (N=459; 67% male), and GDD (N=173; 62% male). Children in the ASD group were stratified as: ASD-A/AA (N=117; 82% male), ASD-DD (N=240; 85% male), and ASD-GDD (N=354; 79% male). Tables 1, 2 depict significant correlations resulting from Spearman correlations comparing SCQ, VABS, and ADI-R Domain Scores to the CBCL Internalizing (INT) and Externalizing (EXT) scores by developmental level and diagnostic group. 

Conclusions: Children in SEED represent a heterogeneous group with various levels of developmental functioning. These correlations demonstrate a pattern suggesting more autism symptoms measured by the SCQ and ADI-R were associated with more behavior problems on the CBCL, for certain groups. In contrast, better adaptive functioning was correlated with fewer behavior problems, for certain groups. Analyses are ongoing and future results will be presented regarding the influence of developmental functioning by developmental and diagnostic group on measures of socialization and behavior. The ASD group will be stratified based on symptom severity to explore the interplay between developmental function and symptom severity.