Complex Autism and Clinical Severity In the Simons Simplex Collection (SSC)

Background: Grouping children with Autism Spectrum Disorders (ASDs) into phenotypically distinct categories creates homogeneous subgroups for studying etiological mechanisms. One way to classify ASDs is by dysmorphic features. In the presence of dysmorphology, ASDs are classified as “complex,” while in the absence of dysmorphology, “essential.” The latter is heritable but not necessarily associated with known medical/genetic syndromes. The Autism Dysmorphology Measure (ADM) is a tool that standardizes the evaluation of dysmorphology across individuals with ASD. Using the ADM, a clinician evaluates 12 body areas for physical anomalies, then compares the results to the individual’s neurotypical family members. Anomalies are recorded along a hierarchical algorithm, which leads to the classification of dysmorphic or nondysmorphic (Miles et al., 2008). Prior work has found an inverse relationship between number of anomalies and cognitive functioning scores (Miles et al., 2005). However, it is unclear whether a comparable relationship exists between dysmorphology and autism severity, as measured with the Calibrated Severity Score (CSS; Gotham et al., 2009). Understanding whether such an association exists would clarify whether an individual’s ASD is linked with dysmorphic traits.

Objectives:  To determine whether a relationship exists between dysmorphology and CSS among children with ASD.

Methods:  Participants are children with ASD from the Simons Simplex Collection (SSC). All received diagnoses of ASD via research-reliable administrations of the ADI-R and ADOS, and all data of interest have been collected. Currently there are 1,887 probands in the SSC; however, ADM data have been collected only at four sites, thus the final sample will contain approximately one third of this pool. The sample will be described in terms of sex, age, race/ethnicity, annual household income, and parent education levels. Point-biserial correlations will be used to determine whether associations exist between ADM dysmorphology status and (a) the CSS and (b) verbal, nonverbal, and full-scale IQ scores.

Results:  We hypothesize that dysmorphic probands will have a higher CSS and lower IQ indices than their nondysmorphic peers. It has been shown previously that CSS does not correlate with measures of IQ, but we will examine whether such a relationship exists and control for IQ in the analysis between dysmorphology and CSS, if necessary. Preliminary results from two SSC sites (University of Missouri, Baylor College of Medicine) indicate that between 14—15% of SSC probands are dysmorphic. We anticipate this proportion to hold true across the collection.

Conclusions:  If a relationship is found between dysmorphology and CSS, this could suggest the need for children with more severe ASD symptoms, based on CSS scores, to be directed for clinical-genetic assessment to identify possible dysmorphic features consistent with genetic causes for their ASD. Likewise, children who display dysmorphic features (i.e., are at risk for greater clinical severity) could be referred to intensive intervention programs that could favorably influence their long-term outcomes. Further implications will be discussed, including next steps for use of the ADM if no association is found between dysmorphology and CSS.