Poor Verbal and Non-Verbal Skills in Children with Abnormal Craniofacial Variability Indices (CVI) and ASD

Background: Physical phenotyping of children with ASD often relies on subjective assessment, as direct measurement of young children is difficult.  Three-dimensional photometry allows objective capture of images quickly and reliably to obtain multiple indirect anthropometric measurements.  Prior studies in older children and adults with ASD show poorer cognitive and verbal abilities in those subjectively rated as having abnormal physical features (Miles, et al. 2005).  This analysis characterizes the verbal and non-verbal functioning of young children with abnormal physical features determined from objective measurements obtained through 3D-photometry.

Objectives: To subgroup children with ASD into those with normal and abnormal Craniofacial Variability Indices (CVI is a composite measure representing facial proportion) and compare groups in terms of verbal and non-verbal abilities using standardized measures.

Methods: Children ages 2-5 years participated in the Autism Phenome Project, a longitudinal study aimed to biomedically and behaviorally phenotype ASD.  Images from 28 Caucasian males were landmarked, and at least 9 facial measurements were compared to published norms to determine z-scores for each measurement.  CVI was calculated as each individual’s standard deviation of all z-scores, with a value >1.2 considered atypical (~95^th percentile in a normal population (Garn, et al. 1985).  CVI therefore provides quantitative description of departure from the norm.  A diagnosis of ASD was confirmed with the ADI-R and the ADOS.  Standard scores within 1 SD on the PPVT-IIIA and EOWPVT defined “average” language level, and "low" language level included scores >1 SD from the mean. Developmental ratios (mental age/chronologic age) were calculated from subtests of the MSEL for verbal and non-verbal abilities. Scores were not normally distributed and sample size was small, so the Wilcoxon Mann-Whitney test was used unless specified.

Results: Children with “low” language levels had statistically significantly higher mean CVI (0.94;SD 0.2  vs. 0.73; SD 0.14) than children with “average” language levels (p=0.04). Two of the 28 boys had abnormal CVI >1.2. Both (100%) boys with abnormal CVI were in the “low” language level compared to 81% (21/26) of children with CVI<1.2 (x²; p=0.5).  Mean verbal ratios were lower for those with abnormal CVI, (0.44;SD 0.25 vs. 0.63;SD 0.13) although this did not reach statistical significance (p=0.3).  Mean non-verbal ratios were lower (0.52;SD 0.19 vs. 0.71;SD 0.08), with a trend toward statistical significance (p=0.07).  ADOS scores were higher for this group (19;SD 1.41 vs. 13.8;SD 4.5), but this was not statistically significant (p=0.16). Conclusions: Our data suggest that children with abnormal CVI perform more poorly on verbal and non-verbal measures and have higher ADOS scores, although this did not reach statistical significance, likely due to the small sample size.  This supports earlier studies that used the presence of atypical physical features as a marker for poorer cognitive and verbal abilities and identifies similar patterns in a younger population based on objective quantitative measures.  Additional subjects are needed to increase statistical power. It is possible that the abnormal CVI may be used as a predictor for poor verbal and non-verbal skills. Such patients may benefit from more thorough genetic evaluation and testing.

Conclusions: