No Sex Differences Found in Autism Phenotype as Measured By the ADOS in Young Children with ASD

Background: The ratio of boys to girls diagnosed with Autism Spectrum Disorder (ASD) is estimated to be 4.3:1.  Although epidemiological studies consistently find male predominance of ASD, few studies have explored sex differences in its clinical presentation.

Objectives: To evaluate sex differences in measures of the autism phenotype in a large sample of children with ASD.

Methods: Data was collected from 18 Autism Treatment Network (ATN) sites between the years 2007 and 2013. The Autism Diagnostic Observation Schedule (ADOS) Module 1 (preverbal/single words) was examined for 2,585 subjects included in the registry.

Results: There were 2,585 children in the sample (2,162 boys [83.6%]; 423 girls [16.4%]).  There was no statistically significant difference in the gender repartition across ATN sites. The mean age at assessment was 4.28 years (SD=2.4) with no difference between boys and girls. No difference was found with respect to race (Caucasian, Black, Asian, “Other”), ethnicity (Hispanic, Non-Hispanic), or parent education (p>0.10). Mean algorithm total scores did not differ between boys and girls for Social Affect (14.58 vs. 14.67; p>.50), Restricted and Repetitive Behaviors (4.41 vs. 4.36; p>.50) or Total scores (18.99 vs. 19.03; p>.50). Similarly, no differences between boys and girls emerged for calibrated severity scores for Social Affect (7.12 vs. 7.10; p>.50), Repetitive Restricted Behaviors (7.62 vs. 7.56; p>.50) and Total (7.27 vs. 7.26; p>.50). Further analyses of 29 ADOS Module 1 individual item scores showed no differences for 27 items. Two non-algorithm items (Imagination/Creativity and Overactivity) showed slightly lower scores for girls than boys (p=.03), but the magnitude of differences was small and could have arisen due only to chance.

Further comparisons of the sample were performed by stratified age bands (< 3 years [N=877; 17.4% girls]; 3 to < 4 years [N=735; 16.3% girls]; 4 to <5 years [N=415; 16.6% girls]; and 5 years or more [N=558; 14.5% girls]) for the 3 algorithm scores and the 3 calibrated severity scores. Of these 24 additional comparisons, only one (girls in the 3 to < 4 years group) had marginally higher Social Affect algorithm scores (14.97 vs. 14.14; p=.04), a finding that could reflect chance alone.

Conclusions: Core features of the autism phenotype as measured by the ADOS Module 1 were strikingly consistent between boys and girls in this large sample of children. It is generally suggested in the literature that typically developing preschool-aged girls demonstrate more advanced social skills (e.g., social competence and social cognition) and communication skills (e.g., verbal abilities and gestures) than their male peers. This developmental difference was not found in the study sample, implying that young girls with ASD may be more impaired as compared to their same-sex typical peers than boys with ASD.  Follow-up analyses will assess whether this pattern is also found for children with higher language levels administered ADOS modules 2 and 3.