Gender Differences in Clinical Presentation of Autism Spectrum Disorders

Background: It is well known that many more males present with autism than females, with an overall male to female ratio of around 4.5:1. Furthermore, this ratio becomes more discrepant among those with higher IQ, with reports of gender ratios as high as 5.7:1.  However, whether there are clinical differences in autism presentation between males and females is less well understood. Earlier studies suggested worse autism symptoms in females, but this was thought to be solely related to IQ.  Studies in high functioning autism have suggested worse social functioning and psychopathology in females as compared with males.  It is difficult to draw conclusions across studies due to the limited number of females available and heterogeneity among subjects between studies.

Objectives: The aim of this study was to examine differences in clinical presentation of Autism Spectrum Disorders (ASDs) between males and females across a range of developmental functioning using standardized measures of symptom severity.  

Methods:  Data were obtained from the Autism Genetics Resource Exchange (AGRE), a national dataset of primarily multiplex families with children with ASDs. There were 1446 males and 343 females over the age of 5 that met Autism Diagnostic Interview (ADI) and Autism Diagnostic Observation Schedule (ADOS) criteria for Autism or Autism Spectrum. We separated subjects into 4 groups based on age and verbal ability, similar to the methods proposed by Gotham et al in their 2007 revised ADOS algorithm: 1) those administered ADOS Module 1 who were nonverbal, 2) those administered ADOS Module 1 who had some words, 3) those administered Module 2 (for those with phrase speech), and 4) those administered Modules 3 or 4 (for those who have fluent speech). Within each group, scores on the Stanford-Binet Intelligence Scales, Vineland Adaptive Behavioral Scales (VABS), and parent-rated Social Responsiveness Scales (SRS) were compared between the genders.

Results:  There were no significant age or IQ differences between genders, within each group. There were no gender-associated clinical differences noted among those administered Modules 1 or 2.  Among those administered Modules 3 or 4, females had lower SRS scores (better social functioning; p<0.05) higher VABS scores (better adaptive functioning; p<0.01), and lower ADOS severity scores (p<0.01) than males. Despite an average IQ of 95 in males and females administered Modules 3 or 4, the VABS Adaptive Behavior Composite was significantly impaired (<75) for both males and females. 

Conclusions:  In this dataset of multiplex families with autism, higher functioning females diagnosed with Autism Spectrum Disorders appear to have better adaptive and social functioning than males.  These results suggest that in the general population females may not be diagnosed with autism due to better social and self-care skills.  However the significant discrepancy between IQ and adaptive functioning noted in higher functioning autistic individuals highlights the importance of appropriately identifying ASDs in order to provide appropriate therapy.  These results are currently being replicated in other datasets, with similar findings.