Examining Autism Sex/Gender Differences in Presentation and Mental Health across a Range of Research Designs

Background:

Women and girls on the autism spectrum have often been excluded from research. This has likely led to male-biased information on which theory and clinical practice are based. New approaches are needed to address this inequality in autism science, to insure our diagnostic and intervention efforts are fair to females.

Objectives:

Four approaches to researching autism gender differences will be described, with illustrative data from recent studies. The aim is to examine possible autism gender differences in presentation and mental health, and to highlight strengths and weaknesses of different approaches.

Methods:

Four contrasting approaches will be described, examining sex differences in: a) adults presenting for diagnosis of autism at a specialist clinic (N=90 ASD, 46 non-ASD, m age 40 years); b) twins (m age c.13 years) meeting diagnostic criteria for ASD (N=135), their non-ASD co-twins (N=55) and comparison twins (N=144); c) twins high in autistic traits, identified from an on-going population-based study (N c.8,000, data from age 8 & 12 years); and d) grandparents (N=43, m age 70 years) of offspring with ASD, hence enriched for the broader autism phenotype. Data from diagnostic (e.g. ADI, ADOS) and trait measures of ASD (e.g., CAST, BAPQ), and mental health (e.g. SDQ, HADS) will be considered across the different designs.

Results:

Strengths and weaknesses of different approaches will be contrasted. Although females in general scored lower than males on autistic trait measures, few sex differences emerged across our different samples in mental health or quality of life. In some cases, sex-differences seen in neurotypical samples were not seen in ASD samples. Our actively ascertained and population-based samples did not show the gender differences in IQ expected from previous clinic-based data.

Conclusions:

Research on gender differences in autism currently relies on diagnostic measures largely designed from and for males. Research with diagnosed samples alone may therefore risk circularity. Examination of trait scores in population-based samples may help us move research forward. A major challenge exists, however, if ASD in females presents very differently from ASD in males, given that ASD is currently a purely behavioural diagnosis with no reliable biomarkers.