Gender Effects in Autism Spectrum Disorders: Symptom and Intelligence Differences in a Population-Based Twin Study

Background: The fact that males show a much higher frequency than females of both Autism Spectrum Disorders (ASDs) and autistic traits is a well established fact. Some researchers hypothesise this gender difference to be x-chromosome linked whereas others highlight the role of intrauterine testosterone level. It is possible that the current diagnostic criteria are slanted towards a male stereotype of ASD, and are less well-suited to recognising ASD in females.

Objectives: The aim of this study was to explore gender differences in ASD and ASD-like traits in a large population-based sample of twins. Because information on symptom/trait levels and intelligence are available independent of diagnosis, this sample allows investigation or possible biases in recognition as well as patterns of genuine male-female differences.  

Methods: The Twins Early Development Study (TEDS) is a representative population-based UK twin study of all twins born in England and Wales in 1994-6, of which approximately 1% meet diagnostic criteria for autism, Asperger Syndrome or atypical autism. 12659 families provided data at first enrolment and have been contacted regularly since. 189 children of actively participating families met diagnostic criteria for ASD in 2007 according to parental interview with the Development and Wellbeing Assessment (DAWBA; Goodman et al., 2000). Data are available for IQ  at ages, 2, 3, 4 and 7 years (with teacher assessments and national curriculum level at age 7), as well as autistic trait data (as assessed by the Childhood Autism Spectrum Test screening questionnaire, Scott et al., 2002) at age 8 (parental rating) and 12 years(parental and teacher rating). Gender differences in IQ and profiles of core symptoms domains (social / communication impairments and restricted, repetitive behaviours or interests) were analysed for diagnosed children and compared to non-diagnosed children with low or high ratings of ASD-like traits.

Results: Our findings suggested different patterns of gender effects in diagnosed versus high ASD-trait non-diagnosed samples. We found the expected disadvantage for IQ in females (versus males) with diagnosed ASD.  However, in girls not diagnosed but with strikingly high ASD-like trait scores, no such IQ disadvantage was found. ASD-traits were more marked in boys than girls across the general population, but not in diagnosed cases or non-diagnosed cases with very high ASD-like traits.

Conclusions: The pattern confirmed previous findings that in the general population total autistic traits were more apparent in boys than girls, and that IQ is lower among diagnosed girls than diagnosed boys with ASD.  However, the fact that non-diagnosed but high ASD-trait girls show no IQ disadvantage may suggest that probability of diagnosis is affected by IQ-level in girls more than in boys.  In other words, ASDs, and especially more subtle forms of ASD, may be harder for clinicians to recognise in girls, especially in the presence of normal or high IQ.