Autism: Genetic or Environmental?

Eastern Virginia Medical School

Dr Harrington is associate professor of pediatrics at Eastern Virginia Medical School and Children’s Hospital of The King’s Daughters in Norfolk, Va. He is also a member of the editorial board of Consultant For Pediatricians.

Guest Commentary

A recent article from Hallmayer and colleagues1 has reinvigorated the debate about whether autism has significant environmental factors affecting its manifestation.1 Through previous twin and sibling studies, experts believed that a broad definition of autism was found 90% of the time in identical twins, and risk of autism in fraternal twins or siblings varied from 3% to 14%.2-5 These studies were based on fairly small sample sizes, case reports, and telephone interviews, which did not allow confirmation of a diagnosis with current standard testing. In addition, studies within the past few years have stated that if autism was so genetically predetermined, we should see more monozygotic twins than dizygotic twins affected, when in fact the opposite is true.6

Hallmayer and colleagues1 identified 1156 total pairs of twins in California with at least 1 twin having the possible diagnosis of autism. This was whittled down to 202 twin pairs who were tested with the Autism Diagnostic Observation Schedule (ADOS) and Autism Diagnostic Interview-Revised (ADIR). Of the 202 twin pairs, 10 did not meet the study criteria (of having at least 1 twin with autism). This left 192 twin pairs (138 dizygotic, 54 monozygotic), which is the largest collection of well described pairs to date. The researchers clarified the diagnosis of autism as “strict” or “broad.” Strict or classic autism meant having all 3 features in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Revised: repetitive and restrictive behaviors, qualitative impairment in use or amount of language, and impairment in social interaction and joint attention. The “broad” definition was more in line with having met criteria for an autism spectrum disorder (Asperger syndrome or pervasive developmental disorder not otherwise specified). When the 192 pairs (384 total subjects) were tested for autism, 171 met “strict” criteria and 71 met “broad” criteria; the remaining 142 were unaffected.

Using the “strict” or “broad” autism phenotype models, there was no difference between male and female twins in heritability or environmental factors that led to autism. However, when the authors used a statistical model of “best fit” (ACE, additive genetic factors, A; common environmental factors, C; and specific environmental factors, E) to extract the amount of liability that genetics or environmental factors play in the development of autism, the results showed a stronger environmental influence. For “strict” autism, genetic heritability was calculated at 37% (95% CI, 8%-84%) and the shared environmental variance component was 55% (95% CI, 9%-81%). For the “broad” phenotype, genetic heritability was estimated to be 38% (95% CI, 14%-67%) and the shared environmental variance component was 58% (95% CI, 30%-80%).

Autism funding has centered on the belief that genetics is the primary influence. The Hallmayer study suggests that these genetic influences may have been overestimated and that environmental factors are probably as strong, if not stronger, in creating susceptibility to autism. Although the ACE model used in the study is one of the basic models of genetic epidemiology, some researchers believe that it may overestimate environmental factors. Twins obviously share the same prenatal environment and, for the most part, postnatal environment. We know that certain prenatal nongenetic factors increase the risk of autism (increased paternal age, low birth weight, and maternal infections during pregnancy).7-9 Perhaps we need to look more closely at postnatal factors related to neurodevelopment. Such factors may include infant or maternal food intake as well as exposure to media and electronic devices. Research should concentrate on these factors rather than on a vaccine-autism link, which has been disproved.10 Regardless, the direction of future autism research over the next decade must now consider a possible environmental component. ■