[Autism: overestimation of the genetic origins]

Since the 1990s the genetic causes of autism have been brought to the fore. Despite scientific efforts and huge fundings, less than 20% of cases of autism have been linked to genetic abnormalities. These results do not slow the spread of discourse on the genetic origin of autism, and the concomitant neglect of risks linked to perinatal factors. A detailed analysis of results and assertions in favour of a strong genetic origin of autism reveals methodological biases, misinterpretations and erroneous approximations, as well as an exaggerated media coverage. Studies on twins are illustrative of these biases. The recent demonstration of an excess of twins among sibling pairs with autism, and especially a 10-fold increase for monozygotic twins compared with the general population frequencies, has questioned the relevance of conclusions from earlier twin studies. Indeed, if being a twin is a risk factor for autism, then there may be an upwards bias in estimates of the genetic contribution to autism, and the intrauterine environment, including competition for nutrients, has been neglected.