Thursday, July 24, 2008

Are the children of first-generation immigrants more likely to be autistic?

I've known about this for about a year, but I never got around to writing about it. I was reminded of it by a recent David Kirby post where he informs us that "an unusually large proportion of Somali-speaking children in Minnesota have autism." I suppose David Kirby is trying to make this about vaccines. The thing is that a link between autism and immigration has been suspected since at least the 1970s, even though it apparently never became an area of research that interested many investigators. Let me just quote from some abstracts, in chronological order.

Harper & Williams (1976): "In a survey on the occurrence of infantile autism in New South Wales it was found that 21-9% of children had at least one foreign-born parent whose native language was not English."

Gillberg et al. (1987): "Urban children with autism more often than age-matched children in the general population had immigrant parents from 'exotic' countries."

Gillberg et al. (1995): "The prevalence for autistic disorder in Göteborg children born to mothers who were born in Uganda was 15% which is almost 200 times higher than in the general population of children."

Gillberg & Gillberg (1996): "Fifteen of these children (27%) were born to parents, at least one of whom had migrated to Sweden."

Bernard-Opitz et al. (2001): "Discussion focuses on possible risk factors and psychosocial adversities for autism such as a high frequency of caregivers who are foreign maids, the use of multiple languages and the high level of punitive educational practices."

Lauritsen et al. (2005): "An increased relative risk of 1.4 was found if the mother was born outside Europe, and in children of parents who were born in different countries."

Maimburg and Vaeth (2006): "The risk of infantile autism was increased for mothers aged >35 years, with foreign citizenship, and mothers who used medicine during pregnancy."

Kolevzon et al. (2007): "The parental characteristics associated with an increased risk of autism and autism spectrum disorders included advanced maternal age, advanced paternal age, and maternal place of birth outside Europe or North America."

There are reports along these lines from Canada as well. In fact, this was discussed in Interverbal's discussion of critiques of Fombonne et al. (2006).

Now, the first thing we need to ask ourselves about these findings is whether they document an actual phenomenon or an artifact. Is there a confound that explains the apparent association?

But if in fact there's an association, what explains it? We don't really know. You will find some unsubstantiated speculation based on old ideas in the cited abstracts, some speculation that I'm sure many readers will find objectionable. It's not surprising to find these types of explanations in old papers, though. If I may engage in some speculation of my own, based on newer ideas, I would say that maternal stress during gestation – see Kinney et al. (2008) – cannot be discounted.

Finally, I'd like to bring attention to Roberts et al. (2007), a study claiming to associate autism with proximity to agricultural pesticide applications in the California Central Valley. The authors stated that they could not dismiss the possibility that the women studied may be disproportionately employed in agriculture. It just so happens that immigrant women also tend to be disproportionately employed in agriculture.