Autism and migration: new clues?

Autism in the Somali population, particularly the migrating Somali population, has been the topic of some debate down the years. I think back to the 2010 study by Barnevik-Olsson and colleagues* detailing an autism prevalence in Somalis living in Sweden of 0.98% compared with 0.21% in children of non-Somali origin alongside other factors related to vitamin D for example. OK, the vitamin D link is still speculative but the migratory aspect remains an interesting association. Enter a new study into the debate from Magnusson and colleagues** detailing some population based findings relating to parental migration status and risk of autism spectrum condition.

The summary:

• Based on an analysis of various records for the total child population of Stockholm County, Sweden (N=589,114) between 2001 and 2007, children with autism defined as high- and low-functioning were identified. I know some people don't like the terms 'high' and 'low' functioning as descriptors of autism, but in this case the authors were referring to cases where intellectual disability was or was not present.
• A total of 4952 cases of autism were identified; 2097 cases comorbid with learning disability (42%) and 2855 cases of autism without (58%). These splits are roughly in line with other data such as that reviewed by O'Brien and Pearson***.  
• When looking at children of migrant parents, there was an increased risk of having a child with autism and learning disability (OR = 1.5) and this risk was particularly pronounced for those migrating from countries with a low human development index and for children whose mothers migrated just before or during pregnancy (OR = 2.3).
• When it came to risk of having a child with autism without intellectual disability, there was actually a decreased risk for migrant mothers compared to children born to indigenous mothers. Whether this was a true reflection of risk or due to the fact that immigrant families might be less likely to seek child development services is unknown.
• These results were not explained by socio-economic, age or obstetric factors.

There are quite a few potentially important implications from this work which require comment.

Over 40% of cases of autism identified in this study were diagnoses of autism and learning disability. One of the blogs I like to follow is Harold Doherty's blog. Some of his more recent discussions have surrounded what will happen to those cases of autism and learning disability in the proposed revisions to the DSM criteria for autism. With the current figures in mind, the question has to be asked whether this kind of relationship will be picked up in any future studies if and when the new guidance is enshrined in diagnostic law.

One also needs to comment on the environmental implications from this latest study. The fact that risk of offspring with autism was elevated just before or during pregnancy implies that genes alone cannot guide this association. I say cannot guide this association but possibly they could if one was to assume that some epigenetic relationship may be evident. I think again back to the prenatal vits and SNPs study on autism from last year. So, assuming that mum is migrating from a warm, sunny environment to a cold, not so sunny (at least not with as strong a sun) environment, where the food may be different, the external environment different (including the 'chemical' environment) and the associated 'stress' that comes with moving to a different part of the world. Mmm.. seems to me that there are plenty of factors ripe for further study there.

Ultimately this is another study of population association and one has to bear in mind the limitations in terms of what information that can bring. Having said that, on top of the other immigrant studies undertaken on autism, there is a growing argument that more focused investigation might yield some interesting clues as to risk and potentially aetiology at least in some cases of autism.

* Barnevik-Olsson M. et al. Prevalence of autism in children of Somali origin living in Stockholm: brief report of an at-risk population. Developmental Medicine & Child Neurology. 2010; 52: 1167-1168

** Magnusson C. et al. Migration and autism-spectrum disorder: population-based study. British Journal of Psychiatry. February 2012.
DOI: 10.1192/bjp.bp.111.095125

*** O'Brien G. & Pearson J. Autism and learning disability. Autism. 2004; 8: 125-140.
DOI: 10.1177/1362361304042718