Are different subtypes of autism spectrum disorders associated with different factors?
- An editorial comment to: van der Ven E, Termorshuizen F, Laan W, Van Os J, Breetvelt E, Selten JP ‘An incidence study of diagnosed autism-spectrum disorders among immigrants to the Netherlands’ 
Van der Ven et al.  in this issue of Acta Psychiatrica Scandinavica distinguished autistic disorders from other pervasive development disorders (PDD) and doing so uncovered an increased risk for narrowly defined autism in children of immigrant parents while the overall risk for autism spectrum disorders (ASD) was similar to that of children born to native-born parents. These results are in line with the findings of two studies conducted in Sweden [2, 3]. Haglund and Källén  compared 157 children with autism and 93 children with Asperger disorder (the few cases of PDD-NOS were excluded from the analysis) to the control population in Malmoe. They found that maternal birth outside Nordic countries was positively associated with autism, but inversely associated with Asperger disorder. Magnusson et al.  included all children up to age 17 living in Stockholm county. ASD cases included 1,649 children with mental retardation (MR) (IQ ≤ 70) and 2,269 children without MR. Compared with children of native Swedish parents, children of immigrants had an increased risk for ASD with MR, but a decreased risk for ASD without MR. Previously, Keen et al.  in the UK had reported more severe forms of autism among children of immigrant parents.
Another interesting finding of the current study  is the association of autism with paternal country of birth. It replicates the association of autism with paternal birth in non-Western countries observed in a Swedish cohort .
Differential diagnosis or misclassification according to parental immigrant status?
It is not possible to exclude a bias owing to differences in referral or diagnosis, the lower risk for Asperger disorder and PDD-NOS being the consequence of underreferral or underdiagnosis of the less severe subtypes of autism because of language or culture barriers. It is noteworthy that in this study  only children of one migrant group, those from Suriname and the Dutch Antilles, had a risk for Asperger disorder and PDD-NOS similar to that of children of native Dutch. This group of immigrants from former Dutch colonies is the closest, by culture and by language, to the native Dutch. Other immigrant parents might be less aware of abnormal behaviour and communication patterns and seek assistance only for children with severe difficulties or disturbances. It is also possible that education or health professionals would not be alerted to subtle symptoms as they would be for children of native-born parents; consequently milder cases of ASD would go unnoticed and undiagnosed.
The possibility that milder cases might be misclassified as more severe in children of immigrant parents is not supported by the finding in this study  of similar ages at diagnosis according to parental immigrant status within each ASD subtype.
Parental immigrant status or ethnic origin?
Most studies on ASD and immigrant status were conducted in developed countries and found that the risk was not different in children of native-born parents vs. those of immigrants from the Western world. This result raises the question: is the variable of interest parental immigrant status or ethnic origin?
Two studies, one in the UK , one in the USA , considered both the generation of immigration and ethnic origin and found contrasting results. In the UK, the risk of ASD was higher for children of Caribbean-born Caribbean mothers than for children of UK-born Caribbean mothers . But among Hispanic children compared with non-Hispanic whites in the USA, the risk for ASD was lower for those with two foreign-born parents and higher for those with two US-born parents . However, these two studies did not differentiate ASD subtypes.
A recent USA study found an overall prevalence of ASD higher in non-Hispanic white than black children . But the analysis distinguishing ASD subtypes showed that black children were more likely to be diagnosed with autistic disorder, but less likely with Asperger disorder. These results are in line with a review of studies conducted in the USA which found a higher risk for the more severe subtypes of ASD (autism strictly defined or associated with MR) in black than in white children, but a lower risk for less severe subtypes (Asperger disorder, PDD-NOS, autism without MR) .
Economic development of parental birth country or skin colour?
This study  found a higher risk for autism in children of parents from both developed (including transition economies) and developing countries. However, it is uncertain whether the variable of interest is economic development or skin pigmentation. Most immigrants come from developing countries, most of these countries are located in lower latitudes, and most of their inhabitants have darkly pigmented skin. Interestingly, the highest estimates for the risk of autism associated with immigrant status were observed for countries where the majority of the population is darkly pigmented: Surinam and Dutch Antilles , sub-Saharan Africa , Africa, Latin America and the Caribbean . These findings underline the importance of maternal skin colour and support the hypothesis recently reviewed by Kočovská et al. , that vitamin D deficiency, either during pregnancy or in early childhood, may be an environmental factor for individuals genetically predisposed to autism.
Future meta-analyses on the risk of autism and other PDD should parallel the strategy adopted by Cantor-Graae and Selten  for estimating the risk for schizophrenia associated with immigrant status: using both the level of development and the predominant skin colour of the country of origin.
Developmental delay or neurodisability?
An association between parental immigrant status and developmental delay independent of ASD might explain the findings of an increased risk in children of immigrant parents for ASD associated with MR . Keen et al.  suggested another interesting possibility, that children of immigrants are overall more at risk for neurodisability: ‘Where this has been explored, it appears that children of immigrant parents are at risk from a greater severity of neurodisability in general, whether physical, intellectual, social, or sensory functioning and this has also been suggested for ASD’. The author of this editorial noticed this higher severity of overall neurodisability in children of dark-skinned immigrants while working as a psychiatrist with handicapped children.
In conclusion, the results of van der Ven et al.  support the suggestion by Haglund and Källén  that different factors are associated with different subtypes of ASD. Today, we have very preliminary results, as this study combined Asperger disorder and PDD-NOS , another excluded PDD-NOS , and the last used a very simple classification of ASD with or without MR . Only one study considered variables other than parental immigrant status and it was probably underpowered . Future studies should distinguish each main subtype and consider not only parental immigrant status but also other variables. It will be interesting to examine the relationship of these subtypes with factors directly related to vitamin D, such as season of birth, latitude, maternal skin colour and rural–urban gradient, as well as other factors, such as maternal and paternal age, pre- and perinatal complications, and maternal stress.