Re: Prenatal and infant exposure to ambient pesticides and autism spectrum disorder in children: population based case-control study
This article has many strengths, not least the ability to link outcomes to an administrative data source in a case control study. The authors do, however, appear to have missed opportunities for a more complete analysis.
For example, they only consider whether or not the cases and controls were exposed to agricultural pesticides even though they did have data on exposure doses. Their arguments for not looking at dose response relations are not at all convincing. Also, they find that the strengths of the associations appear to be somewhat greater when the radius around the address was extended from 2 to 2.5Km, which is surprising, but do not examine what the associations were for a smaller radius with presumably greater exposure e.g. 1Km.
Diagnosed autism increased considerably over the study period and it is likely that the portfolio of agricultural pesticides changed too – it would have been useful to model whether the associations of diagnosed autism with pesticides were constant over time.
Finally, it is always disappointing not to be able to examine the relations between the outcomes and confounders, and this is something the journal editors might address.
Competing interests: No competing interests