Re: Prenatal and infant exposure to ambient pesticides and autism spectrum disorder in children: population based case-control study
I was frankly dismayed to see the title paper published in the BMJ.
In a textbook example of multiple hypothesis testing , the authors examined the effects of estimated exposure to 11 different pesticides during 3 different developmental periods against two different adverse developmental outcomes. From among the 66 evaluated endpoints, they conclude that prenatal exposure to 6 of these pesticides is associated with 10-20% increases in risk of autism disorder, and that prenatal exposure to a partially overlapping list (3 of 6) is associated with autism disorder with intellectual disability.
The study measures the effect of these exposures by comparing outcomes among exposed children to those among randomly selected unexposed children from the same county. However, each of the counties included in the study contain large urban populations. The test group differs from the control group not only in having documented pesticide exposure, but by all the demographic, lifestyle, and ascertainment differences that distinguish urban populations from agricultural ones.
As one might expect given the study's methodological flaws, it reaches a variety of implausible conclusions.
The authors find similar risks associated with pesticides that vary across a vast terrain of chemical structure, uses (both herbicides and insecticides), mechanism of action (inhibitors of plant-specific amino acid pathways, nicotinic antagonists, anti-cholinesterases, and inhibitors of insect-specific ion channels) and ability to cross the blood-brain barrier (avermectin in particular is excluded from the CNS as a substrate of P-glycoprotein). Indeed, the compounds included in the study have no obvious common characteristics except that they are used agriculturally. But somehow they produce the same developmental adverse effects, with very similar relative risk enhancement.
While not discussed in the text, prenatal exposure to the insecticide imidacloprid was found to REDUCE the risk of autism by 10 to 20%, and to REDUCE the risk of autism with intellectual disability by 10 to 27% whether such exposure occurred during pregnancy or during the first year of life. Perhaps, as was once suggested for statins, we should add imidacloprid to the water supply?
Going forward, we need to look at these issues as scientists. A Romanticist viewpoint of "natural = good" and "unnatural = bad" ignores a century of progress in our understanding of biological systems. I would also urge the BMJ to hire a Chief Statistical Editor and to involve her in the review of similar papers that may be submitted in the future.
John Tucker, PhD
Competing interests: No competing interests