Skeptoid PodcastSkeptoid on Facebook   Skeptoid on Twitter   Skeptoid on Stitcher   Skeptoid on Spotify   iTunes   Google Play

Members Portal

Support Us Store


Get a Free Book



Pregnancy, Pesticides, and Autism

by Stephen Propatier

June 27, 2014

Share Tweet Reddit

Donate News media outlets love a good story. Exciting narratives draw eyeballs and sell publications. Unfortunately, people rely on news media outlets for their health and science information. Experience has shown that news media often either get the information wrong, or place far too much emphasis on isolated research, promoting individual studies as the "new knowledge" about a given subject. This week Internet media sites, and even major news outlets such as Time magazine, ran with headlines like "Pesticide Exposure During Pregnancy Linked to Autism," and "Autism Linked to Popular Backyard Product: Check Your Cabinets." Compelling headlines, certainly, but are they just another failure in science reporting?

"A new study shows..."

It's easy to understand why news media would run this story: such assertive titles grab
attention. Realistically this research is one small data point in a sea of research about autism. That, and other context, is often left out of stories like this one.

Here's how these headlines came about: news media are responding to advance publication of "Neurodevelopmental Disorders and Prenatal Residential Proximity to Agricultural Pesticides," by UC Davis environmental epidemiologist Dr. Irva Hertz-Picciotto. The paper is being published in the journal Environmental Health Perspectives. Looking over the methodology, I would like to say that this study is probably data mining. It examines data from the Childhood Autism Risks from Genetics and Environment (CHARGE) study. Dr. Hertz-Picciotto was the primary researcher for an epidemiological study based on regressive analysis of CHARGE data.

The research had so many issues it's hard to highlight just a few. There was no blinding for researcher bias. It used a questionnaire format to determine autism rates based on children currently diagnosed with the disorder. It did not control in any way for any significant co-morbidities. There was no data or bloodwork for intrapartum pesticide levels, or even any good theoretical basis for pesticide level exposure other than geographic location. It merely points out a correlation between diagnosed cases of autism and people who live in a location where they might be exposed to pesticides. Autism may be a completely independent variable from this. There may be another more significant variable here that we're missing by focusing on pesticides. The study was just not well done, unless you were working backward from the assumption that intrapartum pesticide exposure causes autism and the goal was to find that researchers had made an assumption—a study in tautology.

From this analysis Dr. Hertz-Picciotto is making completely unjustified conclusions. Take this comment from an article by CBS News:

"Many of these compounds work on neurons. When they work on the insect, they're dealing with the nervous system of the insect and basically incapacitating it," said study author Irva Hertz-Picciotto, an environmental epidemiologist at the MIND Institute at University of California, Davis.

In adults, the brain is protected from many chemical exposures thanks to special filters that prevent many substances from crossing from the blood into the brain.

Hertz-Picciotto says that in young children, this blood-brain barrier isn't fully formed, which may allow pesticides to reach vulnerable nerve cells just as they are making vital connections to each other.

There is nothing implausible about the above statement, but there is nothing in Dr. Hertz-Picciotto's research that supports it. Her paper shows a correlation between people living close to areas where there is agricultural spraying and autism rates. Science 101: correlation is not causation. Her claims are frankly completely unsupported by the research; there are just too many variables involved in the data sets to propose a mechanism between pesticide exposure during pregnancy and autism. Although the claim is hypothetically plausible, plausibility doesn't make it likely or probable.

If we take this study and place the data in the most favorable light (i.e. a heterogeneous population, identical workplace and home life exposures, identical diets, perfect gradient exposure to pesticides, uniform pregnancy and birthing methods, uniform parent-age distribution, no travel to other areas, identical medications, exercise, BMI, and family history), then we might be able to draw some preliminary conclusions about a correlation. Beyond that, the data say nothing about the blood-brain barrier and even in a wider study of a population we wouldn't be able to make conclusions about a mechanism involving the blood-brain barrier. It is impossible to tell that from the data.

This is a single study that found a correlational factor. It is not proof of anything. It is barely suggestive of any direct link. Despite the researcher's claim of a possible mechanism it's not really her area of expertise.

Here is a more accurate description of the data: there may be a cluster pattern of autism that may or may not have a link with pesticide exposures during pregnancy. I assure you nobody is going to print that headline.

The study made presumptive conclusions, but the primary investigator's comments to the media were unsupported scary extrapolations. This is extremely common. Exaggerated and unsupported claims by investigators attract and exaggerate media attention, but they're often uncorroborated opinions described as facts. I have noted that researchers seem to be offering misinformation with greater frequency. Reporters may be unwitting dupes of primary researchers who wish to foster support and probably funding for their research. In this case, the researcher is formulating an unsupported conclusion, probably not outright fibbing.

What Hertz-Picciotto provides is a correlational study, with unsupported claims, by a primary researcher who is not a neurobiologist, claiming a neurobiological environmental cause for autism. This of course is immediately picked up by the news media since a) she is a scientist who is b) the primary researcher for c) an autism study titled "Neurodevelopmental Disorders and Prenatal Residential Proximity to Agricultural Pesticides." Pesticides cause autism may make a good soundbite or headline, but it's not good science, is not descriptive of what the research says, and frankly it is a disservice to pregnant women.

The Less Sensational but More Reliable Facts About Autism Studies

The vast majority of good epidemiological studies show a clear link towards a genetic predisposition to autism and a declining evidence stream for an environmental source. This is pretty solid evidence thanks to the anti-vaccine community's focus on an environmental source for autism, e.g. vaccines or pesticides. Several systematic evaluations of environmental sources have been done on autism. The paucity of any evidence supporting those hypotheses has produced growing support in pediatric neurology for a genetic source. Although environmental factors may still play a role, it is unlikely at this point that they will be found to be a major determinant factor in whether or not someone's child develops autism.

Dr. Hertz-Picciotto's conclusive advice sounds simple: avoid all pesticide exposure. Such advice always sounds good in isolation. The problem is that in practice advice like that is so dogmatic and vague that it never ends up being very simple. There may be women whose children are developing autistic symptoms, who may have had a roach motel in their house at one time and now have extreme and unnecessary guilt over the suggestion that they made their child autistic. The worst way to educate anyone is to scare the hell out of them with rigid, broad declarations—even more so if such pronouncements have no clear justification.

Let me be clear: pregnant women should minimize their exposure to pesticides. It's always a good idea to wash vegetables, fresh fruits. And I would not recommend that they operate industrial pesticide sprayers. This seems to be overly obvious; exposure to pesticides can be harmful. But scaring a pregnant woman into believing that having a can of Raid in the house may cause a child to suffer neurological deficits is irresponsible.

The Problem of Problem Studies in the News

I'm constantly dismayed and upset by all the social pressures that are put on pregnant women and mothers. Moms receive a dearth of pseudoscientific claims and weak, conflicting scientific information, both almost universally warning them about the dangers of everything. There is a constant drone of media reports terrifying pregnant women, telling them that they alone define their entire baby's future health and one misstep or inadvertent exposure to cheese, a wheat field, or common medical treatments could subject their child to a lifetime of harm.

Although some may ask what the harm in such claims is, autism is a sensitive and controversial medical subject. Proposing another weakly supported environmental cause is like pouring gasoline on a fire. It furthers ignorance and fosters panic, adding more confusion to the barrage of conflicting opinions. Claiming a link between pesticide exposure during pregnancy and autism is good clickbait and good for the researcher. It is not good for pregnant women. Making unsupported and presumptive claims could panic pregnant women and further damage the public's understanding of autism.

A good, simple story gets plenty of play. Nuanced, cautious, and accurate explanations of scientific evidence get much less. The secondary benefits of sensational, declarative stories are clear: they can attract readers, money, and more research grants. People with specific agendas regularly promote such stories, making statements that far exceed the scope of the research and their scientific expertise, in order to further non-scientific aims. If people truly want what's best for pregnant women and mothers (or anyone else), they should be demanding better, more accurate science reporting, rather than simply more exciting science reporting. Researchers should make clarifying statements about what their studies show and what they don't, that studies are preliminary or inconclusive, that their findings bear scrutiny and further examination since they are by no means definitive. News media describe isolated scientific findings as absolute and crystal clear when the scientific process is typically anything but.

The news media is supposed to be informative, not a bogeyman seeking to scare you. It doesn't help anyone by reporting every extravagant thing declared by someone with the title "Doctor." Just because the primary researcher is a scientist doesn't make her a reliable source on the subject. Just because they make a claim that sounds unambiguous—accompanied by a report with numbers and assertions—doesn't mean their methodology is sound or that their data is useful. Just because they present an easy target (such as pesticides) doesn't mean that it's actually the cause or that there's a simple solution.

So wash your vegetables. Don't put your home on the market if you live in a Southern California agricultural region. Pregnant women and infants shouldn't be handling pesticides. And by all means, avoid working in fields if you're pregnant. Outside of that, don't let anxiety take hold of your pregnancy. You and your family will be better for it. As for what ails the news media, their treatment is much tougher.

This article was co-written by our volunteer copy editor Noah Dillon. I couldn't have written this post with out his creative insight. This is as much his work as mine.

Thank you for all your behind the scenes work Noah.


Autism and pesticides


by Stephen Propatier

Share Tweet Reddit

@Skeptoid Media, a 501(c)(3) nonprofit








Want more great stuff like this?

Let us email you a link to each week's new episode. Cancel at any time: