Reports of a French study claimed last week that exposure to certain insecticides during pregnancy were linked to “abnormal behaviour” in children.
Poor behaviour is linked to head lice treatments: Chemicals used to tackle the problem may effect [sic] nerve activity in the brain.
The study’s lead author, Professor Jean-François Viel, told The Conversation he was surprised media reporting focused on head-lice treatment rather than “the overall exposure to pyrethroid insecticides we attempted to address”. You can view his full response at the end of the article.
Published in the Occupational and Environmental Medicine journal, the study suggests an association between exposure to pyrethroids in pregnancy and behavioural issues in six-year-olds. But an association isn’t the same as causation – and as far as associations go, the one in this study was pretty weak.
We asked a chemistry expert to explain, and a toxicologist to review the analysis.
What are pyrethroids?
Pyrethroids are synthetic versions of the naturally occurring insecticide, pyrethrin. They are widely used in agriculture and also in the home to control fleas and head lice.
Pyrethroids are neurotoxins (nerve poisons) that work by making cell membranes more permeable (easy to cross for various molecules). This can impair nerve transmission in insects, as well as humans. Most people are widely exposed to pyrethroids, mainly through diet (as they are used in agriculture) and indoor use.
Pyrethroids enter the body through food, skin and the airways. Once absorbed, they are quickly broken down into byproducts called metabolites. These are excreted through urine and their presence indicates a person has been exposed to pyrethroids.
Earlier work, by the same group of researchers as the study in question, had shown that the presence of pyrethroid metabolites in children, but not in their mothers when pregnant, was negatively associated with children’s verbal comprehension scores.
Several years ago, Canadian research reported that some pyrethroid metabolites – found in 97% of 779 children aged six to 11 years – were associated with parent-reported behavioural problems.
How was this study conducted?
The study had two aims. The first was to test the effect of children’s exposure to pyrethroids in utero; the second tested for exposure to pyrethroids during childhood.
Researchers randomly selected 571 pregnant women from a sample of 3,421 women from an agricultural region of France, who were recruited for a broader study. Of the 571 pregnant women selected, 287 agreed to a neurological (nervous system) and chemical and psychological follow-up when their children were six years old.
Mothers completed a list of 25 questions, drawn from a French version of the International Strengths and Difficulties Questionnaire, to describe their children’s behaviour over the previous six months. Children’s behaviours were also assessed by visiting psychologists.
Researchers tested mothers’ urine at weeks six to 19 of gestation, and their children’s at age 5.99 to 6.27 years. They analysed the urine samples for pyrethroid metabolites.
Pyrethroid metabolites were absent from the urine of 82 mothers and four of their children. So the study only reported results for the remaining 205 women and 283 children. Metabolite concentrations in these women and children were very low – typically in the region of sub-micrograms per litre.
One metabolite (cis-DCCA) was detected in almost all the women and children. Another (3-phenoxy benzoic acid) was detected in some samples, but its concentrations were very low. And in as many as 36% of the childhood urine samples, its levels were not detectable.
Different metabolites arise from different pyrethroids, so the presence or absence of a metabolite is likely to be a consequence of exposure to different pyrethroids rather than different metabolism.
Psychologists assessed children for altruism (social behaviour), internalising disorders (inability to share problems and ask for help) and externalising disorders (defiant and disruptive behaviour). Mothers also sent in reports on their children’s behaviour.
What were the results?
Concentrations of the most commonly observed metabolite (cis-DCCA) in the urine of pregnant mothers in the first trimester was positively correlated with internalising difficulties – such as being anxious or withdrawn – of their six-year-olds.
The variant of cis-DCCA, called trans-DCCA, in the urine of the six-year-olds was associated with reduced externalising behaviours – such as being aggressive or defiant. This is a counter-intuitive finding for which the researchers had no explanation.
Authors also report childhood exposure to the metabolite 3-phenoxy benzoic acid was associated with “increased odds of behavioural disorders”. But this was the metabolite that, in 36% of the urine samples, was below the level of detection.
What are the issues with this study?
There’s a possibility the sample is biased because only about half of the chosen pregnant women agreed to participate. No explanation was offered as to why mothers might have declined the invitation.
Reasons that would affect the study’s validity would include mothers declining due to feeling the research didn’t apply to them as they had never been exposed to pyrethroids, or that participation might lead to revelations of bad parenting.
To their credit, the authors themselves describe the associations they sought in this research as “limited” and advance a number of factors that could have affected them. These included that stored maternal urine samples could have degraded over the six years between their collection and that of the childhood samples.
Second, the authors don’t have a distinct way of knowing whether the behavioural issues seen in some of the children were a direct result of exposure to pyrethroids in utero. Authors used statistical analyses to link childhood behaviour to their potential exposure in utero, but this is not an exact science.
Third, children differ in their ability to metabolise pyrethroids, which makes it difficult to compare like for like. And, finally, there is the possibility of “reverse causality” – children with behavioural problems such as hyperactivity might increase their exposure to pesticides, through deliberately disobeying parents’ orders, for instance.
There is also some evidence that the developing fetus may susceptible to “poisoning” by a particular pesticide for only short windows of time during gestation. If this were true of the pyrethroids and their metabolites, one might not expect to see strong correlations with general measures of exposure like urinary metabolites.
Pyrethroid metabolites are cleared from the body within a few days. This means “snapshots” of their presence at one point in time in children – such as those used in this study – may not serve as good indicators of repeated exposure.
So, what’s the verdict?
Cause and effect is very hard to establish, despite efforts of researchers to control for other factors. And it’s terribly hard to link behaviours back to chemicals. So this study did not establish causation between exposure to pyrethroids and behavioural problems. It only suggested an association.
If parents are concerned about head-lice treatments, they should know there is a non-chemical method of controlling these pests. It uses a fine-tooth comb to separate them from the hair, repeated over a week or so.
As for insect sprays that contain pyrethroids, it is wise to use them sparingly if you are concerned about health impacts. – Ian Rae
Pyrethroid-based pesticides have largely replaced older pesticides, such as DDT, at least in part because they are much less toxic to humans. They are between 1,000 to 30,000 times less toxic to mammals than insects. But they are neurotoxins and there is concern long-term exposure to pyrethroids could affect development of parts of the nervous system.
The observational study published in the Occupational and Environmental Medicine journal sought to address this issue. This Research Check fairly covers the research in question and highlights most of its strengths and limitations.
A strength of the study is that psychologists performed independent tests of child behaviour, rather than self-reports, which are open to bias.
The Research Check correctly points out that what was found is an association between the pyrethroid metabolites and some behaviour, rather than a causal relationship.
The association between the pyrethroid metabolites is weak and does not appear to be dependent on exposure. That is, children with the highest exposure levels do not seem to have more effects on behaviour than those with mid-level exposure.
Importantly, statistical tests performed to link behaviours with exposure were not corrected for the mathematical effects of doing multiple tests. This means the associations may be false positives.
The researchers’ statement that there is “no current explanation for the counter-intuitive association observed between childhood high trans-DCCA concentrations and reduced externalising disorders” is consistent with these results being false positives.
This Research Check’s overall verdict is appropriate. – Ian Musgrave
Full response from study author on media reports
Some headlines refer to “head lice killer” or “chemicals used to kill lice”.
It is true that pyrethroid insecticides are used to treat head lice (but also scabies) in humans and fleas in pets. But they are also used to control pests in residential and agricultural settings.
In our paper, we assessed pyrethroid exposure through concentrations of pyrethroid metabolites in urine (and not through specific usages collected from questionnaires). As a result, we cannot identify and quantify the various sources of exposure. Therefore, this focus on head lice is somewhat surprising to me.
Many pictures show a child with long hair. In the same way, these pictures refer to a single source of exposure (head lice treatment) and not to the overall exposure to pyrethroid insectides that we have attempted to assess.
Professor Jean-François Viel, Department of Epidemiology and Public Health, University of Rennes France