You are here:
-A A +A

Hazardous Pollutants and Chemicals: Comparing Three Environmental Studies

Date first published: February 21, 2008

We know that some medications taken during pregnancy, such as thalidomide 1 and sodium valproate, 2,3,4  have been linked to cases of autism. Few children are exposed to such drugs in the womb anymore, however. What more widespread environmental exposures are autism researchers investigating? 5,6 How do we look for evidence that would support a link between a specific environmental exposure and autism?

As discussed in the Environmental Aspects introduction to this section, our modern world is full of environmental hazards. Below, we discuss and compare three studies that set out to investigate a potential link between an environmental hazard and autism.

Pesticides in Central California

One of the most striking environmental studies was conducted in Central California--a major agricultural area where many pesticides are applied. Entitled "Maternal Residence Near Agricultural Pesticide Applications and Autism Spectrum Disorders Among Children in the California Central Valley," this study was published in July 2007. 7

These researchers explored data on children with autism spectrum disorders (ASDs) born during 1996 through 1998, their mothers' place of residence during pregnancy, and pesticide exposure. They considered many pesticides, both those that were of concern to local communities and governments and those known to interfere with fetal development, especially of the central nervous system.

When their analysis of 249 possible compounds was complete, they found four organochlorine pesticides, including dicofol and endosulfan, to be associated with a definite increase in ASD risk. If an expectant mother was exposed to these chemicals during the first three months of gestation, she was six times more likely to have a child with an ASD. (To help put this in context, note that DDT, long since banned, is also an organochlorine pesticide.)

The researchers found that this ASD risk increased with poundage of organochlorine applied and decreased with distance from field sites.

Because this study evaluated a large number of compounds, it is especially important to confirm these results through other studies. In general, the more items that are studied all at the same time, the more likely you are to get false positive results for at least one of them. (It's like rolling dice, trying to come up with a certain combination. The more times you roll, the more likely you'll get that combination by random chance alone.)

In addition, it is also possible that some other factor, which also tends to be high in agricultural areas where organochlorine pesticides are applied, was responsible for the increase in ASD risk, or that something about the culture, habits, or socioeconomic situation of people living in farming communities increases such risk. The association between organochlorine pesticides and autism tells us there might be a cause-and-effect relationship, but correlation does not equal causation. Only further research can bring us closer to knowing if there is truly a cause-and-effect relationship.

Pollutants Studied in San Francisco and Texas

Other researchers also have tried to measure environmental hazards and autism rates to see if there might be a connection.

One study conducted in the San Francisco Bay Area examined the potential role of exposure to hazardous air pollutants during pregnancy or early life in the development of ASDs. This study, "Autism Spectrum Disorders in Relation to Distribution of Hazardous Air Pollutants in the San Francisco Bay Area," was published in September 2006. 8

These researchers zeroed in on specific pollutants that might be relevant to autism: metals (such as chromium, lead, and mercury), aromatic solvents, and chlorinated solvents. They then considered children born in the San Francisco Bay Area in 1994. They found that higher levels of certain compounds in an area were associated with a higher risk of ASD. For example, San Francisco County, which had by far the highest average levels for six of the compounds (mercury, cadmium, diesel particulate matter, methylene chloride, toluene, and vinyl chloride), also had a higher proportion of autism cases, while Marin County, which had the lowest levels of the same chemicals, had the lowest proportion of autism cases.

The researchers, keeping in mind that correlation does not equal causation, were careful to note that such patterns might be because of other factors, such as a toxin not measured, or differences in diagnostic practices, or care-seeking behavior in different counties. Still, the study provided evidence of the possibility of an environmental connection. The authors stated:

"Our results suggest a potential association between autism and estimated metal concentrations, and possibly solvents, in ambient air around the birth residence, requiring confirmation and more refined exposure assessment in future studies."9

Another study, "Environmental Mercury Release, Special Education Rates, and Autism Disorder: An Ecological Study of Texas," was published in June 2006. 10  These researchers focused on environmentally released mercury, which, according to the U.S. Environmental Protection Agency, is being released into the atmosphere of the United States in the amount of 48 tons per
year. 11  The researchers explain:

"Mercury is released into the environment largely from fossil fuel (mainly coal) combustion by electrical utilities and from municipal and medical waste incinerators. This inorganic mercury becomes airborne and may be carried for miles before being deposited on soil or water. This inorganic form of mercury is then converted to a toxic form (methylmercury) by chemical reactions or by bacteria." 12

In the Texas study, researchers were looking for an association between environmentally released mercury pollution and both autism and special education rates at the county and school district level. They found that "school district autism and special education rates are significantly associated with environmentally released mercury," 13  that is, as one increased, so did the other, in a consistent way. (In one model, for each 1000 pounds of environmentally released mercury, the rate of autism increased by 61%.) They were careful, however, not to claim that one caused the other. For that, more research is needed. (Again, correlation is not causation. It could be that in a polluted area where mercury is high, so is something else -- another risk factor -- and it is this other factor that is contributing to an increase in autism rates. Finding an association provides a clue, not an answer.)

Evaluating Evidence of Causation

To what extent did each of the environmental studies discussed here provide convincing evidence that a certain environmental factor causes autism?

In the Environmental Aspects introduction we provided guidelines for evaluating such evidence. Keeping these guidelines in mind, and for the purpose of helping non-scientists learn to evaluate research, we compared the three environmental studies. (See Table 1, below.)

Table 1. Evaluating Evidence of Causation: Three Environmental Studies Compared
  Texas San Francisco Central California
Substance Studied Environmental Mercury Hazardous Air Pollutant Concentrations Pesticides
Sample Size
(not listed among the "evidence of causation" guidelines, but one general measure of a study's strength)
Autism rates in 254 counties and 1184 school districts, but no consideration of individual children 284 children with ASD
657 controls
465 children with ASD
6975 controls
Timing
Did cause come before effect?
Not necessarily Yes Yes

Strength of Association
To what degree did the substance studied and autism seem to vary with one another?

Stronger --
For each 1000 pounds of environmentally released mercury, the rate of autism increased by 61%.
Strong --
Some elevated risk in those exposed at the highest levels.
Strongest --
Expectant mothers exposed to organochlorine pesticides during the first three months of gestation were six times more likely to have a child with ASD.

Dose-Response Relationship
Did they show that the longer and more intense the exposure, the higher the risk?

None --
Autism rates in an area were studied, but not where mothers lived while pregnant, or where a child lived as a newborn.
Strong --
Assigned exposure by census tract of birth residence.

Strongest --
Assigned exposure at several stages of gestation by specifically looking at mothers' residence while pregnant.

Showed that risk of ASD went up with pounds of pesticide applied to fields and that risk of ASD went down with distance pregnant mother lived from fields.

Replication of Findings
Have other researchers duplicated these results yet?

No No No

Biologic Plausibility
Is the substance under investigation known to cause harm to the human brain or body in a way that could conceivably lead to autism?

Yes Yes Yes

Consideration of Alternative Explanations for the Observed Association
Did they take into account other factors that might have caused the observed effect?

No --
But they did acknowledge, "A causal association between environmentally released mercury and developmental disorders cannot be determined from this cross-sectional data."
Yes --
They discussed other possible explanations such as different care-seeking behaviors or diagnostic practices in different counties.
Yes --
They discussed the fact that they had very limited data on mothers and children and could not tell, for example, who was employed in agriculture (which could have led to more/different risk). Also studied many other compounds besides organochlorine pesticides.

Cessation of Exposure
Were researchers able to see what happened to autism rates if people stopped being exposed to the substances under study?

No No No

If you consider Table 1 (above), the Texas study, which looked at mercury pollution vs. school special education rates, meets two of the guidelines: Strength of Association and Biologic Plausibility. The study found that for every 1000 pounds of environmentally released mercury, the rate of autism increased by 61%. Therefore, risk of autism was much higher where there was more environmentally released mercury in the area. This is biologically plausible, because mercury is a known neurotoxin.

The Texas study is what is known as an ecologic study. It looked at group characteristics, observing whether one factor increased with another on a mass scale. Studies like this link huge group figures, but they do not look at specific people with specific exposures at specific times. They are usually meant to help point the way for future research.

The San Francisco Bay Area study had a more sophisticated design than the Texas study. Instead of looking at autism rates at the school district level, for example, it used specific children's birth residence to measure level of exposure to a variety of chemicals, finding some elevated risk in those exposed at the highest levels.

The Central California pesticide study considered a large number of pesticides, and found that four -- all of which belonged to the same class -- were linked to a six-fold increase in autism rates. It took into account where women lived while pregnant, as well as how far they lived from fields to which pesticides were applied. They also demonstrated a dose-response relationship, showing that risk of ASD increased as pounds of pesticide applied to fields increased, and decreased with a mother's distance from field sites while pregnant.

Each of the environmental studies discussed here yielded important findings. Each will hopefully lead to future research that attempts replication, that is, confirmation and refinement of the original researchers' results. If results cannot be duplicated, major doubt is cast on preliminary findings. If, on the other hand, many different researchers, looking at the same factor, are able to reproduce one another's results, we can begin to believe we may have identified an environmental culprit in autism.

Please rate the helpfulness of this article: 
No votes yet
References: 
  1. Newschaffer, C.J., Croen, L.A., Daniels, J., Giarelli, E., Grether, J.K., Levy, S.E., et al. (2007). The epidemiology of autism spectrum disorders. Annual Reviews of Public Health, 28, 235-258.  View Abstract
  2. Moore, S.J., Turnpenny, P., Quinn, A., Glover, S., Lloyd, D.J., Montgomery, T., & Dean, J.D. (2000). A clinical study of 57 children with fetal anticonvulsant syndromes. Journal of Medical Genetics, 37(7), 489-97.  View Abstract
  3. Williams, G., King, J., Cunningham, M., Stephan, M., Kerr, B., & Hersh, J.H. (2001). Fetal valproate syndrome and autism: Additional evidence of an association. Developmental Medicine & Child Neurology, 43, 202-206.  View Abstract
  4. Rasalam, A.D., Hailey, H., Williams, J.H., Moore, S.J., Turnpenny, P.D., Lloyd, D.J., & Dean, J.C. (2005). Characteristics of fetal anticonvulsant syndrome associated autistic disorder. Developmental Medicine and Child Neurology, 47(8), 551-555.  View Abstract
  5. Newschaffer, C.J., Croen, L.A., Daniels, J., Giarelli, E., Grether, J.K., Levy, S.E., et al. (2007). The epidemiology of autism spectrum disorders. Annual Reviews of Public Health, 28, 235-258.  View Abstract
  6. UC Davis M.I.N.D. Institute Research. (n.d.). The Autism Phenome Project. Retrieved February 25, 2008. View Report
  7. Roberts, E.M., English, P.B., Grether, J.K., Windham, G.C., Somberg, L., & Wolff, C. (2007). Maternal residence near agricultural pesticide applications and autism spectrum disorders among children in the California Central Valley. Environmental Health Perspectives, 115(10), 1482-1489. View Abstract
  8. Windham, G.C., Zhang, L., Gunier, R., Croen, L.A., & Grether, J.K. (2006). Autism spectrum disorders in relation to distribution of hazardous air pollutants in the San Francisco Bay Area. Environmental Health Perspectives, 114(9), 1438-1444.  View Abstract
  9. Windham, G.C., Zhang, L., Gunier, R., Croen, L.A., & Grether, J.K. (2006). Autism spectrum disorders in relation to distribution of hazardous air pollutants in the San Francisco Bay Area. Environmental Health Perspectives, 114(9), 1438-1444, pg1438 in abstract.  View Abstract
  10. Palmer, R.F., Blanchard, S., Stein, Z., Mandell, D., & Miller, C. (2006). Environmental mercury release, special education rates, and autism disorder: An ecological study of Texas. Health & Place, 12, 203-209. (pg. 207)  View Abstract
  11. Environmental Protection Agency. (2007, April 5). Clean Air Mercury Rule. Retrieved May 30, 2007, from http://www.epa.gov/mercuryrule/basic.htm
  12. Palmer, R.F., Blanchard, S., Stein, Z., Mandell, D., & Miller, C. (2006). Environmental mercury release, special education rates, and autism disorder: An ecological study of Texas. Health & Place, 12, 203-209. (pg 204)  View Abstract
  13. Palmer, R.F., Blanchard, S., Stein, Z., Mandell, D., & Miller, C. (2006). Environmental mercury release, special education rates, and autism disorder: An ecological study of Texas. Health & Place, 12, 203-209. (pg. 207)  View Abstract
ianrandomness