Study Watch : The case for developmental neurotoxins contributing to a disease of neurodevelopment

Photo by BEN+_+ (flickr)

Earlier this year there was a study published in NeuroToxicology that looked at whether there was a relationship between the distance from Superfund sites and autism rates in Minnesota.

Ockham’s Razor and autism: The case for developmental neurotoxins contributing to a disease of neurodevelopment

This study was authored by Catherine Desoto, PhD who is somewhat famous (or infamous) in certain circles for finding a mistake in a published study about the relationship between mercury and autism. As a result of the mistake, the study in question went from showing no relation between mercury exposure and autism to showing a modest, but not statistically significant, relationship. See here for more details if you are interested.

This study has two major parts. In the first part the author attempts to make the case for a relationship between toxins in the environment and autism by combining recent research into a coherent theory. While this part is interesting and worth reading on its own there is no new data to back up the theories. As a result I am going to skip this part.

The main part of the study then looks for a relation between the rate of autism in a school district and the number of Superfund sites within a certain distance. For each of the 336 school districts in Minnesota that had over 104 students the rate of autism was calculated. Some of the districts with fewer students and higher number of autism cases were averaged in with the surrounding districts to prevent them from skewing the analysis.

The author hypothesized that school districts that had one or more Superfund site within a ten mile radius would have a higher rate of autism than those that did not.

The geographic center of each district was computed and the number of Superfund sites within a 10 mile radius of this center was calculated using specialized software.

There were 46 school districts that had one or more Superfund sides within a ten mile radius. For these districts the mean rate of autism was 1 in 92 students. For the rest of the districts the mean autism rate was 1 in 132. It is worth noting that both figures are higher than the CDC's recent estimate of 1 in 150.

To double check the results the author redid the comparison and looked at school districts that had a Superfund site within 20 miles - the 97 school districts the fell in range had a higher rate of autism as well.

So the data support the idea that there is a correlation between the rate of autism and the number of Superfund sites nearby.

Unfortunately, the correlation in this case does not mean that the Superfund sites cause autism (correlation vs causation) and the author acknowledges as much.

Some other potential problems are that there is no information about where the families lived during the pregnancy or whether the families moved into particular school districts because of the services offered by the district. The first fact would be important because it is thought that prenatal toxic exposures are more harmful and more likely to result in autism while the second could potentially skew the rate of autism in each district (i.e. better services attract more children with autism).

I think there is something to this idea and that there is a relationship between the toxic chemicals in the environment and autism but this study falls short of showing anything more than a superficial relationship.

This study was also written about earlier this year here and here.