The scientists speculate that the precipitation itself, which might carry pollutants, or its possible consequences, such as increased TV-watching, decreased vitamin D levels or increased exposure to household chemicals, might trigger autism in genetically susceptible children.
"If you look at the autism literature now, they're much more open to an environmental trigger," says lead author Michael Waldman, a Cornell University economist who says his son was diagnosed with an autism spectrum disorder at age 3 but has recovered and is now a normal third-grader.
The new paper, in the Archives of Pediatric and Adolescent Medicine, published by the American Medical Association, is a "more refined" version of a statistical analysis published online two years ago as a National Bureau of Economic Research "working paper," Waldman says.
The controversial 2006 paper, "Does Television Cause Autism?," found that counties with higher precipitation levels and higher percentages of homes with cable TV also had higher autism rates.
Waldman says he decided to investigate a possible link between precipitation levels and autism rates because of findings from a 2003 U.S. Department of Education survey that collected state autism rates data. States with the lowest rates were New Mexico, Mississippi, Colorado, Oklahoma and Tennessee, while those with the highest rates were the more northern states of Minnesota, Oregon, Indiana, Maine and Massachusetts.
The new study looks at data from Oregon and Washington, in which western counties get nearly four times more precipitation than eastern counties, and California, because it has been the focus of many studies of autism rates and also has relatively high precipitation variability among counties, the authors write.
But Lee Grossman, president of the Autism Society of America, says he's skeptical. "It just does not seem plausible," says Grossman, who got a summary of Waldman's findings from a reporter but hadn't yet read the journal article. "It does not match up with any of the demographics that we follow."
Grossman says his organization has 170 chapters and "what's striking is the similarity in terms of the prevalence and incidence of autism."
It makes does make sense that environmental factors play a role, Grossman says, but no one yet has pinpointed what they are and whether exposure before birth or afterward matters. "I'm really surprised that the AMA is publishing such studies," he says.
In fact, the headline of an accompanying editorial asks: "Do these results warrant publication?" While there is no take-home message for pediatricians or the public at large, writes University of Washington epidemiologist Noel Weiss, Waldman and his co-authors "have indeed reported their results responsibly."
Now, Weiss says, scientists should investigate whether any of the suggested environmental factors play a role.