Skip navigation

Calls to Action

San Francisco CADDRE Study Suggests Hazardous Air Pollutants may be Associated with Autism Risk

By Erin Lopes, M.P.H. and CAN Staff
October 04, 2007

The subject of environmental factors, mercury in particular, has been an important and concerning issue for parents of children with autism. What the discussion over mercury has made clear is that there is a great need for studies addressing potential environmental factors that may contribute

to the onset of autism. While several epidemiological studies have attempted to address the role of thimerosal, the mercury preservative contained in some vaccines, much less focus has been put on the contribution of environmental mercury and other contaminants in the development of autism. This issue prompted scientists in California to investigate a potential link between autism and exposure to air pollution, for which existing data were already available. The team led by researcher Gayle Windham, Ph.D. from the California Department of Health Services (CDHS), Environmental Investigations Branch, obtained hazardous air pollutants (HAPs) emission data from the Environmental Protection Agency (EPA), as well as birth records and census tract data, to examine the association between HAPs levels and autism. This study, published this month in Environmental Health Perspectives, found that a greater relative risk of autism was associated with being born in areas with higher levels of some hazardous air pollutants.

Dr. Windham and colleagues conducted their study using data they had collected through the CDC network of Centers for Autism and Developmental Disabilities Research and Epidemiology (
), with partial funding from ATSDR (Agency for Toxic Substance and Disease Research). CADDRE has six centers located throughout the United States including the San Francisco Bay Area where Dr. Windham and colleagues performed their study. The original objective of the CADDRE network was to collect data on the occurrence of autism and to design studies to examine possible risk factors for autism. CADDRE investigators are starting a new case-control study and having multiple CADDRE sites offers researchers the opportunity to generate a study population that is representative of the U.S. population. To do this, CADDRE researchers carefully standardize the study protocols so that ultimately data about risk factors for autism at one CADDRE site can be compared to data at the other CADDRE sites in the U.S.

HAPs are chemicals found in outdoor air that have been identified as potential carcinogens, neurotoxins or endocrine disruptors. The sources of HAPs in outdoor air include diesel fuel exhaust, other vehicle exhaust, common activities like pumping gas or dry cleaning, and industrial emissions like those from power plants. HAPs contain a variety of potentially harmful compounds including metals like mercury and cadmium, aromatic solvents and chlorinated solvents such as vinyl chloride. The Clean Air Act passed in 1990 requires the EPA to track these compounds so that the information can be used to examine the potential health risks of living in areas where HAPs concentrations are high. These pollutants are not routinely measured in air, so the EPA estimates the concentrations based on a complicated model that includes a variety of data about reported and expected use of these chemicals.

Dr. Windham and colleagues paired the EPA's 1996 HAPs data for the San Francisco Bay Area with birth records for those areas from 1994. The researchers then identified autism cases in these regions through the California Department Developmental Services (DDS) records and the Kaiser Permanente Health System. The identified autism cases were linked to the birth certificate in order to identify children with autism born to mothers in these regions. The researchers identified birth addresses of 284 autism cases and 657 controls in the study area and linked the address at birth to the HAPs dataset obtained from the EPA.

After examining the datasets, Dr. Windham and colleagues found that children with an autism diagnosis were 1.5 times or 50% more likely to have a birth residence in an area with higher levels of some HAPs. Metals like mercury, cadmium and nickel were among the HAPs most associated with autism. However, because of the methods used by the EPA to obtain its HAPs estimates, the researchers were not able to separate the effects of each metal associated with autism. The researchers also found weaker associations between autism and the aromatic solvents trichloroethylene and vinyl chloride and diesel exhaust particulates.

Dr. Windham cautions that the results of this study do not indicate a causal link between the HAPs metals and autism and the data should be regarded as preliminary. However, the study does indicate that specific chemicals found in HAPs, like metals, merit further study, in particular, how exposure to them through ambient air may effect early development. Dr. Windham also noted that while their study examined estimated HAPs concentrations at the time of birth, they lacked additional background information about both cases and controls. This includes pertinent data about the mothers of children in the study, such as other exposures during pregnancy, including the mother's smoking status, her diet, or potentially toxic exposures in her workplace. In this study the HAPs estimates obtained by the EPA served only as surrogates for a mother's personal exposures during pregnancy. Furthermore, there was no information about postnatal environmental exposures, as the data studied only allowed a correlation between autism and the place of residence at the time of birth, but did not track how long the child lived in each HAPs area after birth.

According to Dr. Windham, the study is being replicated by researchers from other CADDRE sites. CADDRE sites in Baltimore, North Carolina and New Jersey will use similar datasets to examine associations between HAPs chemicals and autism cases in their regions. If those CADDRE sites see similar association between particular ambient pollutants and autism, this can confirm the findings of the San Francisco study and demonstrate that the association of ambient pollutants to autism is not just particular to this single study or due to chance findings. In doing so, the other CADDRE researchers aim to provide more information on the possible association between ambient pollution and a diagnosis with autism. Such confirmation would then merit a more extensive and fully controlled study of the relationship between autism and pollutants, that would include more individual exposures and activities.

While Dr. Windham's study cannot be used to show a conclusive causal linkage between autism and environmental pollutants, such as heavy metals, it is an important part of the larger effort to develop a comprehensive, nation-wide investigation of the role environmental pollutants may play in the development of disorders such as autism. The findings are suggestive of a connection and reveal new areas of investigation. CAN recognizes the importance of studying how environmental factors may contribute to autism, and continues to target research in this area through its Environmental Initiative. For a listing of some of the environmentally-related projects funded by CAN,
click here


Windham G, Zhang L, Gunier R, Croen L, & Grether J. Autism Spectrum Disorders in Relation to Distribution of Hazardous Air Pollutants in the San Francisco Bay Area. Environmental Health Perspectives. 2006 114(9): 1438-1444.