Health Stream Literature Summary- Issue 50 - June 2008

Atrazine exposure and breast cancer incidence: An ecologic study of Missouri counties.
Hunter, L.D., Gadbury, G.L. and Huang, Y.-W. (2008) Toxicological and Environmental Chemistry, 90(2); 367-376.

The triazine herbicide atrazine is extensively used and in the United States, with approximately 3.4 million kg applied to cropland annually. Numerous studies have implicated atrazine as an endocrine disrupting chemical (EDC) and some epidemiological studies have associated atrazine exposure with breast cancer incidence. An ecological study of breast cancer incidence from 1996 to 2002 was conducted in Missouri where atrazine has been detected in treated drinking water, particularly in the northern section of the state.

The main source of exposure to atrazine was considered to be via drinking water in this study. During the study period the number of water tests of the public water supples for each of the 115 counties was highly variable and water suppliers were not static. As there were changes in populations served from public water sources and a lack of data from private water sources, the combined acreage of corn and sorghum in each county was also used as a surrogate measure of probable atrazine exposure. Combined breast cancer incidence rates for 1996-2002 for each county were obtained and reported as age-adjusted to the 2000 standard population for all stages of breast cancer and invasive breast cancer alone. Median income was used as an indicator of socioeconomic status and percentage black race to account for differing racial composition of the counties. High parity in each county was determined using the incidence of women (per 100) giving birth who already had 4 or more previous live births for the combined years 1996-2002.

There were 19 counties classified has having high atrazine exposure and 25 as having low atrazine exposure. The mean concentration of atrazine for all the highly exposed counties was 1.3 micro g /L, which is well below the USEPA maximum contaminant level of 3 micro g /L. Linear regression analyses were conducted using covariates of median income, % black race and incidence of high parity. Models were generated using the rate of all stages of breast cancer (localised and invasive) and for invasive cancer alone. In the final model, high-parity predictor was statistically significant in the model for all stages of breast cancer (p=0.018) but not in the model for invasive cancer (p =0.077). The atrazine indicator was not statistically significant in either regression equation (p =0.167 for all stages and p =0.371 for invasive breast cancer).
In this ecological study the difference in breast cancer rates between counties classified as high in atrazine exposure and those classified as low atrazine exposure was not statistically significant. This does not rule out the possibility that such an association exists, however the authors note that studies of women where atrazine exposure is documented at the individual level have also not shown a significant association with breast cancer risks.