Authors:
Brody JG., Aschengrau A., McKelvey W., Rudel RA., Swartz CH., Kennedy T

Title:
Breast cancer risk and historical exposure to pesticides from wide-area applications assessed with GIS

Source:
Environmental Health Perspectives 112(8):889-897, 2004.

Summary:
Concerns regarding breast cancer incidence on Cape Cod have previously been cited. Breast cancer incidence rates have been found to be greater on Cape Cod than in the state as a whole, with risk increasing with length of residence on the Cape. Pesticides are considered to be endocrine active compounds, and associations between exposure to these chemicals and breast cancer have been speculated. Previous research has documented the wide use of pesticides to support tourism, cranberry cultivation and other agriculture on the Cape. This study investigated the association between breast cancer and residential proximity to large-scale pesticide use on Cape Cod.

A population-based case-control study of 1,165 cases of invasive breast cancer and 1,016 controls was conducted. Cases who were diagnosed between 1988 and 1995 on the Massachusetts Cancer Registry and who resided for at least 6 months on Cape Cod during this time were selected. Of 1,578 eligible cases 74% agreed to participate in the study. Controls were matched to cases by ten year age group and vital status. Of 1,503 eligible controls, a 68% participation rate was achieved. Exposures were determined by reconstructing wide-scale historical pesticide applications dating back to 1948 using GIS (geographic information system) technology. Factors to account for pesticide drift and deposition were incorporated into models in order to estimate exposure at each subject's place of residence. Differentiation was made between whether the pesticide was considered to be persistent or non-persistent and the purpose for which the pesticide was used (e.g. cranberry bogs, tree pests, other agriculture or mosquito control). For exposures where the type of pesticide was unknown, applications between 1948 and 1974 were considered to be more persistent and applications after 1975 less persistent. Additional potential risk factors (i.e. family history of breast cancer, menstrual and reproductive history, height, weight, physical activity, education, pharmaceutical hormone use, alcohol and tobacco use) were identified by conducting interviews with both the cases and controls.

No overall pattern of association between pesticide use and breast cancer was found.
More cases than controls reported a family history of breast cancer (as was expected). Cases were also more likely to be over the age of 30 at the time of their first birth. Cases and controls were similar with respect to education, but comparability on other factors such as hormone replacement therapy, oral contraceptive use, diethylstilbestrol exposure, home pesticide use etc. was not reported.

There are several methodological limitations which should be considered in the interpretation of the results of this study. One major limitation in the selection of cases and controls was that a minimum length of residence on Cape Cod of only 6 months was required. This is important because length of residence is an indication of duration of exposure. Amount of time living on the Cape ranged from 6 months to 47 years with a median residence time of 16 and 15 years for cases and controls respectively. Thus the study only accounts for a small proportion of lifetime exposure history for participants with only a short residence time on the Cape. Generally, there is a lack of information about exposures during the years when the women lived away from Cape Cod. Also exposures incorporated into the GIS models were limited to those recorded for specific applications. It is quite probable that other non-recorded exposures occurred. Therefore the actual exposures of each woman may vary quite substantially from the exposures estimated by the models; this increases the likelihood that the study will not find any significant effects. In addition, women with a personal history of breast cancer were not excluded from the study. This is a significant limitation as the etiology for a recurrent breast cancer diagnosis is different than that of a primary diagnosis.

Overall, this study reported no association between breast cancer incidence in Cape Cod and wide-scale pesticide exposure. Although a number of limitations were observed, the application of GIS technology and mathematical models to account for the fate of pesticides in the environment is an important methodological development which may be widely used in the future for studies of this type.