Skip to main content
         

Vol. 10 Issue 3, Fall 2005

Research Commentary: Challenges to Estimating Past Exposures to Pesticides
The Ribbon 

Suzanne Snedeker, Ph.D.
BCERF Associate Director for Translational Research Sprecher Institute for Comparative Cancer Research

O'Leary, E., Vena, J.E., Freudenheim, J.L. and Brasure, J. (2004) Pesticide exposure and risk of breast cancer: a nested case-control study of residentially stable women living on Long Island, Environmental Research 94: 134-144.

Researchers are using new approaches and methods to tackle one of the most difficult problems in cancer epidemiology; how can past exposures to environmentally relevant chemicals be estimated when evaluating cancer risks with long latencies? Breast cancer poses a particular challenge, since the disease may take 10, 20 or more years to develop. Therefore, to evaluate whether a chemical has a role in a breast cancer case diagnosed today, researchers have the challenge that they must estimate exposures to chemicals that occurred decades ago.

O'Leary and colleagues made use of a wide variety of methods to estimate historical exposures to pesticides in 105 women diagnosed with breast cancer between 1980 and 1992. Controls included 210 age and race matched controls without the disease. The authors hypothesized that exposures to pesticides that may have a role in the development of breast cancer occurred after 1950 but before the 1970s, when many of the organochlorine pesticides were banned.

One of the greatest strengths of the study was the residential stability of the subjects. On average, they lived at the same address on Long Island, New York for at least 18 years. The researchers used geographic information system (GIS) software to determine the distance between where the women lived (residence) and hazardous waste sites containing pesticides. They also used records of pesticides in wells obtained between 1972 and 1992 to estimate exposures [pesticides detected in well water included 2,4-D, delta-BHC (benzene hexachloride), chlordane, dieldrin, heptachlor epoxide, and 1, 2 dichloropropane]. One question researchers asked was whether past agricultural use of the land may be related to the risk of breast cancer. They used aerial photographs taken nearly 50 years ago between 1947 and 1950, and street maps from 1946 to determine past land use. However, information on specific crops planted and types and amount of pesticides was either incomplete or not available. Land was defined as "farmland," "next to farm land," "residential," "wooded," "sand," "greenhouse," "open land," or "may have been farm land."

There are limitations to this approach. Past land use and distances to hazardous wastes sites can only approximate past exposures to environmental chemicals. This approach is a proxy for exposure, but proximity is not a measure of the actual exposure to a person. This can only be determined by biomonitoring (the internal dose to the body determined by measuring levels in blood, urine, or breast milk). But, again, the challenge with a disease like breast cancer is estimating exposures that occurred many years in the past.

Risk (odds ratio) of breast cancer was not related to residence in a water district where pesticides were detected in well water. However, residence within one mile of a hazardous waste site containing organochlorine pesticides carried nearly three-fold higher risk of breast cancer (odds ratio = 2.9, 95% confidence interval = 1.1-7.2). Breast cancer risk was over six times higher in women living on land previously used for agriculture, but this finding was limited to women who had never had children or were over 26 years old when they had their first child (odds ratio= 6.4, 95% confidence interval 2.2-18.2).

The greatest limitations of this study were the very small number of women who lived within a mile of a hazardous waste site (12 cases and 11 controls), and the lack of information on specific types of pesticide exposure. While information was given on the types of chemicals contained in the various waste sites, other information including quantities and proportions of the chemicals, condition of the site, known leaching, etc. was not available. Therefore, testing of the area around the waste sites for residues could help define the extent of potential exposures to specific chemicals.

The authors did find a higher breast cancer risk in women with a late age of first birth or no children, who also lived on land previously used for agriculture. While the authors hypothesized that these women may have had a higher exposure to environmental chemicals that were estrogenic, this is conjecture and was not confirmed by any analysis. Other factors mentioned by the authors, such a lack of protective effects from early pregnancy and lactation, would be plausible reasons as to why these women may have been more susceptible to the chemical's cancer-causing effects.

Other limitations of the study include a lack of data and appropriate control for other factors known to affect breast cancer risk such as age at menarche, age at menopause, benign breast disease, menopausal status, hormonal therapy use, diet including alcohol consumption, and exposure to ionizing radiation. Authors did control for age, race, years of education, and age at first birth.

Despite these limitations, as methods improve for modeling using GIS systems, the approach used by these authors will become more common in epidemiological studies. Studies in future will, as this study has done, not just focus on overall breast cancer risk, but identify sub-populations that may be at added risk because of chemical exposures. Therefore, identification of higher risk groups, such as women with a late age at first birth or no children, may be particularly important in helping to define the factors that affect the risk of breast and other cancer in women exposed to environmental chemicals.

Back to the top

Last Update 08.30.05