Regression analysis of pesticide use and breast cancer incidence in California Latinas.

Introduction

Hispanic (Latina) females experience lower risk of breast cancer than do non-Hispanic white women. In California, the age-adjusted breast cancer incidence rate in 1999 among Latinas was 84.9/100,000, while among non-Hispanic white females it was 147.3/100,000 (Kwong et al., 2000). These lower rates may reflect known risk factors including fertility, parity, use of exogenous hormones, and socioeconomic status.

Currently established risk factors for breast cancer include older age at first full-term pregnancy, early menarche and late menopause, exogenous hormone use, and family history of breast cancer (Kelsey & Bernstein, 1996). These risk factors, however, account for only about 50 to 60 percent of the variation in incidence rates between high- and low-risk populations (Madigan, Ziegler, Benichou, Byrne, & Hoover, 1995; Robbins, Brescianini, & Kelsey, 1997).

California is the leading agricultural state in the nation. It produces various crops and commodities worth more than $26 billion annually. Pesticide use is intensive, and each year more than 120 million pounds of active pesticide ingredients are applied in agricultural settings. This amount represents nearly one quarter of all pesticide use in the United States (California Department of Pesticide Regulation [DPR], 2000).

Agriculture in California is labor intensive and requires extensive use of seasonal and migrant farm workers. Since a large proportion of the workforce in the agricultural industry is Latina, Latinas are at increased risk of exposure to agricultural chemicals. Hispanics are a rapidly growing segment of the population of the United States and California. In the 1990 U.S. Census, 81.5 percent of all females included in the farm worker occupational category in California were classified as Hispanic (U.S. Census Bureau, 1990).

Farmers have been studied in epidemiologic analyses because they have relatively high levels of workplace exposure to several toxic agents, including ultraviolet radiation from the sun, diesel fumes and exhaust, and fertilizers and pesticides. Several forms of cancer have been associated with work in the agricultural environment including skin, lip, prostate, leukemia, brain, and perhaps testicular cancer, as well as non-Hodgkin's lymphoma (Blair & Zahm, 1995). Female farmers have also been evaluated in some studies for elevated rates of breast cancer (Band et al., 2000; Kocic, Jancovic, Petrovic, & Tiodorvic, 1996; Petralia et al., 1998), but not in others (Blair, Dosemeci, & Heineman, 1993; Settimi et al., 1999; Wang, Lewis-Michael, Fitzgerald, & Stark, 2002; Wicklund & Dich, 1994). None of the previous studies concerning pesticide exposures and breast cancer in the United States have been conducted in an intensively agricultural state where exposure to these pesticides has been significant and where many Hispanic women may have work histories that entail a high potential for pesticide exposure.

The study reported here evaluated breast cancer incidence rates by using the California Cancer Registry (CCR), which includes statewide and county- and race/ethnic--specific data from 1988-1999.

The authors used regression analysis to evaluate breast cancer incidence at the county level as a function of organochlorine (OC) and triazine herbicide (TH) pesticide use after controlling for known risk factors for breast cancer including age, fertility, and socioeconomic status. The OCs and THs were selected because of previous suggestions that they are associated with breast cancer and because of their endocrine-disrupting properties. The analysis was restricted to Latinas diagnosed in California during 1988-1999.

Methods

The CCR is a population-based cancer registry created in 1985 that monitors the occurrence of all newly diagnosed cancers and cancer-related deaths in the population of California (Morris et al., 1999). The CCR includes 10 regional registries that are responsible for collecting the data. These data were largely based on cases abstracted and computerized by hospital cancer registrars in the regions. The methodology of the CCR has been previously described (California Cancer Reporting Standards, 1988). The statewide tumor registry commenced operation in 1988. Current reporting is complete and edited through 1999. The registry accessions all cancers except for non-melanoma skin cancers and in-situ cancers of the uterine cervix. Information on several demographic variables (e.g., age, race, sex, residence); diagnostic variables (including stage at diagnosis, tumor size, histology and grade of tumor); and first course of treatment are collected for all cases. For this analysis all invasive-breast-cancer cases (excluding in situ breast cancers) of all histologic types were included as the numerators of the incidence rates. Cases are routinely coded with regard to anatomic stage of disease according to the general-summary stage schema for 1988-1993 (Surveillance, Epidemiology, and End Results Program [SEER], 1986), and SEER extent-of-disease coding instructions for 1994-1997 (SEER, 1992). For the purposes of this report, attention is focused on the time period 1988-1999.

California's Department of Pesticide Regulation has compiled the nation's most comprehensive pesticide use database. For limited-use pesticides, reporting requirements have been in effect in California since 1950. A regulation change in 1970 required farmers to report applications of restricted materials only. Commercial pest control operators were required to report all pesticides used, including date, location, and crop. Many OC and TH pesticides fall into the class of restricted-use materials. Anyone using a restricted-use chemical must file a pesticide use report with the agricultural commissioner in the county where the application is being made. These reports include the following information: pesticide applied, date and location of the application, and the crop and acres treated. The commissioners forward this information to the state, which has created a computerized database (Department of Pesticide Regulation, 1995).

All pesticide information reported from 1971 to the present is summarized annually in two reports. One report is indexed by crop and the other by active ingredient. The reports cover the following information: pounds of pesticide active ingredients, number of applications, and acres treated on a county-by-county basis.

For this analysis, pesticide use data were evaluated for the years 1970-1988 and were summed by county and pesticide for eight OC and two TH pesticides. The characteristics of the chemicals evaluated are presented in Table 1. They include the chemical class (OC or TH pesticide), the current Carcinogenic Classification, the Chemical Abstracts Service (CAS) number, and the total pounds of active ingredient of the chemical applied in California in 1980. After the pounds of pesticide active ingredients at the county level were summed, the results were found to be skewed. Therefore, the results were log-normalized.

Breast cancer incidence rates were used as the dependent variable in negative-binomial (NB) regression models. Separate models were run for two time periods, 1988-1993 and 1994-1999, to allow for different latency periods and to limit the impact of secular trends in breast cancer incidence. Incidence rates were not calculated for counties in which fewer than 10 breast cancers were diagnosed among Latinas in either time period.

In the regression models, pounds of active ingredients of the OC and TH pesticides were entered as categorical values in quartiles that were created after the distributions of the log-normalized values were examined. Adjustments for age, fertility, and a measure of socioeconomic status were made by inclusion of terms for each of these variables simultaneously with the terms for the OC and TH pesticides. Fertility was expressed as number of births per 1,000 women 15-44 years of age. Socioeconomic status was expressed as the percentage of the Hispanic population below the poverty level (Hall & Richards, 1994). Poisson regression models were used initially but were found to have poor fit with the data and to be characterized by over-dispersion. Therefore, NB regression models were used. NB regression is used when there is evidence of over-dispersion in the data (Cameron & Trevidi, 1998; Long, 1997), and the goodness of fit was found to be much better with NB regression than with Poisson. The results are expressed as incidence rate ratios (IRRs) and 95 percent confidence intervals [CIs]. PROC GENMOD in the SAS statistical software (SAS Institute, 1990) was used to fit the models.

Results

Between 1988 and 1999, a total of 23,513 invasive cancers of the breast were diagnosed in Latinas in California. Using the NB regression model, the authors evaluated the relationship of age, fertility, and socioeconomic status to breast cancer incidence in Latinas for the years 1988-1999 (Table 2). Risk of breast cancer increased with age; women 70-74 years of age experienced a 47-fold increased risk of breast cancer compared with women 25-29 years of age. Fertility rates were inversely associated with breast cancer incidence after age was controlled for. Hispanic women in the lowest quartile of fertility experienced a statistically significant 12 percent increase in breast cancer (IRR = 1.12, 95 percent CI = 1.01-1.24). Women in the highest socioeconomic quartile experienced an 18 percent increased risk of breast cancer in comparison with women in the lowest socioeconomic quartile. This result was also statistically significant.

For the time period 1988-1993, most of the IRRs were not significantly different from 1.00 (Table 3). Two OC chemicals had a positive relationship with breast cancer incidence in Latinas. Toxaphene showed elevated IRR in the second through fourth quartiles, in which the IRR and the 95 percent CI were 1.09 (0.95-1.25), 1.17 (1.01-1.35), and 1.16 (1.01-1.34), respectively. The odds ratios in the two highest quartiles achieved statistical significance. For methoxychlor, the IRRs were 1.00, 1.13, (1.02-1.27), 1.07 (.95-1.20), and 1.18 (1.03-1.35), respectively. For the later time period, 1994-1999, the relationship between breast cancer incidence and methoxychlor persisted, with IRRs of 1.00, 1.06, 1.04, and 1.16, and only the odds ratio in the highest quartile was statistically significant. For simazine and atrazine, no elevation in risk was observed in the period 1988-1993 (Table 4). In the second time period, the IRR for simazine was elevated in the second quartile only (IRR = 1.40; CI = 1.14-1.72), and no dose-response pattern was evident.

Discussion

The strengths of the study reported here include a focus on a population with a tradition of agricultural employment, a large sample size, and a 19-year exposure period (1970-1988) that predated the identification of the newly diagnosed breast cancers (1988-1999). This period allows sufficient latency for potential exposure and disease associations to become manifest. Also, the use of the CCR, an established population-based cancer registry into which more than 95 percent of all cancers are annually accessioned, provided a large number of cancer cases for analysis (N = 23,513), distributed across all 58 counties of California. This resource increased the precision of the regression coefficients and allowed for a wide range of potential exposures among Hispanic females in urban and rural areas of California. Limitations of the study include the ecological nature of the exposure assessment, with the authors assuming that county level measures of socioeconomic status, fertility, and pesticide use were relevant to the individual breast cancer patients in the county in which they resided.

The lack of positive relationships between some pesticides and breast cancer incidence rates could be ascribed to the ecological nature of the exposure assessment. The fact, however, that the current analysis revealed associations between established risk factors for breast cancer such as age, socioeconomic status, and fertility levels strengthens the authors' confidence in the findings related to pesticide use, and suggests that the ecologic methodology is sound and that the NB regression approach is capable of detecting alterations in risk. Recent studies in California have demonstrated a strong association between socioeconomic indices and breast cancer risk within the Latina population (Liu, Deapen, & Bernstein, 1998; Yost, Perkins, Cohen, Morris, & Wright, 2001), and these positive associations were evident in the study reported here. In addition, reproductive factors, including nulliparity and later age at first pregnancy, have been associated with elevated breast cancer risk in Latinas in New Mexico (Gilliland et al., 1998). An alternative interpretation is that the control of confounding by socioeconomic status and fertility may have been inadequate.

A previous correlation analysis was conducted in which pesticide use data from the California Department of Pesticide Regulation and the age-adjusted cancer incidence rates in California counties were evaluated (Mills, 1998). In that study, the number of pounds of active ingredient of several pesticides previously suspected to be associated with cancer risk (namely, atrazine, 2,4-D, Captan, Diazinon, Dicofol, and trifluralin) was correlated (on a sex- and race/ethnicity-specific basis) with the age-adjusted incidence rates of several cancers. Although the resulting coefficients were zero or close to zero in most of the sex/ethnicity subgroups, several positive correlations emerged in Hispanic males. These results suggest that exposure of the Hispanic segment of the population to these pesticides may be associated with increased cancer risk at several sites. In that analysis, however, breast cancer was not evaluated because much of the intercounty variation in breast cancer may be due to established risk factors, including fertility rates and socioeconomic status.

No relationship was observed between pounds of use of DDT and breast cancer risk in Latinas in this analysis. Use of DDT was banned in California in 1972, however, and the pesticide use reports relied on here commenced in 1970. Therefore, only a limited period of use was available for analysis. The IRRs also were not elevated for categories of dieldrin in this analysis, although previous studies of serum concentrations of this OC pesticide have demonstrated increased breast cancer risk (Hoyer, Jorgensen, Brock, & Grandjean, 2000; Wolff et al., 2000)

Several epidemiologic studies have evaluated the role of pesticide residue levels as risk factors for breast cancer. In particular, DDT and its main metabolite DDE have been studied in approximately 30 epidemiologic investigations, with inconsistent results. Reviews of this issue indicate that, at low levels of exposure, these pesticides are unlikely to be associated with increased breast cancer risk (Cocco, 2002; Laden & Hunter, 1998; Longnecker, Rogan, & Lucier, 1997). A study completed in Mexico (where DDT is still used as an insecticide), however, found elevated breast cancer risk in postmenopausal women associated with high levels of DDE (Romieu, Hernandez-Avila, Lazcano-Ponce, Weber, & Dewailly, 2000), and another study conducted in Latin America demonstrated a positive association between serum DDE levels and breast cancer (Olaya-Contreras, Rodriguez-Villamil, Posso-Valencia, & Cortez, 1998). An ecological analysis conducted in the United States, however, found a negative correlation between adipose DDE and breast cancer mortality rates (Cocco, Kazerouni, & Zahm, 2000).

The organochlorine pesticides are extremely persistent in the environment, are lipophilic in human tissue, and have been shown to have estrogenic activity (Soto, Chung, & Sonnenschein, 1994). They have also been shown to be carcinogenic in many animal models (IARC Working Group on the Evaluation of Carcinogenic Risks to Humans, 1999, p. 59).

Previous studies of this issue have been hampered by the fact that they were not conducted in agricultural areas and did not evaluate risk in the Hispanic population of the United States, which traditionally has been engaged in agricultural work more than other segments of the population. A second report from the Centers for Disease Control and Prevention's (CDCs) National Center for Environmental Health (NCEH) demonstrated that Mexican-Americans have higher serum levels of DDE (the main metabolite of DDT) than either non-Hispanic blacks or non-Hispanic whites. The geometric mean of serum DDE for Hispanics in that report was 674 nanograms/gram, while in blacks it was 295 ng/g and in whites it was 217ng/g (NCEH, 2003). A previous study of breast cancer in another agricultural state, North Carolina, did not demonstrate elevated breast cancer risk associated with farming; however, risk was elevated in subgroups who reported being in fields shortly after pesticides were applied and among those who did not use protective clothing when applying pesticides (Duell et al., 2000). Although the results of the recent Long Island breast cancer study were widely reported to have been negative, some positive associations were noted. For example, the odds ratio for breast cancer in women with the highest levels of DDE (the most common metabolite of DDT) was 1.20, or 20 percent higher than the risk in women with the lowest levels. Risk of breast cancer was 37 percent higher in women with the highest level of dieldrin, another organochlorine pesticide, although neither of these results was statistically significant (Gammon et al., 2002).

Risk of breast cancer in this study was elevated approximately 16 percent in Latinas in the highest exposure quartile for toxaphene. In the Factors Affecting Rural Men (FARM) study, toxaphene was found to be associated with a three-fold elevation in risk of 1 (14;18)-positive non-Hodgkin's lymphoma (Schroeder et al., 2001). The U.S. Environmental Protection Agency (U.S. EPA) lists toxaphene as a Group B2 carcinogen, which indicates that the chemical is a probable human carcinogen (U.S. EPA, 2002). Use of toxaphene was discontinued in 1989; the chemical has been shown to have estrogenic properties in some human cell culture studies (Bonefield, Autrup, & Hansen, 1997; Soto et al., 1994), but not others (Arcaro, Yang, Vakharia, & Gierthy, 2000).

The study reported here found a statistically significant 16 to 18 percent elevation in risk of breast cancer among Latinas in the highest exposure quartiles for the OC methoxychlor. Previous epidemiologic studies of pesticides in Iowa and Minnesota revealed elevated risks of leukemia (odds ratio [OR] = 2.2) among men exposed to methoyxchlor (Brown et al., 1990). Methoxychlor, a proestrogenic pesticide, is classified as a Toxicity Class IV compound (practically nontoxic), although there is some evidence that high doses of methoxychlor may have estrogenic or reproductive effects (Cummings, 1997). Carcinogenicity testing has been largely negative, although one strain of mice has been found to have increased incidence of testicular tumors (National Library of Medicine, 1995). U.S. EPA currently classifies methoxychlor as a Group D chemical--not classifiable as to human carcinogenicity (U.S. EPA, 2002). Methoxychlor is moderately persistent in soil, with a half-life of approximately 120 days (Wauchope, Buttler, Hornsby, Augustijn-Becker, & Burt, 1992).

The triazine herbicides (atrazine, simazine, and cyanazine) are the most commonly used herbicides in the United States. These chemicals were introduced in the 1950s and are used to control germinating weeds by inhibiting photosynthesis. They are used extensively in corn, citrus, nut, sugarcane, sorghum, and cotton cultivation. Atrazine is the most commonly used herbicide, and in 1993, 70 to 75 million pounds were applied in the United States to more than 64 million acres of cropland. In that same year, 30 to 35 million pounds of cyanazine and 3 to 6 million pounds of simazine were used (Aspelin, 1997). Exposure to humans occurs via dermal absorption, inhalation, absorption by workers engaged in manufacturing and farming, and contamination of groundwater. In 1995, in California alone, more than a million pounds of simazine were used (primarily in grape cultivation), while about 45,000 pounds of atrazine were used (primarily in corn cultivation) (DPR, 1995).

Atrazine and simazine were evaluated by the International Agency for Research on Cancer (IARC) in 1999. As a result of this review, the overall rating for atrazine was changed from "Group 2B--possibly carcinogenic to humans" to "Group 3--not classifiable as to carcinogenicity in humans," on the basis of data that were inadequate for humans and sufficient for experimental animals. For simazine, human data were found to be inadequate and animal data were limited; therefore, IARC classified simazine as "Group 3-not classifiable as to carcinogenicity in humans" (IARC Working Group on the Evaluation of Carcinogenic Risks to Humans, 1999, p. 59).

In microbial and animal target cells in vivo and in vitro, atrazine has not been found to be genotoxic or mutagenic (Brusic, 1994; Franekic, Hulina, Kniewald, & Alaevic, 1989). Atrazine does, however, induce mammary-gland adenocarcinoma at dose levels at or above the maximum tolerated dose, and lifetime feeding studies have demonstrated an earlier onset of mammary tumors in female Sprague-Dawley rats (Wetzel et al., 1994).

Two ecological analyses of atrazine exposure and breast cancer incidence have been derived from a cancer registry in Kentucky (Hopenhayn-Rich, Stump, & Browning; 2002 Kettles, Browning, Prince, & Hortsman, 1997). The first found a positive association between indices of atrazine use and breast cancer, and the second did not.

Therefore, the relationship between OC and TH pesticide use and breast cancer risk remains controversial. Because of low exposure levels in the general population, it is difficult to identify alterations in risk. In addition, uncontrolled confounding by socioeconomic status and other risk factors for breast cancer have made elucidation of the true relationship difficult.

In conclusion, the study reported here, which utilized a statewide cancer registry and a unique statewide pesticide-reporting system, detected elevated risk of breast cancer attendant to use of two organochlorine pesticides, methoxychlor and toxaphene, and the alterations in risk persisted in a multivariate analysis in which other risk factors for breast cancer, including age, socioeconomic status, and fertility, were taken into account.

The authors recommend that use of persistent organochlorine chemicals, especially those determined to be potentially carcinogenic in animals and humans, be discontinued, and that this practice be replaced with organic and integrated pest management strategies for insect control.

Acknowledgements: Supported by Grant No. 7PB-065 from the California Breast Cancer Research Program.

Corresponding Author: Paul K. Mills, Ph.D., M.P.H., Epidemiologist, Cancer Registry of Central California, 1320 E. Shaw Ave., Suite 160, Fresno, CA 93710. E-mail: [e-mail].

REFERENCES

Arcaro, K.F., Yang, Y., Vakharia, D.D., & Gierthy, J.F. (2000). Toxaphene is antiestrogenic in a human breast-cancer cell assay. Journal of Toxicology and Environmental Health, 59(3), 197-210.

Aspelin, A.L. (1997). Pesticide industry sales and usage: 1994 and 1995 market estimates. Washington, DC: U.S. Environmental Protection Agency, Office of Pesticide Programs.

Band, P.R., Le, N.D., Fang, R., Deschamps, M., Gallagher, R.P., & Yang, P. (2000). Identification of occupational cancer risks in British Columbia: A population-based case-control study of 995 incident breast cancer cases by menopausal status, controlling for confounding factors. Journal of Occupational and Environmental Medicine, 42(3), 284-310.

Blair, A., Dosemeci, M., & Heineman, E. (1993). Cancer and other causes of death among male and female farmers from twenty-three states. American Journal of Industrial Medicine, 23, 729-742.

Blair, A., & Zahm, S.H. (1995). Agricultural exposures and cancer. Environmental Health Perspectives, 103 (Suppl. 8), 205-208.

Bonefeld, J.E.C., Autrup, H., & Hansen, J.C. (1997). Effect of toxaphene on estrogen receptor functions in human breast cancer cells. Carcinogenesis, 18, 1651-1654.

Brown, L.M., Blair, A., Gibson, R., Everett, G.D., Cantor, K., Schuman, L.M., Burmeister, L.F., Van Leir, S.F., & Dick, F. (1990). Pesticide exposures and other agricultural risk factors among men in Iowa and Minnesota. Cancer Research, 50, 6585-6591.

Brusic, D.J. (1994). An assessment of the genetic toxicity of atrazine: Relevance to human health and environmental concerns. Mutation Research, 317(2), 133-144.

California Cancer Reporting system Standards. (1988). Abstracting and coding procedures for hospitals (5th ed.). Sacramento, CA: California Environmental Protection Agency.

California Department of Pesticide Regulation. (1995). Pesticide use reports. Sacramento, CA: California Environmental Protection Agency.

California Department of Pesticide Regulation. (2000). Pesticide use reports. Sacramento, CA: California Protection Agency.

Cameron, A.C., & Trivedi, P.K. (1998). Regression analysis of count data. Cambridge: Cambridge University Press.

Cocco, P. (2000). On the rumors about the silent spring: Review of the scientific evidence linking occupational and environmental pesticide exposure to endocrine disruption health effects. Cad Saude Publica, 18, 379-402.

Cocco, P., Kazerouni, N., Zahm, S.H. (2002). Cancer mortality and environmental exposure to DDE in the United States. Environmental Health Perspectives, 108(1), 1-3.

Cummings, A.M. (1997). Methoxychlor as a model for environmental estrogens. Critical Reviews in Toxicology, 27, 367-379.

Duell, E.J., Milliken, R.C., Savitz, D.A., Newman, B., Smith, J.C., Schell, M.J., Sandler, D.P. (2000). A population-based case-control study of farming and breast cancer in North Carolina. Epidemiology, 11, 523-531.

Franekic, J., Hulina, G., Kniewald, J., & Alaevic, M. (1989). Atrazine and the genotoxicity of its metabolites. Environmental Molecular Mutagenesis, 14(1), 62-71.

Gammon, M.D., Wolff, M.S., Neugat, A.I., Eng, S., Teitelbaum, S., Britton, J., Terry, M., Levin, B., Stellman, S., Kabat, G., Hatch, M., Senie, R., Berkowitz, G., Bradlow, H., Garbowski, G., Maffeo, C., Montalvan, P., Kemeny, M., Citron, M., Schnabel, F., Schuss, A., Hajdu, S., Vinceguerra, V., Niguidula, N., Ireland, K., Santella, R. (2002). Environmental toxins and breast cancer on Long Island (Vol. II): Organochlorine compound levels in blood. Cancer Epidemiology Biomarkers and Prevention, 11, 686-697.

Gilliland, F., Hunt, W.C., Baumgartner, K.B., Crumly, D., Nicholson, C.S., Fetherolf, J., Samet, J. (1998). Reproductive risk factors for breast cancer in Hispanic and non-Hispanic white women. American Journal Epidemiology, 148, 683-692.

Hall, B., & Richards, F. (1994). Health data summary for California counties: 1994 report. Sacramento, CA: State of California, Department of Health Services.

Hopenhayn-Rich, C., Stump, M.L., & Browning, S.R. (2002), Regional assessment of atrazine exposure and incidence of breast and ovarian cancers in Kentucky. Archives of Environmental Contamination and Toxicology, 429(1), 127-136.

Hoyer, A.P., Jorgensen, T., Brock, J.W., & Grandjean, P. (2000). Organochlorine exposure and breast cancer survival, Journal of Clinical Epidemiology, 53, 323-330.

IARC Working Group on the Evaluation of Carcinogenic Risks to Humans, International Agency for Research on Cancer. (1999). Some chemicals that cause tumours of the kidney or urinary bladder in rodents, and some other substances. Monographs on the Evaluation of Carcinogenic Risks to Humans, 73.

Kelsey, J., & Bernstein, L. (1996). Epidemiology and prevention of breast cancer. Annual Review of Public Health. 17, 47-67.

Kettles, M.A., Browning, S.R., Prince, T.S., & Horstman, S.W. (1997). Triazine herbicide exposure and breast cancer incidence: an ecologic study of Kentucky counties. Environmental Health Perspectives, 105, 1222-1227.

Kocic, B., Jankovic, S., Petrovic, B., & Tiodorvic, B. (1996). Some insufficiently recognized risk factors for breast cancer. Srp Arh Celok Lek, 124(7-8), 175-178.

Kwong, S., Perkins, C., Morris, C., Cohen, R., Allen, M., Schlag, R., & Wright, W. (2000). Cancer in California 1988-1998. Sacramento, CA: California Department of Health Services, Cancer Surveillance Section.

Laden, F., & Hunter, D.J. (1998). Environmental risk factors and female breast cancer. Annual Reviews in Public Health, 19, 101-123.

Liu, L., Deapen, D., & Bernstein, L. (1998). Socioeconomic status and cancers of the female breast and reproductive organs: A comparison across racial/ethnic populations in Los Angeles County, California (United States). Cancer Causes Control, 9, 369-380.

Long, J.S. (1997). Regression models for categorical and limited dependent variables: Advanced quantitative techniques in the social sciences, Series 7. Thousand Oaks, CA: Sage Publications.

Longnecker, M., Rogan, W., & Lucier, G. (1997). The human health effects of DDT (dichlorodiphenyl-trichloroethane) and PCBs (polychlorinated biphenyls) and an overview of organochlorines in public health. Annual Reviews in Public Health, 18, 211-244.

Madigan, M.P., Ziegler, R.G., Benichou, J., Byrne, C., & Hoover, R. (1995). Proportion of breast cancer cases in the United States explained by well-established risk factors. Journal National Cancer Institute, 87, 1681-1685.

Mills, P.K. (1998). Correlation analysis of pesticide use data and cancer incidence rates in California counties. Archives Environmental Health, 53, 410-413.

Morris, C.M., Cohen, R., Perkins, C.I., Allen, M., Kwong, S.L., Schlag, R., Wright, W.E. (1999). Cancer in California 1988-1996. Sacramento, CA: California Department of Health Services, Cancer Surveillance Section.

National Center for Environmental Health, Division of Laboratory Sciences. (2003). Second national report on human exposure to environmental chemicals (Report No. 02-0716). Atlanta, GA: Centers for Disease Control and Prevention.

National Library of Medicine. (1995). Hazardous substances databank. Bethesda, MD: Author.

Olaya-Contreras, P., Rodriguez-Villamil, J., Posso-Valencia, H.J., & Cortez, J.E. (1998). Organochlorine exposure and breast cancer risk in Columbian women. Cad Saude Publica, 14(Suppl. 3), 125-132.

Petralia, S.A., Chow, W.H., McLaughlin, J.K., Jin, F., Gao, Y.T., & Dosemeci, M. (1998). Occupational risk factors for breast cancer among women in Shanghai. American Journal Industrial Medicine, 34, 477-483.

Robbins, A.S., Brescianini, S., & Kelsey, J. (1997). Regional differences in known risk factors and the higher incidence of breast cancers in San Francisco. Journal National Cancer Institute, 89, 960-965.

Romieu, I., Hernandez-Avila, M., Lazcano-Ponce, E., Weber, J.P., & Dewailly, E. (2000). Breast cancer, lactation history, and serum organochlorines. American Journal Epidemiology, 152, 363-370.

SAS Institute, Inc. (1990). SAS/Stat users guide (6th ed.). Cary, NC: Author.

Schroeder, J.C., Olshan, A.F., Baric, R., Dent, G.A., Weinberg, C.R., Yount, B., Cerhan, J.R., Lynch, C.F., Schuman, L.M., Tolbert, P., Rothman, N., Cantor, K., & Blair, A. (2001). Agricultural risk factors for t(14;18) subtypes of non-Hodgkin's lymphoma. Epidemiology, 12, 701-709.

Settimi, L., Comba, P., Carrieri, P., Boffetta, P., Terracini, B., Andrion, A., Bosia, S., Ciapinai, C., DeSantis, M., Desideri, E., Fedi, A., Luccoli, L., Maiozzi, P., Masina, A., Perazzo, P., & Axelson, O. (1999). Cancer risk among female agricultural workers: A multi-center case-control study. American Journal Industrial Medicine, 6(1), 135-141.

Soto, A.M., Chung, K.L., & Sonnenschein, C. (1994). The pesticides endosulfan, toxaphene and dieldrin have estrogenic effects on human estrogen-sensitive cells. Environmental Health Perspectives, 102, 380-383.

Surveillance, Epidemiology, and End Results Program. (1986). Summary staging guide for the cancer surveillance, epidemiology, and end results reporting (SEER) program. Bethesda, MD: U.S. Public Health Service, National Institutes of Health.

Surveillance, Epidemiology, and End Results Program. (1992) SEER extent of disease--1988 coding instructions (2nd ed.) (NIH Pub. No. 92-2313). Bethesda, MD: National Cancer Institute.

U.S. Census Bureau. (1990). General population characteristics: California (CP-1-6). Washington, DC: U.S. Department of Commerce, Economics and Statistics Administration, U.S. Census Bureau.

U.S. Environmental Protection Agency. (2002). List of chemicals evaluated for carcinogenic potentials science. Washington, DC: Author, Office of Pesticide Programs, Information Management Branch, Health Effects Division.

Wang, Y., Lewis-Michl, E.L., Fitzgerald, E., & Stark, A. (2002). Cancer incidence among a cohort of female farm residents in New York State. Archives of Environmental Health, 57, 561-567.

Wauchope, R.D., Buttler, T.M., Hornsby, A.G., Augustijn-Becker, P., & Burt, J. (1992). SCS/ARS/CES Pesticide properties database for environmental decision making. Reviews in Environmental Contamination and Toxicology, 123, 1-157.

Wetzel, L.T., Luempert, L.G., Breckenridge, C.B., Tisdel, M., Stevens, J., Thakur, A., & Eldridge, J. (1994). Chronic effects of atrazine on estrous and mammary tumor formation in female Sprague-Dawley and Fischer-344 rats. Journal of Toxicology and Environmental Health, 43(2), 169-182.

Wicklund, K., & Dich, J. (1994). Cancer risks among female farmers in Sweden. Cancer Causes and Control, 5, 449-457.

Wolff, M.S., Berkowitz, G.S., Brower, S., Senie, R., Bleiweiss, I.J., Tartter, P., Pace, B., Roy, N., Wallenstein, S., & Weston, A. (2000). Organochlorine exposures and breast cancer risk in New York City women. Environmental Research, 84(2), 151-161.

Yost, K., Perkins, C., Cohen, R., Morris, C., & Wright, W. (2001). Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes Control, 12, 703-711.

Paul K. Mills, Ph.D., M.P.H.

Richard Yang, M.P.H.

TABLE 1 Characteristics of Chemicals Evaluated in Relation to Breast
Cancer Risk in California Latinas

              Chemical
Pesticide     Class                 Use          Carcinogenicity

Atrazine      Triazine              Herbicide    Not likely to be
                                                 carcinogenic to humans
Simazine      Triazine              Herbicide    Group C -- possible
                                                 human carcinogen
DDT           Organochlorine        Insecticide  Group B2 -- probable
                                                 human carcinogen
Toxaphene     Chlorinated Camphene  Insecticide  Group B2 -- probable
                                                 human carcinogen
Methoxychlor  Organochlorine        Insecticide  Group D -- not
                                                 classifiable as to
                                                 human carcinogenicity
Aldrin        Organochlorine        Insecticide  Group B2 -- probable
                                                 human carcinogen
Chlordane     Organochlorine        Insecticide  Group B2 -- probable
                                                 human carcinogen
Dieldrin      Organochlorine        Insecticide  Group B2 -- probable
                                                 human carcinogen
Endosulfan    Chlorinated Sulfite   Insecticide  Not likely to be
                                                 carcinogenic to humans
Lindane       Organochlorine        Insecticide  Suggestive evidence of
                                                 carcinogenicity, but
                                                 not sufficient to
                                                 assess human
                                                 carcinogenic potential

              CAS        Pounds Applied in
Pesticide     Number     California, 1980

Atrazine      1912-24-9  430,000
Simazine       122-34-9  238,000
DDT             50-29-3  Banned, 1972
Toxaphene     8001-35-2  450,000
Methoxychlor    72-43-5  102,000
Aldrin         309-00-2      865
Chlordane       57-74-9  272,000
Dieldrin        60-57-1      144
Endosulfan     115-29-7  294,000
Lindane         58-89-9   55,000

TABLE 2 Adjusted Incidence Rate Ratios for Age, Fertility, and
Socioeconomic Status and Breast Cancer Risk*

Variable                    IRR    95 Percent CI  p-value

Age
  25-29 (reference)          1.00   -             -
  30-34                      3.04   2.36-3.91     0.0001
  35-39                      6.66   5.23-8.48     0.0001
  40-44                     12.76  10.07-16.18    0.0001
  45-49                     20.01  15.79-25.35    0.0001
  50-54                     23.49  18.51-29.81    0.0001
  55-59                     26.68  21.04-33.82    0.0001
  60-64                     32.84  25.86-41.70    0.0001
  65-69                     44.13  34.75-56.03    0.0001
  70-74                     47.65  37.44-60.62    0.0001
  75-79                     40.88  31.98-52.25    0.0001
  80-84                     39.58  30.72-51.00    0.0001
  85+                       29.71  22.87-38.59    0.0001
Fertility
  Fertility (low)**          1.12   1.01-1.24     0.0304
  Fertility                  1.03   0.92-1.15     0.6355
  Fertility                  1.00   0.90-1.12     0.9668
  Fertility (high)           1.00   -             -
Socioeconomic status (SES)
  SES (high)**               1.18   1.06-1.32     0.0021
  SES                        1.24   1.10-1.39     0.0003
  SES                        1.04   0.93-1.16     0.4868
  SES (low)                  1.00   -             -

*In Hispanic Females, 1988-1999, in California.
**Adjusted for age.

TABLE 3 Adjusted* Incidence Rate Ratios for Organochlorine Chemicals**
and Breast Cancer in Hispanic Women in California***, 1988-1993 and
1994-1999

Chemical                  1988-1993                 1994-1999
                     IRR   95% CI     p-value  IRR   95% CI     p-value

Aldrin (low)         1.00  -          -        1.00  -          -
Aldrin               0.86  0.70-1.07  0.1898   0.96  0.78-1.17  0.7154
Aldrin               0.82  0.67-1.01  0.0584   0.88  0.73-1.07  0.2166
Aldrin (high)        0.83  0.68-1.01  0.0677   0.90  0.75-1.09  0.2772
Chlordane (low)      1.00  -          -        1.00  -          -
Chlordane            1.33  0.72-2.44  0.3579   0.94  0.60-1.47  0.7978
Chlordane            1.20  0.66-2.17  0.5456   0.88  0.57-1.35  0.5538
Chlordane (high)     1.08  0.59-1.94  0.8074   0.83  0.54-1.29  0.4213
DDT (low)            1.00  -          -        1.00  -          -
DDT                  0.94  0.87-1.03  0.2167   0.96  0.88-1.04  0.2982
DDT                  0.87  0.80-0.95  0.0019   0.92  0.85-0.99  0.0449
DDT (high)           0.94  0.85-1.03  0.1558   1.02  0.94-1.11  0.6620
Dieldrin (low)       1.00  -          -        1.00  -          -
Dieldrin             0.80  0.55-1.16  0.2489   1.08  0.79-1.48  0.6069
Dieldrin             0.89  0.62-1.29  0.5538   1.10  0.81-1.50  0.5322
Dieldrin (high)      0.78  0.54-1.14  0.1987   1.05  0.76-1.42  0.7735
Endosulfan (low)     1.00  -          -        1.00  -          -
Endosulfan           1.04  0.88-1.22  0.6364   0.96  0.83-1.11  0.5774
Endosulfan           1.14  0.98-1.34  0.0952   1.04  0.90-1.20  0.5806
Endosulfan (high)    1.12  0.95-1.32  0.1764   1.02  0.88-1.19  0.8135
Lindane (low)        1.00  -          -        1.00  -          -
Lindane              0.93  0.69-1.24  0.6211   0.85  0.64-1.12  0.2402
Lindane              0.94  0.71-1.25  0.6784   0.85  0.65-1.11  0.2374
Lindane (high)       0.88  0.67-1.16  0.3789   0.82  0.63-1.07  0.1461
Methoxychlor (low)   1.00  -          -        1.00  -          -
Methoxychlor         1.13  1.02-1.27  0.0230   1.06  0.96-1.17  0.2733
Methoxychlor         1.06  0.95-1.20  0.2854   1.04  0.94-1.16  0.4299
Methoxychlor (high)  1.18  1.03-1.35  0.0142   1.16  1.02-1.30  0.0190
Toxaphene (low)      1.00  -          -        1.00  -          -
Toxaphene            1.09  0.95-1.25  0.2037   1.01  0.90-1.15  0.8282
Toxaphene            1.17  1.01-1.35  0.0352   1.06  0.93-1.20  0.4094
Toxaphene (high)     1.16  1.01-1.34  0.0330   1.08  0.94-1.22  0.2510

*Adjusted for age, socioeconomic status, and fertility level.
**Pounds of Active Ingredients Applied, 1970-1988.
***Excludes counties with fewer than 10 cases.

TABLE 4 Adjusted* IRRs for Triazine Herbicides** and Breast Cancer in
Hispanic Women in California***, 1988-1993 and 1994-1999

Chemical              1988-1993                 1994-1999
                 IRR   95% CI     p-value  IRR   95% CI     p-value

Atrazine (low)   1.00  -          -        1.00  -          -
Atrazine         0.83  0.66-1.02  0.0802   0.87  0.72-1.04  0.1256
Atrazine         0.94  0.76-1.15  0.5634   0.91  0.76-1.09  0.2898
Atrazine (high)  0.86  0.70-1.05  0.1456   0.87  0.73-1.04  0.1374
Simazine (low)   1.00  -          -        1.00  -          -
Simazine         1.17  0.88-1.56  0.2777   1.40  1.14-1.72  0.0013
Simazine         1.10  0.96-1.25  0.1548   1.06  0.94-1.19  0.3514
Simazine (high)  1.05  0.93-1.19  0.4423   1.07  0.95-1.20  0.2369

*Adjusted for age, socioeconomic status, and fertility level.
**Pounds of active ingredients applied, 1970-1988.
***Excludes counties with less than 10 cases.