Abstract

Pesticides, common environmental exposures, have been examined in relation to breast cancer primarily in occupational studies or exposure biomarker studies. No known studies have focused on self-reported residential pesticide use. The authors investigated the association between reported lifetime residential pesticide use and breast cancer risk among women living on Long Island, New York. They conducted a population-based case-control study of 1,508 women newly diagnosed with breast cancer between August 1996 and July 1997 and 1,556 randomly selected, age-frequency-matched controls. Comprehensive residential pesticide use and other risk factors were assessed by using an in-person, interviewer-administered questionnaire. Unconditional logistic regression was used to calculate odds ratios and 95% confidence intervals. Breast cancer risk was associated with ever lifetime residential pesticide use (odds ratio = 1.39, 95% confidence interval: 1.15, 1.68). However, there was no evidence of increasing risk with increasing lifetime applications. Lawn and garden pesticide use was associated with breast cancer risk, but there was no dose response. Little or no association was found for nuisance-pest pesticides, insect repellants, or products to control lice or fleas and ticks on pets. This study is the first known to suggest that self-reported use of residential pesticides may increase breast cancer risk. Further investigation in other populations is necessary to confirm these findings.

The search for environmental factors associated with breast cancer is of great public interest. Pesticides are common environmental exposures that have been implicated in cancer etiology (1–3). Studies of breast cancer have primarily examined pesticide exposures among occupationally exposed individuals and among women in the general population (i.e., those not occupationally exposed) using biologic markers of exposure as well as indirect exposures such as residential proximity to pesticide exposure sources. However, to our knowledge, no investigation of lifetime, self-reported residential pesticide use has been published.

Numerous pesticides have shown carcinogenicity of varying levels (4–6). They have also been found to be genotoxic, tumor promoters, immunotoxic, and estrogenic (1). Organochlorine pesticides, which include dichlorodiphenyl-trichloroethane (DDT), have been shown to have both estrogenic and carcinogenic properties (1, 7). This biologic plausibility and the fact that long-term exposure to organochlorine pesticides can be inferred from blood or adipose tissue levels account for why they are the most studied pesticide group in relation to breast cancer (1). However, epidemiologic studies have not provided convincing support for an adverse association with biomarkers of organochlorine pesticides (4–8). Many other pesticides commonly used in and around the home may have the potential to influence breast cancer risk but have not yet been studied. The use of biomarkers for many of these pesticides is limited since they may be short term or not available.

Therefore, a population-based, case-control study of the environment and breast cancer was conducted that assessed not only blood levels of organochlorine pesticides (9) but also a wide range of residential pesticide exposures through the use of an in-person, interviewer-administered questionnaire. The organochlorine pesticide biomarker investigation in the Long Island Breast Cancer Study Project (LIBCSP) population did not reveal an association with breast cancer risk (10). The analyses presented here investigate the relation between self-reported lifetime residential pesticide use and breast cancer risk among women living on Long Island, New York.

MATERIALS AND METHODS

Study population

Details of LIBCSP have been described previously (9). In brief, women who were residents of either Nassau or Suffolk Counties in New York State, newly diagnosed with invasive or in situ breast cancer between August 1, 1996, and July 31, 1997, were eligible as cases. Women residents of the same two counties during the same time period who had not been diagnosed with breast cancer were randomly selected as controls. Controls were frequency matched by 5-year age group to the expected age distribution of the cases. Controls were selected through random digit dialing if they were less than 65 years of age (screener response rate for random digit dialing = 77.9 percent) and were selected from Centers for Medicare and Medicaid Services (formerly the Health Care Financing Administration) rosters if 65 years of age or older. At the reference date (date of diagnosis for cases and date of identification for controls), women less than 65 years of age were required to have a residential telephone, and women 65 years of age or older were required to be Medicare participants.

Of 2,030 identified, eligible cases, 193 (9.5 percent) were not contacted because of physician refusal. Of those remaining, 1,508 cases (82.1 percent) were interviewed. Of 2,481 eligible controls, 1,556 (62.8 percent) completed the interview. For women less than 65 years of age, response rates were 89 percent for cases and 76 percent for controls. Among women 65 years of age or older, response rates were 72 percent for cases and 43 percent for controls. The most common reasons for both case and control nonparticipation were refusal and illness. The institutional review boards of all participating institutions approved the study protocol, and all participants provided informed consent.

Data collection

Trained interviewers administered a structured questionnaire that collected information on reproductive and medical history, occupational and residential history, and lifestyle and demographic characteristics. Detailed information on pesticide use, a primary focus of the study, was also obtained (http://epi.grants.cancer.gov/documents/LIBCSP/projects/ques/sectc.pdf). Participants were queried about pesticide use in and around their homes as well as their use of insect repellants, lice control products on themselves and others, and flea and tick control products on their pets. Usage patterns for seven lawn and garden pest categories and eight nuisance-pest categories were ascertained. For a particular pest category, participants identified all persons who applied the products, the various types of products applied, and the average yearly frequency of application and the number of years the products had been used. The number of lifetime applications for each individual category was derived by multiplying yearly application frequency by years of use.

Overall pesticide use (the sum of lifetime applications of all 15 categories), the two combined groups (lawn and garden and nuisance pest), and each of the 15 individual categories were considered in the analyses. Lifetime applications were categorized based on the control distribution. Women in the lowest quintile of overall pesticide use constituted the reference group in analyses of overall pesticide use. For the combined group analyses, women reporting no pesticide use for all individual pest categories within a combined group were the reference group. In this paper, individual category results are presented dichotomized as ever/never use because trends in breast cancer risk within each category were not observed. The use of two different combined group reference categories created reasonably sized groups of women who were unexposed regarding approximately half of the individual pest categories, which reduced exposure misclassification and provided more interpretable odds ratios because they were calculated by using the same reference group for a large number of categories. When odds ratios for individual pest categories were calculated based on all women who had not used pesticides for the category as the reference group, odds ratios were attenuated toward the null.

Patterns of pesticide use were examined in two ways: by applicator (self only/professional only/other or multiple applicators) and by product type (spray only/powder only/liquid only/other product or multiple types). Patterns of use could not be examined for two individual lawn and garden categories (other types of pesticides used outdoors and chemicals used on indoor plants) because of the infrequency of such pesticide use.

The analyses presented are based on 1,505 cases and 1,553 controls because three cases and three controls did not provide any pesticide information. To calculate total lifetime pesticide applications for women who reported pesticide use for an individual pest category but for whom data on use patterns were missing, the median value for the specific pattern variable was used (e.g., median years of weed killer use was used for women without reported duration of use); exclusion of these women did not materially change estimates of association with lifetime applications. For all other analyses, women for whom data were missing were excluded.

Statistical analysis

Unconditional logistic regression was used to calculate odds ratios and 95 percent confidence intervals. Lifetime applications were entered into models as indicator variables for each quintile above the lowest. Characteristics assessed as possible confounders included race, marital status, religion, household income, age at menarche, parity, age at first birth, lactation, menopausal status, oral contraceptive use, hormone replacement therapy use, first-degree family history of breast cancer, history of benign breast disease, body mass index (weight in kilograms divided by height in meters squared) at the reference age and at age 20 years, alcohol use, smoking status, and physical activity. None of these factors were included in the final model because none resulted in at least a 10 percent change in the beta coefficient for the pesticide variables (11, 12). In addition to controlling all models for age, adjustment for education (defined as highest educational level attained: ≤high school graduate/some college/college graduate/postcollege) was included to control for possible confounding by socioeconomic status. Note that adjustment for education did not result in any substantial changes in the estimates of effect. Finally, models assessing one combined group (e.g., lawn and garden) were adjusted for the use of the other combined group (e.g., nuisance pests).

RESULTS

An increased breast cancer risk was associated with lifetime pesticide application for all pest groups combined (age- and education-adjusted odds ratio (OR) = 1.39, 95 percent confidence interval (CI): 1.15, 1.68 for quintiles 2–5 vs. quintile 1; table 1). However, there was no indication of increasing risk with increasing quintile of lifetime applications. Compared with never use of any lawn and garden pesticides, use of pesticides for these types of pests was associated with an elevated breast cancer risk (table 1). Adjustment for lifetime applications of nuisance-pest pesticides did not substantially change risk estimates. Although the age-adjusted odds ratio for ever use of combined nuisance-pest pesticides was elevated when compared with that for women who never used any pesticides for nuisance pests (age-adjusted OR = 1.20, 95 percent CI: 0.88, 1.63), further adjustment for use of lawn and garden pesticides reduced the odds ratio toward the null. No dose-response associations were observed for either combined group, and additional adjustment for education did not affect any odds ratios.

TABLE 1.

Adjusted odds ratios and 95% confidence intervals for breast cancer, according to lifetime applications of pesticides, among 3,058 women in Nassau and Suffolk Counties, New York, 1996–1997

 Lifetime applications (no.) Cases (no.) Controls (no.) Adjusted for age, education, and other combined pest group* 
 Minimum Median Maximum OR 95% CI† 
All pest groups combined        
    Quintile 1 (reference) 16 230 310 1.00 Reference 
    Quintiles 2–5    1,275 1,243 1.39 1.15, 1.68 
    Quintile 2 17 32 50 298 315 1.30 1.03, 1.64 
    Quintile 3 51 77 111 313 307 1.39 1.10, 1.76 
    Quintile 4 112 159 242 347 311 1.49 1.19, 1.88 
    Quintile 5 243 482 20,834 317 310 1.37 1.08, 1.72 
Lawn and garden combined group‡        
    Never used lawn and garden pesticides    240 305 1.00 Reference 
    Ever used lawn and garden pesticides    1,254 1,231 1.34 1.11, 1.63 
    Quartile 1 15 282 303 1.25 0.98, 1.59 
    Quartile 2 16 28 44 341 313 1.44 1.14, 1.82 
    Quartile 3 45 70 108 301 307 1.30 1.02, 1.64 
    Quartile 4 109 180 20,820 330 308 1.38 1.09, 1.75 
Nuisance-pest combined group§        
    Never used nuisance-pest pesticides    100 117 1.00 Reference 
    Ever used nuisance-pest pesticides    1,404 1,436 1.07 0.80, 1.42 
    Quartile 1 290 364 0.88 0.64, 1.20 
    Quartile 2 16 29 338 358 1.07 0.78, 1.46 
    Quartile 3 30 53 96 393 357 1.20 0.88, 1.65 
    Quartile 4 97 223 9,608 383 357 1.16 0.85, 1.58 
 Lifetime applications (no.) Cases (no.) Controls (no.) Adjusted for age, education, and other combined pest group* 
 Minimum Median Maximum OR 95% CI† 
All pest groups combined        
    Quintile 1 (reference) 16 230 310 1.00 Reference 
    Quintiles 2–5    1,275 1,243 1.39 1.15, 1.68 
    Quintile 2 17 32 50 298 315 1.30 1.03, 1.64 
    Quintile 3 51 77 111 313 307 1.39 1.10, 1.76 
    Quintile 4 112 159 242 347 311 1.49 1.19, 1.88 
    Quintile 5 243 482 20,834 317 310 1.37 1.08, 1.72 
Lawn and garden combined group‡        
    Never used lawn and garden pesticides    240 305 1.00 Reference 
    Ever used lawn and garden pesticides    1,254 1,231 1.34 1.11, 1.63 
    Quartile 1 15 282 303 1.25 0.98, 1.59 
    Quartile 2 16 28 44 341 313 1.44 1.14, 1.82 
    Quartile 3 45 70 108 301 307 1.30 1.02, 1.64 
    Quartile 4 109 180 20,820 330 308 1.38 1.09, 1.75 
Nuisance-pest combined group§        
    Never used nuisance-pest pesticides    100 117 1.00 Reference 
    Ever used nuisance-pest pesticides    1,404 1,436 1.07 0.80, 1.42 
    Quartile 1 290 364 0.88 0.64, 1.20 
    Quartile 2 16 29 338 358 1.07 0.78, 1.46 
    Quartile 3 30 53 96 393 357 1.20 0.88, 1.65 
    Quartile 4 97 223 9,608 383 357 1.16 0.85, 1.58 
*

Odds ratios (ORs) for the category of all pest groups combined were adjusted for age and educational status only (≤high school graduate/some college/college graduate/postcollege).

CI, confidence interval.

Data for 11 cases and 17 controls were missing.

§

Data for one case were missing.

Individual lawn and garden categories

Ever use of pesticides for nearly all of the individual lawn and garden categories was associated with increased breast cancer risk (table 2). In general, within each category, no dose response was observed (data not shown). Little variation in breast cancer risk was observed for individual categories of lawn and garden pesticide application when classified according to the person who applied them (table 3). Women who exclusively self-applied lawn and garden pesticides were at a moderately increased risk of breast cancer (e.g., lawn insecticides OR = 1.56, 95 percent CI: 1.01, 2.43; chemicals for insects or diseases of outdoor plants OR = 1.58, 95 percent CI: 1.12, 2.22). However, these risk estimates were in the same range as those for pesticide use without consideration of applicator and were not generally different from those observed for women who had either professionals exclusively apply these pesticides (e.g., lawn insecticides OR = 1.41, 95 percent CI: 1.31, 1.77) or others perform the application (e.g., lawn insecticides OR = 1.32, 95 percent CI: 1.05, 1.67).

TABLE 2.

Adjusted odds ratios and 95% confidence intervals for breast cancer, according to lifetime applications of pesticides for individual categories of pests,* among 3,058 women in Nassau and Suffolk Counties, New York, 1996–1997

Ever used pesticides for individual categories of pests Cases (no.) Controls (no.) Adjusted for age, education,† and other combined pest group 
OR‡ 95% CI§ 
Never used any lawn and garden pesticides 240 305 1.00 Reference 
    Weeds 1,109 1,083 1.43 1.17, 1.75 
    Lawn insects 799 766 1.48 1.20, 1.82 
    Insects or diseases of trees 539 514 1.46 1.17, 1.81 
    Pests in vegetable or fruit gardens 298 259 1.58 1.24, 2.01 
    Insects or diseases of outdoor plants 261 232 1.54 1.20, 1.98 
    Any other type of outdoor pest 70 48 1.13 0.86, 1.49 
    Insects or diseases of indoor plants 121 112 1.48 1.08, 2.02 
Never used any nuisance-pest pesticides 100 117 1.00 Reference 
    Ants, carpenter ants, or cockroaches 1,160 1,171 1.06 0.79, 1.42 
    Bees or wasps 599 610 1.05 0.77, 1.43 
    Flies or mosquitoes 328 310 1.12 0.81, 1.55 
    Moths, silverfish, or caterpillars 388 352 1.19 0.87, 1.64 
    Mice, rats, gophers, or moles 246 268 1.02 0.73, 1.42 
    Fleas or ticks, except on pets 377 394 1.06 0.91, 1.23 
    Termites 731 712 1.10 0.81, 1.48 
    Any other type of pest in the home 50 52 1.06 0.65, 1.71 
Ever used pesticides for individual categories of pests Cases (no.) Controls (no.) Adjusted for age, education,† and other combined pest group 
OR‡ 95% CI§ 
Never used any lawn and garden pesticides 240 305 1.00 Reference 
    Weeds 1,109 1,083 1.43 1.17, 1.75 
    Lawn insects 799 766 1.48 1.20, 1.82 
    Insects or diseases of trees 539 514 1.46 1.17, 1.81 
    Pests in vegetable or fruit gardens 298 259 1.58 1.24, 2.01 
    Insects or diseases of outdoor plants 261 232 1.54 1.20, 1.98 
    Any other type of outdoor pest 70 48 1.13 0.86, 1.49 
    Insects or diseases of indoor plants 121 112 1.48 1.08, 2.02 
Never used any nuisance-pest pesticides 100 117 1.00 Reference 
    Ants, carpenter ants, or cockroaches 1,160 1,171 1.06 0.79, 1.42 
    Bees or wasps 599 610 1.05 0.77, 1.43 
    Flies or mosquitoes 328 310 1.12 0.81, 1.55 
    Moths, silverfish, or caterpillars 388 352 1.19 0.87, 1.64 
    Mice, rats, gophers, or moles 246 268 1.02 0.73, 1.42 
    Fleas or ticks, except on pets 377 394 1.06 0.91, 1.23 
    Termites 731 712 1.10 0.81, 1.48 
    Any other type of pest in the home 50 52 1.06 0.65, 1.71 
*

For all individual categories, some data were missing.

Educational status: ≤high school graduate/some college/college graduate/postcollege.

Each odds ratio (OR) was derived from a separate model. To create a common combined-pest-category reference group, a set of three mutually exclusive indicator variables was used: never used combined-pest-category pesticides (reference), used pesticides for individual categories of pests, used combined-pest-category pesticides but not for individual categories of pests.

§

CI, confidence interval.

TABLE 3.

Adjusted odds ratios and 95% confidence intervals for breast cancer, according to lifetime applications of pesticides, person who applied the pesticides, and type of lawn and garden pesticide product in individual categories,* among 3,058 women in Nassau and Suffolk Counties, New York, 1996–1997

Use of pesticides for individual lawn and garden pest problems Person who applied the pesticides Cases (no.) Controls (no.) Adjusted† Type of product applied Cases (no.) Controls (no.) Adjusted† 
OR‡ 95% CI‡ OR 95% CI 
Never used any lawn and garden pesticides  240 305 1.00 Reference  240 305 1.00 Reference 
Weeds Self only 101 110 1.21 0.88, 1.68 Spray only 292 300 1.30 1.02, 1.65 
 Professional only 363 348 1.36 1.08, 1.71 Powder only 290 298 1.30 1.02, 1.65 
 Other§ 640 623 1.36 1.10, 1.67 Liquid only 77 73 1.40 0.97, 2.03 
      Other¶ 437 408 1.43 1.14, 1.79 
Lawn insects Self only 51 44 1.56  Spray only 197 205 1.32 1.01, 1.72 
 Professional only 402 373 1.41 1.13, 1.77 Powder only 193 195 1.38 1.06, 1.81 
 Other 336 340 1.32 1.05, 1.67 Liquid only 53 40 1.77 1.12, 2.77 
      Other 340 313 1.49 1.18, 1.88 
Insects or diseases of trees Self only 15 30 0.67 0.35, 1.29 Spray only 426 418 1.40 1.12, 1.75 
Professional only 372 343 1.45 1.15, 1.83 Powder only   
Other 148 140 1.42 1.06, 1.9 Liquid only 25 25 1.41 0.78, 2.53 
     Other 83 68 1.69 1.17, 2.44 
Pests in vegetable or fruit gardens Self only 87 85 1.41 0.99, 1.99 Spray only 115 94 1.64 1.18, 2.27 
Professional only 14 2.29 0.94, 5.58 Powder only 107 97 1.50 1.08, 2.09 
Other 195 166 1.56 1.18, 2.04 Liquid only 1.13 0.41, 3.11 
     Other 68 59 1.60 1.08, 2.38 
Insects or diseases of outdoor plants Self only 101 86 1.58 1.12, 2.22 Spray only 127 107 1.58 1.16, 2.17 
Professional only 49 37 1.79 1.12, 2.84 Powder only 31 47 0.91 0.56, 1.49 
Other 106 107 1.29 0.93, 1.78 Liquid only 12 0.93 0.37, 2.33 
     Other 91 66 1.83 1.27, 2.64 
Use of pesticides for individual lawn and garden pest problems Person who applied the pesticides Cases (no.) Controls (no.) Adjusted† Type of product applied Cases (no.) Controls (no.) Adjusted† 
OR‡ 95% CI‡ OR 95% CI 
Never used any lawn and garden pesticides  240 305 1.00 Reference  240 305 1.00 Reference 
Weeds Self only 101 110 1.21 0.88, 1.68 Spray only 292 300 1.30 1.02, 1.65 
 Professional only 363 348 1.36 1.08, 1.71 Powder only 290 298 1.30 1.02, 1.65 
 Other§ 640 623 1.36 1.10, 1.67 Liquid only 77 73 1.40 0.97, 2.03 
      Other¶ 437 408 1.43 1.14, 1.79 
Lawn insects Self only 51 44 1.56  Spray only 197 205 1.32 1.01, 1.72 
 Professional only 402 373 1.41 1.13, 1.77 Powder only 193 195 1.38 1.06, 1.81 
 Other 336 340 1.32 1.05, 1.67 Liquid only 53 40 1.77 1.12, 2.77 
      Other 340 313 1.49 1.18, 1.88 
Insects or diseases of trees Self only 15 30 0.67 0.35, 1.29 Spray only 426 418 1.40 1.12, 1.75 
Professional only 372 343 1.45 1.15, 1.83 Powder only   
Other 148 140 1.42 1.06, 1.9 Liquid only 25 25 1.41 0.78, 2.53 
     Other 83 68 1.69 1.17, 2.44 
Pests in vegetable or fruit gardens Self only 87 85 1.41 0.99, 1.99 Spray only 115 94 1.64 1.18, 2.27 
Professional only 14 2.29 0.94, 5.58 Powder only 107 97 1.50 1.08, 2.09 
Other 195 166 1.56 1.18, 2.04 Liquid only 1.13 0.41, 3.11 
     Other 68 59 1.60 1.08, 2.38 
Insects or diseases of outdoor plants Self only 101 86 1.58 1.12, 2.22 Spray only 127 107 1.58 1.16, 2.17 
Professional only 49 37 1.79 1.12, 2.84 Powder only 31 47 0.91 0.56, 1.49 
Other 106 107 1.29 0.93, 1.78 Liquid only 12 0.93 0.37, 2.33 
     Other 91 66 1.83 1.27, 2.64 
*

For all individual categories, some data were missing.

All models were adjusted for age, education, and combined-nuisance-pest pesticide use.

OR, odds ratio; CI, confidence interval.

§

Any combination of appliers or some person other than self or professional only.

Any combination of product types or some type other than spray, powder, or liquid only.

Overall, examination of breast cancer risk associated with use of different product types in the individual lawn and garden categories did not reveal any specific types as being associated with risk different from that observed without categorization by product type (table 3). For lawn insecticide application, exclusive use of the liquid form was associated with higher risk (OR = 1.77, 95 percent CI: 1.12, 2.77) than other product types. Women who used a combination of product types or some other product type for outdoor plant pest problems were at higher risk than women who used spray only (OR = 1.83, 95 percent CI: 1.27, 2.64).

Individual nuisance-pest categories

Lifetime applications of the individual categories of nuisance-pest pesticides did not appear to be related to breast cancer risk (table 2). Similarly, breast cancer risk was not elevated for nuisance-pest pesticide application when categorized by the person who performed the application or when classified by product types; the vast majority of the odds ratio estimates were at or near the null value (data not shown).

Subgroup analyses

We examined whether the relation of residential pesticide use and breast cancer varied within different subgroups of participants (data not shown). When the population was stratified by menopausal status (pre- vs. post-) or by length of residency (≥15 or <15 years in the current home) or was restricted to participants less than age 65 years, the associations between pesticide use and breast cancer risk in the subgroups were not considerably different from those for the entire sample. The relation was also examined according to stage of disease, that is, invasive cases only (84.4 percent) and in situ cases only. Associations for the invasive cases were nearly identical to those observed for all cases combined. Among in situ cases, the findings for breast cancer risk and use of pesticides for nuisance pests were similar to those found for invasive cases. However, the associations for all pest groups combined and the lawn and garden pest group were stronger for some, but not all, of the quantiles of use (OR = 1.91, 95 percent CI: 1.17, 3.13; OR = 1.97, 95 percent CI: 1.21, 3.21; OR = 2.03, 95 percent CI: 1.25, 3.30; OR = 1.23, 95 percent CI: 0.72, 2.08 for quartiles 1–4, respectively, vs. never use of lawn and garden pesticides). It is important to note that these latter measures were less stable because of the reduced sample size.

Insect repellants, lice control products, and flea and tick products used on pets

Breast cancer risk was not associated with frequent or long-term use of insect repellants (OR = 0.89, 95 percent CI: 0.60, 1.31 for ≥20 years vs. never use) or with use of lice control products (OR = 0.86, 95 percent CI: 0.71, 1.04 for use ≥2 times vs. never use). Likewise, no increased risk of breast cancer was associated with the use of flea and tick products on pets, when examined by frequency of application (OR = 1.08, 95 percent CI: 0.87, 1.33 for use ≥30 times vs. never use), type of product used, or person who applied the product (data not shown).

DISCUSSION

Overall, women who reported the highest (quintiles 2–5 combined) pesticide use in and around their homes had more than a 30 percent increased risk of breast cancer relative to women who reported the lowest use. Lifetime applications and patterns of use of pesticides for nuisance pests were consistently observed to have little or no association with breast cancer risk when examined as either a combined group or by individual categories. On the other hand, use of the combined lawn and garden pesticides as well as the majority of the individual pest categories in this group consistently showed an elevation in breast cancer risk. Finally, use of insect repellants, lice control products, or pet flea and tick control products was not related to breast cancer risk.

Interpretation of these findings in the context of other studies is limited because, to our knowledge, there are no published studies of self-reported residential pesticide use and breast cancer. A recent review of the numerous studies that have examined biologic markers of various organochlorine pesticides concluded that there was little support for a positive association between dichlorodiphenyldichloroethene or dichlorodiphenyltrichloroethane and breast cancer risk (13). Our own investigation of organochlorine levels among the LIBCSP population did not find increased breast cancer risk (10). The absence of an association for treatment of the home for termites in this analysis and in an earlier Long Island study that examined reported termiticide use (14) agrees with the lack of a breast cancer association with chlordane in the LIBCSP population (10). The observed null findings for lice control products is supported by the observed lack of association in the majority of epidemiologic studies that examined beta-hexachlorocyclohexane, a contaminant of lindane (gamma-hexachlorocyclohexane), and breast cancer risk (15–17).

Studies of agricultural workers, an occupational group with a high probability of pesticide exposure, have not shown an increased breast cancer risk among women (18–30). Many of the studies suffered from small sample size or lacked confounder information. Furthermore, occupation is a nonspecific indicator of pesticide exposure, possibly biasing results toward the null. In a recent study, breast cancer in farmers' wives was associated with their husbands' use of 2,4,5-trichlorophenoxypropionic acid, 2,4,5,-trichlorophenoxyacetic acid, and captan (31), pesticides that could have been used on Long Island. Thus, these findings support our observation of increased breast cancer risk and use of pesticides for weeds and fruit tree pests.

Other classes of popular pesticides have replaced organochlorines over the years in a continuing search for less toxic, but effective agents. Organophosphates were formerly among the most widely used household pesticides, accounting for about 22 percent of nonagricultural usage in 2001, so that many women in our study would almost certainly have been exposed in the past. However, because of health concerns, two major organophosphate pesticides, chlorpyrifos (widely used in lawns and against termites—the active ingredient in Dursban (Dow Agrosciences, LLC, Indianapolis, Indiana)) and diazinon, were restricted or banned for residential use after 2001 by the Environmental Protection Agency. Dichlorvos, formerly used in home foggers and aerosols and to control insects in passenger aircraft, is now classified by the Environmental Protection Agency as a “Restricted Use Pesticide” and may be purchased legally by certified applicators only. It is still used in no-pest strips, pet collars, and kennels. Dichlorvos has an Environmental Protection Agency carcinogenicity classification of B2 (probable human carcinogen (32)) and a rating of 2B (possibly carcinogenic to humans) from the International Agency for Research on Cancer. Organophosphates are mostly nonestrogenic, but mixtures of several organophosphates were found to affect birth weight and fetal viability. Chlorpyrifos was regarded as nonestrogenic until recently, when two studies showed possible evidence of weak estrogenicity (33, 34).

Carbamates, another class of widely used insecticides, include the Sevin (Aventis CropScience, Inc., Strasbourg, France) brand of carbaryl. Since Dursban was banned, it has become one of the most popular brands of carbaryl insecticides for home garden use. Carbaryl currently has a carcinogenicity rating of group III (unclassifiable as to human carcinogenicity) from the International Agency for Research on Cancer, although one study has reported an increased risk of non-Hodgkin's lymphoma in farmers who used this pesticide (35). Estrogenicity of carbamates has not been extensively investigated, although there is one report of inhibition of 17β-estradiol and progesterone activity in human breast and endometrial cancer cells (36).

Synthetic pyrethroids, common residential insecticides, have been found to possess estrogenic and antiprogestagenic properties in human breast cell assays (37, 38). These insecticides could have been included among categories of pesticides used for lawn and garden purposes for which an association with breast cancer was observed, such as problems on fruit trees, in vegetable gardens, and on outdoor as well as indoor plants. However, these insecticides could also have been used for nuisance-pest categories—such as ants, cockroaches, wasps, flies, mosquitoes, moths, silverfish, caterpillars, fleas, ticks, and termites—not associated with increased risks. Another insecticide, rotenone, possibly used by participants on fruit trees, vegetable gardens, and indoor plants, has been shown to cause mammary tumors in rats (39) but has also been shown to have anticancer action in human breast cell culture (40).

In LIBCSP, information was not collected on the time frame of pesticide use. This is a possible limitation because it has been proposed that exposures occurring between menarche and first birth may be the most influential in initiating breast cancer (41). However, we were not able to examine the effect of exposures occurring during a particular period of a woman's reproductive life cycle. Additionally, several of the individual pest categories included pests for which different pesticides would be used, and the same pesticide could have been used for several of the individual categories. For example, chlorpyrifos and diazinon could have been used for lawn insects as well as to treat ants and cockroaches. Because we did not ascertain the specific chemicals applied for any of the pest problems, we were not able to assign any observed increased risk to the use of a specific pesticide.

Many women reported a combination of pesticide applicators as well as multiple types of products but did not rank the frequency of use of a particular applicator or product type. Thus, distributing lifetime applications according to multiple reported persons or product types was not possible. Assigning an equal distribution (e.g., if both self and professional were reported, half of the lifetime applications would be attributed to self and half to the professional) assumed information beyond what we collected and was most likely unrepresentative. Therefore, to avoid introducing additional exposure misclassification, we categorized women as combination users or exclusive users of an applicator or type. Future investigations of patterns of pesticide use should consider asking women to rank their use of multiple applicators or product types.

The “lifetime applications” variable represents exposure opportunity rather than an actual exposure dose. The detailed information required to calculate a received dose of pesticides was not, nor could it be, ascertained by a questionnaire alone. The categorized lifetime application variable used as the measure of pesticide exposure in these analyses allowed women to be ranked according to their reported use and allowed for exploration of trends. The variable also enabled both reported frequency and number of years of use to be combined into a single exposure variable.

Finding no dose-dependent relation between exposure and breast cancer risk may be due to the imprecision of the exposure measure. It is possible that ever use of pesticides for individual categories was recalled accurately, but the details of use were not. Thus, the ranking of women according to their lifetime applications may have been incorrect, thereby masking any underlying association. Furthermore, the combination of frequency and duration may not be the optimal approach to assess residential pesticide exposure for some of the pesticide categories. For example, the vast majority of women had only one lifetime application of termite control pesticides; thus, frequency and duration have little meaning.

The reliability of self-reported lifetime residential exposure among the participants must be considered. The design of the questionnaire required women to integrate a considerable amount of information over their lifetime, resulting in potentially imprecise reporting of their pesticide exposures. The time interval between exposure and recall, the amount of detail required, age, how memorable the exposure was, and the social desirability of reporting the exposure may all influence the reproducibility of residential pesticide use recall (42). The hypothesis of an adverse association between pesticides and breast cancer was widely publicized, so it is likely that this heightened awareness influenced cases' reporting. This issue could have resulted in recall bias such that the observed associations are biased away from the null. However, data from our study indicate that cases and controls were equally likely to believe that environmental factors were associated with breast cancer etiology. Examination of self-reported beliefs about the cause of breast cancer revealed no case-control differences in the reporting of at least one environmental factor (69 percent vs. 68 percent, respectively), suggesting that recall was not biased differentially.

The study's design lends many strengths to this investigation, including 1) a population-based design, which allows for generalizability of the results to the population of Long Island, New York, as well as similar populations; 2) a large sample size, which increases the power to detect associations; and 3) a comprehensive, in-person, interviewer-administered questionnaire, which provides well-measured confounding variables. Another strength of this study is long-term Long Island residency; nearly 60 percent of both cases and controls were residents of their current home for at least 15 years. For these long-term residents, pest problems they have encountered are likely to be stable over time (e.g., ant problems in the spring, wasp problems in the summer) so that a woman may have performed the same pesticide application routine for many years, making recall easier. When analyses were restricted to these women, the results were essentially the same, which could indicate that reporting of exposures was not differentially affected by length of residency, or it may reflect that the exposure assessment was not sensitive enough to capture any differences that may have existed.

Not all eligible women participated, and response rates varied by age (9). If the pesticide exposure of women who participated is different from that of those women who did not but does not differ by case-control status, then nondifferential misclassification would occur, biasing the estimates of association between pesticide use and breast cancer toward the null. On the other hand, if the response difference between the cases and controls is somehow related to pesticide use, then differential misclassification would occur, and the estimate of the association could be biased in either direction. The results do not appear to be biased because of age-related nonresponse, since results of analyses restricted to participants less than 65 years of age were not materially different.

A great deal of detail about residential and personal pesticide use was ascertained, which allowed investigation of not only overall pesticide use but also patterns of use for specific pesticide groups. However, information on use of actual chemical products was not ascertained because it was found during pilot testing that women could not recall them. This limitation prevents identification of specific pesticides that require further investigation in relation to breast cancer risk.

To our knowledge, our study is the first to suggest that self-reported residential pesticide use may be associated with elevated breast cancer risk. However, the weak association, the absence of a dose response, the lack of support from studies of biologic measures of exposures, and the possibility of chance findings due to multiple comparisons all indicate the uncertain nature of the observed association. Further investigation in other populations is necessary before any definitive conclusions can be reached.

Abbreviations

    Abbreviations
  • CI

    confidence interval

  • LIBCSP

    Long Island Breast Cancer Study Project

  • OR

    odds ratio

This work was supported by grants from the National Institute of Environmental Health Sciences (NIEHS) (K01 ES12645 and P30ES10126), the National Cancer Institute (NCI)/NIEHS (5 U01CA/ES66572), and the Department of Defense (W81XWH-04-1-0507).

The authors acknowledge the invaluable contributions of the members of the Long Island Breast Cancer Network; the 31 participating institutions on Long Island and in New York City, New York; their collaborators, who assisted with various aspects of data collection, including Dr. Mary Wolff, Dr. Regina Santella, Dr. Geoffrey Kabat, Dr. Maureen Hatch, Gail Garbowski, Dr. H. Leon Bradlow, Martin Trent, Dr. Ruby Senie, Dr. Carla Maffeo, Pat Montalvan, Dr. Gertrud Berkowitz, Dr. Margaret Kemeny, Dr. Mark Citron, Dr. Freya Schnabel, Dr. Allen Schuss, Dr. Steven Hajdu, and Dr. Vincent Vinciguerra, as well as their National Institutes of Health collaborators: Dr. Gwen Collman of NIEHS and Dr. G. Iris Obrams, formerly of the NCI; and the members of the External Advisory Committee to the population-based case-control study: Dr. Leslie Bernstein (Committee chair), Gerald Akland, Barbara Balaban, Dr. Blake Cady, Dr. Dale Sandler, Dr. Roy Shore, and Dr. Gerald Wogan.

The authors also thank Drs. Mary Wolff and Bruce Link for their insightful comments and advice during the writing of this manuscript.

Conflict of interest: none declared.

References

1.
Dich
J
Zahm
SH
Hanberg
A
, et al.  . 
Pesticides and cancer
Cancer Causes Control
 , 
1997
, vol. 
8
 (pg. 
420
-
43
)
2.
Teitelbaum
SL
Questionnaire assessment of nonoccupational pesticide exposure in epidemiologic studies of cancer
J Expo Anal Environ Epidemiol
 , 
2002
, vol. 
12
 (pg. 
373
-
80
)
3.
Laden
F
Hunter
DJ
Environmental risk factors and female breast cancer
Annu Rev Public Health
 , 
1998
, vol. 
19
 (pg. 
101
-
23
)
4.
Occupational exposures in insecticide application, and some pesticides. IARC monographs on the evaluation of carcinogenic risks to humans
1991
, vol. 
Vol 53
 
Lyon, France
International Agency for Research on Cancer
5.
Overall evaluations of carcinogenicity: an updating of IARC monographs volumes 1 to 42. IARC monographs on the evaluation of carcinogenic risks to humans
1987
Suppl 7
Lyon, France
International Agency for Research on Cancer
6.
US Department of Health and Human Services. Report on carcinogens
Public Health Service and National Toxicology Program
 , 
1998
8th ed
Washington, DC
Environmental Health Information Service
7.
Adami
HO
Lipworth
L
Titus-Ernstoff
L
, et al.  . 
Organochlorine compounds and estrogen-related cancers in women
Cancer Causes Control
 , 
1995
, vol. 
6
 (pg. 
551
-
66
)
8.
Snedeker
SM
Diaugustine
RP
Hormonal and environmental factors affecting cell proliferation and neoplasia in the mammary gland
Prog Clin Biol Res
 , 
1996
, vol. 
394
 (pg. 
211
-
53
)
9.
Gammon
MD
Neugut
AI
Santella
RM
, et al.  . 
The Long Island Breast Cancer Study Project: description of a multi-institutional collaboration to identify environmental risk factors for breast cancer
Breast Cancer Res Treat
 , 
2002
, vol. 
74
 (pg. 
235
-
54
)
10.
Gammon
MD
Wolff
MS
Neugut
AI
, et al.  . 
Environmental toxins and breast cancer on Long Island. II. Organochlorine compound levels in blood
Cancer Epidemiol Biomarkers Prev
 , 
2002
, vol. 
11
 (pg. 
686
-
97
)
11.
Hosmer
DW
Lemeshow
S
Applied logistic regression
1989
New York, NY
John Wiley & Sons
12.
Kleinbaum
DG
Kupper
LL
Morgenstern
H
Epidemiologic research: principles and quantitative methods
1982
Belmont, CA
Lifetime Learning Publications
13.
Calle
EE
Frumkin
H
Henley
SJ
, et al.  . 
Organochlorines and breast cancer risk
CA Cancer J Clin
 , 
2002
, vol. 
52
 (pg. 
301
-
9
)
14.
New York State Department of Health (NYSDH)
Long Island Breast Cancer Study, report no. 4
Termiticide use and breast cancer risk
 , 
1992
Albany, NY
NYSDH
15.
Lopez-Carrillo
L
Lopez-Cervantes
M
Torres-Sanchez
L
, et al.  . 
Serum levels of beta-hexachlorocyclohexane, hexachlorobenzene and polychlorinated biphenyls and breast cancer in Mexican women
Eur J Cancer Prev
 , 
2002
, vol. 
11
 (pg. 
129
-
35
)
16.
Dorgan
JF
Brock
JW
Rothman
N
, et al.  . 
Serum organochlorine pesticides and PCBs and breast cancer risk: results from a prospective analysis (USA)
Cancer Causes Control
 , 
1999
, vol. 
10
 (pg. 
1
-
11
)
17.
Hoyer
AP
Grandjean
P
Jorgensen
T
, et al.  . 
Organochlorine exposure and risk of breast cancer
Lancet
 , 
1998
, vol. 
352
 (pg. 
1816
-
20
)
18.
Wiklund
K
Swedish agricultural workers: a group with a decreased risk of cancer
Cancer
 , 
1983
, vol. 
51
 (pg. 
566
-
8
)
19.
Olsen
JH
Jensen
OM
Occupation and risk of cancer in Denmark. An analysis of 93,810 cancer cases, 1970 –1979
Scand J Work Environ Health
 , 
1987
, vol. 
13
 (pg. 
1
-
91
)
20.
Ewertz
M
Risk of breast cancer in relation to social factors in Denmark
Acta Oncol
 , 
1988
, vol. 
27
 (pg. 
787
-
92
)
21.
Kato
I
Tominaga
S
Ikari
A
An epidemiological study on occupation and cancer risk
Jpn J Clin Oncol
 , 
1990
, vol. 
20
 (pg. 
121
-
7
)
22.
Franceschi
S
Barbone
F
Bidoli
E
, et al.  . 
Cancer risk in farmers: results from a multi-site case-control study in north-eastern Italy
Int J Cancer
 , 
1993
, vol. 
53
 (pg. 
740
-
5
)
23.
Rubin
CH
Burnett
CA
Halperin
WE
, et al.  . 
Occupation as a risk identifier for breast cancer
Am J Public Health
 , 
1993
, vol. 
83
 (pg. 
1311
-
15
)
24.
Costantini
AS
Pirastu
R
Lagorio
S
, et al.  . 
Studying cancer among female workers: methods and preliminary results from a record-linkage system in Italy
J Occup Med
 , 
1994
, vol. 
36
 (pg. 
1180
-
6
)
25.
Wiklund
K
Dich
J
Cancer risks among female farmers in Sweden
Cancer Causes Control
 , 
1994
, vol. 
5
 (pg. 
449
-
57
)
26.
Cantor
KP
Stewart
PA
Brinton
LA
, et al.  . 
Occupational exposures and female breast cancer mortality in the United States
J Occup Environ Med
 , 
1995
, vol. 
37
 (pg. 
336
-
48
)
27.
Kristensen
P
Andersen
A
Irgens
LM
, et al.  . 
Incidence and risk factors of cancer among men and women in Norwegian agriculture
Scand J Work Environ Health
 , 
1996
, vol. 
22
 (pg. 
14
-
26
)
28.
Pukkala
E
Notkola
V
Cancer incidence among Finnish farmers, 1979 –93
Cancer Causes Control
 , 
1997
, vol. 
8
 (pg. 
25
-
33
)
29.
Band
PR
Le
ND
Fang
R
, et al.  . 
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
J Occup Environ Med
 , 
2000
, vol. 
42
 (pg. 
284
-
310
)
30.
Duell
EJ
Millikan
RC
Savitz
DA
, et al.  . 
A population-based case-control study of farming and breast cancer in North Carolina
Epidemiology
 , 
2000
, vol. 
11
 (pg. 
523
-
31
)
31.
Engel
LS
Hill
DA
Hoppin
JA
, et al.  . 
Pesticide use and breast cancer risk among farmers' wives in the Agricultural Health Study
Am J Epidemiol
 , 
2005
, vol. 
161
 (pg. 
121
-
35
)
32.
US Environmental Protection Agency
Integrated Risk Information System: dichlorvos
2006
Washington, DC
Environmental Protection Agency
 
(http://www.epa.gov/iris/subst/0151.htm). (Accessed March 13, 2006).
33.
Andersen
HR
Vinggaard
AM
Rasmussen
TH
, et al.  . 
Effects of currently used pesticides in assays for estrogenicity, androgenicity, and aromatase activity in vitro
Toxicol Appl Pharmacol
 , 
2002
, vol. 
179
 (pg. 
1
-
12
)
34.
Kang
HG
Jeong
SH
Cho
JH
, et al.  . 
Chlorpyrifos-methyl shows anti-androgenic activity without estrogenic activity in rats
Toxicology
 , 
2004
, vol. 
199
 (pg. 
219
-
30
)
35.
Zheng
T
Zahm
SH
Cantor
KP
, et al.  . 
Agricultural exposure to carbamate pesticides and risk of non-Hodgkin lymphoma
J Occup Environ Med
 , 
2001
, vol. 
43
 (pg. 
641
-
9
)
36.
Klotz
DM
Arnold
SF
McLachlan
JA
Inhibition of 17 beta-estradiol and progesterone activity in human breast and endometrial cancer cells by carbamate insecticides
Life Sci
 , 
1997
, vol. 
60
 (pg. 
1467
-
75
)
37.
Go
V
Garey
J
Wolff
MS
, et al.  . 
Estrogenic potential of certain pyrethroid compounds in the MCF-7 human breast carcinoma cell line
Environ Health Perspect
 , 
1999
, vol. 
107
 (pg. 
173
-
7
)
38.
Garey
J
Wolff
MS
Estrogenic and antiprogestagenic activities of pyrethroid insecticides
Biochem Biophys Res Commun
 , 
1998
, vol. 
251
 (pg. 
855
-
9
)
39.
Kamrin
MA
Pesticide profiles: toxicity, environmental impact, and fate
1998
New York, NY
Lewis Publishers
40.
Fang
N
Casida
JE
Anticancer action of cube insecticide: correlation for rotenoid constituents between inhibition of NADH:ubiquinone oxidoreductase and induced ornithine decarboxylase activities
Proc Natl Acad Sci U S A
 , 
1998
, vol. 
95
 (pg. 
3380
-
4
)
41.
Rosner
B
Nurses'
Colditz GA
Health Study: log-incidence mathematical model of breast cancer incidence
J Natl Cancer Inst
 , 
1996
, vol. 
88
 (pg. 
359
-
64
)
42.
Coughlin
SS
Recall bias in epidemiologic studies
J Clin Epidemiol
 , 
1990
, vol. 
43
 (pg. 
87
-
91
)