Breast Cancer Risk after Occupational Solvent Exposure: the Influence of Timing and Setting

Organic solvents are chemicals that are ubiquitous in occupational settings. Solvents are present in the workplace through the use of adhesives, paints, degreasing agents, and cleaning products. Solvents are also used in industrial processes, including the manufacturing of computer components, plastics, pharmaceuticals, and textiles. In the United States, millions of workers are exposed to solvents on a daily basis (1). Solvents such as benzene and trichloroethylene have been recognized by the International Agency for Research on Cancer and the US National Toxicology Program as known human carcinogens; however, there are limited data on solvents as human breast carcinogens (2, 3).

Animal studies have provided strong evidence for an association between organic solvents and breast cancer (4). Benzene, methylene chloride, styrene, and vinyl chloride have been shown to cause mammary tumors in mice (5), and different mechanisms for breast carcinogenesis have been hypothesized. Solvents are lipophilic compounds that can accumulate in the adipose tissue of the breast where they are stored, biotransformed, and excreted into the parenchyma of the breast. If present in the parenchyma for a considerable amount of time, solvents or their metabolites can initiate or promote carcinogenesis through genotoxic mechanisms (6). Several chemicals used as solvents have exhibited endocrine-disrupting properties (7–9). Endocrine-disrupting chemicals can bind to estrogen receptors, disrupt estrogen-mediated pathways, and alter gene expression, thus making the mammary gland more vulnerable to tumor cell proliferation (10, 11).

Several epidemiologic studies have reported associations between occupational exposure to solvents and increased breast cancer risk (12–17). In general, these studies have assessed risk associated with overall lifetime exposures. A few studies have observed elevated breast cancer risks among women who were occupationally exposed to solvents at young ages (12, 17), suggesting that early solvent exposure, before full differentiation of breast tissue, may result in increased vulnerability for breast cancer.

In this study, we examined the relationship between occupational exposure to solvents and breast cancer among women enrolled in the Sister Study, a large prospective cohort of women from the United States and Puerto Rico who have a family history of breast cancer. We used detailed occupational histories to study the risk of breast cancer among women who were occupationally exposed to solvents before their first full-term birth, a critical period of breast tissue differentiation. To study the impact of occupation-specific exposures, we also examined the relationship between self-reported exposure to solvents and breast cancer risk by occupation.

The Sister Study is a prospective cohort study of environmental and genetic risk factors for breast cancer (18). The cohort consists of 50,884 initially breast cancer–free sisters of women who had been diagnosed with breast cancer. Participants enrolled in the study between 2003 and 2009. Baseline activities included a computer-assisted telephone interview and self-administered questionnaires that included questions about suspected risk factors for breast cancer. The Sister Study was approved by the institutional review boards at the National Institute of Environmental Health Sciences and Copernicus Group. Written informed consent was provided by study participants.

Lifetime occupational history and self-reported information on occupational exposures at each job were collected during the baseline interview. After excluding a vanguard group of women with incomplete occupational data (n = 2,297), women without an occupational history (n = 832), and women who were diagnosed with breast cancer before the baseline interview (n = 94), the present analysis included 47,661 women who reported ever working at least one job. For each job held after the age of 18 years, study participants were asked whether they had worked with “solvents, degreasers, or other cleaning agents” (hereafter referred to as “solvents”) on the job and the first and last date of employment for each job.

In addition to lifetime solvent job exposure, several measures of exposure were determined for each participant who reported working with solvents: the duration of solvent job exposure in years, the weekly frequency of exposure to solvents, the time period of first solvent job, the age at first solvent job, and the time in years since last solvent job at baseline interview. For parous women, we determined the timing of first solvent job relative to first full-term birth, and for those women who reported working at a solvent job before their first full-term birth, we determined the duration of solvent job exposure before the birth.

The most commonly reported solvent-exposed occupations (minimum 50 study participants with self-reported solvent exposure) were classified and assigned into the following major occupational categories according to the 2010 United States Standard Occupational Classification system: building and grounds cleaning and maintenance; education, training, and library; food preparation and serving related; healthcare practitioner and technical; management; office administrative and support; and production occupations.

Study participants reported incident breast cancer diagnoses via annual health updates and biennial self- or telephone-administered questionnaires. If a study participant reported a breast cancer diagnosis, we followed up with the participant about diagnostic details such as tumor subtype and estrogen receptor status and requested permission to contact their physician for medical records and pathology reports. We used data from self-report breast cancer characteristics only when medical record or pathology report data were not available. Participants were followed prospectively from baseline until date of invasive breast cancer diagnosis. Women reporting in situ breast cancer were censored at time of diagnosis. Women lost to follow-up or deceased from other causes were censored as appropriate.

HRs and 95% confidence intervals (CI) for invasive breast cancer and estrogen receptor tumor subtypes were estimated using multivariable Cox proportional hazards regressions models with the study participant's age as the time scale. We used directed acyclic graphs (DAG) to identify potential confounders (19). Briefly, through a review of the literature, we identified potential confounders that were associated with both occupational solvent exposure and breast cancer but were not direct descendants of both or in the causal pathway (20, 21). The final multivariable model adjusted for race/ethnicity (non–Hispanic White, Black, Hispanic, other), education (<high school graduate/GED, high school graduate/GED, some college, college/post graduate), income (<$50,000, $50,000 to <$100,000, $100,000+), parity (nulliparous, 1, 2, 3+ births), and age at first birth (<21, 21 to <24, 24 to <28, 28+). Where appropriate, trends in incidence were examined across exposure categories.

To explore exposures within occupations, proportional hazards regressions analyses were carried out separately for each occupation with women who had never worked in the occupation as a reference group. For each occupation, HRs were determined for two groups: solvent-exposed women and solvent-unexposed women. The statistical software package SAS (version 9.3, SAS Institute) was used for all analyses.

A total of 1,798 breast cancer diagnoses were reported during the follow-up period (mean follow-up of 4.7 years). Among women with available information about tumor subtypes, approximately 70% (n = 1,255) of breast cancer tumors were invasive and 77% (n = 968) of invasive tumors were estrogen receptor positive. Characteristics of invasive breast cancer cases and the remainder of the cohort are displayed in Table 1. Women diagnosed with invasive breast cancer were more likely than non-cases to be older, non-Hispanic White, have higher educational attainment and income, and have a later age at first-term pregnancy.

The HRs for invasive breast cancer associated with solvent job exposure are shown in Table 2. Overall, there was no increased risk of invasive breast cancer among women exposed occupationally to solvents (HR, 1.04; 95% CI, 0.88–1.24). Longer duration of work and more frequent work with solvents and in solvent jobs were not associated with an increased risk of breast cancer. The breast cancer risk for women whose first solvent job was before 1980 was not different from the risk among women who first solvent job was in 1980 or later (a time period after regulatory interventions in the late 1970s). In latency analyses, the risk for women whose last solvent job was 5 or more years before the baseline interview was not significantly different from women who were recently exposed to solvents.

We evaluated the risk of invasive breast cancer associated with occupational solvent exposure for different periods of initial exposure. Age of first solvent exposure was not associated with breast cancer risk. Women who were exposed to solvents before their first birth had a nonsignificant increased risk of breast cancer (HR, 1.24; 95% CI, 0.95–1.63). Among parous women, there was a significant positive trend for duration of solvent work before first birth and breast cancer (P_trend = 0.04).

Results by estrogen receptor status are presented in Table 3. The overall association between any occupational exposure to solvents and invasive breast cancer incidence was elevated, but nonsignificant, for estrogen receptor–positive tumors (HR, 1.15; 95% CI, 0.95–1.39). Estrogen receptor–positive breast cancer was not associated with either frequency or duration of solvent exposure. There was, however, an increased risk for estrogen receptor–positive breast cancer among women whose first solvent job was before 1980 (HR, 1.28; 95% CI, 1.01–1.62).

Compared with women who were never exposed to solvents, women who were exposed to solvents before their first birth had an increased risk of estrogen receptor–positive breast cancer (HR, 1.39; 95% CI, 1.03–1.86). Women who started working in solvent jobs after their first birth (HR, 1.05; 95% CI, 0.79–1.40) and nulliparous women with solvent job exposure (HR, 1.03; 95% CI, 0.69–1.54) did not have an increased risk for estrogen receptor–positive breast cancer. Parous women who were exposed to solvents for more than 5 years before their first birth had a borderline excess risk for developing estrogen receptor–positive breast cancer compared with women who were never exposed to solvents (HR, 1.56; 95% CI, 0.97–2.50). No statistically significant associations were observed between solvent exposure and estrogen receptor–negative breast cancer risk.

A total of 44 occupations had 50 or more women who were exposed to solvents. The results for women in occupations with a minimum of five breast cancer cases are presented in Supplementary Table. A significant increase in estrogen receptor–positive breast cancer risk was observed among clinical laboratory technologists and technicians who worked with solvents (HR, 2.00; 95% CI, 1.07–3.73). Nonsignificant elevated risks were also observed for solvent-exposed maids and housekeeping cleaners (HR, 1.82; 95% CI, 0.90–3.67) and solvent-exposed women in production occupations (HR, 1.66; 95% CI, 0.86–3.22). Additional adjustment for alcohol consumption, environmental tobacco smoke exposure, smoking, and work at night did not have an impact on occupation-specific point estimates (data not shown).

In this large cohort of women, there was no overall association between occupational exposure to solvents and invasive breast cancer risk. We observed an elevated risk of estrogen receptor–positive tumors among women who were exposed to solvents before their first full-term birth, independent of known confounding factors. Our findings support the hypothesis that chemical exposure before breast tissue differentiation may be a risk factor for breast cancer.

This study examined the association between an occupational exposure before first birth and breast cancer. An increased risk associated with pre-birth occupational exposure to solvents is biologically plausible. The years before the first birth of a child have been identified as a critical period of exposure for breast tissue susceptibility (22). Furthermore, studies of women exposed to cigarette smoke (23–25), DDT (26), ionizing radiation (27, 28), and traffic-related polycyclic aromatic hydrocarbons (29) have reported elevated risks among younger women, providing evidence that developing breast tissue may be more vulnerable to potential carcinogens. We did not observe an association between solvent exposure and breast cancer among nulliparous women. Our results suggest that solvent exposure before and in combination with the proliferative changes that occur during pregnancy may play a role in breast cancer development.

We are aware of only one other study that reported an association between occupational chemical exposures before first full-term pregnancy and breast cancer risk. In a recent case–control study with 1,005 cases, Brophy and colleagues observed an increased risk for Canadian women who worked with automotive plastics between menarche and the first full-term pregnancy (30). Two previous case–control studies of women occupationally exposed to solvents have reported increased breast cancer risks for women exposed at young ages. A Danish study reported an elevated risk of breast cancer among women aged 55 years or younger in solvent-using industries (12), and a Canadian study reported an increased risk for breast cancer among women who worked with solvents before the age of 36 (17). Our study is the first with a cohort design to report an association between early solvent exposure and breast cancer risk.

Null results for lifetime solvent exposure may reflect our assessment of solvents as a group of chemicals, rather than individual compounds or mixtures. Occupational studies of breast cancer risk in women have reported significant associations with several types of solvents. A significant exposure–response breast cancer trend was observed in a cohort of Finnish women exposed to aromatic hydrocarbon solvents (13). In a large study of active duty military women, women in occupations with moderate high volatile organic compound (VOC) exposure had significantly higher incidence rates for breast cancer compared with low unexposed VOC exposure (31). Trichloroethylene, but not any of the other organic solvents assessed in a cohort of civilian maintenance workers, was associated with a nonsignificant excess risk for breast cancer (32).

Several studies have reported inconsistent associations between occupational solvent exposure and breast cancer by hormone receptor status. One study, composed of 2,383 incident cases, found that solvent exposure was significantly associated with 120 estrogen and progesterone receptor–negative tumors (16). Others have reported no difference in risk among hormone receptor types associated with lifetime solvent (17) and benzene (16) exposure. A case–control study with 556 cases reported increased risks for estrogen receptor–positive and progesterone receptor–negative tumors in women who were exposed before the age of 36 years (17) and another study of only 56 cases and 35 controls found a greater risk for estrogen receptor–positive tumors among premenopausal women exposed to benzene (33). We had limited power to evaluate results jointly by both estrogen receptor and progesterone receptor status in this study; however, risk estimates tended to be higher for estrogen receptor–positive tumors than for estrogen receptor–negative tumors.

We identified occupations that had elevated breast cancer risks associated with solvent exposure. A significantly elevated risk for estrogen receptor–positive breast cancer was associated with solvent exposure among clinical laboratory technologists and technicians. Studies of breast cancer among laboratory workers have produced inconsistent results, primarily due to incomplete information about reproductive and lifestyle characteristics, but also because of incomplete information about workplace exposures. In addition to laboratory solvents, clinical laboratory workers may be exposed to elevated levels of solvents and degreasers through routine tasks like applying topical cleansers and antiseptics, sterilizing instruments, and other general cleaning activities. Other possible explanations for the observed excess risk could include inadequate personal protection measures or unmeasured confounding due to exposure to potentially carcinogenic or endocrine-disrupting compounds in the workplace.

Solvent-exposed maids and housekeeping cleaners had an elevated, nonsignificant risk for breast cancer. Although shift work and work at night is common in this population, only those reporting solvent exposure had an increased breast cancer risk; thus, irregular work schedule is an unlikely explanation for our observation. Maids and housekeeping cleaners are an understudied population that may be at risk for breast cancer. Women in this occupation are exposed to a variety of chemicals, including solvent-containing cleaning products and disinfectants. Cleaning product use has been associated with increased breast cancer risk in women (34).

A nonsignificant increase in breast cancer risk was also observed among solvent-exposed women in production occupations. Previous epidemiologic studies have reported elevated breast cancer risks in manufacturing occupations such as automotive plastics manufacturing, rubber and plastic manufacturers, metal work, and textile workers (30, 31, 35). Women in these settings may have had concurrent exposure to solvents, metals, dusts, and other potential carcinogens. This study, like many previous occupational studies, was unable to evaluate the impact of chemical mixtures on breast cancer risk.

The major strengths of this study include its large sample size, comprehensive occupational history data, and extensive information on important breast cancer covariates. We were able to prospectively examine the association between occupational solvent exposures and breast cancer risk, adjusting for established demographic, socioeconomic, and reproductive risk factors. This study had a sizeable proportion of working women with detailed individual information about exposure to solvents, and we used these data to evaluate the exposure–response relationship between duration of occupational solvent exposure and breast cancer incidence. In addition, we were able to evaluate the impact of solvent exposure during a critical period of breast development in women. The results of this study support a more biologically based exposure assessment approach to studies of chemical exposures and breast cancer.

Our study has several potential limitations associated with exposure assessment. We assessed solvent exposure via self-report, and participants may have had difficulty recalling solvent exposures in their workplace. Any misclassification of lifetime solvent exposure, however, is likely to be nondifferential, thus leading to an underestimation of effect. Similarly, unmeasured household or environmental solvent exposures would not explain the elevated risks we observed as it is unlikely that residential or environmental exposures differed by occupational solvent exposure status.

We combined self-reported exposures and occupational coding of job titles to create a composite index of solvent exposure and identify occupations that had elevated breast cancer risks associated with solvent exposure. Although this composite measure reduced the potential for solvent exposure misclassification, our results by occupation should be interpreted with caution as they involved multiple comparisons and relatively small numbers of breast cancer cases for some occupations. Adjusting for alcohol use, smoking, environmental tobacco smoke, and work at night did not alter our occupation-specific results. Nevertheless, because these data were based on retrospective self-reports, we cannot fully rule out the potential confounding effects of these risk factors or other unmeasured socioeconomic factors that may act as negative confounders. Also, despite the size of our cohort, some occupational groups with elevated solvent-exposed risks may not have been identified because of small group sizes and low exposure prevalence.

In this study, we did not control for simultaneous exposures to other potential occupational carcinogens in the workplace that could have been associated with breast cancer. Additional exposure studies focusing on the high-risk occupations identified in this study may yield new insights into the types of solvents used across different occupational settings and the role of other occupational exposures as carcinogens potentially associated with breast cancer.

Our results suggest that occupational exposure to solvents before first full-term birth is associated with an increased risk of breast cancer. Further large-scale epidemiologic studies are warranted to explore the relationship between the timing of occupational chemical exposures and breast cancer incidence.

No potential conflicts of interest were disclosed.

Conception and design: C.C. Ekenga, D.P. Sandler

Development of methodology: C.C. Ekenga, D.P. Sandler

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L.A. DeRoo, D.P. Sandler

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C.C. Ekenga, C.G. Parks, A.A. D'Aloisio, L.A. DeRoo

Writing, review, and/or revision of the manuscript: C.C. Ekenga, C.G. Parks, A.A. D'Aloisio, L.A. DeRoo, D.P. Sandler

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.A. D'Aloisio, L.A. DeRoo

Study supervision: C.G. Parks, D.P. Sandler

This research was supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (Z1A ES044005).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.