Abstract
Background: There is increasing concern that environmental exposures, such as air pollution, may be related to increasing rates of breast cancer; however, results from cohort studies have been mixed. We examined the association between particulate matter (PM) and measures of distance to roadway with the risk of incident breast cancer in the prospective nationwide Nurses' Health Study II (NHSII) cohort.
Methods: Incident invasive breast cancer from 1993 to 2011 (N = 3,416) was assessed among 115,921 women in the NHSII cohort. Time-varying Cox proportional hazards models were used to calculate HRs and 95% confidence intervals (95% CI) for increases in ambient exposures to PM10, PM2.5–10, and PM2.5 and residential roadway proximity categories.
Results: In multivariable adjusted models, there was little evidence of an increased risk of breast cancer (or any of the receptor-specific subtypes) overall or by menopausal status with PM exposure. There was, however, a suggestion of increased risks among women living <50 m of the largest road type (HR = 1.60; 95% CI, 0.80–3.21) or within <50 m of the two largest road types (1.14; 95% CI, 0.84–1.54) compared with women living farther (≥200 m) away.
Conclusions: Among women in the NHSII, we found no statistically significant associations between particulate matter exposures and incidence of breast cancer overall, by menopausal status, or by hormone receptor subtype. There was, however, a suggestion that residential proximity to major roadways may be associated with increased risk.
Impact: These results suggest no elevated breast cancer risk with increasing exposures to particulate matter air pollution, but that other traffic-related exposures may be important. Cancer Epidemiol Biomarkers Prev; 25(8); 1274–6. ©2016 AACR.
Introduction
The International Agency for Research on Cancer monograph on the carcinogenicity of ambient air pollution presented mixed evidence for air pollution and breast cancer risk (1). However, some recent studies have suggested associations (2–7). Studies with large numbers of breast cancer cases, information on specific breast cancer subtypes, and the ability to control for other breast cancer risk factors are needed. Therefore, we examined the association of particulate matter (PM) exposures and roadway proximity on risk of overall, hormone receptor–specific, and menopausal status–specific breast cancer risk among women in the nationwide, prospective Nurses' Health Study II (NHSII) cohort.
Materials and Methods
The NHSII is a cohort of 116,430 female nurses with no previous history of cancer (except nonmelanoma skin cancer) enrolled in 1989 when they were aged 25 to 42 years. Mailed biennial follow-up questionnaires update information on risk factors and medical history. Self-reports of incident cancers are confirmed by medical record review.
Biennial residential addresses were geocoded, and roadway proximity was assigned as a proxy for traffic-related exposures. Specifically, we calculated the distance from each address to the nearest street segments in the ESRI StreetPro 2007 data layer listed as U.S. Census Feature class codes A1, A2, or A3. On the basis of case distributions and exposure studies, we categorized distance to road as 0 to 49 m, 50 to 199 m, and ≥200 m, for all three road types together, for the two largest road types (A1, A2) and for the largest road type (A1). Forty-eight–month moving average and cumulative average exposures to three size fractions of particulate matter, PM10, PM2.5–10, and PM2.5, were assigned to each participant based on monthly spatiotemporal prediction models applied to the geocoded address history (8). For participants who moved, we assumed they changed addresses at the beginning of the biennial cycle.
Time-varying Cox proportional hazards models on a biennial time scale were used to calculate HRs and 95% confidence intervals (95% CI) for each roadway proximity category compared with the farthest, and for each 10 μg/m3 increase in each of the PM size fractions, after assessing linearity using splines. All models were adjusted for age, race, and calendar year; multivariable models were additionally adjusted for established breast cancer risk factors. Separate models assessed risk of overall breast cancer, and hormone receptor–specific subtypes (ER+/PR+ and ER−/PR−). We also stratified models by menopausal status.
Results
Eligible participants (N = 115,921) were 47 years old on average during follow-up, with an average body mass index (BMI) of 26 kg/m2. Sixteen percent were nulliparous, 10% were current users of postmenopausal hormone therapy, and 64% were never smokers. There was no association between exposures to 48-month PM with breast cancer overall or for specific hormone receptor subtypes, or by menopausal status (Tables 1 and 2); estimates were similar for cumulative average exposures (not shown). Although case numbers were small, there was a suggestion of increasing risk with residential proximity to the two largest road types, with a multivariable adjusted HR of 1.60 (95% CI, 0.80–3.21) for women living <50 m of the largest road type and 1.14 (95% CI, 0.84–1.54) for women closest to the two largest road types compared with those living ≥200 m away. These risks were consistent across the various breast cancer subtypes
Risk of overall and hormone receptor subtype–invasive breast cancer with exposures to particulate matter and roadway proximity among 115,921 women in the NHSII followed 1993–2011
Risk of invasive breast cancer by menopausal status with exposures to particulate matter and roadway proximity among women in the NHSII
Discussion
We did not observe an association between PM exposures and breast cancer incidence in this prospective cohort study. This is in contrast to an earlier case–control study in Western New York, where total suspended particulate exposures were associated with increased odds of postmenopausal breast cancer, but not premenopausal breast cancer (detailed in ref. 1). Our results are similar to a recent analysis in the Sister Study, where no elevations in overall or hormone receptor–specific breast cancer risk were observed with increasing PM10 or PM2.5 exposures (2).
Our suggestive findings of an elevated risk with roadway proximity, although not statistically significant and based on small numbers of exposed cases, are in contrast to several studies that have found no association (1, 3) but do agree with some studies using other measures of traffic exposure (traffic density, motor vehicle density; refs. 1, 7). Our findings also agree with numerous studies that have observed positive associations between exposures to nitrogen oxides or other traffic-related pollutants and breast cancer (1–3, 5–7).
Our analyses in a large, nation-wide, prospective cohort has many strengths, including fine control for potential confounders, high spatial and temporal resolution estimates of PM exposure, and large numbers of cases in many subcategories of interest (e.g., premenopausal, ER−/PR−). However, we only had information on adult exposures, which may not be an important etiologic period. The findings in this cohort may not be generalizable to populations with more racial/ethnic diversity or a broader range of socioeconomic status. Future studies should examine the association of traffic-related exposures in these groups.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: J.E. Hart, K.A. Bertrand, V.M. Vieira, R.M. Tamimi, F. Laden
Development of methodology: J.E. Hart, R.M. Tamimi, F. Laden
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.E. Hart, F. Laden
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.E. Hart, K.A. Bertrand, R.M. Tamimi, F. Laden
Writing, review, and/or revision of the manuscript: J.E. Hart, K.A. Bertrand, N. DuPre, P. James, V.M. Vieira, R.M. Tamimi, F. Laden
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R.M. Tamimi
Study supervision: R.M. Tamimi, F. Laden
Grant Support
All authors were funded by Susan G. Komen for the Cure grant IIR13264020. The NHSII cohort is funded by NIH grant UM1 CA176726. J.E. Hart and F. Laden received R01 ES017017, J.E. Hart received P30 ES000002, and P. James and N. DuPre were supported by T32 CA09001.
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.
Acknowledgments
The authors thank the participants and staff of the Nurses' Health Study II and following state cancer registries: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY.
Footnotes
R.M. Tamimi and F. Laden share senior authorship of this article.
- Received March 24, 2016.
- Accepted May 10, 2016.
- ©2016 American Association for Cancer Research.
Volume 25, Issue 8
