Abstract
Background: Environmental tobacco smoke (ETS) exposure is hypothesized to influence survival after breast cancer, but few studies have examined this association.
Methods: A population-based cohort of women (N = 1,508) diagnosed with first primary invasive or in situ breast cancer in 1996 to 1997 was interviewed shortly after diagnosis and again approximately 5 years later to assess ETS exposure, and women were followed for more than 18 years using the National Death Index; 597 deaths (237 associated with breast cancer) were identified. Multivariable Cox regression was used to estimate adjusted HRs and 95% confidence intervals (CI) for mortality among women with breast cancer as related to at-diagnosis and at-/postdiagnosis changes in ETS exposure.
Results: There was little or no association between at-diagnosis ETS exposure and all-cause (HR = 1.04; 95% CI, 0.78–1.40) or breast cancer–specific (HR = 0.98; 95% CI, 0.63–1.52) mortality. Mortality was elevated among women who reported cessation in postdiagnosis ETS exposure up to 1 year before the follow-up assessment, for all-cause (HR = 1.81; 95% CI, 0.87–3.74) and breast cancer mortality (HR = 1.89; 95% CI, 0.68–5.24); however, estimates were imprecise.
Conclusions: We found little evidence of an association between at-diagnosis ETS exposure and mortality after breast cancer. Postdiagnosis cessation of ETS exposure was positively associated with mortality, although we could not rule out chance and reverse causation as possible explanations.
Impact: Exposure to ETS does not appear to influence mortality after breast cancer. Cancer Epidemiol Biomarkers Prev; 26(2); 278–80. ©2016 AACR.
Introduction
Few studies (1–4) have examined whether environmental tobacco smoke (ETS) exposure increases the risk of mortality among women with breast cancer, and no studies to date have prospectively examined the impact of postdiagnosis changes in ETS exposure on mortality. This study examined whether ETS exposure was associated with long-term all-cause and breast cancer–specific mortality among a population-based sample of women.
Materials and Methods
Participants of the Long Island Breast Cancer Study Project (LIBCSP), a population-based cohort of women newly diagnosed with breast cancer, were interviewed shortly after diagnosis and again about 5 years later and now continue to be followed for vital status. Details of the LIBCSP have been published previously (2, 5). Institutional Review Board approval was obtained from of all participating institutions.
ETS exposure assessment
ETS exposure was determined via structured interviews (2). Women were asked to report whether any members of the household smoked in their presence, the relationship of the smoker, the participant's ages at first/last exposure, and any time periods the household member did not smoke. Duration of exposure was categorized as <15 years, ≥15 to <30 years, and ≥30 years of exposure. Recency of exposure was categorized as <5 years, ≥5 to <10 years, and ≥10 years.
Covariate assessment
Covariates assessed via questionnaire included age, menopausal status, annual household income, education, marital status, body mass index (BMI), physical activity, intake of alcoholic beverages, cigarette smoking, and treatment. Estrogen receptor status and nodal involvement were determined by medical record. Tumor size was obtained from the NY State Cancer Registry.
Outcome assessment
Vital status of the 1,508 women diagnosed with breast cancer was determined using the National Death Index. Follow-up for mortality occurred from the date of diagnosis in 1996 to 1997 until December 31, 2014 (median = 17.61 years). We identified 597 deaths; 234 were associated with breast cancer.
Statistical analysis
Using multivariable Cox proportional regression models, we estimated HR and 95% confidence intervals (CI) for the associations between at-diagnosis as well as at-/postdiagnosis changes in ETS exposure and mortality following breast cancer. Models restricted to women with invasive cancer only yielded similar results from those of all women; only the latter are shown. All analyses were conducted using IBM SPSS Statistics Version 22.0 (IBM Corp.).
In analyses of at-diagnosis ETS exposure, survival time began at the date of breast cancer diagnosis and ended on the date of death or, if alive, December 31, 2014. In analyses examining postdiagnosis ETS exposure, survival time began at the date of completion of the follow-up questionnaire and ended on the date of death or, if alive, December 31, 2014. Missing covariates were imputed in SPSS using 25 imputations with 1,000 iterations. The imputation models included age at diagnosis, menopausal status, income, education, marital status, BMI, physical activity, alcohol intake, smoking status, post-diagnosis ETS exposure, disease characteristics (stage, tumor size, nodal involvement estrogen receptor status), treatment (radiation, chemotherapy, and hormone therapies), and the outcome (the event indicator and the Nelson–Aalen cumulative hazard estimator).
Results
Approximately 15% of women reported ETS exposure in the year before diagnosis and 14% reported current exposure at the follow-up questionnaire.
At-diagnosis ETS exposure
There was little or no association between current ETS exposure and all-cause (HR = 1.04; 95% CI, 0.78–1.40) or breast cancer–specific (HR = 0.98; 95% CI, 0.63–1.52) mortality after adjustment for covariates (Table 1). Risk of mortality was slightly elevated for all-cause (HR = 1.17; 95% CI, 0.74–1.86) and breast cancer–specific (HR = 1.13; 95% CI, 0.57–2.27) mortality when we restricted the analyses to never smokers, although the corresponding estimates were imprecise.
Cox regression HRs and 95% CIs for the association between prediagnosis and at-diagnosis ETS exposure and mortality in the LIBCSP women diagnosed with breast cancer in 1996–1997 (N = 1,508)
At-/postdiagnosis ETS exposure
Although no associations were observed among women with ongoing ETS exposure, HRs were elevated 81% (HR = 1.81; 95% CI, 0.87–3.74) for all-cause mortality and 89% (HR = 1.89; 95% CI, 0.68–5.24) for breast cancer–specific mortality among women who reported cessation in postdiagnosis ETS exposure up to the year before the follow-up assessment (Table 2).
Cox regression HRs and 95% CIs for the association between at-/postdiagnosis ETS exposure and mortality in the LIBCSP women diagnosed with breast cancer in 1996–1997 (N = 1,339).
Discussion
Exposure to the constituents of tobacco smoke, either through active smoking or exposure to ETS, is hypothesized to influence breast cancer progression through several mechanisms, including directly by influencing cell proliferation and metastasis (6) and indirectly by disrupting the endocrine system (7). In addition, because up to 70% of tar in ETS is in the vapor phase, whereas all of the tar in direct smoking is in the particulate phase, ETS may be an important source of exposure to carcinogens as particulate smoke is cleared into the mouth and swallowed, but vapor phase constituents are inhaled and absorbed into the bloodstream and lymph system (8). Despite these hypothesized mechanisms, the few studies conducted to date (1–4), including the sufficiently powered study reported here, provide limited or no evidence of an association between ETS exposure and survival after breast cancer. Although we observed an elevated risk of mortality among women with postdiagnosis cessation of ETS exposure, we could not rule out chance and reverse causation as possible explanations.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: H. Parada Jr, M.D. Gammon
Development of methodology: H. Parada Jr, P.T. Bradshaw, L.S. Engel, M.D. Gammon
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P.T. Bradshaw, S.L. Teitelbaum, A.I. Neugut, R.M. Santella, M.D. Gammon
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Parada Jr, P.T. Bradshaw, L.S. Engel, S.E. Steck, A.I. Neugut, M.D. Gammon
Writing, review, and/or revision of the manuscript: H. Parada Jr, P.T. Bradshaw, L.S. Engel, K. Conway, S.E. Steck, S.L. Teitelbaum, A.I. Neugut, R.M. Santella, M.D. Gammon
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.D. Gammon
Study supervision: M.D. Gammon
Grant Support
The Long Island Breast Cancer Study Project was funded by the NCI and the National Institute of Environmental Health Sciences grants UO1CA/ES66572 and UO1CA66572 (awarded to M.D. Gammon). H. Parada Jr was supported by T32 ES007018 and R25 CA057726.
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.
- Received August 15, 2016.
- Accepted August 26, 2016.
- ©2016 American Association for Cancer Research.
Volume 26, Issue 2
