Smoking and survival after breast cancer diagnosis Michelle D. Holmes 1 * , Susan Murin 2,3 , Wendy Y. Chen 1,4 , Candyce H. Kroenke 5 , Donna Spiegelman 1,6 and Graham A. Colditz 1,6 1 Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA 2 UC Davis School of Medicine, Sacramento, CA 3 Department of Internal Medicine, Northern California Veterans Administration Health System, Sacramento, CA 4 Department of Medical Oncology, Dana Farber Cancer Institute, Boston, MA 5 Robert Wood Johnson Health and Society Scholar, University of California, San Francisco and Berkeley, San Francisco, CA 6 Department of Epidemiology, Harvard School of Public Health, Boston, MA We examined whether a history of smoking is associated with an increased risk of death from any cause or from breast cancer, among women diagnosed with breast cancer. This was a prospec- tive observational study among 5,056 women from the Nurses’ Health Study with Stages I–III invasive breast cancer diagnosed between 1978 and 2002 and for whom we had information on smoking, and who were followed until January 2002 or death, whichever came first. Subjects were classified as current, former or never smokers based upon smoking status at the biennial ques- tionnaire immediately preceding the breast cancer diagnosis. In multivariate-adjusted analyses, compared with never smokers, women who were current smokers had a 43% increased adjusted relative risk (RR) [95% confidence interval (95% CI): 1.24–1.65] of death from any cause. A strong linear gradient was observed with the number of cigarettes per day smoked, p-trend <0.0001; the RR (95% CI) for 1–14, 15–24 and 25 or more cigarettes per day was 1.27 (1.01–1.61), 1.30 (1.08–1.57) and 1.79 (1.47–2.19). In contrast, there was no association with current smoking and breast cancer death; the RR (95% CI) was 1.00 (0.83–1.19). Cur- rent and past smokers were more likely than never smokers to die from primary lung cancer, chronic obstructive pulmonary disease and other lung diseases. We conclude that a history of smoking increased mortality following diagnosis with breast cancer, but did not increase mortality from breast cancer. ' 2007 Wiley-Liss, Inc. Key words: breast neoplasms; survival; smoking; cohort studies; women Although smoking has not been clearly demonstrated to increase the incidence of breast cancer, 1,2 cigarette smoke or its constituents have been found to affect the metastatic propensity of tumor cells, cell adhesiveness and stimulation of angiogenesis. 3 In a mouse model of mammary cancer, cigarette smoke exposure was associated with an increase in the total pulmonary metastatic burden. 4 Therefore, it is plausible that smoking could increase the risk of breast cancer recurrence and death. Several epidemiological studies have examined the relationship between smoking and the natural history of breast cancer. In cross-sectional studies, Daniell et al. found that smokers with breast cancer had more and larger lymph node metastases than nonsmokers, after controlling for primary tumor size and other variables. 5,6 A case–control 7 and a retrospective cohort study 8 also found smoking to be associated with an increased risk of developing metastases to the lung. However, these studies could not definitively distinguish lung metastases from primary lung cancers. Four prospective studies have focused specifically upon the association of smoking and breast cancer survival. 9–12 Although a study of 1,774 Danish women with breast cancer showed no asso- ciation of smoking with survival [relative risk (RR), 95% confi- dence interval (95% CI) 5 1.05, 0.87–1.26], 12 the other 3 studies all reported an increased risk of total mortality associated with current smoking, 9–11 with RRs for fatal breast cancer ranging from 1.26 (95% CI: 1.05–1.50) among 808 cases in the American Cancer Society (ACS) Cohort 10 to 2.14 (95% CI: 1.47–3.10) among 792 cases in a Swedish mammography screening trial. 9 It should be noted that the Swedish mammography trial grouped to- gether in situ with invasive carcinomas. The ACS study, because the outcome was fatal breast cancer, could not distinguish an effect on incidence from a true effect of smoking on survival. Studies on smoking and survival after breast cancer, which can distinguish both incidence from survival, and breast cancer metas- tases from primary cancers due to smoking but unrelated to breast cancer, are needed. Based on the biological evidence that smoking may influence breast cancer progression, in our prospective cohort of more than 5,000 women with breast cancer and smoking information, we hypothesized that smoking would be associated with both an increased risk of death from all causes, and an increased risk of death from breast cancer. Methods Study subjects and identification of breast cancer In 1976, the Nurses’ Health Study (NHS) cohort was established when 121,700 female registered nurses from across the United States, aged 30–55 years, answered a mailed questionnaire on risk factors for cancer and cardiovascular disease. Every 2 years since, we have sent follow-up questionnaires to NHS participants. For any report of breast cancer on the biennial questionnaire, written permission was obtained from study participants to review their medical records. Physicians, blinded to exposure information from questionnaires, subsequently reviewed medical and pathol- ogy reports. Overall, 99% of self-reported invasive breast cancers for which medical records were obtained have been confirmed. We included women with Stages I–III invasive breast cancer diagnosed between 1978 and 2002, and confirmed by medical record review. Women were excluded if they were diagnosed with any cancer (other than nonmelanoma skin cancer) prior to 1978, had missing information on smoking or disease stage, or had in situ or Stage IV disease at diagnosis. Women with 4 or more positive nodes but lacking a complete metastatic workup were also excluded because of concerns for occult metastatic disease. A complete metastatic workup consisted of a negative chest X-ray (or chest computer tomography), bone scan and liver function tests (or liver scan) or documentation from a treating physician that the patient did not have metastatic dis- The authors have no conflicts of interest to disclose. Grant sponsor: National Institutes of Health; Grant number: CA87969; Grant sponsor: American Lung Association; Grant number: CG-014-N. *Correspondence to: Channing Laboratory, 181 Longwood Avenue, Boston, MA 02115, USA. Fax: 617-525-2008. E-mail: michelle.holmes@channing.harvard.edu Received 22 August 2006; Accepted after revision 8 December 2006 DOI 10.1002/ijc.22575 Published online 2 February 2007 in Wiley InterScience (www.interscience. wiley.com). Int. J. Cancer: 120, 2672–2677 (2007) ' 2007 Wiley-Liss, Inc. Publication of the International Union Against Cancer