Abstract
Background: The 2014 US Surgeon General’s report noted research gaps necessary to determine a causal relationship between active cigarette smoking and invasive breast cancer risk, including the role of alcohol consumption, timing of exposure, modification by menopausal status and heterogeneity by oestrogen receptor (ER) status.
Methods: To address these issues, we pooled data from 14 cohort studies contributing 934 681 participants (36 060 invasive breast cancer cases). Cox proportional hazard regression models were used to calculate multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs).
Results: Smoking duration before first birth was positively associated with risk (P-value for trend = 2 × 10–7) with the highest HR for initiation >10 years before first birth (HR = 1.18, CI 1.12–1.24). Effect modification by current alcohol consumption was evident for the association with smoking duration before first birth (P-value=2×10–4); compared with never-smoking non-drinkers, initiation >10 years before first birth was associated with risk in every category of alcohol intake, including non-drinkers (HR = 1.15, CI 1.04–1.28) and those who consumed at least three drinks per day (1.85, 1.55–2.21). Associations with smoking before first birth were limited to risk of ER+ breast cancer (P-value for homogeneity=3×10–3). Other smoking timing and duration characteristics were associated with risk even after controlling for alcohol, but were not associated with risk in non-drinkers. Effect modification by menopause was not evident.
Conclusions: Smoking, particularly if initiated before first birth, was modestly associated with ER+ breast cancer risk that was not confounded by amount of adult alcohol intake. Possible links with breast cancer provide additional motivation for young women to not initiate smoking.
Keywords: tobacco smoking, alcohol, breast cancer
Key Messages
In a pooled analysis of 14 prospective cohort studies of nearly one million women, smoking >10 years before first birth was associated with an 18% higher risk of breast cancer, the strongest association of all the smoking characteristics compared with never smokers.
The association with smoking varied by alcohol consumption, but was evident in every category of alcohol intake. Smoking >10 years before first birth was associated with an 85% higher risk of breast cancer in the heaviest smokers and drinkers.
Associations with years of smoking initiation were stronger for risk of ER+ breast cancer, estimates ranging from 1.02 to 1.21, than for risk of ER– breast cancer, estimates ranging from 0.95 to 1.08.
The association of smoking characteristics with breast cancer risk varied little by menopausal status and age at menopause.
Introduction
Determining whether there is a causal relationship between active cigarette smoking and breast cancer risk has been controversial. There are biological data linking active smoking, particularly at young ages, with breast carcinogenesis, which include the induction of mammary cancers by 20 tobacco smoke compounds in rodents1–3 and detectable tobacco metabolites,4,5 and smoking-specific DNA adducts and p53 mutation in human breasts,6–10 although limitations and inconsistencies do exist.11 Despite the biological data and the positive associations in a large number of epidemiologic studies published before 2012,1,2,11–16 the recent US Surgeon General’s report concluded that ‘the evidence is suggestive but not sufficient to infer a causal relationship between active smoking and breast cancer’.11 The report noted lingering epidemiological issues concerning the assessment of this relationship, including whether the association is due to: (i) the timing of exposure at early ages and/or long duration of smoking, (ii) confounding or effect modification by alcohol intake, (iii) modification by menopausal status or (iv) differences by oestrogen receptor (ER) status.
The two prevailing concerns are residual confounding by alcohol intake and timing of smoking initiation relative to first birth.1,2,11–16 Alcohol consumption is an established risk factor for breast cancer; even consumption at low levels is associated with increased risk.17 Some have argued that the association with active smoking can only be evaluated in never drinkers because of the potential correlation between cigarette smoking and alcohol consumption.18 The Collaborative Group on Hormonal Factors in Breast Cancer concluded that there was no association of ever smoking [relative risk (RR)=1.03, 95% confidence interval (CI) 0.98–1.07]; however, among drinkers, the RR with ever smoking was 1.09 (95% CI 1.05–1.13), which was attenuated to 1.05 (95% CI 1.01–1.09) after adjustment for amount of alcohol consumed.18 However, some of the recent studies19–21 published after the Surgeon General’s report have found associations with smoking even among non-drinkers when these women were examined separately. Analysis of smoking initiation relative to first birth is also an important issue because, before the first full-term birth, the undifferentiated breast epithelium is particularly susceptible to carcinogens.22 Whereas previous cohort studies have shown that the strongest smoking association with breast cancer risk is among women who initiated smoking prior to first birth,19–21,23–26 no studies have examined associations of initiation relative to first birth stratified by alcohol consumption.
In this study, we pooled data from 14 prospective cohorts and undertook a unified analytical approach to overcome the lingering epidemiological issues related to assessing the association between smoking and breast cancer risk.
Methods
Study population
Member studies of the National Cancer Institute (NCI) Cohort Consortium with smoking data and ≥500 incident breast cancer cases were invited to participate; 14 cohorts (Supplementary Table 1, available as Supplementary data at IJE online) agreed. Investigators from each cohort provided individual-level data for the entire cohort after excluding those who were male, had a personal history of cancer at baseline (except non-melanoma skin cancer), had missing information on smoking status at baseline or had other cohort-specific exclusions. Data for 934 681 women were included in this analysis. Written informed consent was obtained from study participants at entry into each cohort or was implied by participants’ return of the enrolment questionnaire. The present investigation was approved by the Institutional Review Board (IRB) at each participating institution or was considered within the scope of the original IRB protocol.
Exposure information
De-identified data from the baseline questionnaire were provided for active cigarette smoking, current alcohol consumption (former alcohol drinkers were distinguished for only 6 of 14 studies) and other characteristics. Smoking status (never, former, current) is defined as at the time the baseline questionnaire was completed. Data were harmonized and variables were categorized a priori. Initiation of smoking relative to first birth, defined among parous women, is based on the number of years between age at smoking initiation and age at first pregnancy.
Case definition
In our primary analyses, cases were defined as incident, invasive breast cancers diagnosed after enrolment and identified through self-report, cancer registry linkage, medical record/pathology report or death certificate. In the latter situation, breast cancer had to be listed as a primary or contributory cause of death (ICD-9: 174 or ICD-O, ICD-10: C50). Incident in situ breast tumours were excluded from the case definition, because risk factors for in situ breast cancer might differ from invasive breast cancer.27 Tumours of unknown invasiveness were assumed to be invasive.
Statistical analysis
Person-time was calculated from the date of the return of the baseline survey until the date of the first-occurring event: breast cancer diagnosis, death or last follow-up. Women diagnosed with carcinoma in situ of the breast were censored at the time of diagnosis. In pooled analyses, Cox proportional hazard regression models were used to calculate minimally adjusted and multivariable-adjusted hazard ratios (HRs) and 95% CI. All models controlled for study as a covariate and were stratified on age at enrolment. Multivariable-adjusted models included breast cancer risk factors (categorized as listed in Table 1, including categories for missing variables) and are shown with and without control for alcohol consumption. Models of time since quitting also included smoking duration. Linear trend tests were based on the P-value of the continuous smoking variables, excluding never smokers.28 In addition to pooled analyses, a meta-analytic approach was used assuming a random effects model and weighting the cohorts based on the inverse of the cohort size. Between-study heterogeneity was assessed using the I2 statistic,29,30 in which I2 of 0–40% was considered evidence of minimal heterogeneity, 30–60% moderate heterogeneity, 50–90% substantial heterogeneity and 75–100% considerable heterogeneity.31
Table 1.
Distribution of key factors as reported on the baseline survey, a pooled analysis of active smoking and invasive breast cancer risk among 14 cohorts in the NCI Cohort Consortium
| Never smoker (n = 494 194) | Former smoker (n = 288 895) | Current smoker (n = 151 592) | |
|---|---|---|---|
| Years of follow-up, mean (SD) | 13.5 (7.1) | 12.2 (7.5) | 14.7 (7.4) |
| Age at baseline, mean (SD) | 54.1 (13.6) | 55.5 (12.6) | 49.7 (11.5) |
| Age at menarche, mean (SD) | 12.8 (1.5) | 12.8 (1.4) | 12.9 (1.5) |
| Age at first birth, mean (SD) | 24.4 (4.2) | 24.2 (4.2) | 23.4 (4.2) |
| Number of births, mean (SD) | 2.5 (1.4) | 2.4 (1.4) | 2.4 (1.4) |
| Age at menopause, mean (SD) | 48.1 (5.8) | 47.6 (5.8) | 46.7 (6) |
| Race (%) | |||
| White | 88.6 | 95.3 | 95.8 |
| Black | 2.2 | 2.4 | 1.8 |
| Asian | 7.4 | 1.0 | 1.7 |
| Other/missing | 1.8 | 1.4 | 0.8 |
| Ethnicity (%) | |||
| Not Hispanic | 98.1 | 98.6 | 99.2 |
| Hispanic | 1.3 | 0.9 | 0.6 |
| Missing | 0.6 | 0.5 | 0.2 |
| Education level (%) | |||
| Less than high school | 16.7 | 12.5 | 22.4 |
| High school | 19.3 | 19.5 | 21.5 |
| Some college/college graduate | 62.4 | 66.0 | 53.6 |
| Missing | 1.6 | 2.0 | 2.4 |
| Parity (%) | |||
| No | 13.1 | 11.8 | 12.1 |
| Yes | 85.0 | 87.4 | 86.9 |
| Missing | 1.9 | 0.8 | 1.1 |
| Family history of breast cancer (%) | |||
| No | 78.0 | 75.4 | 84.4 |
| Yes | 9.5 | 11.1 | 7.8 |
| Missing | 12.5 | 13.5 | 7.8 |
| Benign breast disease (%) | |||
| No | 46.7 | 42.4 | 38.9 |
| Yes | 14.1 | 14.5 | 11.3 |
| Missing | 39.2 | 43.1 | 49.8 |
| Ever oral contraceptive use (%) | |||
| Never | 50.6 | 46.2 | 43.0 |
| Ever | 46.9 | 52.1 | 54.7 |
| Missing | 2.5 | 1.7 | 2.4 |
| Menopausal status (%) | |||
| Pre-menopausal | 31.5 | 27.5 | 43.2 |
| Peri-menopausal | 0.2 | 0.2 | 0.1 |
| Post-menopausal | 61.8 | 64.7 | 47.3 |
| Missing | 6.5 | 7.7 | 9.4 |
| Ever post-menopausal hormone use (%) | |||
| Never | 62.1 | 54.1 | 69.3 |
| Ever | 32.3 | 41.3 | 26.1 |
| Missing | 5.7 | 4.6 | 4.6 |
| Body mass index (kg/m2, %) | |||
| <18.5 | 1.9 | 1.3 | 3.3 |
| 18.5–22.4 | 26.9 | 25.1 | 35.0 |
| 22.5–24.9 | 24.2 | 24.2 | 25.7 |
| 25.0–29.9 | 27.9 | 29.0 | 23.9 |
| 30+ | 15.8 | 16.7 | 9.6 |
| Missing | 3.4 | 3.7 | 2.6 |
| Alcohol consumption at baseline (%) | |||
| Not current drinker/non-drinker | 39.6 | 20.0 | 23.2 |
| <1 drink per day | 39.7 | 49.9 | 54.9 |
| 1–2 drinks per day | 4.9 | 7.8 | 6.7 |
| >2 drinks per day | 1.0 | 2.5 | 3.9 |
| Current drinker, amount missing | 14.7 | 19.8 | 11.3 |
| Missing | |||
| Oestrogen receptor status among cases | |||
| Positive | 67.6 | 69.0 | 63.7 |
| Negative | 14.7 | 13.9 | 14.2 |
| Missing | 17.7 | 17.2 | 22.1 |
Interaction analyses were conducted using a common reference group to evaluate for effect modification on the multiplicative scale. A P-value for interaction was calculated comparing the –2 log likelihood estimates of models with and without the interaction term(s). The interaction term was the cross-product of the two categorical variables with missing values excluded. Interaction analyses with menopausal status and age at menopause were conducted using a joint variable considering both exposures and excluded peri-menopausal women. Associations were evaluated for subgroups defined by ER status using a joint Cox proportional hazards model.32 Since data on ER status were not available for all cases, we compared main effect associations using cases with and without ER status to ensure those with data were not a biased sample. In sensitivity analysis, the influence of changes in smoking patterns prior to breast cancer diagnosis were examined by excluding cases that were diagnosed within the first two years of follow-up; this exclusion did not appreciably alter associations (data not shown). Reported P-values are two-sided. All programming was performed in R version 3.2.4 (3–10–2016), including the package survival (v2.38–3) for the aforementioned statistical models and the package forestplot (v1.4) for the forest plots.
Results
Among 934 681 study participants, 36 060 invasive breast cancer cases were diagnosed. The average age at baseline was 53.9 years, age at first birth was 24.0 years and number of births was 2.5. Most women were White (91.8%), had at least some college education (62.1%) and were post-menopausal at baseline (60.3%). Current smokers at baseline accounted for 16.2% of participants and they smoked an average of 15.3 cigarettes per day. Former smokers accounted for 30.9% of the participants and they quit smoking, on average, at 37.7 years of age. Most smokers (73.9%) started smoking before their first birth. Parous smokers who started before first birth, compared with those who started after first birth, were more likely to have smoked for longer (mean 23 vs 20 years, P-value < 1×10–3) and to have had their first birth at an older age (mean 24 vs 21 years of age, P-value < 1×10–3). Known breast cancer risk factors had expected associations (Supplementary Table 2, available as Supplementary data at IJE online).
Current smokers were more likely than never smokers to be less educated, pre-menopausal, current alcohol drinkers, oral contraceptive users, never users of menopausal hormone therapy and have a lower body mass index (BMI) (Table 1). Former smokers were more likely to have used menopausal hormone therapy than never smokers. Current and former smokers were more likely to drink alcohol at baseline than never smokers.
There was little confounding of the association between smoking status and invasive breast cancer (Table 2): controlling for alcohol intake at baseline changed the HR for current smoking from 1.09 to 1.07 (95% CI 1.04–1.10), compared with never smokers. There was only minimal evidence of between-study heterogeneity (Table 2 and Supplementary Figure 1, available as Supplementary data at IJE online). Associations of cigarettes per day and duration in current smokers and years since quitting and duration in former smokers showed linear trends (P-value < 0.02; Table 2); however, categorical HRs were within a narrow range of values (e.g. HRs for duration in current smokers ranged from 1.01 to 1.11). There was minimal heterogeneity between studies for most of the categories of these aforementioned smoking characteristics (I2 < 40%) with the exception of the lowest exposure categories of cigarettes per day (I2=47%) and duration in current smokers (I2=55%). Smoking 40 or more cigarettes per day at baseline had the strongest association with risk of breast cancer (HR=1.17, 95% CI 1.05–1.31; P for trend=4×10–3).
Table 2.
Multivariate-adjusted* hazard ratios (HRs) and 95% confidence intervals (CIs) for the association of smoking characteristics with invasive breast cancer risk, a pooled analysis among 14 cohorts in the NCI Cohort Consortium
| Cases | Rates | Age- and cohort- adjusted |
Multivariable- adjusted* excluding alcohol intake |
Multivariable- adjusted* including alcohol intake |
I2 | ||||
|---|---|---|---|---|---|---|---|---|---|
| HR | 95% CI | HR | 95% CI | HR | 95% CI | ||||
| Smoking status | |||||||||
| Never | 18 647 | 254 | 1.00 | 1.00 | 1.00 | ||||
| Former | 11 585 | 296 | 1.11 | (1.08, 1.13) | 1.08 | (1.06, 1.11) | 1.06 | (1.04, 1.09) | <1% |
| Current | 5828 | 266 | 1.05 | (1.02, 1.08) | 1.09 | (1.06, 1.13) | 1.07 | (1.04, 1.10) | 39% |
| Cigarettes per day in current smokers | |||||||||
| Never smoker | 18 647 | 254 | 1.00 | 1.00 | 1.00 | ||||
| <10 | 1238 | 247 | 1.05 | (0.99, 1.11) | 1.10 | (1.03, 1.16) | 1.08 | (1.01, 1.14) | 47% |
| 10–19 | 2033 | 260 | 1.04 | (1.00, 1.09) | 1.10 | (1.04, 1.15) | 1.07 | (1.02, 1.13) | 9% |
| 20–29 | 1570 | 274 | 1.03 | (0.97, 1.08) | 1.06 | (1.01, 1.12) | 1.04 | (0.98, 1.09) | 15% |
| 30–39 | 592 | 295 | 1.10 | (1.02, 1.20) | 1.13 | (1.04, 1.22) | 1.09 | (1.00, 1.18) | 27% |
| ≥40 | 315 | 314 | 1.19 | (1.07, 1.33) | 1.22 | (1.09, 1.37) | 1.17 | (1.05, 1.31) | <1% |
| P-value for linear trenda | 2×10–3 | 4×10–3 | |||||||
| Years of smoking in current smokers | |||||||||
| Never smokers | 18 647 | 254 | 1.00 | 1.00 | 1.00 | ||||
| <10 | 94 | 248 | 0.93 | (0.76, 1.15) | 1.02 | (0.83, 1.25) | 1.01 | (0.82, 1.23) | 55% |
| 10–19 | 794 | 250 | 1.10 | (1.02, 1.19) | 1.13 | (1.05, 1.22) | 1.11 | (1.03, 1.20) | <1% |
| 20–39 | 3632 | 274 | 1.04 | (1.00, 1.08) | 1.09 | (1.05, 1.13) | 1.06 | (1.02, 1.10) | 5% |
| ≥40 | 1231 | 242 | 1.08 | (1.02, 1.15) | 1.12 | (1.05, 1.19) | 1.09 | (1.03, 1.16) | 10% |
| P-value for linear trenda | 9×10–3 | 0.02 | |||||||
| Years of smoking in former smokers | |||||||||
| Never smoker | 18 647 | 254 | 1.00 | 1.00 | 1.00 | ||||
| <5 | 1588 | 264 | 1.04 | (0.99, 1.09) | 1.04 | (0.99, 1.10) | 1.03 | (0.98, 1.09) | <1% |
| 5–10 | 1541 | 290 | 1.08 | (1.03, 1.14) | 1.05 | (1.00, 1.11) | 1.04 | (0.98, 1.09) | <1% |
| 10–20 | 3263 | 305 | 1.14 | (1.10, 1.19) | 1.11 | (1.07, 1.15) | 1.09 | (1.05, 1.13) | <1% |
| 20–30 | 2314 | 323 | 1.11 | (1.06, 1.16) | 1.08 | (1.04, 1.13) | 1.06 | (1.02, 1.11) | <1% |
| 30–40 | 1592 | 348 | 1.15 | (1.09, 1.21) | 1.12 | (1.06, 1.18) | 1.09 | (1.04, 1.15) | <1% |
| >40 | 541 | 231 | 1.18 | (1.08, 1.29) | 1.17 | (1.08, 1.28) | 1.14 | (1.05, 1.24) | <1% |
| P-value for linear trenda | 1×10–4 | 8×10–4 | |||||||
| Years since quitting in former smokers | |||||||||
| Never smoker | 18 647 | 254 | 1.00 | 1.00 | 1.00 | ||||
| ≥31 | 2251 | 318 | 1.08 | (1.03, 1.13) | 1.05 | (1.00, 1.09) | 1.03 | (0.98, 1.07) | 35% |
| 21–30 | 3357 | 300 | 1.11 | (1.07, 1.15) | 1.09 | (1.05, 1.13) | 1.07 | (1.03, 1.11) | 26% |
| 11–20 | 3815 | 290 | 1.14 | (1.10, 1.18) | 1.13 | (1.09, 1.17) | 1.11 | (1.07, 1.15) | <1% |
| 1–10 | 5828 | 266 | 1.05 | (1.02, 1.09) | 1.10 | (1.06, 1.13) | 1.07 | (1.04, 1.11) | 42% |
| P-value for linear trenda | 3×10–4 | 1×10–3 | |||||||
| Age of smoking initiation in ever smokers | |||||||||
| Never smoker | 18 647 | 254 | 1.00 | 1.00 | 1.00 | ||||
| <16 | 1523 | 283 | 1.10 | (1.04, 1.16) | 1.03 | (0.98, 1.09) | 1.01 | (0.96, 1.07) | <1% |
| 16–20 | 10 251 | 290 | 1.10 | (1.07, 1.12) | 1.10 | (1.07, 1.12) | 1.07 | (1.04, 1.10) | 31% |
| 21–25 | 3956 | 293 | 1.12 | (1.08, 1.16) | 1.13 | (1.09, 1.17) | 1.11 | (1.07, 1.15) | <1% |
| ≥25 years | 1373 | 256 | 0.99 | (0.94, 1.04) | 1.02 | (0.97, 1.08) | 1.01 | (0.95, 1.06) | <1% |
| P-value for linear trenda | 0.02 | 0.07 | |||||||
| Smoking initiation relative to first birth in ever parous smokers | |||||||||
| Never smoker | 16 025 | 250 | 1.00 | 1.00 | 1.00 | ||||
| After first birth | 1767 | 246 | 0.99 | (0.95, 1.04) | 1.07 | (1.02, 1.12) | 1.05 | (1.00, 1.11) | <1% |
| ≤5 years before birth | 5913 | 270 | 1.03 | (1.00, 1.07) | 1.08 | (1.04, 1.11) | 1.06 | (1.02, 1.09) | 34% |
| 6–10 years before birth | 5201 | 302 | 1.16 | (1.12, 1.20) | 1.12 | (1.09, 1.16) | 1.10 | (1.06, 1.14) | 55% |
| >10 years before birth | 2065 | 328 | 1.35 | (1.29, 1.41) | 1.20 | (1.15, 1.26) | 1.18 | (1.12, 1.24) | 36% |
| P-value for linear trenda | 1×10–8 | 2×10–7 | |||||||
Multivariable-adjusted models included age, cohort, race/ethnicity, education, birth year, family history of breast cancer, history of benign breast disease, ever use of oral contraceptives, menopausal status and age at menopause, age at menarche, ever use of menopausal hormone therapy, age at first birth and number of live births, body mass index, and amount and frequency of alcohol use. aLinear trend was evaluated using a continuous version of the variable excluding never smokers.
Age of smoking initiation in ever smokers was not associated with breast cancer risk (Table 2). However, in parous women, those who initiated smoking more than 10 years before their first birth had the highest risk of breast cancer compared with never smokers (HR=1.18, 95% CI 1.12–1.24; Table 2). These results had only minimal evidence of between-study heterogeneity (Table 2 and Supplementary Figure 2, available as Supplementary data at IJE online). To evaluate whether this association was driven by residual confounding from later age at first birth or longer smoking duration, we conducted stratified analyses by these factors and found similar effects for smoking initiation in each strata (P-values for interaction=0.51 and 0.74, respectively; Supplementary Table 3a and b, available as Supplementary data at IJE online).
Smoking patterns differed by current alcohol consumption (results not otherwise shown): among non-drinkers, 68% of women were never smokers, 12% were current smokers and 20% were former smokers. Among people who drank more than two drinks per day, 28% were never smokers, 33% were current smokers and 39% were former smokers. Smoking for ≥40 years was more prevalent among women who reported drinking more than two drinks per day than not currently drinking: 11.5% vs 4.1%, respectively.
The associations of smoking characteristics of timing and duration with breast cancer risk were modified by current alcohol intake (P-values for interaction < 0.05; Table 3). Using a common reference group of never smokers, non-drinkers, former and current smoking was not associated with breast cancer risk among non-drinkers; however, current drinkers consuming two or more drinks per day who were former (HR=1.33, 95% CI 1.19–1.49) or current smokers (HR=1.32, 95% CI 1.16–1.49) were at slightly greater risk than expected under a multiplicative model (expected HRs=1.22 and 1.22, respectively). Smoking more than 10 years before first birth was associated with breast cancer risk among non-drinkers (HR=1.15, 95% CI 1.04–1.28) and in every stratum of alcohol intake (Table 3). Current drinkers consuming at least three drinks per day who smoked >10 years prior to first birth (HR=1.85, 95% CI 1.55–2.21; Table 3) were at noticeably greater risk than expected under a multiplicative model (expected HR=1.43).
Table 3.
Multivariate-adjusted* hazard ratios (HRs) and 95% confidence intervals (CIs) for the association of smoking characteristics with invasive breast cancer risk by amount of alcohol intake, a pooled analysis among 14 cohorts in the NCI Cohort Consortium
| |
Non-current drinker |
<1 drink per day |
1–2 drinks per day |
≥3 drinks per day |
P-values for interaction | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Cases | HR | 95% CI | Cases | HR | 95% CI | Cases | HR | 95% CI | Cases | HR | 95% CI | ||
| Smoking status | |||||||||||||
| Never smoker | 7746 | 1 | 8075 | 1.01 | (0.98, 1.05) | 1110 | 1.06 | (0.99, 1.14) | 303 | 1.21 | (1.06, 1.38) | 0.03 | |
| Former | 2396 | 1.01 | (0.97, 1.06) | 6239 | 1.09 | (1.05, 1.13) | 1415 | 1.25 | (1.18, 1.34) | 528 | 1.33 | (1.19, 1.49) | |
| Current | 1262 | 1.01 | (0.95, 1.08) | 3207 | 1.11 | (1.06, 1.16) | 556 | 1.21 | (1.11, 1.33) | 417 | 1.32 | (1.16, 1.49) | |
| Cigarettes per day in current smokers | |||||||||||||
| Never smoker | 7746 | 1 | 8075 | 1.01 | (0.98, 1.05) | 1110 | 1.06 | (0.99, 1.15) | 303 | 1.2 | (1.04, 1.38) | 0.02 | |
| <10 | 225 | 0.94 | (0.82, 1.07) | 763 | 1.11 | (1.02, 1.20) | 116 | 1.4 | (1.16, 1.69) | 38 | 1.43 | (1.03, 2.00) | |
| 10–19 | 437 | 1.1 | (0.99, 1.21) | 1188 | 1.11 | (1.04, 1.19) | 169 | 1.19 | (1.01, 1.39) | 71 | 1 | (0.78, 1.29) | |
| 20–29 | 366 | 0.95 | (0.86, 1.06) | 815 | 1.07 | (1.00, 1.16) | 166 | 1.18 | (1.01, 1.38) | 147 | 1.3 | (1.08, 1.56) | |
| 30–39 | 139 | 1 | (0.85, 1.18) | 266 | 1.15 | (1.02, 1.31) | 79 | 1.22 | (0.97, 1.53) | 82 | 1.28 | (1.01, 1.62) | |
| ≥40 | 85 | 1.18 | (0.95, 1.47) | 121 | 1.17 | (0.97, 1.40) | 24 | 0.89 | (0.59, 1.33) | 79 | 1.85 | (1.45, 2.36) | |
| P-value for linear trenda | 0.16 | 0.07 | 0.55 | 4×10–3 | |||||||||
| Years of smoking in current smokers | |||||||||||||
| Never smoker | 7746 | 1 | 8075 | 1.01 | (0.98, 1.05) | 1110 | 1.06 | (0.98, 1.14) | 303 | 1.19 | (1.03, 1.37) | 0.03 | |
| <10 | 29 | 1.25 | (0.87, 1.80) | 49 | 0.92 | (0.7, 1.23) | 3 | 0.63 | (0.20, 1.96) | 1 | 0.35 | (0.05, 2.51) | |
| 10–19 | 188 | 1.11 | (0.96, 1.29) | 499 | 1.15 | (1.04, 1.26) | 46 | 1.09 | (0.81, 1.46) | 27 | 1.09 | (0.74, 1.60) | |
| 20–39 | 725 | 0.96 | (0.89, 1.04) | 2097 | 1.11 | (1.05, 1.17) | 338 | 1.15 | (1.03, 1.30) | 282 | 1.31 | (1.13, 1.51) | |
| ≥40 | 294 | 1.07 | (0.95, 1.21) | 545 | 1.07 | (0.97, 1.17) | 146 | 1.45 | (1.23, 1.72) | 99 | 1.42 | (1.13, 1.79) | |
| P-value for linear trenda | 0.25 | 0.12 | 3×10–3 | 7×10–3 | |||||||||
| Years of smoking in former smokers | |||||||||||||
| Never smoker | 7746 | 1 | 8075 | 1.02 | (0.98, 1.05) | 1110 | 1.06 | (0.98, 1.14) | 303 | 1.2 | (1.04, 1.38) | ||
| <5 | 380 | 1 | (0.90, 1.10) | 950 | 1.08 | (1.01, 1.16) | 112 | 1.06 | (0.87, 1.28) | 35 | 1.06 | (0.75, 1.50) | |
| 5–10 | 312 | 1.03 | (0.92, 1.15) | 883 | 1.02 | (0.95, 1.10) | 174 | 1.33 | (1.14, 1.55) | 68 | 1.54 | (1.20, 1.99) | |
| 10–20 | 586 | 0.99 | (0.91, 1.08) | 1876 | 1.13 | (1.07, 1.19) | 388 | 1.34 | (1.20, 1.49) | 153 | 1.35 | (1.13, 1.62) | |
| 20–30 | 444 | 1 | (0.91, 1.10) | 1265 | 1.12 | (1.06, 1.20) | 268 | 1.14 | (1.01, 1.30) | 103 | 1.13 | (0.91, 1.40) | |
| 30–40 | 357 | 1.1 | (0.99, 1.23) | 751 | 1.08 | (1.00, 1.17) | 212 | 1.29 | (1.12, 1.48) | 100 | 1.46 | (1.17, 1.82) | |
| >40 | 124 | 1.1 | (0.92, 1.32) | 214 | 1.08 | (0.94, 1.24) | 80 | 1.59 | (1.27, 1.99) | 44 | 1.59 | (1.16, 2.18) | 2×10–4 |
| P-value for linear trenda | 0.01 | 0.04 | 0.36 | 0.43 | |||||||||
| Years since quitting smoking in former smokers | |||||||||||||
| Never smoker | 7746 | 1 | 8075 | 1.01 | (0.98, 1.05) | 1110 | 1.06 | (0.99, 1.14) | 303 | 1.2 | (1.05, 1.38) | 0.05 | |
| ≥31 | 474 | 1.02 | (0.93, 1.12) | 1209 | 1.05 | (0.98, 1.11) | 276 | 1.22 | (1.08, 1.38) | 95 | 1.19 | (0.95, 1.47) | |
| 21–30 | 634 | 0.97 | (0.89, 1.05) | 1914 | 1.09 | (1.04, 1.15) | 378 | 1.25 | (1.12, 1.39) | 168 | 1.47 | (1.24, 1.74) | |
| 11–20 | 789 | 1.07 | (0.99, 1.15) | 2177 | 1.14 | (1.09, 1.2) | 377 | 1.25 | (1.13, 1.40) | 190 | 1.31 | (1.12, 1.54) | |
| 1–10 | 1262 | 1.01 | (0.95, 1.07) | 3207 | 1.11 | (1.06, 1.16) | 556 | 1.21 | (1.10, 1.32) | 417 | 1.31 | (1.16, 1.48) | |
| P-value for linear trenda | 0.02 | 1×10–3 | 0.48 | 0.95 | |||||||||
| Age of smoking initiation in ever smokers | |||||||||||||
| Never smoker | 7746 | 1 | 8075 | 1.02 | (0.98, 1.05) | 1110 | 1.06 | (0.98, 1.13) | 303 | 1.2 | (1.05, 1.37) | ||
| <16 | 327 | 1.04 | (0.93, 1.16) | 817 | 1.01 | (0.94, 1.09) | 128 | 1.13 | (0.95, 1.35) | 84 | 1.34 | (1.06, 1.69) | |
| 16–20 | 2046 | 1.03 | (0.98, 1.08) | 5576 | 1.08 | (1.04, 1.12) | 1243 | 1.3 | (1.22, 1.39) | 639 | 1.35 | (1.21, 1.50) | |
| 21–25 | 776 | 0.96 | (0.90, 1.04) | 2294 | 1.2 | (1.14, 1.26) | 388 | 1.2 | (1.08, 1.33) | 155 | 1.32 | (1.11, 1.57) | |
| ≥25 | 399 | 1.03 | (0.93, 1.14) | 687 | 1.04 | (0.96, 1.12) | 120 | 1.12 | (0.93, 1.34) | 48 | 1.11 | (0.83, 1.48) | 5×10–6 |
| P-value for linear trenda | 0.79 | 0.75 | 0.04 | 0.13 | |||||||||
| Smoking initiation relative to first birth in parous ever smokers | |||||||||||||
| Never smoker | 6697 | 1 | 7024 | 1.03 | (1.00, 1.07) | 860 | 1.05 | (0.97, 1.14) | 258 | 1.24 | (1.08, 1.43) | 2×10–4 | |
| After first birth | 497 | 1.01 | (0.92, 1.10) | 907 | 1.12 | (1.04, 1.20) | 125 | 1.13 | (0.94, 1.35) | 61 | 1.18 | (0.91, 1.53) | |
| ≤5 years before birth | 1275 | 1 | (0.94, 1.07) | 3195 | 1.09 | (1.05, 1.14) | 591 | 1.24 | (1.14, 1.36) | 314 | 1.35 | (1.18, 1.54) | |
| 6–10 years before birth | 963 | 1.05 | (0.98, 1.12) | 2977 | 1.14 | (1.09, 1.19) | 665 | 1.34 | (1.23, 1.46) | 301 | 1.33 | (1.16, 1.52) | |
| >10 years before birth | 386 | 1.15 | (1.04, 1.28) | 1163 | 1.19 | (1.11, 1.27) | 197 | 1.3 | (1.12, 1.50) | 148 | 1.85 | (1.55, 2.21) | |
| P-value for linear trenda | 0.06 | 6×10–4 | 0.02 | 5×10–4 | |||||||||
Multivariable-adjusted models included age, cohort, race/ethnicity, education, birth year, family history of breast cancer, history of benign breast disease, ever use of oral contraceptives, menopausal status and age at menopause, age at menarche, ever use of menopausal hormone therapy, age at first birth and number of live births, body mass index and amount of alcohol use.
Linear trend was evaluated using a continuous version of the variable excluding never smokers.
Interactions with menopausal status and age at menopause were also examined (Supplementary Table 4, available as Supplementary data at IJE online). The associations with smoking status and smoking initiation relative to first birth did not meaningfully vary by menopause status and age at menopause.
We also examined whether the association with smoking status was different for subtypes defined by ER status (Table 4). Most breast cancers were ER + (67.4%); the association of smoking characteristics of timing and duration with breast cancer risk overall was similar for those with and without hormone receptor data (results not shown). The association with smoking before first birth differed by ER status, in which the association was stronger for risk of ER+ breast cancer than risk of ER– breast cancer (P-value for tumour heterogeneity=3×10–3). Modification of the associations of smoking initiation relative to first birth with breast cancer risk by alcohol intake also was stronger for risk of ER+ breast cancer (data not in tables). Current drinkers consuming two or more drinks per day who smoked >10 years prior to first birth (HR=2.02, 95% CI 1.64–2.49) were also at greater risk than expected under a multiplicative model (expected HR=1.50).
Table 4.
Multivariate-adjusted* hazard ratios (HRs) and 95% confidence intervals (CIs) for the association of smoking characteristics with invasive breast cancer risk by oestrogen receptor (ER) status, a pooled analysis among 14 cohorts in the NCI Cohort Consortium
| ER + |
ER– |
P-value for homogeneity | |||||
|---|---|---|---|---|---|---|---|
| Cases | HR | 95% CI | Cases | HR | 95% CI | ||
| Smoking status | |||||||
| Never smoker | 11 616 | 1 | 2563 | 1 | 0.05 | ||
| Former | 7297 | 1.08 | (1.05, 1.11) | 1455 | 0.99 | (0.93, 1.06) | |
| Current | 3466 | 1.05 | (1.01, 1.10) | 766 | 1 | (0.92, 1.08) | |
| Cigarettes per day in current smokers | |||||||
| Never smoker | 11 616 | 1 | 2563 | 1 | 0.55 | ||
| <10 | 708 | 1.07 | (0.99, 1.16) | 151 | 0.94 | (0.80, 1.10) | |
| 10–19 | 1121 | 1.03 | (0.97, 1.09) | 266 | 1.04 | (0.92, 1.17) | |
| 20–29 | 998 | 1.05 | (0.98, 1.12) | 215 | 0.98 | (0.86, 1.13) | |
| 30–39 | 388 | 1.11 | (1.01, 1.23) | 82 | 0.98 | (0.78, 1.22) | |
| ≥40 | 204 | 1.15 | (1.00, 1.32) | 49 | 1.18 | (0.89, 1.56) | |
| P-value for linear trenda | 0.03 | 0.11 | |||||
| Years of smoking in current smokers | |||||||
| Never smoker | 11 616 | 1 | 2563 | 1 | 0.68 | ||
| <10 | 51 | 1.05 | (0.81, 1.36) | 14 | 0.95 | (0.56, 1.60) | |
| 10–19 | 466 | 1.12 | (1.01, 1.23) | 113 | 1.02 | (0.84, 1.24) | |
| 20–39 | 2167 | 1.05 | (1.00, 1.10) | 509 | 1.01 | (0.92, 1.11) | |
| ≥40 | 742 | 1.06 | (0.99, 1.14) | 121 | 0.96 | (0.81, 1.14) | |
| P-value for linear trenda | 0.1 | 0.2 | |||||
| Years of smoking in former smokers | |||||||
| Never smoker | 12 608 | 1 | 2740 | 1 | |||
| <5 | 1007 | 1.02 | (0.95, 1.09) | 235 | 1.02 | (0.89, 1.16) | |
| 5–10 | 1077 | 1.09 | (1.02, 1.16) | 206 | 0.91 | (0.79, 1.04) | |
| 10–20 | 2273 | 1.13 | (1.08, 1.18) | 445 | 0.98 | (0.89, 1.09) | |
| 20–30 | 1626 | 1.08 | (1.02, 1.14) | 317 | 1 | (0.89, 1.13) | |
| 30–40 | 1102 | 1.08 | (1.01, 1.15) | 227 | 1.14 | (0.99, 1.30) | |
| >40 | 375 | 1.14 | (1.02, 1.26) | 75 | 1.27 | (1.01, 1.60) | 0.04 |
| P-value for linear trenda | 0.07 | 4×10–4 | |||||
| Years since quitting smoker in former smokers | |||||||
| Never smoker | 11.616 | 1 | 2563 | 1 | 0.07 | ||
| ≥31 | 1441 | 1.06 | (1.00, 1.12) | 264 | 0.91 | (0.81, 1.03) | |
| 21–30 | 2147 | 1.1 | (1.05, 1.15) | 427 | 0.99 | (0.90, 1.09) | |
| 11–20 | 2345 | 1.1 | (1.05, 1.15) | 542 | 1.11 | (1.02, 1.22) | |
| 1–10 | 3466 | 1.05 | (1.01, 1.09) | 766 | 1 | (0.92, 1.08) | |
| P-value for linear trenda | 0.12 | 3×10–4 | |||||
| Age of smoking initiation in ever smokers | |||||||
| Never smoker | 12 608 | 1 | 2740 | 1 | |||
| <16 | 1018 | 1.01 | (0.95, 1.08) | 224 | 0.94 | (0.82, 1.08) | |
| 16–20 | 6927 | 1.08 | (1.05, 1.12) | 1425 | 0.99 | (0.93, 1.06) | |
| 21–25 | 2634 | 1.11 | (1.06, 1.15) | 569 | 1.08 | (0.99, 1.19) | |
| ≥25 | 923 | 1 | (0.93, 1.07) | 182 | 0.96 | (0.82, 1.11) | 0.16 |
| P-value for linear trenda | 0.07 | 0.28 | |||||
| Smoking initiation relative to first birth in parous ever smokers | |||||||
| Never smoker | 9936 | 1 | 2239 | 1 | 3×10–3 | ||
| After first birth | 1030 | 1.02 | (0.96, 1.09) | 226 | 1.06 | (0.93, 1.20) | |
| ≤5 years before birth | 3548 | 1.05 | (1.01, 1.09) | 798 | 1.03 | (0.95, 1.11) | |
| 6–10 years before birth | 3389 | 1.13 | (1.09, 1.18) | 644 | 0.95 | (0.87, 1.04) | |
| >10 years before birth | 1248 | 1.21 | (1.14, 1.28) | 271 | 1.08 | (0.95, 1.22) | |
| P-value for linear trenda | 5×10–7 | 0.29 | |||||
Multivariable-adjusted models included age, cohort, race/ethnicity, education, birth year, family history of breast cancer, history of benign breast disease, ever use of oral contraceptives, menopausal status and age at menopause, age at menarche, ever use of menopausal hormone therapy, age at first birth and number of live births, body mass index and amount of alcohol use.
Linear trend was evaluated using a continuous version of the variable excluding never smokers.
Discussion
In this large pooled analysis, we addressed the key lingering epidemiologic issues raised in the 2014 US Surgeon General’s report,11 including the importance of duration and timing of smoking initiation, the role of alcohol intake, modification by menopausal status and differences in risk by ER status. We confirmed modest associations of current and former smoking with invasive breast cancer risk. We also showed that the timing of smoking initiation was the smoking characteristic most strongly associated with risk, with initiation more than 10 years before first birth associated with an 18% increased risk of breast cancer among parous ever smokers, and was evident in non-drinkers and in every stratum of alcohol intake. Those who both initiated smoking more than 10 years before first birth and who were the heaviest drinkers (two or more drinks per day) were at greatest risk from early smoking initiation (85% higher risk of breast cancer) compared with never-smoking non-drinkers. Furthermore, this association was stronger for risk of ER+ breast cancer.
Based on their meta-analysis of 22 case–control and prospective studies,11 the Surgeon General’s report concluded that the association for smoking before and after first birth did not differ. In our pooled analysis, we found a dose response with the number of years parous women smoked before their first birth. The association was evident in all strata of age at first birth and of duration of smoking, indicating that the association was independent of these factors, as previously suggested.11 The inconsistencies among prior studies in the Surgeon General’s report might be due to the small number of cases, especially if examining the number of years of smoking prior to first birth in individual studies and the recent maturity of prospective studies that included birth cohorts who initiated smoking at a young age. For example, recent publications from other cohorts (not included in our analysis) also support stronger associations among parous women who smoke before first birth in highly screened women,25 in racial/ethnic subgroups21,26 and a comparatively large analysis of Europeans.19 Our pooled analysis included at least a portion of the data from six of the prospective studies included in the Surgeon General’s report.11 Our results for initiation years before first birth are also consistent with models of carcinogenesis of the breast.22
We further addressed the concern of residual confounding effects by alcohol intake on the association of smoking with breast cancer risk1,2,11–16 by using statistical control in multivariable models and by stratifying the association into categories of alcohol intake with additional statistical control for amount of alcohol within categories. Statistical control for alcohol intake at baseline only slightly attenuated the RRs for smoking characteristics in our study. Consistently with the Collaborative Group on Hormonal Factors in Breast Cancer,18 we observed no association between smoking status at baseline and breast cancer risk among non-drinkers. However, we did observe associations with initiating smoking more than 10 years prior to first birth in non-drinkers and in every stratum of alcohol intake at baseline. We cannot, however, eliminate the possibility that this association is further confounded by alcohol consumption during early adult life, which was not captured by the majority of the studies in our analysis. In the Nurses’ Health Study II, women who had higher alcohol intake between menarche and first birth also reported higher intakes at baseline and after first pregnancy compared with women who abstained between menarche and first birth. In their multivariable analyses, alcohol intake before first birth, compared with those abstaining before first birth, was associated with breast cancer risk (per 10-g increase: HR=1.11, 95% CI 1.00–1.23), independently of cumulative alcohol intake after first pregnancy.33
Stronger associations between smoking and breast cancer in pre-menopausal women has been hypothesized because the morphology of the breast and endogenous hormone levels undergo significant changes during the menopausal transition and other breast cancer risk factors are modified by menopausal status.11 However, we found no modification by menopausal status, which is consistent with prior studies.34,35 In our analysis, smoking initiation prior to first birth and smoking status were not modified by menopausal status, which is consistent with the large cohort of European women.19 Although the interaction was statistically significant for duration of smoking in current smokers, the lower risk estimated for the women who had smoked for ≥40 years and who had not gone through menopause was based on very few cases (n=26) and the adjacent exposure category of women who had smoked for 20–39 years did not differ across strata of menopause.
Despite initial published hypotheses and early evidence from small studies supporting a stronger association of smoking with ER– breast cancer,11 the findings from our pooled analysis suggest smoking, particularly initiation >10 years prior to first birth, suggest a positive association with risk of ER+, but not ER–, breast cancer. Results from the large cohort study of European women19 also suggested stronger positive associations with smoking for risk of hormone-receptor-positive disease. Although more recent data summarized in the Surgeon General’s report found stronger associations for ER+ breast cancer, they listed a number of limitations among the published studies, including use of case–control data, incomplete control for confounders and bias due to incomplete assessment of ER status.11 Our pooled analysis was based on prospective cohort data, we controlled for a large number of known breast cancer risk factors and we found no bias in the association with smoking among the cases who had ER status compared with those who did not.
Although data pooling provided a large number of study participants, it also presented limitations. Variables were harmonized to be inclusive of all participating cohorts, and we were not able to harmonize all variables. We were not able to define a reference group that excluded passive smokers or lifelong never drinkers, which likely biased our results toward the null. Despite this limitation, we did not find evidence of between-study heterogeneity in the associations (Supplementary Figures 1 and 2, available as Supplementary data at IJE online). We only collected baseline data on smoking status and covariate information. Although current smokers at baseline may have quit during follow-up, one of the studies included here did not find differences in results of baseline or updated smoking status.23
Consistently with previous studies published after the Surgeon General’s report,19,23,25,26,36 our estimate of the magnitude of association between smoking status and breast cancer risk is modest. However, this association did not appear to be confounded by alcohol intake; rather, our results support a synergistic relationship between smoking initiation and adult alcohol drinking. Furthermore, we found that longer duration of smoking prior to first birth was associated with risk, and this association persisted in both drinkers and non-drinkers. The associations with smoking were more consistently associated with risk of ER+ breast cancer. Other lingering epidemiologic issues mentioned in the recent US Surgeon General’s report that we addressed in this manuscript do not appear to have a major influence. Whereas the association with breast cancer might be modest relative to the more profound health effects of smoking on lung and other cancers, the number of breast cancer cases attributable to smoking might increase over time, as the prevalence of adolescent smoking in the USA has remained stable since the 1930s37,38 and globally smoking initiation at young ages is increasing39 and a greater proportion of women are delaying childbirth.40 Continued research in this area is warranted to further support public health campaigns aimed at preventing smoking and encouraging early cessation.
Supplementary Data
Supplementary data are available at IJE online.
Funding
The pooling project was sponsored by the NCI’s Cohort Consortium and funded by the American Cancer Society. The Agricultural Health Study (AgHealth) was funded by the Intramural Program of the NIH, NCI (Z01 P010119) and the National Institute of Environmental Health Sciences (Z01 ES 049030–11). The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study-II (CPS-II) cohort. The Iowa Women’s Health Study (IWHS) is supported by a grant from the NCI (R01 CA39742). The Melbourne Collaborative Cohort Study (MCCS) receives core funding from the Cancer Council Victoria and is additionally supported by grants from the Australian NHMRC (209057, 251533, 396414 and 504715). The NIH-AARP Diet and Health Study (AARP) was supported by the Intramural Research Program of the National Cancer Institute, National Institutes of Health. The Nurses’ Health Study 1 (NHS1) was supported by grants from the NCI (UM1CA176726 and P01CA087969–09). The Nurses’ Health Study 2 (NHS2) was supported by a grant from the NCI (R01 CA050385). The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial is supported by contracts from the NCI. The Singapore Chinese Health Study (SCHS) is supported by grants from the NCI (UM1 CA182876 and R01 CA144034). The Swedish Mammography Cohort (SMC) was supported by the Swedish Research Council, Swedish Council for Working Life and Social Research and the Swedish Cancer Foundation. The Swedish Women’s Lifestyle and Health Study (SWLH) was supported by the Swedish Research Council (grant number 521–2011–2955) and a Distinguished Professor Award at Karolinska Institutet to Hans-Olov Adami, grant number: 2368/10–221. VITamin and Lifestyle (VITAL) study: Dr White was supported by the National Institutes of Health grant K05-CA154337 (National Cancer Institute and Office of Dietary Supplements). The Women’s Health Initiative (WHI) Program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, US Department of Health and Human Services through contracts HHSN268201100046C, HHSN268201100001C, HHSN268201100002C, HHSN268201100003C, HHSN268201100004C and HHSN271201100004C.
Supplementary Material
Acknowledgements
The authors thank the study participants from all the cohorts for their invaluable contributions to this research. The Cancer Prevention Study-II (CPS-II) investigators thank the Study Management Group for their invaluable contributions to this research, and acknowledge the contribution to this study from central cancer registries supported through the Centers for Disease Control and Prevention National Program of Cancer Registries, and cancer registries supported by the National Cancer Institute Surveillance Epidemiology and End Results program. The Iowa Women’s Health Study (IWHS) investigators thank the IWHS study participants for their contributions, and Ching Ping Hong, MS for her assistance in creating the data set for this project. The Melbourne Collaborative Cohort Study (MCCS) was made possible by the contribution of many people, including the original investigators and the diligent team who recruited the participants and who continue working on follow-up. We would also like to express our gratitude to the many thousands of Melbourne residents who continue to participate in the study. We thank Drs. Christine Berg and Philip Prorok (Division of Cancer Prevention, NCI), the Screening Center investigators and staff of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial, Tom Riley and staff (Information Management Services, Inc.), Barbara O’Brien and staff (Westat, Inc.) and Jackie King (Bioreliance, Rockville) for their contributions to making this study possible. For the Singapore Chinese Health Study (SCHS), we thank Siew-Hong Low of the National University of Singapore for supervising the field work of the Singapore Chinese Health Study, the Singapore Cancer Registry for assistance with the identification of cancer outcomes. We are grateful to the radiologic technologists who have participated in the US Radiologic Technologists (USRT) study; Jerry Reid of the American Registry of Radiologic Technologists for continued support of the study; Diane Kampa and Allison Iwan of the University of Minnesota for cohort follow-up and data management; and Jeremy Miller of Information Management Systems, Inc. for computation and database support. The authors thank the Women’s Health Initiative (WHI) investigators and staff for their dedication, and the study participants for making the program possible. A full listing of WHI investigators can be found at https://www.whi.org/researchers/Documents%20%20Write%20a%20Paper/WHI%20Investigator%20Short%20List.pdf.
Conflict of interest: The authors have no conflicts of interest to declare.
References
- 1. International Agency for Research on Cancer. Tobacco Smoke and Involuntary Smoking. International Agency for Research on Cancer, 2004. [Google Scholar]
- 2. US Department of Health and Human Services. The Health Consequences of Smoking: A Report of the Surgeon General US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2004.
- 3. Hecht SS. Tobacco smoke carcinogens and breast cancer. Environ Mol Mutagen 2002;39:119–26. [DOI] [PubMed] [Google Scholar]
- 4. Petrakis NL, Gruenke LD, Beelen TC. et al. Nicotine in breast fluid of nonlactating women. Science 1978;199:303–5. [DOI] [PubMed] [Google Scholar]
- 5. Petrakis NL, Maack CA, Lee RE. et al. Mutagenic activity in nipple aspirates of human breast fluid. Cancer Res 1980;40:188–9. [PubMed] [Google Scholar]
- 6. Conway K, Edmiston SN, Cui L. et al. Prevalence and spectrum of p53 mutations associated with smoking in breast cancer. Cancer Res 2002;62:1987–95. [PubMed] [Google Scholar]
- 7. Li D, Zhang W, Sahin AA. et al. DNA adducts in normal tissue adjacent to breast cancer: a review. Cancer Detect Prev 1999;23:454–62. [DOI] [PubMed] [Google Scholar]
- 8. Rundle A, Tang D, Hibshoosh H. et al. The relationship between genetic damage from polycyclic aromatic hydrocarbons in breast tissue and breast cancer. Carcinogenesis 2000;21:1281–9. [PubMed] [Google Scholar]
- 9. Li D, Wang M, Dhingra K. et al. Aromatic DNA adducts in adjacent tissues of breast cancer patients: clues to breast cancer etiology. Cancer Res 1996;56:287–93. [PubMed] [Google Scholar]
- 10. Firozi PF, Bondy ML, Sahin AA. et al. Aromatic DNA adducts and polymorphisms of CYP1A1, NAT2, and GSTM1 in breast cancer. Carcinogenesis 2002;23:301–6. [DOI] [PubMed] [Google Scholar]
- 11. US Department of Health and Human Services. The Health Consequences of Smoking: 50 Years of Progress: A Report of the Surgeon General US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2014.
- 12. California Environmental Protection Agency. Health Effects Assessment for ETS: Final. California Environmental Protection Agency, 2005. [Google Scholar]
- 13. Collishaw NE,, Boyd NF,, Cantor KP. et al. Canadian Expert Panel on Tobacco Smoke and Breast Cancer Risk. Ontario Tobacco Research Unit, 2009. [DOI] [PubMed] [Google Scholar]
- 14. Secretan B, Straif K, Baan R. et al. A review of human carcinogens—Part E: tobacco, areca nut, alcohol, coal smoke, and salted fish. Lancet Oncol 2009;10:1033–4. [DOI] [PubMed] [Google Scholar]
- 15. US Department of Health and Human Services. Women and Smoking: A Report of the Surgeon General US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2001.
- 16. US Department of Health and Human Services. The Health Consequences of Involuntary Exposure to Tobacco Smoke: A Report of the Surgeon General US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2006.
- 17. Seitz HK, Pelucchi C, Bagnardi V. et al. Epidemiology and Pathophysiology of Alcohol and Breast Cancer: Update 2012. Alcohol Alcohol, 2012. [DOI] [PubMed] [Google Scholar]
- 18. Hamajima N, Hirose K, Tajima K. et al. Alcohol, tobacco and breast cancer—collaborative reanalysis of individual data from 53 epidemiological studies, including 58,515 women with breast cancer and 95,067 women without the disease. Br J Cancer 2002;87:1234–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Dossus L, Boutron-Ruault MC, Kaaks R. et al. Active and passive cigarette smoking and breast cancer risk: results from the EPIC cohort. Int J Cancer 2014;134:1871–88. [DOI] [PubMed] [Google Scholar]
- 20. Gram IT,, Little MA,, Lund E. et al. The fraction of breast cancer attributable to smoking: the Norwegian women and cancer study 1991–2012. Br J Cancer 2016;115:616–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Gram IT, Park SY, Kolonel LN. et al. Smoking and risk of breast cancer in a racially/ethnically diverse population of mainly women who do not drink alcohol: the MEC Study. Am J Epidemiol 2015;182:917–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Russo J, Hu YF, Yang X. et al. Developmental, cellular, and molecular basis of human breast cancer. J Natl Cancer Inst Monogr 2000;27:17–37. [DOI] [PubMed] [Google Scholar]
- 23. Gaudet MM, Gapstur SM, Sun J. et al. Active smoking and breast cancer risk: original cohort data and meta-analysis. J Natl Cancer Inst 2013;105:515–25. [DOI] [PubMed] [Google Scholar]
- 24. Nyante SJ, Gierach GL, Dallal CM. et al. Cigarette smoking and postmenopausal breast cancer risk in a prospective cohort. Br J Cancer 2014;110:2339–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Catsburg C, Miller AB, Rohan TE.. Active cigarette smoking and risk of breast cancer. Int J Cancer 2015;136:2204–9. [DOI] [PubMed] [Google Scholar]
- 26. Rosenberg L, Boggs DA, Bethea TN. et al. A prospective study of smoking and breast cancer risk among African-American women. Cancer Causes Control 2013;24:2207–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Trentham-Dietz A, Nichols HB, Egan KM. et al. Cigarette smoking and risk of breast carcinoma in situ. Epidemiology 2007;18:629–38. [DOI] [PubMed] [Google Scholar]
- 28. Breslow NE, Day NE. Statistical Methods in Cancer Research: Volume II—The Design and Analysis of Cohort Studies. IARC Sci Publ, 1987, pp. 1–406. [PubMed] [Google Scholar]
- 29. Higgins JP, Thompson SG.. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539–58. [DOI] [PubMed] [Google Scholar]
- 30. Higgins JP, Thompson SG, Deeks JJ. et al. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Cochrane Handbook for Systematic Reviews of Interventions 2011, cited 2016. http://handbook.cochrane.org/front_page.htm (15 July 2016, date last accessed).
- 32. Xue X, Kim MY, Gaudet MM. et al. A comparison of the polytomous logistic regression and joint cox proportional hazards models for evaluating multiple disease subtypes in prospective cohort studies. Cancer Epidemiol Biomarkers Prev 2013;22:275–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Liu Y, Colditz GA, Rosner B. et al. Alcohol intake between menarche and first pregnancy: a prospective study of breast cancer risk. J Natl Cancer Inst 2013;105:1571–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Egan KM, Stampfer MJ, Hunter D. et al. Active and passive smoking in breast cancer: prospective results from the Nurses’ Health Study. Epidemiology 2002;13:138–45. [DOI] [PubMed] [Google Scholar]
- 35. Terry PD, Rohan TE.. Cigarette smoking and the risk of breast cancer in women: a review of the literature. Cancer Epidemiol Biomarkers Prev 2002;11:953–71. [PubMed] [Google Scholar]
- 36. Catsburg C, Kirsh VA, Soskolne CL. et al. Active cigarette smoking and the risk of breast cancer: a cohort study. Cancer Epidemiology 2014;38:376–81. [DOI] [PubMed] [Google Scholar]
- 37. Centers for Disease Control and Prevention. Tobacco product use among middle and high school students—United States, 2011 and 2012. MMWR 2013;62:893–7. [PMC free article] [PubMed] [Google Scholar]
- 38. Warren GW, Alberg AJ, Kraft AS. et al. The 2014 Surgeon General’s report: ‘The health consequences of smoking—50 years of progress’: a paradigm shift in cancer care. Cancer 2014;120:1914–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Glantz SA, Johnson KC.. The surgeon general report on smoking and health 50 years later: breast cancer and the cost of increasing caution. Cancer Epidemiol Biomarkers Prev 2014;23:37–46. [DOI] [PubMed] [Google Scholar]
- 40. Martinez G,, Daniels K,, Chandra A.. Fertility of men and women aged 15–44 years in the United States: National Survey of Family Growth, 2006–2010. National Health Statistics Reports 2012;51:1–28. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.