Abstract
Tobacco smoke is an established carcinogen, but the association between tobacco smoking and cancer risk in BRCA mutation carriers is not clear. The aim of this study was to evaluate prospectively the association between tobacco smoking and cancer incidence in a cohort of BRCA1 and BRCA2 mutation carriers. The study population consisted of unaffected BRCA mutation carriers. Information on lifestyle including smoking histories, reproductive factors, and past medical histories was obtained through questionnaires. Incident cancers were updated biennially via follow-up questionnaires. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using time-dependent Cox regression models. There were 700 incident cancers diagnosed over 26,711 person-years of follow-up. The most frequent cancers seen in BRCA mutation carriers were breast (n = 428; 61%) and ovarian (n = 109; 15%) cancer. Compared to non-smokers, (ever) smoking was associated with a modest increased risk of all cancers combined (HR = 1.17; 95%CI 1.01–1.37). Women in the highest group of total pack-years (4.3–9.8) had an increased risk of developing any cancer (HR = 1.27; 95%CI 1.04–1.56), breast cancer (HR=1.33, 95%CI 1.02–1.75), and ovarian cancer (HR = 1.68; 95%CI 1.06–2.67) compared to never smokers. The associations between tobacco smoking and cancer did not differ by BRCA mutation type or by age at diagnosis. This prospective study suggests that tobacco smoking is associated with a modest increase in the risks of breast and ovarian cancer among women with BRCA1 or BRCA2 mutation.
Keywords: BRCA1, BRCA2, smoking, breast cancer, ovarian cancer
Introduction
Cigarette smoking has been categorized as a group 1 carcinogen by the International Agency for Research, in particular for cancers of the lung, bladder, liver, and colon.1, 2 Although smoking and the risk of developing breast or ovarian cancer is not conclusive, recent meta-analyses of the prospective studies suggest a 13% increased risk of breast cancer (relative risk [RR] = 1.13; 95% confidence intervals [CI] 1.09–1.17) and no increase in the risk of ovarian cancer (RR = 1.06; 95%CI 0.99–1.12) for current smokers compared to never smokers.3, 4 These risk estimates are derived from studies conducted in the general population.
Women with inherited mutations in BRCA1 or BRCA2 face high lifetime risks of developing breast and ovarian and other cancers.5, 6 The incomplete penetrance of a BRCA mutation implies that exogenous factors may modify cancer risk.7, 8 The identification of modifiable exposures is critical for this high-risk population, given that effective prevention strategies are currently limited to prophylactic surgery. Cigarette smoking is of interest in this respect because it is modifiable. Furthermore, because of the important role of the BRCA proteins in the repair of double stranded DNA breaks as well as damage due to oxidative stress, mutation carriers may be more susceptible to the relevant carcinogens.9
It is not clear if cigarette smoking is a risk factor for cancer in this high-risk population. A number of earlier reports, including two from our group, have yielded inconclusive results and have been limited by small sample sizes and retrospective study designs, which are vulnerable to selection bias and information bias.10–16 We undertook a prospective evaluation of cigarette smoking and overall cancer incidence, as well as breast and ovarian cancer incidence, in a large cohort of BRCA1 and BRCA2 mutation carriers.
Materials and Methods
Study population
The study population was selected from a multicenter longitudinal cohort of BRCA1 and BRCA2 mutation carriers from 80 participating centers in 17 countries including North America, Europe, Asia, the Caribbean, and Latin America. These women sought genetic testing for a BRCA1 or BRCA2 mutation because of a personal or family history of breast and/or ovarian cancer and genetic counselling was conducted for all study subjects (with the exception of some from the University of Utah and University of California Irvine). Mutation detection was conducted using a range of techniques, but all nucleotide sequences were confirmed by direct sequencing of DNA. Informed consent was obtained from all participants and the study protocol was approved by the institutional ethics review board of each host institution.
Data collection and ascertainment of incident cancers
All study subjects completed a baseline questionnaire at the individual center at the time of a clinic appointment or at their home at a later date. The questionnaire requested information on family and personal history of cancer, reproductive and medical histories, including preventive oophorectomy and mastectomy. Follow-up questionnaires were completed every two years thereafter and administered by mail or completed over the phone by a genetic counsellor or research assistant. At each follow-up, participants provided updates on exposures, preventive surgery and cancer incidence. The questionnaire also collected detailed information on past and current regular cigarette smoking history including: age first started to smoke, age last smoked, and the average number of cigarette packs smoked per week. Incident breast cancers (invasive breast cancer only) and ovarian cancers were identified by self-report, and 85% of breast cancer cases and 63% of ovarian cancer cases were confirmed by review of pathology reports. In the contributing Korean institutions, incident cancer cases were ascertained by the national cancer registry.
Study Subjects Available for Analysis
There were 15,731 women with a BRCA1 or BRCA2 mutation who were potentially eligible for inclusion in the current study. Of the 15,731 women, we excluded subjects who had a previous diagnosis of any cancer (n = 8,541), had no information on smoking history (n = 542), or had no follow-up information (n = 2,372). After these exclusions, a total of 4,276 subjects qualified for inclusion in the analysis. For the analysis of tobacco smoking with breast and ovarian cancer, we additionally excluded the subjects who had a prophylactic mastectomy (n = 209) and oophorectomy (n = 854) at baseline, respectively, and we excluded those who had missing information on prophylactic surgery during the follow-up period (n = 147). There was a total of 3,920 subjects in the analysis for breast cancer and 3,275 subjects for ovarian cancer.
Statistical analysis
Time dependent Cox regression analysis was used to estimate the hazard ratios (HRs) and 95% confidence intervals (CIs) of all cancer as well as breast and ovarian cancer, with smoking. All the smoking exposures, as well as the covariates, included in the analyses were treated as time-dependent variables and updated to reflect the any changes following completion of the baseline questionnaire or at any point in the follow-up. In the analysis of smoking and all cancers (as well as breast cancer specific incidence), the continuous variables of exposures were categorized into quartiles groups based on the distribution in the entire cohort. For the analysis of ovarian cancer, continuous variables of exposures were categorized into 3 groups (never smoker, and dichotomization of exposed group by median) based on the distribution in the entire cohort.
We created the following smoking variables: 1) smoking history (ever, never), 2) current smoking history (non-smoker, past smoker, current smoker), 3) age first started to smoke (never, ≥ 20 years, 18–19 years, ≤ 17 years for overall and breast cancer analysis; never, > 19 years, ≤ 19 years for ovarian cancer analysis), 4) average packs smoked per week (never, ≤ 2 packs, 2 < packs ≤ 5, >5 packs for overall and breast cancer analysis; never, ≤ 3 packs, > 3 packs for ovarian cancer analysis), 5) smoking duration (never, ≤ 8 years, 8 < years ≤ 18, >18 years for overall and breast cancer analysis; never, ≤ 10 years, > 10 years for ovarian cancer analysis), and 6) total pack-years (never, ≤ 2.3 pack-years, 2.3 < pack-years ≤ 9.8, > 9.8 pack-years for overall and breast cancer analysis; never, ≤ 4.3 pack-years, > 4.3 pack-years for ovarian cancer analysis). Total pack-years smoked was estimated by the following equation: (cigarette packs smoked per day) × (total years smoked). 7) duration of smoking before first birth among parous women (never smoker, smoking initiation after first birth, ≤ 5 years, and > 5 years). For the analysis of overall cancer incidence, participants were followed from the date of baseline questionnaire until either the: 1) date of completion of the last follow-up questionnaire, 2) date of cancer diagnosis, or 3) date of death. For the analysis of breast cancer incidence, participants were followed from the date of baseline questionnaire until either the: 1) date of completion of the last follow-up questionnaire, 2) date of breast cancer or ovarian cancer diagnosis, 3) date of prophylactic mastectomy, or 4) date of death and similarly for ovarian cancer incidence except for censoring at date of prophylactic oophorectomy instead of date of prophylactic mastectomy.
The multivariate model was adjusted for age (continuous), country of residence (Canada and United States/Europe/others [including Asia and Latin America]), BRCA mutation type (BRCA1 or BRCA2), pregnancy history (ever/never), oral contraceptive use (ever/never), regular alcohol consumption (ever/never) and obesity (normal, less than 25 kg/m2/overweight, 25–29.9 kg/m2/obese, 30 kg/m2 or more). We also performed analyses stratified by BRCA1 and BRCA2 mutation, age at breast cancer diagnosis (<50 vs. ≥50 years), and ER status of breast cancer cases.
The P for trend was estimated by treating the numerical values of the categorical variables as a score in time-dependent Cox regression model to assess dose-response relationships. All analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC, USA) and a P value less than 0.05 was considered as statistical significance.
Results
A total of 4,276 women with a BRCA1 or BRCA2 mutation were included in the current prospective analysis. The age-adjusted baseline characteristics by smoking history are shown in Table 1. Women who smoked were, on average, older (40.4 years vs. 38.1 years), more likely to be BRCA1 mutation carriers (78% vs. 21%), to have been pregnant (43% vs. 31%), and to have breastfed for a longer duration (compared to women who never smoked (all comparisons P < 0.01). Smokers were more likely to have used oral contraceptives (42% vs. 36%), to have a greater BMI (mean 24.7 vs. 24.0 kg/m2), to live in Europe (54% vs. 43%), to consume alcohol (45% vs. 28%) and to consume coffee regularly (46% vs. 24%) compared with never smokers (all comparisons P < 0.01).
Table 1.
Baseline characteristics of BRCA mutation carriers by smoking status
| Variables | Non-smoker (n=2,566) | Smoker (n=1,710) | P-value* |
|---|---|---|---|
| Age, years (mean ± SD) | 38.1 ± 12.6 | 40.4 ± 11.6 | < 0.01 |
| BMI, kg/m2 (mean ± SD) | 24.0 ± 4.8 | 24.7 ± 5.0 | < 0.01 |
| Country, n (%) | |||
| US/Canada | 1,248 (48.6) | 747 (43.7) | |
| Europe | 1,113 (43.4) | 923 (54.0) | |
| The others | 205 (8.0) | 40 (2.3) | < 0.01 |
| BRCA mutation, n (%) | |||
| BRCA1 mutation | 1,918 (74.8) | 1,338 (78.3) | |
| BRCA2 mutation | 627 (24.4) | 360 (21.1) | |
| BRCA1 and BRCA2 mutation or unknown | 21 (0.8) | 12 (0.7) | < 0.01 |
| Pregnancy history, n (%) | |||
| Never | 781 (30.5) | 349 (20.5) | |
| Ever | 1,782 (69.5) | 1,357 (79.5) | < 0.01 |
| Breastfeeding duration, n (%) | |||
| Never | 1,033 (44.6) | 624 (40.1) | |
| Less than 6 months | 430 (18.6) | 398 (25.6) | |
| More than 6 months | 854 (36.9) | 533 (34.3) | < 0.01 |
| Menarche age, n (%) | |||
| ≤ 13 years | 886 (35.6) | 581 (34.7) | |
| 14 years | 713 (28.6) | 467 (27.9) | |
| ≥15 years | 892 (35.8) | 627 (37.4) | 0.75 |
| Menopausal status, n (%) | |||
| Premenopausal | 1,841 (72.1) | 1,114 (65.2) | |
| Postmenopausal | 711 (27.9) | 596 (34.8) | 0.40 |
| Oral contraceptive use, n (%) | |||
| Never | 1,027 (40.2) | 576 (33.9) | |
| Ever | 1,529 (59.8) | 1,121 (66.1) | < 0.01 |
| HRT use, n (%) | |||
| Never | 2,192 (86.1) | 1,398 (82.7) | |
| Ever | 354 (13.9) | 292 (17.3) | 0.37 |
| Tamoxifen use, n (%) | |||
| Never | 2,325 (97.8) | 1,619 (98.0) | |
| Ever | 51 (2.2) | 33 (2.0) | 0.21 |
| Regular alcohol drinker, n (%) | |||
| No | 867 (34.3) | 339 (20.2) | |
| Yes | 1,661 (65.7) | 1,340 (79.8) | < 0.01 |
| Regular coffee consumer, n (%) | |||
| No | 818 (34.2) | 253 (15.6) | |
| Yes | 1,574 (65.8) | 1,365 (84.4) | < 0.01 |
| Prophylactic mastectomy history, n (%) | |||
| No | 2,407 (95.1) | 1,589 (94.9) | |
| Yes | 123 (4.9) | 86 (5.1) | 0.95 |
| Oophorectomy history, n (%) | |||
| No | 2,063 (81.2) | 1,325 (77.9) | |
| Yes | 477 (18.8) | 377 (22.1) | 0.75 |
adjusted for age using logistic regression
BMI, body mass index; HRT, hormone replacement therapy
Average follow up time for smokers and nonsmokers was 6.5 and 6.1 years, respectively. Average age at diagnosis of any cancer for smokers and nonsmokers was 49.3 and 47.4 years old, respectively. During the entire follow-up period of 26,711 person-years, 700 incident cancers were identified. These are summarized by BRCA mutation type in Supplemental table 1. Among the 3,256 BRCA1 mutation carriers, there were 544 incident cancers and among the 987 BRCA2 mutation carriers, there were 153 incident cancers. The most commonly diagnosed cancers in BRCA mutation carriers were breast cancer (n=339; 10.4% in BRCA1 mutation carriers and n=87; 8.8% in BRCA2 mutation carriers) and ovarian cancer (n=103; 3.2% in BRCA1 mutation carriers and n=6; 0.6% in BRCA2 mutation carriers). Other commonly diagnosed cancers included ductal carcinoma in situ or lobular carcinoma in situ, as well as cancers of the skin, peritoneal and fallopian tube. With respect to smoking related cancers, there were nine pancreatic cancers, seven lung cancers and one bladder cancer diagnosed in the entire cohort.
The association between cigarette smoking and the risk of developing cancer at any site is shown in Table 2. Ever smokers had a 17% increased risk of developing any cancer compared to never smokers (HR=1.17; 95% CI 1.01–1.37). Current smokers had a 31% increased risk of developing cancer compared to never smokers (HR=1.31; 95% CI 1.07–1.59; P trend = 0.01). Earlier age at first smoking was associated with a significant increasing risk of developing cancer compared to never smoking (P-trend = 0.03). Increasing duration of smoking was associated with a significant increasing risk of developing cancer compared to never smoking (P-trend < 0.01). Women in the highest category of total pack-years of smoking (i.e., >9.8) had a 27% increased risk of cancer compared to women who never smoked (HR=1.27; 95% CI 1.04–1.56; P -trend = 0.02). Average packs of cigarette smoked per week (P-trend = 0.13) were not associated with the risk of any cancer.
Table 2.
Association between smoking and risk of all cancer incidence among BRCA mutation carriers
| Person-years (Total 26,711) | Incident cases (N = 700) | Crude HR (95% CI) | Adjusted HR* (95% CI) | |
|---|---|---|---|---|
| Smoking history | ||||
| Never | 15,335 | 359 | Reference | Reference |
| Ever | 11,377 | 341 | 1.28 (1.11–1.49) | 1.17 (1.01–1.37) |
| Current smoking history | ||||
| Past smoker | 6,243 | 184 | 1.27 (1.06–1.51) | 1.09 (0.91–1.31) |
| Current smoker | 5,050 | 153 | 1.29 (1.07–1.56) | 1.31 (1.07–1.59) |
| P for trend = | < 0.01 | 0.01 | ||
| Age at first smoking | ||||
| ≥ 20 years | 3,767 | 115 | 1.31 (1.06–1.61) | 1.08 (0.87–1.34) |
| 18–19 years | 3,129 | 93 | 1.27 (1.01–1.60) | 1.20 (0.95–1.51) |
| ≤ 17 years | 4,247 | 119 | 1.20 (0.98–1.48) | 1.22 (0.99–1.51) |
| P for trend = | < 0.01 | 0.03 | ||
| Average packs smoked/week | ||||
| <2 packs | 3,006 | 80 | 1.14 (0.90–1.46) | 1.09 (0.85–1.39) |
| 2 ≤ packs < 5 | 3,757 | 101 | 1.15 (0.92–1.43) | 1.10 (0.88–1.38) |
| 5 ≤ packs | 3,685 | 115 | 1.34 (1.08–1.65) | 1.17 (0.95–1.45) |
| P for trend = | 0.01 | 0.13 | ||
| Smoking duration | ||||
| ≤ 8 years | 3,796 | 70 | 0.79 (0.61–1.02) | 0.88 (0.68–1.14) |
| 8 < duration ≤ 18 | 3,400 | 93 | 1.17 (0.93–1.47) | 1.18 (0.93–1.48) |
| 18 < duration | 3,968 | 166 | 1.80 (1.50–2.17) | 1.37 (1.13–1.66) |
| P for trend = | < 0.01 | < 0.01 | ||
| Total pack-years | ||||
| ≤ 2.3 pack-years | 3,424 | 71 | 0.89 (0.69–1.15) | 0.96 (0.75–1.25) |
| 2.3< pack-years ≤ 9.8 | 3,342 | 91 | 1.16 (0.92–1.46) | 1.12 (0.89–1.41) |
| > 9.8 pack-years | 3,534 | 131 | 1.59 (1.30–1.95) | 1.27 (1.04–1.56) |
| P for trend = | < 0.01 | 0.02 |
adjusted for age, country, BRCA mutation, pregnancy history, contraceptive drug use, alcohol drinking and obesity using time-dependent Cox regression model
Table 3 summarizes the results for the analysis of cigarette smoking and breast cancer risk. Ever smoking was not significantly associated with the risk of breast cancer (HR = 1.16; 95%CI 0.96–1.41). Current smokers had a 28% increased risk of breast cancer compared to never smokers (HR=1.28; 95% CI 1.00–1.64; P -trend = 0.07). Women in the highest quartile of smoking duration had a 34% increased risk of developing breast cancer compared to women who never smoked (HR=1.34; 95% CI 1.04–1.73; P -trend = 0.02). Women in the highest quartile of total pack-years had a 33% increased risk of developing breast cancer compared to never smokers (HR=1.33; 95% CI 1.02–1.75; P-trend = 0.08). There was no significant association between age at smoking initiation (P-trend = 0.39), duration of smoking before the first birth among parous women (P-trend = 0.77) or average packs smoked per week (P-trend = 0.22) and the risk of developing breast cancer.
Table 3.
Association between smoking and risk of breast cancer among BRCA mutation carriers
| Person-years (Total 21,098) | Incident cases (N = 420) | Crude HR (95% CI) | Adjusted HR* (95% CI) | |
|---|---|---|---|---|
| Smoking history | ||||
| Never | 11,990 | 217 | Reference | Reference |
| Ever | 9,108 | 203 | 1.22 (1.01–1.48) | 1.16 (0.96–1.41) |
| Current smoking history | ||||
| Past smoker | 4,835 | 104 | 1.16 (0.92–1.47) | 1.04 (0.82–1.32) |
| Current smoker | 4,250 | 95 | 1.24 (0.98–1.58) | 1.28 (1.00–1.64) |
| P for trend = | 0.06 | 0.07 | ||
| Age at first smoking | ||||
| ≥ 20 years | 3,165 | 75 | 1.29 (0.99–1.67) | 1.15 (0.88–1.51) |
| 18–19 years | 2,584 | 52 | 1.10 (0.82–1.49) | 1.08 (0.79–1.47) |
| ≤ 17 years | 3,223 | 65 | 1.11 (0.84–1.45) | 1.12 (0.84–1.48) |
| P for trend = | 0.34 | 0.39 | ||
| Average packs smoked/week | ||||
| <2 packs | 2,413 | 47 | 1.06 (0.77–1.45) | 1.04 (0.76–1.43) |
| 2 ≤ packs < 5 | 3,086 | 59 | 1.05 (0.79–1.41) | 1.04 (0.77–1.39) |
| 5 ≤ packs | 2,783 | 67 | 1.32 (1.01–1.74) | 1.22 (0.92–1.61) |
| P for trend = | 0.08 | 0.22 | ||
| Smoking duration | ||||
| ≤ 8 years | 3,050 | 39 | 0.71 (0.50–0.99) | 0.76 (0.54–1.07) |
| 8 < duration ≤ 18 | 2,750 | 60 | 1.20 (0.90–1.60) | 1.21 (0.90–1.61) |
| 18 < duration | 3,172 | 93 | 1.59 (1.24–2.02) | 1.34 (1.04–1.73) |
| P for trend = | < 0.01 | 0.02 | ||
| Total pack-years | ||||
| ≤ 2.3 pack-years | 2,800 | 43 | 0.84 (0.61–1.17) | 0.89 (0.64–1.24) |
| 2.3< pack-years ≤ 9.8 | 2,680 | 51 | 1.05 (0.77–1.42) | 1.01 (0.74–1.38) |
| > 9.8 pack-years | 2,720 | 76 | 1.53 (1.18–1.98) | 1.33 (1.02–1.75) |
| P for trend = | 0.01 | 0.08 | ||
| Duration of smoking before the first birth among parous women | ||||
| Never smoker | 7,905 | 168 | Reference | Reference |
| Initiation after first birth | 952 | 27 | 1.33 (0.88–1.99) | 1.29 (0.85–1.96) |
| ≤ 5 years | 2,606 | 48 | 0.86 (0.63–1.19) | 0.86 (0.62–1.19) |
| > 5 years | 2,858 | 69 | 1.14 (0.86–1.50) | 1.12 (0.84–1.50) |
| P for trend = | 0.69 | 0.77 | ||
adjusted for age, country, BRCA mutation, pregnancy history, contraceptive drug use, alcohol drinking and obesity using time-dependent Cox regression model
Past smokers had a 69% increased risk of developing ovarian cancer compared to never smokers (HR 1.69; 95% CI 1.06–2.71)(Table 4). Current smoking was not significantly associated with risk (HR = 1.25; 95%CI 0.73–2.12). Women who smoked for 10 or more years had a 63% increased risk of developing ovarian cancer compared to women who never smoked (HR=1.63; 95% CI 1.04–2.54). Total pack-years of more than 4.3 was associated with the increased risk of ovarian cancer compared with never smoking (HR=1.68; 95% CI 1.06–2.67). There was a significant dose-response relationship between both smoking duration and total pack-years and the risk of developing ovarian cancer (P-trend = 0.03).
Table 4.
Association between smoking and risk of ovarian cancer among BRCA mutation carriers
| Person-years (Total 15,720) | Incident cases (N = 100) | Crude HR (95% CI) | Adjusted HR* (95% CI) | |
|---|---|---|---|---|
| Smoking history | ||||
| Never | 9,260 | 49 | Reference | Reference |
| Ever | 6,460 | 51 | 1.48 (1.00–2.19) | 1.45 (0.96–2.19) |
| Current smoking history | ||||
| Past smoker | 3,036 | 30 | 1.84 (1.16–2.90) | 1.69 (1.06–2.71) |
| Current smoker | 3,375 | 21 | 1.17 (0.70–1.96) | 1.25 (0.73–2.12) |
| P for trend = | 0.25 | 0.22 | ||
| Age at first smoking | ||||
| > 19 years | 2,531 | 30 | 2.22 (1.41–3.50) | 1.69 (1.05–2.72) |
| ≤ 19 years | 3,627 | 19 | 0.98 (0.58–1.66) | 1.18 (0.68–2.05) |
| P for trend = | 0.52 | 0.28 | ||
| Average packs smoked/week | ||||
| ≤ 3 packs | 3,417 | 27 | 1.48 (0.93–2.37) | 1.72 (1.05–2.81) |
| 3 packs < | 2,700 | 19 | 1.30 (0.76–2.20) | 1.17 (0.68–2.01) |
| P for trend = | 0.20 | 0.33 | ||
| Smoking duration | ||||
| ≤ 10 years | 3,205 | 15 | 0.88 (0.50–1.58) | 1.22 (0.67–2.22) |
| 10 years < | 3,154 | 36 | 2.12 (1.38–3.27) | 1.63 (1.04–2.54) |
| P for trend = | < 0.01 | 0.03 | ||
| Total pack-years | ||||
| ≤ 4.3 pack-years | 3,187 | 14 | 0.83 (0.46–1.50) | 1.11 (0.60–2.04) |
| 4.3 pack-years < | 2,871 | 32 | 2.06 (1.32–3.22) | 1.68 (1.06–2.67) |
| P for trend = | 0.01 | 0.03 | ||
adjusted for age, country, BRCA mutation, pregnancy history, contraceptive drug use, and alcohol drinking using time-dependent Cox regression model
The associations between tobacco smoking and overall cancer, as well as breast and ovarian cancer were similar in analyses stratified by BRCA mutation type (P heterogeneity > 0.05, Supplementary tables 2–4) and age at diagnosis (P heterogeneity > 0.05; data not shown).
Discussion
In this prospective evaluation of cigarette smoking and cancer risk among women with an inherited BRCA1 or BRCA2 mutation, we observed a borderline significant increased risk of developing any cancer, as well as breast and ovarian cancer specifically, with a history of smoking. Compared to never smokers, both increasing duration of smoking and pack-years of smoking were associated with a significant 1.3-fold increased risk of all cancers and of breast cancer, as well as a 1.6-fold increased risk of ovarian cancer. Although we found no significant associations between age at first use or average packs smoked per week, there was a consistent and significant dose-response relationship between lifetime dose of smoking (measured as total pack-years) and long-term exposure (measured as duration of smoking) with cancer risk. The results from this large prospective analyses implicate smoking as a likely carcinogen in this high-risk population, which we estimate confers a 30% and 60% increased risk of developing BRCA-associated breast and ovarian cancer, respectively.
Among studies conducted in the general population, causal links between cigarette smoking and 17 different types of cancers of the lung, bladder, liver, colorectal cancer has previously confirmed.2 Although the number of smoking-related cancers in the current study was low, we observed a 1.3-fold increased risk of cancer overall with smoking history (including duration of smoking and total pack-years) which is in accordance with risk estimates from epidemiologic reports stemming from women in the general population. To our knowledge, there are no prior evaluations of smoking and total BRCA-cancer incidence.
Although the evidence on tobacco smoking for breast cancer is not sufficient to infer a causal relationship in general population, a recent meta-analysis of the prospective studies estimated a significant 13% increased risk of breast cancer with current smoking (summary RR = 1.13; 95% CI 1.09–1.17).3 The observed adverse effect of smoking was greater than this in our cohort of high-risk women, which current smoking was associated with 28% increased risk of breast cancer. Also we observed a 33%-34% increased risk of BRCA-associated breast cancer with increasing duration and pack-years. A retrospective cohort study of smoking and cancer in BRCA mutation carriers, which included 990 women, reported a 2.1-fold increased risk of breast cancer among women with more than 20 pack-years of smoking.15 Limitations of the latter analysis is that follow-up was initiated at birth and women were followed until a cancer diagnosis or the date of interview and prevalent cases were included.
Our current findings contrast with our earlier reports that similarly assessed the association between smoking and breast cancer among BRCA mutation carriers (a subset of which were also included in the current analysis).12, 13 In the first report by our group which included 186 matched cases and controls, we reported a significant reduction in BRCA-breast cancer risk among women who smoked more than four pack-years (odds ratio [OR] = 0.46; 95% CI 0.27–0.80).14 In an updated analysis with a substantially larger number of BRCA mutation carriers (1,097 matched cases and controls), Ghadirian et al., found no association between smoking and the risk of breast cancer compared to never smoking (OR for ever smoking = 1.05; 95% CI 0.88–1.25 and OR for > 20 pack-years = 0.81; 95% CI 0.57–1.15).12 Finally, in our most recent report with 2,538 matched pairs, ever smoking were not associated with the risk of breast cancer regardless of BRCA1 or BRCA2 mutation, although a past history of smoking was associated with a significantly increased risk of breast cancer among BRCA1 carriers but not BRCA2 mutation carriers.13
In a small retrospective cohort study of BRCA1 mutation carriers (176 cases and 140 non-cases), ever smoking was associated with a decreased risk of breast cancer (HR = 0.63; 95% CI 0.47–0.87) and this inverse association was attenuated in the women with a 28 repeat allele for steroid receptor co-activator gene AIB1 (hazard ratio = 0.19; 95% CI 0.07–0.54).11 A case-control study composed of BRCA mutation aged less than 50 years old with reported that smoking history and high pack-years of smoking were associated with an increased risk of breast cancer with an estimated average of 7% increase per pack-year of smoking in BRCA mutation carriers.10 These prior reports were retrospective analyses and limited by a relatively smaller number of BRCA mutation carriers. Furthermore, in the case-control studies published by our team, there was an average of eight years between the date of diagnosis to date of interview or genetic testing, and thus, the analyses were vulnerable to survival and recall bias.
A recent pooled analysis in the general population from 14 cohort studies found an 18% increased risk of developing breast cancer among women who smoked 10 or more years prior to their first pregnancy compared to women who never smoked (HR = 1.18; 95% CI 1.12–1.24).17 Duration of smoking before the first birth among parous women was emerging risk factor for breast cancer in relation to the timing of smoking initiation.18, 19 Given the early age at breast cancer onset among BRCA mutation carriers (between ages 30 and 50), it is plausible that early exposures may significantly impact risk. Thus, we also evaluated the effect of smoking by age at first use, duration of smoking before first birth, and age at diagnosis; however, we found no evidence for an effect of this classification of smoking exposure on risk. Our study showed that heavy smoking with long time exposure was the important attributor in the development of breast cancer regardless of exposure timing, diagnosed age for breast cancer, and BRCA mutation type.
We observed a borderline significant 1.6-fold increased risk of ovarian cancer among women in the highest tertile of both smoking duration and total pack-years compared to never smokers. One other study has evaluated the relationship between smoking and ovarian cancer among BRCA mutation carriers.20 In this matched case-control study of Polish women with BRCA1 mutations (150 cases and 150 controls), smoking was not associated with ovarian cancer (OR = 1.3; 95% CI 0.8–2.3).20 Among women in the general population the relationship between smoking and ovarian cancer varies by histologic subtype; current smoking is associated with an increased risk of mucinous ovarian cancer compared to never smoking (RR=1.79; 95% CI 1.47–2.17), a decreased risk of endometrioid ovarian cancer (RR=0.81; 95% CI 0.70–0.94)4 but is not associated with the serous subtype (RR=0.99; 95% CI 0.91–1.08) or the clear-cell subtype (RR=0.80; 95% CI 0.63–1.01) in meta-analysis from 51 studies. Histologically, mucinous ovarian cancer are similar to cancers of the intestinal epithelium, which are also well-known smoking-related cancer.2 The findings from studies of women in the general population (who are likely not mutation carriers) contrasts with our current findings given that BRCA-ovarian cancers are predominantly of the serous or endometrioid subtypes.4, 21
Tobacco contains more than 7,000 chemical compounds including benzo[a]pyrene, hydrocarbons, tobacco-specific nitrosamines, benzene, formaldehyde, carbon monoxide cyanide, and polonium, the majority of which are well-established carcinogens and toxins.16 These components of tobacco smoke have been shown to contribute to carcinogenesis via multiple pathways, including DNA binding and consequent induction of mutation, inflammation, oxidative stress, and epigenetic mechanisms.22 Given the important role of the BRCA proteins in the repair of double-stranded DNA breaks, this population of women may be more susceptible to damage caused which could be caused from oxygen species by smoking23 The underlying mechanisms to explain the carcinogenic effects of cigarette smoking for breast cancer include the detection of circulating metabolites such as polycyclic aromatic hydrocarbons, aromatic amines, and N-nitrosamines in breast tissue and p53 mutation.24, 25 In contrast, anti-carcinogenic effects of smoking for breast cancer have also been suggested because of the anti-estrogenic effects of tobacco26, 27 Specifically, smoking has been shown to have inverse association with mammographic density which is a biomarker of breast cnacer.28, 29
Our study had several limitations including the ascertainment of incident cancers via questionnaire and our inability to confirm all reported cancers by medical record review. Although this may have resulted in misclassification of the outcome variables, self-report of cancer incidence has previously been shown to be valid and reliable, particularly for breast cancer.30–32 The sensitivity of self-reported cancer in women has been estimated between 0.61 and 0.89, while the sensitivity for breast cancer (the major incident cancer in our population) is high (range 0.85–0.91).30–32 Loss to follow-up is inevitable in prospective studies and can result in selection bias attributed to differences in the distribution of the exposure variable among subjects with and without follow-up information.33 In our study population, we did not have follow-up information on 36 % (2,372 subjects of 6,648 subjects) of the baseline participants, and thus, we could not ascertain the cancer incidence. Nevertheless, the proportion study subjects with follow-up information did not differ by smoking status in our population of BRCA mutation carriers (i.e., 66% in ex-smokers or current smokers) vs. 63% in never smokers. Also we could not find obvious association in the stratified analysis because of small sample size.
Despite these limitations, the current study had several strengths including the large sample of BRCA mutation carriers (n = 4,276) with a relatively long follow-up period (median = 5.4 years and interquartile range = 3.1 to 8.8 years). The prospective nature of our study design allowed for data collection prior to disease incidence, minimizing both selection and information bias. We reduced the influence of measurement error by using biennial follow-up questionnaires to update both exposures and covariates.
In conclusion, findings from this large prospective analysis of smoking and BRCA-associated cancers clearly implicates smoking as a risk factor, particularly for breast and ovarian cancer, the most common type of cancers diagnosed in this high-risk population. The adverse health effects of smoking should be discussed by clinicians and genetic counsellors managing women with BRCA-associated cancers. The impact of smoking on cancers incidence in male BRCA mutation carriers warrants further study.
Supplementary Material
Novelty Impact Statement.
Findings from this large prospective analysis of smoking and BRCA-associated cancers clearly implicates smoking as a risk factor, particularly for breast and ovarian cancer, the most common type of cancers diagnosed in this high-risk population. The adverse health effects of smoking should be discussed by clinicians and genetic counsellors managing women with BRCA-associated cancers. The impact of smoking on cancers incidence in male BRCA mutation carriers warrants further study.
Acknowledgments
Sources of Support: JK is the recipient of a Cancer Care Ontario Research Chair in Population Studies and a Canadian Cancer Society Career Development Award in Prevention. SAN is the recipient of a Tier I Canada Research Chair. This work was supported by National Cancer Institute at the National Institutes of Health (R01CA74415 to S.L.N.) and Canadian Cancer Society Research Institute (703058 to S.A.N.).
References
- 1.International Agency for Research Center. IARC monograph on the evaluation of carcinogenic risks to humans; Personal habits and indoor combustions. 2012 [PMC free article] [PubMed] [Google Scholar]
- 2.Surgeon General’s Report. The Health Consequences of Smoking: 50 Years of Progress. Centers for Disease Prevention and Control: 2014. [PubMed] [Google Scholar]
- 3.Macacu A, Autier P, Boniol M, Boyle P. Active and passive smoking and risk of breast cancer: a meta-analysis. Breast Cancer Res Treat. 2015;154:213–24. doi: 10.1007/s10549-015-3628-4. [DOI] [PubMed] [Google Scholar]
- 4.Collaborative Group on Epidemiological Studies of Ovarian C. Beral V, Gaitskell K, Hermon C, Moser K, Reeves G, Peto R. Ovarian cancer and smoking: individual participant meta-analysis including 28,114 women with ovarian cancer from 51 epidemiological studies. Lancet Oncol. 2012;13:946–56. doi: 10.1016/S1470-2045(12)70322-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Miki Y, Swensen J, Shattuck-Eidens D, Futreal PA, Harshman K, Tavtigian S, Liu Q, Cochran C, Bennett LM, Ding W, et al. A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science. 1994;266:66–71. doi: 10.1126/science.7545954. [DOI] [PubMed] [Google Scholar]
- 6.Chen S, Parmigiani G. Meta-analysis of BRCA1 and BRCA2 penetrance. J Clin Oncol. 2007;25:1329–33. doi: 10.1200/JCO.2006.09.1066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tryggvadottir L, Sigvaldason H, Olafsdottir GH, Jonasson JG, Jonsson T, Tulinius H, Eyfjord JE. Population-based study of changing breast cancer risk in Icelandic BRCA2 mutation carriers, 1920–2000. Journal of the National Cancer Institute. 2006;98:116–22. doi: 10.1093/jnci/djj012. [DOI] [PubMed] [Google Scholar]
- 8.Nkondjock A, Ghadirian P. Epidemiology of breast cancer among BRCA mutation carriers: an overview. Cancer Lett. 2004;205:1–8. doi: 10.1016/j.canlet.2003.10.005. [DOI] [PubMed] [Google Scholar]
- 9.Gudmundsdottir K, Ashworth A. The roles of BRCA1 and BRCA2 and associated proteins in the maintenance of genomic stability. Oncogene. 2006;25:5864–74. doi: 10.1038/sj.onc.1209874. [DOI] [PubMed] [Google Scholar]
- 10.Breast Cancer Family R, Kathleen Cuningham Consortium for Research into Familial Breast C, Ontario Cancer Genetics N. Smoking and risk of breast cancer in carriers of mutations in BRCA1 or BRCA2 aged less than 50 years. Breast Cancer Res Treat. 2008;109:67–75. doi: 10.1007/s10549-007-9621-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Colilla S, Kantoff PW, Neuhausen SL, Godwin AK, Daly MB, Narod SA, Garber JE, Lynch HT, Brown M, Weber BL, Rebbeck TR. The joint effect of smoking and AIB1 on breast cancer risk in BRCA1 mutation carriers. Carcinogenesis. 2006;27:599–605. doi: 10.1093/carcin/bgi246. [DOI] [PubMed] [Google Scholar]
- 12.Ghadirian P, Lubinski J, Lynch H, Neuhausen SL, Weber B, Isaacs C, Baruch RG, Randall S, Ainsworth P, Friedman E, Horsman D, Tonin P, et al. Smoking and the risk of breast cancer among carriers of BRCA mutations. Int J Cancer. 2004;110:413–6. doi: 10.1002/ijc.20106. [DOI] [PubMed] [Google Scholar]
- 13.Ginsburg O, Ghadirian P, Lubinski J, Cybulski C, Lynch H, Neuhausen S, Kim-Sing C, Robson M, Domchek S, Isaacs C, Klijn J, Armel S, et al. Smoking and the risk of breast cancer in BRCA1 and BRCA2 carriers: an update. Breast Cancer Res Treat. 2009;114:127–35. doi: 10.1007/s10549-008-9977-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Brunet JS, Ghadirian P, Rebbeck TR, Lerman C, Garber JE, Tonin PN, Abrahamson J, Foulkes WD, Daly M, Wagner-Costalas J, Godwin A, Olopade OI, et al. Effect of smoking on breast cancer in carriers of mutant BRCA1 or BRCA2 genes. Journal of the National Cancer Institute. 1998;90:761–6. doi: 10.1093/jnci/90.10.761. [DOI] [PubMed] [Google Scholar]
- 15.Lecarpentier J, Nogues C, Mouret-Fourme E, Stoppa-Lyonnet D, Lasset C, Caron O, Fricker JP, Gladieff L, Faivre L, Sobol H, Gesta P, Frenay M, et al. Variation in breast cancer risk with mutation position, smoking, alcohol, and chest X-ray history, in the French National BRCA1/2 carrier cohort (GENEPSO) Breast Cancer Res Treat. 2011;130:927–38. doi: 10.1007/s10549-011-1655-3. [DOI] [PubMed] [Google Scholar]
- 16.U.S. Department of Health and Human Services. How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis for Smoking-Attributable Disease: A Report of the Surgeon General. Centers for Disease Control and Prevention; 2014. [PubMed] [Google Scholar]
- 17.Gaudet MM, Carter BD, Brinton LA, Falk RT, Gram IT, Luo J, Milne RL, Nyante SJ, Weiderpass E, Beane Freeman LE, Sandler DP, Robien K, et al. Pooled analysis of active cigarette smoking and invasive breast cancer risk in 14 cohort studies. Int J Epidemiol. 2016 doi: 10.1093/ije/dyw288. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Dossus L, Boutron-Ruault MC, Kaaks R, Gram IT, Vilier A, Fervers B, Manjer J, Tjonneland A, Olsen A, Overvad K, Chang-Claude J, Boeing H, et al. Active and passive cigarette smoking and breast cancer risk: results from the EPIC cohort. Int J Cancer. 2014;134:1871–88. doi: 10.1002/ijc.28508. [DOI] [PubMed] [Google Scholar]
- 19.Gram IT, Park SY, Kolonel LN, Maskarinec G, Wilkens LR, Henderson BE, Le Marchand L. 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: 10.1093/aje/kwv092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Gronwald J, Byrski T, Huzarski T, Cybulski C, Sun P, Tulman A, Narod SA, Lubinski J. Influence of selected lifestyle factors on breast and ovarian cancer risk in BRCA1 mutation carriers from Poland. Breast cancer research and treatment. 2006;95:105–9. doi: 10.1007/s10549-005-9051-5. [DOI] [PubMed] [Google Scholar]
- 21.Wentzensen N, Poole EM, Trabert B, White E, Arslan AA, Patel AV, Setiawan VW, Visvanathan K, Weiderpass E, Adami HO, Black A, Bernstein L, et al. Ovarian Cancer Risk Factors by Histologic Subtype: An Analysis From the Ovarian Cancer Cohort Consortium. J Clin Oncol. 2016;34:2888–98. doi: 10.1200/JCO.2016.66.8178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.International Agency for Research Center. Tobacco smoke and involuntary smoking. IARC; 2004. IARC monograph on the evaluation of carcinogenic risks to humans. [PMC free article] [PubMed] [Google Scholar]
- 23.Stern MC, Umbach DM, van Gils CH, Lunn RM, Taylor JA. DNA repair gene XRCC1 polymorphisms, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev. 2001;10:125–31. [PubMed] [Google Scholar]
- 24.Petrakis NL, Maack CA, Lee RE, Lyon M. Mutagenic activity in nipple aspirates of human breast fluid. Cancer Res. 1980;40:188–9. [PubMed] [Google Scholar]
- 25.Conway K, Edmiston SN, Cui L, Drouin SS, Pang J, He M, Tse CK, Geradts J, Dressler L, Liu ET, Millikan R, Newman B. Prevalence and spectrum of p53 mutations associated with smoking in breast cancer. Cancer Res. 2002;62:1987–95. [PubMed] [Google Scholar]
- 26.Perera FP, Estabrook A, Hewer A, Channing K, Rundle A, Mooney LA, Whyatt R, Phillips DH. Carcinogen-DNA adducts in human breast tissue. Cancer Epidemiol Biomarkers Prev. 1995;4:233–8. [PubMed] [Google Scholar]
- 27.Baron JA, La Vecchia C, Levi F. The antiestrogenic effect of cigarette smoking in women. Am J Obstet Gynecol. 1990;162:502–14. doi: 10.1016/0002-9378(90)90420-c. [DOI] [PubMed] [Google Scholar]
- 28.Butler LM, Gold EB, Conroy SM, Crandall CJ, Greendale GA, Oestreicher N, Quesenberry CP, Jr, Habel LA. Active, but not passive cigarette smoking was inversely associated with mammographic density. Cancer Causes Control. 2010;21:301–11. doi: 10.1007/s10552-009-9462-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Jacobsen KK, Lynge E, Vejborg I, Tjonneland A, von Euler-Chelpin M, Andersen ZJ. Cigarette smoking and mammographic density in the Danish Diet, Cancer and Health cohort. Cancer Causes Control. 2016;27:271–80. doi: 10.1007/s10552-015-0704-3. [DOI] [PubMed] [Google Scholar]
- 30.Bergmann MM, Calle EE, Mervis CA, Miracle-McMahill HL, Thun MJ, Heath CW. Validity of self-reported cancers in a prospective cohort study in comparison with data from state cancer registries. Am J Epidemiol. 1998;147:556–62. doi: 10.1093/oxfordjournals.aje.a009487. [DOI] [PubMed] [Google Scholar]
- 31.Navarro C, Chirlaque MD, Tormo MJ, Perez-Flores D, Rodriguez-Barranco M, Sanchez-Villegas A, Agudo A, Pera G, Amiano P, Dorronsoro M, Larranaga N, Quiros JR, et al. Validity of self reported diagnoses of cancer in a major Spanish prospective cohort study. J Epidemiol Community Health. 2006;60:593–9. doi: 10.1136/jech.2005.039131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Stavrou E, Vajdic CM, Loxton D, Pearson SA. The validity of self-reported cancer diagnoses and factors associated with accurate reporting in a cohort of older Australian women. Cancer Epidemiol. 2011;35:e75–80. doi: 10.1016/j.canep.2011.02.005. [DOI] [PubMed] [Google Scholar]
- 33.Greenland S. Response and follow-up bias in cohort studies. Am J Epidemiol. 1977;106:184–7. doi: 10.1093/oxfordjournals.aje.a112451. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.