Short paper 209
Alcohol consumption and breast cancer risk subtypes in the E3N-EPIC cohort Guy Fagherazzia,b,c, Alice Viliera,b,c, Marie-Christine Boutron-Ruaulta,b,c, Sylvie Mesrinea,b,c and Franc¸oise Clavel-Chapelona,b,c The aim of this study was to obtain an overview of the associations between alcohol consumption and breast cancer risk at adulthood, by type of alcohol and subtype of breast cancer. Between 1993 and 2008, 66 481 women from the French E3N-EPIC cohort were followed up and asked to report their alcohol consumption, by type of alcohol, through a 208-item diet-history questionnaire. A total of 2812 breast cancer cases were validated during the follow-up session. No association was found between high alcohol consumption, whatever its type, and increase in breast cancer risk in the premenopausal period. During the postmenopausal period, a linear association between total alcohol consumption and breast cancer risk was found (P < 0.0001), mainly driven by the associations with wine and beer [hazard ratio = 1.33 (1.11–1.58) and 1.85 (1.19–2.89)] for more than two glasses per day of wine and beer, respectively, compared with nondrinkers] and with ER + /PR + breast cancer subtypes. In the postmenopausal period, we observed interactions between total alcohol and folate intake levels (P = 0.1192) and BMI (P = 0.0367), with higher increased risks observed for high alcohol intake
among women with low folate intake or who were overweight or obese. Our results make precise the current body of knowledge on the relationship between alcohol and breast cancer subtypes. Interactions between alcohol and other factors should further be taken into account in public health nutrition programs. European Journal of c 2015 Wolters Cancer Prevention 24:209–214 Copyright Kluwer Health, Inc. All rights reserved.
Introduction
Materials and methods
To date, alcohol consumption is the only dietary risk factor for breast cancer clearly established in the literature. The World Cancer Research Fund has classified alcohol consumption as a convincing risk factor in the premenopausal and postmenopausal period (Hamajima et al., 2002; Wiseman, 2008), and it has been estimated that between 5 and 7% of breast cancer cases are attributable to alcohol consumption in the USA and Europe (Schutze et al., 2011). However, some questions have still been understudied. In only a few studies were the data stratified according to the menopausal status of the women, or to the hormone receptor status of the tumor (Nasca et al., 1994; Enger et al., 1999; Lew et al., 2009; Li et al., 2010). The influence of the type of alcohol is not well known, although some studies have already reported no difference in risks (Tjonneland et al., 2004). In addition, some interactions, such as use of menopausal hormone therapy (MHT), BMI, or folate intake, need further investigation to better understand the underlying biological mechanisms. Therefore, we proposed to obtain an overview of the associations between alcohol consumption and breast cancer risk at adulthood, by type of alcohol and subtype of breast cancer, in the large French E3N-EPIC cohort.
Study population
c 2015 Wolters Kluwer Health, Inc. All rights reserved. 0959-8278 Copyright
European Journal of Cancer Prevention 2015, 24:209–214 Keywords: alcohol, beer, breast cancer, cohort, estrogen receptor, risk, wine a Center for Research in Epidemiology and Population Health (CESP), INSERM (French National Institute of Health and Medical Research), bParis-South Univ, Villejuif Cedex and cIGR Institut Gustave Roussy, Villejuif, France
Correspondence to Franc¸oise Clavel-Chapelon, PhD, Center for Research in Epidemiology and Population Health (CESP), INSERM (French National Institute of Health and Medical Research) U1018 Team 9: Nutrition, Hormones and Women’s Health, Institut Gustave Roussy, 39 Rue Camille Desmoulins, 94805 Villejuif Cedex, France Tel: + 33 1 42 11 41 48; fax: + 33 1 42 11 40 00; e-mail: [email protected] Received 4 March 2014 Accepted 12 March 2014
The E3N cohort includes 98 995 women living in France and covered by a national health insurance scheme primarily involving school teachers. Participants were between 40 and 65 years old when they first entered the cohort between June 1990 and November 1991. At inclusion, they filled in a baseline questionnaire, which contained questions on established risk factors for breast cancer, including aspects of reproductive life, menopausal status, history of benign breast disease, breast cancer in first-degree relatives, and a variety of lifestyle characteristics. Follow-up questionnaires were sent out B2–3 years thereafter. Follow-up started in 1993, the date of return of the questionnaire through which information on diet was requested. Responders (n = 74 531) contributed personyears of follow-up until the date of cancer diagnosis (other than a basal cell carcinoma), the date of the last completed questionnaire, or July 2008 (date of mailing of the last available questionnaire), whichever occurred first. We excluded women with any prevalent cancer (other than a basal cell carcinoma; n = 4686), women with unknown statuses for estrogen and progesterone receptors (n = 800), women with no follow-up or who did not DOI: 10.1097/CEJ.0000000000000031
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
210 European Journal of Cancer Prevention 2015, Vol 24 No 3
answer the diet-history questionnaire (n = 1228), women with unknown menopausal status (n = 5), and those with extreme values (i.e. the lowest and highest percentiles) for the ratio between energy intake and required energy (n = 1331). Among the remaining 66 481 women, 2812 breast cancer cases were diagnosed during follow-up. Identification of breast cancer cases
All questionnaires inquired about cancer occurrence and type, requesting addresses of physicians and permission to contact them. A small number of breast cancer cases were further identified from insurance files and death certificates. Pathology reports were obtained for 99.4% of incident cases considered in the present analysis. Information on estrogen receptor (ER) and progesterone receptor (PR) statuses was extracted from those reports, and invasive breast cancer cases were classified into four categories on the basis of receptor positivity or negativity: ER + /PR + , ER + /PR – , ER – /PR + , and ER – /PR – . Assessment of alcohol consumption
Usual diet over the previous year was assessed using a validated 208-item diet-history questionnaire in 1993, structured according to the French meal pattern (Van Liere et al., 1997). Briefly, questions were asked about the time of consumption, from breakfast to after-dinner snacks. The frequency and portion sizes of 66 food groups or items were recorded by meal. Participants were asked to report the frequency and usual serving size of their alcoholic drinks by type of alcohol: wine, beer/cider, fortified wine, and spirits. Consumption values were then converted into standard glasses of 150 ml for wine, 250 ml for beer, 70 ml for fortified wine, and 40 ml for spirits, with one standard glass corresponding to B10 g of ethanol. Statistical analyses
Cox proportional hazards regression models with age as the time-scale were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Models were adjusted for the following variables: education level (undergraduate/graduate or postgraduate), total physical activity (MET-h/week, continuous, recorded at baseline), age at menarche (year, continuous), nulliparity (yes/no) and age at first full-term pregnancy (before/after 30 years), breastfeeding (yes/no), use of oral contraceptives (ever/never), family history of breast cancer (yes/no), mammography during the previous follow-up period (yes/no/unknown, time-dependent), history of benign breast disease (never/ever, time-dependent), use of oral progestagens alone before menopause (current/nonuser as a time-dependent variable), age at menopause (continuous for postmenopausal women only), menopausal status and use of MHT (premenopausal/postmenopausal with current use of MHT/postmenopausal and nonuser of MHT/postmenopausal and unknown use of MHT, time-dependent), BMI (< 20, [20–25], [25–30], Z 30 kg/m2, time-dependent), and age at menarche. Models were further stratified according to year of birth ([1925–1930], [1930–1935], [1935–1940],
[1940–1945], [1945–1950]). Interactions between main exposures and BMI, and folate intake and current use of MHT were tested. All P-values were two-tailed. All analyses were carried out using SAS software, version 9.3 (SAS Institute Inc., Cary, North Carolina, USA).
Results Characteristics of the study population
Baseline characteristics of the women are shown in Table 1. Women of the study population had an average age of 52.7 years at baseline (SD = 6.6) and were mostly highly educated (86.1% were graduates or postgraduates). At inclusion, 80.7% of them had a BMI lower than 25 kg/m2 and 55.3% were in their postmenopausal period. The mean daily total alcohol consumption was 11.6 g/day (1.2 standard glasses/day), and 13.2 g/day (1.3 standard glasses/day) among consumers only (12.6% of the population were abstainers). The mean consumption was as follows for alcohol subtypes (among alcohol consumers): 106.3 ml/day for wine (0.8 standard glasses/ day), 29.2 ml/day for beer (0.1 standard glasses/day), 7.9 ml/day (0.1 standard glasses/day) for fortified wine, and 9.5 ml/day for spirits (0.2 standard glasses/day). Alcohol consumption and breast cancer risk In the overall, premenopausal, and postmenopausal populations
We observed an interaction between alcohol intake and menopausal status (P = 0.0266; Table 2). We thus presented associations in both the overall population and in premenopausal and postmenopausal women. Results did not differ significantly between age-adjusted and multivariate models. We observed a positive trend in breast cancer risk with increasing alcohol intake in the overall population (Ptrend = 0.0003). Women who consumed more than two standard drinks per day had a significant 19% increase in risk [HR = 1.19 (1.04–1.36)]. This association was limited to postmenopausal women [Ptrend < 0.0001 and HR = 1.24 (1.07–1.44) for women who drank more than two standard alcohol glasses per day]. In terms of types of alcohol, high consumption of wine or beer was found to be associated with an increased breast cancer risk in the overall population [HR = 1.25 (1.06–1.47) and HR = 1.90 (1.29–2.79), respectively, for a consumption of more than two standard glasses per day]. Positive trends were observed for both types of alcohol (Ptrend = 0.001 for wine and Ptrend < 0.0001 for beer). These associations were limited to postmenopausal women. No association or trend in risks was seen for consumption of spirits or fortified wine in both premenopausal and postmenopausal women. Risks for ER + /PR + and ER – /PR – breast cancer
Stratified analyses by hormone receptor status in the premenopausal period did not reveal any significant
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Alcohol consumption and breast cancer risk subtypes Fagherazzi et al. 211
Table 1
Baseline characteristics of the study population
Characteristics Age (years) Age at menarche (years) Age at menopause (postmenopausal women only) Menopausal status and use of MHT Premenopause Menopause, nonuse of MHT Menopause, current use of MHT Menopause, unknown use of MHT Daily energy intake (except alcohol; kcal/day) Physical activity (MET-h/week) Schooling Undergraduate Graduate/postgraduate Use of oral contraceptives Never Ever Family history of breast cancer No Yes Personal history of benign breast disease No Yes BMI (kg/m2) < 20 [20–25] [25–30] Z 30 Alcohol intake (among alcohol drinkers; g/day) Wine (among alcohol drinkers; ml/day) Beer (among alcohol drinkers; ml/day) Fortified wine (among alcohol drinkers; ml/day) Spirits (among alcohol drinkers; ml/day)
Mean (SD) or N (%) 52.7 (6.6) 12.8 (1.4) 50.5 (3.7) 29 740 22 502 14 058 181 2130.8 54.9
(44.7) (33.8) (21.2) (0.3) (543.7) (30.2)
9264 (13.9) 57 217 (86.1) 25 935 (39.0) 40 546 (61.0) 58 785 (88.4) 7696 (11.6) 47 086 (70.8) 19 395 (29.2) 9861 43 827 10 489 2304 13.2 106.3 29.2 7.9 9.5
(14.8) (65.9) (15.8) (3.5) (14.1) (132.1) (83.6) (15.2) (21.6)
E3N cohort data (N = 66 481).
association or trend in risk. Therefore, only the stratifications among postmenopausal women were tabulated (Table 3). We observed a significant relationship between high alcohol intake and ER + /PR + breast cancer risk [HR = 1.32 (1.08–1.60)]. Similar results were observed for ER – /PR – , although the associations did not reach significance, possibly because of a lack of power. Results by type of alcoholic beverage (not tabulated) revealed that the increase in ER + /PR + breast cancer risk was driven by wine [HR = 1.54 (1.23–1.93) for two drinks or more per day, compared with nondrinkers]. Positive significant trends in ER + /PR + breast cancer risk were observed for total alcohol intake (Ptrend = 0.0006) and wine (Ptrend < 0.0001). For ER – /PR – cases, we also observed positive trends in risk for total alcohol intake (Ptrend = 0.0057), wine (Ptrend = 0.0397), beer (Ptrend = 0.0003), and fortified wine (Ptrend = 0.0236). High beer consumption was significantly associated with ER – /PR – breast cancer risk [HR = 2.52 (1.00–6.36), but with only five cases in the higher category]. Interactions with folate intake, BMI, and use of MHT
We observed interactions, in the postmenopausal period only, between total alcohol and folate intake (P = 0.1192), alcohol and BMI (P = 0.0367), and alcohol and current MHT use (P = 0.1123). Stratifications
according to the median intake of folate (411 g/day), BMI (cut-off point at 25 kg/m2), and current use of MHT were performed, and results are tabulated in Table 3. The associations previously observed between total alcohol intake and breast cancer risk were restricted to the low folate intake category in which consumption of more than two drinks per day was associated with a significantly increased breast cancer risk [HR = 1.35 (1.10–1.67)], whereas there was no association of total alcohol intake with breast cancer risk in the high folate intake category [HR = 1.16 (0.94–1.44)]. However, the trends in risk tended to persist in both low and high folate intake categories (Ptrend = 0.0049 and 0.0077, respectively). Stratification on the basis of BMI revealed that overweight/obese women who consumed two glasses or more per day of total alcohol had a slightly greater risk of postmenopausal breast cancer than women with a BMI under 25 kg/m2. Women who were current users of MHT and were alcohol consumers were not associated with an increased risk for postmenopausal breast cancer [HR = 1.07 (0.88–1.30)], among drinkers of two or more glasses daily compared with nondrinkers. In contrast, non-MHT-users who consumed more than two glasses per day were at a highly significantly increased risk compared with nonconsumers [HR = 1.52 (1.21–1.92)].
Discussion We found a strong direct association between baseline high consumption of alcohol and overall and postmenopausal breast cancer risk. This relation was mainly driven by wine and beer consumption and was restricted to ER + /PR + breast cancers, women with low folate intake, or non-MHT-users. The strength of the association between alcohol intake and breast cancer risk was greater among overweight and obese women. Our results are consistent with the literature, according to which most of the studies concluded that there was an increased risk for breast cancer (Wiseman, 2008; Bagnardi et al., 2012) – mostly cancers occurring in the postmenopausal period and with an ER + status (Lew et al., 2009; Li et al., 2010). Previous meta-analyses and the World Cancer Research Fund review (Hamajima et al., 2002; Wiseman, 2008) have found an increased risk of premenopausal breast cancer associated with alcohol consumption that we did not confirm in our study, probably because of a relatively low number of premenopausal women included. In addition, we found that the association between total alcohol intake and breast cancer was driven by the consumption of wine and beer; previous studies (Hamajima et al., 2002; Key et al., 2006; Wiseman, 2008) did not conclude a differential effect of alcohol depending on the type of beverage.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
212 European Journal of Cancer Prevention 2015, Vol 24 No 3
Table 2 Risk of breast cancer according to baseline total alcohol consumption and types of alcohol in the overall, premenopausal, and postmenopausal populations Multivariateb
Overall population Age-adjusteda N Alcohol (1 drink = 10 g of ethanol) Non-alcohol-consumer 318 < 0.5 drinks/day 806 0.5–1 drinks/day 469 1–2 drinks/day 587 Z 2 drinks/day 632 Ptrend Wine (1 drink = 150 ml) Non-alcohol-consumer 318 < 0.5 drinks/day 1307 0.5–1 drinks/day 495 1–2 drinks/day 422 Z 2 drinks/day 270 Ptrend Beer (1 drink = 250 ml) Non-alcohol-consumer 318 < 0.5 drinks/day 2297 0.5–1 drinks/day 111 1–2 drinks/day 58 Z 2 drinks/day 28 Ptrend Fortified wine (1 drink = 70 ml) Non-alcohol-consumer 318 < 0.5 drinks/day 2382 0.5–1 drinks/day 87 1–2 drinks/day 20 Z 2 drinks/day 5 Ptrend Spirits (1 drink = 40 ml) Non-alcohol-consumer 318 < 0.5 drinks/day 2133 0.5–1 drinks/day 216 1–2 drinks/day 103 Z 2 drinks/day 42 Ptrend
HR (95%CI)
Multivariateb HR (95% CI)
Premenopause N
HR (95% CI)
Postmenopause b
N
HR (95% CI)b
1 (reference) 1.01 (0.88–1.14) 1.04 (0.90–1.20) 1.03 (0.99–1.08) 1.24 (1.08–1.41) < 0.0001
1 1.00 1.01 1.02 1.19
(reference) (0.88–1.14) (0.88–1.17) (0.97–1.07) (1.04–1.36) 0.0003
59 157 89 108 94
1 0.98 0.94 1.00 0.95
(reference) (0.72–1.32) (0.68–1.31) (0.90–1.12) (0.68–1.32) 0.8531
259 649 380 479 538
1 (reference) 1.00 (0.86–1.15) 1.02 (0.87–1.20) 1.02 (0.97–1.07) 1.24 (1.07–1.44) < 0.0001
1 1.02 1.14 1.11 1.30
1 1.01 1.10 1.08 1.25
(reference) (0.89–1.14) (0.96–1.27) (0.93–1.24) (1.06–1.47) 0.001
59 266 90 59 33
1 0.99 1.06 0.84 0.84
(reference) (0.75–1.32) (0.76–1.48) (0.58–1.20) (0.55–1.29) 0.2225
259 1041 405 363 237
1 (reference) 1.00 (0.87–1.15) 1.10 (0.94–1.29) 1.12 (0.95–1.31) 1.33 (1.11–1.58) < 0.0001
1 (reference) 1.06 (0.95–1.20) 1.37 (1.11–1.71) 1.33 (1.00–1.76) 1.89 (1.29–2.79) < 0.0001
1 (reference) 1.04 (0.92–1.17) 1.32 (1.06–1.64) 1.29 (0.98–1.71) 1.90 (1.29–2.79) < 0.0001
59 404 27 10 7
1 0.94 1.49 0.99 2.04
(reference) (0.72–1.24) (0.94–2.36) (0.51–1.94) (0.93–4.48) 0.0174
259 1893 84 48 21
1 1.05 1.26 1.36 1.85
(reference) (0.92–1.20) (0.98–1.61) (1.00–1.86) (1.19–2.89) 0.0005
1 1.08 1.39 1.01 1.11
(reference) (0.96–1.21) (1.09–1.76) (0.64–1.59) (0.46–2.68) 0.1972
1 1.05 1.33 0.96 1.06
(reference) (0.93–1.18) (1.04–1.68) (0.61–1.51) (0.44–2.57) 0.3029
59 431 14 2 1
1 0.97 1.08 0.48 1.26
(reference) (0.74–1.28) (0.60–1.94) (0.12–1.96) (0.17–9.12) 0.7548
259 1951 73 18 4
1 1.06 1.37 1.07 1.02
(reference) (0.93–1.21) (1.06–1.78) (0.67–1.73) (0.38–2.73) 0.2116
1 1.07 1.17 1.17 1.15
(reference) (0.95–1.21) (0.98–1.39) (0.94–1.46) (0.83–1.58) 0.1919
1 1.05 1.12 1.12 1.11
(reference) (0.93–1.18) (0.94–1.33) (0.90–1.40) (0.80–1.53) 0.3237
59 378 39 25 6
1 0.96 0.93 1.33 0.76
(reference) (0.73–1.27) (0.62–1.40) (0.83–2.12) (0.33–1.75) 0.7944
259 1755 177 78 36
1 1.06 1.16 1.06 1.19
(reference) (0.93–1.21) (0.95–1.40) (0.82–1.37) (0.84–1.69) 0.3215
(reference) (0.91–1.16) (0.99–1.32) (0.96–1.29) (1.10–1.53) 0.0001
E3N cohort data (N = 66 481). Bold values indicate P < 0.05. CI, confidence interval; HR, hazard ratio; MHT, menopausal hormone therapy. a Models were age-adjusted. b Models were adjusted for schooling (undergraduate/graduate or postgraduate), physical activity (MET-h/week), age at puberty, nulliparous (yes/no/missing), and age at first pregnancy (first pregnancy before 30/after 30/missing), breastfeeding (yes/no), use of oral contraceptives (ever/never), family history of breast cancer (yes/no), mammograms during the last follow-up period (yes/no), personal history of benign breast disease (yes/no), BMI (< 20/20–25/25–30/ Z 30 kg/m2), use of progestagens in premenopause (current/nonuser/unknown), age at menopause (menopausal women only), menopausal women and use of MHT (premenopause/postmenopause and current use of MHT/postmenopause and non-MHT-user/postmenopause and unknown use of MHT).
The existing literature failed to answer some questions on the interactions of alcohol with other factors. We confirmed an interaction with folate intake, which was suggested by previous studies (Lew et al., 2009). Existing papers did not support interactions between alcohol and BMI or MHT use (Hamajima et al., 2002). In our study, we observed an association only for non-MHT-users and, although the associations were of a higher magnitude among women with BMI greater than 25 kg/m2, trends in risk were present in the both strata of BMI. Biological mechanisms The estrogen pathway
The increase in postmenopausal and ER + /PR + breast cancer risks with alcohol intake can be explained by the associated increase in estrogen production, as high alcohol intake has been previously shown to be an enhancer of
estrogen synthesis (Ginsburg et al., 1996; Purohit, 1998; Nielsen and Gronbaek, 2008). The aromatization increase may also explain the higher increase in risk observed among overweight and obese women in our study. However, our study failed to confirm the hypothesis of a synergistic effect of alcohol and MHT on breast cancer risk in the postmenopausal period, as our relative risks were not higher among MHT users. Other mechanisms
Alcohol was found to be associated with an increase in mammographic density (Conroy et al., 2012; Rauh et al., 2012). Alcohol consumption has also been shown to promote mammary tumor growth and insulin sensitivity (Hong et al., 2010), breast cancer cell invasion by regulating the Nm23-ITGA5 pathway (Wong et al., 2011), and, more recently, HER2 breast cancer development through
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Alcohol consumption and breast cancer risk subtypes Fagherazzi et al. 213
Table 3 Risk of postmenopausal breast cancer according to baseline alcohol consumption (overall population and stratified by hormone receptor status, folate intake, BMI, and use of hormone replacement therapy)
ER + /PR + Alcohol consumption (drinks/day)a
N (cases)
HR (95% CI)
Low folate intake (Vit B9 < 411 g/day)
ER – /PR – N (cases)
Nonconsumers < 0.5 0.5–1 1–2 Z2 Ptrend
147 387 226 278 328
1 (reference) 1.05 (0.87–1.27) 1.07 (0.86–1.31) 1.03 (0.96–1.10) 1.32 (1.08–1.60) 0.0006 BMI < 25 kg/m2
45 92 58 78 92
Nonconsumers < 0.5 0.5–1 1–2 Z2 Ptrend
190 438 269 336 373
1 (reference) 0.92 (0.77–1.09) 0.99 (0.82–1.20) 1.00 (0.94–1.06) 1.20 (1.01–1.43) 0.0003
69 211 111 143 165
HR (95% CI)
N (cases)
1 (reference) 0.80 (0.56–1.15) 0.90 (0.61–1.33) 1.00 (0.89–1.13) 1.24 (0.87–1.78) 0.0057 BMI Z 25 kg/m2 1 (reference) 1.23 (0.94–1.61) 1.12 (0.82–1.51) 1.08 (0.98–1.19) 1.37 (1.03–1.82) 0.0569
HR (95% CI)
132 341 183 211 265
1 (reference) 1.06 (0.87–1.30) 1.11 (0.88–1.38) 1.01 (0.94–1.09) 1.35 (1.10–1.67) 0.0049 Non-current-MHT user
108 292 157 190 231
1 (reference) 1.17 (0.94–1.46) 1.16 (0.91–1.49) 1.06 (0.98–1.15) 1.52 (1.21–1.92) 0.0002
High folate intake (Vit B9 Z 411 g/day) N (cases)
HR (95% CI)
127 308 197 268 273
1 (reference) 0.94 (0.7–1.16) 0.96 (0.77–1.21) 1.03 (0.96–1.11) 1.16 (0.94–1.44) 0.0077 Current MHT user
151 357 223 289 307
1 (reference) 0.88 (0.73–1.07) 0.93 (0.75–1.14) 0.99 (0.93–1.06) 1.07 (0.88–1.30) 0.0272
E3N cohort data (N = 66 481). Bold values indicate P < 0.05. CI, confidence interval; HR, hazard ratio; MHT, menopausal hormone therapy. a Models were adjusted for schooling (undergraduate/graduate or postgraduate), physical activity (MET-h/week), age at puberty, nulliparous (yes/no/missing) and age at first pregnancy (first pregnancy before 30/after 30/missing), breastfeeding (yes/no), use of oral contraceptives (ever/never), family history of breast cancer (yes/no), mammograms during the last follow-up period (yes/no), personal history of benign breast disease (yes/no), BMI (< 20/20–25/25–30/ Z 30 kg/m2, except for models stratified on BMI), use of progestagens in premenopause (current/nonuser/unknown), age at menopause (menopausal women only), menopausal women and use of MHT (premenopause/postmenopause and current use of MHT/postmenopause and non-MHT-user/postmenopause and unknown use of MHT, except for models stratified on MHT use).
the estrogen signaling pathway in animal models (Wong et al., 2012). In our study, we found that both lifetime and baseline alcohol exposures were independently associated with an increased risk of postmenopausal breast cancer. These findings are in line with the hypothesis of Brooks and Zakhari (2013), who suggested that alcohol could act on breast carcinogenesis not only through a pure tumorpromoter-type mechanism, but also through a weak cumulative carcinogen mechanism throughout life. Strengths and limitations
This study has some limitations. Participants in the E3N cohort belong to a homogeneous health-conscious population, thus reducing the variability of certain characteristics such as alcohol consumption. Moreover, information on diet was not updated during follow-up, although dietary habits may have changed over time. Although this might reduce our power to show differences between groups, it would not bias our estimates. This study also has several strengths. To our knowledge, we are the first prospective cohort study to obtain such an overview of the associations between alcohol consumption – with details on types of alcohol – and breast cancer risk subtypes. The long followup period yielded sufficient statistical power for the analysis of stratified populations and subgroups of breast cancers defined by menopausal status and tumor hormone receptor status. Moreover, we were able to adjust for most of the risk factors for breast cancer. Conclusion
We showed that alcohol consumption was associated with an increased risk of postmenopausal and ER + /PR + breast cancer and that these associations could be modulated by
folate intake and, to a lesser extent, by BMI. Our study adds significant information to the body of knowledge on the complex benefit–risk association of alcohol consumption. We encourage public health authorities to further take into account interactions between alcohol and other factors in public health nutrition programs.
Acknowledgements The authors thank all participants for providing data and practitioners for providing pathology reports. This work was supported by the Institut National du Cancer, the Mutuelle Ge´ne´rale de l’Education Nationale, the Institut de Cance´rologie Gustave Roussy, and the Institut National de la Sante´ et de la Recherche Me´dicale. The study sponsors played no role in the design of the study, analysis or interpretation of data, writing of the manuscript, or the decision to submit the manuscript for publication. Conflicts of interest
There are no conflicts of interest.
References Bagnardi V, Rota M, Botteri E, Tramacere I, Islami F, Fedirko V, et al. (2013). Light alcohol drinking and cancer: a meta-analysis. Ann Oncol 24:301–308. Brooks PJ, Zakhari S (2013). Moderate alcohol consumption and breast cancer in women: from epidemiology to mechanisms and interventions. Alcohol Clin Exp Res 37:23–30. Conroy SM, Koga K, Woolcott CG, Dahl T, Byrne C, Nagata C, et al. (2012). Higher alcohol intake may modify the association between mammographic density and breast cancer: an analysis of three case-control studies. Cancer Epidemiol 36:458–460. Enger SM, Ross RK, Paganini-Hill A, Longnecker MP, Bernstein L (1999). Alcohol consumption and breast cancer oestrogen and progesterone receptor status. Br J Cancer 79:1308–1314.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
214 European Journal of Cancer Prevention 2015, Vol 24 No 3
Ginsburg ES, Mello NK, Mendelson JH, Barbieri RL, Teoh SK, Rothman M, et al. (1996). Effects of alcohol ingestion on estrogens in postmenopausal women. JAMA 276:1747–1751. Hamajima N, Hirose K, Tajima K, Rohan T, Calle EE, Heath CW Jr, et al. (2002). 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 87:1234–1245. Hong J, Holcomb VB, Tekle SA, Fan B, Nunez NP (2010). Alcohol consumption promotes mammary tumor growth and insulin sensitivity. Cancer Lett 294:229–235. Key J, Hodgson S, Omar RZ, Jensen TK, Thompson SG, Boobis AR, et al. (2006). Meta-analysis of studies of alcohol and breast cancer with consideration of the methodological issues. Cancer Causes Control 17: 759–770. Lew JQ, Freedman ND, Leitzmann MF, Brinton LA, Hoover RN, Hollenbeck AR, et al. (2009). Alcohol and risk of breast cancer by histologic type and hormone receptor status in postmenopausal women: the NIH-AARP Diet and Health Study. Am J Epidemiol 170:308–317. Li CI, Chlebowski RT, Freiberg M, Johnson KC, Kuller L, Lane D, et al. (2010). Alcohol consumption and risk of postmenopausal breast cancer by subtype: the Women’s Health Initiative observational study. J Natl Cancer Inst 102:1422–1431. Nasca PC, Liu S, Baptiste MS, Kwon CS, Jacobson H, Metzger BB (1994). Alcohol consumption and breast cancer: estrogen receptor status and histology. Am J Epidemiol 140:980–988.
Nielsen NR, Gronbaek M (2008). Interactions between intakes of alcohol and postmenopausal hormones on risk of breast cancer. Int J Cancer 122: 1109–1113. Purohit V (1998). Moderate alcohol consumption and estrogen levels in postmenopausal women: a review. Alcohol Clin Exp Res 22:994–997. Rauh C, Hack C, Ha¨berle L, Hein A, Engel A, Schrauder MG, et al. (2012). Percent mammographic density and dense area as risk factors for breast cancer. Geburtsh Frauenheilk 72:727–733. Schutze M, Boeing H, Pischon T, Rehm J, Kehoe T, Gmel G, et al. (2011). Alcohol attributable burden of incidence of cancer in eight European countries based on results from prospective cohort study. BMJ 342:d1584. Tjonneland A, Christensen J, Thomsen BL, Olsen A, Stripp C, Overvad K, et al. (2004). Lifetime alcohol consumption and postmenopausal breast cancer rate in Denmark: a prospective cohort study. J Nutr 134:173–178. Van Liere MJ, Lucas F, Clavel F, Slimani N, Villeminot S (1997). Relative validity and reproducibility of a French dietary history questionnaire. Int J Epidemiol 26 (Suppl 1):S128–S136. Wiseman M (2008). The second World Cancer Research Fund/American Institute for Cancer Research expert report. Food, nutrition, physical activity, and the prevention of cancer: a global perspective. Proc Nutr Soc 67:253–256. Wong AW, Paulson QX, Hong J, Stubbins RE, Poh K, Schrader E, et al. (2011). Alcohol promotes breast cancer cell invasion by regulating the Nm23-ITGA5 pathway. J Exp Clin Cancer Res 30:75. Wong AW, Dunlap SM, Holcomb VB, Nunez NP (2012). Alcohol promotes mammary tumor development via the estrogen pathway in estrogen receptor alpha-negative HER2/neu mice. Alcohol Clin Exp Res 36:577–587.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Alcohol consumption and breast cancer risk subtypes in the E3N-EPIC cohort Guy Fagherazzia,b,c, Alice Viliera,b,c, Marie-Christine Boutron-Ruaulta,b,c, Sylvie Mesrinea,b,c and Franc¸oise Clavel-Chapelona,b,c The aim of this study was to obtain an overview of the associations between alcohol consumption and breast cancer risk at adulthood, by type of alcohol and subtype of breast cancer. Between 1993 and 2008, 66 481 women from the French E3N-EPIC cohort were followed up and asked to report their alcohol consumption, by type of alcohol, through a 208-item diet-history questionnaire. A total of 2812 breast cancer cases were validated during the follow-up session. No association was found between high alcohol consumption, whatever its type, and increase in breast cancer risk in the premenopausal period. During the postmenopausal period, a linear association between total alcohol consumption and breast cancer risk was found (P < 0.0001), mainly driven by the associations with wine and beer [hazard ratio = 1.33 (1.11–1.58) and 1.85 (1.19–2.89)] for more than two glasses per day of wine and beer, respectively, compared with nondrinkers] and with ER + /PR + breast cancer subtypes. In the postmenopausal period, we observed interactions between total alcohol and folate intake levels (P = 0.1192) and BMI (P = 0.0367), with higher increased risks observed for high alcohol intake
among women with low folate intake or who were overweight or obese. Our results make precise the current body of knowledge on the relationship between alcohol and breast cancer subtypes. Interactions between alcohol and other factors should further be taken into account in public health nutrition programs. European Journal of c 2015 Wolters Cancer Prevention 24:209–214 Copyright Kluwer Health, Inc. All rights reserved.
Introduction
Materials and methods
To date, alcohol consumption is the only dietary risk factor for breast cancer clearly established in the literature. The World Cancer Research Fund has classified alcohol consumption as a convincing risk factor in the premenopausal and postmenopausal period (Hamajima et al., 2002; Wiseman, 2008), and it has been estimated that between 5 and 7% of breast cancer cases are attributable to alcohol consumption in the USA and Europe (Schutze et al., 2011). However, some questions have still been understudied. In only a few studies were the data stratified according to the menopausal status of the women, or to the hormone receptor status of the tumor (Nasca et al., 1994; Enger et al., 1999; Lew et al., 2009; Li et al., 2010). The influence of the type of alcohol is not well known, although some studies have already reported no difference in risks (Tjonneland et al., 2004). In addition, some interactions, such as use of menopausal hormone therapy (MHT), BMI, or folate intake, need further investigation to better understand the underlying biological mechanisms. Therefore, we proposed to obtain an overview of the associations between alcohol consumption and breast cancer risk at adulthood, by type of alcohol and subtype of breast cancer, in the large French E3N-EPIC cohort.
Study population
c 2015 Wolters Kluwer Health, Inc. All rights reserved. 0959-8278 Copyright
European Journal of Cancer Prevention 2015, 24:209–214 Keywords: alcohol, beer, breast cancer, cohort, estrogen receptor, risk, wine a Center for Research in Epidemiology and Population Health (CESP), INSERM (French National Institute of Health and Medical Research), bParis-South Univ, Villejuif Cedex and cIGR Institut Gustave Roussy, Villejuif, France
Correspondence to Franc¸oise Clavel-Chapelon, PhD, Center for Research in Epidemiology and Population Health (CESP), INSERM (French National Institute of Health and Medical Research) U1018 Team 9: Nutrition, Hormones and Women’s Health, Institut Gustave Roussy, 39 Rue Camille Desmoulins, 94805 Villejuif Cedex, France Tel: + 33 1 42 11 41 48; fax: + 33 1 42 11 40 00; e-mail: [email protected] Received 4 March 2014 Accepted 12 March 2014
The E3N cohort includes 98 995 women living in France and covered by a national health insurance scheme primarily involving school teachers. Participants were between 40 and 65 years old when they first entered the cohort between June 1990 and November 1991. At inclusion, they filled in a baseline questionnaire, which contained questions on established risk factors for breast cancer, including aspects of reproductive life, menopausal status, history of benign breast disease, breast cancer in first-degree relatives, and a variety of lifestyle characteristics. Follow-up questionnaires were sent out B2–3 years thereafter. Follow-up started in 1993, the date of return of the questionnaire through which information on diet was requested. Responders (n = 74 531) contributed personyears of follow-up until the date of cancer diagnosis (other than a basal cell carcinoma), the date of the last completed questionnaire, or July 2008 (date of mailing of the last available questionnaire), whichever occurred first. We excluded women with any prevalent cancer (other than a basal cell carcinoma; n = 4686), women with unknown statuses for estrogen and progesterone receptors (n = 800), women with no follow-up or who did not DOI: 10.1097/CEJ.0000000000000031
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
210 European Journal of Cancer Prevention 2015, Vol 24 No 3
answer the diet-history questionnaire (n = 1228), women with unknown menopausal status (n = 5), and those with extreme values (i.e. the lowest and highest percentiles) for the ratio between energy intake and required energy (n = 1331). Among the remaining 66 481 women, 2812 breast cancer cases were diagnosed during follow-up. Identification of breast cancer cases
All questionnaires inquired about cancer occurrence and type, requesting addresses of physicians and permission to contact them. A small number of breast cancer cases were further identified from insurance files and death certificates. Pathology reports were obtained for 99.4% of incident cases considered in the present analysis. Information on estrogen receptor (ER) and progesterone receptor (PR) statuses was extracted from those reports, and invasive breast cancer cases were classified into four categories on the basis of receptor positivity or negativity: ER + /PR + , ER + /PR – , ER – /PR + , and ER – /PR – . Assessment of alcohol consumption
Usual diet over the previous year was assessed using a validated 208-item diet-history questionnaire in 1993, structured according to the French meal pattern (Van Liere et al., 1997). Briefly, questions were asked about the time of consumption, from breakfast to after-dinner snacks. The frequency and portion sizes of 66 food groups or items were recorded by meal. Participants were asked to report the frequency and usual serving size of their alcoholic drinks by type of alcohol: wine, beer/cider, fortified wine, and spirits. Consumption values were then converted into standard glasses of 150 ml for wine, 250 ml for beer, 70 ml for fortified wine, and 40 ml for spirits, with one standard glass corresponding to B10 g of ethanol. Statistical analyses
Cox proportional hazards regression models with age as the time-scale were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Models were adjusted for the following variables: education level (undergraduate/graduate or postgraduate), total physical activity (MET-h/week, continuous, recorded at baseline), age at menarche (year, continuous), nulliparity (yes/no) and age at first full-term pregnancy (before/after 30 years), breastfeeding (yes/no), use of oral contraceptives (ever/never), family history of breast cancer (yes/no), mammography during the previous follow-up period (yes/no/unknown, time-dependent), history of benign breast disease (never/ever, time-dependent), use of oral progestagens alone before menopause (current/nonuser as a time-dependent variable), age at menopause (continuous for postmenopausal women only), menopausal status and use of MHT (premenopausal/postmenopausal with current use of MHT/postmenopausal and nonuser of MHT/postmenopausal and unknown use of MHT, time-dependent), BMI (< 20, [20–25], [25–30], Z 30 kg/m2, time-dependent), and age at menarche. Models were further stratified according to year of birth ([1925–1930], [1930–1935], [1935–1940],
[1940–1945], [1945–1950]). Interactions between main exposures and BMI, and folate intake and current use of MHT were tested. All P-values were two-tailed. All analyses were carried out using SAS software, version 9.3 (SAS Institute Inc., Cary, North Carolina, USA).
Results Characteristics of the study population
Baseline characteristics of the women are shown in Table 1. Women of the study population had an average age of 52.7 years at baseline (SD = 6.6) and were mostly highly educated (86.1% were graduates or postgraduates). At inclusion, 80.7% of them had a BMI lower than 25 kg/m2 and 55.3% were in their postmenopausal period. The mean daily total alcohol consumption was 11.6 g/day (1.2 standard glasses/day), and 13.2 g/day (1.3 standard glasses/day) among consumers only (12.6% of the population were abstainers). The mean consumption was as follows for alcohol subtypes (among alcohol consumers): 106.3 ml/day for wine (0.8 standard glasses/ day), 29.2 ml/day for beer (0.1 standard glasses/day), 7.9 ml/day (0.1 standard glasses/day) for fortified wine, and 9.5 ml/day for spirits (0.2 standard glasses/day). Alcohol consumption and breast cancer risk In the overall, premenopausal, and postmenopausal populations
We observed an interaction between alcohol intake and menopausal status (P = 0.0266; Table 2). We thus presented associations in both the overall population and in premenopausal and postmenopausal women. Results did not differ significantly between age-adjusted and multivariate models. We observed a positive trend in breast cancer risk with increasing alcohol intake in the overall population (Ptrend = 0.0003). Women who consumed more than two standard drinks per day had a significant 19% increase in risk [HR = 1.19 (1.04–1.36)]. This association was limited to postmenopausal women [Ptrend < 0.0001 and HR = 1.24 (1.07–1.44) for women who drank more than two standard alcohol glasses per day]. In terms of types of alcohol, high consumption of wine or beer was found to be associated with an increased breast cancer risk in the overall population [HR = 1.25 (1.06–1.47) and HR = 1.90 (1.29–2.79), respectively, for a consumption of more than two standard glasses per day]. Positive trends were observed for both types of alcohol (Ptrend = 0.001 for wine and Ptrend < 0.0001 for beer). These associations were limited to postmenopausal women. No association or trend in risks was seen for consumption of spirits or fortified wine in both premenopausal and postmenopausal women. Risks for ER + /PR + and ER – /PR – breast cancer
Stratified analyses by hormone receptor status in the premenopausal period did not reveal any significant
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Alcohol consumption and breast cancer risk subtypes Fagherazzi et al. 211
Table 1
Baseline characteristics of the study population
Characteristics Age (years) Age at menarche (years) Age at menopause (postmenopausal women only) Menopausal status and use of MHT Premenopause Menopause, nonuse of MHT Menopause, current use of MHT Menopause, unknown use of MHT Daily energy intake (except alcohol; kcal/day) Physical activity (MET-h/week) Schooling Undergraduate Graduate/postgraduate Use of oral contraceptives Never Ever Family history of breast cancer No Yes Personal history of benign breast disease No Yes BMI (kg/m2) < 20 [20–25] [25–30] Z 30 Alcohol intake (among alcohol drinkers; g/day) Wine (among alcohol drinkers; ml/day) Beer (among alcohol drinkers; ml/day) Fortified wine (among alcohol drinkers; ml/day) Spirits (among alcohol drinkers; ml/day)
Mean (SD) or N (%) 52.7 (6.6) 12.8 (1.4) 50.5 (3.7) 29 740 22 502 14 058 181 2130.8 54.9
(44.7) (33.8) (21.2) (0.3) (543.7) (30.2)
9264 (13.9) 57 217 (86.1) 25 935 (39.0) 40 546 (61.0) 58 785 (88.4) 7696 (11.6) 47 086 (70.8) 19 395 (29.2) 9861 43 827 10 489 2304 13.2 106.3 29.2 7.9 9.5
(14.8) (65.9) (15.8) (3.5) (14.1) (132.1) (83.6) (15.2) (21.6)
E3N cohort data (N = 66 481).
association or trend in risk. Therefore, only the stratifications among postmenopausal women were tabulated (Table 3). We observed a significant relationship between high alcohol intake and ER + /PR + breast cancer risk [HR = 1.32 (1.08–1.60)]. Similar results were observed for ER – /PR – , although the associations did not reach significance, possibly because of a lack of power. Results by type of alcoholic beverage (not tabulated) revealed that the increase in ER + /PR + breast cancer risk was driven by wine [HR = 1.54 (1.23–1.93) for two drinks or more per day, compared with nondrinkers]. Positive significant trends in ER + /PR + breast cancer risk were observed for total alcohol intake (Ptrend = 0.0006) and wine (Ptrend < 0.0001). For ER – /PR – cases, we also observed positive trends in risk for total alcohol intake (Ptrend = 0.0057), wine (Ptrend = 0.0397), beer (Ptrend = 0.0003), and fortified wine (Ptrend = 0.0236). High beer consumption was significantly associated with ER – /PR – breast cancer risk [HR = 2.52 (1.00–6.36), but with only five cases in the higher category]. Interactions with folate intake, BMI, and use of MHT
We observed interactions, in the postmenopausal period only, between total alcohol and folate intake (P = 0.1192), alcohol and BMI (P = 0.0367), and alcohol and current MHT use (P = 0.1123). Stratifications
according to the median intake of folate (411 g/day), BMI (cut-off point at 25 kg/m2), and current use of MHT were performed, and results are tabulated in Table 3. The associations previously observed between total alcohol intake and breast cancer risk were restricted to the low folate intake category in which consumption of more than two drinks per day was associated with a significantly increased breast cancer risk [HR = 1.35 (1.10–1.67)], whereas there was no association of total alcohol intake with breast cancer risk in the high folate intake category [HR = 1.16 (0.94–1.44)]. However, the trends in risk tended to persist in both low and high folate intake categories (Ptrend = 0.0049 and 0.0077, respectively). Stratification on the basis of BMI revealed that overweight/obese women who consumed two glasses or more per day of total alcohol had a slightly greater risk of postmenopausal breast cancer than women with a BMI under 25 kg/m2. Women who were current users of MHT and were alcohol consumers were not associated with an increased risk for postmenopausal breast cancer [HR = 1.07 (0.88–1.30)], among drinkers of two or more glasses daily compared with nondrinkers. In contrast, non-MHT-users who consumed more than two glasses per day were at a highly significantly increased risk compared with nonconsumers [HR = 1.52 (1.21–1.92)].
Discussion We found a strong direct association between baseline high consumption of alcohol and overall and postmenopausal breast cancer risk. This relation was mainly driven by wine and beer consumption and was restricted to ER + /PR + breast cancers, women with low folate intake, or non-MHT-users. The strength of the association between alcohol intake and breast cancer risk was greater among overweight and obese women. Our results are consistent with the literature, according to which most of the studies concluded that there was an increased risk for breast cancer (Wiseman, 2008; Bagnardi et al., 2012) – mostly cancers occurring in the postmenopausal period and with an ER + status (Lew et al., 2009; Li et al., 2010). Previous meta-analyses and the World Cancer Research Fund review (Hamajima et al., 2002; Wiseman, 2008) have found an increased risk of premenopausal breast cancer associated with alcohol consumption that we did not confirm in our study, probably because of a relatively low number of premenopausal women included. In addition, we found that the association between total alcohol intake and breast cancer was driven by the consumption of wine and beer; previous studies (Hamajima et al., 2002; Key et al., 2006; Wiseman, 2008) did not conclude a differential effect of alcohol depending on the type of beverage.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
212 European Journal of Cancer Prevention 2015, Vol 24 No 3
Table 2 Risk of breast cancer according to baseline total alcohol consumption and types of alcohol in the overall, premenopausal, and postmenopausal populations Multivariateb
Overall population Age-adjusteda N Alcohol (1 drink = 10 g of ethanol) Non-alcohol-consumer 318 < 0.5 drinks/day 806 0.5–1 drinks/day 469 1–2 drinks/day 587 Z 2 drinks/day 632 Ptrend Wine (1 drink = 150 ml) Non-alcohol-consumer 318 < 0.5 drinks/day 1307 0.5–1 drinks/day 495 1–2 drinks/day 422 Z 2 drinks/day 270 Ptrend Beer (1 drink = 250 ml) Non-alcohol-consumer 318 < 0.5 drinks/day 2297 0.5–1 drinks/day 111 1–2 drinks/day 58 Z 2 drinks/day 28 Ptrend Fortified wine (1 drink = 70 ml) Non-alcohol-consumer 318 < 0.5 drinks/day 2382 0.5–1 drinks/day 87 1–2 drinks/day 20 Z 2 drinks/day 5 Ptrend Spirits (1 drink = 40 ml) Non-alcohol-consumer 318 < 0.5 drinks/day 2133 0.5–1 drinks/day 216 1–2 drinks/day 103 Z 2 drinks/day 42 Ptrend
HR (95%CI)
Multivariateb HR (95% CI)
Premenopause N
HR (95% CI)
Postmenopause b
N
HR (95% CI)b
1 (reference) 1.01 (0.88–1.14) 1.04 (0.90–1.20) 1.03 (0.99–1.08) 1.24 (1.08–1.41) < 0.0001
1 1.00 1.01 1.02 1.19
(reference) (0.88–1.14) (0.88–1.17) (0.97–1.07) (1.04–1.36) 0.0003
59 157 89 108 94
1 0.98 0.94 1.00 0.95
(reference) (0.72–1.32) (0.68–1.31) (0.90–1.12) (0.68–1.32) 0.8531
259 649 380 479 538
1 (reference) 1.00 (0.86–1.15) 1.02 (0.87–1.20) 1.02 (0.97–1.07) 1.24 (1.07–1.44) < 0.0001
1 1.02 1.14 1.11 1.30
1 1.01 1.10 1.08 1.25
(reference) (0.89–1.14) (0.96–1.27) (0.93–1.24) (1.06–1.47) 0.001
59 266 90 59 33
1 0.99 1.06 0.84 0.84
(reference) (0.75–1.32) (0.76–1.48) (0.58–1.20) (0.55–1.29) 0.2225
259 1041 405 363 237
1 (reference) 1.00 (0.87–1.15) 1.10 (0.94–1.29) 1.12 (0.95–1.31) 1.33 (1.11–1.58) < 0.0001
1 (reference) 1.06 (0.95–1.20) 1.37 (1.11–1.71) 1.33 (1.00–1.76) 1.89 (1.29–2.79) < 0.0001
1 (reference) 1.04 (0.92–1.17) 1.32 (1.06–1.64) 1.29 (0.98–1.71) 1.90 (1.29–2.79) < 0.0001
59 404 27 10 7
1 0.94 1.49 0.99 2.04
(reference) (0.72–1.24) (0.94–2.36) (0.51–1.94) (0.93–4.48) 0.0174
259 1893 84 48 21
1 1.05 1.26 1.36 1.85
(reference) (0.92–1.20) (0.98–1.61) (1.00–1.86) (1.19–2.89) 0.0005
1 1.08 1.39 1.01 1.11
(reference) (0.96–1.21) (1.09–1.76) (0.64–1.59) (0.46–2.68) 0.1972
1 1.05 1.33 0.96 1.06
(reference) (0.93–1.18) (1.04–1.68) (0.61–1.51) (0.44–2.57) 0.3029
59 431 14 2 1
1 0.97 1.08 0.48 1.26
(reference) (0.74–1.28) (0.60–1.94) (0.12–1.96) (0.17–9.12) 0.7548
259 1951 73 18 4
1 1.06 1.37 1.07 1.02
(reference) (0.93–1.21) (1.06–1.78) (0.67–1.73) (0.38–2.73) 0.2116
1 1.07 1.17 1.17 1.15
(reference) (0.95–1.21) (0.98–1.39) (0.94–1.46) (0.83–1.58) 0.1919
1 1.05 1.12 1.12 1.11
(reference) (0.93–1.18) (0.94–1.33) (0.90–1.40) (0.80–1.53) 0.3237
59 378 39 25 6
1 0.96 0.93 1.33 0.76
(reference) (0.73–1.27) (0.62–1.40) (0.83–2.12) (0.33–1.75) 0.7944
259 1755 177 78 36
1 1.06 1.16 1.06 1.19
(reference) (0.93–1.21) (0.95–1.40) (0.82–1.37) (0.84–1.69) 0.3215
(reference) (0.91–1.16) (0.99–1.32) (0.96–1.29) (1.10–1.53) 0.0001
E3N cohort data (N = 66 481). Bold values indicate P < 0.05. CI, confidence interval; HR, hazard ratio; MHT, menopausal hormone therapy. a Models were age-adjusted. b Models were adjusted for schooling (undergraduate/graduate or postgraduate), physical activity (MET-h/week), age at puberty, nulliparous (yes/no/missing), and age at first pregnancy (first pregnancy before 30/after 30/missing), breastfeeding (yes/no), use of oral contraceptives (ever/never), family history of breast cancer (yes/no), mammograms during the last follow-up period (yes/no), personal history of benign breast disease (yes/no), BMI (< 20/20–25/25–30/ Z 30 kg/m2), use of progestagens in premenopause (current/nonuser/unknown), age at menopause (menopausal women only), menopausal women and use of MHT (premenopause/postmenopause and current use of MHT/postmenopause and non-MHT-user/postmenopause and unknown use of MHT).
The existing literature failed to answer some questions on the interactions of alcohol with other factors. We confirmed an interaction with folate intake, which was suggested by previous studies (Lew et al., 2009). Existing papers did not support interactions between alcohol and BMI or MHT use (Hamajima et al., 2002). In our study, we observed an association only for non-MHT-users and, although the associations were of a higher magnitude among women with BMI greater than 25 kg/m2, trends in risk were present in the both strata of BMI. Biological mechanisms The estrogen pathway
The increase in postmenopausal and ER + /PR + breast cancer risks with alcohol intake can be explained by the associated increase in estrogen production, as high alcohol intake has been previously shown to be an enhancer of
estrogen synthesis (Ginsburg et al., 1996; Purohit, 1998; Nielsen and Gronbaek, 2008). The aromatization increase may also explain the higher increase in risk observed among overweight and obese women in our study. However, our study failed to confirm the hypothesis of a synergistic effect of alcohol and MHT on breast cancer risk in the postmenopausal period, as our relative risks were not higher among MHT users. Other mechanisms
Alcohol was found to be associated with an increase in mammographic density (Conroy et al., 2012; Rauh et al., 2012). Alcohol consumption has also been shown to promote mammary tumor growth and insulin sensitivity (Hong et al., 2010), breast cancer cell invasion by regulating the Nm23-ITGA5 pathway (Wong et al., 2011), and, more recently, HER2 breast cancer development through
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
Alcohol consumption and breast cancer risk subtypes Fagherazzi et al. 213
Table 3 Risk of postmenopausal breast cancer according to baseline alcohol consumption (overall population and stratified by hormone receptor status, folate intake, BMI, and use of hormone replacement therapy)
ER + /PR + Alcohol consumption (drinks/day)a
N (cases)
HR (95% CI)
Low folate intake (Vit B9 < 411 g/day)
ER – /PR – N (cases)
Nonconsumers < 0.5 0.5–1 1–2 Z2 Ptrend
147 387 226 278 328
1 (reference) 1.05 (0.87–1.27) 1.07 (0.86–1.31) 1.03 (0.96–1.10) 1.32 (1.08–1.60) 0.0006 BMI < 25 kg/m2
45 92 58 78 92
Nonconsumers < 0.5 0.5–1 1–2 Z2 Ptrend
190 438 269 336 373
1 (reference) 0.92 (0.77–1.09) 0.99 (0.82–1.20) 1.00 (0.94–1.06) 1.20 (1.01–1.43) 0.0003
69 211 111 143 165
HR (95% CI)
N (cases)
1 (reference) 0.80 (0.56–1.15) 0.90 (0.61–1.33) 1.00 (0.89–1.13) 1.24 (0.87–1.78) 0.0057 BMI Z 25 kg/m2 1 (reference) 1.23 (0.94–1.61) 1.12 (0.82–1.51) 1.08 (0.98–1.19) 1.37 (1.03–1.82) 0.0569
HR (95% CI)
132 341 183 211 265
1 (reference) 1.06 (0.87–1.30) 1.11 (0.88–1.38) 1.01 (0.94–1.09) 1.35 (1.10–1.67) 0.0049 Non-current-MHT user
108 292 157 190 231
1 (reference) 1.17 (0.94–1.46) 1.16 (0.91–1.49) 1.06 (0.98–1.15) 1.52 (1.21–1.92) 0.0002
High folate intake (Vit B9 Z 411 g/day) N (cases)
HR (95% CI)
127 308 197 268 273
1 (reference) 0.94 (0.7–1.16) 0.96 (0.77–1.21) 1.03 (0.96–1.11) 1.16 (0.94–1.44) 0.0077 Current MHT user
151 357 223 289 307
1 (reference) 0.88 (0.73–1.07) 0.93 (0.75–1.14) 0.99 (0.93–1.06) 1.07 (0.88–1.30) 0.0272
E3N cohort data (N = 66 481). Bold values indicate P < 0.05. CI, confidence interval; HR, hazard ratio; MHT, menopausal hormone therapy. a Models were adjusted for schooling (undergraduate/graduate or postgraduate), physical activity (MET-h/week), age at puberty, nulliparous (yes/no/missing) and age at first pregnancy (first pregnancy before 30/after 30/missing), breastfeeding (yes/no), use of oral contraceptives (ever/never), family history of breast cancer (yes/no), mammograms during the last follow-up period (yes/no), personal history of benign breast disease (yes/no), BMI (< 20/20–25/25–30/ Z 30 kg/m2, except for models stratified on BMI), use of progestagens in premenopause (current/nonuser/unknown), age at menopause (menopausal women only), menopausal women and use of MHT (premenopause/postmenopause and current use of MHT/postmenopause and non-MHT-user/postmenopause and unknown use of MHT, except for models stratified on MHT use).
the estrogen signaling pathway in animal models (Wong et al., 2012). In our study, we found that both lifetime and baseline alcohol exposures were independently associated with an increased risk of postmenopausal breast cancer. These findings are in line with the hypothesis of Brooks and Zakhari (2013), who suggested that alcohol could act on breast carcinogenesis not only through a pure tumorpromoter-type mechanism, but also through a weak cumulative carcinogen mechanism throughout life. Strengths and limitations
This study has some limitations. Participants in the E3N cohort belong to a homogeneous health-conscious population, thus reducing the variability of certain characteristics such as alcohol consumption. Moreover, information on diet was not updated during follow-up, although dietary habits may have changed over time. Although this might reduce our power to show differences between groups, it would not bias our estimates. This study also has several strengths. To our knowledge, we are the first prospective cohort study to obtain such an overview of the associations between alcohol consumption – with details on types of alcohol – and breast cancer risk subtypes. The long followup period yielded sufficient statistical power for the analysis of stratified populations and subgroups of breast cancers defined by menopausal status and tumor hormone receptor status. Moreover, we were able to adjust for most of the risk factors for breast cancer. Conclusion
We showed that alcohol consumption was associated with an increased risk of postmenopausal and ER + /PR + breast cancer and that these associations could be modulated by
folate intake and, to a lesser extent, by BMI. Our study adds significant information to the body of knowledge on the complex benefit–risk association of alcohol consumption. We encourage public health authorities to further take into account interactions between alcohol and other factors in public health nutrition programs.
Acknowledgements The authors thank all participants for providing data and practitioners for providing pathology reports. This work was supported by the Institut National du Cancer, the Mutuelle Ge´ne´rale de l’Education Nationale, the Institut de Cance´rologie Gustave Roussy, and the Institut National de la Sante´ et de la Recherche Me´dicale. The study sponsors played no role in the design of the study, analysis or interpretation of data, writing of the manuscript, or the decision to submit the manuscript for publication. Conflicts of interest
There are no conflicts of interest.
References Bagnardi V, Rota M, Botteri E, Tramacere I, Islami F, Fedirko V, et al. (2013). Light alcohol drinking and cancer: a meta-analysis. Ann Oncol 24:301–308. Brooks PJ, Zakhari S (2013). Moderate alcohol consumption and breast cancer in women: from epidemiology to mechanisms and interventions. Alcohol Clin Exp Res 37:23–30. Conroy SM, Koga K, Woolcott CG, Dahl T, Byrne C, Nagata C, et al. (2012). Higher alcohol intake may modify the association between mammographic density and breast cancer: an analysis of three case-control studies. Cancer Epidemiol 36:458–460. Enger SM, Ross RK, Paganini-Hill A, Longnecker MP, Bernstein L (1999). Alcohol consumption and breast cancer oestrogen and progesterone receptor status. Br J Cancer 79:1308–1314.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
214 European Journal of Cancer Prevention 2015, Vol 24 No 3
Ginsburg ES, Mello NK, Mendelson JH, Barbieri RL, Teoh SK, Rothman M, et al. (1996). Effects of alcohol ingestion on estrogens in postmenopausal women. JAMA 276:1747–1751. Hamajima N, Hirose K, Tajima K, Rohan T, Calle EE, Heath CW Jr, et al. (2002). 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 87:1234–1245. Hong J, Holcomb VB, Tekle SA, Fan B, Nunez NP (2010). Alcohol consumption promotes mammary tumor growth and insulin sensitivity. Cancer Lett 294:229–235. Key J, Hodgson S, Omar RZ, Jensen TK, Thompson SG, Boobis AR, et al. (2006). Meta-analysis of studies of alcohol and breast cancer with consideration of the methodological issues. Cancer Causes Control 17: 759–770. Lew JQ, Freedman ND, Leitzmann MF, Brinton LA, Hoover RN, Hollenbeck AR, et al. (2009). Alcohol and risk of breast cancer by histologic type and hormone receptor status in postmenopausal women: the NIH-AARP Diet and Health Study. Am J Epidemiol 170:308–317. Li CI, Chlebowski RT, Freiberg M, Johnson KC, Kuller L, Lane D, et al. (2010). Alcohol consumption and risk of postmenopausal breast cancer by subtype: the Women’s Health Initiative observational study. J Natl Cancer Inst 102:1422–1431. Nasca PC, Liu S, Baptiste MS, Kwon CS, Jacobson H, Metzger BB (1994). Alcohol consumption and breast cancer: estrogen receptor status and histology. Am J Epidemiol 140:980–988.
Nielsen NR, Gronbaek M (2008). Interactions between intakes of alcohol and postmenopausal hormones on risk of breast cancer. Int J Cancer 122: 1109–1113. Purohit V (1998). Moderate alcohol consumption and estrogen levels in postmenopausal women: a review. Alcohol Clin Exp Res 22:994–997. Rauh C, Hack C, Ha¨berle L, Hein A, Engel A, Schrauder MG, et al. (2012). Percent mammographic density and dense area as risk factors for breast cancer. Geburtsh Frauenheilk 72:727–733. Schutze M, Boeing H, Pischon T, Rehm J, Kehoe T, Gmel G, et al. (2011). Alcohol attributable burden of incidence of cancer in eight European countries based on results from prospective cohort study. BMJ 342:d1584. Tjonneland A, Christensen J, Thomsen BL, Olsen A, Stripp C, Overvad K, et al. (2004). Lifetime alcohol consumption and postmenopausal breast cancer rate in Denmark: a prospective cohort study. J Nutr 134:173–178. Van Liere MJ, Lucas F, Clavel F, Slimani N, Villeminot S (1997). Relative validity and reproducibility of a French dietary history questionnaire. Int J Epidemiol 26 (Suppl 1):S128–S136. Wiseman M (2008). The second World Cancer Research Fund/American Institute for Cancer Research expert report. Food, nutrition, physical activity, and the prevention of cancer: a global perspective. Proc Nutr Soc 67:253–256. Wong AW, Paulson QX, Hong J, Stubbins RE, Poh K, Schrader E, et al. (2011). Alcohol promotes breast cancer cell invasion by regulating the Nm23-ITGA5 pathway. J Exp Clin Cancer Res 30:75. Wong AW, Dunlap SM, Holcomb VB, Nunez NP (2012). Alcohol promotes mammary tumor development via the estrogen pathway in estrogen receptor alpha-negative HER2/neu mice. Alcohol Clin Exp Res 36:577–587.
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.
