Coffee consumption and breast cancer risk among BRCA1 and BRCA2 mutation carriers
Article first published online: 19 JUL 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 118, Issue 1, pages 103–107, 1 January 2006
How to Cite
Nkondjock, A., Ghadirian, P., Kotsopoulos, J., Lubinski, J., Lynch, H., Kim-Sing, C., Horsman, D., Rosen, B., Isaacs, C., Weber, B., Foulkes, W., Ainsworth, P., Tung, N., Eisen, A., Friedman, E., Eng, C., Sun, P. and Narod, S. A. (2006), Coffee consumption and breast cancer risk among BRCA1 and BRCA2 mutation carriers. Int. J. Cancer, 118: 103–107. doi: 10.1002/ijc.21296
- Issue published online: 26 OCT 2005
- Article first published online: 19 JUL 2005
- Manuscript Accepted: 8 APR 2005
- Manuscript Received: 6 JAN 2005
- Canadian Cancer Society
- Canadian Breast Cancer Research Alliance and National Cancer Institute. Grant Number: P30CA16058
- breast cancer;
- case-control study
Although there are several plausible biologic mechanisms whereby coffee consumption might influence the risk of breast cancer, epidemiologic evidence is limited. We assessed the association between coffee consumption and breast cancer risk among high-risk women who carry BRCA mutations. We performed a matched case-control analysis on 1,690 women with a BRCA1 or BRCA2 mutation from 40 centers in 4 countries. Average lifetime coffee consumption was estimated via a self-administered questionnaire. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using conditional logistic regression. After adjustment for potential confounders, the ORs for breast cancer in BRCA carriers who habitually drank 0, 1–3, 4–5 and 6 or more cups of coffee were 1.00, 0.90 (95% CI 0.72–1.12), 0.75 (95% CI 0.47–1.19) and 0.31 (95% CI 0.13–0.71; p-trend = 0.02). The effect was limited to the consumption of caffeinated coffee. These results suggest that among women with BRCA gene mutation, coffee consumption is unlikely to be harmful and that high levels of consumption may in fact be related to reduced breast cancer risk. © 2005 Wiley-Liss, Inc.
Coffee is among the most widely consumed beverages in the world and is a major source of dietary caffeine. Over the last 2 decades, coffee and caffeine have been studied with regard to their potential role in breast cancer etiology. A number of animal studies have reported caffeine to both stimulate and to suppress breast tumors, depending upon the species and the phase of administration.1 Coffee consumption (or caffeine intake) has been directly associated with plasma estradiol, estrone and sex hormone-binding globulin levels, and inversely associated with testosterone.2, 3 Coffee consumption also induces cell differentiation4 and may inhibit mitosis.5 These effects suggest that coffee consumption may affect breast cancer risk, but epidemiologic evidence supporting the association between coffee consumption and breast cancer risk is scant. A decrease in breast cancer risk in women with high levels of coffee consumption has been reported in several studies,6, 7 but other investigators have detected no relationship between coffee consumption and breast cancer risk.1, 8
It has been estimated that the risk of developing breast cancer among women with a mutation in BRCA1 or BRCA2 is as high as 80% by the age of 70.9 Whether or not coffee affects BRCA-related breast cancer risk is currently unknown. To address this issue, we examined the association between coffee consumption and the risk of breast cancer among women with BRCA1 and BRCA2 gene mutations from 40 participating centers in 4 countries.
Material and methods
Study population and design
The study population has been described elsewhere.10 Briefly, eligible study subjects included women who were currently alive and were known to be carriers of deleterious mutations of the BRCA1 or BRCA2 gene. These women were identified from 40 participating clinical cancer genetics centers in 4 countries and were participants in prior and ongoing clinical research protocols at the host institutions. All study subjects received counseling and provided written informed consent for genetic testing.
Our study was approved by the institutional review boards of the host institutions. In most cases, testing was initially offered to women who had been affected with breast or ovarian cancer. When a BRCA1 or BRCA2 mutation was identified in a proband or relative, genetic testing was offered to other at-risk women in the family. In a few families (<10% of the total), only unaffected carriers were identified. Mutation detection was performed using a range of techniques, but all nucleotide sequences were confirmed by direct sequencing of DNA. A woman was eligible for the current study when the molecular analysis established that she was a carrier of a pathogenic mutation. Most (>95%) of the mutations identified in the study subjects were nonsense mutations, deletions, insertions or small frameshifts.
There was information on cancer history and mutation carrier status for a total of 6,053 women who carried BRCA1 or BRCA2 mutations. A total of 3,579 were excluded because information regarding coffee consumption was not collected (n = 3,464) or if they had a history of ovarian cancer or other cancer prior to the diagnosis of breast cancer (n = 110). After these exclusions, there was a total of 2,474 eligible women, including 1,371 women with a history of invasive breast cancer diagnosed between 1970 and 2002 (potential cases) and 1,103 women who never had breast cancer and who were carriers of a mutation in the BRCA1 or BRCA2 gene (potential controls).
Control subjects were women who never had breast cancer but were known to be carriers of a mutation in the BRCA1 or BRCA2 gene. A single control subject was selected for each case subject matched according to mutation in the same gene (BRCA1 or BRCA2), year of birth (within 1 year) and the country of residence (USA, Quebec, other Canada, Poland and Israel). We did not match by family because in most cases a suitable control from within the family could not be found. A diagnosis of ovarian or other form of cancer in the control had to be after the year of diagnosis of the matched case subject. In addition, the date of interview of the controls was after the breast cancer diagnosis of the case. A total of 845 matched case-control pairs were generated for the analysis, including 652 pairs with BRCA1 mutations and 193 pairs with BRCA2 mutations. The questionnaires were completed between 1977 and 2002. For the cases, an average of 7.8 years had elapsed from the date of diagnosis of breast cancer until the date the questionnaire was completed (range 0–42 years).
Case and control subjects at all participating centers completed a standardized questionnaire that asked for all relevant information regarding family history, reproductive and medical histories and selected lifestyle factors including smoking history and use of oral contraceptives. Questionnaires were administered by each of the individual centers at the time of a clinic appointment or at their home at a later date. Additional variables of interest included information on demography, ethnicity, as well as alcohol and coffee consumption. The questionnaire was completed at the time blood was drawn for genetic testing or within a year of receiving the test result. The study subjects were asked if they ever consumed coffee, if they currently consumed coffee, at what age they first began coffee consumption, at what age they stopped drinking coffee and their average daily coffee consumption over this period. Questions were asked separately for caffeinated and decaffeinated coffee. The information on smoking history was collected similarly to coffee use. The questionnaire also asked for information regarding each pregnancy (year of pregnancy and outcome). With respect to alcohol use, information regarding alcohol use (ever/never) as well as the number of alcoholic drinks per week was available; however, age first started alcohol use was not. Therefore, in the analyses, parity, smoking status and coffee use were all censored at least one calendar year prior to the diagnosis of the case, but this was not possible for alcohol use.
The Student's t-test was used to compare continuous variables between case and control subjects. The χ2 test was used to test for differences in categorical variables. The odds ratios (ORs) and 95% confidence intervals (CIs) for breast cancer associated with coffee consumption prior to breast cancer diagnosis of the case were estimated using conditional logistic regression for matched sets. A multivariate analysis was carried out to control for the potential confounding effects of oral contraceptive (OC) use, smoking, alcohol use, parity and body mass index (BMI) at age 30 (<25, 25–29 and ≥30 kg/m2). Smoking use was coded as ever or never smoker; OC and alcohol use was coded as ever or never user; and parity was coded as 0, 1, 2, 3 and ≥4 or more births. Data analyses were performed for BRCA1 and BRCA2 mutation carriers separately and for caffeinated and decaffeinated coffee. All tests of statistical significance were 2-sided.
Cases were similar to controls in terms of age, parity, age at first birth, oral contraceptive use, alcohol use and BMI (Table I). Cases were more likely to have earlier menarche than controls (p = 0.02). Coffee consumption was measured as the average number of cups of coffee that had been drunk per day (caffeinated and decaffeinated) during the time of consumption. There was no significant difference in the mean age the cases and controls first began drinking coffee or at what age they stopped. On average, cases drank less caffeinated coffee than controls (1.38 cups vs. 1.54 cups) but drank equivalent amounts of decaffeinated coffee.
|Characteristic||Cases(n = 845) n (%)||Controls(n = 845) n (%)||p-value1|
|<30||135 (16.0)||149 (17.6)|
|30–34||53 (6.3)||61 (7.2)|
|35–39||129 (15.3)||135 (16.0)|
|45–49||156 (18.4)||166 (19.6)|
|50–54||181 (21.4)||156 (18.5)|
|55–59||96 (11.4)||82 (9.7)|
|60–64||95 (11.2)||96 (11.4)||0.59|
|Place of residence|
|USA||325 (38.5)||325 (38.5)|
|Canada||294 (34.8)||294 (34.8)|
|Poland||207 (24.5)||207 (24.5)|
|Israel||19 (2.2)||19 (2.2)|
|Type of mutation|
|BRCA1||652 (77.2)||652 (77.2)|
|BRCA2||193 (22.8)||193 (22.8)|
|Age at menarche, mean||12.8||13.0||0.02|
|Age at first birth, mean||25.3||25.5||0.64|
|Yes||192 (22.8)||168 (20.0)|
|No||651 (77.2)||671 (80.0)||0.17|
|Oral contraceptive use|
|Ever||526 (63.0)||516 (62.2)|
|Never||309 (37.0)||313 (37.8)||0.75|
|Ever||382 (45.2)||341 (40.4)|
|Never||463 (54.8)||503 (59.6)||0.05|
|Ever||484 (57.3)||488 (57.8)|
|Never||361 (43.7)||355 (42.2)||0.84|
|Body mass index, kg/m2|
|<25||687 (81.2)||690 (83.8)|
|25–29.9||110 (13.3)||102 (12.4)|
|≥30||29 (3.5)||31 (3.8)||0.83|
|Caffeinated coffee consumption|
|Yes||547 (64.7)||562 (66.5)|
|No||298 (35.3)||283 (33.5)||0.44|
|Decaffeinated coffee consumption|
|Yes||107 (12.7)||112 (13.3)|
|No||738 (87.3)||733 (86.7)||0.72|
|Total coffee consumption|
|Yes||581 (68.8)||599 (70.9)|
|No||264 (31.2)||246 (29.1)||0.34|
We next examined the association between coffee consumption and breast cancer risk (Table II). After adjustment for parity, smoking, OC use, alcohol use and BMI, the ORs for breast cancer in women who drank 0, 1–3, 4–5 and ≥6 cups of coffee were 1.00, 0.90 (95% CI 0.72–1.12), 0.75 (95% CI 0.47–1.19) and 0.31 (95% CI 0.13–0.71) (p-trend = 0.02). When the analysis was restricted to decaffeinated coffee, no trend was evident. The multivariate odds ratios for breast cancer for women with a daily consumption of 0, 1–3 and ≥4 decaffeinated cups were 1.00, 0.99 (95% CI 0.72–1.36) and 1.14 (95% CI 0.30–4.31) (p-trend = 1.00). After stratifying by mutation status, the multivariate ORs associated with daily consumption of ≥6 cups of caffeinated coffee per day were 0.25 (95% CI 0.09–0.71) and 0.40 (95% CI 0.09–1.73) for BRCA1 and BRCA2 mutation carriers, respectively (Table III). Among women with a BRCA1 mutation, there was a significant protective trend for increasing consumption of caffeinated coffee (p-trend = 0.009). This was not observed among BRCA2 mutation carriers (p-trend = 0.84) but the sample size was small. The protective effect was similar in all countries; per additional cup of coffee drunk daily, the reduction in breast cancer risk was 15% in Canada (95% CI 0–34%), 17% in Poland (95% CI 0–48%) and 24% in the USA (95% CI 2–41%) (there were too few pairs from Israel to conduct a subgroup analysis).
|Type of coffee||Coffee consumption||p-trend|
|0 cup/day||1–3 cups/day||4–5 cups/day||≥6 cups/day|
|Univariate OR||1.00||0.95 (0.78–1.17)||0.88 (0.57–1.35)||0.42 (0.20–0.89)||0.10|
|Multivariate OR (95% CI)||1.00||0.90 (0.72–1.12)||0.75 (0.47–1.19)||0.31 (0.13–0.71)||0.02|
|Univariate OR||1.00||0.93 (0.68–1.27)||1.24 (0.33–4.62)||NA|
|Multivariate OR (95% CI)||1.00||0.99 (0.72–1.36)||1.14 (0.30–4.31)||NA||1.00|
|Total (caffeinated + decaffeinated)|
|Univariate OR||1.00||0.93 (0.75–1.15)||0.80 (0.54–1.18)||0.62 (0.34–1.15)||0.10|
|Multivariate OR (95% CI)||1.00||0.89 (0.70–1.13)||0.73 (0.48–1.10)||0.51 (0.26–0.98)||0.03|
|BRCA mutation||Coffee consumption||p-trend|
|0 cup/day||1–3 cups/day||4–5 cups/day||≥6 cups/day|
|BRCA1 mutation carriers|
|Univariate OR||1||0.90 (0.72–1.14)||0.87 (0.52–1.45)||0.33 (0.13–0.84)||0.07|
|Multivariate OR (95% CI)||1||0.82 (0.64–1.06)||0.67 (0.39–1.16)||0.25 (0.09–0.71)||0.009|
|BRCA2 mutation carriers|
|Univariate OR||1||1.15 (0.74–1.80)||0.95 (0.44–2.08)||0.71 (0.20–2.59)||0.88|
|Multivariate OR (95% CI)||1||1.26 (0.78–2.08)||1.17 (0.48–2.83)||0.40 (0.09–1.73)||0.84|
We also analysed the protective effect by age of breast cancer diagnosis (Table IV). The protective effect was significant for women diagnosed under the age of 50, and a trend was apparent. A significant risk reduction was not seen for women diagnosed with breast cancer after age 50, but this study group was small.
|Age at diagnosis||Coffee consumption||p-trend|
|0 cup/day||1–3 cups/day||4–5 cups/day||≥6 cups/day|
|Age at diagnosis ≤50|
|Univariate OR||1||0.96 (0.78–1.20)||0.84 (0.54–1.30)||0.41 (0.19–0.92)||0.09|
|Multivariate OR (95% CI)||1||0.91 (0.72–1.15)||0.72 (0.45–1.17)||0.30 (0.12–0.72)||0.02|
|Age at diagnosis >50|
|Univariate OR||1||0.84 (0.43–1.61)||2.21 (0.36–13.7)||0.78 (0.05–11.9)||0.89|
|Multivariate OR (95% CI)||1||0.74 (0.35–1.56)||1.66 (0.23–12.0)||0.62 (0.04–9.96)||0.61|
Our study indicates that women with BRCA1 or BRCA2 mutations who consumed at least 6 cups of coffee per day appeared to have a statistically significant reduction in breast cancer risk, relative to BRCA mutation carriers who have never drunk coffee. The stronger effect observed among BRCA1 mutation carriers is likely attributed to a larger sample size. There was no clear protective effect of coffee consumed at levels less than this. Ours is the first study to our knowledge of coffee consumption and breast cancer risk in BRCA mutation carriers—previous studies have been in the nontested population (which are mostly noncarriers). No consistent association has been seen among the case-control studies that have evaluated the relationship between coffee (or caffeine) consumption and the incidence of sporadic breast cancer,1, 11, 12 although a few studies have reported a statistically significant inverse association.6, 7, 13, 14 Preliminary results from the Nurses' Health Study suggest a weak inverse association between caffeine intake (RR = 0.85; 95% CI 0.72–1.00 for the fifth vs. the first quintile) and the risk of breast cancer.15 A cohort of 14,593 lean Norwegian women who drank ≥5 cups of coffee per day experienced a statistically significant 50% decrease in the risk of breast cancer compared to those who drank ≤2 cups.16 No significant associations between coffee or caffeine intake and breast cancer incidence were reported in 5 other cohort studies.8, 17, 18
Recently, a protective association between coffee consumption and liver cancer was reported in Japan.21, 22 In one study, daily coffee consumption was found to reduce the risk of primary liver cancer by 42% (95% CI 4–64%), and in another study a risk reduction of 51% (95% CI 33–64%) was reported.22
The protective effect associated with coffee consumption may be due to the ability of caffeine to influence estrogen metabolism. Metabolism of estradiol (E2; the principal estrogen produced and secreted by the ovaries of premenopausal women) occurs via 2 hydroxylation pathways yielding products with contrasting estrogenic properties. The first pathway yields 2-hydroxyestrone (OHE), the less potent estrogen,23 whereas the alternate route favors the production of 16 α-OHE, the more potent estrogen metabolite.24 Two-hydroxylation of endogenous estradiol is catalyzed by CYP1A1 and CYP1A2,25 whereas, CYP3A4 is responsible for 16-hydroxylation of endogenous estradiol.26 Several cohort and case-control studies have reported a low risk of (either pre- or postmenopausal) breast cancer in women with a high ratio of 2-OHE to 16α-OHE.27, 28, 29, 30, 31, 32 A recent prospective study of 10,876 Italian women showed that premenopausal women with a 2-OHE to 16α-OHE ratio in the highest vs. the lowest quintile experienced a reduced risk of breast cancer, but the association was not significant (OR = 0.58; 95% CI 0.25–1.34).30 A case-control study of postmenopausal women reported a strong inverse association between the 2-OHE/16α-OHE ratio and risk of breast cancer (OR = 32.7; 95% CI 3.4–3.19 and OR = 9.7; 95% CI 1.3–7.5) for the lowest and for the intermediate tertiles, respectively, relative to the highest.27 Collectively, the evidence suggests an inverse relationship between a low 2-OHE/16α-OHE ratio and the risk of breast cancer, indicating the potential significance of this ratio as a predictive biomarker of breast cancer risk.
Nongenetic or environmental modifiers of breast cancer risk in BRCA-mutation carriers include hormonal factors, particularly those related to estrogen exposure or depletion.33 Currently, nonsurgical chemoprevention options available for BRCA carriers are based on interrupting the estrogen-signalling pathway and implicate an important role of sex hormones in the etiology of the disease. Although the majority of BRCA1 (but not BRCA2) tumours are estrogen receptor negative, oophorectomy34, 35, 36 and tamoxifen37 have been shown to reduce the risk of breast cancer among carriers of either mutation. Based on this observation, alternative dietary and lifestyle strategies that minimize exposure to estrogen or alter the 2-OHE/16α-OHE ratio may also impact overall risk.38
Cigarette smoking, certain drugs, cruciferous vegetables and caffeine have all been shown to induce CYP1A1 and CYP1A2 activity, subsequently enhancing the production of 2-OHE and limiting the production of 16α-OHE.39, 40 Coffee consumption and cigarette smoking may be positively correlated.41, 42 We previously found no association between smoking and risk of breast cancer in this study population;43 nevertheless, we did take into account the possible confounding effect of smoking in our analysis and found no modification of the coffee effect. Recently, CYP1A2 activity was positively related to caffeine intake among premenopausal women.44 We found a significant positive correlation between plasma 2-OHE/16α-OHE levels and daily coffee consumption among young women from 4 different ethnic groups.45 Although caffeine appears to be an inducer of CYP1A2 activity,46 future studies are necessary to evaluate whether coffee (or caffeine) may protect against breast cancer development in BRCA-mutation carriers by specifically altering the 2-OHE/16α-OHE ratio via enzyme induction or through an alternate pathway. If this is indeed the case, interventions with dietary supplements such as diindolylmethane (DIM), which have been shown to upregulate CYP activity, may be a better approach to altering the ratio of estrogen metabolites and possibly reducing breast cancer risk.47 In our study, a protective effect of coffee was not seen until a daily level of 6 cups of caffeinated coffee was reached. Ideally, a compound such as DIM can provide the same level of benefit as coffee, without the side effects associated with excess caffeine consumption. However, further studies need be conducted to confirm the beneficial effect of coffee and to establish the mechanism of protection.
Coffee is an important source of phytoestrogens;48, 49 these may exhibit chemoprotective effects and may have a beneficial influence on the risk of breast cancer.50 The mechanism by which phytoestrogens may beneficially influence the risk of breast cancer has predominantly been attributed to their structural similarity to endogenous estrogens and their ability to bind to estrogen receptors.50, 51 Coffee consumption has been associated with a higher level of circulating sex hormone binding globulin, which decreases the level of bioavailable estrogen.2, 3, 45
The major strength of our study is the large number of known BRCA carriers. A matched analysis was performed so that possible confounding effects, including variation in patterns of coffee consumption between countries, could be minimized. The role of bias and confounding must also be taken into account when interpreting the observed results. The retrospective assessment of coffee intake may have introduced errors in measurement. Recall bias is an inherent weakness of case-control studies, and women with breast cancer might be more likely to overestimate their coffee intake if they believe it is related to their disease. However, this would have led to a spurious positive association between coffee and breast cancer risk. On the contrary, we reported a protective effect.
Our analyses were limited to coffee consumption of cases and controls prior to the age of breast cancer diagnosis of the case to avoid the potential bias due to disease diagnosis prompting a change in the dietary habits. There was no difference in the age at which cases and controls first started drinking coffee or at what age they stopped, thus, it is likely that duration of coffee consumption was not influential, whereas mean daily consumption was. We cannot completely rule out the potential residual confounding effects of unmeasured dietary factors, such as total energy intake but this information was not collected at baseline. Similarly, we did not have information on other dietary sources of caffeine, such as tea, chocolate and soft drinks, which might also be important.
In conclusion, our findings suggest that relatively high coffee consumption may reduce breast cancer risk among women with BRCA1 and/or BRCA2 gene mutations. It will be of importance to confirm this association in other populations, and if confirmed, to explore the biologic basis for this observation.
A.N. is a Research Fellow funded by the Canadian Cancer Society through an award from the National Cancer Institute of Canada. C.E. is a Doris Duke Distinguished Clinical Scientist. This work is supported by a grant from the Canadian Breast Cancer Research Alliance and National Cancer Institute grant P30CA16058 to The Ohio State University Comprehensive Cancer Center. We also thank Drs. M. Daly, D. Fishman, J. Garber, D. Gilchrist, B. Karlan, E. Lemire, W. McKinnon, J. McLennan, S. Merajver, W.S. Meschino, O.I. Olopade, M. Osborne, D. Provencher, H. Saal, K. Sweet, E. Warner, M. Wood and G. Weitzel for their contribution of patients to our study.
- 1World Cancer Research Fund. Food, nutrition and the prevention of cancer: a global perspective. 1997. Washington, DC: American Institute for Cancer Research.
- 15A prospective study of caffeine, coffee, tea, and breast cancer. Am J Epidemiol 1992; 136: 1000–1 (Abstract)., , , , .
- 32Estrogen metabolites and breast cancer: a prospective study. Era of Hope Department of Defense Breast Cancer Research Program Meeting 2000; II: 471., , , , , .
- 40Estrogen metabolism and the diet-cancer connection: rationale for assessing the ratio of urinary hydroxylated estrogen metabolites. Altern Med Rev 2002; 7: 112–29., , .
- 41Alcohol and tobacco: relations and possible interactions in health and disease. In: AvogaroP, SirtoriC, TremoliE, eds. Metabolic effects of alcohol. The Netherlands: North-Holland Biomedical Press, 1979; 143–54., , .