Supplemental Material can be found at: http://jn.nutrition.org/content/suppl/2014/08/20/jn.114.19696 4.DCSupplemental.html
The Journal of Nutrition Nutritional Epidemiology
Evidence for an Association of Dietary Flavonoid Intake with Breast Cancer Risk by Estrogen Receptor Status Is Limited1–3 Ying Wang,4* Susan M. Gapstur,4 Mia M. Gaudet,4 Julia J. Peterson,5 Johanna T. Dwyer,6 and Marjorie L. McCullough4 6
Epidemiology Research Program, American Cancer Society, Atlanta, GA; and 5Friedman School of Nutrition Science and Policy and School of Medicine and Jean Mayer USDA Human Nutrition Center on Aging, Tufts University, Boston, MA
Abstract Background: Results from preclinical studies suggest that flavonoids, which are ubiquitous in plant-based diets, lower breast cancer risk. Epidemiologic studies of flavonoid intake and breast cancer risk, however, are limited, and few investigated associations with the more aggressive estrogen receptor (ER)–negative (ER2) tumors. Objective: We examined the associations between 7 subclasses of dietary flavonoids and invasive postmenopausal breast cancer risk overall and by ER status in a U.S. prospective cohort. Methods: In 1999–2000, 56,630 postmenopausal women completed detailed self-administered questionnaires, among whom 2116 invasive breast cancers were verified during a mean follow-up period of 8.5 y. Cox proportional hazards regression was used to calculate multivariable-adjusted HRs and 95% CIs. Results: Total flavonoid intake was not associated with breast cancer risk. However, there was a modest inverse association between flavone intake and overall breast cancer risk (fifth vs. first quintile HR: 0.88; 95% CI: 0.76, 1.01; P-trend = 0.04) and between flavan-3-ol intake and risk of ER2 breast cancer (for an increment of 40 mg/d; HR: 0.81; 95% CI: 0.67, 0.97) but not for ER-positive (ER+) breast cancer risk. Conclusion: The inverse association of flavan-3-ol intake with ER2 but not ER+ breast cancer is consistent with other studies that suggest a beneficial role of plant-based diets in ER2 breast cancer risk. J. Nutr. 144: 1603–1611, 2014.
Introduction Despite advances in the early detection, treatment, and chemoprevention of breast cancer, this malignancy remains the most commonly diagnosed non–skin cancer and the second most common cause of cancer mortality among women in the United States (1). Breast cancer is considered a heterogeneous disease in that hormone receptor–positive and –negative breast cancers differ in their etiology, response to treatment, and prognosis (2– 4). Several hormone-related risk factors such as early menarche, late age at first birth, and late age at menopause are more strongly related to estrogen receptor (ER)7–positive (ER+) than ER-negative 1 Supported in part with resources from the USDA, Agricultural Research Service, under agreement 58-1950-7-707. The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study II cohort. 2 Author disclosures: Y. Wang, S. M. Gapstur, M. M. Gaudet, J. J. Peterson, J. T. Dwyer, and M. L. McCullough, no conflicts of interest. 3 Supplemental Table 1 is available from the ‘‘Online Supporting Material’’ link in the online posting of the article and from the same link in the online table of contents at http://jn.nutrition.org. 7 Abbreviations used: CPS-II, Cancer Prevention Study II; ER, estrogen receptor; ER+, estrogen receptor positive; ER2, estrogen receptor negative: PR, progesterone receptor. * To whom correspondence should be addressed. E-mail: ying.wang@cancer. org.
(ER2) breast cancers, which are more aggressive and carry a poorer 5-y prognosis (2,4,5). There are no known strategies for reducing the risk of the difficult to treat ER2 breast cancers. However, recent studies suggest that a higher intake of fruits and vegetables is associated with a lower risk of breast cancers (6), especially ER2 breast cancers (7). Large pooled analyses showed inverse associations of both serum (8) and dietary carotenoids (9) with risk of ER2 but not ER+ tumors. In addition to carotenoids, flavonoids, including proanthocyanidins, are also abundant in plant foods. Comprising a large group of polyphenolic compounds, >4000 flavonoids have been identified, with 6 principal subclasses commonly found in human diets (anthocyanins, flavan-3-ols, flavanones, flavones, flavonols, and isoflavones). Proanthocyanidins, polymers of flavan-3-ols, are concentrated in fruits (except for citrus), chocolate or cocoa, and red wine (10,11). Although they have been less studied than the 6 subclasses of monomeric flavonoids because of lack of information to quantify their presence in food, a higher degree of polymerization seems to be associated with stronger inverse associations with cancer risk (12,13). These compounds are potentially effective anticancer agents in vitro and in vivo, possibly due to their antioxidant, anti-inflammatory, antiproliferative, and apoptosisinducing properties as shown in preclinical studies of breast cancer
ã 2014 American Society for Nutrition. Manuscript received May 14, 2014. Initial review completed June 17, 2014. Revision accepted July 25, 2014. First published online August 20, 2014; doi:10.3945/jn.114.196964.
1603
Downloaded from jn.nutrition.org at CALIFORNIA STATE POLYTECH UNIVERSITY POMONA on January 11, 2015
4
Participants and Methods Study population. Participants in this study were drawn from the ;98,000 women enrolled in the CPS-II Nutrition Cohort, a prospective cohort study of cancer incidence and mortality initiated in 1992 (24). The CPS-II Nutrition Cohort is a subset of ;1.2 million U.S. adults in the CPS-II, a prospective cohort study of cancer mortality established by the American Cancer Society in 1982 (25). In 1992–1993, participants in the CPS-II cohort who resided in 21 states with population-based cancer registries were invited to participate in the CPS-II Nutrition Cohort study. At enrollment, participants completed a 10-page, selfadministered questionnaire that included anthropometric, demographic, dietary, lifestyle, and medical information. Follow-up questionnaires to update exposure information and ascertain newly diagnosed cancers were mailed in 1997 and every other year thereafter. All aspects of the Nutrition Cohort study were approved by the Emory University Institutional Review Board. Analytic cohort. Follow-up for this analysis began on the date of completion of the 1999 follow-up questionnaire, when a 152-item modified Willett FFQ, which included specific questions on flavonoidrich foods and beverages, was first administered (24). At baseline in 1999–2000, 73,640 women returned the FFQ. We excluded from the analysis women who were lost to follow-up (n = 1752), who reported prevalence of any cancer except for nonmelanoma skin cancer on or before the 1999 questionnaire (n = 13,655), and who were pre- or perimenopausal or with missing menopausal status (n = 224). We further excluded women who reported extreme energy intakes (>3500 or 70 line items missing on the 1999 FFQ (n = 916) or $50% flavonoid-rich food items missing (n = 144). Women who were in the top 0.1% of individual flavonoid subclass intake were also excluded to limit over-reporting (n = 319). A total of 56,630 postmenopausal women were included in the analytic cohort. Exposure assessment. For each line item on the FFQ, a common food or beverage serving size was specified [e.g., 1 banana or 8 ounces (;237 mL) soy milk]. Participants were asked how often, on average, they had consumed this amount over the previous year. The possible frequency responses ranged from ‘‘never’’ to ‘‘5 or more per week,’’ or to ‘‘6 or more times per day,’’ depending on the item. Dietary intake of 6 subclasses of flavonoids and total proanthocyanidins were derived from 3 USDA databases (20,26,27) as well as other literature (28–30) as described elsewhere (31). Total proanthocyanidins were the sum of monomers, dimers, trimers, 4–6 mers, 7–10 mers, and >10 mers as well as theaflavins and thearubigins, which are derived tannins. Total flavonoids were defined as the sum of the 7 subclasses (anthocyanidins, flavones, 1604
Wang et al.
flavanones, flavan-3-ols, flavonols, isoflavones, and proanthocyanidins), including both monomeric and polymeric flavonoids. Percentage contributions of food sources to flavonoids were calculated among all participants as the ratio of total intake of flavonoids from a food source divided by the total amount of flavonoids consumed in the population (Supplemental Table 1). Outcome ascertainment. Follow-up for the analytic cohort started from the return of the 1999 FFQ and ended at the date of breast cancer diagnosis or breast cancer death, date of other death, date of last survey returned, or 30 June 2009, whichever came first. During a mean followup period of 8.5 y, 2116 invasive breast cancers were identified, including 1498 ER+ cases and 218 ER2 cases. A majority of the incident cases (n = 2075) were self-reported and verified subsequently through medical records (n = 1671) or through linkage with state cancer registries (n = 404). Three self-reported cases that were not verified were identified through linkage with the National Death Index (32). Another 13 cases that were not self-reported were identified through medical records (n = 2) or linkage with state cancer registry (n = 11) during the process of verifying another cancer. Fatal breast cancer cases (n = 25) that were not reported were identified through linkage with the National Death Index, and 17 of these were verified through state cancer registries. Statistical analysis. Energy-adjusted flavonoid intakes, calculated by using the residual method (33), were examined as categorical variables in quintiles based on distributions in the entire cohort as well as continuous variables with ;1 SD as the incremental unit. Cox proportional hazards regression models were used to calculate HRs and 95% CIs for breast cancer risk by quintile of flavonoid intake and by continuous form. When ER+ and ER2 cases were examined as separate outcomes, cases that were non-ER+ or non-ER2, including those missing ER status, were censored. Age was adjusted in all Cox models by stratifying on single year of age at enrollment. For other covariates, missing values 4.5–9 kg, >9–18 kg, >18–25 kg, >25 kg, unknown), education (less than high school graduate, high school graduate, some college, college graduate or higher/unknown), combination of parity and age at first live birth (nulliparous; 1–2,
The Journal of Nutrition Nutritional Epidemiology
Evidence for an Association of Dietary Flavonoid Intake with Breast Cancer Risk by Estrogen Receptor Status Is Limited1–3 Ying Wang,4* Susan M. Gapstur,4 Mia M. Gaudet,4 Julia J. Peterson,5 Johanna T. Dwyer,6 and Marjorie L. McCullough4 6
Epidemiology Research Program, American Cancer Society, Atlanta, GA; and 5Friedman School of Nutrition Science and Policy and School of Medicine and Jean Mayer USDA Human Nutrition Center on Aging, Tufts University, Boston, MA
Abstract Background: Results from preclinical studies suggest that flavonoids, which are ubiquitous in plant-based diets, lower breast cancer risk. Epidemiologic studies of flavonoid intake and breast cancer risk, however, are limited, and few investigated associations with the more aggressive estrogen receptor (ER)–negative (ER2) tumors. Objective: We examined the associations between 7 subclasses of dietary flavonoids and invasive postmenopausal breast cancer risk overall and by ER status in a U.S. prospective cohort. Methods: In 1999–2000, 56,630 postmenopausal women completed detailed self-administered questionnaires, among whom 2116 invasive breast cancers were verified during a mean follow-up period of 8.5 y. Cox proportional hazards regression was used to calculate multivariable-adjusted HRs and 95% CIs. Results: Total flavonoid intake was not associated with breast cancer risk. However, there was a modest inverse association between flavone intake and overall breast cancer risk (fifth vs. first quintile HR: 0.88; 95% CI: 0.76, 1.01; P-trend = 0.04) and between flavan-3-ol intake and risk of ER2 breast cancer (for an increment of 40 mg/d; HR: 0.81; 95% CI: 0.67, 0.97) but not for ER-positive (ER+) breast cancer risk. Conclusion: The inverse association of flavan-3-ol intake with ER2 but not ER+ breast cancer is consistent with other studies that suggest a beneficial role of plant-based diets in ER2 breast cancer risk. J. Nutr. 144: 1603–1611, 2014.
Introduction Despite advances in the early detection, treatment, and chemoprevention of breast cancer, this malignancy remains the most commonly diagnosed non–skin cancer and the second most common cause of cancer mortality among women in the United States (1). Breast cancer is considered a heterogeneous disease in that hormone receptor–positive and –negative breast cancers differ in their etiology, response to treatment, and prognosis (2– 4). Several hormone-related risk factors such as early menarche, late age at first birth, and late age at menopause are more strongly related to estrogen receptor (ER)7–positive (ER+) than ER-negative 1 Supported in part with resources from the USDA, Agricultural Research Service, under agreement 58-1950-7-707. The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study II cohort. 2 Author disclosures: Y. Wang, S. M. Gapstur, M. M. Gaudet, J. J. Peterson, J. T. Dwyer, and M. L. McCullough, no conflicts of interest. 3 Supplemental Table 1 is available from the ‘‘Online Supporting Material’’ link in the online posting of the article and from the same link in the online table of contents at http://jn.nutrition.org. 7 Abbreviations used: CPS-II, Cancer Prevention Study II; ER, estrogen receptor; ER+, estrogen receptor positive; ER2, estrogen receptor negative: PR, progesterone receptor. * To whom correspondence should be addressed. E-mail: ying.wang@cancer. org.
(ER2) breast cancers, which are more aggressive and carry a poorer 5-y prognosis (2,4,5). There are no known strategies for reducing the risk of the difficult to treat ER2 breast cancers. However, recent studies suggest that a higher intake of fruits and vegetables is associated with a lower risk of breast cancers (6), especially ER2 breast cancers (7). Large pooled analyses showed inverse associations of both serum (8) and dietary carotenoids (9) with risk of ER2 but not ER+ tumors. In addition to carotenoids, flavonoids, including proanthocyanidins, are also abundant in plant foods. Comprising a large group of polyphenolic compounds, >4000 flavonoids have been identified, with 6 principal subclasses commonly found in human diets (anthocyanins, flavan-3-ols, flavanones, flavones, flavonols, and isoflavones). Proanthocyanidins, polymers of flavan-3-ols, are concentrated in fruits (except for citrus), chocolate or cocoa, and red wine (10,11). Although they have been less studied than the 6 subclasses of monomeric flavonoids because of lack of information to quantify their presence in food, a higher degree of polymerization seems to be associated with stronger inverse associations with cancer risk (12,13). These compounds are potentially effective anticancer agents in vitro and in vivo, possibly due to their antioxidant, anti-inflammatory, antiproliferative, and apoptosisinducing properties as shown in preclinical studies of breast cancer
ã 2014 American Society for Nutrition. Manuscript received May 14, 2014. Initial review completed June 17, 2014. Revision accepted July 25, 2014. First published online August 20, 2014; doi:10.3945/jn.114.196964.
1603
Downloaded from jn.nutrition.org at CALIFORNIA STATE POLYTECH UNIVERSITY POMONA on January 11, 2015
4
Participants and Methods Study population. Participants in this study were drawn from the ;98,000 women enrolled in the CPS-II Nutrition Cohort, a prospective cohort study of cancer incidence and mortality initiated in 1992 (24). The CPS-II Nutrition Cohort is a subset of ;1.2 million U.S. adults in the CPS-II, a prospective cohort study of cancer mortality established by the American Cancer Society in 1982 (25). In 1992–1993, participants in the CPS-II cohort who resided in 21 states with population-based cancer registries were invited to participate in the CPS-II Nutrition Cohort study. At enrollment, participants completed a 10-page, selfadministered questionnaire that included anthropometric, demographic, dietary, lifestyle, and medical information. Follow-up questionnaires to update exposure information and ascertain newly diagnosed cancers were mailed in 1997 and every other year thereafter. All aspects of the Nutrition Cohort study were approved by the Emory University Institutional Review Board. Analytic cohort. Follow-up for this analysis began on the date of completion of the 1999 follow-up questionnaire, when a 152-item modified Willett FFQ, which included specific questions on flavonoidrich foods and beverages, was first administered (24). At baseline in 1999–2000, 73,640 women returned the FFQ. We excluded from the analysis women who were lost to follow-up (n = 1752), who reported prevalence of any cancer except for nonmelanoma skin cancer on or before the 1999 questionnaire (n = 13,655), and who were pre- or perimenopausal or with missing menopausal status (n = 224). We further excluded women who reported extreme energy intakes (>3500 or 70 line items missing on the 1999 FFQ (n = 916) or $50% flavonoid-rich food items missing (n = 144). Women who were in the top 0.1% of individual flavonoid subclass intake were also excluded to limit over-reporting (n = 319). A total of 56,630 postmenopausal women were included in the analytic cohort. Exposure assessment. For each line item on the FFQ, a common food or beverage serving size was specified [e.g., 1 banana or 8 ounces (;237 mL) soy milk]. Participants were asked how often, on average, they had consumed this amount over the previous year. The possible frequency responses ranged from ‘‘never’’ to ‘‘5 or more per week,’’ or to ‘‘6 or more times per day,’’ depending on the item. Dietary intake of 6 subclasses of flavonoids and total proanthocyanidins were derived from 3 USDA databases (20,26,27) as well as other literature (28–30) as described elsewhere (31). Total proanthocyanidins were the sum of monomers, dimers, trimers, 4–6 mers, 7–10 mers, and >10 mers as well as theaflavins and thearubigins, which are derived tannins. Total flavonoids were defined as the sum of the 7 subclasses (anthocyanidins, flavones, 1604
Wang et al.
flavanones, flavan-3-ols, flavonols, isoflavones, and proanthocyanidins), including both monomeric and polymeric flavonoids. Percentage contributions of food sources to flavonoids were calculated among all participants as the ratio of total intake of flavonoids from a food source divided by the total amount of flavonoids consumed in the population (Supplemental Table 1). Outcome ascertainment. Follow-up for the analytic cohort started from the return of the 1999 FFQ and ended at the date of breast cancer diagnosis or breast cancer death, date of other death, date of last survey returned, or 30 June 2009, whichever came first. During a mean followup period of 8.5 y, 2116 invasive breast cancers were identified, including 1498 ER+ cases and 218 ER2 cases. A majority of the incident cases (n = 2075) were self-reported and verified subsequently through medical records (n = 1671) or through linkage with state cancer registries (n = 404). Three self-reported cases that were not verified were identified through linkage with the National Death Index (32). Another 13 cases that were not self-reported were identified through medical records (n = 2) or linkage with state cancer registry (n = 11) during the process of verifying another cancer. Fatal breast cancer cases (n = 25) that were not reported were identified through linkage with the National Death Index, and 17 of these were verified through state cancer registries. Statistical analysis. Energy-adjusted flavonoid intakes, calculated by using the residual method (33), were examined as categorical variables in quintiles based on distributions in the entire cohort as well as continuous variables with ;1 SD as the incremental unit. Cox proportional hazards regression models were used to calculate HRs and 95% CIs for breast cancer risk by quintile of flavonoid intake and by continuous form. When ER+ and ER2 cases were examined as separate outcomes, cases that were non-ER+ or non-ER2, including those missing ER status, were censored. Age was adjusted in all Cox models by stratifying on single year of age at enrollment. For other covariates, missing values 4.5–9 kg, >9–18 kg, >18–25 kg, >25 kg, unknown), education (less than high school graduate, high school graduate, some college, college graduate or higher/unknown), combination of parity and age at first live birth (nulliparous; 1–2,
