Polyphenols are a large and diverse group of secondary plant metabolites that occur ubiquitously in plant-based foods, such as fruits, vegetables, tea, wine, coffee and cocoa.1 Over 8000 different polyphenolic compounds have been described in nature and are divided chemically into flavonoids and nonflavonoids. Flavonoids are formed from a common skeleton of diphenylpropanes (C6-C3-C6).1 Nonflavonoids comprise all other polyphenols and have a broad variety of chemical structures. Of these, phenolic acids, lignans, and stilbenes (resveratrol) are of greatest dietary importance.1
Flavonoids exert a wide range of biological activities in vitro that may explain their potential role in the prevention of chronic diseases, including antioxidant, anti-inflammatory, anticarcinogenic, and other bioactivities.2–5 Epidemiologic studies suggest a protective effect of flavonoids against cardiovascular disease (CVD),6,7 neurodegenerative diseases,8,9 and some cancers,10 particularly gastrointestinal cancers.11
Isoflavones and lignans are phytoestrogens that may exhibit a weak estrogenic-like activity.12 This effect may contribute to an observed inverse association between intake and risk of CVD13 and of some cancers, particularly hormone-related cancers such as breast, prostate, ovarian and endometrial cancers.14,15
Overall mortality was inversely associated with both anthocyanidin and flavone intake in a US study,16 whereas dietary quercetin (the most abundant flavonol) and matairesinol (a minor lignan) were inversely related to total mortality in Finnish17 and Dutch18 studies, respectively. To our knowledge, the US study is the only one that has evaluated the relationship between total flavonoid intake and all-cause mortality.16 Nevertheless, there is reasonable evidence that either fruit and vegetable consumption19 or consumption of a Mediterranean diet20 (a dietary pattern characterized by an abundance of flavonoid-rich plant foods) reduces the risk of overall mortality.
We assessed the relation between the intakes of both total and individual subclasses of flavonoids and lignans and all-cause and specific-cause mortality in the adult Spanish cohort from the European Prospective Investigation into Cancer and Nutrition (EPIC). Our hypothesis was that a higher intake of flavonoids/lignans would decrease mortality.
Study Design and Population
The study population was the EPIC-Spain cohort, a large ongoing prospective European study designed to investigate the associations among diet, lifestyle, genetic factors, and chronic diseases, particularly cancer.21 The EPIC-Spain cohort includes 41,438 subjects (15,632 men and 25,806 women), 29–70 years old, recruited between October 1992 and July 1996 from five Spanish regions: three from the north (Asturias, Navarra, and Gipuzkoa) and two from the south (Murcia and Granada).22 Most participants were blood donors and were selected from a range of social backgrounds and both urban and rural areas. At enrolment, all participants gave their informed consent, and the study was approved by the Ethical Committee of the Spanish Carlos III Institute.
Follow-up and Mortality Ascertainment
Mortality data (date and underlying cause of death) were obtained from regional mortality registries and the National Death Index. Causes of death were coded using International Classification of Diseases (ICD)-9 until 1999 and ICD-10 from 1999 to 2008. For this analysis, the follow-up for vital status was completed between May 2008 and August 2009, depending on the center. The mean follow-up period was 13.6 years. After the exclusion of 816 subjects due to implausible dietary information (the lowest and highest 1% of the ratio of total energy intake to energy requirement),23 the final population studied consisted of 40,622 subjects (15,324 men and 25,298 women), of whom 1915 (1157 men and 758 women) died during the follow-up period.
Trained dieticians collected information on usual food intake during the year before recruitment, taking into account seasonal variations of the interview was conducted in person using a computerized diet history questionnaire developed and validated specifically for the EPIC study in Spain.24 Data on food processing and preparation methods, average frequency of consumption, and usual portion size for each food consumed at least twice per month (or once per month for seasonal foods) were collected. Portion size was assessed through a photo series, natural units, and household measures. The questionnaire included a list of more than 600 foods and beverages and about 150 regional recipes. For each food item, the final amount consumed was calculated, taking into account the cooking method and the edible part consumed. Total energy (kcal/day), alcohol (g/day), total dietary fiber (g/day), and vitamin C (mg/day) intakes were estimated using a Spanish food composition table.25
Quantities of flavonoids and lignans subclasses consumed were estimated using our food composition database based on similar databases from the US Department of Agriculture,26–28 the Phenol-Explorer,29 and the UK Food Standards Agency.30 Our food composition database on flavonoids and lignans is detailed elsewhere.31 Each flavonoid subclass is defined as the sum of its individual compounds: anthocyanidins (cyanidin, delphinidin, malvidin, pelargonidin, peonidin, petunidin), flavan-3-ol monomers (catechin, epigallocatechin, epicatechin, epicatechin 3-gallate, epigallocatechin 3-gallate, gallocatechin, catechin 3-gallate), proanthocyanidins (PA) (dimers, trimers, 4–6 monomers, 7–10 monomers, >10 monomers), theaflavins (theaflavin, theaflavin 3,3’-digallate, theaflavin 3’-gallate, theaflavin 3-gallate), flavones (apigenin, luteolin), flavonols (isorhamnetin, kaempferol, myricetin, quercetin), flavanones (eriodictyol, hesperetin, naringenin), and isoflavones (daidzein, genistein, glycetin, biochanin A, formononetin, equol). Flavanols were calculated as the sum of flavan-3-ol monomers, PAs, and theaflavins. Lignans were defined as the sum of secoisolariciresinol, matairesinol, lariciresinol, pinoresinol, enterolactone, and enterodiol. Enterodiol and enterolactone are synthesized from lignans by the gut microbiota of vegetable consuming animals; therefore, these metabolites can be found only in foods of animal origin.32
Information was obtained by trained interviewers on lifestyle characteristics, consisting of level of education, total physical activity (combining both occupational and leisure time activities), and smoking history.21,33 Anthropometric measurements (weight and height) were also measured at recruitment using standardized procedures. Body mass index (BMI) (in kg/m2) was calculated. The history of chronic diseases was assessed by face-to-face interview, in which participants reported previous chronic diseases and regular use of medication.
Polyphenolic-related variables were categorized into quintiles to obtain a better description of the range and variation of the dietary intake of flavonoids and lignans. Comparisons among quintiles were performed for the entire cohort using two-sided chi-square and Kruskal–Wallis tests for categorical and continuous variables, respectively. Associations between dietary variables and mortality were assessed using Cox proportional hazards models. Age was used as the primary time variable, with age at recruitment as the entry time, and age at the date of death or the end of follow-up (whichever came first) as the exit time. The assumption of proportional hazards over time was assessed by a test based on scaled Schoenfeld residuals. The hazard ratio (HR) with corresponding 95% confidence intervals (CIs) was estimated for each quintile of dietary variable, using the lowest quintile as the reference. Trend tests were calculated by assigning the medians of each quintile as scores. Flavonoid and lignan intakes were also analyzed as continuous, after a log2 transformation, with the HR indicating the increase in risk of death with a doubling of flavonoid or lignan intake (mg/day). Interactions between sex, BMI (≤18.5, 18.6–24.9, 25–29.9, and ≥30 kg/m2), smoking status (never, former, current smokers), alcohol consumption (never, former, 0.1–24 g/day, >24 g/day), and flavonoid intake in relation to mortality were evaluated using a likelihood ratio test based on models with and without an interaction term.
All models were stratified by EPIC center, sex and age of entry at study, and adjusted for BMI (≤18.5, 18.6–24.9, 25–29.9, and ≥30 kg/m2), education level (no formal education, primary school, technical or professional training, secondary school, university, and unspecified), physical activity level (inactive, moderately inactive, moderately active, and active), smoking status (never, former from ≤10 years, former from 11–20 years, former from >20 years, current 1–15 cigarettes/day, current 16–25 cigarettes/day, current >25 cigarettes/day, current not specified, and not specified), lifetime alcohol consumption (never consumer, former consumer, >0–6, >6–12, >12–24, >24–60, >60–96, >96 g/day), total energy intake (continuous, kcal/day), vitamin C (continuous, mg/day), and fiber consumptions (continuous, g/day). Sensitivity analyses were carried out by additional adjustment for the presence of diabetes, hyperlipidemia, or hypertension. Additional models were created excluding participants with chronic diseases at recruitment (coronary heart disease, stroke, diabetes mellitus, cancer) and excluding deaths in the first 2 years of follow-up.
A competing risk analysis was carried out to evaluate the association between flavonoid and lignan intakes and cause-specific mortality. A stratified competing risk regression that allows the baseline hazard to vary across levels of the stratification covariate were fitted to estimate the cause-specific mortality HRs.34 Stratification covariates were age at baseline, sex, and center. All models were also adjusted for the same confounders as the previous Cox models.
R statistical software (R Development Core Team. A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2006. http://www.r-project.org/) was used for all analyses.
During the mean 13.6 years of follow-up, 1915 (4.7%) of 40,622 participants died, of which 416 deaths were because of CVD, 956 because of cancer, and 417 because of other causes, with the remaining 126 missing information on cause of death. General baseline characteristics of the study population by quintile of total flavonoid intake are presented in Table 1. Subjects in the highest quintile of flavonoid intake tended to be male, older, and overweight, and to consume more energy and alcohol, but they were more physically active.
Table 2 shows the mean dietary flavonoid and lignan intakes and their distributions. In the EPIC-Spain cohort, mean dietary total flavonoid and lignan intakes were 387 mg/d and 1.0 mg/d, respectively. As indicated by the large differences between means and medians, the distributions were right skewed. Proanthocyanidins were the most important contributor (66%) to total flavonoid intake, followed by flavanones (11%), flavan-3-ol monomers (9%), anthocyanidins (7%), and flavonols (6%). Finally, the contributions of flavones (1%), isoflavones (0.1%), and theaflavins (<0.1%) were minor. The main food sources of total flavonoids were apples (23%), red wine (21%), fruit not specified (13%), and oranges (9%).35
Table 3 presents data on the associations between all-cause mortality and total and subclasses of flavonoid and lignan intakes, stratified for age at recruitment, center and sex, and adjusted for education level, BMI, physical activity, smoking status, lifetime alcohol consumption, total energy, vitamin C and fiber intake. Multivariate-adjusted HRs comparing the fifth to the first quintiles of intake of flavanones and flavonols were 0.60 (95% CI = 0.38–0.94) and 0.59 (0.40–0.88), respectively. There was a slight decrease in mortality with increased intake of flavanols and proanthocyanidins as a continuous variable, but there was not a clear dose response. The suggestion of an inverse relation between total flavonoid and all-cause mortality was observed (highest vs. lowest quintile HR 0.71 [0.49–1.03]), and for the continuous variable (HR for log2 0.94 [0.89–1.00]). There was no evidence of an association for lignans and the remaining flavonoid subclasses. There was also no substantial effect modification by sex (P for interaction = 0.65), BMI (P for interaction = 0.17), alcohol (P for interaction = 0.43), and tobacco (P for interaction = 0.80). Therefore, the associations between mortality and flavonoid intakes divided by sex, BMI, alcohol or tobacco consumption were not presented. We did not consider theaflavins because the proportion of theaflavin consumers in the EPIC-Spain population was too low (4%).
In cause-specific mortality analyses using competing risk regressions (Table 4), doubling the dietary intakes of total flavonoids, flavanones, flavonols, flavanols, and PAs were inversely associated with risk of mortality from CVD. Slight negative associations were observed between a doubling of the dietary intake of flavones and lignans and mortality from CVD. There was no evidence of an association between dietary flavonoid or lignan intake and mortality from cancer or other causes.
In sensitivity analyses that included additional adjustments for diabetes, hypertension, or hyperlipidemia, there was little change in the associations between all-cause mortality and flavanone (HR for log2 0.95 [95% CI = 0.93–0.98]) and total flavonoid (HR for log2 0.94 [0.88–0.99]) intakes. The exclusion of 10,330 subjects with chronic diseases at recruitment (cancer, CVD, or diabetes) also had little effect on the associations between overall mortality and flavanone (HR for log2 0.95 [0.92–0.98]) and total flavonoid intakes (HR for log2 0.95 [0.89–1.03]). In further sensitivity analysis, the exclusion of 109 subjects who had died within the first 2 years of follow-up did not substantially alter the associations between overall mortality and flavanone intake (HR for log2 0.96 [0.94–0.99]) and total flavonoid intake (HR for log2 0.95 [0.89–1.01]).
In this large prospective study, flavanone and flavonol intakes were inversely related to all-cause mortality. We found a 40% and 41% lower risk of mortality for the highest versus the lowest quintile of flavonone and flavonol intake, respectively. Total flavonoid intake was associated with a modest reduction in all-cause mortality, whereas lignans and the remaining flavonoid subclasses were not related to overall mortality.
Only a few epidemiological studies have systematically investigated flavonoid intake in relation to all-cause mortality. Overall mortality was inversely associated with increasing intake of anthocyanidins and flavones in female health professionals enrolled in the Women’s Health Study, but this effect was not seen for total flavonoids or other flavonoid subclasses.16 However, in the Finnish Mobile Clinic Health Examination Survey, all-cause mortality was related only to quercetin intake, although this study evaluated only flavonols and flavanones.17 With regard to nonflavonoid compounds, only one study conducted in The Netherlands has investigated the relationship between lignan intake and mortality.18 Matairesinol, which accounted for around 1% of lignan intake,18 was the only lignan whose intake was inversely associated with all-cause and specific-cause mortality, after adjustment for other effects of its main food sources. In the present study, there was little evidence for an association between total lignan intake and overall mortality.
In relation to mortality from CVD, we found inverse associations with total flavonoid, flavanol, PA, flavonol, and flavanone intakes and slight associations with flavone and lignan intakes. Several epidemiological studies have examined the relationship between flavonoid intake and fatal or nonfatal CVD such as coronary heart disease, stroke, myocardial infarct, and peripheral arterial occlusive disease.6,16 Flavonols, flavones, and flavanones have generally displayed protective effects against CVD, followed by flavan-3-ol monomers, proanthocyanidins, and anthocyanidins. Furthermore, a meta-analysis of randomized clinical trials on CVD and flavonoid and flavonoid-rich foods confirmed that individual flavonoid subclasses vary in their ability to protect against CVD-related mortality.7 Despite the evidence suggesting that flavonoid intake may be inversely related to both CVD incidence and mortality, further large cohort studies are needed to clarify which flavonoid subclasses have more protective effects upon CVD. Mechanisms of action responsible for the beneficial effects of flavonoids on CVD risk could be antioxidant, anti-inflammatory and estrogenic activity, protection from DNA cleavage, enzyme modulation, increased cytokine production (thus regulating immune responses), lipid peroxidation, decreased capillary permeability and fragility, and membrane strengthening.2,4,5
We found no evidence of association between intake of falvonoids and lignans and mortality from cancer. To our knowledge, there are no previous studies assessing cancer mortality and flavonoid/lignan intakes. However, several epidemiological studies support a protective role in the incidence of some cancers such as colorectal, gastric, esophageal and pancreatic,11,36 lung,37 breast, and prostate cancers.30,38 Furthermore, some flavonoids may be related to increased survival of patients with breast cancer39 or decreased recurrence of colorectal adenocarcinoma.40 In vitro, flavonoids showed strong anticarcinogenic properties, particularly their ability to modulate carcinogen metabolism (eg, phase I and II metabolic enzymes), regulate inflammatory pathways (eg, nuclear transcription factor κβ, cyclooxygenase-1, and cyclooxygenase-2), inhibit cell proliferation, and induce apoptosis (eg, intracellular protein β-catenin).3,11
Several epidemiological studies suggest a protective role of flavonoid/lignan intakes on chronic disease risk, although the findings are still inconsistent, particularly among different populations. These could be due to differences in study designs, participant characteristics, dietary assessment, and nutrient database used for analyses. Also, levels of dietary flavonoid vary considerably among populations. The diet in Spain, as in other Mediterranean countries, is rich in fruits and vegetables and therefore high in dietary flavonoid and lignan intakes.35 A recent study found similar results for total dietary flavonoid intakes but significant differences between Mediterranean and non-Mediterranean countries in flavonoid subclass intakes and food sources.31 In the case of flavanones (the most protective subclass in our study), the intake in Spain is two- to three-fold higher than in Scandinavian countries31 or the United States.41
The following limitations should be taken into account when interpreting these data. Firstly, we had only a single baseline assessment of diet and other lifestyle variables. Therefore, changes in diet and lifestyle could not be taken into account in these analyses. Secondly, dietary measurement errors could have led to inaccuracies in the estimation of flavonoid and lignan intake. However, the use of a validated diet history questionnaire,24 which yields fewer measurements errors than food frequency questionnaires, should have reduced these inaccuracies.24 Some underestimation of intake is likely due to unknown composition data, but, to date, our food composition database, which was compiled using recent databases on flavonoids and lignans,26–30 has only 10% missing values in 1877 food items. The omission of dietary supplements in this analysis also potentially reduced the dietary estimation of flavonoids and lignans. Nevertheless, few consumers of herb/plant supplements participated in this study (<2%).42 After adjustments for fruit and vegetable intake, all associations were attenuated, so these protective associations with flavonoids and lignans could be partly explained by other compounds also presented in fruits and vegetables. Another limitation is that information on medication use was not available and so we could not account for nonsteroidal anti-inflammatory drug use, a potentially important confounder particularly in CVD. Another limitation is related to the reverse causality in subjects with CVD or cancer, or with early symptoms of these diseases. These participants may have already altered their diet or lifestyle as a consequence of disease onset. However, adjusting for the presence of diabetes, hypertension, or hyperlipidemia or excluding deaths during the first 2 years of follow-up or excluding subjects with CVD, cancer, or diabetes at baseline did not notably alter the estimators.
In conclusion, this large prospective study shows substantial inverse associations of dietary flavanone and flavonol intakes with all-cause mortality. These results also suggest a modest association between total flavonoid intake and reduced overall mortality. No persuasive associations were observed with lignans or other remaining flavonoid subclasses. This reduction in overall mortality was due entirely to a decrease in mortality from CVD and not from cancer or other causes.
1. Beecher GR. Overview of dietary flavonoids: nomenclature, occurrence and intake. J Nutr. 2003;133:3248S–3254S
2. Cazarolli LH, Zanatta L, Alberton EH, et al. Flavonoids: prospective drug candidates. Mini Rev Med Chem. 2008;8:1429–1440
3. Kampa M, Nifli AP, Notas G, Castanas E. Polyphenols and cancer cell growth. Rev Physiol Biochem Pharmacol. 2007;159:79–113
4. Yao LH, Jiang YM, Shi J, et al. Flavonoids in food and their health benefits. Plant Foods Hum Nutr. 2004;59:113–122
5. Marzocchella L, Fantini M, Benvenuto M, et al. Dietary flavonoids: molecular mechanisms of action as anti- inflammatory agents. Recent Pat Inflamm Allergy Drug Discov. 2011;5:200–220
6. Peterson JJ, Dwyer JT, Jacques PF, McCullough ML. Associations between flavonoids and cardiovascular disease incidence or mortality in European and US populations. Nutr Rev. 2012;70:491–508
7. Hooper L, Kroon PA, Rimm EB, et al. Flavonoids, flavonoid-rich foods, and cardiovascular risk: a meta-analysis of randomized controlled trials. Am J Clin Nutr. 2008;88:38–50
8. Engelhart MJ, Geerlings MI, Ruitenberg A, et al. Dietary intake of antioxidants and risk of Alzheimer disease. JAMA. 2002;287:3223–3229
9. Letenneur L, Proust-Lima C, Le Gouge A, Dartigues JF, Barberger-Gateau P. Flavonoid intake and cognitive decline over a 10-year period. Am J Epidemiol. 2007;165:1364–1371
10. Neuhouser ML. Dietary flavonoids and cancer risk: evidence from human population studies. Nutr Cancer. 2004;50:1–7
11. Pierini R, Gee JM, Belshaw NJ, Johnson IT. Flavonoids and intestinal cancers. Br J Nutr. 2008;99 E(suppl 1):ES53–ES59
12. Setchell KD. Phytoestrogens: the biochemistry, physiology, and implications for human health of soy isoflavones. Am J Clin Nutr. 1998;68:1333S–1346S
13. van der Schouw YT, Kreijkamp-Kaspers S, Peeters PH, Keinan-Boker L, Rimm EB, Grobbee DE. Prospective study on usual dietary phytoestrogen intake and cardiovascular disease risk in Western women. Circulation. 2005;111:465–471
14. Ward HA, Kuhnle GG. Phytoestrogen consumption and association with breast, prostate and colorectal cancer in EPIC Norfolk. Arch Biochem Biophys. 2010;501:170–175
15. Horn-Ross PL, John EM, Canchola AJ, Stewart SL, Lee MM. Phytoestrogen intake and endometrial cancer risk. J Natl Cancer Inst. 2003;95:1158–1164
16. Mink PJ, Scrafford CG, Barraj LM, et al. Flavonoid intake and cardiovascular disease mortality: a prospective study in postmenopausal women. Am J Clin Nutr. 2007;85:895–909
17. Knekt P, Kumpulainen J, Järvinen R, et al. Flavonoid intake and risk of chronic diseases. Am J Clin Nutr. 2002;76:560–568
18. Milder IE, Feskens EJ, Arts IC, Bueno-de-Mesquita HB, Hollman PC, Kromhout D. Intakes of 4 dietary lignans and cause-specific and all-cause mortality in the Zutphen Elderly Study. Am J Clin Nutr. 2006;84:400–405
19. Agudo A, Cabrera L, Amiano P, et al. Fruit and vegetable intakes, dietary antioxidant nutrients, and total mortality in Spanish adults: findings from the Spanish cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC-Spain). Am J Clin Nutr. 2007;85:1634–1642
20. Buckland G, Agudo A, Travier N, et al. Adherence to the Mediterranean diet reduces mortality in the Spanish cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC-Spain). Br J Nutr. 2011;106:1581–1591
21. Riboli E, Hunt KJ, Slimani N, et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr. 2002;5(6B):1113–1124
22. González CA, Navarro C, Martínez C, et al. The European prospective investigation about cancer and nutrition (EPIC). Rev Esp Salud Publica. 2004;78:167–176
23. Ferrari P, Slimani N, Ciampi A, et al. Evaluation of under- and overreporting of energy intake in the 24-hour diet recalls in the European Prospective Investigation into Cancer and Nutrition (EPIC). Public Health Nutr. 2002;5(6B):1329–1345
24. EPIC Group of Spain. . Relative validity and reproducibility of a diet history questionnaire in Spain. I. Foods. Int J Epidemiol. 1997;26(suppl 1):S91–S99
25. Slimani N, Deharveng G, Unwin I, et al. The EPIC nutrient database project (ENDB): a first attempt to standardize nutrient databases across the 10 European countries participating in the EPIC study. Eur J Clin Nutr. 2007;61:1037–1056
26. U.S. Departament of Agriculture. USDA Database for the Proanthocyanidin Content of Selected Foods. 2004 Beltsville, MD U.S. Departament of Agriculture
27. U.S. Departament of Agriculture. USDA Database for the Flavonoid Content of Selected Foods. 2007 Beltsville, MD U.S. Departament of Agriculture
28. U.S. Departament of Agriculture. USDA Database for the Isoflavone Content of Selected Foods. 2008 Beltsville, MD U.S. Departament of Agriculture
29. Neveu V, Perez-Jiménez J, Vos F, et al. Phenol-Explorer: an online comprehensive database on polyphenol contents in foods. Database (Oxford). 2010;2010:bap024
30. Ward HA, Kuhnle GG, Mulligan AA, Lentjes MA, Luben RN, Khaw KT. Breast, colorectal, and prostate cancer risk in the European Prospective Investigation into Cancer and Nutrition-Norfolk in relation to phytoestrogen intake derived from an improved database. Am J Clin Nutr. 2010;91:440–448
31. Zamora-Ros R, Knaze V, Luján-Barroso L, et al. Differences in dietary intakes, food sources and determinants of total flavonoids between Mediterranean and non-Mediterranean countries participating in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Br J Nutr. 2013;109:1498–1507
32. Kuhnle GG, Dell’Aquila C, Aspinall SM, Runswick SA, Mulligan AA, Bingham SA. Phytoestrogen content of foods of animal origin: dairy products, eggs, meat, fish, and seafood. J Agric Food Chem. 2008;56:10099–10104
33. Wareham NJ, Jakes RW, Rennie KL, et al. Validity and repeatability of a simple index derived from the short physical activity questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Public Health Nutr. 2003;6:407–413
34. Zhou B, Latouche A, Rocha V, Fine J. Competing risks regression for stratified data. Biometrics. 2011;67:661–670
35. Zamora-Ros R, Andres-Lacueva C, Lamuela-Raventós RM, et al. Estimation of dietary sources and flavonoid intake in a Spanish adult population (EPIC-Spain). J Am Diet Assoc. 2010;110:390–398
36. Zamora-Ros R, Agudo A, Luján-Barroso L, et al. Dietary flavonoid and lignan intake and gastric adenocarcinoma risk in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Am J Clin Nutr. 2012;96:1398–1408
37. Tang NP, Zhou B, Wang B, Yu RB, Ma J. Flavonoids intake and risk of lung cancer: a meta-analysis. Jpn J Clin Oncol. 2009;39:352–359
38. Magee PJ, Rowland IR. Phyto-oestrogens, their mechanism of action: current evidence for a role in breast and prostate cancer. Br J Nutr. 2004;91:513–531
39. Fink BN, Steck SE, Wolff MS, et al. Dietary flavonoid intake and breast cancer survival among women on Long Island. Cancer Epidemiol Biomarkers Prev. 2007;16:2285–2292
40. Bobe G, Sansbury LB, Albert PS, et al. Dietary flavonoids and colorectal adenoma recurrence in the Polyp Prevention Trial. Cancer Epidemiol Biomarkers Prev. 2008;17:1344–1353
41. Chun OK, Chung SJ, Song WO. Estimated dietary flavonoid intake and major food sources of U.S. adults. J Nutr. 2007;137:1244–1252
42. Skeie G, Braaten T, Hjartåker A, et al. Use of dietary supplements in the European Prospective Investigation into Cancer and Nutrition calibration study. Eur J Clin Nutr. 2009;63(suppl 4):S226–S238