Intake of specific flavonoid classes and coronary heart disease—a case–control study in Greece


Objective: Dietary intake of flavonoids has been reported to protect against coronary heart disease (CHD) risk, but associations of specific classes of flavonoids with CHD have not been adequately studied.

Design: Hospital-based case–control study relying on interviewer administered questionnaires.

Setting: Cardiology Department of the University of Athens Medical School in the Hippokrateion General Hospital (1990–1991).

Subjects: Cases were 329 patients with electrocardiographically confirmed first coronary infarct or a first positive coronary arteriogram, or both (participation fraction 93%). Controls were 570 patients admitted to the same hospital for minor conditions unrelated to nutrition (participation fraction 95%). All cases and controls were interviewed in the hospital wards by experienced interviewers, and a 110-item food frequency questionnaire was administered.

Results: There was statistically significant evidence (P0.03) for an inverse association between intake of flavan-3-ols and CHD risk, an increase of about 21 mg per day corresponding to a 24% decrease in CHD risk. The inverse association between flavan-3-ols and CHD risk was largely accounted for by the intake of wine and to a lesser extent tea. For none of the other flavonoid classes was there statistically significant evidence of an association.

Conclusion: Flavan-3-ols, which are largely found in wine and tea, are inversely associated with, and may be protective against, coronary heart disease.


Flavonoids occur naturally in plants and have strong antioxidant, anti-inflammatory, antithrombotic and endothelial protection properties (Gryglewski et al, 1987; Soleas et al, 1997; Peterson and Dwyer, 1998; Puddey and Croft, 1999; Giugliano, 2000; Middleton et al, 2000). Six classes of flavonoids common in foods are usually considered: flavones, flavonols, flavan-3-ols, flavanones, anthocyanidins and isoflavones. Flavones, particularly apigenin and luteolin, are mainly present in leafy vegetables and herbs. Flavonols are found in many plant foods; they include the glycosides of quercetin in fruits, leaves and vegetables, kaempferol in many fruits and leafy vegetables, isorhamnetin in onions and pears, and myricetin in berries, maize and tea. Flavan-3-ols, specifically the catechins, are abundant in ripe fruits, leaves, tea and red wine. The major sources of the flavanone class in foods are citrus fruits and juices. The anthocyanidins are most abundant in fruits and red wine and less frequent in vegetables. The isoflavones include the compounds daidzein and genistein and are found mainly in soya and soya products (Peterson and Dwyer, 1998).

The biological properties of flavonoids and the inverse association of several flavonoid-rich foods (including wine and tea) with coronary heart disease (CHD) risk have led to the hypothesis that dietary intake of flavonoids or specific flavonoid classes may protect against CHD. Several studies have examined the intake of various combinations of flavonoids, thought to reflect total flavonoid intake (Hertog et al, 1993, 1995; Knekt et al, 1996; Kromhout et al, 1996; Yochum et al, 1999) in relation to CHD risk and the results were generally compatible with the hypothesis that at least some flavonoids convey a degree of protection against CHD. More detailed investigation of the effects of specific classes of flavonoids has been hindered by the fact that flavonoid databases have not been generally available until recently. In 2002 and 2003, however, the United States Department of Agriculture published food composition data for the previously indicated six classes of flavonoids as well as for specific flavonoid compounds within these six classes (US Department of Agriculture-Iowa State University, 2002; US Department of Agriculture, 2003). Relying on these databases, we have studied the association of each major flavonoid class with CHD risk using data from a case–control study undertaken in Greece in the early 1990s (Kalandidi et al, 1992; Tzonou et al, 1993).

Patients and methods

Between January 1990 and April 1991, 355 patients with a first definite diagnosis of CHD were hospitalized at the Cardiology Department of the University of Athens Medical School in the Hippokrateion General Hospital and had survived for at least 4 days after hospitalization. Either a first positive coronary arteriogram or an electrocardiographically confirmed first coronary infarct, or both, determined CHD diagnosis. Of these, 329 (93%) agreed to participate in a study of the dietary etiology of CHD. Eligible controls were 603 patients admitted to the same hospital just before or after the CHD cases for diseases considered as unrelated to diet (minor surgery; eye, ear, nose problems; urologic problems; respiratory problems unrelated to CHD). Of these control patients, 570 (95%) agreed to participate in the study (Kalandidi et al, 1992; Tzonou et al, 1993).

Four trained research assistants interviewed CHD and control patients in the hospital wards for approximately 1 h. Data concerning demographic, socioeconomic and biomedical variables were recorded, and dietary intakes were assessed. Specifically, all patients were asked to indicate the average frequency of consumption of 110 food items or beverage categories over a period of 1 y before the onset of the present disease. An expanded version of this questionnaire has been subsequently validated (Gnardellis et al, 1995). Details about the study, as well as results concerning major nondietary and dietary CHD risk factors have been previously reported (Kalandidi et al, 1992; Tzonou et al, 1993).

Food frequencies were translated into quantities using typical portion sizes for each food item or beverage category, and quantities were converted into average daily intakes of energy and macronutrients using a standard nutrient database (Polychronopoulou-Trichopoulou, 1989; Katsouyanni et al, 1991; Trichopoulou, 1992; Tzonou et al, 1993). Food and beverage contents of flavones, flavonols, flavan-3-ols, flavanones, anthocyanidins and isoflavones were obtained from the US Department of Agriculture databases (US Department of Agriculture-Iowa State University Database, 2002; US Department of Agriculture, 2003).

Initially, the data were studied through simple cross-tabulations, using marginal quintiles for the six studied flavonoid categories, as well as for total flavonoid intake, defined as the sum of the intake of the six evaluated categories. Subsequently, the data were modeled through multiple logistic regression (Breslow and Day, 1980), using the SPSS statistical package (Statistical Package for Social Sciences v. 11.5, Chicago, IL, USA). As the effects of several established nondietary and dietary risk factors for CHD were evident in the data for this study (Kalandidi et al, 1992; Tzonou et al, 1993) a core model was used that included age (in years, continuously; a study design factor), gender (male vs female; a design factor), place of residence (categorically; a design factor), interviewer (categorically; a design factor), years of schooling (continuously), body mass index (continuously), systolic blood pressure (hypertensive vs normotensive), regular exercise (yes vs no), smoking habits (never and ex-smokers vs current smokers and number of cigarettes per day for current smokers – continuously), alcoholic beverage intake (in glasses per day; ≤1, >1 and <3, ≥3, categorically), coffee intake (in cups per day; <1, ≥1 and <3, ≥3, categorically), as well as total energy, saturated lipid and unsaturated lipid intake (all three continuously). We did not include in the models blood glucose or cholesterol levels, because we consider them as intermediate variables reflecting nutritional intakes and because there were many missing values among controls (8 and 50%, respectively). CHD incidence odds ratios were expressed per increment of each of the evaluated categories of flavonoids equal to one standard deviation.


Table 1 shows the demographic and lifestyle characteristics of the 329 patients with coronary heart disease and the 570 control subjects. The data in this table are shown for descriptive purposes, but they are not directly interpretable because of mutual confounding. After mutual adjustment (Kalandidi et al, 1992; Tzonou et al, 1993), there was a higher proportion of men and central Athens residents among cases than among controls, cases had a higher body mass index, were more educated, more frequently hypertensive and more frequently coffee drinkers in comparison to controls (P<0.05 in all these instances), whereas there was no significant difference with respect to the remaining variables listed in Table 1. Nevertheless, all the variables in Table 1 were adjusted for in the analyses focusing on the relation of flavonoid intake with CHD risk.

Table 1 Demographic and lifestyle characteristics of 329 patients with coronary heart disease and 570 control subjectsa

Table 2 shows the distribution of cases and controls by quintiles of intake of the studied flavonoid classes, as well as total flavonoids. On the basis of the median values, flavanones are the most abundant flavonoids in the Greek diet, followed by flavan-3-ols. In contrast, there is minimal intake of isoflavones in this diet. There is no evidence of a significance difference between cases and controls in the intake of any of the studied flavonoid classes, after adjustment for age and gender, but these results do not take into account possible confounding by other dietary or nondietary factors.

Table 2 Distribution of 329 patients with CHD and 570 control subjects by marginal quintiles of the indicated nutritional variables

In Table 3, multiple logistic regression-derived odds ratios for coronary heart disease per one standard deviation increment of each of the major classes of flavonoids are presented. There is statistically significant evidence (P0.03) for an inverse association between intake of flavan-3-ols and CHD risk, an increase of about 21 mg per day corresponding to a 24% decrease in CHD risk. Since intake of alcohol, particularly red wine, is an important determinant of flavan-3-ol intake, we also performed the analysis without this variable among the potential confounders, but this changed the odds ratio for flavan-3-ols only slightly to 0.75 (95% confidence interval 0.59–0.96). For none of the other flavonoid classes, nor for total flavonoid intake, was there statistically significant evidence of an association with CHD, although the evidence with respect to anthocyanidins may be considered as suggestive. Mutually controlling among the six categories of flavonoids further strengthened the inverse association between flavan-3-ols and coronary heart disease, the respective odds ratio becoming 0.56 (95% confidence interval 0.35–0.90), while no other significant association with any flavonoid category was noted. In further analyses, the inverse association between flavan-3-ols and CHD risk was found to be largely accounted for by the intake of wine and to a lesser extent tea. Thus, controlling for wine but not tea gave an odds ratio of 0.81 (95% confidence interval 0.61–1.07), controlling for tea but not wine an odds ratio of 0.78 (95% confidence interval 0.60–1.03), while controlling for both tea and wine gave an odds ratio for flavan-3-ols of 0.86 (95% confidence interval 0.62–1.20).

Table 3 Multiple logistic regression-derived odds ratios for coronary heart disease per one standard deviation increment of the major categories of flavonoids


In this case–control study in the Greek population, which is characterized by a high intake of flavonoid-containing foods (except isoflavones), we have found evidence that flavan-3-ols, largely derived from wine and tea, are significantly inversely associated with CHD risk. The apparent protection is considerable, an increase in the intake of about 21 mg per day (1 standard deviation) being associated with a 24% reduction in the risk of CHD. Mutual adjustment among flavonoid categories indicated an even sharper reduction amounting to 44%. There was also some evidence, which did not reach statistical significance, that intake of anthocyanidins was inversely associated with risk for CHD. None of the other flavonoid classes was significantly or suggestively associated with this disease.

Few previous investigations have focused on specific flavonoid classes, mainly because suitable databases were not generally available until recently. Huxley and Neil (2003) have carried out an exemplary meta-analysis of seven cohort studies in which, however, intake of total flavonoids, flavonols and other flavonoid classes was not always distinguishable; they reported a weak but statistically significant inverse association for flavonols. In one of the studies included in the meta-analysis of Huxley and Neil, Rimm et al (1996) found no association between intake of flavonols and flavones on one hand and incidence of CHD on the other, but they noted a protective effect of these flavonoid classes on subsequent coronary events among men who had previously been diagnosed with CHD. Arts et al (2001) reported that intake of catechins (an important compound of the flavan-3-ols class) mostly derived from wine and apples is significantly inversely associated with CHD risk. In the uncertain context created by previous investigations, our study contributes by pointing to flavan-3-ols, rather than flavonols, as the flavonoid class is most likely to convey protection against CHD.

Several biological mechanisms have been invoked to explain the apparent protection against CHD conveyed by flavonoids in general, and flavan-3-ols in particular. Like all flavonoids, flavan-3-ols inhibit lipid peroxidation and accumulation of oxidized low- density lipoproteins in macrophages, and play a beneficial role on endothelial function (Soleas et al, 1997; Giugliano, 2000). Flavan-3-ols have been reported to inhibit mast cell secretion, reduce mitochondrial injury in ischemia and have particularly strong antioxidant potency (Van Jaarsvelt et al, 1996; Middleton et al, 2000). Red wine, which is rich in catechin, an important flavan-3-ol, has been shown to inhibit ligand binding of the platelet-derived growth factor beta receptor, which plays a critical role in the pathogenesis of arteriosclerosis (Demrow et al, 1995; Rosenkraunz et al, 2002).

The present study has the strengths and weaknesses of hospital-based case–control studies of the nutritional etiology of disease (MacMahon and Trichopoulos, 1996). Response rates were high among both cases and controls and all interviews were conducted by trained health professionals in a medical environment with very little time pressure. However, the case–control study was not population based and CHD cases first manifested as sudden death could not be included. Moreover, the existence of a particular disease (CHD or other) may have generated some type of information bias, even though the questionnaire focused on the period preceding the onset of the symptoms for the present disease. However, known or suspected selection factors for hospitalization were accounted for in the analysis as were nutritional or non-nutritional factors with confounding potential. Moreover, there is no evidence that CHD cases manifested as sudden death have a nutritional etiology different from that of cases with other clinical manifestations and, notwithstanding a reasonable cause for concern, it has not been shown that information bias represents a major problem in case–control studies probing the etiologic role of diet in the causation of disease. In this context, it is reassuring that the results of earlier analyses concerning the risk profile of CHD in this investigation were in line with those reported from other studies (Kalandidi et al, 1992). A shortcoming of our study was the application of a USA flavonoid database on Greek food data, but this was unavoidable, since there is no complete flavonoid database for Greek foods. This investigation, however, did not focus on the absolute intake of flavonoids, but rather on the effect of relative differences between cases and controls which can be adequately captured in this study design.

In conclusion, we have found evidence that intake of flavan-3-ols, mostly from wine and tea, is inversely associated with risk of CHD. We cannot exclude the possibility that anthocyanidins also convey some protection. There was no evidence that flavanones, flavonols, flavones and isoflavones are associated with CHD.


  1. Arts IC, Jacobs DR, Harnack LJ, Gross M & Folsom AR (2001): Dietary catechins in relation to coronary heart disease death among postmenopausal women. Epidemiology 12, 668–675.

  2. Breslow NE & Day NE (1980): Statistical methods in cancer research. Volume I—The analysis of case–control studies. IARC Sci. Publ. 32, 5–338.

  3. Demrow HS, Slane PR & Folts JD (1995): Administration of wine and grape juice inhibits in vivo platelet activity and thrombosis in stenosed canine coronary arteries. Circulation 91, 1182–1188.

  4. Gnardellis C, Trichopoulou A, Katsouyanni K, Polychronopoulos E, Rimm EB & Trichopoulos D (1995): Reproducibility and validity of an extensive semi-quantitative food frequency questionnaire among Greek school teachers. Epidemiology 6, 74–77.

  5. Giugliano D (2000): Dietary antioxidants for cardiovascular prevention. Nutr. Metab. Cardiovasc. Dis. 10, 38–44.

  6. Gryglewski RJ, Korbut R, Robak J & Swies J (1987): On the mechanism of antithrombotic action of flavonoids. Biochem. Pharmacol. 36, 317–322.

  7. Huxley RR & Neil HA (2003): The relation between dietary flavonol intake and coronary heart disease mortality: a meta-analysis of prospective cohort studies. Eur. J. Clin. Nutr. 57, 904–908.

  8. Hertog MG, Kromhout D, Aravanis C, Blackburn H, Buzina R, Fidanza F, Giampaoli S, Jansen A, Menotti A, Nedeljkovic S, Pekkarinen M, Simic BS, Toshima H, Feskens EJM, Hollman PCH & Katan MB (1995): Flavonoid intake and long-term risk of coronary heart disease and cancer in the seven countries study. Arch. Intern. Med. 155 (4), 381–386.

  9. Hertog MG, Feskens EJ, Hollman PC, Katan MB & Kromhout D (1993): Dietary antioxidant flavonoids and risk of coronary heart disease: The Zutphen Elderly Study. Lancet 342 (8878), 1007–1011.

  10. Kalandidi A, Tzonou A, Toupadaki N, Lan S-J, Koutis C, Drogari P, Notara V, Hsieh C-C, Toutouzas P & Trichopoulos D (1992): A case–control study of coronary heart disease in Athens, Greece. Int. J. Epidemiol. 21, 1074–1080.

  11. Katsouyanni K, Skalkidis Y, Petridou E, Polychronopoulou-Trichopoulou A, Willett W & Trichopoulos D (1991): Diet and peripheral arterial occlusive disease: the role of poly-, mono-, and saturated fatty acids. Am. J. Epidemiol. 133, 24–31.

  12. Knekt P, Jarvinen R, Reunanen A & Maatela J (1996): Flavonoid intake and coronary mortality in Finland: a cohort study. BMJ 312, 478–481.

  13. Kromhout D, Bloemberg BP, Feskens EJ, Hertog MG, Menotti A & Blackburn H (1996): Alcohol, fish, fibre and antioxidant vitamins intake do not explain population differences in coronary heart disease mortality. Int. J. Epidemiol. 25, 753–759.

  14. MacMahon B & Trichopoulos D (1996): Epidemiology: Principles and Methods 2nd Ed, pp 229–299. Boston: Little Brown.

  15. Middleton Jr E, Kandaswami C & Theoharides TC (2000): The effects of plant flavonoids on mammalian cells: implications for inflammation, heart disease and cancer. Pharmacol. Rev. 52, 673–751.

  16. Peterson J & Dwyer J (1998): Flavonoids: Dietary occurrence and biochemical activity. Nutr. Res. 18, 1995–2018.

  17. Polychronopoulou-Trichopoulou A (1989): Composition of Greek Foods and Recipes. Athens: Parisianos.

  18. Puddey IB & Croft KD (1999): Alcohol, stroke and coronary heart disease. Are there antioxidants and pro-oxidants in alcoholic beverages that might influence the development of atherosclerotic cardiovascular disease? Neuroepidemiology 18, 292–302.

  19. Rimm EB, Katan MB, Ascherio A, Stampher MJ & Willett WC (1996): Relation between intake of flavonoids and risk for coronary heart disease in male health professionals. Ann. Intern. Med. 125, 384–389.

  20. Rosenkraunz S, Knirel D, Dietrich H, Flesch M, Erdmann E & Bohm M (2002): Inhibition of the PDGF receptor by red wine flavonoids provides a molecular explanation for the ‘French paradox’. FASEB J. 16, 1958–1960.

  21. Soleas GJ, Diamantides EP & Goldberg DM (1997): Wine as a biological fluid: history, production, and role in the disease prevention. J. Clin. Lab. Anal. 11, 287–313.

  22. Trichopoulou A (1992): Composition of Greek Foods and Dishes. Athens, Greece: Athens School of Public Health.

  23. Tzonou A, Kalandidi A, Trichopoulou A, Hsieh C-C, Toupadaki N, Willett W & Trichopoulos D (1993): Diet and coronary heart diseases: a case–control study in Athens, Greece. Epidemiology 4, 511–516.

  24. US Department of Agriculture (2003): USDA Database for the Flavonoid Content of Selected Foods. Beltsville, MD: USDA.

  25. US Department of Agriculture (2002): Iowa State University Database on the Isoflavone Content of Foods, Release 1.3—2002. Beltsville, MD: USDA.

  26. Van Jaarsvelt H, Kuyl JM, Schulenburg DH & Wiid NM (1996): Effect of flavonoids on the outcome of myocardial mitochondrial ischemia/reperfusion injury. Res. Commun. Mol. Pathol. Pharmacol. 91, 65–75.

  27. Yochum L, Kushi LH, Myer K & Folsom AR (1999): Dietary flavonoid intake and risk of cardiovascular disease in postmenopausal women. Am. J. Epidemiol. 149, 943–949.

Download references


This study was partially supported by the University of Athens and a grant to Harvard University by the Samourkas Foundation. The project was also funded in part with Federal funds from the US Department of Agriculture Research Service under contract number 58-1950-9-001. The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government. Partial support was also provided by the State of Florida, Department of Citrus.

Author information

Guarantor: D Trichopoulos.

Contributors: PL: lead investigator, supervisor of all activities and manuscript developer. ES: Statistical analysis. AL: Epidemiologist and contributor in manuscript drafting. AT: Statistical analysis. AK: Design of the original study and supervision of data collection. JP: Contribution in development and application of flavonoid content data. JD: Contribution in development and application of flavonoid content data. DT: Design of the original study; senior epidemiologist.

Correspondence to P Lagiou.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lagiou, P., Samoli, E., Lagiou, A. et al. Intake of specific flavonoid classes and coronary heart disease—a case–control study in Greece. Eur J Clin Nutr 58, 1643–1648 (2004).

Download citation


  • flavonoids
  • flavan-3-ols
  • coronary heart disease
  • diet

Further reading