Flavonoids and heart disease ’2

Flavonoids are secondary plant metabolites; plants use them to attract pollinators and repel predators, to color flowers and fruits, and for other functions. Subclasses include isoflavones or phytoestrogens, catechins, anthocyanidins, and flavonols. Flavonoids contain benzene rings with multiple hydroxyl groups and they are therefore known as polyphenols. In 1936, Rusznyak and Szent-Gyorgyi (1) proposed that intake of flavonoids decreased the capillary fragility seen in scurvy and named flavonoids vitamin P. Flavonoids were stripped of vitamin status around 1950, but interest in them has resurged because of possible links with cancer and cardiovascular disease. Epidemiologic studies have shown an inverse association between fruit and vegetable intake and cancer risk; plants may therefore contain anticarcinogenic compounds. Animal experiments made the flavonol quercetin a plausible candidate. The antioxidant properties of flavonols and catechins suggested that their intake might prevent atherosclerosis, and the low incidence of heart disease in France has been ascribed to a high intake of flavonoids from red wine. The hypothesis that flavonoids prevent heart disease received a boost when Hertog et al (2) reported an inverse association between the intake of the flavonol quercetin and coronary mortality in men in Zutphen, Netherlands. The major sources of quercetin and related flavonols turned out to be tea and onions rather than fruits or green vegetables. Great Britain, with its high tea intake, was therefore an obvious choice for further studies. In this issue, Hertog et al (3) report on the association of flavonol intake at baseline with morbidity and mortality over 14 y in 1900 men in Caerphilly, Wales, United Kingdom. There was no association of flavonol or tea intake with ischemic heart disease incidence. On the contrary, heart disease and total mortality rose with tea intake. The authors offer two explanations for this. One is that subjects used milk in their tea and that the proteins from milk form complexes with flavonoids that inhibit their absorption. Intake of black tea indeed raised plasma antioxidant capacity in volunteers whereas tea with milk did not (4). However, plasma concentrations of catechins and quercetin in volunteers given tea were the same whether the tea contained milk or not (K Van het Hof, J Weststrate, L Tijburg, personal communication, 1997; PCH Hollman, K Van het Hof, personal communication, 1997). Thus, lack of absorption is an insufficient explanation. The second explanation is confounding. The men with the highest intake of tea and flavonols tended to be manual workers; they smoked more, consumed less alcohol, and ate more fat (Table 1 in reference 3). Hertog et al corrected for these confounders, but confounders that have been measured with less than perfect precision cannot be completely eliminated by multivariate analysis. Measurements of diet, smoking, and alcohol intake were necessarily crude, and no data were available on physical activity or psychosocial factors. Therefore, the most plausible explanation for the higher mortality in tea drinkers is confounding by unmeasured or imperfectly measured coronary risk factors that clustered with tea intake. Where does this leave flavonoids? The food table is the central paradigm of nutrition science and health effects of foods should be explained in terms of defined food components. However, different sources of flavonols show different associations with disease: onion intake is associated with decreased cancer risk (5) but tea intake is not (6), and in the present study, the incidence of heart disease mortality went up with tea but down with onion consumption (Table 3 in reference 3). Such differences can be due to confounding or to other substances present in these foods, but the same substance could also be present in different forms. Metabolic studies show that quercetin compounds from onions have a higher bioavailability than do those from tea. Such studies also suggest that in vivo antioxidant activity of dietary quercetin remains feasible; quercetin is absorbed well from foods and plasma concentrations 0.5 mol/L can be reached (7). Addition of flavonols to blood plasma in vitro reduces the oxidizability of low-density lipoprotein (LDL), although effects of administration in vivo are less clear. Flavonols probably do not accumulate inside LDL particles, which explains why consumption of tea by volunteers has only minimal effects on the oxidizability of LDL isolated from plasma. Addition of red wine to blood plasma reduced the oxidizability of LDL, but in vivo data are conflicting (8, 9). The epidemiologic findings on flavonols and cardiovascular disease are summarized in Table 1. Inverse associations were seen in Zutphen (2, 10) and in elderly persons in Rotterdam, Netherlands, where high tea consumption was inversely associated with noninvasively assessed measures of atherosclerosis (JCM Witteman, personal communication, 1997). A modest protective effect was also seen in Finland ( I I ). However, in a large cohort of US health professionals there was no associa-