The red wine hypothesis: from concepts to protective signalling molecules

Transcription

1 European Heart Journal (2007) 28, doi: /eurheartj/ehm149 Review The red wine hypothesis: from concepts to protective signalling molecules Lionel H. Opie* and Sandrine Lecour Department of Medicine, Faculty of Health Sciences, Hatter Cardiovascular Research Institute, University of Cape Town, Private Bag 3, 7935 Observatory, South Africa Received 30 January 2007; revised 28 March 2007; accepted 5 April 2007; online publish-ahead-of-print 7 June 2007 KEYWORDS Wine; Alcohol; Resveratrol; Molecular mechanisms; Sirtuin Introduction The harmful effects of risk factors for cardiovascular disease as often linear, as in the case of cigarette smoking and blood lipid abnormalities. 1 For alcohol, however, the situation is different, with data for moderate-dose cardiovascular benefits 2 and high-dose harm. This article will first review the increasing evidence for this dose-related protection, then evaluate the proposal that wine is more protective than other alcoholic drinks, and thereafter examine the hypothesis that red wine is more protective than white. Important arguments favouring this hypothesis derive from the role of specific protective chemical polyphenols found both in red wine and de-alcoholized red wine. Putative molecular mechanisms involve newly described signals and paths that may help to explain the proposed protective qualities of moderate amounts of red wine. The J-shaped mortality curve: still controversial? The major and consistent finding is that alcohol consumption, from whatever source, appears to have a J-shaped curve, whereby a modest intake is beneficial and either no intake or an increased intake is harmful. This relation has * Corresponding author. Tel: þ ; fax: þ address: lionel.opie@uct.ac.za We review evidence for and against the red wine hypothesis, whereby red wine is more likely to confer cardiovascular benefits than white. As background, there is a strong epidemiological and mechanistic evidence for J-shaped relation between alcohol intake and total mortality. However, epidemiological data favouring a specific benefit of red over white wine are not strong and the French paradox could at least in part be explained by confounding factors. More convincing evidence is that human studies with de-alcoholized red but not white wine show short-term cardiovascular benefits. The specific components of the de-alcoholized wine that are active on cardiovascular endpoints, are the polyphenols found in red wine, especially resveratrol. The effects of resveratrol on isolated tissues or organs are well-described including molecular mechanisms leading to decreased arterial damage, decreased activity of angiotensin-ii, increased nitric oxide, and decreased platelet aggregation. Anti-ischaemic effects include stimulation of prosurvival paths, decreased LDL-oxidation, atheroma, and on the ischaemic-beneficial metabolic changes. Most recently, the agonist effect of resveratrol on the anti-senescence factor sirtuin has lessened cell death in myocytes from failing hearts. Mechanistic feasibility strengthens the case for prospective therapeutic trials of alcohol vs. red wine vs. resveratrol, for example in those with heart failure. been established in a series of studies starting with the Framingham study, 3 followed by the British Doctors Study, 4 the Cancer Prevention study on about persons, 5 the Nurses Health Study, 6 the Physicians Health Study, 7 the British Heart Study, 8 the Multiple Risk Factor Intervention Trial, 9 and the Copenhagen City Heart Trial, 4,5,7,10,11 However, the issue of confounding factors has bedevilled these findings as shown by two opposing views. 12,13 Of particular relevance is a meta-analysis of mortality in 34 outcome studies on more than 1 million subjects. 14 Nearly deaths were analysed to show a clear J-relationship with decreased total mortality associated with two to four daily drinks for men and one to two for women vs. abstinence or higher drinking rates. 14 Of note in relation to the controversy regarding the significance of confounding factors, adjustment for social factors, and dietary markers made the J-shaped dip shallower and shorter, yet still evident. 14 The J-shaped mortality curve might in part be explained by less myocardial infarction 6,15 and in part by less incident heart failure, 16 but not by decreased hypertension. 17 Regarding mechanistic studies, these are many and difficult to explain by the confounding hypothesis. One established mechanism of the cardiovascular benefit is through a consistently reported increase in plasma high-density lipoprotein (HDL)-cholesterol levels, thought to be the result of increased hepatic synthesis. The definitive study shows that & The European Society of Cardiology All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 1684 L.H. Opie and S. Lecour a genetic variation in hepatic alcohol dehydrogenase, which slows the rate of ethanol metabolism, is associated with higher plasma HDL-levels and lower rates of myocardial infarction. 18 This was a prospective study with a high rate of completeness of follow-up data, and part of the carefully monitored Physicians Health and Nurses Health studies. The benefits found in moderate drinkers were profoundly altered by the presence or absence of the ADH3 genotype, which is contrary to the confounding hypothesis. About half of the benefit of moderate drinking can be related to increased HDL-levels. 19 Other possible protective mechanisms include anti-platelet, 20 anti-coagulatory, 19,21 improved glucose control, 19 and anti-inflammatory effects as shown in the MONICA study. 22 The latter were more evident in moderate drinkers than in abstainers or heavy drinkers. When given acutely for 8 weeks, ethanol reduced triglyceride and insulin levels. 23 Of these effects, those on HDL, fibrinogen, and insulin sensitivity have been established in randomized prospective studies, 19 which are especially difficult to explain by the confounding hypothesis. Alcohol also has detrimental high-dose ( g/kg body weight) pro-oxidant effects in humans, 24 which together with liver damage may contribute to the right-hand upward slope of the J-curve. Is no alcohol really harmful? In a large multinational study, no alcohol intake was a harmful risk factor for myocardial infarction, especially in women. 2 The first and most obvious explanation would be the lack of the benefit of moderate alcohol intake on HDL-cholesterol and, as will be discussed, the absence of the postulated non-ethanolic benefits of red wine. Alcohol has beneficial effects on insulin and triglyceride levels, 23 and abstinence is a risk factor for new type 2 diabetes. 25 Alcohol gives experimental cardioprotection against ischaemic-reperfusion injury in isolated hearts. 26,27 There is, however, another explanation for the harm of no alcohol intake, which is the recurring problem of confounding factors which are difficult totally to exclude in epidemiological studies. Nonetheless, confounding was discounted in the study on British doctors, 12 and also in a very large and up-to-date meta-analysis. 14 Yet, Americans who take no alcohol have a different and generally less cardioprotective lifestyle from moderate drinkers. 13,28 Regular alcohol intake, generally as wine and with meals, is part of the Mediterranean diet 29 and the beneficial Mediterranean lifestyle. In France, moderate alcohol and wine drinkers smoke less and have a higher socio-economic status than abstainers. 30 However, confounding factors do not mean that the J-shaped protection is lost. Thus men with a healthy lifestyle not only have a lower rate of myocardial infarction but achieve increased protection by moderate alcohol intake. 15 Those who take no alcohol might be disadvantaging themselves, not only for lack of alcohol, but also for lack of a beneficial lifestyle. Alcohol and blood pressure Here there is no J-shaped curve. A common opinion is that wine intake reduces blood pressure (BP) because wine is a relaxant. This only holds for an early effect within hours. 31 In long term, the opposite holds. A series of recent articles show that in normal persons, alcohol but not de-alcoholized red wine elevates BP by about 2 3 mmhg Conversely, alcohol reduction in those men who are fairly heavy drinkers (3 drinks per day) reduces BP by a similar degree. 34 Even such small differences may have relatively important epidemiological implications. In hypertensive men, the alcohol-induced rise is even greater. 35 These adverse BP data neither exclude alternative benefits such as decreased aortic stiffness, 36 nor changes in the pattern of arterial wave deflections with decreased central systolic BP. 37 Overall, for any given level of BP, a lower all-cause mortality is associated with a moderate wine intake (,3 glasses daily). 38 This being an observational study, confounding factors are again difficult to exclude. Patterns of drinking Its regular intake is protective, 19 especially in males. 39 Binge-drinking (more than three drinks per hour of any variety of alcohol) is associated with increased mortality after myocardial infarction. 19 Binge-drinking is one of several lifestyle factors that may counter the J-shaped curve. 40 Regarding the BP, there is little or no pressure rise if alcohol is taken with meals. 41 Fitting in with the effect of wine in relation to meals, basic research suggests that red wine could protect against some effects of post-prandial lipaemia. 42 While some other indirect evidence also supports the concept of the benefits of wine with meals, 29 there was no relation to meals in American men who predominantly drink liquor and beer. 43 Wine vs. other alcoholic drinks As already mentioned, many studies have related alcohol to the J-shaped beneficial curve. 14 An important claim is that wine may have benefits that, for example, beer or spirits lack. For example, Gronbaek found all-cause mortality reduced in wine but not in non-wine drinkers in his study on Danes over 10 years. 10 In a meta-analysis on persons, consumption of wine but not that of beer was protective. 11 Klatsky followed-up adults in Northern California for up to 20 years. 44 The J-shaped curve was once again confirmed. Drinkers of any type of wine had a lower mortality risk than did beer or liquor drinkers, in part because wine drinkers were the lighter drinkers. Renaud s group studied healthy French men, also to show that moderate wine but not other alcohol types reduced all-cause mortality over years. 45 These impressive data from three entirely different countries must be balanced against other studies that differ and the recurrent problem of confounding factors. Thus, the type of alcohol made no difference to its protective effect on myocardial infarction in the participants in the US male health professions study over 12 years of follow-up. 43 Moderate consumption of both wine and beer equally decreased inflammatory markers such as C-reactive protein. 22 The real problem is that there are very few long-term prospective studies comparing the possible beneficial effects of the types of alcohol on cardiovascular outcomes or total mortality while fully allowing for confounding factors. According to such factors, wine drinkers have a more healthy diet than beer drinkers with increased intake of fruit, vegetables, and fibre. 28,30,46 In the British

3 The red wine hypothesis 1685 Regional Heart study, 7735 men drawn at random were followed-up for an average of 16.8 years. 47 Wine drinkers had a lower age-adjusted risk of coronary heart disease and all-cause mortality than did beer and spirits drinkers, but the wine drinkers also had a better life style (for example, much less smoking). Thus the wine vs. beer data suggest, but do not conclusively give the benefit to wine. The red wine hypothesis It is often thought that epidemiological data from France strongly suggests a protective effect of red wine despite a high-fat diet (the French paradox ). This idea goes back to St Leger et al. 48 who in 1979 found an inverse relation between coronary heart disease mortality and wine consumption, with France having the lowest mortality. The jump to the specific benefit of red wine was made in 1993 when Frankel et al. showed that red wine diluted 1000-fold and contained 10 mmol/l of total phenols that inhibited the oxidation of LDL more than a-tocopherol, an established antioxidant. 49 The antioxidant mechanisms involved are scavenging of peroxyl radicals as well as chelation of redox metal ions such a copper. 50 Indirect evidence favouring the French paradox and the red wine hypothesis is that the French habitually drink wine with their meals (which are often fatty) and this wine is most often red. Furthermore, Alsace, a white-wine drinking region of France, has a much higher mortality (about 50% higher) than red wine-drinking Mediterranean areas, 51 despite having a lower mean serum cholesterol level. 20 However, such comparisons need to take into account the many confounding inter-regional differences in diet and life style. 30 Furthermore, there are several other explanations for the French paradox, including the time lag hypothesis which states that the French diet had low fat in the past and that it takes about 30 years for any dietary pattern to manifest itself in mortality data. 52 A PubMed search could uncover no good epidemiological study firmly to establish any clinical outcome benefit of red vs. white wine drinking in Europe. Most large-scale epidemiological data coming from North America suggest that there is no difference between white and red wine, 43,44,53 thus concluding that alcohol intake rather than wine colour predicts the eventual cardiovascular outcome. Again there may be confounding factors. The pattern of wine drinking could be different, with less wine with meals and more prior to meals in America than in Europe. Furthermore, American and French diets are typically very different, for example with much more wine used in cooking in France. Marinating meat in red wine substantially decreases formation of potentially toxic heterocyclic amines 54 and of lipid peroxides. 55 French vs. German reds: effects on nitric oxide French red wine is better than German reds in the induction of human endothelial nitric oxide synthase (enos) in human umbilical vein endothelial cells. 56 Presumably, this is a class effect although only six French and three German wines were studied. In view of the similar effects on nitric oxide production by two French reds, the one barrel-fermented and the other not, 56 it is very unlikely that better French than German oak can explain these differences. While the human inclination is to add such basic observations to the known better aroma, palate, and after-taste of French vs. German reds, we still await a large comparative epidemiological study proving the proposed superior cardioprotective effects of French red wines. Red vs. white wine or vodka: comparative studies In normal human volunteers, drinking an unspecified amount of red wine but not white or vodka increased coronary flowvelocity reserve. 57 Drinking red wine raised HDL-cholesterol and plasma antioxidant status more than did an equivalent dose of white wine when given to healthy volunteers. 58 Furthermore, in vascular smooth muscle cells, pre-incubation with red wine but not white inhibited the platelet-derived growth factor beta-receptor (PDGF) that is crucial in the development of atherosclerosis. 59 Conversely, incubating white wine with grape skins (that contain the polyphenols found in red wine) and alcohol endowed white wine with a similar power of inhibition of LDL-oxidation to that possessed by red wine. 60 In human mononuclear cells incubated with VLDL (very low-density lipoproteins), drinking moderate doses of red wine but not vodka inhibited the activation of nuclear factor (NF)-kappa-B 61 during post-prandial lipaemia, thus providing an anti-inflammatory mechanism. 42 This effect was simulated by antioxidants contained in red wine. These comparative studies support the red wine hypothesis. Red wine: protective effects beyond alcohol The conclusive studies showing that red wine has qualities beyond alcohol are those on de-alcoholized red wine, which has cardiovascular protective effects in short-term studies on humans. 37,62 In patients with coronary artery disease, 250 ml of de-alcoholized Greek red wine decreased arterial stiffness and improved the augmentation index, as derived from arterial wave reflection patterns. 37 A similar dose of de-alcoholized red wine decreased adverse post-smoking arterial wave reflections and lessened the rise in systolic BP. 62 Brachial artery flow-mediated vasodilation was improved by ml of de-alcoholized red wine. 63,64 Ingestion of purple grape juice in patients with coronary artery disease (about 640 ml/day for 14 days) similarly improved brachial flow. 65 The purple grape juice also reduced LDL-cholesterol susceptibility to oxidation, with the benefits being attributed to flavonoids. Alcohol-free red but not white wine greatly increased plasma antioxidant capacity 50 min after ingestion. 66 Red wine incubated with human plasma strongly inhibited LDL-oxidation and uptake of LDL by macrophages, and the active components were catechins and flavonols, whereas ethanol had no effect. 50 All these human studies are among the most powerful arguing for the benefit of red (but not white) wine beyond alcohol. Supporting data comes from dogs with stenosed coronary arteries, in whom administration of French blended red wine, apparently verging on vintage in age, eliminated cyclic flow reductions caused by periodic acute plateletmediated thrombus formation. 67 This markedly beneficial effect was also achieved by purple grape juice, but about 2.5 times greater volume was required than in the case of red wine. White wine was not protective, so that the explanation must again lie in the effects of red wine beyond alcohol. Experimentally, the benefit of the red wine lies in the polyphenols, found in de-alcoholized red wine extracts

4 1686 L.H. Opie and S. Lecour which improve endothelial nitric oxide release, 68 and revert the prothrombotic effects of a cholesterol-rich diet. 69 De-alcoholized red but not white wine decreased atherosclerosis in apolipoprotein E gene-deficient mice. 70 However, there is no strict clinical proof that these beyond alcohol effects operate in red wine drinkers to give clinically significant cardiovascular protection. Rather, there is only indirect observational evidence in a case control study that links high dietary anthocyanidin intake to reduce the risk for acute myocardial infarction independent of alcohol intake. 71 Specific protective chemicals in red wine: molecular effects Polyphenol extracts Red wine polyphenol extracts have experimental cardioprotective properties, 72 and may counter one of the mechanisms underlying atherosclerosis, namely thrombin-induced matrix invasion of vascular smooth muscle cells. 73 An extract of Cabernet Sauvignon suppressed endothelin-1 synthesis and release in bovine aortic endothelial cells. 74 Another polyphenol extract protected against angiotensin- II-induced hypertension in rats by blunting endothelial dysfunction and promoting formation of nitric oxide. 75 Yet, another polyphenol extract increased enos promoter and nitric oxide release in human umbilical endothelial cells. 68 Of the possible polyphenol components tested, the activity only resided in resveratrol which is a stilbene derivative found in grape skins, and hence found much more in red than white wine. Resveratrol also was an active component of a red wine polyphenol extract that transcriptionally inhibited the expression of endothelial vascular cell adhesion molecule, VCAM Besides the polyphenols currently attracting most attention are resveratrol and the procyanidins (the latter are flavan-3-ols and hence part of the flavonoid family). Resveratrol Antioxidant effect Dosing animals with resveratrol gives a variety of cardioprotective results (Table 1) of which the strong antioxidant properties are perhaps the best known, 26 and could explain the inhibition of LDL oxidation. 49 The antioxidant effect on LDL is also linked to increased expression of the paraoxonase-1 gene. 77 Other antioxidant effects are widespread, including cardioprotection where it offsets the adverse pro-oxidant qualities of alcohol. 26 Furthermore, resveratrol inhibits the inflammatory response by suppressing prostaglandin biosynthesis 78 and by downregulating the gene expression of intercellular adhesion molecule-1 (ICAM-1) and NF-k-B, the latter taken as an indirect biomarker of oxidative stress. 79 The NF-k-B signalling pathway can be evoked by oxidative stress and is in turn inhibited by resveratrol 80 (Figure 1). Not surprisingly, TNF-a-induced activation of NF-k-B is attenuated by resveratrol and this sequence is in part mediated by an antioxidant effect, 81 which in turn reduces platelet activation. 82 This sequence can also explain the resveratrol-induced inhibition of oxidative stress in platelets induced by TNF-a. 83 Resveratrol inhibits angiotensin-ii (A-II)-induced cardiomyocyte hypertrophy, which is clearly linked to its antioxidant effects, because it inhibited the A-II-induced production of reactive oxygen species (ROS). 84 As the cellular redox state influences so many reactions, and is sensitive to both free radicals and to resveratrol, 80 it is likely but usually assumed rather than directly proven that the antioxidant effect of resveratrol could play a major role in at least some of the other protective mechanisms now to be discussed. Proposed vascular protective mechanisms Resveratrol acts through multiple mechanisms (Figure 1), including the antioxidant effects on platelet formation already discussed. Resveratrol vasodilates isolated human small arteries, 61 in response to oestrogen-receptor stimulation with formation of nitric oxide. The vasodilation is absent in arteries from patients with coronary artery disease. 61,85 The mechanism involves increased formation and release of nitric oxide associated with increased enos activity. 68,86 Although there is no direct evidence linking such increased enos activity to platelets, there is a wellknown effect of nitric oxide in inhibiting the activation of platelets. Links between resveratrol and reduced platelet aggregation 87 could also be mediated by TNF-a. 88 Resveratrol at a concentration of only 0.1 mmol/l increases the activity of NF-k-B, which lies on the TNF-a signalling pathway, in human endothelial cells after overnight incubation, thereby providing the mechanism for a possible anti-inflammatory and anti-platelet-effect. 89 Increased post-infarct neovascularization is a long-term benefit. 90 Resveratrol interferes with angiotensin-ii and epidermal growth factor signalling in vascular smooth muscle cells, which may be a long-term mechanism for inhibition of atherosclerosis. 91 Furthermore, inhibition of angiotensin-ii and increased synthesis of nitric oxide could, in part, explain renoprotective effects 92 and BP reduction in some renal models. 93 Cardiac fibroblasts Resveratrol (5 25 mmol/l) pretreatment inhibited growth paths stimulated by angiotensin-ii, epidermal growth factor, or transforming growth factor-b, which are essential in fibroblast proliferation and differentiation. 94 Thus resveratrol is potentially anti-fibrotic. Metabolic effects of resveratrol Resveratrol in high doses shifts the physiology of middle-aged obese mice, with features resembling the metabolic syndrome in humans, towards those of non-obese mice on a standard diet. 95 The metabolic changes included increased insulin sensitivity, decreased circulating free fatty acids, decreased insulin-like growth factor (IGF-1), increased activity of the energy-sensing enzyme, AMP-activated protein kinase (AMPK), increased activity of peroxisome proliferatoractivated receptor-gamma coactivator-alpha (PGC-1a), and increased mitochondrial number (Figure 2). These diverse effects are mediated by sirtuin (see next paragraph). Whether similar trends can be achieved by the very much lower resveratrol dose associated with red wine drinking is a subject for future research. Pharmacological preconditioning by resveratrol In several studies, resveratrol decreases ischaemic-reperfusion damage (Table 1). A burst of short perfusion by resveratrol or pretreatment before the onset of ischaemic-reperfusion injury stimulates a plethora of preconditioning-like survival paths

5 The red wine hypothesis 1687 Table 1 Concentrations of resveratrol that have relevant cardiovascular effects on isolated tissues or organs Tissue Concentration Effect Reference Human coronary artery endothelial mmol/l Inhibits TNF-a-induced NF-k-B activation and 81 cells mg/l inflammatory markers Human umbilical vein endothelial 0.1 mmol/l Increased the inhibitor of TNF-a (NF-k-B) 89 cells Human umbilical vein endothelial 1 10 mmol/l Increased enos promoter and nitric oxide release 68 cells Human umbilical vein endothelial mmol/l Increased enos promoter activity and enos protein 86 cells levels Human umbilical vein endothelial mmol/l Inhibition of tissue factor activity 132 cells Human endothelial progenitor cells 1.0 mmol/l Increased activity vs. decrease with 50 mm 96 Human hepatoma cells 5 10 mmol/l Increased expression of paraoxonase-1 gene 77 Human platelets, ASA-resistant 10 mmol/l Reduced platelet aggregation 87 Human small arteries, subcutaneous mmol/l Dilation but not in those with coronary heart disease 61 Rat neonatal ventricular myocytes mmol/l Inhibited A-II-induced phosphorylation of ERK and 84 A-II-induced increase of ROS levels Rat aortic smooth muscle cells mmol/l Inhibition of A-II and epidermal growth factor-mediated 91 growth signalling path Rat cardiac fibroblasts 5 25 mmol/l Inhibition of signalling paths mediated by A-II, epidermal 94 and transforming growth factors Ischaemic-reperfused rat heart 10 mmol/l Expression of anti-apoptotic Bcl-2 99 Ischaemic-reperfused rat heart 10 mmol/l Preconditioning of heart via adenosine-3 receptor 99 and survival paths Ischaemic-reperfused rat heart, Gavage 2.5 mg/kg Less infarct size and apoptosis via survival paths 97 after pretreatment Ischaemic-reperfused rat heart, 18.5 mmol/l in blood after Less post-reperfusion depressed function 133 after pretreatment i.p injection Ischaemic-reperfused rat heart 10 mmol/l Decreased infarct size 26 Ischaemic-reperfused rat heart 10 mmol/l Less ischaemic-reperfusion injury 134 Myocytes from failing mice hearts 0.5 mmol/l mg/l Inhibits falling levels of sirtuin (inhibitor of cell death) 104 Note the wide range of concentrations used of resveratrol, always in trans form (molecular weight ¼ 228 g/mol); for relation to circulating concentrations likely to be found after red wine intake in humans, see text. TNF, tumour necrosis factor; NF, nuclear factor; A-II, angiotensin-ii; ERK, extracellular signalregulated kinase; ROS, reactive oxygen species. Figure 1 Proposed multiple vascular protective mechanisms of resveratrol. ROS, reactive oxygen species; NF-k-B, nuclear factor kappa B; enos, endothelial nitric oxide synthase; TNFa, tumour necrosis factor alpha; Ang-II, angiotensin II; VSMC, vascular smooth muscle cells. involving signalling molecules such phosphatidyl-inositol-3- phosphate (P-I-3)-kinase-Akt (Figure 2). Other preconditioningassociated proteins include mitogen-activated protein (MAP) kinase kinase, 96 and p38-map kinase-beta. 97 These signals are, at least in part, mediated by the adenosine 3-receptor, 98,99 and involve nitric oxide. 98,100 An important prosurvival signal is activation of the antiapoptotic enzyme Bcl The overlap of some of these paths (P-I-3-kinase-Akt) with those stimulated by insulin, may explain the insulin-like effect of resveratrol in diabetic rats. 101 Effects on sirtuin (anti-senescent) proteins These effects of resveratrol are the focus of current research (Figure 3). A high resveratrol diet opposes the effects of a high fat diet in mice, increasing insulin sensitivity, decreasing myocardial and aortic damage, and prolonging lifespan. 95 Mechanistically, at least some of these effects can be related to the recent discovery that resveratrol is an agonist of the sirtuins (also called SIRT proteins, Silent Information Regulator Two) which belong to the histone deacetylase family. 102 By deacetylation they inactivate the histone senescence factor p53. Conversely, overexpression of sirtuins limits premature cellular senescence. 103 Activation of sirtuins by resveratrol can experimentally prolong the life of myocytes from failing hearts and decrease myocyte death induced by angiotensin-ii. 104 Furthermore, resveratrol protects against Alzheimer-like neurodegeneration by increasing sirtuin activity and, thereby, inhibiting adverse NF-k-B signalling, 105 thus, experimentally protecting both heart and brain via sirtuins in high doses. 104 Resveratrol: male vs. female differences Resveratrol stimulates the oestrogen E-beta receptors in vascular endothelial cells, 61 and functions as an agonist for

6 1688 L.H. Opie and S. Lecour It would be intriguing to imagine that women could offset this gender disadvantage of alcohol by drinking red rather than white wine, but there are no supporting data. Figure 2 Proposed anti-ischaemic protective mechanisms of resveratrol. Those myocardial signalling paths involved in resveratrol pharmacological preconditioning are shown on the left, and those giving metabolic protection on the right. Note multiple antioxidant effects. P-I-3 kinase, phosphatidyl-inositol-3 phosphate-kinase; Akt, protein kinase B; BcL-2, mitochondrial anti-apoptotic protein; NF-k-B, nuclear factor kappa-b; LDL, low-dentistry lipoprotein; PGC-1a, peroxisome proliferator-activated receptorgamma coactivator-alpha; IGF-1, insulin-like growth factor; AMPK, AMPactivated protein kinase; mito, mitochondria. Outlines of heart and atheroma modified from Opie LH Heart Physiology, from Cell to Circulation, Lippincott, Williams and Wilkins, Philadelphia, 2004, with permission. Figure 3 Proposed effect of resveratrol in countering heart failure by acting as an agonist to sirtuin, considered as a longevity factor. Hypothetically, heart failure induces overactivity of poly(adp-ribose) polymerase (PARP), a multifunctional DNA-bound nuclear enzyme, that inhibits NAD-dependent paths thereby decreasing the activity of sirtuin. The proposal is that resveratrol, by increasing the activity of sirtuin, deacetylates and inhibits the pro-apoptotic effector p53 (a transcription factor). The result is that resveratrol protects cells from failing hearts against PARP-mediated cell death. For concepts and data, see Pillai et al. 104 and Smith. 131 NAD, nicotinamide adenine dinucleotide. Figure of heart from Drugs for the Heart, 6th edn., (Eds) Opie LH and Gersh B, Elsevier Saunders, Philadelphia, 2005, with permission. oestrogen receptor-mediated transcriptions albeit at a much higher concentration than oestradiol. 85 In general, women are more susceptible to the harmful effects of alcohol, with limits of two drinks/day vs. four in men, 14 because lower levels of the enzyme alcohol dehydrogenase give higher blood levels than with comparable doses in males. Resveratrol pharmacokinetics Are the protective concentrations used in such studies within the range that could be expected after drinking red wine? (Tables 1 and 2). Resveratrol itself exists in two forms, cis- and trans-forms that are in equilibrium and have similar biological activities, so that the total resveratrol concentration is relevant. However, it varies substantially between wines (Table 2). Resveratrol is only one of several stilbenes in red wine although the best studied. In Brazilian red wine, the total stilbene consumption from drinking half a bottle of some types of red wine per day is about 11 mg, 106 which if all were absorbed could give a peak total stilbene concentration of about 10 mmol/l, or about 4 mmol/l of resveratrol. This potential peak value compares well with the lowest in vitro resveratrol concentration having an in vitro biological effect ( mmol/l; Table 1), and raises the theoretical possibility that therapeutic resveratrol levels could be reached by drinking highresveratrol red wines. Proof of this hypothesis requires studies in humans of red wine dosage and related resveratrol blood levels. In one study, after 25 mg of resveratrol, as would be found in 2.5 L of some red wines (Table 2), peak total (free and conjugated) resveratrol was about 450 mg/l or about 2 mmol/l, fold more than the lowest resveratrol concentrations experimentally active (Table 1). Thus half a bottle of such wine should still give enough resveratrol to be active. However, the calculation assumes that resveratrol conjugates are active, which has not yet been studied. The peak free circulating unconjugated resveratrol was just about 5 nmol/l, 107 which is 20-fold below the lowest effective concentration reported. In another study, a dose of 1 g as in 10 L of some red wines, gave higher peak free Table 2 Wine varietal Concentrations of resveratrol in various wine varietals Country of origin Resveratrol concentration Pinot noir Australian 13.4 mg/l¼ mmol/l Pinot noir Californian 5.5 mg/l 135 Pinot noir Burgundy 4.4 mg/l 135 Pinot noir Spanish 5.1 mg/l 136 Cabernet Australian 1.7 mg/l 135 Sauvignon Cabernet California 0.9 mg/l 135 Sauvignon Rhone valley France 3.6 mg/l 135 reds Bordeaux reds France 3.9 mg/l 135 Red, varietal Switzerland mg/l 135 not stated Red, varietal Brazil 18.0 mg/l 135 not stated Red grape juice 0.5 mg/l 110 Reference Concentration of resveratrol in red wines for comparison with concentrations used in various isolated tissue or organ experiments shown in Table 1.

7 The red wine hypothesis 1689 resveratrol levels of 1 2 mmol/l. 108 Thus, 1 L of such red wine would give levels of nmol/l, just within the bottom end of the therapeutic range (Table 2). Oral administration of resveratrol to rats in doses similar to a reasonable wine intake gives plasma resveratrol levels of 0.1 mmol/l, 109 about one-third of the lowest concentration that can relax isolated human small arteries, 61 and one-fifth of the lowest concentration that protects mycocytes from failing hearts, 104 but about twice the lowest level that interferes with signalling by angiotensin-ii. 91 There are two further relevant issues. First, the plasma concentrations of the conjugated glucoside form, piceid, can be about six times higher than that of free resveratrol. 110 The biological activity of these conjugated compounds is not known. Secondly, with prolonged daily dosing of rats with an amount of resveratrol found in 250 ml of some red wines, there is substantial accumulation in the liver, and further metabolic fates are unknown. 109 Overall, resveratrol concentrations reached by drinking within the safe limits of some red wines might be just within or below the therapeutic range, judged by isolated tissue experiments. Procyanidins Resveratrol is not the only protective agent in red wine. Red wine contains a large number of other potentially active flavonoids, especially the flavan-3-ols which include procynanidins and proanthocyanidins (Table 3). Proanthocyanidin, like resveratrol, is a strong antioxidant and both equally reduce experimental myocardial infarct size and apoptosis in isolated rat hearts. 26 The mechanisms include a preconditioning-like effect, shown by induction of heat shock protein (HSP) 70. Taking inhibition of endothelin-1 production by aortic endothelial cells as another endpoint, oligomeric procyanidins inhibit ET-1 synthesis at concentrations similar to those in some red wines, with some correlation between the procyanidin activity of the wine varietal and the longevity. 111 Oligomeric procyanidins are also inhibitors of ACE activity, in concentrations similar to those found in red wines. 112 The antioxidant effect of red wine on LDL oxidation is mediated chiefly by catechins but also by the Table 3 Potentially active polyphenols in wine 125 Polyphenols have multiple aromatic rings possessing hydroxyl groups A. Flavonoids Anthocyanins: cyanidin, delphindin (both as glycosides) Flavonols: quercetin, as glycoside Flavan-3-ols (not found as glycosides): Monomers: catechins Oligomers: procyanidins, 111 Polymers: proanthocyanidins (oligomers and polymers contain catechins and epicatechin) B. Non-flavonoids Resveratrol Peceid, conjugated resveratrol glycoside Gallic acid Others The italicized polyphenols are those most clearly linked to beneficial changes in cardiovascular experiments. flavonol and polymeric anthocyanidins. 50 However, there are few strict studies relating post-ingestion blood levels found in humans to cardioprotective concentrations in isolated tissues. In this regard, flavonoids of the catechin family inhibited the PDGF beta-receptor when used in concentrations similar to those found in the blood after drinking red wine. 59 Other active ingredients of red wine Of the many other polyphenolic components of red wine, the less well-studied flavonoids include quercetin and catechin, 75 which also have properties of potential cardiovascular importance. In a dietary study in over persons for 1 year, those with a higher flavonoid intake, particularly quercetin, had less coronary heart disease. 113 Both quercetin and catechin given orally can retard progression of atherosclerosis in apolipoprotein E-deficient mice, acting by reduction of the susceptibility of LDL to oxidation and aggregation. 114 In humans, plasma catechin rises to about 75 nmol/l soon after the equivalent of one to two glasses of red wine. 115 These levels contrast with those experimentally protective, such as 50 mmol/l. 114 De-alcoholized red wine decreased atherosclerosis in apolipoprotein E genedeficient mice, acting by phenolic acids not listed in Table The mechanism involved decreased LDL oxidation associated with decreased alpha-tocopheroxyl radicals, the oxidation product of vitamin E. Thus protective effects of vitamin E could be facilitated, overcoming the otherwise adverse excess pro-oxidant effects. In mice, a daily dose of Cabernet Sauvignon similar in dose to half a bottle of wine per day for humans, and with a long list of potentially active polyphenols, was neuroprotective against experimental Alzheimer s disease. 116 Multiple mechanisms of benefit of various ingredients seem probable. What is still missing are human data on the pharmacokinetics and bioavailability of the polyphenols found in red wine. 117 Nonetheless of note, at least one laboratory effect of polyphenols, namely inhibition of thrombin-induced matrix invasion of vascular smooth muscle cells, can be found in concentrations similar to those achieved after moderate red wine ingestion. 73 Proposed protective components in white wine and beer White wine may have different protective qualities Red and white wine have equal effects on fibrinolytic factors in a short-term cross-over study. 118 Equal effects on collagen-induced platelet aggregation are also reported. 119 In addition, white wines may contain anti-inflammatory protective substances that lessen cytokine release, as also found in extra virgin olive oil. 120 Thereby, the inflammatory reaction that promotes atherosclerosis could be lessened. De-alcoholized beer may also give some protection Some epidemiological studies have not shown less protection by beer than wine drinking. 22 De-alcoholized beer inhibits platelet and thrombogenic activity in young adults. 121 However, the dose was high (3 L) so that if translated into the amount of alcohol in standard beer, such intake would be harmful.

8 1690 L.H. Opie and S. Lecour Overall perspective There are several sequential proposals leading from observational population studies. Thus (i) only modest alcohol intake gives a J-shaped pattern of protection; (ii) wine is more protective than beer; and (iii) red wine is more protective than white. Of these, the J-shaped curve is best established, but in every case confounding factors are difficult to exclude. 107,122 However, in the case of the J-shaped curve, the genetic links between alcohol dehydrogenase activity, increased HDL-cholesterol, and decreased myocardial infarction, and the prospective trial data on HDL, fibrinogen, and insulin, are powerful arguments against confounding. Moderate red wine drinking could be, at least in part, a surrogate for a healthy life style. In this case the cardiovascular data with de-alcoholized red but not white wine and studies on isolated tissues or organs give decisive mechanistic insights that strengthen the case already made for randomized prospective controlled trials with cardiovascular endpoints. 122 The first trial should simply compare alcohol with red wine with abstinence, for example in postinfarct patients. 107 Similar prospective studies would be needed conclusively to prove the red-better-than-white wine hypothesis. As these studies may never be done, attention has shifted to the beneficial components of red wine. Resveratrol remains as the most powerful of the polyphenols and one of the most likely to give biological protection. 123 The recent discovery that resveratrol is an agonist of the sirtuin (SIRT) proteins may lead to a dose-response trial of resveratrol and red wine in heart failure patients. However, pharmacokinetic data remain incomplete. Experimentally, other polyphenolic components of red wine also play a protective role. One hypothesis would be that no single component of red wine confers all the benefit, but rather several components contribute by different mechanisms. For example, resveratrol and procyanidins are both antioxidants, but resveratrol interacts with SIRT, whereas the procyanidins inhibit endothelin-1 synthesis. Given that there will always be wine drinkers, do current data help in choice of wine? First, all drinkers of any alcohol should avoid overdosing and binge-drinking. Then they should consider the incomplete but emerging evidence that red wine is more cardioprotective (and possibly, neuroprotective) than white. Furthermore, as red wine-drinkers can choose a specific varietal, pinot noir is high both in resveratrol 124 and in procyanidins. 125 The price of pinot is high enough to ensure drinking only in moderate amounts and preferably with meals. Some may be influenced by the neuroprotective qualities of Cabernet Sauvignon given in low doses to rodents. And others, especially those influenced by American epidemiological data, will settle for ethanol as the major cardioprotective agent in wine. Teetotallers will be consoled by the substantial data linking de-alcoholized red wine to experimental cardiovascular benefits and turn to purple grape juice. Hopefully, in the future, low alcohol yet still protective wines will make their debut. 126 Special patient groups who need abstinence or specific limitations Those with gout are traditionally warned against red wine, which does increase blood uric acid levels. 127 Those with alcoholic cardiomyopathy may not take any alcohol, despite the ambiguous effects of ethanol on myocardial function. 128 Those with atrial fibrillation should be safer with less than 35 drinks/week. 19 Hypertensive males may need to limit drinks to not more than two/day 34,129 with, presumably, even fewer for hypertensive females. However, as alcohol only elevates BP by a few millimetres of mercury, a reasonable alternative would be to monitor by BP measurements the contrasting policies of (i) allowing the normal moderate amounts of alcohol while watching the BP and, if needed, adjusting doses of anti-hypertensive drugs; and (ii) keeping the medication unchanged but cutting the alcohol. Post-myocardial infarction, moderate alcohol does not impose a greater risk of overt heart failure on those already having systolic dysfunction. 130 Also, it is important to remember that alcohol interacts with many drugs such as nitroglycerin, anxiolytics, neuroleptics, and oral contraceptives. Summary There are strong epidemiological data favouring the view that modest alcohol but neither zero nor more than modest intake beneficially reduces total mortality and cardiovascular risk. As there is only a narrow window of opportunity for optimal intake, it becomes important to consider which is the most beneficial type of alcohol. The red-better-than white hypothesis rests on 1 human studies showing short-term cardiovascular benefits achieved by de-alcoholized red but not white wine; and 2 many studies on isolated tissues or organs assessing the effects of polyphenols in red wine, especially resveratrol. Thus, red wine potentially has beneficial effects beyond alcohol. The molecular mechanisms involved are many and complex as revealed by studies on isolated tissues or organs. Besides, the well known antioxidant effects of polyphenols, resveratrol effects include activation of preconditioning-like prosurvival signalling paths similar to those responding to insulin, anti-apoptotic protection mediated by the antisenescence factor sirtuin, and inhibition of adverse signalling by TNF-a, NF-k-B, angiotensin-ii, and growth factors (epidermal and transforming growth factor-beta). Mechanistic feasibility strengthens the case for prospective therapeutic trials. Acknowledgements We would like to thank D. Blackhurst for very helpful comments. Conflict of interest: none declared. References 1. Kannel WB, McGee D, Gordon T. A general cardiovascular risk profile: the Framingham Study. Am J Cardiol 1976;38: Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen M, Budaj A, Pais P, Varigos J, Lisheng L. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case control study. Lancet 2004;364: Gordon T, Kannel WB. Drinking and mortality. The Framingham Study. Am J Epidemiol 1984;120: Doll R, Peto R, Hall E, Wheatley K, Gray R. Mortality in relation to consumption of alcohol: 13 years observations on male British doctors. Brit Med J 1994;309:

9 The red wine hypothesis Thun MJ, Peto R, Lopez AD, Monaco JH, Henley SJ, Heath CW Jr, Doll R. Alcohol consumption and mortality among middle-aged and elderly U.S. adults. N Engl J Med 1997;337: Fuchs CS, Stampfer MJ, Colditz GA, Giovannucci EL, Manson JE, Kawachi I, Hunter DJ, Hankinson SE, Hennekens CH, Rosner B. Alcohol consumption and mortality among women. N Engl J Med 1995;332: Gaziano JM, Gaziano TA, Glynn RJ, Sesso HD, Ajani UA, Stampfer MJ, Manson JE, Hennekens CH, Buring JE. Light-to-moderate alcohol consumption and mortality in the Physicians Health Study enrollment cohort. J Am Coll Cardiol 2000;35: Emberson JR, Shaper AG, Wannamethee SG, Morris RW, Whincup PH. Alcohol intake in middle age and risk of cardiovascular disease and mortality: accounting for intake variation over time. Am J Epidemiol 2005; 161: Suh I, Shaten BJ, Cutler JA, Kuller LH. Alcohol use and mortality from coronary heart disease: the role of high-density lipoprotein cholesterol. The Multiple Risk Factor Intervention Trial Research Group. Ann Intern Med 1992;116: Gronbaek M, Becker U, Johansen D, Gottschau A, Schnohr P, Hein HO, Jensen G, Sorensen TI. Type of alcohol consumed and mortality from all causes, coronary heart disease, and cancer. Ann Intern Med 2000; 133: Di Castelnuovo A, Rotondo S, Iacoviello L, Donati MB, De Gaetano G. Meta-analysis of wine and beer consumption in relation to vascular risk. Circulation 2002;105: Doll R, Peto R, Hall E, Wheatley K, Gray R. Alcohol and coronary heart disease reduction among British doctors: confounding or causality? Eur Heart J 1997;18: Naimi TS, Brown DW, Brewer RD, Giles WH, Mensah G, Serdula MK, Mokdad AH, Hungerford DW, Lando J, Naimi S, Stroup DF. Cardiovascular risk factors and confounders among nondrinking and moderate-drinking U.S. adults. Am J Prev Med 2005;28: Di Castelnuovo A, Costanzo S, Bagnardi V, Donati MB, Iacoviello L, de Gaetano G. Alcohol dosing and total mortality in men and women: an updated meta-analysis of 34 prospective studies. Arch Intern Med 2006;166: Mukamal KJ, Chiuve SE, Rimm EB. Alcohol consumption and risk for coronary heart disease in men with healthy lifestyles. Arch Intern Med 2006;166: Bryson CL, Mukamal KJ, Mittleman MA, Fried LP, Hirsch CH, Kitzman DW, Siscovick DS. The association of alcohol consumption and incident heart failure: the Cardiovascular Health Study. J Am Coll Cardiol 2006;48: Beilin LJ, Puddey IB. Alcohol and hypertension: an update. Hypertension 2006;47: Hines LM, Stampfer MJ, Ma J, Gaziano JM, Ridker PM, Hankinson SE, Sacks F, Rimm EB, Hunter DJ. Genetic variation in alcohol dehydrogenase and the beneficial effect of moderate alcohol consumption on myocardial infarction. N Engl J Med 2001;344: Mukamal KJ, Jensen MK, Gronbaek M, Stampfer MJ, Manson JE, Pischon T, Rimm EB. Drinking frequency, mediating biomarkers, and risk of myocardial infarction in women and men. Circulation 2005; 112: Renaud S, de Lorgeril M. Wine, alcohol, platelets, and the French paradox for coronary heart disease. Lancet 1992;339: Mukamal KJ, Jadhav PP, D Agostino RB, Massaro JM, Mittleman MA, Lipinska I, Sutherland PA, Matheney T, Levy D, Wilson PW, Ellison RC, Silbershatz H, Muller JE, Tofler GH. Alcohol consumption and hemostatic factors: analysis of the Framingham Offspring cohort. Circulation 2001; 104: Imhof A, Woodward M, Doering A, Helbecque N, Loewel H, Amouyel P, Lowe GD, Koenig W. Overall alcohol intake, beer, wine, and systemic markers of inflammation in western Europe: results from three MONICA samples (Augsburg, Glasgow, Lille). Eur Heart J 2004;25: Davies MJ, Baer DJ, Judd JT, Brown ED, Campbell WS, Taylor PR. Effects of moderate alcohol intake on fasting insulin and glucose concentrations and insulin sensitivity in postmenopausal women: a randomized controlled trial. JAMA 2002;287: Meagher EA, Barry OP, Burke A, Lucey MR, Lawson JA, Rokach J, FitzGerald GA. Alcohol-induced generation of lipid peroxidation products in humans. J Clin Invest 1999;104: Hu FB, Manson JE, Stampfer MJ, Colditz G, Liu S, Solomon CG, Willett WC. Diet, lifestyle, and the risk of type 2 diabetes mellitus in women. N Engl J Med 2001;345: Sato M, Maulik N, Das DK. Cardioprotection with alcohol: role of both alcohol and polyphenolic antioxidants. Ann NY Acad Sci 2002;957: Guiraud A, de Lorgeril M, Boucher F, Berthonneche C, Rakotovao A, de Leiris J. Cardioprotective effect of chronic low dose ethanol drinking: insights into the concept of ethanol preconditioning. J Mol Cell Cardiol 2004;36: Barefoot JC, Gronbaek M, Feaganes JR, McPherson RS, Williams RB, Siegler IC. Alcoholic beverage preference, diet, and health habits in the UNC Alumni Heart Study. Am J Clin Nutr 2002;76: Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to a Mediterranean diet and survival in a Greek population. N Engl J Med 2003;348: Ruidavets JB, Bataille V, Dallongeville J, Simon C, Bingham A, Amouyel P, Arveiler D, Ducimetiere P, Ferrieres J. Alcohol intake and diet in France, the prominent role of lifestyle. Eur Heart J 2004;25: McFadden CB, Brensinger CM, Berlin JA, Townsend RR. Systematic review of the effect of daily alcohol intake on blood pressure. Am J Hypertens 2005;18: Zilkens RR, Burke V, Hodgson JM, Barden A, Beilin LJ, Puddey IB. Red wine and beer elevate blood pressure in normotensive men. Hypertension 2005;45: Wildman RP, Gu D, Muntner P, Huang G, Chen J, Duan X, He J. Alcohol intake and hypertension subtypes in Chinese men. J Hypertens 2005; 23: Xin X, He J, Frontini MG, Ogden LG, Motsamai OI, Whelton PK. Effects of alcohol reduction on blood pressure: a meta-analysis of randomized controlled trials. Hypertension 2001;38: Potter JF, Beevers DG. Pressor effect of alcohol in hypertension. Lancet 1984;1: van den Elzen AP, Sierksma A, Oren A, Vos LE, Witteman JC, Grobbee DE, Hendriks HF, Uiterwaal CS, Bots ML. Alcohol intake and aortic stiffness in young men and women. J Hypertens 2005;23: Karatzi KN, Papamichael CM, Karatzis EN, Papaioannou TG, Aznaouridis KA, Katsichti PP, Stamatelopoulos KS, Zampelas A, Lekakis JP, Mavrikakis ME. Red wine acutely induces favorable effects on wave reflections and central pressures in coronary artery disease patients. Am J Hypertens 2005;18: Renaud SC, Gueguen R, Conard P, Lanzmann-Petithory D, Orgogozo JM, Henry O. Moderate wine drinkers have lower hypertension-related mortality: a prospective cohort study in French men. Am J Clin Nutr 2004; 80: Tolstrup J, Jensen MK, Tjonneland A, Overvad K, Mukamal KJ, Gronbaek M. Prospective study of alcohol drinking patterns and coronary heart disease in women and men. BMJ 2006;332: Fuchs FD, Chambless LE, Folsom AR, Eigenbrodt ML, Duncan BB, Gilbert A, Szklo M. Association between alcoholic beverage consumption and incidence of coronary heart disease in whites and blacks: the Atherosclerosis Risk in Communities Study. Am J Epidemiol 2004;160: Stranges S, Wu T, Dorn JM, Freudenheim JL, Muti P, Farinaro E, Russell M, Nochajski TH, Trevisan M. Relationship of alcohol drinking pattern to risk of hypertension: a population-based study. Hypertension 2004;44: Blanco-Colio LM, Valderrama M, Alvarez-Sala LA, Bustos C, Ortego M, Hernandez-Presa MA, Cancelas P, Gomez-Gerique J, Millan J, Egido J. Red wine intake prevents nuclear factor-kappab activation in peripheral blood mononuclear cells of healthy volunteers during postprandial lipemia. Circulation 2000;102: Mukamal KJ, Conigrave KM, Mittleman MA, Camargo CA Jr, Stampfer MJ, Willett WC, Rimm EB. Roles of drinking pattern and type of alcohol consumed in coronary heart disease in men. N Engl J Med 2003;348: Klatsky AL, Friedman GD, Armstrong MA, Kipp H. Wine, liquor, beer, and mortality. Am J Epidemiol 2003;158: Renaud SC, Gueguen R, Siest G, Salamon R. Wine, beer, and mortality in middle-aged men from eastern France. Arch Intern Med 1999;159: Johansen D, Friis K, Skovenborg E, Gronbaek M. Food buying habits of people who buy wine or beer: cross-sectional study. BMJ 2006;332: Wannamethee SG, Shaper AG. Type of alcoholic drink and risk of major coronary heart disease events and all-cause mortality. Am J Publ Health 1999;89: