OUP user menu


Dharam P. Agarwal
DOI: http://dx.doi.org/10.1093/alcalc/37.5.409 409-415 First published online: 1 September 2002


— There is abundant epidemiological and clinical evidence showing that light–moderate drinking is associated with a reduced risk of coronary heart disease (CHD), total and ischaemic stroke and total mortality in middle-aged and elderly men and women. The epidemiological evidence suggests a J- or U-shaped relationship between alcohol and CHD. However, the apparent benefits of moderate drinking on CHD mortality are offset at higher drinking levels by increasing risk of death from other types of heart diseases (cardiomyopathy, arrhythmia etc.), neurological disorders, cancer, liver cirrhosis, and traffic accidents. The plausible mechanisms for the putative cardioprotective effects include increased levels of high-density lipoprotein cholesterol, decreased levels of low-density lipoprotein cholesterol, prevention of clot formation, reduction in platelet aggregation, and lowering of plasma apolipoprotein(a) concentration. Thus, alcohol reduces the risk of coronary vascular diseases both by inhibiting the formation of atheroma and decreasing the rate of blood coagulation.


Drinking excessive amounts of alcohol regularly for years is toxic to almost every tissue of the body. Many of the toxic effects of alcohol are due to disturbances of a wide variety of metabolic functions and organ damage. Long-term alcohol use increases the risk of liver disease, heart disease, peptic ulcers, certain types of cancers, complicated pregnancies, birth defects, and brain damage (Agarwal and Seitz, 2001). Heavy or binge drinking may even result in respiratory depression and death. Alcohol use can also cause mood changes and loss of inhibitions as well as violent or self-destructive behaviour.

On the other hand, epidemiological and clinical evidence shows that light–moderate drinking is associated with a reduced risk of coronary heart disease (CHD), total and ischaemic stroke and total mortality in middle-aged and elderly men and women (Doll, 1997; Grobbee et al., 1999; Rimm et al., 1999; Klatsky, 2001; Rotondo et al., 2001; van Tol and Hendriks, 2001). The evidence suggests a J- or U-shaped relationship between alcohol and CHD. This article reviews the epidemiological evidence for alcohol’s putative cardioprotective effects and discusses the plausible underlying biological mechanisms.


Alcohol intake is frequently expressed in ‘drinks’ or ‘units’ that vary with beverage type, culture and era. Alcohol drinking may be divided into ‘light’, ‘moderate’ and ‘heavy’ categories, depending upon the amount of alcohol consumed in terms of pure ethanol per day (Dufour, 1999; Kalant and Poikolainen, 1999). ‘Light–moderate’ alcohol intake is usually defined as an average consumption of 1 to 2 drinks per day; ‘heavy drinking’ indicates usual intake of ≥3 drinks per day of wine, liquor, or beer. In absolute terms, amounts of alcohol consumed <30 g/day are taken as light–moderate drinking, amounts of pure alcohol consumed >30 g per day are considered as heavy drinking. Whereby drinking pattern, type of beverage, gender, and age are important parameters in defining moderate and heavy drinking.


Risk of death from all causes has been found to be significantly lower among men who drink moderately, compared to abstainers. Several epidemiological investigations have shown that a low to moderate level of alcohol intake has a definitive protective role against CHD and stroke. Such conclusions have been based upon epidemiological studies on the risks for heart disease, coronary artery disease and death in individuals with low or moderate alcohol intake, when compared with the corresponding risks in persons who do not consume alcohol at all (Rimm et al., 1999; Corrao et al., 2000; Meister et al., 2000; Agarwal and Srivastava, 2001). The dose–response curve usually is found to be J- or U-shaped, i.e. the risk is higher when alcohol consumption is high, lower when alcohol consumption is low or moderate, and tends to go up again in individuals not consuming any alcohol (Andreasson, 1998; San Jose et al., 1999). Level of alcohol consumption that has been associated with lower risk for CHD ranges as widely as from 1 drink daily to ∼3 drinks per day (Rimm et al., 1999; Gronbaek et al., 1999). When all cohort data of the above-mentioned studies are combined, there appears to be a decline in the risk for myocardial infarction at doses up to 1 drink per day, with little further change in risk associated with increased alcohol intake (Rimm et al., 1999). Berger et al. (1999) found that light–moderate alcohol consumption reduced the overall risk of stroke and risk of ischaemic stroke in men. The benefit was apparent with as little as 1 drink per week. Greater consumption up to 1 drink per day did not increase the observed benefit. In a Finnish study, Makela et al. (1997) observed that, among men aged 30–69 years, the beneficial effects of light– moderate alcohol consumption ‘prevented’ some 400 CHD deaths each year, which corresponds to 12–14% of the observed CHD deaths. Rimm et al. (1999) in their meta-analysis concluded that alcohol intake (30 g of alcohol per day) is causally related to 24.7% reduction in risk of CHD through changes in lipids, lipoproteins and fibrinogen.

In a more recent meta-analysis relating to alcohol consumption with the risk of CHD (Corrao et al., 2000), the risk decreased from drinking levels of 0 to 20 g/day (RR = 0.80; 95% CI); there was evidence of a protective effect up to 72 g/day (RR = 0.96; 95% CI) and increased risk at ≥89 g/day (RR = 1.05; 95% CI). Lower protective effects and harmful effects were found in women, in men living in countries outside the Mediterranean area and in studies where fatal events were used as the outcome.


The strongest inverse correlation between moderate drinking and CHD has been shown among both men and women who consumed 1–2 drinks per day on 5–6 days per week (McElduff and Dobson, 1997). Rimm et al. (1999) observed that men who reported drinking, on average, on 3–4 days per week had a relative risk of 0.66, compared with men who drank less than 1 day a week. Alcohol drinking pattern may also have a profound influence on the blood pressure effects of alcohol. Intervention studies in men have shown acute increases in blood pressure in men who drink predominantly at weekends, compared to longer-term pressor effects in regular daily drinkers (Marques-Vidal et al., 2001). The binge-drinking pattern observed among Northern Irish drinkers leads to physiologically disadvantageous consequences regarding blood pressure levels, whereas no such fluctuations in blood pressure levels were found for regular consumption noted among French drinkers (Puddey et al., 1999). Hence, exploration of any protective association of alcohol against CHD needs to consider carefully the implications of pattern of drinking for the relationship.


The low CHD mortality rate observed in Mediterranean populations in association with red wine consumption and a high saturated fat intake has given rise to what is now popularly termed the ‘French paradox’. This phenomenon refers to people residing in certain parts of France and other Mediterranean countries where red wine is customarily consumed during meals. These populations show a low CHD mortality, despite living a lifestyle considered to have comparably high CHD risks, like those in the USA and many other developed countries (Criqi, 2001). This relationship has been observed in both men and women and in different age groups. Many investigators have claimed that wine is the significant factor explaining the French paradox (see below).


Most of the prospective cohort studies have consistently shown that moderate alcohol intake has a protective effect against CHD (Leighton et al., 1997; McKee and Britton, 1998; Rimm et al., 1999; Gaziano et al., 2000). As shown in Table 1, several factors have been proposed to explain the beneficial effects of moderate alcohol consumption on the development of CHD and atherosclerosis (Srivastava et al., 1994; Grobbee et al., 1999; Goldberg and Soleas, 2001; Rotondo et al., 2001). These include effects of alcohol on high-density lipoprotein cholesterol (HDL), low-density lipoprotein cholesterol (LDL), plasma apolipoprotein(a) [Lp(a)] levels, platelet aggregability, blood fibrinolytic activity, insulin sensitivity, oestrogen levels, and stress. The following is a discussion of these factors.

View this table:
Table 1.

Putative biological mechanisms underlying cardioprotection by low–moderate alcohol consumption

ParameterCardioprotective effect of moderate alcohol intake
HDL, high-density lipoprotein; LDL, low-density lipoprotein.
Lipid and lipoprotein profileIncreases protective (‘good’) HDL-cholesterol
Inhibits oxidation of harmful (‘bad’) LDL-cholesterol
ThrombosisReduces platelet aggregation
Reduces fibrinogen levels
Increases fibrinolysis (the process by which clots dissolve)
Cardiovascular systemIncreases coronary blood flow
Reduces blood pressure (<1–2 drinks per day)
HormonesReduces blood insulin levels
Increases blood insulin sensitivity
Increases oestrogen levels
LifestyleReduces stress
Other effectsIncreases paraoxonase activity
Decreases plasma homocysteine levels

Increase in HDL

HDL-cholesterol is inversely related to CHD and hence it has been proclaimed that alcohol’s effect is mediated through increased HDL (De Oliviera et al., 2000; Sillanaukee et al., 2000; Hannuksela and Savolainen, 2001). Cholesteryl ester transfer protein (CETP) mediates the transfer of cholesteryl esters from HDL into very-low-density lipoproteins (VLDL) and LDL with a reciprocal exchange of triglycerides (Krause and Auerbach, 2001). A low transfer rate may reduce the reverse transport of cholesterol (Castilho et al., 2001). The plaque regression has been demonstrated angiographically in those persons who received drugs that lower cholesterol and increase HDL, indicating that the reverse cholesterol transport can be enhanced by raising the HDL levels and thereby overcoming the inhibitory effects of oxidized LDL upon the atherogenic process. While it was earlier thought that alcohol increased only HDL-3 and not HDL-2, recent observations have shown that both classes contribute equally and cooperatively to the overall efficiency of reverse transport of cholesterol (Hannuksela and Savolainen, 2001). Hence, alcohol intake may raise plasma HDL levels either by altering the synthesis or clearance of HDL or by effects on enzymes and proteins influencing HDL metabolism. According to Rimm et al. (1999), a 16.8% reduction in CHD is directly attributable to increased HDL from consuming 30 g alcohol per day.

Decreased LDL oxidation

Many observational studies have provided strong evidence that oxidation of LDL plays an important role in the progression of atherosclerotic vascular disease (Griffin, 1999). In fact, the antioxidative effect of alcohol, in all probability, is exerted by affecting the oxidation of LDL thereby hampering atherosclerotic plaque formation (Serafini et al., 2000). Wine contains compounds with antioxidant capacity that could account for its postulated stabilizing effect on LDL. Grape-derived beverages supply a large number of nutritional antioxidants, because of their high content of polyphenols. This might be one of the mechanisms behind the supposed beneficial effect of red wine. Indeed, polyphenolic antioxidants present in red wine have been shown to inhibit the oxidation of human LDL (Puddey et al., 1998; van Golde et al., 1999). It has recently been shown that distilled alcohol on wood ageing acquires significant amounts of antioxidants which could be different from those present in red wine (Puddey and Croft, 1999). These findings suggest that there might not be large differences between the capabilities of wine (red wine especially) and distilled alcohol in the protection and/or lower risk of heart disease (Goldberg et al., 1999).

Alcohol-related changes in apolipoprotein profile

Apolipoprotein AI is distributed within high-density lipoprotein (HDL) between different types of particles. Alcohol drinking increases the plasma concentrations of apolipoprotein AI and apolipoprotein AII, the main components of HDL particles (De Oliviera et al., 2000). Based upon epidemiological data, it has been estimated that an average individual consuming 30 g of alcohol per day would show an 8 mg/dl increase in the plasma concentration of apolipoprotein AI, primarily due to increased synthesis in liver (Rimm et al., 1999).

Reduction in apolipoprotein(a)

Recent observational studies showed that Lp(a) plays a significant role in atherosclerosis and is one of the major risk factors for cardiovascular disease (Paassilta et al., 1998; Fontana et al., 1999). Epidemiologically, Lp(a) is a strong positive risk factor for CHD, the presumed mechanism being the inhibition of fibrinolysis by reduced plasminogen levels due to elevated Lp(a) levels (Testa and Marcovina, 1999). A number of intervention studies in human subjects support the notion that alcohol consumption reduces plasma Lp(a) concentrations. As reported by Kervinen et al. (1993), cessation or reduction in beverage alcohol consumption was accompanied by a significant increase in plasma Lp(a). We have earlier demonstrated that CHD patients have higher levels of Lp(a) and those who were alcohol drinkers showed significantly lower Lp(a) levels (Vasisht et al., 1996). Further, Paassilta et al. (1998) have shown that social drinking is associated with low Lp(a) lipoprotein concentration in middle-aged men and concluded that low Lp(a) lipoprotein concentration may be one factor explaining low mortality and retarded progression of coronary artery disease in social drinkers.

Reduction in blood clotting and platelet aggregation

Increased HDL-cholesterol levels can explain only a part of the protective effect of alcoholic beverages (Rimm et al., 1999). The other part may be related to decreased platelet activity and other clotting factors. Moderate alcohol consumption may affect several haemostatic factors, including fibrinogen concentration, platelet aggregability and the fibrinolytic factors: tissue-type plasminogen activator and plasminogen activator inhibitor (Djousse et al., 2000; Mukamal et al., 2001; van de Wiel et al., 2001). Plasma fibrinogen concentrations are decreased by moderate alcohol consumption (Hendriks and van der Gaag, 1998; Mennen et al., 1999; Lacoste et al., 2001). Hence, alcohol has been shown to reduce blood platelet aggregability (Ruf, 1999). According to Rimm et al. (1999), the projected percentage reduction in the risk of CHD attributed to effects of alcohol on concentrations of fibrinogen were 4.3 (unadjusted) and 12.5 (adjusted for intraindividual variability), respectively.

The antiplatelet activity of wine is explained not only by ethanol, but also by the polyphenolic components with which red wines are richly endowed. It appears that wine and wine phenolics in particular could significantly inhibit platelet aggregation and that this could explain, at least in part, the protective effect of red wine against atherosclerosis and CHD.

Light–moderate alcohol consumption reduces blood pressure

Epidemiological data clearly show higher mean blood pressure and/or hypertension with increasing alcohol drinking (Grobbee et al., 1999). Blood pressure effects of alcohol vary according to chronicity and amount of intake. On the other hand, short-term lowering of blood pressure by alcohol is known. A J-shaped relation between alcohol and blood pressure has been suggested with moderate drinkers having a lower blood pressure level, with the lowest levels in consumers of 1 to 3 drinks per day (Gillman et al., 1995; Beilin et al., 1996).

Alcohol reduces insulin resistance and increases insulin sensitivity

Regular moderate alcohol consumption is associated with decreased insulin resistance and this may partly explain the cardioprotective effect of alcohol (Bell et al., 2000; Flanagan et al., 2000). Moreover, light–moderate alcohol consumption promotes insulin sensitivity of skeletal muscle (Facchini et al., 1994). Conceivably, this benefits the protective effects of moderate drinking on vascular health and risk for obesity and diabetes. The mechanism responsible for alcohol’s insulin-sensitizing activity remains obscure. McCarty (2001) has proposed that metabolism of acetate in peripheral tissues generates sufficient levels of AMP to temporarily stimulate the AMP-activated protein kinase, which in turn induces the synthesis of certain long-lived proteins that act to boost insulin sensitivity and possibly aid the efficiency of fat oxidation as well. Moreover, improvement of insulin sensitivity results in higher HDL-cholesterol levels; it is suggested that this is one of the routes taken by alcohol to act upon HDL metabolism (Kiechl et al., 1996). Whether interrelated or not, both the HDL-cholesterol increase and enhanced insulin sensitivity are considered to have a beneficial effect on the process of atherosclerosis (Lazarus et al., 1997).

Reduction in plasma homocysteine

Recent studies suggest that high plasma homocysteine concentrations are an independent risk factor for coronary, cerebral and peripheral arterial occlusive diseases (Danesh and Lewington, 1998; Chambers et al., 2000). Its effect appears to depend upon its direct toxicity for endothelial cells. Endothelial dysfunction is associated with atherogenesis and oxidative stress in humans. Some observational studies have reported that light–moderate alcohol intake is associated with lower plasma total homocysteine level (Ubbink et al., 1998; de Bree et al., 2001). However, it was recently reported that serum homocysteine increases even after moderate alcohol consumption in social drinkers (Bleich et al., 2001). Thus, the cardioprotective role of moderate alcohol consumption, particularly in relation to homocysteine, remains debatable and the concept of the ‘French paradox’ needs further investigation (Badawy, 2001).

Increase in paraoxonase activity

Recent studies have implicated paraoxonase in providing protection against LDL oxidation, thus affecting the risk of CHD in the general population (van der Gaag et al., 1999). The human serum HDL-linked paraoxonase enzyme limits LDL peroxidation by preventing transformation of LDL into biologically active atherogenic particles. Fasting paraoxonase activity was higher after intake of wine, beer, and spirits than after water consumption, but did not differ significantly between the three alcoholic beverages (van der Gaag et al., 1999). These findings and those from other recent studies (Durrington et al., 2001) suggest that increased serum paraoxonase may be one of the biological mechanisms underlying the reduced CHD risk in moderate alcohol consumers.

Moderate alcohol consumption raises oestrogen levels

Alcohol consumption may increase blood oestradiol levels in postmenopausal women who are on oestrogen replacement therapy, and this may increase the risk of breast cancer (Purohit, 1998). Moderate alcohol intake exerts a major influence not only on oestradiol, testosterone, and the estimate of aromatization of testosterone to oestradiol, but also on the oestrogen-responsive pituitary hormones in normal postmenopausal women (Gavaler et al., 1993). These findings suggest that moderate alcohol use is an important factor for postmenopausal oestrogen status and may offer a partial explanation for the reported protective effect of moderate alcohol consumption with respect to postmenopausal cardiovascular disease risk. However, there are data suggesting that the use of both alcohol and oestrogen may increase breast cancer risk more than the use of either agent alone (Ginsburg, 1999).

Moderate alcohol consumption reduces stress

Light–moderate drinkers have less depression in the presence of stress, than persons in other more extreme drinking categories. According to Lipton (1994), moderate alcohol use may serve as a proxy for a spectrum of generally moderate behaviours that either attenuate the effect of stress on depression or suppress the effects of stress. A review of the literature (Baum-Baicker, 1985) on the positive psychological benefits of light and moderate alcohol consumption suggests that alcohol in moderate amounts is effective in reducing stress. Low and moderate doses of alcohol increase overall affective expression, happiness, euphoria, conviviality and pleasant and carefree feelings (Vasse et al., 1998). Tension, depression and self-consciousness have been reported to decrease with equal doses. Low alcohol doses have been found to improve certain types of cognitive performance. Included here are problem-solving and short-term memory (Baum-Baicker, 1985). Heavy drinkers and abstainers have higher rates of clinical depression than do regular moderate drinkers. Alcohol in low and moderate doses has been effective in the treatment of geropsychiatric problems (Baum-Baicker, 1985).


The putative health benefits of light–moderate alcohol drinking may vary among individuals. While the association between light–moderate drinking and reduced risk of CHD mortality is well established, the issues of causality and magnitude of effect are more problematic. A wide range of confounding factors may indirectly explain the apparent variability in cardiovascular benefits of light–moderate drinking. Published findings have, nevertheless, shown that alcohol’s beneficial effects are independent of dietary and other known risk factors for heart disease, such as smoking and obesity.

Epidemiological studies have shown that the cardioprotective effect of alcohol does not include men aged <40 years or premenopausal women (Meister et al., 2000). There is also some evidence that the cardioprotective effect cannot be achieved if a person waits to reach middle age before starting to drink (Wannamethee and Shaper, 2002). Moreover, the putative benefit could be inflated because of the use of reference groups with different lifestyles and health profiles (Shaper and Wannamethee, 2000). Alcohol drinking patterns (occasional, regular, binge, with or without food, etc.) are also of importance in relation to the effects of intake.

Despite the overwhelming epidemiological evidence of an inverse relationship between moderate alcohol consumption and CHD, there remains a pending question. Is alcohol the direct cause of the observed cardioprotective effects or is it due to some other factors. Alcohol may directly increase the hepatic production and secretion of apolipoproteins and lipoprotein particles, increase triglyceride lipase concentrations, and decrease removal of circulating HDL-cholesterol. It appears that about half the apparent protective effect of alcohol against CHD is due to higher HDL-cholesterol, leading to an enhancement of reverse cholesterol transport, thereby reducing the risk of CHD (Goldberg and Soleas, 2001). In addition, alcohol may inhibit blood clotting and platelet aggregation, perhaps due to the presence of polyphenolic antioxidants in alcoholic beverages, partcularly in wine. Type of alcoholic beverage intake may play a role in the effects on CHD. Red wine has been shown ex vivo to inhibit low-density lipoprotein oxidation, increase antioxidant capacity in humans, and reduce susceptibility of human plasma to lipid peroxidation. The non-alcoholic fraction of wine, represented mainly by phenolic compounds, may be the primary factor responsible for this protective effect.

Those who do not benefit from alcohol could have an even lower risk if they do not drink, because of the elevation by alcohol of the cardiovascular risk substance homocysteine. Though most of the evidence points to increased homocysteine after alcohol drinking (Bleich et al., 2001), an inverse relationship between moderate alcohol consumption and plasma total homocysteine levels has been observed in other studies, at least in men (Ubbink et al., 1998; de Bree et al., 2001; Ueland et al., 2001). In particular, serum homocysteine increased after moderate consumption of red wine and spirits, but not after moderate consumption of beer (van der Gaag et al., 2000). Vitamin B6, vitamin B2 and folate present in beer seem to prevent an alcohol-induced rise in serum homocysteine (de Bree et al., 2001; Mayer et al., 2001).

Taken together, in comparison with non-drinkers, individuals who consume light–moderate amounts of alcohol have, on average, lower death rates due to cardiovascular disorders. However, it is suggested that individuals who are currently non-drinkers should not be encouraged to start drinking alcohol because of its putative cardioprotective properties.


Financial support from the Volkswagen Foundation, Hannover, Germany is gratefully acknowledged.


View Abstract