Portal de Periódicos da CAPES

Alcohol and Coronary Heart Disease: Can We Learn More?

2001; Lippincott Williams & Wilkins; Volume: 12; Issue: 4 Linguagem: Inglês

10.1097/00001648-200107000-00006

1531-5487

Eric B. Rimm,

Substance Abuse Treatment and Outcomes

Research on the association between alcohol consumption and coronary heart disease has spanned more than 3 decades. Over 100 studies, including ecological, case-control, and cohort studies, have addressed this complex relationship. There has also been an almost equal number of reviews, editorials, meta-analyses, and letters. 1–6 The excessive use of alcohol is linked to cirrhosis, aerodigestive cancers, motor vehicle deaths, and is frequently implicated in homicides and domestic disputes. Nevertheless, the evidence supporting an inverse association between moderate alcohol consumption and coronary heart disease is overwhelming and consistent. With the exception of the link between smoking and lung cancer, few other associations are so uniformly reported in the literature despite diverse population samples, varying exposure and outcome measurements, and inconsistent control for confounding. In this issue of Epidemiology, Brenner et al report a substantial reduction in risk of coronary heart disease (CHD) associated with alcohol consumption. 7 These results from Germany are similar to those reported in China, Japan, France, the United States, and the United Kingdom, to name a few. 8–13 A long-term, large-scale clinical trial of alcohol and coronary heart disease would be difficult to conduct. Thus, guidelines and clinical advice must be based on evidence from these types of non-experimental studies and from our understanding of the effect of short-term administration of alcohol from experimental studies. A summary of this body of data suggests strongly that abstaining from alcohol is a risk factor for coronary heart disease; yet, advice to abstainers to drink alcohol in moderation is generally not prudent on a population level but may be on an individual basis. If the evidence is so strong in favor of an inverse association between alcohol and coronary heart disease, what can we learn from new research in this area? If alcohol (ethanol) were always consumed in the same manner and as part of the same beverage, there would be less to learn; however, recent evidence indicates that drinking patterns (including beverage choice), other dietary constituents, and even genetic predisposition to ethanol metabolism may modify the underlying association between moderate alcohol consumption and coronary heart disease. 14–18 Beverage preference has received the most attention because it has direct public health relevance. An equal number of prospective studies reported benefit for wine as did for beer or spirits. In this current analysis from Germany, Brenner and colleagues had sufficient drinkers of beer only and wine only to examine beverage-specific associations. After adjusting for potential confounders, the inverse association for beer-only drinkers (OR = 0.50, 95% CI = 0.30–0.84) was much stronger than the inverse association for drinkers of wine only (OR = 0.95, 95% CI = 1.57–1.59). Conversely, in a large cohort study in Denmark by Grønbaek et al., 19 very different associations were reported: drinkers of wine had substantially lower risk of coronary heart disease than drinkers of beer or spirits. In other populations, benefits are found only among spirits drinkers, while in others benefits are found for all forms of alcohol-containing beverages. 20 What can be the explanation for such divergent findings for beverage choice when almost all agree that ethanol lowers risk of coronary heart disease? Some might attribute the differences to chance. A second explanation for the divergent findings could be that wine intake is more protective against coronary heart disease in some populations, while beer or spirits is more beneficial in others. This explanation is very unlikely and would be difficult to study. We know from more than 30 metabolic studies that ethanol intake from any source raises HDL-lipoprotein cholesterol and decreases the blood's propensity to clot by affecting several coagulation factors. 21 Therefore, drinking wine, beer, or spirits should be inversely associated with coronary heart disease to some extent in all populations. Thus, the lack of an inverse association for spirits in the Danish study and for wine in the Brenner et al study, if not attributable to chance, must be due to factors not measured or accounted for in these analyses. A possible third explanation for differences among studies is that confounding and exposure modification has not been adequately addressed. In the original studies from the Danish cohorts, wine was highlighted as being more beneficial, but diet and several other lifestyle factors were not assessed. As Grønbaek and colleagues have pointed out in further study of other populations, 19,22 Danish wine drinkers have twice the intake of fruit and vegetables than drinkers of beer or spirits. In the study from Germany, where beer is the most common beverage and is presumably consumed in a regular fashion as part of a "healthy lifestyle," diet was also not accounted for. In work from several of our large cohort studies, we have found that diet and related lifestyle factors may confound the association between beverage choice and coronary heart disease and, in some cases, strongly modify the benefits attributed to moderate alcohol consumption. 8,15,23 In future studies, investigators should assess other lifestyle factors, such as diet, which frequently co-vary with beverage preference. Brenner et al incorporated in their analysis of alcohol and coronary heart disease several potential biological mediators of the effect of alcohol and coronary heart disease. In this type of study, first described by Criqui et al24 and Langer et al, 25 researchers can estimate the proportion of the association between alcohol and CHD attributable to one or a set of biological markers. These useful and enlightening studies provide a link between the metabolic studies that show short-term effects of ethanol on biomarkers and the long-term prospective studies that demonstrate an inverse association between alcohol and CHD. After controlling for HDL-lipoprotein cholesterol, apolipoproteins, and fibrinogen, the odds of CHD associated with alcohol consumption was attenuated from 0.55 (95% CI = 0.37–0.83) to 0.76 (95% CI = 0.48–1.20). As with other studies, 24–28 approximately 50% to 70% of the benefits of moderate alcohol consumption can be attributed to these biologic parameters. In a meta-analysis of experimental studies of alcohol consumption, we used regression analysis to predict the reduction of risk in CHD that would be anticipated based on changes in lipid and hemostatic factors caused by drinking 30 g/day of ethanol. We found that at least half of the beneficial effects of moderate alcohol on CHD were explained by elevation of HDL-lipoprotein cholesterol. A further 20% to 30% was attributed to beneficial changes in hemostatic factors. We need to explore in much more detail the cellular and molecular basis for these effects. 29–31 Much work is also needed to elucidate other factors besides the standard lipid and hemostatic parameters usually assessed. Recent evidence suggests ethanol may be important in modifying inflammatory response, 32 insulin sensitivity, 33 and nutrient profile. 18 The evidence indicates that the association between moderate alcohol consumption and lower risk of coronary heart disease is causal and that abstaining from alcohol could be considered a risk factor for coronary heart disease. As epidemiologists, we need to move on to expand our knowledge by incorporating detailed assessments of drinking patterns, by exploring new genetic and biological parameters that may influence or be influenced by the effects of ethanol, and by collaborating with molecular biologists and geneticists to obtain a deeper understanding of the true biological effects of alcohol consumption.