Should Patients With Heart Disease Drink Alcohol?

Numerous epidemiological studies have shown that light to moderate drinkers of alcohol are at lower risk than abstainers for fatal or nonfatal coronary heart disease (CHD). A recent meta-analysis¹ that included 51 (43 cohort) studies estimated a 20% risk reduction for consumption of 0 to 20 g of alcohol (0-2 drinks) per day and some risk reduction for intake up to 72 g (6 drinks) per day. The lower CHD risk in drinkers also has been observed in a wide variety of patient populations, including those with diabetes,^{2 - 3} hypertension,^{4 - 5} and prior myocardial infarction (MI).⁶ The consistency of these findings and the growing evidence that alcohol might protect against CHD via higher high-density lipoprotein (HDL) cholesterol levels, antithrombotic actions, or reduced insulin resistance argue for a causal protective effect of moderate drinking.^{7 - 11} Genetic factors probably play an important role, for example, moderate drinkers who are homozygous for the ADH3 allele (associated with slower rate of ethanol metabolism) have higher HDL levels and decreased risk of MI.¹²

Residual controversy about the benefits of alcohol in reducing CHD risk persists on the basis that correlates of abstinence and lighter drinking could explain the higher risk of CHD among abstainers. Without data from randomized trials, these doubts cannot be resolved completely. Some prospective population-based observational studies separated lifelong abstainers from former drinkers, carefully controlled for baseline CHD risk, or both. These studies attempted to control for established CHD risk factors and other possible confounders, including dietary habits, exercise, and demographic and psychosocial traits. Thus, there appears to be little need for additional observational studies in general population groups.

In contrast, research examining risks or benefits of alcohol consumption in patients with CHD is needed to assist clinicians in advising certain patients. The prospective analysis of moderate drinking and mortality after acute MI in this issue of THE JOURNAL by Mukamal et al¹³ is thus important. Compared with abstainers, patients who reported alcohol intake in the year prior to MI had reduced total and cardiovascular mortality during a median follow-up of nearly 4 years. The 20% to 30% lower mortality risk among drinkers was similar to the estimate from the meta-analysis.¹ Although there were major pre-MI risk differences related to alcohol use, with abstainers at higher risk than drinkers, rigorous attempts to control for these differences did not eliminate the reduced risk of moderate drinkers. The authors concede the impossibility of ruling out residual confounding in these observational data and point out the limitation imposed by knowledge of only pre-MI drinking. It is also possible that drinkers, who had higher socioeconomic status, might have been more compliant with post-MI treatment and preventive practices.

Risk factors for CHD generally are operative both before and after evident disease, often with greater absolute risk reduction after MI. Specifically, the biological mechanisms by which alcohol might protect are likely to have an effect after as well as before MI. One report of the alcohol-mortality relationship after MI showed virtually identical reduced risk (about 20%) in those with or without a history of MI.⁶ Another study showed a slight nonsignificant (about 15%) lower risk of CHD death.¹⁴ All considered, it appears that moderate drinkers fare better than abstainers after MI.

Moderate drinking may carry lower risk of other vascular conditions, including ischemic stroke^{10 - 11} and peripheral vascular disease,¹⁵ and may be associated with lower incidence of type 2 diabetes,^{16 - 17} a powerful predictor of atherosclerotic disease and of heart failure. However, heavy drinking (>3 drinks per day) increases the risk of other cardiovascular problems, including hypertension, arrhythmias, hemorrhagic stroke, and cardiomyopathy.^{7 - 11} Chronic heavy drinking causes myocardial toxicity in susceptible persons, leading to structural and functional impairment and, in some instances, heart failure. Although alcoholic cardiomyopathy does not appear related to the alcohol-CHD relationship or to moderate drinking,¹⁸ this consequence of alcohol abuse has appropriately led to great caution when giving advice about alcohol consumption to patients with heart failure or left ventricular dysfunction.

Heart failure syndrome, increasingly common in the aging population, can be caused by several conditions with relationships to alcohol—hypertension, CHD, and cardiomyopathy. Of these, only CHD is inversely related to moderate drinking but is the most common etiology of heart failure in developed countries. Accordingly, the prospective study of moderate drinking and heart failure risk reported by Abramson et al¹⁹ in this issue of THE JOURNAL has both scientific interest and practical importance. In this study of 2235 community-based older persons, increasing levels of moderate alcohol consumption were associated with decreasing risk of heart failure. Almost 50% less heart failure was found in those reporting at least 1.5 drinks per day, with only minor change when clinical CHD and several CHD risk traits were controlled in this analysis. These data suggest that reduced risk of heart failure in moderate drinkers is not entirely mediated by reducing clinically evident CHD. The authors acknowledge that preclinical CHD cannot be ruled out as an explanation. In a study showing apparent benefit of moderate drinking in persons with left ventricular systolic dysfunction, the prognostic benefit was clear only for those with CHD, but no harm was evident in those without CHD.²⁰ Other explanations offered by Abramson et al¹⁹ include favorable modulation of several neurohormonal factors, such as norepinephrine, arginine vasopressin, and atrial natriuretic peptide. To these might be added possible lowered insulin resistance in moderate drinkers,¹¹ a key part of the "metabolic syndrome" associated with adverse cardiovascular outcomes. Whatever the mechanism(s), and whether reduced CHD related to moderate drinking is responsible, the study by Abramson et al suggests that such drinking appears safe or beneficial for older individuals and probably not harmful in patients with heart failure.

Moreover, the studies by Abramson et al and Mukamal et al show no differences in reduced risk related to choice of wine, liquor, or beer, thereby strengthening the probability that alcohol is the protective agent. This finding also supports the opinion that there is no consensus that any beverage type is clearly superior for cardiovascular benefit.²¹

How should these studies influence advice given by health professionals? Specifically, should persons who have had MI or those with left ventricular dysfunction be advised to drink small amounts of alcohol? The reports in this issue, plus other studies directly or indirectly supporting lower risk of heart failure or death after MI associated with moderate drinking, seem to support a "yes" response. Importantly, these studies provide no evidence suggesting harm to these persons by moderate drinking. In the area of advice, the need for great caution is universally recognized. One widely used "safe limit" definition is no more than 2 drinks per day for men or 1 per day for women,^{8 ,11} amounts associated with evidence of lower risk of CHD or heart failure. The number and size of drinks comprising the safe limit should always be specified. Advice must avoid inducement of heavy drinking or even a basis for rationalization of such behavior.

Of course, moderate drinking carries risk of progression to problem drinking. Unresolved issues with respect to moderate drinking include the association with risk of fetal alcohol syndrome, hemorrhagic stroke, and large bowel cancer, and, perhaps most worrisome, the link with possible increased risk of breast cancer²² (especially since women <50 years old are generally at very low CHD risk). On the other hand, older persons with a history of MI or who are at high risk of heart failure represent population subsets likely to derive a high benefit-risk ratio from moderate drinking.

The need to individualize advice about health effects of drinking is an increasingly consistent theme in the medical literature.^{16 ,23 - 24} For each patient, age, sex, family history, personal drinking history, and specific medical history all must be known to determine the individual benefit-risk equation for alcohol consumption. From the health viewpoint, the only easy rules are that heavy drinkers would be better off to reduce drinking or abstain, and that all persons, whether abstainers or light to moderate drinkers, should be warned to avoid heavy drinking. In light of current knowledge, it seems that established moderate drinkers with a history of MI or decreased left ventricular function generally should not be advised to abstain from alcohol. The data do not justify advising lifelong nondrinkers with these conditions to start drinking for health, especially because most have good reasons for abstinence. There may be exceptions when clinicians might advise certain abstainers to resume alcohol consumption. For instance, former light to moderate drinkers who quit drinking because of an acute coronary syndrome or heart failure but who then recover or stabilize might benefit from small amounts of alcohol. The studies by Mukamal et al and Abramson et al add important information to the evidence on alcohol and health and should be helpful to clinicians in making individualized, judicious recommendations about alcohol drinking for patients with heart disease.