The Alcohol Paradox

Abstract & Commentary

By Jonathan Abrams, MD, Professor of Medicine, Division of Cardiology, University of New Mexico, Albuquerque. Dr. Abrams serves on the speaker's bureau for Merck, Pfizer, and Parke-Davis.

Synopsis: Moderate alcohol consumption is associated with a decreased risk for MI and heart failure.

Sources: Djoussé L, Gaziano JM. Alcohol Consumption and Risk of Heart Failure in the Physicians' Health Study I. Circulation. 2007;115:34-39. Beulens JW, et al. Alcohol Consumption and Risk for Coronary Heart Disease among Men with Hypertension. Ann Int Med. 2007;146:10-19.

Many epidemiologic- and population-based studies have convincingly shown that mild to moderate alcohol consumption (no more than 2 drinks per day) is associated with decreased cardiac mortality and morbidity. However, physicians and guidelines do not support the use of alcohol in secondary or primary prevention. Two pertinent studies have recently been reported; one in a large population of hypertensive men, the other in a cohort of individuals without congestive heart failure. The results, not surprisingly, support the many previous studies and databases confirming that mild to moderate alcohol consumption on a regular basis is cardio-protective. One report comes from the Physicians Health Study I (PHS I), which enrolled 22,000 males, free of heart failure, followed prospectively from 1982 to 2005 with annual questionnaires. Participants in the PHS I longitudinal study received serial questionnaires about alcohol consumption; those with incident heart failure were contacted every 6 months. Frequent alcohol consumption was associated with older age, smoking, hypertension, and atrial fibrillation. Over an average period of 18.4 years, 904 incident cases of HF (heart failure) occurred, with a decreased incidence in men consuming more than 7 drinks per week. The hazard ratio for these moderate drinkers was 0.62 (0.41-0.96, p = 0.012 with adjustments for multiple risk factors). Of note, there was no significant association between HF and moderate alcohol consumption in HF patients who did not have underlying coronary artery disease. The incidence rates of heart failure were highest for zero consumption; those in the highest category of alcohol consumption had the greatest event-free survival. Adjusted hazard ratios were 0.88-1.05 for no drinking, 0.80 (0.68-0.94) for 1-4 drinks per week, and 0.62 (0.4-0.95) for greater than 7 drinks per week. Myocardial infarction incidence tracked with alcohol use rates; the lowest hazard ratio occurred in the highest-alcohol-consumption group. Multivariate analysis demonstrated comparable results, with relative risk of 1.1, 0.94, and 0.69 from the lowest to highest alcohol groups, p = 0.003. Prior myocardial infarction and coronary artery disease correlated with heart failure far better than those with no history of coronary artery disease. In fact, HF without antecedent myocardial infarction had a weaker inverse association with alcohol. The authors point out that the lifetime risk of heart failure is estimated to be 20% in individuals over 40 years of age.

These results are concordant with data from the Kaiser Permanente Cardio-Vascular Health Study, which showed a 40% lower risk of coronary artery disease related heart failure and much less for non-coronary artery disease HF. Other reports have been less robust. The authors identify a number of biologic mechanisms responsible for these observations, including beneficial effects on HDL, insulin sensitivity, inflammation, endothelial function, coagulation, and atrial natriuretic peptide. No differences were found between the different types of alcoholic beverages. The authors conclude that their data shows "an inverse association between moderate alcohol consumption and incident heart failure." They speculate that alcohol use "may lower the risk of heart failure, especially in coronary artery disease-related heart failure."

The second study was of alcohol use in men with hypertension and came to similar conclusions. These data are from the Health Professional Follow-Up study (HPFS), a group of almost 12,000 men with hypertension studied between 1986 and 2002. The enrollees were all health professionals, although physicians were excluded. Subjects were categorized by preexisting hypertension or the development of hypertension after entering the study. Overall, 9,000 men with hypertension were diagnosed during the follow-up period, 2,700 men had hypertension at baseline. Thus, 11,700 patients with were eligible for this observational study. Patients were comprehensively evaluated by extensive questionnaires regarding alcohol consumption. Ethanol intake was calculated as were the number of days per week that individuals typically drank. Estimated mean alcohol intake was 12.5 grams per day (one drink). Primary end-points were incident non-fatal MI and fatal coronary heart disease (CHD) or stroke. CHD risk factors were assessed. Alcohol consumption and other dietary variables were updated every 4 years and anti-hypertensive medication was tracked. Of interest, only liquor consumption was associated with improved CHD outcomes. Approximately half of the study population were on hypertensive medications, more with greater alcohol use. Heavier alcohol consumption was associated with smoking and a lower prevalence of diabetes. Liquor and beer were used in the greatest quantity. During follow up of 16 years there were 653 myocardial infarctions (MI) (279 fatal, 374 non fatal). Low-quantity drinkers were comparable to abstainers. There was an approximate dose-dependant-inverse relationship between alcohol consumption and the risk for MI in adjusted models. The hazard ratio for consumption of 15-30 grams of alcohol per day was 0.70 (CI, 0.53-0.97). There was an inverse association for MI as low as 0.51 for consumption of 10-15 grams per day. There was no significant association for ischemic stroke. Liquor consumption was most strongly related to a lower risk for MI, (hazard ratio 0.59). Hazard ratio for red wine was 1.01, and 0.89 for white wine. Drinking frequency was inversely associated with risk for MI, with an adjusted hazard ratio of 0.64 for daily drinking. Alcohol consumption decreased somewhat after the initial diagnosis of hypertension, but 80% of men did not alter the intake by more than an average of one half drink per day following hypertension diagnosis.

In summary, their analysis of 12,000 men with hypertension indicates that alcohol use was associated with a decreased risk for fatal and non fatal MI but not for total death or CHD death. There was no statistical increase in other medical conditions, particularly cancer. The authors suggest that "approximately one half of the association of alcohol consumption with CHD is mediated by an increase in HDL." Controlling for age, smoking, BMI, etc., did not change the results. The authors conclude that moderate alcohol consumption is associated with a decrease risk for MI in men with or without hypertension, with no adverse affect from alcohol. "Men with hypertension who drink moderate and safely may not need to change their drinking habits."


These data are reassuring for those of us who like to have one or 2 glasses of wine on a daily basis. It has been long recognized and recommended that 1-2 drinks per day is a healthy maximum, and these reports robustly confirm these recommendation. In fact, in the modest number of individuals who were heavy drinkers, there appeared to be little adverse health effects (although benefits seen for moderate drinking were greater). These 2 very large data sets utilized careful and detailed statistical methods. Confounding information could be missed, as the alcohol intake records were done by repeated questionnaire. Nevertheless, there appears to be no "smoking gun" that would be responsible for these results being false or misleading.

It has long been a dilemma as to whether physicians should recommend a daily glass of wine to high-risk or overt CHD individuals. The concern about alcoholism has so far overpowered the argument that such a recommendation should be made by the physician. Nevertheless, it seems that patients who are well known to their doctor, have no addictive behaviors, and who are most interested in maximizing health benefits, could be safely advised of the data in the literature, including these 2 studies, which confirm a rather striking benefit in habitual, low-quantity drinkers, particularly if HDL is low. In addition to the general population studies of alcohol consumption, we now know these data reflect a lower likelihood of congestive heart failure in patients who have underlying coronary disease, as well as clear-cut benefits for individuals who have or will develop hypertension. The alcohol paradox remains and is unlikely to disappear. Perhaps it is reasonable to revisit our policies and guidelines with a multi-representative task force to see if some of the benefits of alcohol can be safely steered toward individuals at risk for CHD.