The Epidemiology of Alcohol and Cardiovascular Diseases | to pdf >>

by Arthur L. Klatsky, MD

Introduction
Disparity in the relation of alcohol consumption to various cardiovascular (CV) conditions has become evident.¹ Heavier drinking is related to higher prevalence of cardiomyopathy (CM), hypertension (HTN), hemorrhagic stroke, and cardiac arrhythmias. Lighter drinking is related to lower prevalence of coronary artery disease (CAD), ischemic stroke, and sudden cardiac death. The composite of these relations in several population studies of overall CV mortality is a U-shaped curve (lighter drinkers at lower risk than abstainers or heavier drinkers), although several other studies show all drinkers, lighter and heavier, at lower CV mortality risk than abstainers. Increased non-CV mortality among heavier drinkers is found in all studies, with a J-curve (heavier drinkers at highest; lighter drinkers at lowest risk) for the all-cause alcohol-mortality relation.

Definitions of Moderate and Heavy Drinking
Any definition of moderate drinking is arbitrary. The operational definition here used is based upon the level of drinking in epidemiologic studies above which net harm is usually seen. Thus, less than three drinks per day is called "lighter" or "moderate" drinking, and three or more drinks per day, "heavy" drinking. Sex, age, and individual factors lower the upper limit for some persons and raise it for others. In data based upon surveys, systematic "underestimation" (lying) probably tends to lower the apparent threshold for harmful alcohol effects.

Fortunately, the amount of alcohol in a standard-sized drink of wine, liquor, or beer is approximately the same. Since people think in terms of "drinks," not milliliters or grams of alcohol, it seems to this author best to describe alcohol consumption in terms of drinks per day or week. When talking with patients, health professionals should always remember the importance of defining the size of drinks.

Alcoholic Cardiomyopathy (ACM)
The concept of an independent, direct cardiotoxic effect of alcohol has become accepted.¹ The circumstantial evidence is substantial, but the absence of specific diagnostic tests seriously impedes epidemiologic study. Alcohol-associated CM cannot be distinguished clinically or pathologically from dilated CM of unknown cause(s). Historical episodes suggest synergistic myocardial toxicity of alcohol with arsenic and cobalt; other cofactors in alcoholic heart disease remain speculative. A role for thiamine deficiency in low-output chronic heart failure has never been established, although an interaction with alcohol cardiotoxicity might exist in malnourished persons.

The most convincing circumstantial evidence for ACM is the extensive data, in animals and humans, of nonspecific cardiac abnormalities related to alcohol. These include structural abnormalities in autopsy and biopsy studies and demonstration of acute and chronic functional and metabolic derangements by several techniques. A possible nonoxidative metabolic pathway for alcohol has been reported by Laposata and Lange² in the heart, muscle, pancreas, and brain, related to fatty acid metabolism. Accumulation of fatty acid ethyl esters was shown to be related to blood alcohol levels and to mitochondrial metabolism. A report by Urbano-Marquez et al³ showed in alcoholics a clear relation of lifetime alcohol consumption to structural and functional myocardial and skeletal muscle abnormalities. The amounts of alcohol were largethe equivalent of 120 grams alcohol per day for 20 years.

As of 1997, a majority of all cases of CM are considered to be of unknown cause. The proportion of CM cases attributed to alcohol varies markedly in reports, probably due mostly to differences in the alcohol consumption habits of the populations under study. Thus, recent reports include alcohol-attributable proportions ranging from 3.4% at Johns Hopkins Hospital⁴ to 41.9% at the Philadelphia VA Hospital.⁵

The lack of specific diagnostic tests for ACM necessitates exclusion of other CV conditions for diagnosis. However, the probability of synergistic damage includes possibly enhanced alcohol cardiotoxity in the presence of other myocardial damage. For this reason, persons with heart muscle impairment or major arrhythmias should be especially strongly advised to limit alcohol intake to less than three drinks per day.

Hypertension (HTN)
An association between heavier alcohol consumption and HTN reported by Lian in French servicemen in 1915⁶ was largely ignored for the next 60 years. Since the mid-1970s, largely because of epidemiologic studies in developed countries, alcohol ingestion has joined other correlates of hypertension, such as obesity and salt intake, as a major focus in research about possible HTN risk factors. An alcohol-HTN link has been shown in almost all of more than 50 cross-sectional and 10 prospective population studies in ambulatory persons in a number of countries.^7,8 Studies differ about whether the alcohol-HTN link is linear or nonlinear in men (i.e., is a consumption threshold present?); in women, the curve which represents the relation is J-shaped, or present only at higher alcohol intake. Studies of hospitalized alcoholics or problem drinkers have been conflicting with respect to HTN. It is possible that chronic alcohol-related conditions such as malnutrition, cirrhosis, and cardiomyopathy lower blood pressure in some persons.

Two Kaiser Permanente studies^9,10 are among the largest of the cross-sectional population surveys. The first⁹ showed a J-curve in women and a threshold relation in men, with higher blood pressures at three or more drinks per day in both sexes (Figure 1). The findings were independent of age, sex, and race and, by direct cross-classification (examination of the alcohol-HTN relationship in population subcategories), of smoking, coffee intake, reported past heavy drinking, education, adiposity, and habitual salt use. HTN (greater than 160/95 mmHg) prevalence was doubled in white men and women reporting consumption of six or more drinks per day. The second Kaiser Permanente study¹⁰ showed similar findings in an analysis adjusted simultaneously for age, adiposity, smoking, coffee, tea, and seven blood tests (Figure 2). Ex-drinkers did not have higher blood pressure than lifelong abstainers. Study of drinking variability and intake in the week before examination suggested rapid regression of alcohol-associated HTN with abstinence.

Figure 1. Mean systolic blood pressures (upper half) and mean diastolic blood pressures (lower half) for white, black, or Asian men and women with known drinking habits. Small circles represent data based on fewer than 30 persons. (From Klatsky AL, Friedman GD, Siegelaub AB, Gerard MJ. Alcohol consumption and blood pressure. Kaiser Permanente Multiphasic Examination data. N Engl J Med 1977;296:1194-1200. Used by permission.)

Figure 2. Adjusted mean systolic and diastolic blood pressures (mmHg) according to alcohol consumption by three age groups (top left, white men; lower left, black men; top right, white women; lower right, black women). Dashed lines and open circles indicate 10 < n < 25. Data omitted from figure for categories with n < 11 (white women aged 40 to 59 years, nine or more drinks per day and aged > 59 years, six to eight and nine or more drinks per day; black men aged > 59 years, six to eight and nine or more drinks per day; black women aged 40 to 59 years, nine or more drinks per day and aged >59 years, three to five, six to eight, and nine or more drinks per day). (From Klatsky AL, Friedman GD, Armstrong MA. The relationship between alcoholic beverage use and other traits to blood pressure: A new Kaiser Permanente study. Circulation 1986;73:628-636. By permission of the American Heart Association, Inc.

Several intervention studies suggest a short-term (develops in days to several weeks) pressor effect of three to eight alcoholic drinks per day, and decreases in blood pressure upon abstention or marked reduction in alcohol intake.⁸ No elevations of blood pressure due to withdrawal have been seen in these studies. A few studies present data showing independence of the alcohol-blood pressure association from intake of salt, physical activity, and psychosocial stress. Even without confirmation in long-term trials, the intervention studies support a cause-effect relation between alcohol intake and HTN. Estimates of possible population-attributable risk (proportion of HTN due to alcohol) range from 5% to 30%.⁸ Even if only 5% of HTN is attributable to alcohol, this may be the commonest cause of reversible HTN in developed societies.

The inconsistent acute effects of alcohol on blood pressure as reported in human and animal studies may not be directly relevant to the epidemiologic relation in humans.⁸ There is no known animal model for chronic studies. There is no proof of a sustained effect in humans via the renin-angiotensin mechanism, cortisol, catecholamines, increased cardiac output, "hypermetabolic state," central nervous system actions, or autonomic nervous system effects. A recently reported experiment¹¹ in normal humans used intraneural microelectrodes to demonstrate increased sympathetic activation in response to I.V. alcohol with a delayed (second hour) blood pressure rise. Inhibition by dexamethasone suggested a central mechanism via corticotropin-releasing hormone. There is some current interest in a possible direct effect upon peripheral vascular tone via a calcium transport mechanism. Explanations for the alcohol-HTN association remain speculative; this fact is the major deficiency in the case for causality. Studies of HTN sequelae (coronary disease, stroke, congestive heart failure, renal insufficiency, etc.) are greatly complicated by the independent relations of alcohol use to several common hypertension sequelae.⁸

It is likely that the alcohol-HTN link is causal. Reduction of intake in some heavier drinkers is probably therapeutic, and avoidance of heavier drinking will probably prove to have an important role in primary prevention of HTN.¹²

Coronary Artery Disease (CAD)
Data showing that major CAD events are more likely to develop in abstainers than in alcohol drinkers include international comparisons, time-trend analyses, case-control studies, and longitudinal studies.^13,14 Most studies of CAD hospitalizations show heavier drinkers have a risk of CAD hospitalization similar to or lower than that of lighter drinkers (i.e., no U-shaped curve). Several population studies using CAD mortality as an endpoint also show a progressive inverse relation to amount of alcohol consumption, but others show a U-shaped curve. Those studies which separate lifelong abstainers from past drinkers suggest that both subsets of nondrinkers are at higher risk of CAD than drinkers, but some would still dispute this. Many population studies were not able to distinguish these subsets of nondrinkers. Where available, data about choice of type of alcoholic beverage suggest that beverage choice is a minor factor in CAD risk. Studies of sudden cardiac death, due mostly to CAD, also show an inverse relation to alcohol use.

There are plausible mechanisms by which alcohol drinking might protect against CAD.^13,14 These include a favorable effect on HDL cholesterol concentration (an increased level), a similarly favorable effect upon apolipoproteins, and an antithrombotic action. Controversy about protection persists, however, on the grounds that correlates of abstinence and lighter drinking could explain the higher risk of abstainers. For example, a much publicized hypothesis advanced by Shaper et al¹⁵ suggested that movement of persons at high CAD risk into the abstainer referent group could explain the U-shaped curve shown in their work and in that of other investigators.

A prospective Kaiser Permanente study of alcohol habits in relation to CAD hospitalizations¹⁶ showed that ex-drinkers and infrequent (less than 1/month) drinkers were at a risk similar to that of lifelong abstainers. A lower CAD risk was present among all other drinkers with no U-shaped curve, independent of a number of potential confounders (Table 1). These relations were independent of base line CAD risk at examination (Table 2) and beverage choice. The data suggested a protective effect of alcohol against risk of hospitalization for CAD.

In a prospective Kaiser Permanente study of total CV mortality,¹⁷ ex-drinkers had higher age-adjusted CAD and overall CV mortality risk than lifelong abstainers, but the difference disappeared when adjusted for other traits. Among drinkers, there were U-shaped mortality curves relating amounts of alcohol and both CVD and CAD, with a nadir at one to two and at three to five drinks per day. Subsets free of baseline risk had similar alcohol-CAD and alcohol-CV mortality curves. The study demonstrated the expected disparities between alcohol and various CV conditions (Table 3). A number of features of the analysis argued against a spurious inverse alcohol-CAD relation, including:
  1) independence from CAD risk at baseline examination;
  2) absence of higher CAD risk among persons reducing alcohol intake for medical reasons;
  3) evidence that the higher unadjusted CAD risk of ex-drinkers is due to confounding;
  4) absence of a relation among ex-drinkers between CAD risk and maximal past intake;
  5) absence of a relation between infrequent (less than 1/month) drinking and CAD risk;
  6) similar reduction of CAD risk among drinkers of wine, liquor, and beer.

Another large prospective study among women free of CAD at examination¹⁸ showed a progressive inverse relation of alcohol use to major CAD events, independent of prior reduction in alcohol intake and of nutrient intake (the latter was analyzed in detail). The relative risk of CAD events in women reporting daily alcohol intake of 25 or more grams per day was 0.4, similar to the findings for women in the Kaiser Permanente study. Further analysis of these data in women¹⁹ demonstrated that net beneficial effects of moderate alcohol use in women was limited by adverse effects to persons clearly at above-average CAD risk (i.e., those above 50 years of age).

Large prospective studies in men also confirm the lower CAD risk of drinkers, independent of confounders or disease when alcohol habits were determined.^20,21 The American Cancer Society Study²⁰ was a 12-year prospective mortality study of 276,802 white men; there was a U-shaped curve for CAD mortality, with a RR of 0.8 (vs. abstainers) at one to two drinks per day. The Health Professional Followup Study of 51,529 men²¹ was well controlled for dietary habits; newly diagnosed CAD was inversely related to increasing alcohol intake. A study in both sexes, the Auckland Heart Study,²² was designed to study the hypothesis that persons at high CAD risk are likely to become nondrinkers; the analysis showed that moderate drinkers had lower CAD risk than both lifelong abstainers and ex-drinkers, thus supporting the hypothesis that alcohol protects against CAD.

Reduced risk of CAD is present at various ages, although the impact upon total mortality in a Kaiser Permanente study was clearest in older age brackets and the adverse effects of alcohol were greater among younger persons.²³ Among persons > 60 years of age, overt or latent CAD may play a role in risk of death from causes other than CAD.²³

The hypothesis that the apparent protective effect of alcohol against CAD is mediated by higher HDL cholesterol levels in drinkers has been examined quantitatively in three separate studies.^24-26 All three analyses yielded similar findings suggesting that higher HDL levels in drinkers mediated about half of the lower CAD risk. One of these studies²⁶ suggests that both HDL2 and HDL3 are involved. HDL3 may be more strongly related to lighter alcohol intake but is probably related as strongly as HDL2 to lower CAD risk. There are no similar data about protective mechanisms other than the HDL link, but some data support several possible antithrombotic mechanisms.^13,14,26 Thus, multiple mechanisms may play a role.

International comparison studies^27-29 suggest that wine confers more protection against CAD than beer or liquor. The "French paradox" concept has arisen from these data; it refers to the fact that France tends to be an outlier on graphs of mean dietary fat intake vs. CAD mortality, unless adjusted for wine alcohol intake.^28,29 Reports of nonalcohol antioxidant phenolic compounds^30-32 or antithrombotic substances^33-36 in wine, especially red wine, have appeared. Inhibition of oxidative modification of low-density-lipoprotein cholesterol is probably anti-atherogenic, although prospective clinical trials of antioxidant supplements are not yet conclusive.³⁷ Thus, antioxidant substances in wine are an attractive hypothetical explanation for CAD protection. However, the prospective population studies provide no consensus that wine has additional benefits, and various studies show benefit for wine, beer, liquor, or all three major beverage types.^13,14 In Kaiser Permanente studies, all three major beverage types show evidence of protection against CAD^16,17; wine drinkers fare best with respect to CAD mortality, but drinkers of red and non-red wine fare equally well.³⁸ Because the beverages differ in user traits, with wine drinkers having the most favorable CAD risk profile,³⁹ a noncausal explanation was favored for the lower CAD risk of wine drinkers. Drinking-pattern differences among the beverage types are another hypothetical factor. The wine/liquor/beer issue is unresolved at this time, but it seems likely that ethyl alcohol is the major factor with respect to lower CAD risk.

It remains theoretically possible that lifelong abstainers could differ from drinkers in psychological traits, dietary habits, physical exercise habits, or some other way which could be related to CAD risk, but there is no good evidence for such a trait. The various studies indicate that such a correlate would need to be present in persons of both sexes, various countries, and multiple racial groups. Although it remains possible that other factors play a role, a causal, protective effect of alcohol is a simpler and more plausible explanation.^13,14,40,41

Cerebrovascular Disease
Several reports suggest that alcohol use, especially heavier drinking, is associated with higher risk of stroke. Some studies examined only drinking sprees; others did not differentiate between hemorrhagic and occlusive strokes. Several studies have suggested that alcohol was related only to hemorrhagic stroke. The Nurse's Health Study¹⁸ showed drinkers to be at higher risk of subarachnoid hemorrhage but at lower risk of occlusive stroke.

A Kaiser Permanente study looked at the relations between reported alcohol use and the incidence of hospitalization for several types of cerebrovascular disease.⁴² Daily consumption of 3 or more drinks, but not lighter drinking, was related to higher hospitalization rates for hemorrhagic cerebrovascular disease, especially intracerebral hemorrhage. Higher blood pressure appeared to be a partial mediator of this relation. Alcohol use was associated with lower hospitalization rates for occlusive cerebrovascular disease; an inverse relation was present for both sexes, for whites and blacks, and for extracranial and intracerebral occlusive lesions (Table 4). The data suggest that heavier drinking increases the risk of hemorrhagic cerebrovascular events but that alcohol use may lessen the risk of occlusive lesions.

At this time there is no consensus about the relations of alcohol drinking to the various types of cerebrovascular disease and agreement only that more study of this important area is needed.⁴³

Cardiac Arrhythmias
Increased ventricular ectopic activity has been documented after ingestion of substantial amounts of alcohol, although epidemiologic studies have not shown a higher risk of sudden death in drinkers. Various atrial dysrhythmias have been reported to be associated with spree drinking. A Kaiser Permanente study⁴⁴ compared atrial dysrhythmias in 1,322 persons reporting six or more drinks per day to dysrhythmias in 2,644 light drinkers. The relative risk in the heavier drinkers was at least doubled for atrial fibrillation, atrial flutter, supraventricular tachycardia, and atrial premature complexes (Table 5).

Conclusion
This brief survey documents the evidence for disparity in the relations of alcohol and CV disorders. Published reviews are available.^1,46 Table 6 summarizes the relations, with emphasis on the disparity between the overall favorable relations of lighter drinking and the overall unfavorable relations of heavier drinking.

Acknowledgment: Some of the material here reported was supported by research performed with a grant from the Alcoholic Beverage Medical Research Foundation, Baltimore, MD.

References:
1. Klatsky AL. Cardiovascular effects of alcohol. Sci Am Sci Med 1995;2:28-37.
2. Laposata EA, Lange LG. Presence of nonoxidative ethanol metabolism in human organs commonly damaged by ethanol abuse. Science 1986;231:497-99.
3. Urbano-Marquez A, Estrich R, Navarro-Lopez F, Grau JM, Mont L, Rubin E. The effects of alcoholism on skeletal and cardiac muscle. N Engl J Med 1989;320:409-15.
4. Kasper EK, Agema WR, Hutchins GM, Deckers JW, Hare JM, Baughman KL. The causes of dilated cardiomyopathy: clinicopathologic review of 673 consecutive patients. J Am Coll Cardiol 1994;23:586-90.
5. Wang RY, Alterman AI, Searles JS, McLellan AT. Alcohol abuse in patients with dilated cardiomyopathy: laboratory vs clinical detection. Arch Intern Med 1990;150:1079-82.
6. Lian C. L'alcoholisme cause d'hypertension artérielle. Bull Acad Med (Paris) 1915;74:525-28.
7. Klatsky, AL. Blood pressure and alcohol consumption. In: Izzo, JL, Black, HL, eds. Hypertension Primer. Dallas: American Heart Association; 1993:164-76
8. Klatsky AL. Blood pressure and alcohol intake. In: Laragh JH, Brenner BM, editors. Hypertension: pathophysiology, diagnosis and management. 2nd ed. New York: Raven Press; 1995:2649-67.
9. Klatsky AL, Friedman GD, Siegelaub AB, Gerard MJ. Alcohol consumption and blood pressure. Kaiser-Permanente Multiphasic Health Examination data. N Engl J Med 1977;296:1194-200.
10. Klatsky AL, Friedman GD, Armstrong MA. The relationships between alcoholic beverage use and other traits to blood pressure: a new Kaiser Permanente study. Circulation 1986;73:628-36. 11. Randin D, Vollenweider P, Tappy L, Jequier E, Nicod P, Scherred U. Suppression of alcohol-induced hypertension by dexamethasone. N Engl J Med 1995;332:1733-37.
12. Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure (JNC-V). The fifth report of the Joint National Committee. Arch Intern Med 1993;153:154-83.
13. Klatsky AL. Epidemiology of coronary heart disease: influence of alcohol. Alcohol Clin Exp Res 1994;18:88-96.
14. Renaud S, Criqui MH, Farchi G, Veenstra J. Alcohol drinking and coronary heart disease. In: Verschuren PM,ed. Health issues related to alcohol consumption. Washington, DC: ILSI Press; 1993:81-124.
15. Shaper AG, Wannamethee G, Walker M. Alcohol and mortality in British men: explaining the U-shaped curve. Lancet 1988;2:1267-73.
16. Klatsky AL, Armstrong MA, Friedman GD. Relations of alcoholic beverage use to subsequent coronary artery disease hospitalization. Am J Cardiol 1986;58:710-14.
17. Klatsky AL, Armstrong MA, Friedman GD. Risk of cardiovascular mortality in alcohol drinkers, ex-drinkers, and nondrinkers. Am J Cardiol 1990;66:1237-42.
18. Stampfer MJ, Colditz GA, Willett WC, Speizer FE, Hennekens CH. A prospective study of moderate alcohol consumption and the risk of coronary disease and stroke in women. N Engl J Med 1988;319:267-73.
19. Fuchs CS, Stampfer MJ, Colditz GA, Giovannucci EL, Manson JE, Kawachi I, et al. Alcohol consumption and mortality among women. N Engl J Med 1995;332:1245-50. [published erratum appears in N Engl J Med 1997 Feb 13;336(7):523]
20. Boffetta P, Garfinkel L. Alcohol drinking and mortality among men enrolled in an American Cancer Society prospective study. Epidemiology 1990;1:342-48.
21. Rimm EB, Giovannucci EL, Willett WC, Colditz GA, Ascherio A, Rosner B, et al. Prospective study of alcohol consumption and risk of coronary heart disease in men. Lancet 1991;338:464-468.
22. Jackson R, Scragg R, Beaglehole R. Alcohol consumption and risk of coronary heart disease. BMJ 1991;303:211-16.
23. Klatsky AL, Armstrong MA, Friedman GD. Alcohol and mortality. Ann Intern Med 1992;117:646-54.
24. Criqui MH, Cowan LD, Tyroler HA, Bangdiwala S, Heiss G, Wallace RB, et al. Lipoproteins as mediators for the effects of alcohol consumption and cigarette smoking on cardiovascular mortality: results from the Lipid Research Clinics Follow-up Study. Am J Epidemiol 1987;126:629-37.
25. Suh I, Shaten BJ, Cutler JA, Kuller KH. Alcohol use and mortality from coronary heart disease: the role of high-density-lipoprotein cholesterol. Ann Intern Med 1992;116:881-87.
26. Gaziano JM, Buring JE, Breslow JL, Goldhaber SZ, Rosner B, VanDenburgh M, et al. Moderate alcohol intake, increased levels of high-density lipoprotein and its subfractions, and decreased risk of myocardial infarction. N Engl J Med 1993;329:1829-34.
27. St. Leger AS, Cochrane AL, Moore F. Factors associated with cardiac mortality in developed countries with particular reference to the consumption of wine. Lancet 1979;1:1017-20.
28. Renaud S, de Lorgeril M. Wine, alcohol, platelets, and the French paradox for coronary heart disease. Lancet 1992;339:1523-26.
29. Criqui MH, Ringel BL. Does diet or alcohol explain the French paradox? Lancet 1994;344:1719-23.
30. Siemann EH, Creasy LL. Concentration of the phytoalexin resveratrol in wine. Am J Enol Viticulture 1992;43:1-4.
31. Frankel EN, Kanner J, German JB, Parks E, Kinsella JE. Inhibition of oxidation of human low-density lipoprotein by phenolic substances in red wine. Lancet 1993;341:454-57.
32. Maxwell S, Cruickshank A, Thorpe G. Red wine and antioxidant activity in serum {letter}. Lancet 1994;344:193-94.
33. Fitzpatrick DF, Hirschfield DL, Coffey RG. Endothelium-dependent vasorelaxing activity of wine and other grape products. Am J Physiol 1993;265(2 Pt 2):H774-H778.
34. Kluft C, Veenstra J, Schaafsma G, Pikaar NA. Regular moderate wine consumption for five weeks increases plasma activity of the plasminogen activator inhibitor-1 (PAI-1) in healthy young volunteers. Fibrinolysis 1990;4(suppl 2):69-70.
35. Demrow HS, Slane PR, Folts JD. Adminstration of wine and grape juice inhibits in vivo platelet activity and thrombosis in stenosed canine coronary arteries. Circulation 1995;91:1182-88.
36. Seigneur M, Bonnet J, Dorian B, et al. Effect of the consumption of alcohol, white wine, and red wine on platelet function and serum lipids. J Appl Cardiol 1990;5:215-22.
37. Hoffman RM, Garewal HS. Antioxidants and the prevention of coronary heart disease. Arch Intern Med 1995;155:241-46.
38. Klatsky AL, Armstrong MA. Alcoholic beverage choice and risk of coronary artery disease mortality: Do red wine drinkers fare best? Am J Cardiol 1993;71:467-69.
39. Klatsky AL, Armstrong MA, Kipp H. Correlates of alcoholic beverage preference: traits of persons who choose wine, liquor or beer. Br J Addict 1990 Oct;85:1279-89.
40. Marmot M, Brunner E. Alcohol and cardiovascular disease: the status of the U-shaped curve. BMJ 1991;303:565-68.
41. Friedman GD, Klatsky AL. Is alcohol good for your health? [editorial] N Engl J Med 1993;329:1882-83.
42. Klatsky AL, Armstrong MA, Friedman GD. Alcohol use and subsequent cerebrovascular disease hospitalizations. Stroke 1989;20:741-46.
43. Van Gign J, Stampfer MJ, Wolfe C, Algra A. The association between alcohol consumption and stroke. In: Verschuren PM, ed. Health issues related to alcohol consumption. Washington, DC: ILSI Press; 1993:43-80.
44. Cohen EJ, Klatsky AL, Armstrong MA. Alcohol use and supraventricular arrhythmia. Am J Cardiol 1988;62:971-73.
45. Davidson DM. Cardiovascular effects of alcohol. West J Med 1989;151:430-39.