Critique 155: The effects of alcohol consumption on the risk of developing heart failure — 27 January 2015

Gonḉalves A, Claggett B, Jhund PS, Rosamond W, Deswal A,  Aguilar D, Shah AM, Cheng S, Solomon SD. Alcohol consumption and risk of heart failure: the Atherosclerosis Risk in Communities Study.  European Heart Journal, 2015, pre-publication; doi:10.1093/eurheartj/ehu514.

Authors’ Abstract

Aim  Alcohol is a known cardiac toxin and heavy consumption can lead to heart failure (HF). However, the relationship between moderate alcohol consumption and risk for HF, in either men or women, remains unclear.

Methods and Results  We examined 14 629 participants of the Atherosclerosis Risk in Communities (ARIC) study (54+6 years, 55% women) without prevalent HF at baseline (1987–89) who were followed for 24+1 years.  Self-reported alcohol consumption was assessed as the number of drinks/week (1 drink = 14 g of alcohol) at baseline, and updated cumulative average alcohol intake was calculated over 8.9 ± 0.3 years.  Using multivariable Cox proportional hazards models, we examined the relation of alcohol intake with incident HF and assessed whether associations were modified by sex. Overall, most participants were abstainers (42%) or former drinkers (19%), with 25% reporting up to 7 drinks per week, 8% reporting ≥7 to 14 drinks per week, and 3% reporting ≥14–21 and ≥21 drinks per week, respectively.  Incident HF occurred in 1271 men and 1237 women.  Men consuming up to 7 drinks/week had reduced risk of HF relative to abstainers (hazard ratio, HR 0.80, 95% CI 0.68–0.94, P = 0.006); this effect was less robust in women (HR 0.84, 95% CI 0.71–1.00, P = 0.05).  In the higher drinking categories, the risk of HF was not significantly different from abstainers, either in men or in women.

Conclusion  In the community, alcohol consumption of up to 7 drinks/week at early-middle age is associated with lower risk for future HF, with a similar but less definite association in women than in men.  These findings suggest that despite the dangers of heavy drinking, modest alcohol consumption in early-middle age may be associated with a lower risk for HF.

Forum Comments

It has been known for many decades that excessive alcohol intake can result in alcoholic cardiomyopathy, causing heart failure, and it was previously believed that heart failure, in general, would be made worse by any alcohol consumption. As reviewed by Djousse and Gaziano in 2008, however, most studies have shown that moderate drinkers are at lower risk for  the development of heart failure (Walsh et al)(Djousse and Gaziano 2007)(Abrahmson et al)(Klatsky et al, 2005)(Bryson et al)(Kloner et al)(Cooper et al)(Aquilar et al).  For example, in a population of more than 100,000 subjects, Klatsky and associates found that alcohol drinking was inversely related to risk of heart failure related to coronary heart disease (CHD) (e.g., at one or two drinks per day; RR, 0.6; 95% CI, 0.5-0.7), with consistency across subgroups of age, gender, ethnicity, education, smoking status, interval to diagnosis, and presence or absence of baseline heart disease or systemic hypertension.  In the Klatsky et al study, for heart failure not associated with CHD, moderate drinking was inversely related significantly only in subjects who had diabetes mellitus. 

For patients who already have depressed myocardial function, Cooper and associates showed in the Studies of Left Ventricular Dysfunction (SOLVD) that moderate drinkers had lower subsequent all-cause mortality than did abstainers.  Many studies have provided data supporting such a finding in that people who have had a myocardial infarction who drink moderately afterwards tend to have fewer cardiovascular problems and live longer (e.g., Pai et al).

Some studies have not shown a reduction in the risk of cardiac events from moderate alcohol intake.  For example, in the Survival And Ventricular Enlargement (SAVE) trial in subjects after myocardial infarction, Aguilar and coworkers did not show a significant effect of moderate drinking on the development of heart failure or on survival.

Specific comments on this paper:  Forum members consider that ARIC is a very well-managed, population-based cohort study, and that this paper was well done.  As described by reviewer Finkel, and very similarly by others, “This is a well-done study with adequate statistical power, at least for populations short of heavy drinkers.”  The particular strengths of this study include the fact that it is based on a large biracial community-based cohort with excellent long-term follow up for 24 ± 1 years; the investigators had data on most potential confounders; in sensitivity analyses they accounted for the potential effects related to mortality from non-heart-failure causes; they evaluated both baseline and cumulative average of alcohol intake as exposures; and they used appropriate statistical methods, testing for both linear and curvilinear relations between alcohol and heart failure. 

In their analyses, the authors adjusted for a previous history of coronary artery disease (CAD), but did not present data specifically comparing results from subjects who did versus those who did not develop a myocardial infarction or other evidence of CAD during follow up.  At baseline, there were relatively few subjects with previously diagnosed CAD in this cohort, but it is unclear from the data presented how the development of CAD during follow up affected the relation of alcohol intake to the risk of HF.

Overall 19% of subjects were classified as “former drinkers,” but how many of these were previous heavy drinkers (versus former light or occasional drinkers) is not presented.  Baseline values of characteristics such as smoking and HDL-cholesterol for subjects in the former drinking category were more similar to those of abstainers than to those of heavy drinkers; further, adjusted hazard ratios for developing heart failure based on cumulative average alcohol intake were not that different between abstainers and former drinkers (HR =  0.91 for men and 1.02 for women).  These findings may suggest that the former drinking category contained many light drinkers and perhaps relatively few former heavy drinkers or alcoholics.

For the effects of current heavy drinking, the implications from this study are limited because of the small number of such drinkers in this cohort; this was especially the case for women.  As an example, the majority of Black women were abstainers or former drinkers (and there were only 8 women in the highest category of current alcohol consumption).  This is reflected in the wide confidence intervals for estimates of effect among heavy drinkers, and indicates, as Finkel commented: “This analysis provides important data on the population studied, except that heavy drinkers were too sparse for the results to be convincing.” 

In response to a question about the small number of heavy drinkers that are usually found in prospective epidemiologic studies, reviewer Ellison suggested one reason may be that heavy drinkers, especially alcoholics, do not volunteer to be in such cohorts, or drop out early during follow up.  Reviewer Orgogozo added: “Another main reason for the low numbers of heavy drinkers in prospective studies of the the elderly is ‘left censorship’: these folks frequently do not reach old age.  For example, in our population-representative cohort (PAQUID) of 3,700 subjects in France who were 65-years old or more at baseline, only 3% were heavy drinkers (more than 4 glasses/day), while mild and moderate drinkers made up 53% of the cohort.”

While the authors used time-updated reports of alcohol intake to estimate cumulative average intake, Reviewer Barrett-Connor and others emphasized that the investigators did not include the the pattern of drinking in their analyses.  It has been shown repeatedly that consuming small amounts of alcohol on a regular (even daily) basis has favorable health effects, while binge drinking generally has adverse effects on health; this may occur even when the total weekly amount of alcohol consumed is the same.  “Responsible” drinking does not include people who “save up” their alcohol quota for the week and drink it all at once!

Reviewer Djoussé commented: “This is a well-written paper, but I wish the investigators had enough data in the >7 drinks/week, and the pattern of alcohol consumption, to examine the relations of patterns, and perhaps patterns X amount interactions in this cohort.  Beverage type has been neglected in many studies due to lack of adequate data.  It would have been nice to have enough data to provide detailed analyses on beer, spirits, and wine in this large cohort.”

Forum member Skovenborg agreed with other members that this was a good study, adding: “The results are plausible even if the physiological mechanisms are complex and not well understood given that the lower risk of HF associated with moderate drinking is independent of incidence of CAD and MI.  Among the strengths of the study is the large group of abstainers (42%) making confounding associated with abstaining less plausible.”

Could under-reporting of alcohol intake have biased the results of this study?  Forum member Thelle commented on another aspect of the findings in this paper: “I would like to draw your attention to the flatness of the J-shaped curve.  Or rather, ask whether some of the heavy drinkers have been misclassified as moderate drinker, whereby the curve should be flatter than as observed.  This may be due to a discrepancy between observed and reported alcohol consumption.  In an upcoming epidemiology book, I have noted previous research in which exact sales of alcohol in a small community could be measured, and were related to that reported by people in the community (Høyer et al).  Responses on alcohol consumption in the health survey accounted for approximately 40% of total consumption based on sales of alcoholic beverages. The problem with this result is that we do not know who is underestimating their intake.  If everybody reported less than what they consumed this would not introduce an important bias, but perhaps it is only heavy drinkers who under-report their intake and thereby invalidate any possible associations.”  Reviewer Ellison commented that Klatsky et al (2014) have recently reported that subjects who are “likely underreporters” of alcohol consumption can be reasonably well identified in prospective studies with repeated assessments of alcohol and outcomes by using other data in the medical records that suggest heavy drinking or alcohol misuse.

Effects of alcohol consumption on total mortality:  The finding in the present study of no significant effects of baseline (or later) drinking on total mortality is in conflict with many previous studies.  For example, in a study of more than 270,000 men by the American Cancer Society (Bofetta & Garfinkel), the risk of dying of any cause was 16% lower for those reporting one drink per day than for abstainers. Thun et al evaluated alcohol and mortality among 490,000 Americans; all-cause mortality rates were 21% lower among men and women reporting about one drink daily than among nondrinkers.  Doll et al reported that British physicians who were moderate drinkers had lower total mortality rates than lifetime abstainers. 

Di Castelnuovo and colleagues, in a meta-analysis based on more than one million subjects from 56 independent prospective studies, found that the relation between alcohol intake and total mortality is J-shaped, with about 16% reduced risk for light drinkers and increased  mortality for heavy drinkers.  More recently, in large prospective studies from Australia of men aged 60-79 and women aged 70-75 years of age, men consuming up to 4 drinks/day and women up to 2 drinks/day had considerably lower risk of dying (total mortality) than did non-drinkers (McCaul et al).  In those studies, for subjects reporting 1-2 drinks/day, the total mortality risk was about 20-30% lower than that of abstainers. 

Plunk et al, in an analysis based on data from more than 110,000 subjects in the USA, evaluated the relation of “heavy drinking” and “nonheavy drinking” to the risk of all-cause mortality.  The key findings of their analyses were that there was a positive and linear increase in risk of mortality for subjects consuming heavy amounts of alcohol, but for nonheavy drinkers, there was a J-shaped relation with mortality.  Further, Sun et al showed from the Nurses’ Health Study that many aspects of “successful ageing,” in addition to just survival, are favorably affected by regular, moderate consumption of alcohol.

Commenting on a recent paper on alcohol and mortality from the EPIC study by Bergmann et al, in an accompanying editorial Banks concluded: “If taken as causal, these findings are consistent with most public health advice about alcohol, except that most advice recommends an upper limit to alcohol consumption, but does not actually encourage drinking.  In fact, the evidence goes further than this and indicates that, in later life, on average and bearing in mind the priorities and risks of specific individuals, drinking at least some alcohol, but not too much, is likely to minimize the overall risk of death.” 

It is unclear why the present study did not show such a decrease in mortality from moderate drinking.  In comparison with abstainers, the investigators report a significant increase in total mortality for men and women who were former drinkers and those in the highest categories of alcohol intake, but the risks of death among categories of moderate drinkers were not significantly lower than those of abstainers.  The authors note these findings, but do not suggest reasons why their data did not show the expected “J-shaped” curve for total mortality.    

References from Forum Review

Abramson JL, Williams SA, Krumholz HM, Vaccarino V.  Moderate alcohol consumption and risk of heart failure among older persons.  JAMA 2001;285:1971.

Aguilar D, Skali H, Moye LA, et al.  Alcohol consumption and prognosis in patients with left ventricular systolic dysfunction after a myocardial infarction.  J Am Coll Cardiol 2004;43:2015.

Banks E.  Commentary: Lifetime alcohol consumption and mortality: have some, but not too much.  Int J Epidemiol 2013;42:1790–1792; doi:10.1093/ije/dyt218.

Bergmann MM, Rehm J, Klipstein-Grobusch K, et al (38 authors).  The association of pattern of lifetime alcohol use and cause of death in the European Prospective Investigation into Cancer and Nutrition (EPIC) study.  Int J Epidemiol 2013;42:1772-1790. (Reviewed by Forum:

Boffetta P, Garfinkel L.  Alcohol drinking and mortality among men enrolled in an American Cancer Society Prospective Study.  Epidemiology 1990;1:342.

Bryson CL, Mukamal KJ, Mittleman MA, et al.  The association of alcohol consumption and incident heart failure: the Cardiovascular Health Study.  J Am Coll Cardiol 2006;48:305.

Cooper HA, Exner DV, Domanski MJ.  Light-to-moderate alcohol consumption and prognosis in patients with left ventricular systolic dysfunction.  J Am Coll Cardiol 2000;35:1753.

Di Castelnuovo A, Costanzo S, Bagnardi V, Donati MB, Iacoviello L, de Gaetano G.  Alcohol dosing and total mortality in men and women: an updated meta-analysis of 34 prospective studies.  Arch Intern Med 2006;166:2437.

Djousse L, Gaziano JM (2007).  Alcohol consumption and risk of heart failure in the Physicians’ Health Study I. Circulation 2007;115:34.

Djousse L, Gaziano JM (2008).  Alcohol consumption and heart failure: a systematic review.  Curr Atheroscler Rep 2008;10:117.

Doll R, Peto R, Boreham J, Sutherland I.  Mortality in relation to alcohol consumption: a prospective study among male British doctors.  Int J Epidemiol 2005;34:199.

Hoyer G, Nilssen O, Brenn T, Schirmer H.  The Svalbard study 1988–89: a unique setting for validation of self–reported alcohol consumption.  Addiction 1995;90:539-544. 

Klatsky AL, Chartier D, Udaltsova N, et al (2005).  Alcohol drinking and risk of hospitalization for heart failure with and without associated coronary artery disease.  Am J Cardiol 2005;96:346.

Klatsky AL, Udaltsova N, Li Y, Baer D, Nicole Tran H, Friedman GD (2014).   Moderate alcohol intake and cancer: the role of underreporting.  Cancer Causes Control 2014;25:693-699. doi: 10.1007/s10552-014-0372-8. (Reviewed by Forum:

Kloner RA, Rezkalla SH.  To drink or not to drink? That is the question.  Circulation 2007;116:1306.

McCaul KA, Almeida OP, Hankey GJ, Jamrozik K, Byles JE, Flicker L.  Alcohol use and mortality in older men and women.  Addiction 2010;105:1391-1400.  (Reviewed by Forum:

Pai J, Mukamal K, Rimm E.  Long-term alcohol consumption in relation to all-cause and cardiovascular mortality among survivors of myocardial infarction: the Health Professionals Follow-up Study.  Eur Heart J 2012;33:1598–1605. (Reviewed by Forum:

Plunk AD, Syed-Mohammed H, Cavazos-Rehg P, Bierut LJ, Grucza RA.  Alcohol consumption, heavy drinking, and mortality: Rethinking the J-shaped curve.  Alcohol Clin Exp Res 2014;38:471–478. DOI:0.1111/acer.12250.  (Reviewed by Forum:

Sun Q, Townsend MK, Okereke OI, Rimm EB, Hu FB, Stampfer MJ, Grodstein F.  Alcohol consumption at midlife and successful ageing in women: A prospective cohort analysis in the Nurses’ Health Study.  PLoS Med 2011;8: doi:10.1371/journal.pmed.1001090 e1001090. (Reviewed by Forum:

Thun MJ, Peto R, Lopez AD, et al.  Alcohol consumption and mortality among middle-aged and elderly U.S. adults.  N Engl J Med 1997;337:1705.

Walsh CR, Larson MG, Evans JC, Djoussé L, Ellison RC, Vasan RS, Levy D.  Alcohol consumption and risk for congestive heart failure in the Framingham Heart Study, Ann Intern Med 2002;136:181.

Forum Summary

An analysis of data from the large Atherosclerosis Risk in Communities (ARIC) study evaluated the effects of alcohol consumption at baseline, and the cumulative average intake based on several later assessments during a 24 year follow-up period, on the risk of the development of heart failure (HF).  The authors conclude: “In this community, alcohol consumption of up to 7 drinks/week at early-middle age is associated with lower risk for future HF, with a similar but less definite association in women than in men.  These findings suggest that despite the dangers of heavy drinking, modest alcohol consumption in early-middle age may be associated with a lower risk for HF.”  The authors reported further that the risk of total mortality was higher than that of abstainers for former drinkers and for the heaviest drinkers, but did not demonstrate a “J-shaped” curve for total mortality (i.e., a lower risk for moderate drinkers in comparison with abstainers).

Forum members considered this to be a well-done analysis, with findings for a lower risk of HF among light-to-moderate drinkers consistent with many earlier reports.  They noted that these findings may be somewhat affected by the inability of the investigators to include data on the pattern of alcohol consumption (regular moderate intake versus periodic binge drinking of similar total amounts) to disease.  It would be expected that the relations between alcohol intake and HF (as well as with total mortality) would be stronger for regular moderate drinkers than for those who consumed their alcohol in binges (even if the total weekly average intake was the same).  This could be one reason why a “J-shaped” curve was not seen for the relation of alcohol intake to total mortality in this study.         

While coronary artery disease (CAD) is known to be a major precursor of HF, it is not clear from the data presented if the alcohol-HF relation may have differed between subjects with and those without CAD.  For the effects of current heavy drinking, the implications from this study are limited because of the small number of such drinkers in this cohort; this was especially the case for women.  The authors do not discuss the potential effects on their results of under-reporting of alcohol intake among their subjects; other studies have suggested means for identifying such subjects and ways of adjusting for such.

This study adds support to other large studies, including one just released as a pre-publication (see Addendum), that suggest that moderate alcohol consumption may reduce the risk of the development of heart failure.  Some of this putative protective effect surely relates to a decrease in the risk of coronary artery disease among moderate drinkers.  Heavy alcohol intake, especially among alcoholics, may increase the risk of both heart disease and heart failure.


Following the preparation of this critique, the pre-publication version of another paper has appeared: Larsson SC, Orsini N, Wolk A.  Alcohol consumption and risk of heart failure: a dose–response meta‐analysis of prospective studies.  European Journal of Heart Failure 2015; January, 2015. 10.1002/ejhf.228.  According to its abstract, this paper is based on “a meta-analysis that included eight prospective studies, with a total of 202,378 participants and 6,211 cases of HF.”

While the Forum has not yet had an opportunity for a careful review of this new paper, the authors’ conclusions appear to provide strong support for the paper by Gonḉalves et al.  The authors of the Larsson et al paper conclude their abstract by stating that when carrying out a dose-response meta-analysis, they “observed a non-linear relationship between alcohol consumption and risk of HF.  Compared with non-drinkers, the RRs (95% CI) across levels of alcohol consumption were 0.90 (0.84–0.96) for 3 drinks/week, 0.83 (0.73–0.95) for 7 drinks/week, 0.84 (0.72–0.98) for 10 drinks/week, 0.90 (0.73–1.10) for 14 drinks/week, and 1.07 (0.77–1.48) for 21 drinks/week.  Alcohol consumption in moderation is associated with a reduced risk of HF.”

                                               *                *                  *

This critique by the International Scientific Forum on Alcohol Research was based on data provided by the following members of the Forum:

Elizabeth Barrett-Connor, MD, Chief of the Division of Epidemiology, Distinguished Professor in the Departments of Family and Preventive Medicine & Medicine, University of California, San Diego, La Jolla, CA, USA

Giovanni de Gaetano, MD, PhD, Research Laboratories, Catholic University, Campobasso, Italy

Luc Djoussé, MD, DSc, Dept. of Medicine, Division of Aging, Brigham & Women’s Hospital and Harvard Medical School, Boston, MA, USA

R. Curtis Ellison, MD.  Section of Preventive Medicine & Epidemiology, Department of Medicine, Boston University School of Medicine, Boston, MA, USA

Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA

Tedd Goldfinger, DO, FACC, Desert Cardiology of Tucson Heart Center, University of Arizona School of Medicine, Tucson, AZ, USA

Jean-Marc Orgogozo, MD, Professor of Neurology and Head of the Neurology Divisions, the University Hospital of Bordeaux, Pessac, France

Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark

Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Health and Regulatory Information Manager, Australian Wine Research Institute, Glen Osmond, South Australia, Australia.

Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway

Dag S. Thelle, MD, PhD, Senior Professor of Cardiovascular Epidemiology and Prevention, University of Gothenburg, Sweden; Senior Professor of Quantitative Medicine at the University of Oslo, Norway

Gordon Troup, MSc, DSc, School of Physics, Monash University, Victoria, Australia

David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa