Critique 173: A Large Study of the Association of Alcohol with the Risk of Acute Myocardial Infarction — 15 October 2015

Gémes K, Janszky I, Laugsand LE, Lászlo KD, Ahnve S, Vatten LJ, Mukamal KJ.  Alcohol consumption is associated with a lower incidence of acute myocardial infarction: results from a large prospective population-based study in Norway. J Intern Med 2015; doi: 10.1111/joim.12428.

Authors’ Abstract

Aims. Compelling evidence suggests that light-to-moderate alcohol consumption is associated with a reduced risk of acute myocardial infarction (AMI), but several issues from previous studies remain to be addressed. The aim of this study was to investigate some of these key issues related to the association between alcohol consumption and AMI risk, including the strength and shape of the association in a low-drinking setting, the roles of quantity, frequency and beverage type, the importance of confounding by medical and psychiatric conditions, and the lack of prospective data on previous drinking.

Methods. A population-based prospective cohort study of 58 827 community-dwelling individuals followed for 11.6 years was conducted. We assessed the quantity and frequency of consumption of beer, wine and spirits at baseline in 1995–1997 and the frequency of alcohol intake approximately 10 years earlier.

Results. A total of 2966 study participants had an AMI during the follow-up period. Light-to-moderate alcohol consumption was inversely and linearly associated with AMI risk. After adjusting for major cardiovascular disease risk factors, the hazard ratio for a one-drink increment in daily consumption was 0.72 (95% confidence interval 0.62–0.86). Accounting for former drinking or comorbidities had almost no effect on the association. Frequency of alcohol consumption was more strongly associated with lower AMI risk than overall quantity consumed.

Conclusions. Light-to-moderate alcohol consumption was linearly associated with a decreased risk of AMI in a population in which abstaining from alcohol is not socially stigmatized. Our results suggest that frequent alcohol consumption is most cardioprotective and that this association is not driven by misclassification of former drinkers.

Forum Comments

This was a very large, population-based study evaluating the risk of acute myocardial infarction (AMI) associated with alcohol consumption. The study had high levels of participation (69.5% of those invited to take part), so the results are likely to reflect findings of the entire population where it was carried out (a typically low-alcohol culture, with 41% abstainers).  As the authors stated: “Abstainers represent a considerably less disadvantaged subset of the population in Norway . . . reducing the likelihood of uncontrolled confounding by social pressures and integration.”  The Forum has recently commented upon the effects of socioeconomic status on the association between alcohol intake and disease, reviewing studies by Jones et al and Sjölund et al.  For poorly understood reasons, individuals with lower education, income, and other indices of low socioeconomic status tend to show more adverse effects from alcohol as do subjects with higher status.

The present study shows a clear inverse association between alcohol and AMI.  It reports that the frequency of drinking had a greater effect that the quantity of alcohol, even in this low-drinking population.  (As pointed out by Forum member Lanzmann-Petithory, this population differs from other regions of Norway, where heavier drinking is common.)  In beverage-specific analyses, which were adjusted for the other types of beverage consumed, the results suggested greater protection against AMI from wine and spirits than from beer.

Forum members considered this to be a very well-done study, with good data on confounders. There was good ascertainment of outcome, with two-thirds of cases being confirmed by hospital records.  While the present study has few subjects reporting heavy drinking, it gives important information on the health effects of light drinking.

As stated by Forum member Skovenborg, “The thorough adjustment for several important confounding factors is especially important in this study from Norway, where an often-heard opinion is that moderate drinking is an indicator of a healthy lifestyle, not the cause of it (Fekjær).”  Forum member Waterhouse stated: “I like the fact that the authors also disclose total mortality, so that the effect on AMI is not used to hide other problems.”

Reviewer Van Velden stated: “I find this publication quite interesting because the population studied is uniform and the analysis is well done.  The results confirm the well-established observation that moderate alcohol consumption has a protective effect on AMI.”  The authors had responses from the CAGE questionnaire that allowed them to identify problem drinkers (such drinkers had much higher risk of death from liver disease).  Further, the investigators carried out analyses adjusting for competing causes of death; also they did appropriate sensitivity analyses by the level of confirmation of the diagnosis, checking for reverse causation, restricting analyses to individuals without other chronic diseases, etc.  The study had reasonable data on previous drinking among current abstainers, but found that it had little effect on the risk of AMI.

An inverse association or a J-shaped curve?  Reviewer de Gaetano noted that “The authors failed to observe the traditional J-shaped curve: this may be due to a relatively low consumption of alcohol in the region where the study was performed.  Thus, heavy drinking has not influenced the dose –response curve. (I was recently told by a Polish expert of alcohol that in Poland, no J-curve is observed, due to the opposite reason, i.e., too few low to moderate drinkers in respect to heavy binge drinkers.”  Forum member Ursini added: “I endorse such comments.  A ‘J’ is still a ‘J’ although some parts are missing . . .”  Other members agreed that there were not enough heavy drinkers in this population to bring the curve back up, so we only see the early, descending part of the ‘j” in the curve resulting from a decrease in risk with light to moderate drinking.

Noting the “protective” effects against AMI from small amounts of alcohol:  Reviewer Estruch commented: “I am surprised on the effects of so little amounts of alcohol on the cardiovascular system.  Perhaps there is a problem of underreporting, but in the PREDIMED, a prospective nutrition trial in which the participants answered a food frequency questionnaire practically every year during a mean of 5 years, we observed similar results (Esturch et al)”  Reviewer Skovenborg noted: “Regarding the effect of a very light alcohol consumption, I agree with other reviewers that some underreporting is to be expected in this largely non-drinking population.”

Forum member Thelle stated: “This study is closely linked to a series of large population studies undertaken in Norway with standardised questionnaires and laboratory methods.  The validity of the study is in general considered to be high.  I support the views expressed by other Forum members.  Still, the issue of underreporting may be larger in this rural (and puritan?) population than in other studies.”

Estruch commented further: “Drinking pattern is very important and I think that it should be taken into account in the analysis.  The effects of a ‘moderate’ alcohol intake of 1 drink a day (total: 7 drinks/week) are different from one based on 7 drinks during the weekend, where the total is also 7 drinks/week.  Further, duration of intake is also very important and the authors only registered the current alcohol consumption during the last month.  Recent intake may act in preventing effects on thrombosis or arrhythmias, but not on atherosclerosis, the underlying basis of cardiovascular disease; this somewhat limits the conclusions reached by the authors.”  Added Ellison, “On the other hand, alcohol consumption ‘during the previous month’ is usually closely correlated with longer-term intake.”

Reviewer Finkel asked: “Isn’t frequency of drinking both a correlate to the good: the antithesis of pooling all one’s drinks into a binge; and to the bad: an abusive drinker will usually drink more often than a modest drinker?  Although the results in general could have been expected by those without bias, this appears to me a nicely done study, whose lesson is that frequent moderate alcohol consumption is healthy: not only by reducing the risk of heart attack, but by decreasing total mortality.”  As noted by other Forum members, the results of this study are in line with most previous epidemiologic studies over many decades, and support the axiom: “Drink frequently, but in small amounts.”

Is the association with myocardial infarction different for beer drinkers than for consumers of other beverages?  Reviewer Estruch questions the reported finding of less of a decrease in the risk of AMI for beer drinkers: “I do not understand why ‘moderate beer consumption’ does not exert a protective effects on AMI, as wine and spirits did.  Since food and nutrients were not evaluated (a mistake), perhaps different dietary patterns may explain these differences among beer and wine consumers.”  Reviewer Finkel asked: “Were the beer drinkers different?  For obvious example, did they have fat bellies, a demonstrated risk factor?”  Skovenborg stated: “The difference between the effects of beer vs. wine and spirits is very modest, and the reason might well be a difference in the diet of beer drinkers in comparison to the other groups (Johansen et al).”  Reviewer de Gaetano added: “I agree with other reveiwers that beer consumers in Norway, similarly to other Northern European Countries, might have different (less healthy) dietary habits than wine drinkers.”

An important aspect of this study is the identification and classification of abstainers as life-long, former or occasional drinkers. The inclusion of the latter two categories in the control group in several previous studies has been a matter of concern for data analysis and interpretation.  However, somewhat surprisingly, in the presesnt study there was only a modestly increased risk for current non-drinkers who had previously reported some alcohol consumption.

References from Forum critique

Estruch R, Ros E, Salas-Salvadó J, et al, for the PREDIMED Study Investigators.  Primary Prevention of Cardiovascular Disease with a Mediterranean Diet.  N Engl J Med 2013; 368:1279-1290; DOI: 10.1056/NEJMoa1200303.

Fekjær HO.  Alcohol—a universal preventive agent?  A critical analysis.  Addiction 2013;108:2051-2057.

Johansen D, Friis K, Skovenborg E, Grønbæk M.  Food buying habits of people who buy wine or beer: cross sectional study.  BMJ 2006;332(7540):519-522. doi:10.1136/bmj.38694.568981.80

Jones L, Bates G, McCoy E, Bellis MA.  Relationship between alcohol-attributable disease and socioeconomic status, and the role of alcohol consumption in this relationship: a systematic review and meta-analysis.  BMC Public Health 2015;15:400. DOI 10.1186/s12889-015-1720-7.  (Reviewed by Forum on 14 May 2015:

Sjölund S, Hemmingsson T, Gustafsson JE, Allebeck P.  IQ and alcohol-related morbidity and mortality among Swedish men and women: the importance of socioeconomic position.  Journal of Epidemiology and Community Health 2015;69:858-864.  (Reviewed by Forum on 25 August 2015:

Forum Summary

This paper presents prospective data from a large population-based cohort from rural Norway, a region with typically light alcohol drinking and many abstainers who were not ex-heavy drinkers.  It relates reported alcohol intake, assessed on two occasions, with the risk of developing an acute myocardial infarction (AMI).  Its key findings are that, even in this very light-drinking population, drinkers had significantly lower risk (about 20% to 30%) of developing a MI than non-drinkers.  As the authors conclude: “Light-to-moderate alcohol consumption was linearly associated with a decreased risk of AMI in a population in which abstaining from alcohol is not socially stigmatized.  Our results suggest that frequent alcohol consumption is most cardioprotective and that this association is not driven by misclassification of former drinkers.”

As this was a rather homogeneous rural cohort without huge differences among subjects in socioeconomic status, it allows much better control of confounding.  The fact that these results show essentially a continuous inverse association (rather than the usually seen “J-shaped” curve) is not unexpected given the very few heavy drinking subjects in this cohort.   In addition to cardiovascular disease, the investigators also reported effects on total mortality, which showed a “J-shaped” curve.

Forum members considered this to be a very well-done study, and supports almost all previous research showing a protective effect of moderate drinking on the risk of coronary heart disease.  The authors had good data on alcohol exposure and the occurrence of AMI.  A key result from this study is that frequent light drinking, even less than an average of one drink/day, is associated with a significant decrease in the risk of coronary heart disease.  Based on this study and most previous research, the clear message to the public remains: if you have no contraindications to alcohol use and decide to consume an alcoholic beverage, the healthiest approach is to drink frequently but in small amounts.


Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following members:

Ramon Estruch, MD, PhD.  Associate Professor of Medicine, University of Barcelona, Spain

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

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

Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy

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

Andrew L. Waterhouse, PhD, Department of Viticulture and Enology, University of California, Davis, USA

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

Fulvio Mattivi, MSc, Head of the Department of Food Quality and Nutrition, Research and Innovation Centre, Fondazione Edmund Mach, in San Michele all’Adige, Italy

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

Dominique Lanzmann-Petithory,MD, PhD, Nutrition/Cardiology, Praticien Hospitalier Hôpital Emile Roux, Paris, France

Fulvio Ursini, MD, Dept. of Biological Chemistry, Universityof Padova, Padova, Italy

R. Curtis Ellison, MD, Professor of Medicine & Public Health, Boston University School of Medicine, Boston, MA, USA