Critique 185: An innovative assessment of alcohol consumption and the risk of atrial fibrillation – 28 April 2016

Tolstrup JS, Wium-Andersen MK, Ørsted DD, Nordestgaard BG. Alcohol consumption and risk of atrial fibrillation: Observational and genetic estimates of association.  European  Journal of Preventive Cardiology 2016; pre-publication.  DOI: 10.1177/2047487316641804

Authors’ Abstract

Background: The aim of this study was to test the hypothesis that alcohol consumption, both observational (self-reported) and estimated by genetic instruments, is associated with a risk of atrial fibrillation and to determine whether people with high cardiovascular risk are more sensitive towards alcohol than people with low risk.

Methods: We used data for a total of 88,782 men and women from the Copenhagen City Heart Study 1991–1994 and 2001–2003 and the Copenhagen General Population Study 2003–2010. Information on incident cases of atrial fibrillation was obtained from a validated nationwide register. As a measure of alcohol exposure, both self-reported consumption and genetic variations in alcohol metabolizing genes (ADH1B/ADH1C) were used as in instrumental variables. The endpoint was admission to hospital for atrial fibrillation as recorded in a validated hospital register.

Results: A total of 3493 cases of atrial fibrillation occurred during follow-up. High alcohol consumption was associated with a risk of atrial fibrillation among men, but not among women. Among the men who drank 28–35 and 35+ drinks/week, the hazards ratios were 1.40 (95% confidence interval 1.09–1.80) and 1.62 (95% confidence interval 1.27–2.05) compared with men who drank <1 drink/week. Using genotypes as instrumental variables did not reveal a higher risk. Associations in those with high cardiovascular risk were similar to those at lower risk.

Conclusions: Observational alcohol consumption was associated with a higher risk of atrial fibrillation in men. In women, only a high alcohol intake (28+ drinks/week) was associated with a higher risk. Participants with a high cardiovascular risk were no more sensitive towards alcohol than those at low risk. Genetic analysis did not support a causal relationship of linear association between alcohol intake and atrial fibrillation.

Forum Comments

Because of the ageing population, and the fact that many people survive an acute myocardial infarction or other initial manifestation of coronary heart disease, there are more elderly people alive with cardiovascular disease. Such people are at increased risk for developing atrial fibrillation (AF), and its prevalence is increasing world-wide.

The usual risk factors for AF are coronary disease or cardiac valvular disease, but alcohol use has been suggested as another factor that may affect its occurrence. Heavy alcohol ingestion, especially binge drinking, has been known to increase the risk of acute episodes of AF (as part of the “holiday heart syndrome”, Ettinger et al).  However, most prospective epidemiologic studies have found that regular light drinkers are not at increased risk for AF, while regular heavy drinkers or binge drinkers may be at increased risk (Djoussé et al, Kodama et al, Liang et al, Larsson et al).  The Larsson et al paper included a meta-analysis suggesting a slight increase in risk even with light drinking; this was also suggested by Samokhvalov et al, although the latter authors stated: “Based on the categorical analyses, we could not exclude the existence of a threshold (3 drinks a day for men and 2 drinks a day for women).”

The present large study from Denmark provides important data on the risk of AF (with almost 3,500 incident cases occurring in this study) in association with a previous history of alcohol intake. It also evaluates the relation of genetic variants in alcohol metabolizing genes (ADH1B/ADH1C) to the occurrence of the disease.  AF cases were detected by the nationwide register of subjects admitted to hospital or seen in hospital outpatient facilities.  The cohort included only native-born Danes, as it was considered that the distribution of genetic factors would be different in foreign-born citizens of other ethnicities.

The conclusion of the authors is that, in comparison with a referent group consisting of light drinkers or non-drinkers, high alcohol consumption is associated with an increased risk of atrial fibrillation among men, but not among women. Specifically, they found no increase in risk among men for consumption of 1-6 or 7-13 drinks/week, a significant increase for those reporting 14-20 drinks/week (RR = 1.28, 95% CI 1.03 – 1.58), but no significant increase for the 21-27 drinks/week group (RR = 1.17, 95% CI 0.93 – 1.48); thus, there was not a completely smooth dose-response curve of effect.  For heavier drinking, those reporting 28-34 drinks per week and those reporting 35 or more drinks/week, significant increases in the risk of AF was seen among men (RRs of 1.40 and 1.62, respectively).

For women, there were no significant increases in risk of AF regardless of reported alcohol, although a non-significant increase was noted for women reporting 28 + drinks/week (RR = 1.36, CI 0.98-1.98). Thus, overall, these data suggest that there may not be an increase in risk of AF among men who do not consume in excess of the usual guidelines of an average of no more than 2 drinks per day.  For women in this study, there was a suggestion of an increase in risk for heavy drinkers (those reporting an average consumption exceeding 4 drinks/day) but no increase with more moderate drinking.

When the association with the risk of AF was evaluated separately according to the cardiovascular risk of individuals, significant increases were seen only for men consuming 35 or more drinks/week regardless of their risk; no evidence was seen for an increase among women, regardless of their prior cardiovascular risk.

As expected, the genetic markers for ADH1B related to the reported alcohol; the lowest amount of alcohol was found in subjects with “fast metabolism” genotypes (ADH1B 2/2), but this genotype was exceedingly rare (only 14 of about 30,000 men and 19 of about 35,000 women). There were over 2,500 subjects (1,245 men and 1,394 women) with the intermediate ADH1B form (1/2), and they showed 17% higher alcohol intake for men and 32% higher intake for women than the “fast metabolism” group.

The vast majority of subjects were slow metabolizers of alcohol (ADH1B 1/1), and they showed 45% and 60% greater alcohol intake than fast metabolizers for men and women, respectively. There was much less relation between ADH1C genotypes and reported alcohol consumption.  These genetic markers of metabolism of alcohol showed no relation to the risk of developing AF.  (As stated by the authors, this might discourage others from using these genetic markers alone for Mendelian randomization studies of the effects of alcohol consumption.)

Specific comments on paper by Forum members: Overall, Forum members considered this to be a very well-done analysis on a large group of subjects with good ascertainment of the development of atrial fibrillation.  Forum member Thelle, who has published extensively on AF, stated: “The validity of this study, as any observational analysis, depends upon ascertainment of exposure variables, end-points and proper control of confounding variables.  In the present paper we have the usual problems with alcohol as an exposure variable, but atrial fibrillation as an end-point is a different matter altogether compared to, for instance, stroke or myocardial infarction.  Atrial fibrillation is a fugitive end-point, and under-reporting is likely to be large (perhaps >25%), partly due to the fact that many cases are episodic events.  To this must be added that the causal pathways leading to atrial fibrillation encompass obesity, diabetes, hypertension, genes, thyroid disorders, coronary heart disease and other heart diseases.  My main interest has been lone fibrillation (or atrial fibrillation without any known cause — a disputable concept though) and endurance physical training.

“The present paper shows an association between alcohol and atrial fibrillation for men but the Mendelian design makes it unlikely to represent an etiological pathway, thus the hunt for confounders is on. But to this must be added that the mixed bag of atrial fibrillation, either due to other heart disorders, diabetes and obesity, added to the cases of otherwise healthy “lone atrial fibbers” may dilute associations (including etiological) and increase heterogeneity.”

Reviewer Thelle added: “Among the issues which should be considered in future research on this issue is also the effect of the temperature of the beverage taken. It is well known among ‘atrial fibbers’ that cold beverages or ice-cream may trigger bouts of fibrillation.  Thus the cold beer after endurance training is a sure winner here.  One possible explanation is that the thickness of the tissue wall between the esophagus and the left atrium where the trigger areas are found is very small, as low as <5 mm in 40% of subjects in one study, so that the low temperature of a beverage affects the conductivity of the cells.  Another speculation is that sudden low temperature triggers vagus activity and induces bradycardia, which by itself is a mechanism leading to atrial fibrillation in lone fibrillation.  For future research one would want a clearer distinction between different types of atrial fibrillation, and that other characteristics of the beverages were considered as the mediating factors besides the alcohol per se.”

Forum member Goldfinger did not think that the limitation only to subjects seen in hospital was necessarily a problem in this study. “Patients with ‘new onset’ atrial fibrillation are usually hospitalized for rate control, assessment for anticoagulation needs, and potential cardioversion.  Few patients will be completely asymptomatic and found serendipitously during a routine visit to a physician (or a PA or NP . . . more common these days). So I believe that the population studied is valid.”

Forum member Stockley wrote: “From a recent literature review, I concluded that the relationship between atrial fibrillation, a common chronic cardiac arrhythmia, and alcohol appears to be a linear and dose-response relationship, which may reflect alcohol-induced electrophysiological changes in atrial cells (Habuchi et al, Steinbigler et al, Marcus et al, Samokhvalov et al, Kodama et al). The results of this study and paper are consistent with that conclusion.”

Reviewer Goldfinger added: “The value of this paper is the finding that light to moderate drinking, associated with decreased ischemic cardiovascular events and increased longevity, is not clearly a risk factor for atrial fibrillation. It might have been interesting if patients with structural or functional heart disease such as valvular heart disease or cardiomyopathies were teased out as these patients are at a very heightened risk for AF from a substrate potential.  Might these subsets be more sensitive even to light alcohol compared to the ‘lone atrial fibbers’.”  However, it is noted that this study showed that, in terms of alcohol consumption and AF, “Associations in those with high cardiovascular risk were similar to those at lower risk.”

Reviewer Ellison considered it unfortunate that the authors did not have any data on the pattern of drinking (regular versus binge drinking). It would be expected that high peaks of blood alcohol levels associated with binge drinking might be more likely to increase the risk of AF than the generally much lower blood alcohol levels from regular moderate drinking, especially when the alcohol intake is in conjunction with food intake (which blunts the increase in blood alcohol concentration after alcohol consumption).

The Mendelian randomization analyses: The study showed that the presence of pre-existing cardiovascular disease or its risk factors (present in 39% of men and 38% of women) did not affect the association between alcohol intake and the risk of AF. And importantly, while alcohol intake was greater among subjects with intermediate or slow metabolism genes than among fast metabolizers, these genetic patterns did not relate to the risk of AF from alcohol consumption.

The authors state that “From analyses using the genetic variation in alcohol metabolizing genes (ADH1B and ADH1C) as instrumental variables for alcohol consumption, we found no evidence to support causality of the observational findings.” The authors added, however: “A limitation is that the ADH1B and ADH1C genotypes are not strong instruments, meaning that the influence on the amount of alcohol consumption is limited; the lack of a positive finding does not preclude that alcohol is causally associated with atrial fibrillation.”

Forum member Finkel said: “I remain uneasy about the Mendelian hocus-pocus, and fear that the restriction of AF subjects to those who have been hospitalized may be a serious defect of this study.” Reviewer Keil was also disturbed by the fact that only hospital subjects were included, and added: “Studies having applied the ADH gene polymorphisms for obtaining a ‘more valid’ estimate of alcohol’s effects have been very disappointing so far, since the first study by Hines et al appeared in the NEJM in 2001.”

Forum member Goldfinger, a cardiologist, concluded his remarks with overall comments on AF: “Atrial fibrillation, for the most part, is easily rate controlled if chronic, and poses a low risk for cardio-embolic events when the appropriate patients are placed on anticoagulant therapy. Only a small percentage of the most seriously afflicted with a low ejection fraction are thrown into heart failure by loss of atrial systole.  So, most cardiologists, I think, look at it as more of a nuisance than a life threatening disease.  In proper perspective, the findings of this study are reassuring, but even if there were a small risk associated with light drinking (as in hypertensive patients with a rise in blood pressure), the big picture is still one of longer life and decreased risk of myocardial infarction for light to moderate drinkers.”


Djoussé L, Levy D, Benjamin EJ, Blease SJ, Russ A, Larson MG, Massaro JM, D’Agostino RB, Wolf PA, Ellison RC. Long-term alcohol consumption and the risk of atrial fibrillation in the Framingham study.  Am J Cardiol 2004;93:710-713.

Ettinger PO, Wu CF, De La Cruz C, Weisse AB, Ahmed SS, Regan TJ. Arrhythmias and the “Holiday Heart”: alcohol-associated cardiac rhythm disorders.  Am. Heart J 1978;95:555–562. doi:10.1016/0002-8703(78)90296-X.

Habuchi Y, Furukawa T, Tanaka H, Lu L, Morikawa J, Yoshimura M. Ethanol inhibition of Ca2+ and Na+ currents in the guinea-pig heart.  Eur J Pharmacol 1995;292:143–149.

Hines LM, Meir SM, Stampfer J, et al. Genetic Variation in Alcohol Dehydrogenase and the Beneficial Effect of Moderate Alcohol Consumption on Myocardial Infarction.  N Engl J Med 2001; 344:549-555; DOI: 10.1056/NEJM200102223440802.

Kodama S, Saito K, Tanaka S, et al. Alcohol Consumption and Risk of Atrial Fibrillation.  J Am Coll Cardiol 2011;57:427-436.  doi:10.1016/j.jacc.2010.08.641.

Larsson SC, Drca N, Wolk A. Alcohol Consumption and Risk of Atrial Fibrillation: A Prospective Study and Dose-Response Meta-Analysis.  J Am Coll Cardiol 2014;64:281–289.

Liang Y, Mente A, Yusuf S, et al. Alcohol consumption and the risk of incident atrial fibrillation among people with cardiovascular disease.  CMAJ 2012;184:E857–E866. doi: 10.1503/cmaj.120412.

Marcus GM, Smith LM, Whiteman D. Alcohol intake is significantly associated with atrial flutter in patients under 60 years of age and a shorter right atrial effective refractory period.  Pacing Clin Electrophysiol 2008;31:266-272.

Samokhvalov AV, Irving HM, Rehm J. Alcohol Consumption as a Risk Factor for Atrial Fibrillation: a Systematic Review and Meta-Analysis.  Eur J Cardiovasc Prev Rehabil 2010;17:706–712.  doi: 10.1097/HJR.0b013e32833a1947.

Steinbigler P, Haberl R, Konig B, Steinbeck G. P-wave signal averaging identifies patients prone to alcohol-induced paroxysmal atrial fibrillation.  Am J Cardiol 2003;91:491-494.

Forum Summary

A study from Denmark has been carried out to test the hypothesis that alcohol consumption, both observational (self-reported) and estimated by genetic instruments, is associated with the risk of atrial fibrillation (AF) and to determine whether people with high cardiovascular risk are more sensitive towards alcohol than people with low risk. It was based on a large cohort of subjects (more than 88,000) with a mean follow-up period of 6.1 years; there were almost 3,500 cases of AF diagnosed from hospital records during follow up.  Unfortunately, the authors did not have data to identify binge drinkers, which tend to show greater adverse cardiovascular events than regular moderate drinkers whose weekly intake may be the same.

The main results of the study were that men consuming more than 14 drinks/week, especially those consuming more than 28 drinks/week, had an increase in risk of AF, but no significant increase in risk was seen for any level of alcohol intake among women. When genotypes affecting alcohol metabolism (AHD1B, ADH1C) were studied in a Mendelian randomization analysis, the authors state that they “found no evidence to support causality of the observational findings.”

Forum reviewers of this article considered it to be a well-done study with appropriate analyses. Its results reflect the findings of most previous prospective studies and meta-analyses of little effect of light drinking on AF, but an increase in risk for heavier drinkers.  The study also showed that the effects of alcohol consumption on the risk of AF were not different between subjects who had cardiovascular disease or were at high-risk of cardiovascular disease than for other subjects.

While Mendelian randomization using genetic factors affecting alcohol metabolism has been touted as an unbiased approach for judging causal health effects of alcohol, there are questions about the adequacy of such instruments for judging effects. In the present study, their use did not suggest that the relations shown by the self-report of alcohol by subjects necessarily indicated a causal association of alcohol with AF.  Overall, current data suggest that heavy drinking may increase the risk of AF, but there is little evidence of a meaningful increase in risk from light drinking. Luckily, such levels of alcohol intake (some guidelines suggest no more than 2 drinks/day for men or 1 drink/day for women) have been shown from many previous studies to significantly lower the risk of cardiovascular disease and total mortality.

*                      *                      *                      *                      *                      *                      *                     *

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

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

Dag S. Thelle, MD, PhD, Department of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, Norway; Section for Epidemiology and Social Medicine, Sahlgrenska Academy, University of Gothenburg, Sweden

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

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

Ulrich Keil, MD, PhD, Professor Emeritus, Institute of Epidemiology & Social Medicine, University of Muenster, Germany

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

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

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