Alcohol: Better for Heart Attacks, Worse for Atrial Fibrillation? Interview with:

Dr. Gregory M. Marcus MD Associate Professor UCSF School of Medicine

Dr. Gregory Marcus

Gregory M Marcus, MD, MAS, FACC, FAHA, FHRS
Director of Clinical Research
Division of Cardiology
Endowed Professor of Atrial Fibrillation Research
University of California, San Francisco What is the background for this study?

Response: Multiple epidemiologic studies have demonstrated that alcohol consumption likely increases the risk for atrial fibrillation and reduces the risk for myocardial infarction. However, the results have been conflicting, they generally all rely on self-report of alcohol consumption (which is known to be unreliable, particularly in those that drink more heavily), and there is almost certainly confounding related to an individual’s choice to consume alcohol (which in most settings is ubiquitously available). In addition, the relationship between alcohol and heart failure remains poorly understood, with evidence suggesting there may be both harmful and beneficial effects. Finally, the relationship between alcohol consumption and these various cardiovascular diseases (atrial fibrillation, myocardial infarction, and heart failure) have not been examined within the same cohort of individuals in a simultaneous fashion. What are the main findings?

Response: We sought to approach this important question in a novel way. Specifically, we leveraged the fact that Texas includes counties that prohibit all sales of alcohol (what we termed “dry” counties) and others that have no restrictions (what we termed “wet counties”), and, perhaps most importantly, 7 counties changed from dry to wet during our 5 year study period. We examined all of the diagnoses listed for all hospitalization in those counties from 2005 to 2010. In an important validation analysis, we found that residents of wet counties exhibited more diagnoses of alcohol abuse and alcoholic liver disease, suggesting that alcohol consumption is indeed greater in the wet counties.

After adjusting for potential confounders, we found that residents of wet counties exhibited significantly more atrial fibrillation and yet significantly less myocardial infarctions and heart failure. Among the counties that changed their alcohol access laws, both alcohol abuse and alcoholic liver disease became significantly more common within only a median of about 2 years of follow-up; during that same time, atrial fibrillation became significantly more common after (versus before) liberalization of alcohol sales laws. Over that short period of time and contrary to the between-county analyses, there were no detectable differences in myocardial infarction while heart failure increased. While we cannot exclude residual confounding as an explanation for the discrepant results, it may be that the protective effects of alcohol with regards to myocardial infarction and heart failure are more chronic (the between-county differences had been present for decades), whereas quick change does not yet have a chance to manifest protection against myocardial infarction and may more acutely increase heart failure due to direct myopathic effects. Regardless, in both cases, atrial fibrillation was more common. What should readers take away from your report?

Response: This very different approach to studying the effects of alcohol on heart disease revealed several findings that likely bolster previous research in favor of alcohol as a real and important determinant of atrial fibrillation and yet protective from myocardial infarction. Because any confounding in our current study would have been expected to be quite different than confounding in previous studies, these consistent results strongly favor actual causality. Our report also confirms the notion that the relationship between alcohol and heart failure is quite complex. Ultimately, as it appears alcohol may be both protective and harmful when it comes to heart disease, it is very likely that one size does not fit all—future research will be important to identify which patients are more or less prone to these different forms of heart disease to help us counsel individuals in an informed fashion. Finally, these data demonstrate that laws governing access to controlled substances can have health effects, both helpful and harmful. What recommendations do you have for future research as a result of this study?

Response: Ultimately, a prospective randomized trial would be most helpful, albeit difficult to accomplish. Using platforms such as our Health eHeart Study (where large numbers of people can be enrolled remotely and efficiently, including any adult living anywhere with an internet connection) may help with such research (note- any reader can enroll at Such platforms and new genomic research will likely be important for “personalized medicine,” enabling us to identify individual patients prone to alcohol-induced atrial fibrillation and/ or those with the most to gain from moderate alcohol consumption to protect against myocardial infarction. Thank you for your contribution to the community.


Dukes Jonathan W, Dewland Thomas A, VittinghoffEric, Olgin Jeffrey E, Pletcher Mark J, Hahn Judith Aet al. Access to alcohol and heart disease among patients in hospital: observational cohort study using differences in alcohol sales laws BMJ 2016;353 :i2714

Note: Content is Not intended as medical advice. Please consult your health care provider regarding your specific medical condition and questions.

More Medical Research Interviews on

[wysija_form id=”5″]

Last Updated on June 26, 2016 by Marie Benz MD FAAD