For many of us, a beer after work is an enjoyable way to relax. And though drinking alcohol comes with known health risks, some data suggest that low to moderate drinking may actually decrease the risk of ischemic heart disease (IHD) – also called coronary heart disease – defined by reduced blood flow to the heart, typically due to narrowed coronary arteries.
The jury is still out
Is this good news about IHD true? Unfortunately, the evidence is unclear. Recent meta-analyses synthesising data from large numbers of existing observational studies – which typically rely on participant self-reports of alcohol use – have indicated that average consumption may indeed reduce IHD risk. But results from Mendelian Randomisation (MR) studies – a new type of research relying on participants’ genetic profiles to predict their alcohol use rather than solely depending on self-report – either show no association between moderate drinking and IHD risk or show a harmful relationship.
The devil is in the details
Why do these different types of studies yield different results? In each type of study, the true relationship can get obscured by the presence of bias or confounding in the data. Simply put, in the observational studies, there may be reasons other than the absence of alcohol use that underlie current non-drinkers showing a higher risk of IHD. For instance, some of them may in fact be former drinkers, even “sick quitters” who quit because they developed preclinical symptoms due to excessive drinking. Or there may be other characteristics common to either drinkers or non-drinkers – like diet or emotional/mental health factors – that have a hidden impact on the association observed. Additionally, self-report measures of alcohol use are prone to inaccuracy, including intentional misreporting.
Although MR studies avoid issues of self-report and confounding, they too can suffer from bias. For instance, the genetic variant used to predict alcohol consumption may affect IHD risk through pathways other than drinking, or the variant may actually be most closely linked to heavy episodic drinking, which is likely to affect IHD risk differently than low-to-moderate drinking.
Re-examining the existing data
To re-examine the relationship between alcohol and IHD risk, we used newly developed “Burden of Proof” methods that systematically combine a wide variety of data differing across factors like alcohol use definitions and degree of adjusting for confounders and generate a conservative estimate of the alcohol–IHD association, taking into account how widely findings vary across input studies. Our careful analysis confirms previous results that self-reported versus genetically predicted alcohol use data yield conflicting findings about the alcohol–IHD relationship.
Our analysis of two different types of observational studies (cohort and case-control) revealed an inverse relationship between alcohol and IHD. Our synthesis of 95 cohort studies showed that any alcohol consumption between >0 and 100 grams/day was related to lower IHD risk, compared to drinking no alcohol. The lowest risk was observed at an intake of about 23 grams/day.
RR relative risk.
Based on data from 27 case-control studies, we found that alcohol consumption of up to 61 grams/day was related to a lower risk of IHD, compared to no alcohol consumption, while a slightly higher IHD risk was observed at consumption levels above 61 grams/day.
RR relative risk.
Our analysis of data from four MR studies showed no relationship between alcohol consumption and IHD risk.
RR relative risk.
Teasing out an answer to this critical public health question
The differing results leave us questioning which study design to trust. But there is no simple answer, as each design has inherent biases and limitations. So, what’s next?
MR is a rapidly evolving field. With new methods and expanding genetic databases, MR should clearly be further used to study the alcohol–IHD relationship.
In 2018, the first long-term randomised controlled trial (RCT) to study the cardiometabolic effects of low-to-moderate alcohol consumption – the Moderate Alcohol and Cardiovascular Health Trial (MACH15) – was initiated. Although it was soon terminated by the US National Institutes of Health due to concerns about the study design and irregularities in the development of funding opportunities, MACH15 revisited the possibility of using RCTs as a way to better understand the alcohol–IHD relationship. The use of RCTs to study alcohol use is, however, fraught with potential ethical issues due to the growing number of studies suggesting increased disease risk, including cancer, associated with alcohol use even at very low levels.
In 2023, the European Research Council approved funding for the University of Navarra Alumni Trialists Initiative (UNATI), a four-year RCT examining the impact of advice to moderate consumption on major diseases and mortality. UNATI only enrols participants who already drink alcohol and creates a control group by advising some of them to abstain.
An alternative to conducting an RCT is to use observational data to “emulate” a trial. This involves first describing the (hypothetical) trial, including the participant population, treatments, and health outcomes, and then applying this trial design to an existing observational database. This helps avoid the biases inherent to observational studies. In contrast to RCTs, emulating trials will not entail safety or other ethical concerns because the data are already collected. There is also the potential to address questions that likely will never be answered with an RCT, such as how long-term heavy (episodic) drinking affects health. For example, MACH15 or UNATI could be emulated following the published protocol but with relaxed eligibility criteria, longer follow-up, and the assessment of drinking patterns beyond moderate intake.
While results from UNATI will be invaluable, we should continue to move forward by advancing our use of MR and emulating trials to more conclusively answer whether alcohol and the heart are friend or foe.
Written by Sinclair Carr, PhD Student in Epidemiology at the Harvard T.H. Chan School of Public Health, and Dr. Susan A. McLaughlin, Institute for Health Metrics and Evaluation, University of Washington.
All IAS Blogposts are published with the permission of the author. The views expressed are solely the author’s own and do not necessarily represent the views of the Institute of Alcohol Studies.