LONDON, UK — There's more evidence that even low levels of alcohol consumption don't help reduce the risk of cardiovascular disease, this time from a unique study that authors say may be the closest possible approximation to a randomized clinical trial[1]. The implications are sobering for anyone planning the drinks list for their summer barbecue: reduced alcohol consumption, even in light drinkers, "is beneficial for cardiovascular health."
The report is published online July 10 in the BMJ.
Dr Michael Holmes (University of Pennsylvania, Philadelphia) and Dr Caroline Dale (London School of Hygiene & Tropical Medicine, UK), with senior author Dr Juan P Casas (University College London, UK), collaborated with more than 100 other investigators internationally on a meta-analysis of 56 epidemiological studies to look at the link between levels of alcohol consumption and coronary heart disease. But unlike the countless observational studies that have tackled this topic in the past, Holmes et al tried something new.
"For decades, observational studies have suggested that drinking moderate amounts of alcohol compared with no drinking is beneficial to cardiovascular disease risk, but the problem is that those studies rely on observational data, and it's hard to make cause-and-effect deductions from them," Holmes told heartwire . "Ideally we'd do a randomized controlled trial, which is the gold standard, to understand causal mechanisms, but we know that alcohol causes so much harm in terms of other diseases."
Alcohol is the fifth leading risk factor for death and disability, the authors note, and is an established cause of liver cirrhosis, injuries, and many cancers.
"To expose people to alcohol in a randomized controlled trial is probably unethical, so the question we are faced with is: we have observational data suggesting cardioprotection, and we can't do a randomized controlled trial, so how can we bring the quality of evidence up to the next level?"
Their answer was to conduct a Mendelian randomization, comparing rates of coronary heart disease and stroke in subjects who were or were not carriers of a genetic variant known to be associated with lower levels of alcohol consumption. The A-allele of ADH1B rs1229984, they write, causes flushing, dizziness, and nausea in response to alcohol and is associated with lower levels of usual alcohol consumption and blood ethanol.
As Holmes explained to heartwire , carriers of this variant "feel the effects" of alcohol at much lower levels of consumption and so tend to drink less than people who are not carriers.
Alcohol and Genes
The authors were ultimately able to obtain information on more than 260 000 subjects of European descent, a monumental task that combined studies in which the specific single nucleotide polymorphism (SNP) for the ADH1B variant was either previously measured as part of a gene chip or by contacting individual study groups to have their samples newly genotyped for this specific SNP.
Carriers and noncarriers were then compared for the primary outcome of fatal and nonfatal coronary heart disease and the secondary outcomes of stroke and type 2 diabetes.
Across the sample, carriers of the rs1229984 A-allele consumed, on average, 17.2% fewer British units (10-mL ethanol) of alcohol per week than noncarriers. They were significantly less likely to be in the top third of the sample by drinking volume, were less likely to report binge drinking, and were more likely to report abstaining from alcohol.
Carriers were also more likely than noncarriers to have better CVD risk-factor profiles, including lower rates of hypertension, lower inflammatory markers, lower adiposity, and slightly lower non-HDL cholesterol.
Asked whether this specific genetic variant might also confer other cardioprotective qualities, Holmes observed: "In a Mendelian randomization, you want to make sure the gene variant is specific for the exposure that you're interested in, so we considered this [possibility], and we are pretty confident that it has no other effects."
Strikingly, across all levels of alcohol consumption, carriers of the rs1229984 A-allele had reduced odds of developing coronary heart diseases and ischemic stroke, as compared with noncarriers. No such association was seen for type 2 diabetes.
No U-Shaped Curve
The findings are at odds with the oft-cited U-shape for the alcohol-CVD association, derived from observational studies, that has long boosted the theory that alcohol, in moderation, is heart healthy. If this were true, the authors argue, their analysis would have detected an uptick in CV events among subjects drinking less than 12 to 25 units per week and a downturn in risk among those drinking this many units or slightly higher amounts.
Instead, "contrary to these expectations," the authors write, "we found that individuals below the nadir with genetic predisposition to consume less alcohol had lower odds of developing coronary heart disease at all categories of alcohol consumption, bringing the hypothesized cardioprotective effect of alcohol into question."
To heartwire , Holmes said this unique study lends support to a growing number of studies that are challenging the view that alcohol, in moderation, may be part of a heart-healthy diet. "The interpretation from the paper would be, don't drink alcohol for your heart and if you are concerned about your cardiovascular risk and you do drink, then reduce your alcohol consumption, and that's irrespective of how much you drink."
Commenting on the particular design of the Holmes et al study for heartwire , Dr Sanjay Kaul (Cedars Sinai Medical Center, Los Angeles, CA) noted that "inferences from Mendelian randomization studies are more reliable than observational studies because of minimization of confounding and reverse causation, two key biases that confound the results of epidemiological research."
Kaul agrees that randomized trials represent the "ideal" approach, but like Holmes believes an RCT of reduced vs high alcohol consumption is "unlikely to be feasible."
As such, the Mendelian design "does provide a higher level of evidence compared with observational data."
Holmes, Dale, and colleagues plan to get access to large data sets such as the UK Biobank in order to replicate their findings in an even larger sample.
Holmes disclosed funding by a UK Medical Research Council population health scientist fellowship. Disclosures for the coauthors are listed in the paper.
Heartwire from Medscape © 2014
Cite this: Lowest Alcohol Intake, Least CVD: Novel Study Challenges CV Benefits of Booze - Medscape - Jul 11, 2014.
Comments