It’s a new year, which means it’s time to take stock of old habits. For many that means resolutions to get healthier and a month long break from booze, in observation of Dry January.
Yet as 2025 begins, what hasn’t changed is the muddled messaging over the health effects of moderate drinking. Two recently released federal documents offer contrasting perspectives on alcohol and health–underscoring a long-standing, ongoing scientific debate that’s reverberated through some of the most prestigious scientific journals and institutions.
A review published in December by the National Academies of Science, Engineering, and Medicine tows what some see as an outdated line: suggesting that up to a drink or two a day is associated with health benefits like reduced risks of heart disease and death. The National Academies’ analysis will likely inform revised national dietary guidelines set to come out later this year. (Current U.S. guidelines state that alcohol consumption should be limited to one standard drink a day for women and up to two drinks a day for men. This is also considered the upper limit on “moderate consumption” by many U.S. researchers and agencies.)
Then, on January 3rd, the office of the U.S. Surgeon General released an official advisory unequivocally stating that drinking alcohol, at levels as low as one drink per day, can cause certain cancers. The advisory calls for updating the existing health warning labels on alcoholic beverages to include a warning about the cancer risks.
So what gives? How can any substance be simultaneously associated with reduced risk of death and also increased cancer risk? How can big studies in well-regarded journals disagree on something as basic as harm or benefit? Why is the U.S. government seemingly arguing with itself? Is any amount of alcohol healthy?
Answers to all of the above are complicated. There’s statistical biases and confounding factors that muddy the data, the financial interests and outside influence, and a lack of consensus on how to define terms such as risk and moderate. Researchers sometimes disagree with each other about how to interpret the available evidence, and what guidance is best for boosting health. But there are some things we know for sure. Here’s what’s clear and what remains murky, when it comes to alcohol and health.
Consensus, straight up
Popular Science spoke with six expert sources for this article, including some who endorse the idea that moderate drinking may be associated with health benefits and those who say that drinking at every level carries only health risks. Every single source agreed that alcohol consumption exceeding one standard drink a day for women or two drinks a day for men can have negative health consequences, and that drinking beyond that level brings significantly increased risks of accidental injury and death, certain cancers, heart problems, liver disease, cognitive impairments, and more.
Internationally, the definition of a standard drink varies. But in the U.S. and Canada it’s defined as 0.6 fluid ounces or about 14 grams of pure alcohol, which is equivalent to 12 oz of a 5% beer, 5 oz of a 12% wine, or 1.5 oz of a 40% distilled spirit.
All sources agree that binge drinking, heavy drinking, and alcohol use disorder are serious public health problems, and none directly endorse starting or increasing alcohol consumption for any health reason.
“I wouldn’t want to recommend drinking, especially to someone who isn’t otherwise going to drink,” says Gregory Marcus, a cardiologist and professor of medicine at the University of California, San Francisco. “Even if there are protective effects, which I acknowledge is possible, I don’t think that the level of evidence is high enough to recommend initiating it,” he adds. That’s despite his own, ongoing research trying to tease out potential heart health benefits of moderate drinking and his perspective that “light drinking may have salutary effects that are biologically plausible.”
Further, it’s firmly established that alcohol is an addictive substance with systemic effects in the body that can lead to physiological dependence. Use at low levels carries the risk of increasing and excessive consumption. Upwards of 20% of Americans who drink will experience an alcohol use disorder in their lifetime, according to data from the National Epidemiologic Survey on Alcohol and Related Conditions.
There is no debate when it comes to the harms of excessive alcohol consumption. Instead the ongoing scientific discussion is about how to research and classify the health effects of relatively low alcohol consumption and how to communicate to the public about those effects.
Increasingly, there’s also burgeoning scientific agreement that the risk-relationship for alcohol and certain cancers is straightforward: every additional volume of alcohol consumed is associated with increased risk of cancer. “Alcohol is a carcinogen. That link is very well established,” says Adam Sherk, a senior scientist at the Canadian Center on Substance Use and Addiction, an NGO founded and overseen by the Canadian government.
Different analyses disagree about what level of alcohol consumption poses what level of risk for what cancers. But even the new National Academies report, which critics say is conservative in its assessment of health harms, found a 10% increase in breast cancer risk among moderate drinkers compared with those who reported never drinking alcohol.
According to the National Cancer Institute (NCI), there is robust evidence linking moderate drinking with increased risk of head and neck, esophageal, colorectal, and breast cancers. Light drinking (variably defined, but loosely: a few standard drinks a week, and less than a drink a day) is associated with measurable increased risk of esophageal and breast cancers, per NCI.
Finally, sources generally agree that, when it comes to assessing the cumulative health effects of occasional or light drinking–whatever benefits or harms may exist are modest. It’s entirely possible that light or moderate alcohol consumption has no meaningful health impact for many people, according to Marcus. “These are not major, large effects,” he says.
Tim Stockwell, a psychologist and professor at the University of Victoria in British Columbia, also views the effects of low level drinking as small. Stockwell studies the measurement and prevention of alcohol related harms, and stations himself squarely in the camp of researchers who believe no level of alcohol consumption is without risk. Despite his firm stance that evidence indicates drinking in any volume is “net bad” for health, averaged over a population, he explains each drink is likely equivalent to only a few minutes of lost life. “If you [consistently] have six drinks a week, it’s probably about two or three months of lost life expectancy,” he adds. At something like two drinks per week, “it’s a very small risk,” Stockwell says.
Where things gets fuzzy
Yet a small risk is a very different thing from a small benefit, at the public health level, and that’s where the consensus breaks down. Despite the fact that individual effects of low and moderate drinking may be small, Stockwell notes that they can add up across the population. “We shouldn’t just focus on the high end of the spectrum,” he says. Unfortunately, the broader view remains confusing.
Beginning in the 1990’s reports about the purported health benefits of red wine began to lend moderate drinking a “health halo”–wherein the prevailing assumption among the drinking public became that a small amount of alcohol is good for you, especially for the heart. Subsequent studies of alcohol consumption and all-cause mortality found what’s known as a j-shaped curve: Those who drank some had a lower risk of dying than those who reported entirely abstaining. Meanwhile heavy drinkers’ mortality risk was far higher than both groups. This furthered the idea that a little bit of alcohol every day is healthy, despite the harms of overconsumption.
But in ensuing decades, research into all-cause mortality and other health metrics has yielded inconsistent results. For instance, a 2017 study found that even moderate drinking is associated with brain volume loss. Some reviews, like a meta-analysis co-authored by Stockwell in 2023, conclude that low-level drinking offers no protective effect. In 2022 the World Health Organization surveyed the available research and concluded, “No level of alcohol consumption is safe for our health.”
Whereas other assessments, like the recent National Academies meta-analysis, continues to report measurable benefits associated with alcohol consumption at low and moderate levels. In one particularly complicating instance, The Lancet–among the highest impact and most esteemed medical journals–published a 2018 analysis on alcohol’s global burden of disease reporting no j-shaped curve and no health benefit of drinking. Two years later, the same research group published a revision, stating that the data does, in fact, support a reduction in cardiovascular risk and all-cause death associated with “small amounts of alcohol consumption” among certain age groups.
Again, it’s worth noting that all of these review studies combine research that defines “moderate” and “low levels” of drinking in different ways. In the National Academies study, a few of the studies used to inform the all-cause mortality finding set the upper threshold for moderate drinking at 0.7 U.S. standard drinks per day. Only one set an upper limit on moderate drinking as high as how it’s defined by existing U.S. dietary guidelines. That means that the reported findings on all-cause mortality related to moderate drinking don’t necessarily apply evenly to all amounts of alcohol consumption between 0 and 2 drinks a day.
Regardless, the basic idea is that, even if alcohol increases cancer or other health risks, at low or moderate levels (somewhere between 0 and 1-2 drinks per day) it may simultaneously reduce cardiovascular risk. And, in the U.S. where heart disease is the biggest killer, that could seem a worthwhile trade-off.
Proposed mechanisms for alcohol’s cardiovascular health boost include that it might increase the amount of the body’s “good” HDL-cholesterol, inhibit clotting, or help moderate blood sugar levels–though it’s hard to disentangle the effects of alcohol itself from the other compounds present in beverages like red wine. Some also suggest that the social aspect of alcohol may mean those who engage in moderate drinking have richer social lives and are less lonely. “Social interaction is a positive thing,” says Ned Calonge, a physician, professor of epidemiology, and an associate dean in the school of public health at Colorado University. Calonge chaired the consensus committee that wrote the National Academies report. Perhaps, he suggests, “there are things associated with moderate drinking that have nothing to do with alcohol,” which could account for some of the observed heart and lifespan benefits.
The key thing to know here is that nearly all studies of alcohol in humans are observational, meaning they look for correlations among large research cohorts who self-report habits. These aren’t randomized control studies, which serve as the gold standard for understanding the effects of pharmaceuticals. Instead, observational studies can only really demonstrate associations, not causal relationships. Drinking habits are complex and heavily influenced by non-random factors. Observational research can easily be inadvertently biased.
One of the primary difficulties is “abstainer bias,” wherein studies may lump together never drinkers with those who previously drank and then quit. People who stop drinking often do so, in part, because of health problems–and so they may constitute an inherently less healthy group compared with those who keep drinking moderately. Calonge notes that the National Academies report sought to eliminate this problem by only including studies that separate never drinkers from quitters.
Yet that, along with other stringent selection criteria, left Calonge and his co-authors with only a small handful of studies to assess. Their primary all-cause mortality finding was based on just eight research papers and some of their all-cause and cancer findings on as few as two. For many of the questions they asked, they concluded “insufficient evidence”.
“They looked at probably about 1% of the studies that have been published and said there’s not enough evidence,” says Stockwell. “I thought that was very misleading.”
Further, abstainer bias can cut both ways: by eliminating those who quit drinking for health reasons from a study, you may artificially be inflating the health profile of drinkers, Stockwell explains. Plus, self-reporting is an unreliable metric of actual drinking habits, different studies (even those equated in meta-analyses) define moderate drinking at different levels, and people who are true never drinkers may be an inherently less healthy group because lifelong health conditions prevent them from drinking. Confounding demographic factors like age and socioeconomic status also play a role. Some research, including Stockwell’s own, has shown that heart and health risk factors among non-drinkers and moderate drinkers vary for reasons completely unrelated to alcohol.
In the U.S., about 5% of deaths are attributed to alcohol annually. Yet somehow, the National Academies report found a 16% lower risk of all-cause mortality for moderate drinkers. “I think that’s logically impossible… alcohol would have to have an immense benefit,” says Sherk, and other evidence doesn’t support that. “I just feel like it has to be an artifact, because otherwise every single study would send immensely clear messages that low volume use is hugely protective.”
All-cause mortality as a metric has come into question and fallen out of favor for many scientists because it’s so difficult to control for data artifacts, bias, and confounders, says Jürgen Rehm a senior scientist and professor at the Center for Addiction and Mental Health at the University of Toronto who contributed to the 2022 WHO guidance. “All cause mortality studies use middle class samples who die of completely different causes of death than the rest of the population [and] they’re very specific per country,” he explains. One major flaw in the National Academies review is that they include some observational studies of non-U.S. cohorts to make a judgement about Americans’ risk, Rehm says. “It’s nonsense…South Korea has a completely different profile of causes of death and you cannot do that if you want to create guidelines for the U.S.”
Some research has begun to embrace alternate methods beyond observational cohort studies. There are, for instance, studies that attempt to randomize participants based on genetics, which Rehm says have more potential to illuminate causal relationships. (Though Marcus thinks this type of research is often over-interpreted.) And there are a handful of randomized control studies looking at the effects of alcohol dosage on health. To clear the ethical hurdles inherent in assigning people to drink alcohol, most of these studies instead take a group of moderate drinkers and prescribe some portion to drink less. So far, some of this research has found low to moderate levels of alcohol consumption has no significant effect on factors like blood pressure or mild negative effects on abnormal heart rhythms.
For now, the state of the research leaves even experts unconvinced one way or the other on the potential benefits of alcohol. “There is no consensus of scientists on the cardiovascular effects,” says Rehm.
“I’m quite agnostic… I’m willing to let well-designed studies decide,” says Sherk. And again, those studies might not yet exist.
What’s at stake?
Money is often a talking point for those on both sides of the moderate drinking debate. The alcohol industry and its lobbying arm have funded, advised, and promoted research endorsing health benefits associated with drinking. In 2018, an National Institutes of Health study was shut down because of an investigation revealing financial conflicts of interest. Similar issues have cropped up since.
On the flipside, drink-defenders note that this accounts for a small portion of the total research, according to one 2020 analysis. And, at the same time, temperance and religious organizations have a vested interest in promoting research demonstrating alcohol’s harms. Stockwell admits that, at different points in his career, he’s accepted small amounts of conference and travel reimbursement from both an alcohol industry group and, more recently, Movendi International, a Swedish temperance group.
It’s clear that official guidance, online trends, and shifting research tides do change peoples’ drinking and buying behavior, so any group with a moral position or financial stake has good reason to try to involve itself with scientists. Following a “60 Minutes” broadcast promoting the idea of red wine’s health benefits in 1991, sales of red wine spiked. Amid the current health and wellness wave, which tends to endorse drinking less, wineries are experiencing a downturn. But ideally, our health decisions would be well-informed and free from the influence of profit or proselytizers.
Public health organizations and researchers try to help, by parsing the evidence simplifying what’s known and unknown for the public. “Our role is really to raise awareness and to put the evidence out there and support people in understanding what the risks are,” says Carina Ferreira-Borges, a regional program manager for WHO Europe who focuses on alcohol. Then, she says, groups like WHO advocate for harm reduction policies accordingly. Labels, regulations, and consumption guidelines can all be tools for doing so.
Yet even official guidelines can’t capture the nuance and individual factors that might shift the outcomes of drinking for any one person at any one time. If, perhaps, you know you’re at high risk of developing cancer, then cutting out alcohol entirely might make sense. If, on the other hand, you’re more worried about a heart attack, then maybe that’s not a reason to become a teetotaler, says Marcus.
“If you believe in both the [positive and negative] effects, you should try to look at the reasonable risks you’re willing to take,” says Rehm. Whether or not to drink is a personal choice, and sometimes science has limits.
Win the Holidays with PopSci’s Gift Guides
