UCare generously supports MinnPost’s Second Opinion coverage; learn why.

Risks of drinking alcohol far outweigh any potential benefits, study authors conclude

REUTERS/Mike Blake
The study authors believe that when alcohol’s risks and benefits are weighed, only one conclusion can be reached: “The safest level of drinking is none.”

We’ve been told for years that drinking alcohol in moderation — having a glass of wine or a beer with dinner each day — is not only harmless to our health, but may actually protect our heart and thus help us live longer.  

In recent years, however, some health experts have begun to question that message, pointing to a growing number of studies that suggest moderate drinking may not be so innocuous after all. 

The authors of a major new study, published last Friday in The Lancet, definitely fall into that latter group of experts. After examining the evidence gathered from around the world, they believe that when alcohol’s risks and benefits are weighed, only one conclusion can be reached: “The safest level of drinking is none.”

“The health risks associated with alcohol are massive,” said Emmanuela Gakidou of the Institute for Health Metrics and Evaluation (IHME) at the University of Washington in Seattle and the senior author of the study, in a released statement. “Our findings are consistent with other recent research, which found clear and convincing correlations between drinking and premature death, cancer, and cardiovascular problems. 

“Zero alcohol consumption minimizes the overall risk of health loss,” she added.  

From the media headlines that quickly followed — “No amount of alcohol is good for your health,” “No alcohol safe to drink,” “How much alcohol is safe to drink? There is no healthy amount” — you would think a dramatic turning point had occurred in the debate over what level of alcohol consumption is safe for your health. 

But, as is often the case with research on health-related risk factors, the conclusions that can be drawn from the new study, which used observational data, are not that clear-cut. 

The study’s findings certainly add to the growing evidence that current U.S. recommendations regarding “safe” levels of alcohol consumption need to be revisited. Those recommendations may be too generous. But the study’s findings don’t prove that abstinence from alcohol is the only healthy option for all individuals. 

Hundreds of studies

For the study, which was funded by the Bill and Melinda Gates Foundation, IHME researchers relied heavily on the 2016 Global Burden of Disease report. They used that report’s data to analyze alcohol-related consumption patterns and health risks between 1990 and 2016 for 195 countries and territories.

The consumption data was derived from 694 datasets, while the health-risk data came from 592 studies.

The researchers focused on 23 alcohol-related health problems for people aged 15 and older. The problems included cardiovascular diseases (such as heart disease and stroke), cancers (such as breast, colorectal, liver and mouth), other noncommunicable diseases (such as diabetes, epilepsy and cirrhosis of the liver), communicable diseases (such as lower respiratory infections and tuberculosis), intentional injuries (such as interpersonal violence and self-harm), unintentional injuries (such as poisonings, burns and drownings) and transportation-related injuries (such as ones resulting from motor-vehicle crashes). 

The analysis revealed that, globally, about one in three people, or about 2.4 billion individuals, drink alcohol — 25 percent of women and 39 percent of men. It also showed that the global average number of daily standard drinks is 0.7 for women and 1.7 for men. (For purposes of the study, the researchers defined a standard alcoholic drink as 10 grams of alcohol. That’s less than the 14 grams used in the United States to define a standard drink.)

The researchers then assessed the health risks, including premature death, associated with alcohol consumption. They estimated that 6.8 percent of all deaths among men and 2.2 percent of all deaths among women in 2016 were linked to alcohol-related health problems.

Those numbers would make alcohol consumption the seventh leading risk factor for premature death and disability worldwide — and the leading one for people aged 15 to 49 years. Indeed, the researchers linked alcohol consumption to about one in 10 deaths in the 15-49 age group in 2016, primarily from tuberculosis (1.4 percent of deaths), road injuries (1.2 percent) and self-harm (1.1 percent). 

For people aged 50 and older, however, cancer was the leading cause of death associated with alcohol consumption, accounting for 27.1 percent of deaths among women and 18.9 percent of deaths among men, according to the analysis.

Alcoholic-related cancers were most common in high-income countries. In low-income countries, the leading cause of alcohol-related premature death and disability was tuberculosis, while in high-to-medium-income countries the leading cause was stroke.

Risks vs. potential benefits

The researchers did find that low levels of alcohol consumption (less than the study’s definition of a standard drink per day) may offer some modest protection against heart disease and, perhaps, against stroke and diabetes.

But the health risks associated with alcohol consumption — even at the recommended “moderate” levels — far outweigh any minor potential benefits, the study’s authors insist.

“The widely held view of the health benefits of alcohol needs revising, particularly as improved methods and analyses continue to show how much alcohol use contributes to global death and disability,” they write.  

“Our results show that the safest level of drinking is none,” they add. “This level is in conflict with most health guidelines, which espouse health benefits associated with consuming up to two drinks per day.” 

Putting risk into context

Of course, the problem with this study — as with earlier studies that have suggested that low levels of alcohol are beneficial to health — is that it is based on observational data, which can’t prove cause and effect.

Also, it describes its findings in terms of relative risk. To put the study’s findings in perspective, we need to know the absolute health risk associated with various levels of alcohol consumption. Only then can we understand the true size of the risk to individual drinkers. 

A press release issued by The Lancet does, however, what the study doesn’t. It explains the results in absolute risk: 

Specifically, comparing no drinks with one drink a day the risk of developing one of the 23 alcohol-related health problems was 0.5% higher  —  meaning 914 in 100,000 15–95 year olds would develop a condition in one year if they did not drink, but 918 people in 100,000 who drank one alcoholic drink a day would develop an alcohol-related health problem in a year. 

This increased to 7% in people who drank two drinks a day (for one year, 977 people in 100,000 who drank two alcoholic drinks a day would develop an alcohol-related health problem) and 37% in people who drank five drinks every day (for one year, 1,252 people in 100,000 who drank five alcoholic drinks a day would develop an alcohol-related health problem).

As David Spielgelhalter, a statistician at the University of Cambridge, explains in a blog post, those absolute-risk findings mean that for one person among the one-drink-a-day group to experience an extra [health] problem, 25,000 people need to drink 10 grams of alcohol a day for a year — about the annual equivalent of 16 bottles of gin each.

“That’s a total of 400,000 bottles of gin among 25,000 people, being associated with one extra health problem. Which indicates a rather low level of harm in these occasional drinkers,” Spielgelhalter writes.

Caution and common sense

Still, alcohol is not a benign substance. It takes the lives of more than 88,000 Americans each year, according to the Centers for Disease Control and Prevention.

In fact, alcohol is the second deadliest drug in America. (Tobacco is first; opioids are third.)

So, caution is needed. But so is some common-sense risk assessment. 

“There is no safe level of driving, but governments do not recommend that people avoid driving,” Spielgelhalter told Guardian reporter Sarah Boseley. “Come to think of it, there is no safe level of living, but nobody would recommend abstention.”

FMI: You can read the study in full at The Lancet’s website. You can read Spielgelhalter’s stastical critique of it at his Cambridge University website.

You can also learn about all our free newsletter options.

Comments (13)

  1. Submitted by Colin Brownlow on 08/27/2018 - 09:09 am.



    Thanks for posting the absolute risk numbers. Relative risk can be awfully misleading when the incidence rates are low. My take away from this is that the increased risk from a glass of wine with dinner or beer with friends is minimal. On the other hand the net health benefit from that glass of wine is minimal at best and that low to moderate alcohol consumption should not be promoted as a health protective measure.

  2. Submitted by Ray Schoch on 08/27/2018 - 10:06 am.

    Risk and reward

    A teetotaler myself, I’m aware that there’s ample historical evidence that society-wide prohibition simply doesn’t work, though – as a teetotaler – I wish it were otherwise. Wishing won’t make it so, however, and as long as alcohol continues to be big business in this and other societies, we’ll have to continue to deal with the consequences of its use and abuse.

    For the most part, I’m inclined to agree with Colin Brown: the risks of low-level and occasional use appear to be minimal. However, the health benefits of similar low levels of consumption appear to be minimal, as well.

  3. Submitted by Jim Lit on 08/27/2018 - 11:14 am.

    Take it with a grain of salt

    These studies about health & in this one about alcohol consumption need to taken with a large grain of salt. I live in a senior gated community. My neighbor who will be 93 plays golf 3 times a week, actively involved with the Ladies golf program here & still maintains 2 homes in two different states. Just don’t take here daily glass of wine away from her. My wife who is in her low 80’s, plays golf 4 times a week, is treasurer of the local women’s club plus assists with our summer golf program & again don’t take her 5 p.m. glass of wine away from her or she will be grouchy all night.
    I could go on and on about many of our fellow seniors who enjoy an alcoholic beverage or two, but my point is there are so many other factors dealing with an individuals health that to make a blanket statement regarding normal alcohol drinking affecting your health to me is a stretch.
    One other example & I will get off my soap box. I lost two family members recently, not blood related, both in their mid 70’s to Alzheimer.
    Both never drank.

  4. Submitted by Bill Kahn on 08/27/2018 - 12:17 pm.

    Dunno, but I feel certain that without a good stiff drink for some of us once in a while in this age of Trump, we’d see much worse. We all die of something and life is never safe; fun, beautiful, ugly, and relentless in the experience that substance abuse may enhance or dull, depending on what one figgers one needs, but sometimes some of us want a drink or two once in a while in spite of “lies, damn lies, and statistics.” Okay?

  5. Submitted by ian wade on 08/27/2018 - 12:45 pm.

    Sorry, Susan

    but as long as Donald Trump is in office, I’m going to take my chances.

  6. Submitted by chuck holtman on 08/27/2018 - 02:04 pm.


    I read a summary of research from the reputable Jack Daniels Institute of Medical Research and it appears to arrive at some very different conclusions. I guess we just don’t know whom to believe.

    Until it gets sorted out, I’ll just have myself a little refreshment, thanks.

  7. Submitted by Robert Owen on 08/27/2018 - 03:37 pm.

    Good to know

    So the science isn’t exactly settled on this topic.

  8. Submitted by Pat Terry on 08/27/2018 - 05:50 pm.


    Having an occasional beer or two does me a lot of good. Maybe not for my body, but definitely for my mental health.

  9. Submitted by Ron Gotzman on 08/27/2018 - 07:45 pm.

    Who cares about science?

    All it takes is a lawyer to convince a jury about the “evils” of alcohol and the booze companies and all will be paying more and government will be the beneficiaries.

  10. Submitted by Susan Lesch on 08/28/2018 - 07:37 am.

    Thank you for citing Dr. Spiegelhalter who put some sanity back into this story. Epidemiology is rarely worth the time of day.

  11. Submitted by Paul Udstrand on 08/28/2018 - 10:12 am.


    Epidemiology is indeed a complex matter. For instance, look at the alcohol “related” finding: 914 (per 100,000) non drinkers will develop a certain medical condition, while 918 drinkers will develop the same condition; but that same condition is attributed to alcohol consumption in one group but not the other? In reality the 914 of the 918 would have the same chances of developing the condition whether they drank or not, you can’t attribute 100% of the disease rate to drinking, just because that group drinks. You have to have actual data that the drinkers with the same conditions didn’t acquire those conditions the same way the non-drinkers acquired it. Basically this would be a case control study that yields a very low odds ratio for drinking, 4 more cases out of a 100,000. That’s .004%

    I know the addiction industry is invested in a narrative that no amount of alcohol is “risk” free, but the fact is always that some people develop problems and some don’t, that doesn’t mean statistically that EVERYONE who drinks is at risk.

    As for the risks outweighing the “benefits”, that almost has to be a garbage analysis because the benefits, if any, would not accrue to populations, but rather to individuals, and they can’t be clearly defined. What “benefit” are you going to look for?

    • Submitted by Colin Brownlow on 08/28/2018 - 03:45 pm.

      Here’s the fallacy in your argument

      If in fact alcohol has a health benefit, due to the biological activity of the alcohol, then there should be a clear pattern of health improvement or benefit that is measurable across the population. That being the case you can in fact compare magnitude of health risk and magnitude of health benefit.

      • Submitted by Paul Udstrand on 08/29/2018 - 08:53 am.

        Fallacy shmallacy

        You’re logic is impeccable and correct sir. However, this study didn’t try to confirm health benefits, it measured risks compared to benefits. It didn’t conclude that there are NO benefits, merely that whatever benefits there may be, are nullified by the risk.

        I don’t think most people drink alcohol for their health, the way they take vitamins or exercise for instance. These studies that find some health benefit usually find a very small possible benefit, and when they pop up in places like Minnpost, we just say: “cheers” and go about our business. The health benefit claim is almost irrelevant from an empirical perspective. It’s the risk analysis that we’re concerned with, and the finding that consuming any alcohol at all puts anyone who does so at risk… is a dubious finding. Statistical probability doesn’t tell you what will actually happen in any individual case, hence the 95 year old beer drinker or the non-drinker who acquires the same medical condition some drinkers acquire.

        Look: I’m not saying alcohol isn’t a social problem, or a medical problem for some people, but over-generalizing the risk probably isn’t helpful either.

Leave a Reply