Alcohol, health benefits and health risks
For decades we were told that moderate drinking was good for us. That claim has been steadily unwound, while the risks have only become clearer. Here is what the evidence actually shows now — so you can decide with your eyes open.
The idea that a moderate amount of alcohol is healthful has been repeated so often, for so long, that it came to feel like settled fact. But three questions were always doing a lot of quiet work underneath it: how much is "moderate," which benefits were real, and how those benefits stacked up against the risks. As researchers have gone back and answered them more carefully, the comfortable version of the story has not survived.
The case that was built
From the 1970s onward, a large body of research reported that people who drank moderately lived longer and had less heart disease than people who did not drink at all. Red wine drew particular attention for its antioxidant compounds, and the so-called "French paradox" — rich diets alongside relatively low heart-disease rates — became shorthand for the whole idea. Studies pointed to plausible mechanisms: moderate drinking seemed to raise HDL (the "good" cholesterol), reduce blood clotting, and correlate with lower rates of coronary heart disease. Some large reviews reported reductions in heart-disease risk of 30 percent or more among moderate drinkers, and the finding appeared to hold across sexes, regions, and backgrounds. It was on this foundation that "a glass a day" became conventional wisdom.
Why that case no longer holds
The problem lay in who those studies were comparing drinkers against. The "non-drinker" group often included people who had stopped or cut down precisely because they were already ill or ageing. Counting these former drinkers as lifelong abstainers made the abstaining group look sicker than it was — and made moderate drinkers look healthier by comparison. This flaw is known as the abstainer, or "sick-quitter," bias.
When researchers corrected for it, the benefit faded. A 2023 systematic review in JAMA Network Open pooled 107 studies covering nearly 4.8 million people; once the bias was accounted for, the apparent protective effect of low and moderate drinking on death rates disappeared. Studies using Mendelian randomization — a genetic method that avoids much of the confounding in ordinary surveys — likewise found no protective effect, and saw risk rising even at low intakes. In 2025 the American Heart Association, long associated with the heart-health message, advised clinicians to step back from telling people to drink for their hearts.
There is a further point that reframes the whole "which drink is healthiest" debate: the harm, and whatever effect there is, comes from the alcohol itself — not from the wine, the beer, or the phenols. The resveratrol in red wine is real, but the doses used in laboratory studies are far beyond anything a person could drink, and taking it as a supplement has not delivered the promised benefits. Ethanol is ethanol, whatever glass it arrives in.
How much is "moderate"?
Definitions vary, which is part of the confusion. In the United States, one standard drink contains about 14 grams of pure alcohol — roughly a 5-ounce glass of wine, a 12-ounce beer, or 1.5 ounces of spirits. US dietary guidance defines moderate drinking as up to one drink a day for women and up to two for men, and stresses that people who don't drink shouldn't start.
Elsewhere the limits have dropped sharply. Canada's 2023 guidance concluded that risk climbs beyond just two drinks per week, and the World Health Organization's position is blunter still: when it comes to health, no amount of alcohol is safe. The gap between these numbers is itself a signal of how much the thinking has shifted.
The risks, which are well established
Where the benefits have grown uncertain, the risks are not in doubt — and this is where eyes-open drinking matters most.
Cancer. Alcohol is classified by the International Agency for Research on Cancer as a Group 1 carcinogen, the highest category, alongside tobacco and asbestos. It is linked to at least seven cancers, including cancers of the bowel and the female breast, and the risk is dose-dependent, beginning at low levels. The US Surgeon General's 2025 advisory estimated that alcohol causes roughly 100,000 cancer cases and 20,000 cancer deaths a year in the United States, and called for cancer warning labels on alcoholic drinks. For breast cancer in particular, risk rises measurably with each additional drink, which is why women with a family history have specific reason for caution.
Dependence and social harm. Alcohol is the most widely used drug in the country, and misuse is common. According to the 2024 National Survey on Drug Use and Health, about 27.9 million Americans aged 12 and older had an alcohol use disorder in the past year — and only a small fraction received any treatment.
Drink-driving. In 2023, about 12,400 people died in alcohol-impaired driving crashes in the United States — close to a third of all traffic deaths. Getting behind the wheel after drinking is the most immediate and preventable harm alcohol causes.
Liver disease. The link between heavy, sustained drinking and liver damage has been understood for generations. Cirrhosis (scarring of the liver) and alcoholic hepatitis (inflammation) are common and well-documented results of drinking too much for too long.
Pregnancy. Alcohol during pregnancy can cause fetal alcohol spectrum disorders, which may involve low birth weight, developmental delays, organ problems, and lifelong effects. Guidance is that there is no known safe amount of alcohol during pregnancy.
Sleep and weight. Even short of these serious harms, alcohol disrupts sleep — which in turn dulls memory, reaction time, and mood — and it is calorie-dense. A 5-ounce glass of wine carries roughly 120 calories and a 1.5-ounce measure of spirits around 100, and mixers add more; the calories add up quickly and carry no nutritional benefit.
So where does that leave you?
None of this is a verdict on whether you should pour a glass tonight. Wine and beer and spirits are woven into food, celebration, and culture in ways that have little to do with biomarkers, and plenty of people will weigh the pleasure against the risk and decide it's worth it. The point is simply that the decision should be made with the real picture in view: the health halo that once tipped the scales toward "good for you" has come off, and what remains is a genuine, personal trade-off.
Risk rises with how much and how often you drink, and it varies from person to person — anyone who is pregnant, has a family or personal history of cancer or liver disease, or takes medication that interacts with alcohol has more reason to be careful. The oldest advice still holds best: don't rely on any single article, this one included. Read widely, weigh it against your own circumstances, and if it matters to your health, talk it through with your doctor. Then the choice is yours to make.
This article is general information, not medical advice.
The studies behind the old consensus
It is worth seeing where the "moderate drinking is good for you" idea came from, because the studies were real, large, and published in respected journals. They were also observational, and they share the abstainer-bias weakness described above. Three were especially influential:
- The Nurses' Health Study followed some 87,500 female nurses aged 34 to 59 and reported that moderate drinkers had roughly a 40 percent lower risk of coronary heart disease. (Stampfer et al., New England Journal of Medicine, 1988.)
- The Physicians' Health Study followed 22,071 male physicians aged 40 to 84 and reported about a 30 to 35 percent lower risk of angina and heart attack among moderate drinkers. (Camargo et al., Annals of Internal Medicine, 1997.)
- An American Cancer Society cohort of roughly 490,000 men and women aged 30 to 104 reported a 30 to 40 percent lower risk of cardiovascular death among moderate drinkers over nine years. (Thun et al., New England Journal of Medicine, 1997.) Notably, the same study found higher cancer death rates among heavier drinkers.
These findings shaped a generation of health advice. But they compared drinkers against a mixed group of non-drinkers that included people who had quit because they were already ill — the flaw that later, better-controlled work set out to fix. When a 2023 review corrected for it across 107 studies and nearly 4.8 million people, the mortality benefit disappeared. The older studies are best read today as a record of what was believed, and why, rather than as current guidance.
Drink responsibly
Whatever you decide, the sensible course has not changed: if you drink, do so in moderation, never before driving, and never while pregnant.