Key Takeaways
  • The new study is a model, not a trial. It combined evidence from more than 7,200 papers with US health data, found no net protective effect at any level, and estimated a 1 in 25 lifetime risk of an alcohol-caused death for men averaging 14 drinks a week
  • The "J-curve" has faded. Non-drinker groups often included people who quit for health reasons. In a 2023 meta-analysis of 107 cohorts and 4.8 million people, the apparent benefit of light drinking was no longer significant once study quality was accounted for
  • Cancer is the clearest harm. Alcohol is a Group 1 carcinogen. The US Surgeon General links it to nearly 100,000 US cancer cases and about 20,000 cancer deaths a year, whatever the drink
  • Guidelines are converging on less. Canada calls 1 to 2 drinks a week low risk. The 2025โ€“2030 US Dietary Guidelines dropped numeric limits and say "consume less alcohol for better overall health"
  • Practical upshot: don't start drinking for health. If you drink, fewer is lower risk, and cutting back lowers the risk of some cancers

"Even moderate drinking linked to cancer and early death," ran a ScienceDaily headline last week. It reports a US study published in the summer, and it lands on an old belief: isn't a glass of red wine good for the heart? Here's what the study showed, why the "J-curve" has faded, and what official guidance now says. It isn't medical advice: if you take medication, have a health condition or worry about your drinking, talk to your doctor.

1 in 25
Estimated lifetime risk of an alcohol-caused death for US men averaging 14 drinks a week, the old upper limit. A modelling estimate (Journal of Studies on Alcohol and Drugs, 2026)
Part one ยท The news

What the New Study Actually Did

Evidence: Modelling study โ€” population-level estimates

The Alcohol Intake and Health Study, by a team including Katherine Keyes of Columbia University and Kevin Shield of the University of Toronto, appeared in the July 2026 Journal of Studies on Alcohol and Drugs. It was originally commissioned by the US federal government to inform the new Dietary Guidelines (EurekAlert; ASCO Post).

It is not a cohort study, and it doesn't report hazard ratios. Experts reviewed more than 7,200 scientific articles to set a risk level for each alcohol-related disease and injury. The team applied those risks to US survey and census data, CDC death data and IHME illness estimates, and modelled lifetime risk at different average intakes (George, J Stud Alcohol Drugs 2026).

The findings:

The authors concluded that US guidance should be tightened to no more than 1 drink a day for both men and women. An accompanying editorial by Robert M. Vincent argued that "despite the study's adherence to its mandate, its findings were sidelined" when the new US guidelines were written (ScienceDaily).

The headlineEven moderate drinking is linked to cancer and early death.
What the study showsA model, not a new cohort. Risk rises from low levels with no net benefit anywhere. These are population averages with wide uncertainty, not a forecast for any one person.
Part two ยท The J-curve

Why the J-Curve Looks Weaker Now

Evidence: Large meta-analyses and genetic studies

For decades, studies drew a J-shaped curve: light drinkers died at slightly lower rates than non-drinkers, with risk climbing at higher intakes. The catch is who the "non-drinkers" were. Many studies failed "to remove 'sick quitters' or former drinkers, many of whom cut down or stop for health reasons" (Zhao, JAMA Netw Open 2023). Put sick people in the comparison group and light drinking looks protective even if it isn't.

107 cohorts, 4.8 million people

A 2023 meta-analysis pooled 107 cohort studies with 4,838,825 participants and 425,564 deaths, comparing drinkers with lifetime non-drinkers. Before adjusting for study design, low-volume drinkers (1.3 to 24 g of alcohol a day, under two US drinks) had a 15% lower risk of death (relative risk 0.85). After adjusting for study characteristics, including abstainer biases, cohort age and sex, and whether studies controlled for smoking, it was 0.93, not statistically significant. Risk was significantly higher for women drinking 25 g a day or more and men drinking 45 g or more. "Low-volume alcohol drinking was not associated with protection against death from all causes," the authors wrote.

Figure 1 ยท Zhao et al., 2023

Risk of death vs lifetime non-drinkers, by daily intake

Relative risk of death (non-drinkers = 1.00) Low-volume before adjustment 0.85 Occasional <1.3 g/day 0.96* Low-volume 1.3โ€“24 g/day 0.93* Medium 25โ€“44 g/day 1.05* High 45โ€“64 g/day 1.19 Higher 65+ g/day 1.35

Zhao et al., JAMA Network Open 2023. Grey bar: before adjustment for study characteristics; other bars fully adjusted. *Not statistically significant. A US drink is about 14 g. Observational data.

Wood 2018: 599,912 current drinkers

Another way around the sick-quitter problem is to leave non-drinkers out. A 2018 Lancet analysis combined data on 599,912 current drinkers without cardiovascular disease from 83 prospective studies in 19 high-income countries; 40,310 died. The authors excluded non-drinkers partly because "ex-drinkers include people who might have abstained from alcohol owing to poor health itself" (Wood, Lancet 2018).

The lowest risk of death was at about 100 g of alcohol a week or less, roughly seven US drinks. Compared with drinkers having up to 100 g a week, those drinking 100 to 200 g, 200 to 350 g and more than 350 g "had lower life expectancy at age 40 years of approximately 6 months, 1โ€“2 years, or 4โ€“5 years, respectively." Each extra 100 g a week was linked to higher risk of stroke (hazard ratio 1.14) and heart failure (1.09), though slightly lower risk of heart attack (0.94). The authors called for lower limits than "most current guidelines."

What the genes say

Genetic (Mendelian randomisation) studies use gene variants that nudge people to drink more, which are less tangled up with lifestyle. In a 2022 JAMA Network Open study of 371,463 UK Biobank participants, light-to-moderate drinkers exercised more, ate more vegetables and smoked less than abstainers, and accounting for a few such factors substantially reduced alcohol's apparent benefit. In the genetic analysis, cardiovascular risk rose minimally from 0 to 7 drinks a week, much more from 7 to 14, and was especially high at 21 or more (Biddinger, JAMA Netw Open 2022).

Global Burden of Disease: "zero", then a nuance

The Global Burden of Disease (GBD) team's 2018 Lancet analysis ranked alcohol the seventh leading risk factor for death and disability worldwide in 2016 and concluded that "the level of consumption that minimises health loss is zero" (GBD 2016 Alcohol Collaborators, Lancet 2018).

The 2022 update looked at age and region separately, using a 10 g standard drink. For people aged 15 to 39, the healthiest level ranged from 0 to 0.603 drinks a day. For people 40 and over, it ranged from 0.114 to 1.87 drinks a day, depending on region, sex and age (GBD 2020 Alcohol Collaborators, Lancet 2022). In this older group, small amounts were linked to lower risks of conditions such as ischaemic heart disease, stroke and diabetes (News-Medical). Of the 1.34 billion people drinking harmful amounts in 2020, 59.1% were aged 15 to 39. Senior author Emmanuela Gakidou: "Our message is simple: young people should not drink, but older people may benefit from drinking small amounts" (News-Medical). That possible benefit rests on the same kind of observational heart data questioned above, and tops out below two 10 g drinks a day.

Part three ยท Cancer

Alcohol and Cancer Risk

Evidence: Strong โ€” Group 1 carcinogen

The International Agency for Research on Cancer (IARC) classifies alcohol as a Group 1 carcinogen, which the Surgeon General calls "the highest level of classification." In January 2025, the US Surgeon General's advisory called alcohol the third leading preventable cause of cancer in the US, after tobacco and obesity. It contributes to "nearly 100,000 cancer cases and about 20,000 cancer deaths each year," across at least seven cancers: breast (in women), colorectum, oesophagus, liver, mouth, throat and voice box. Risk holds "regardless of the type of alcohol (e.g., beer, wine, and spirits)," and for breast, mouth and throat cancers "this risk may start to increase around one or fewer drinks per day" (US Surgeon General, 2025). Only 45% of Americans recognised this risk in a 2019 survey.

One of the best-established mechanisms is acetaldehyde, which alcohol breaks down into: "a metabolite that causes cancer by binding to DNA and damaging it." Alcohol also drives inflammation, alters hormones such as oestrogen, and helps carcinogens in tobacco smoke get absorbed.

Figure 2 ยท US Surgeon General, 2025

Lifetime risk of an alcohol-related cancer, by daily drinks

% who develop one, over a lifetime WOMEN <1 drink/week 16.5% 1 drink/day 19.0% 2 drinks/day 21.8% MEN <1 drink/week 10.0% 1 drink/day 11.4% 2 drinks/day 13.1%

US Surgeon General's Advisory (2025), recalculated from an Australian cohort of 226,162 adults aged 45 and over. Risk up to age 80 of any of seven alcohol-related cancers. Very light drinkers, not non-drinkers, are the baseline, to avoid the sick-quitter effect.

Breast cancer, drink by drink

A 2017 report from the American Institute for Cancer Research and World Cancer Research Fund, covering 119 studies with data on 12 million women and 260,000 breast cancer cases, found that about 10 g of alcohol a day, "the equivalent of a small glass of wine or beer," was linked to a 5% higher risk of breast cancer before menopause and 9% after (AICR). In absolute terms, the Surgeon General's advisory puts a woman's lifetime breast cancer risk at about 11.3% with under one drink a week, 13.1% with one a day and 15.3% with two.

Does cutting back help?

For some cancers, yes. In 2023, an IARC panel of 17 experts from eight countries found "sufficient evidence that reduction or cessation of consumption of alcoholic beverages can reduce risk" of mouth and oesophageal cancer, and "limited evidence of risk reduction for laryngeal, colorectal and breast cancers" (National Cancer Institute, on Gapstur, NEJM 2023).

Part four ยท The guidelines

What Official Guidance Says Now

Evidence: Official guidance

The US Dietary Guidelines for Americans, 2025โ€“2030, released in January 2026, dropped the old limits of up to two drinks a day for men and one for women and give no number. They say: "Consume less alcohol for better overall health." They say some people should avoid alcohol completely (see below), and tell those with a family history of alcoholism to "be mindful" (USDA/HHS, 2026). Timothy Naimi, a co-author of the new study, said the guidelines "contain a useful 'less-is-best' message" but "provide no quantitative framework" (EurekAlert).

Canada's Guidance on Alcohol and Health (2023) describes a continuum: "1โ€“2 standard drinks per week is low risk," "3โ€“6 standard drinks per week is a moderate risk," and "7 or more standard drinks per week is an increasingly high risk." If you drink, it advises no more than 2 drinks on any day (CCSA). The UK Chief Medical Officers (2016) advise men and women that "it is safest not to drink more than 14 units a week on a regular basis," spread over three or more days, and say "there is no level of regular drinking that can be considered as completely safe in relation to some cancers" (UK CMOs).

Figure 3 ยท National guidance

Drinking guidelines, side by side

SourceWeekly guidanceStandard drink
US Dietary Guidelines 2025โ€“2030No number: "consume less alcohol" old limit: up to 2/day men, 1/day women14 g 12 oz beer, 5 oz wine, 1.5 oz spirits
Canada, 20231โ€“2 low risk; 3โ€“6 moderate; 7+ increasingly high no more than 2 on any day341 ml beer at 5%, 142 ml wine at 12%, 43 ml spirits at 40%
UK Chief Medical Officers, 2016No more than 14 units about 112 g; spread over 3+ days1 unit = 8 g 14 units โ‰ˆ 6 pints of average beer
Alcohol Intake and Health Study authors, 2026No more than 1 drink a day men and womenUS drinks

Sources: Dietary Guidelines for Americans 2025โ€“2030; NIAAA; CCSA (2023); UK CMOs (2016); NHS. 14 UK units equal about 8 US standard drinks.

Part five ยท What to do

Drinking Sensibly, If You Drink

Know what a drink is

A US standard drink contains about 14 g of pure alcohol: 12 oz of beer at about 5%, 5 oz of wine at about 12%, or 1.5 oz of spirits at about 40%. The NIAAA notes these amounts "may not reflect customary serving sizes": a 12 oz beer at 10% is two drinks, and a glass of wine filled well past 5 oz is more than one (NIAAA).

The red wine myth

Red wine's reputation leans on resveratrol, a grape-skin compound. In the InCHIANTI study of 783 adults aged 65 and over in Tuscany, followed from 1998 to 2009, people with the highest levels of resveratrol breakdown products in their urine were no less likely to die than those with none, and resveratrol wasn't linked to inflammation, cardiovascular disease or cancer (Semba, JAMA Intern Med 2014). Lead author Richard Semba of Johns Hopkins said: "The thinking was that certain foods are good for you because they contain resveratrol. We didn't find that at all" (ScienceDaily). And for cancer, the type of drink doesn't matter.

Who shouldn't drink

The US guidelines say people who should completely avoid alcohol include pregnant women, "people who are recovering from alcohol use disorder or are unable to control the amount they drink," and "people taking medications or with medical conditions that can interact with alcohol." Ask your pharmacist about your medications. The CDC and NIAAA also say not to drink if you are under 21 (CDC), and NIAAA adds anyone planning to drive or operate machinery (NIAAA). GBD's estimates suggest under-40s have the least to gain.

Cutting back

Frequently Asked Questions

Is a little alcohol good for your heart?
Probably not in the way once thought. Light drinkers looked healthier in older studies, but a 2023 meta-analysis of 107 cohorts found no significant drop in death rates for low-volume drinkers once study quality was taken into account. A large genetic study found cardiovascular risk rising only slightly from 0 to 7 drinks a week and more steeply after that.
How many drinks a week is low risk?
No level is completely safe for some cancers, according to the UK's Chief Medical Officers. Canada calls 1 to 2 standard drinks a week low risk, 3 to 6 moderate and 7 or more increasingly high. The UK advises no more than 14 units (112 g of alcohol) a week. The authors of a 2026 US study recommend no more than 1 drink a day.
Does even light drinking raise cancer risk?
For some cancers, yes. The US Surgeon General's 2025 advisory says breast, mouth and throat cancer risk may start to rise at around one drink a day or less. Over a lifetime, a woman's risk of breast cancer goes from about 11.3% with under one drink a week to 13.1% with one a day.
Is red wine healthier than beer or spirits?
Not for cancer: the Surgeon General's advisory says the risk applies regardless of the type of drink. And in a study of 783 older Italians, people with the most resveratrol, the much-hyped red wine compound, in their urine did not live longer than those with none.
What counts as one drink?
In the US, a standard drink has about 14 grams of pure alcohol: 12 oz of 5% beer, 5 oz of 12% wine or 1.5 oz of 40% spirits. Many pours are bigger, and a 12 oz beer at 10% counts as two drinks. A UK unit is 8 grams.
The Bottom Line

Less is better, none is fine.

The idea that a daily drink protects your health has not held up well. Account for sick quitters and healthier lifestyles and the benefit shrinks or disappears, while cancer risk starts at low levels and rises with intake. A small benefit for older adults is the most generous reading, and it is uncertain.

If you enjoy a drink, the added risk at a few drinks a week is small, though not zero. Know what a standard drink is, keep most days alcohol-free, keep each occasion small, and talk to your doctor about what's right for you.

Sources

  1. George S, Naimi TS, Keyes K, et al.; Shield K (senior author). Alcohol Intake and Health Study: no protective effect at low levels, with mortality increasing to 1 in 25 at 14 drinks per week. Journal of Studies on Alcohol and Drugs 2026;87(4):621โ€“638 (doi:10.15288/jsad.25-00435). See also the news release (EurekAlert), ScienceDaily and ASCO Post.
  2. Zhao J, Stockwell T, Naimi T, Churchill S, Clay J, Sherk A. Association between daily alcohol intake and risk of all-cause mortality: a systematic review and meta-analyses. JAMA Network Open 2023;6(3):e236185 (doi:10.1001/jamanetworkopen.2023.6185).
  3. Wood AM, Kaptoge S, Butterworth AS, et al. Risk thresholds for alcohol consumption: combined analysis of individual-participant data for 599 912 current drinkers in 83 prospective studies. The Lancet 2018;391(10129):1513โ€“1523; published online 14 April 2018 (doi:10.1016/S0140-6736(18)30134-X).
  4. Biddinger KJ, et al. Association of habitual alcohol intake with risk of cardiovascular disease. JAMA Network Open, published 25 March 2022 (doi:10.1001/jamanetworkopen.2022.3849). As reported in the news release (EurekAlert).
  5. GBD 2016 Alcohol Collaborators. Alcohol use and burden for 195 countries and territories, 1990โ€“2016: a systematic analysis for the Global Burden of Disease Study 2016. The Lancet 2018;392(10152):1015โ€“1035 (doi:10.1016/S0140-6736(18)31310-2).
  6. GBD 2020 Alcohol Collaborators. Population-level risks of alcohol consumption by amount, geography, age, sex, and year: a systematic analysis for the Global Burden of Disease Study 2020. The Lancet 2022;400(10347):185โ€“235; published online 14 July 2022 (doi:10.1016/S0140-6736(22)00847-9). See also News-Medical summary.
  7. Office of the US Surgeon General. Alcohol and Cancer Risk: The US Surgeon General's Advisory. 2025; released 3 January 2025 (Pharmacy Times). HHS overview page.
  8. American Institute for Cancer Research / World Cancer Research Fund. Continuous Update Project Report: Diet, nutrition, physical activity and breast cancer. 2017. AICR news release; Breastcancer.org summary.
  9. Gapstur SM, et al. The IARC perspective on alcohol reduction or cessation and cancer risk. New England Journal of Medicine 2023;389(26):2486โ€“2494; published 28 December 2023 (doi:10.1056/NEJMsr2306723). As summarised by the National Cancer Institute; IARC Handbook Volume 20A summary.
  10. Semba RD, Ferrucci L, Bartali B, et al. Resveratrol levels and all-cause mortality in older community-dwelling adults. JAMA Internal Medicine 2014;174(7):1077โ€“1084 (doi:10.1001/jamainternmed.2014.1582). See also ScienceDaily.
  11. US Department of Agriculture and US Department of Health and Human Services. Dietary Guidelines for Americans, 2025โ€“2030. January 2026. See also AASLD statement.
  12. Canadian Centre on Substance Use and Addiction. Canada's Guidance on Alcohol and Health (news release, 17 January 2023); guidance poster.
  13. UK Chief Medical Officers. UK Chief Medical Officers' Low Risk Drinking Guidelines. August 2016. NHS, Calculating alcohol units.
  14. National Institute on Alcohol Abuse and Alcoholism. What is a standard drink?; Rethinking Drinking, What are the U.S. guidelines for drinking?
  15. Centers for Disease Control and Prevention. About Moderate Alcohol Use. Last reviewed 14 January 2025.