Modeling study finds no net benefit from light drinking, and rising risk

What happened

If you drink a glass of wine most nights, the headline is that the heart-health rationale does not hold up. A study modeling lifetime alcohol risks found no overall health benefit at low levels of drinking, and a higher chance of alcohol-caused death as intake rises.

The coverage is as first reported by ScienceDaily · Health & medicine, which rewrote a Columbia Mailman School press release. That release was posted on EurekAlert on June 9, 2026, about five months earlier. ScienceDaily ran the same study a week before under a different headline.

The paper appeared in the Journal of Studies on Alcohol and Drugs. Lead author Kevin Shield of the University of Toronto worked with Katherine Keyes of Columbia and Timothy Naimi of the University of Victoria.

What the study found

This is a systematic review plus modeling, not a new group of people followed over time. Researchers reviewed scientific articles on alcohol-related diseases and injuries to set the risk for each condition. The researchers then applied those risks to large national US health datasets and modeled lifetime outcomes.

The study's published protocol in BMJ Open describes the method: relative risks from meta-analyses, combined with estimated risks for lifetime abstainers, to produce lifetime absolute-risk curves by cause, age and sex. It also adjusts for under-reporting of drinking in medical studies and for alcohol that is spilled or wasted, each previously estimated at 10%.

Here is a nuance the ScienceDaily summary flattens. The authors say low-level drinking may be associated with a lower risk of ischemic heart disease and stroke. But across all outcomes combined, they found no significant protective effect at any level, and they say the cardiovascular benefit is outweighed by other risks even at 7 drinks a week.

So no overall benefit is accurate. Heart disease risk rising at low levels is not what the release says.

What it does not show

We could not review the full paper, so we worked from the press release, the protocol and secondary coverage. The release gives no sample size, confidence intervals or per-cancer risk figures, and those matter for judging how firm the estimates are.

• It does not track drinkers. The risks are assembled from existing research and modeled, so the output is only as good as the inputs and assumptions.
• It is not personal risk. The authors say these are population-level estimates, and genetics, lifestyle and drinking pattern all shift an individual's odds.
• It is not a full accounting. The authors name esophageal, oral and breast cancers, cardiovascular disease, liver disease and injury as established causal conditions. Evidence on others, such as pancreatic cancer, is still emerging and not captured.
• It does not say one glass causes cancer. The 1-in-25 figure applies to an average of two drinks a day.

What it means for you

The most defensible takeaway is that alcohol is not a health product, and that less is safer than more. If you drink for the supposed heart benefit, this analysis suggests it does not net out in your favor.

If you drink lightly, the release does not describe a cliff. It describes a slope, with risk increasing above one drink a day.

If you have a personal or family history of breast cancer, liver disease or other conditions tied to alcohol, that is a conversation for your own clinician rather than a headline.

Context: funding, politics and guidelines

The study was funded by Synergy Enterprises Inc. and the US Substance Abuse and Mental Health Services Administration. The release says the author reports no conflicts of interest. ScienceDaily's October piece says the project was originally commissioned by the federal government to help shape the dietary guidelines.

An accompanying editorial by Robert M. Vincent, a former SAMHSA associate administrator, says the findings were sidelined in that process. That is one person's opinion, and we could not verify it independently.

The 2025–30 Dietary Guidelines, released January 7, 2026, dropped the old numeric limits of two drinks a day for men and one for women. A Lancet Gastroenterology & Hepatology commentary calls the replacement vague and notes it omits the cancer link, which the 2020–25 edition stated explicitly, even below one drink a day. The same commentary says alcohol contributes to nearly 100,000 new US cancer cases and 20,000 cancer deaths a year. Exact wording of the new guidance varies by source, as one Personal View highlights.

Public awareness has not moved much. An Annenberg survey in February 2026 found 53% of Americans say regular drinking raises cancer risk, statistically unchanged from 56% a year earlier.