Alcohol causes cancer at every level of drinking. Most people do not know it.

In January 2025 the US Surgeon General issued a public advisory stating that alcohol is a leading preventable cause of cancer, and that most Americans do not know it [s2]. The advisory is not the first authority to reach that conclusion, and the evidence it rests on is not new. What is new is that a top government health office said it plainly, and asked that alcoholic drinks carry a cancer warning label [s2].

What the advisory claims

The advisory attributes roughly 100,000 cancer cases and about 20,000 cancer deaths in the United States each year to alcohol, which it places third among preventable causes of cancer, behind tobacco and obesity [s2]. It links alcohol to seven cancer types: cancers of the mouth, throat, larynx, oesophagus, liver, breast and colorectum [s2].

Its central public-health point is about awareness rather than biology. The advisory states that a majority of Americans are unaware of the alcohol-cancer link, and calls for updated warning labels, revised guidelines and public-education efforts to close that gap [s2].

The global numbers behind it

The advisory's US figures sit inside a larger picture. A population-based study in The Lancet Oncology estimated that 741,300 new cancer cases worldwide in 2020 — 4.1% of all new cancers — were attributable to alcohol [s1]. Men accounted for 568,700 of those cases, or 76.7% [s1].

The cancers alcohol drives most are the ones the tissue meets directly or the liver processes: oesophageal cancer (189,700 cases), liver cancer (154,700) and breast cancer (98,300) were the largest contributors [s1]. The attributable fraction varied sharply by region, from 0.3% in northern Africa to 5.7% in eastern Asia [s1].

The finding that matters most is about dose

The intuitive assumption is that alcohol-related cancer is a disease of heavy drinkers. The data do not fully support that. Heavy drinking, defined as more than 60 g of alcohol a day, accounted for the largest share — 346,400 cases, or 46.7% of the alcohol-attributable total [s1]. But risky drinking of 20 to 60 g a day added 291,800 cases (39.4%), and moderate drinking of less than 20 g a day still contributed 103,100 cases (13.9%) [s1]. Drinking of up to 10 g a day — roughly one small drink — was linked to an estimated 41,300 cases on its own [s1].

That is the point the Surgeon General's advisory and a 2023 statement in The Lancet Public Health both stress: the cancer risk begins at low levels of consumption and rises with the amount, without a threshold below which it disappears [s2][s3]. The Lancet Public Health authors note that alcoholic beverages have been classified by the International Agency for Research on Cancer as a Group 1 carcinogen — the category reserved for agents with sufficient evidence of causing cancer in humans, alongside tobacco and asbestos [s3].

Why the link is real, and why it is under-recognised

The biological case is not circumstantial. Ethanol is metabolised to acetaldehyde, a compound that damages DNA, and the Lancet Public Health statement identifies this and related pathways as the mechanisms by which alcohol contributes to cancer [s3]. The dose-response pattern in the global data is consistent with a causal relationship rather than a statistical artefact [s1].

The recognition gap is the striking part. The Surgeon General's advisory frames low public awareness as itself a health problem, because people cannot weigh a risk they do not know exists [s2]. Alcohol's cultural position — legal, social, marketed as part of ordinary life — sits awkwardly next to its classification alongside the most established human carcinogens [s3].

What the evidence does not say

The numbers are estimates built on population attributable fractions, not counts of individuals, and they carry wide uncertainty intervals: the global total of 741,300 cases has a range of 558,500 to 951,200 [s1]. Attributable-fraction methods assign a share of cancers to alcohol across a population; they do not predict whether any one person will develop cancer, and most people who drink moderately will not [s1].

The relative increase in risk from light drinking is small in proportional terms, even though it scales to a large absolute number of cases across a whole population [s1]. That distinction — between a modest personal risk and a substantial population burden — is one the advisory's warning does not spell out, and it is the part a reader most needs in order to make sense of the figure [s2].

This article describes what the advisory and the studies found. It is not medical advice, and questions about personal cancer risk belong with a clinician.

Sources

• Global burden of cancer in 2020 attributable to alcohol consumption: a population-based study — The Lancet Oncology, 2021-07-30
• Alcohol and Cancer Risk: The U.S. Surgeon General's Advisory — US Surgeon General, HHS, 2025-01-03
• Health and cancer risks associated with low levels of alcohol consumption — The Lancet Public Health, 2023-01-01