The GBD 2016 Alcohol Collaborators pooled 694 data sources and 592 prospective and retrospective studies covering 195 countries and territories. They found that in 2016, alcohol use was responsible for 2.8 million deaths globally - about 6.8% of all male deaths and 2.2% of all female deaths. Among people aged 15 to 49, alcohol was the single leading risk factor for death and disability. The researchers calculated the "theoretical minimum risk exposure level" and concluded it was zero standard drinks per week. In other words, no amount of alcohol was associated with reduced overall health risk once all harms (cancer, injuries, infectious disease, cardiovascular disease) were combined. Published in The Lancet (Vol. 392, pp. 1015-1035).

Anderson, Berdzuli, Ilbawi, and colleagues published a WHO position statement clarifying that when it comes to cancer risk, there is no identified threshold below which alcohol becomes safe. The statement noted that alcohol is classified as a Group 1 carcinogen (the highest certainty category) by the International Agency for Research on Cancer and is linked to at least seven types of cancer. The authors argued that any potential cardiovascular benefits at light-to-moderate levels do not outweigh the cancer risk for individual consumers. This statement formalized the WHO's official position that the safest level of alcohol consumption for health is none at all. Published in The Lancet Public Health (Vol. 8, pp. e6-e7, 2023).

Bagnardi, Rota, Botteri, and colleagues conducted a comprehensive dose-response meta-analysis pooling data from 572 studies that included 486,538 cancer cases. They found that even light alcohol consumption (up to one standard drink per day) was associated with a statistically significant increased risk of several cancers: oral and pharyngeal cancer risk rose by 13%, esophageal squamous cell carcinoma by 26%, and breast cancer by 5%. At heavy drinking levels, the risks were much larger - oral cancer risk increased by about 5 times and esophageal cancer by about 5 times compared to non-drinkers. The study confirmed a clear dose-response relationship, meaning the more alcohol consumed, the higher the cancer risk, with no safe lower threshold identified. Published in the British Journal of Cancer (Vol. 112, pp. 580-593).

Biddinger, Emdin, Haas, and colleagues used Mendelian randomization - a method that uses genetic variants to separate cause from correlation - to study 371,463 UK Biobank participants with a mean age of 57. In standard observational analysis, light-to-moderate drinkers appeared to have lower cardiovascular risk. But when the researchers used genetic instruments to control for confounding lifestyle factors (light drinkers tend to exercise more, eat better, and have higher income), the protective effect vanished. Instead, each standard-deviation increase in genetically predicted alcohol consumption was linked to 1.3 times higher risk of hypertension and 1.4 times higher risk of coronary artery disease. Heavier consumption showed exponentially increasing cardiovascular risk. This study is important because it suggests the "heart-healthy" effect of moderate drinking was a statistical illusion caused by healthy lifestyle confounders. Published in JAMA Network Open (Vol. 5, e223849, 2022).

Daviet, Aydogan, Jagannathan, and colleagues studied 36,678 generally healthy adults from the UK Biobank who had undergone brain MRI scans. They found a negative association between alcohol intake and both gray matter and white matter volume that was already apparent at just 1 to 2 daily alcohol units. In 50-year-olds, going from 1 drink per day to 2 was associated with brain changes equivalent to 2 years of aging. Going from 2 to 3 drinks was equivalent to 3.5 years of aging. The relationship was not linear but exponential - each additional drink caused progressively more damage. The fornix, a brain structure critical for memory that connects the hippocampus, was one of the most affected regions. This study's large sample size and use of objective brain imaging rather than self-reported outcomes make it particularly strong evidence. Published in Nature Communications (Vol. 13, p. 1175, 2022).

Topiwala, Wang, Ebmeier, and colleagues studied 20,965 UK Biobank participants using brain MRI scans and cognitive tests. They found that consuming 7 or more alcohol units per week (roughly 1 drink per day) was associated with higher iron levels in the basal ganglia, a region deep in the brain involved in movement and decision-making. Elevated iron in this region has been linked to Alzheimer's disease and Parkinson's disease. Participants with higher basal ganglia iron also performed worse on tests of executive function (the ability to plan and switch between tasks), had lower fluid intelligence (the ability to solve new problems), and had slower reaction times. The researchers used both observational analysis and Mendelian randomization (genetic analysis) and found consistent results with both methods, strengthening the case that alcohol is the cause rather than a coincidence. Published in PLOS Medicine (Vol. 19, e1004039, 2022).

Zhao, Stockwell, Naimi, and colleagues conducted a meta-analysis of 107 cohort studies covering more than 4.8 million participants. After adjusting for key study characteristics including median age and sex distribution, they found no significant reduction in all-cause mortality risk for drinkers consuming less than 25 grams of ethanol per day (roughly 2 standard drinks). Significantly increased death risk appeared among female drinkers at 25 or more grams per day and male drinkers at 45 or more grams per day. The researchers showed that previous studies reporting a protective effect of moderate drinking suffered from methodological biases, most notably including former drinkers (who may have quit due to illness) in the non-drinking reference group. Once these biases were corrected, the survival advantage of moderate drinking disappeared. Published in JAMA Network Open (Vol. 6, e236185, 2023).

Sohi, Rehm, Engeland, and colleagues systematically reviewed 23 prospective cohort publications examining alcohol and breast cancer. They found a clear dose-response relationship starting below a single daily drink. Women who consumed less than 1 standard drink per day had a 4% higher risk of breast cancer compared to non-drinkers. At 1 drink per day, the risk increased by 10%. At 2 drinks per day, the risk increased by 18%. At 3 drinks per day, the risk increased by 22%. Postmenopausal women were more vulnerable, showing a 10% increased risk per additional daily drink compared to 3% in premenopausal women. Since breast cancer is the most common cancer in women worldwide, even these seemingly small percentage increases translate to large numbers of additional cases at the population level. Published in Alcohol: Clinical and Experimental Research (Vol. 48, pp. 2268-2281, 2024).

The U.S. Surgeon General released an official advisory in January 2025 stating that alcohol consumption is causally linked to at least seven types of cancer: mouth, throat, voice box, esophagus, liver, breast, and colorectal. The advisory noted that alcohol contributes to approximately 75,000 new cancer cases and nearly 19,000 cancer deaths each year in the United States, making it the third leading preventable cause of cancer after tobacco and obesity. The advisory called for updated cancer warning labels on alcoholic beverages, noting that current labels (unchanged since 1988) do not mention cancer risk. Even light-to-moderate drinking was identified as contributing to cancer risk, particularly for breast cancer. The Surgeon General emphasized that most Americans are unaware of the alcohol-cancer connection. Published by the U.S. Department of Health and Human Services.

Stockwell, Zhao, Panwar, and colleagues analyzed 87 studies covering 3,998,626 individuals and 367,103 deaths. They demonstrated that the widely reported finding that moderate drinkers live longer than abstainers is largely a result of a methodological error called "abstainer bias" or "sick quitter" bias. Many studies classified former drinkers (people who stopped drinking, often because of health problems) as "abstainers," which made the non-drinking group look sicker than it actually was. When the researchers limited their analysis to studies that properly separated lifetime abstainers from former drinkers and used younger cohorts (less affected by survivorship bias), the apparent protective effect of low-volume drinking vanished entirely. The adjusted risk for low-volume drinkers was 0.98 (essentially no difference from abstainers), compared to the biased estimate of 0.84 (a seemingly large 16% benefit). Published in the Journal of Studies on Alcohol and Drugs (Vol. 77, pp. 185-198, 2016).