Wine with dinner, a Friday beer, the drink poured “just to toast.” Alcohol has been woven into social life across half the planet for centuries, and for just as long it has raised the same uncomfortable question: how much is too much, and is there even an amount that’s safe for the brain? A team of researchers at the Affiliated Hospital of Heilongjiang University of Chinese Medicine has published, in Internal Medicine Journal, the most complete meta-analysis to date on alcohol and dementia risk, and the answer it offers isn’t a simple yes or no.
The study pooled 21 investigations and nearly 24.2 million people. The authors combed PubMed, Embase, Cochrane Library and Web of Science for studies published between 1992 and 2024 linking alcohol consumption to dementia risk. The final analysis combined 21 studies (19 cohort studies and 2 case-control studies) carried out in 14 countries across Europe, North America, Asia and Oceania, covering 24,191,921 participants in total (10,563,581 men and 13,628,340 women) and 1,233,833 recorded dementia cases. Few meta-analyses on this topic have ever worked with this much data.
Taken as a whole, alcohol doesn’t seem to change dementia risk at all. When researchers compared drinkers of any kind against non-drinkers without breaking down quantities, they found no statistically significant association for dementia from any cause (relative risk, RR, of 1.03), Alzheimer’s (RR 0.97), vascular dementia (RR 1.09), or other types of dementia (RR 0.62). At first glance, alcohol would look irrelevant to brain health. But that overall picture hid something far more interesting that only showed up once drinkers were split by how much they actually drank.
Moderate drinking was linked to less Alzheimer’s and less dementia overall. When participants were split into light-to-moderate drinkers (fewer than three drinks a day, or under 49.9 grams of pure alcohol daily), heavy drinkers (three or more drinks a day, or 50 grams or more) and people with alcohol use disorder, the pattern changed completely. Light-to-moderate drinkers showed a 12% lower risk of dementia from any cause (RR 0.88) and a 12% lower risk of Alzheimer’s (RR 0.88) compared with non-drinkers. Vascular dementia, however, showed no clear relationship at this level of consumption (RR 0.96).
Heavy drinking, at the opposite extreme, sent the risk of every type of dementia soaring. Heavy drinkers had an 18% higher risk of dementia from any cause (RR 1.18), a 29% higher risk of Alzheimer’s (RR 1.29) and a 25% higher risk of vascular dementia (RR 1.25). And people who met the criteria for alcohol use disorder (dependence or sustained hazardous drinking) had a risk of dementia from any cause more than three times higher (RR 3.35), the highest figure in the entire analysis.
The benefit of moderation was especially pronounced in Europe and among people aged 60 to 69. Looking only at light-to-moderate drinkers by region and age group, Europe stood out with a 16% reduction in risk of dementia from any cause and 41% for Alzheimer’s. By age, the 60-69 group showed the strongest protection, with 30% less risk of overall dementia and 32% less for Alzheimer’s. The authors point out that alcohol consumption and its cultural patterns vary widely by region (Europe has some of the highest consumption levels in the world), and that the risk profile shifts with age: younger drinkers accumulate more risk through accidents and injuries, while at older ages cardiovascular risk weighs more heavily, something low doses of alcohol appear to modulate.
Why would one drink help while a whole bottle hurts? The authors themselves describe a U-shaped or J-shaped relationship: the lowest risk clusters around light-to-moderate consumption, while both total abstinence and heavy drinking are linked to higher risk. Earlier research suggests moderate drinking may increase total brain volume, reduce white matter lesions and silent brain infarcts, improve blood lipid profiles and lower inflammation markers, all of which could protect against cognitive decline. Heavy, prolonged drinking, by contrast, causes well-documented brain damage: thiamine deficiency and the direct neurotoxic effects of ethanol and acetaldehyde damage neurons over time.
The important caveat: this is observational data, not a trial that proves causation. The authors themselves stress that many people who don’t drink at all include former drinkers who quit precisely because they already had health problems, which can artificially inflate the apparent risk of abstinence (a bias known as reverse causation). The studies also used different definitions of a “standard drink” depending on the country (8 grams of alcohol in the UK, 10 in Australia, 14 in the US), which introduces some imprecision when comparing quantities across studies, and nearly all relied on participants self-reporting how much they drank, a method always vulnerable to inaccurate recall or under-reporting.
The team itself is clear this isn’t an invitation to start drinking. The authors stress that even though light-to-moderate consumption was linked to lower dementia risk, alcohol remains linked to more than 200 diseases, including several cancers, whose risk can rise with even a single daily drink. Their conclusion is explicit: people who don’t drink shouldn’t start in hopes of protecting their brain, and people who already drink heavily or show signs of dependence now have one more, and weighty, reason in dementia risk to cut back.
Source: Zhang R, Li B, Miao Y. “Alcohol consumption and risk of dementia: a systematic review and meta-analysis.” Internal Medicine Journal, 2025. DOI: 10.1111/imj.70288.
