Even light alcohol drinking raises dementia risk, according to largest genetic study to date

A large-scale international study combining genetic data and observational records suggests that alcohol consumption of any amount may raise the risk of developing dementia. While earlier research indicated light drinking could be protective, this new study provides evidence that such findings may reflect reverse causation rather than a true beneficial effect. The authors conclude that reducing alcohol consumption could be an effective public health strategy to lower dementia rates.

The research was led by Anya Topiwala of the University of Oxford’s Nuffield Department of Population Health, alongside colleagues from Yale University, Harvard University, the University of Cambridge, and other institutions. Their findings were published in BMJ Evidence Based Medicine.

By analyzing data from more than half a million participants in the United States and the United Kingdom, and combining this with genetic data from over two million individuals, the researchers sought to clarify a longstanding question: does alcohol truly protect against dementia at low levels, or does any amount carry risk?

Past observational studies have consistently reported what is called a U-shaped or J-shaped relationship between alcohol and dementia. These studies suggested that people who drank lightly or moderately had a lower risk of dementia compared to both non-drinkers and heavy drinkers. This pattern led some to propose that small amounts of alcohol might be protective for brain health. But this idea has been controversial.

Critics have pointed out that many non-drinkers in these studies may have stopped drinking due to health problems, making them appear at higher risk. This could create a misleading picture, where abstainers seem worse off not because they avoid alcohol, but because they have other underlying health issues. Additionally, some studies may have lacked enough data on heavy drinkers to properly capture the full range of alcohol’s effects. To address these issues, the researchers used a combination of traditional and genetic methods, allowing for a more rigorous test of causality.

To investigate this relationship, the researchers drew on two major population-based cohorts: the United States Million Veteran Program and the UK Biobank. Together, these included over 559,000 adults aged 56 to 72 at baseline. During the follow-up periods, which ranged from 4 to 12 years depending on the cohort, more than 14,000 participants developed dementia and over 48,000 died.

Initial observational analyses suggested that both heavy drinkers and non-drinkers had higher dementia rates than light drinkers. Specifically, individuals who drank more than 40 drinks per week or had a history of alcohol use disorder showed increased risk, as did those who abstained entirely. At first glance, this pattern echoed past findings suggesting a possible benefit of light drinking.

However, the researchers then applied a genetic method known as Mendelian randomization. This technique uses genetic variants associated with alcohol consumption as a proxy to estimate lifetime exposure to alcohol. Unlike traditional observational methods, it is less likely to be distorted by reverse causation or unmeasured confounding factors.

These analyses told a different story. They showed a steady increase in dementia risk as alcohol intake increased, without any sign of benefit at lower levels. A genetically predicted increase in weekly alcohol consumption was linked to a 15 percent higher dementia risk. Likewise, genetic risk for alcohol use disorder was associated with a 16 percent increase in dementia risk.

The research team also analyzed how alcohol use changed over time in people who eventually developed dementia. Using repeated clinical alcohol screening data from the Million Veteran Program, they found that individuals who went on to develop dementia tended to reduce their alcohol intake in the years before diagnosis.

This suggests that lower drinking levels in people with early signs of cognitive decline may have contributed to the misleading appearance of a protective effect in past observational studies. In other words, rather than alcohol preventing dementia, it may be that the onset of dementia leads people to drink less.

The genetic analyses were bolstered by data from over 2.4 million individuals and considered a wide range of alcohol-related traits. This included both how much people drank and whether they exhibited problematic or dependent drinking patterns. Results were consistent across different types of alcohol exposure, and the findings held up in several sensitivity tests. Notably, the researchers found no evidence of a non-linear or U-shaped relationship in their genetic analyses. The more alcohol a person was genetically predisposed to consume, the higher their risk of dementia appeared to be.

By combining traditional observational data with genetic approaches, the study offers a more nuanced understanding of alcohol’s potential impact on the brain. The authors suggest that public health policies aiming to reduce the prevalence of alcohol use disorder could help lower dementia rates across populations. In their calculations, halving the number of people with alcohol use disorder could reduce dementia cases by up to 16 percent.

While the study has several strengths, including its large sample size, diverse ancestry representation, and use of multiple analytic methods, it is not without limitations. For example, dementia diagnoses were based on medical records, which may not always be accurate or complete. The genetic data also reflect lifelong tendencies rather than specific drinking patterns during certain life stages. In addition, the strongest findings came from participants of European ancestry, with somewhat weaker signals in individuals of African or Latin American descent, likely due to smaller sample sizes in these groups.

Another limitation relates to how alcohol intake was measured. In many cases, drinking behavior was self-reported, which can lead to underestimation or misclassification. There was also limited information on lifetime drinking history, which might influence risk in ways not captured by current or recent behavior alone.

“Authors rightly acknowledge several important limitations of the study. Self-reported alcohol use may not be accurate, particularly if people have memory problems in early stages of dementia, and the genetic markers used as predictors of both alcohol intake and dementia are not perfect,” Tara Spires-Jones, the director of the Centre for Discovery Brain Sciences at the University of Edinburgh, who was not involved in the study, told the Science Media Centre.

“Neither part of the study can conclusively prove that alcohol use directly causes dementia, but this adds to a large amount of similar data showing associations between alcohol intake and increased dementia risk, and fundamental neuroscience work has shown that alcohol is directly toxic to neurons in the brain.”

Despite these caveats, the study’s combination of approaches provides stronger evidence than most previous efforts. The researchers emphasize that their results challenge the popular belief that light or moderate alcohol consumption is safe—or even beneficial—for brain health. While this notion has been widely accepted for years, it may have been based on misinterpreted data. The genetic evidence in this study suggests that any level of alcohol exposure could increase the risk of dementia, and that public health messages should reflect this possibility.

Future research may explore whether specific types of alcohol, patterns of drinking, or interactions with other lifestyle or genetic factors influence dementia risk differently. Researchers may also investigate whether the mechanisms linking alcohol to cognitive decline involve direct neurotoxicity, vascular damage, inflammation, or other biological pathways.

For now, the findings support a more cautious stance toward alcohol use, particularly among middle-aged and older adults. While moderate drinking has long been portrayed as compatible with healthy aging, the evidence from this study suggests that even small amounts of alcohol might have long-term consequences for brain function.

The study, “Alcohol use and risk of dementia in diverse populations: evidence from cohort, case–control and Mendelian randomisation approaches,” was authored by Anya Topiwala, Daniel F. Levey, Hang Zhou, Joseph D. Deak, Keyrun Adhikari, Klaus P. Ebmeier, Steven Bell, Stephen Burgess, Thomas E. Nichols, Michael Gaziano, Murray Stein, and Joel Gelernter.