Wasting Your Brain Away in Margaritaville

Written by: JENN DOWD, PHD

Originally posted here on SEP 30, 2025

I grew up in the red wine-loving, “French Paradox” era of alcohol and health messaging, which told us that a little bit of alcohol was actually healthier than completely abstaining.

You know what they say about things that sound too good to be true…

In recent years, the tide has gradually been turning against the idea that moderate alcohol consumption is good for your health. One challenge was methodological. Many alcohol abstainers have quit for health reasons or because of past problematic drinking, making the better health of light to moderate drinkers a bit of a mirage.

A new study throws more cold beer on the notion of any protective effects of alcohol, this time for dementia.

The study

The researchers looked at data from more than 500,000 people in the U.S. Million Veteran Program and the UK Biobank. They analyzed how self-reported alcohol consumption at the baseline interview was associated with dementia diagnoses over several years of follow-up.

In their first analysis, the authors replicated the familiar U-shaped relationship with alcohol and health, with the lowest risk of dementia among those with low or moderate consumption and a higher risk for those who drank none or drank heavily.

But knowing the potential for “sick-quitter” bias, the authors followed up with a Mendelian randomization (MR) analysis meant to better identify true cause and effect. Mendelian randomization, as the name implies, attempts to mimic an experiment based on the “random” draw of genes you get from your parents. Some people get a mix of genes that predict higher alcohol consumption in large genetic studies. In theory, whether you get these particular genes should not be associated with your level of education or other things that might confound the association between alcohol and health. (In reality, kids get both their genes and social environments from their parents, so genes are not as much of a random “lottery” as they may seem).

What they found

Using this genetically informed approach, the protective effect of alcohol went away. Instead, each level of increasing alcohol intake was associated with higher dementia risk.

But there is some good news-the magnitude of the effect was not huge: an increase from five to sixteen drinks per week was associated with ~15% increased risk of dementia. And having double the genetic risk of alcohol use disorder was associated with a 16% relative increase in the risk of dementia. The overall cumulative risk of dementia incidence in these cohorts was around 2% over the follow-up period, so these relative increases are on top of a small baseline rate.

How strong is the evidence?

To me, this study is a small but not groundbreaking addition to the accumulating evidence that even moderate drinking is likely not good for our bodies or brains. The findings from the Mendelian Randomization analysis do suggest that previous observational studies have been biased by sick-quitter bias (ex-drinkers lumped in with lifelong abstainers), socioeconomic confounding (well-off people like their wine) as well as reverse causation (early dementia leading people to reduce drinking).

The fine print

Mendelian randomization analyses rely on some pretty strong assumptions. An important one is that the genes increasing the propensity for alcohol consumption don’t affect dementia through any other pathways. Since we know that the same genes often affect many different outcomes (especially genes associated with behavioral traits like alcohol consumption), I’m not 100% sold that this MR analysis is identifying a pure causal effect of alcohol on dementia.

Even if we do buy the MR analysis, the effect sizes aren’t enormous-far from smoking-level hazards, for example. And at the lowest levels of drinking, the MR estimates become less precise, making the “no safe level” headline a bit tidier than the statistical reality.

All of that said, the study does a commendable job trying to overcome the statistical challenges inherent in studying the health effects of a behavior like alcohol consumption. Any behavior that is strongly patterned by social factors (whether it is alcohol consumption, kombucha drinking, or exercise) is difficult to study and confidently identify cause and effect. And this study deserves a Nobel Prize compared to another recent UK Biobank study using Mendelian Randomization and champagne drinking that I covered here:

But like many things in health, we don’t rely entirely on observational studies to build our knowledge. We also have fundamental neuroscience showing that alcohol is toxic to neurons, which lends credence to the idea that any amount of alcohol is probably not doing your brain any favors. Somehow, I doubt this comes as a surprise to any of my fellow aging Gen-Xers!

Bottom line

The comforting idea that a little tipple is neuroprotective is probably not true. Alcohol likely isn’t helping your brain, and it may be hurting it, even in moderation.

So if you’re drinking, do it safely and because you enjoy it, not based on any belief that it’s good for you.