A new meta-analysis returned an overall relative risk of 1.03 for alcohol and all-cause dementia. Essentially noise. I find that aggregate number almost useless on its own - the interesting part comes when you cut by how much people actually drink.

"Alcohol and brain health" gets argued from both sides with equal conviction. One study says neuroprotective. The next says neurotoxic. A meta-analysis doesn't settle that - it just forces both signals onto a single number. And that number depends entirely on dose.

What they did

Researchers searched PubMed, Embase, Cochrane Library, and Web of Science for studies published through July 2024. Four categories of dementia: all-cause dementia (ACD), Alzheimer's disease (AD), vascular dementia (VD), and other. Study quality was scored using the Newcastle-Ottawa Scale (NOS), a standard checklist for cohort and case-control work. Subgroup analyses cut the data by drinking level, geographic region, and age group.

What they found

Overall - all drinkers against non-drinkers - no significant association with any dementia type:

  • ACD: RR (relative risk) = 1.03, 95% CI (confidence interval): 0.84-1.27
  • AD: RR = 0.97, 95% CI: 0.86-1.08
  • VD: RR = 1.09, 95% CI: 0.95-1.26

The "other dementia" category returned an RR of 0.62, but the confidence interval was wide enough to include 1 (0.33-1.15), so that number doesn't tell us much.

Break it down by drinking level, and the picture changes sharply:

  • Light/moderate. 12% lower risk of ACD and AD (RR = 0.88 for both). Stronger in European populations and in the 60-69 age group.
  • Heavy. Risk up across every category: ACD +18% (RR = 1.18), AD +29% (RR = 1.29), VD +25% (RR = 1.25).

What it means

The J-shaped curve shows up in alcohol research often enough that I'm not surprised to see it here. A little, and the association tilts protective. A lot, and it tips the other way.

The overall null finding is worth pausing on. When you lump all drinkers together, the heavy drinkers raising risk and the moderate drinkers potentially lowering it average out to near-zero. This is why a single "alcohol and dementia" number is nearly meaningless.

I'd hold the protective signal loosely. One reason: reverse causality. People in early cognitive decline often stop drinking before they're diagnosed, which makes abstainers look like a higher-risk group by default. Another reason: definitional drift. "Light," "moderate," and "heavy" aren't standardized across studies - the definitions shift from paper to paper, so the threshold where protective flips to harmful isn't fixed. Self-reported consumption also tends to run low.

The European and age-specific patterns don't have a clean explanation. The protective signal was stronger in European populations and clearest in the 60-69 cohort. The paper doesn't offer a mechanism, and I won't speculate. But a real causal effect would show up more uniformly across regions and ages. That it doesn't makes me more skeptical of the protective reading, not less.

The heavy-drinking end of the curve is harder to dismiss. Three different dementia outcomes, all pointing the same direction. The Alzheimer's number is especially tight: RR = 1.29, 95% CI: 1.21-1.36. That's not noise.

The practical boundary is the line between "moderate" and "heavy" - where the curve flips. AlcoBalance gives you pace control in real time: you see your current rate and your peak, and you can ease off the moment you notice yourself speeding up, while you're still steering the evening.

Source: Internal Medicine Journal, DOI