Coffee vs. Dementia: Harvard Tracked 132,000 People for 40 Years
Authors: Yu Zhang, Daniel Wang
Why This Matters
Every three seconds, someone in the world is diagnosed with dementia. Alzheimer’s disease and related conditions already affect more than 55 million people — and the numbers keep climbing. There is still no treatment that stops or reverses the process.
That’s why prevention research matters so much. And that’s why a new Harvard study published in JAMA on February 9, 2026, has attracted so much attention: for the first time, researchers followed more than 130,000 people for up to 43 years — and found a consistent link between daily coffee consumption and reduced dementia risk.
Not in a lab mouse. In real people, living real lives, drinking real coffee.
The Core Idea
The study found that people who drank 2–3 cups of coffee or 1–2 cups of tea per day had a dementia risk roughly 18% lower than those who consumed little or none.
Dementia — a group of conditions (including Alzheimer’s, vascular dementia, and others) characterized by progressive decline in memory, thinking, and the ability to live independently.
The detail that makes this finding convincing: decaffeinated coffee showed no protective effect. This points directly to caffeine — or compounds closely tied to it — as the active ingredient, rather than some general «coffee drinker lifestyle.»
How the Study Was Done
This is not a randomized trial — no one was assigned to drink coffee against their will. It’s a large-scale cohort study: researchers tracked what participants ate and drank over many years and recorded who developed dementia and when.
Cohort study — a method in which a group of people is observed over time, with their habits and health outcomes recorded. It can identify statistical associations but cannot prove cause and effect.
Data came from two landmark cohorts:
- Nurses’ Health Study — followed since 1976
- Health Professionals Follow-Up Study — men in health professions, followed since 1986
Together they provided 131,821 participants and 11,033 documented dementia cases over the observation period. Participants regularly completed dietary questionnaires; cognitive status was assessed through standardized tests and medical records.
The paper underwent full peer review at JAMA — one of the most authoritative medical journals in the world.
Results
| Beverage | Optimal dose | Dementia risk reduction |
|---|---|---|
| Caffeinated coffee | 2–3 cups/day | ~18% |
| Tea | 1–2 cups/day | ~18% |
| Decaffeinated coffee | any | no effect |
Beyond clinical dementia, the study also looked at earlier signs of cognitive change:
- Subjective cognitive decline (when people notice their own memory slipping): 7,8% in caffeine consumers vs. 9,5% in others
- The effect held regardless of genetic predisposition to dementia — suggesting caffeine may be beneficial even for those at higher genetic risk
Why Decaf Doesn’t Work — And What That Means
This is arguably the most interesting result. If the effect were simply about «coffee drinker lifestyle» — active, social people with good metabolic health — decaf would show a similar benefit. It doesn’t.
Researchers propose two main mechanisms:
1. Caffeine as a neuroprotectant. Caffeine blocks adenosine receptors in the brain — the same receptors responsible for building «sleep pressure.» One hypothesis: chronic blockade of these receptors reduces the accumulation of beta-amyloid, the toxic protein considered a key driver of Alzheimer’s.
Adenosine receptors — proteins on the surface of neurons that, when activated, slow brain activity (causing fatigue and sleepiness). Caffeine competes with adenosine for these receptors — hence the alerting effect.
2. Polyphenols as anti-inflammatory agents. Coffee and tea are rich in polyphenols — plant compounds with antioxidant properties. These may reduce chronic neuroinflammation, which is associated with neurodegenerative disease.
Tea appears effective at lower doses than coffee, possibly because of additional polyphenols (catechins, theanine) that act through complementary pathways.
Critical Analysis
Disclaimer: This is an analysis based on publicly available data and the official publication. The paper was peer-reviewed in JAMA, but interpretation of results requires careful consideration of methodological limitations.
Strengths:
- Largest sample in the history of coffee-dementia research — 131,821 participants
- 40+ years of follow-up provides statistical power impossible to achieve in short studies
- Publication in JAMA — the highest tier of peer review
- Decaf as a built-in «control condition» strengthens plausibility
Limitations:
- Observational design — causality cannot be proven: healthy people may simply drink more coffee
- Reverse causation: people with early cognitive changes may avoid caffeine due to anxiety or sleep disruption
- Self-reported dietary data — people recall their coffee intake imprecisely; systematic error is unavoidable
- Participants are healthcare workers: educated, with stable incomes and medical access — how well do results generalize to broader populations?
Open questions:
- What is the exact molecular mechanism by which caffeine acts on the brain?
- Do results replicate in populations with different dietary patterns — for example, in Asia, where tea dominates and coffee intake differs?
Takeaway
Dr. Daniel Wang, the study’s senior author, is deliberately measured: «The effect size is small and coffee is just one piece of the puzzle alongside other prevention strategies.» That’s honest.
But an 18% reduction in risk across 131,821 real people followed for 40+ years is not a fluke. It’s a signal worth taking seriously.
Drinking coffee specifically to «prevent Alzheimer’s» would be an oversimplification. But if you already drink 2–3 cups a day — this study gives you one of the best scientific arguments yet to keep doing it.
Just, please: not decaf.
Original publication: Zhang Y. et al. «Coffee and Tea Intake, Dementia Risk, and Cognitive Function.» JAMA, February 9, 2026. Harvard Gazette