The shingles shot you may already be due for keeps showing up, study after study, next to a lower rate of dementia. That is worth your attention. It is not yet worth your certainty.
The newest number arrived in June 2026, in Annals of Internal Medicine. Brown University researchers followed 509,926 older adults who had passed through a nursing home. Over four years, about 246 of every 1,000 unvaccinated people were diagnosed with dementia, against about 188 of every 1,000 who had received the recombinant shingles vaccine, sold as Shingrix. That is a 24 percent lower relative risk. The authors estimate it could mean one dementia case avoided for every 17 people vaccinated.
The method is better than most. They used target trial emulation, which arranges the records to imitate the randomized trial nobody ran. But look at who got the shot: only 8,843 of more than half a million. People who seek out a vaccine are, on average, healthier and more tethered to the health system than people who do not, and no statistical adjustment fully erases that. This is the healthy-vaccinee problem, and it is the shadow over every study here. The Brown work was also funded by GSK, which makes Shingrix, though the disclosure says the company had no hand in the design or the results.
So one study, taken alone, would not move me. What is striking is that it does not stand alone.
An Oxford group reported in 2025, across 436,788 people, that the shingles vaccine and the RSV vaccine, which both carry the AS01 adjuvant, the immune-boosting compound added to a vaccine, were associated with a lower 18-month dementia risk, and by roughly the same margin as each other. The similarity implicates the adjuvant itself rather than the shingles part. Different vaccines, different countries, different methods, and the arrow keeps pointing the same way.
Consistency is not the same as proof. A shared bias produces consistently wrong answers just as neatly as a real effect produces consistently right ones.
The closest thing we have to an experiment is a 2025 Nature study out of Wales. Eligibility for the shingles vaccine there was set by an arbitrary line: born before September 2, 1933, and you were shut out for life; born just after, and you were let in. People a week apart in age are otherwise alike, so the cutoff sorted them almost as a coin toss would. On the eligible side, dementia diagnoses over seven years fell by 3.5 percentage points, which is about a fifth of the cases in that group, with the effect larger in women. One detail complicates the tidy adjuvant story from the paragraph above: Wales was handing out the older live vaccine, Zostavax, which does not contain AS01 at all. So the signal turns up with the adjuvant and without it, which deepens the puzzle rather than resolving it. That arbitrary line is still what makes the finding hard to wave away.
Hard to wave away is not the same as settled. The trial that could settle it is running now. DAN-ZOSTER is randomizing roughly 162,000 Danish adults over 65 to Shingrix or nothing and tracking dementia and heart events through the national registries. That is the test.
Until it reports, we have a set of tea leaves that happen to all point in one direction. The vaccine already earns its place by sparing you shingles, which is reason enough to consider it. Whether it also guards the brain is a question we do not get to answer by wanting to.
