On 1 September 2013, Wales began offering a shingles vaccine to older adults, and to ration supply it drew a line through the calendar. If you were born on or after 2 September 1933, you could have it. If you were born a day earlier, you could not, ever.
Nobody designing that policy was thinking about dementia. But the rule accidentally created something medicine almost never gets: two huge groups of people, identical in every way that matters, separated only by whether a vaccine was available to them. Twelve years later that accident has produced the most persuasive evidence anyone has for preventing dementia with something that already exists.
It is worth understanding exactly why this evidence is better than the usual “people who do X get less dementia” story, and exactly where it still falls short.
Why ordinary vaccine studies cannot answer this
Several observational studies had already found that people who get a shingles vaccine are diagnosed with dementia less often. The problem is obvious: people who turn up for vaccines are different. They see doctors more, exercise more, smoke less and are generally the kind of people who get less of everything.
You can adjust for the things you measured. You cannot adjust for the things you did not. As the lead researcher, Pascal Geldsetzer of Stanford, put it to neurologists in 2026 when describing the limits of health records: do we have perfect measures of physical activity in these data records? The answer is never.
That matters because the history of dementia prevention is full of things that looked protective in observational data and then failed when tested properly. One report on this research notes that many drugs that appear to protect against dementia in observational studies fail in clinical trials, as was recently the case with the GLP-1 drug semaglutide.
What the birthday rule made possible
The Welsh rule sidesteps the problem. In the paper’s words, those born before 2 September 1933 were ineligible and remained ineligible for life. Someone born in the last week of August 1933 and someone born in the first week of September 1933 are, on average, the same kind of person. They did not choose which side of the line to be born on.
The cut-off had a dramatic effect on who got vaccinated. Vaccination rates increased sharply from 0.01% among those just too old to qualify to 47.2% among those eligible. Nothing else changed at that date. So any jump in dementia rates at the same point in the calendar can be pinned on the vaccine with far more confidence than usual.
There was a jump. Using the health records of a large cohort of older adults, the researchers found that receiving the vaccine reduced the probability of a new dementia diagnosis over a follow-up period of 7 years by 3.5 percentage points. That corresponds to a relative reduction of about 20%.
Two details strengthen it. This protective effect was stronger among women than men. And the team checked the result in a different population with different data, using death certificates across England and Wales, and found the same signal in deaths where dementia was the primary cause.
The authors describe the result carefully: evidence of a dementia-preventing or dementia-delaying effect that is less vulnerable to confounding and bias than the existing associational evidence.
The catch: that vaccine is no longer used
The Welsh study was of Zostavax, an older live vaccine that is no longer used in the United States, where it has been replaced by Shingrix, a newer vaccine that is much better at preventing shingles.
Whether Shingrix shares the dementia effect cannot be answered with the same elegant design, because no birthday rule governed its rollout. The evidence so far is encouraging but conventional. A 2024 analysis in Nature Medicine of medical records from more than 100,000 patients suggested the newer vaccine was associated with even better protection. In June 2026, a Brown University team reported in Annals of Internal Medicine that older adults who received a shingles vaccine after a stay in a skilled nursing facility had a 24% lower risk of being diagnosed with dementia over a four-year period.
That study used a method designed to imitate a trial, and it is a serious piece of work. It also shows the difficulty: only 8,843 of 509,926 participants received the vaccine. When fewer than 2% of a group get vaccinated, the vaccinated are unlikely to be typical, however careful the statistics.
Why would a shingles vaccine protect the brain?
Nobody knows, and there are two competing ideas.
The first is about the virus. Shingles is caused by the chickenpox virus, which lies dormant in nerve cells for life and can reactivate. Proposed mechanisms include the virus contributing to blood vessel damage, triggering inflammation in the brain and promoting the build-up of the proteins associated with Alzheimer’s disease. Fewer reactivations would mean less of that, and in the Welsh data the vaccine did reduce shingles cases.
The second is about the immune system. Vaccines, particularly live ones, may have broad off-target effects on immunity that have nothing to do with their intended target. If that is the explanation, the benefit might belong to the old live vaccine specifically and not carry over to the new one.
Until that is resolved, nobody can say whether Shingrix does the same job, a better one, or a different one.
What is not known
Whether it holds up in a randomised trial. A natural experiment is the next best thing, not the same thing, and no completed trial exists.
Whether it prevents dementia or delays diagnosis. A seven-year follow-up cannot distinguish a disease that never arrives from one that arrives a few years later.
Whether it applies to younger people. The Welsh participants were around 80 at vaccination. Nothing in the data says what happens if you are vaccinated at 50 and followed for thirty years.
How big the benefit is for an individual. One epidemiologist not involved in the work called the individual-level effect modest, while noting that across millions of eligible adults it could avert a substantial number of cases.
What this means for you
If you are a 35-year-old developer, there is nothing to act on personally. Shingles vaccination is not routinely offered at your age, and no evidence concerns people like you.
If you are 50 or over, or have parents who are, the calculation is simple. Shingrix is recommended for adults from around age 50, given in two doses, because shingles is common, painful and sometimes disabling. That case stands on its own. A possible reduction in dementia risk is a bonus that might be real, stacked on top of a benefit that certainly is.
This is not individual medical advice. Eligibility differs between countries, and decisions about vaccination belong with a clinician.
Common questions
Does the shingles vaccine prevent dementia?
Probably in part, but it is not proven. A natural experiment in Wales found that vaccination reduced new dementia diagnoses over 7 years by 3.5 percentage points, about a 20% relative reduction. That design is far stronger than ordinary observational research but is not a randomised trial.
Does Shingrix lower dementia risk like the old vaccine?
The strongest study was of Zostavax, which is no longer available in the United States. Observational studies of Shingrix point the same way, including a 2026 analysis finding 24% lower risk over four years, but those designs are more open to bias.
Why would a shingles vaccine affect the brain?
There are two theories: that preventing reactivation of the chickenpox virus spares the brain from inflammation and vascular damage, or that vaccination has broader off-target effects on the immune system. Neither has been confirmed.
Should I get the shingles vaccine early to protect my brain?
There is no evidence for that. The Welsh participants were about 80 years old, and the CDC recommends Shingrix from age 50 to prevent shingles. Get it when you are eligible, for the reason it was designed.
The bottom line
A bureaucratic cut-off date handed researchers a near-perfect comparison, and the result is the best evidence so far that anything prevents dementia: roughly a fifth fewer diagnoses over seven years. The caveats are real. It was an older vaccine, in 80 year olds, without a randomised trial, by a mechanism nobody understands. None of them changes the practical advice, which is pleasantly boring: when you or your parents become eligible for the shingles vaccine, take it.
Did this answer your question?
Thanks. That helps us decide what to fix.
Leave a comment