Shingles vaccine tied to 24% lower dementia risk in large study of older adults

Among more than 500,000 U.S. adults 66 and older who entered a nursing or rehabilitation facility, those who received the recombinant shingles vaccine were 24% less likely to be diagnosed with dementia over the next four years, according to a target-trial-emulation study published this week in Annals of Internal Medicine. The researchers' own estimate: roughly one dementia case prevented for every 17 people vaccinated.

That is the headline. The honest question is what it covers, and what it does not.

The study is the largest single look yet at the recombinant zoster vaccine (sold as Shingrix, the only shingles vaccine on the U.S. market since 2017) and later dementia. Researchers, led by Kaley Hayes, a pharmacoepidemiologist at Brown University, used Medicare claims and health records to mimic a randomized trial in more than 502,000 adults age 66 and older who were admitted to a skilled nursing or rehabilitation facility and had not previously been up to date on shingles vaccination. Recipients had a relative risk of dementia diagnosis of 0.76 (95% confidence interval 0.69 to 0.84) compared with unvaccinated peers, the team reports, and the effect held across subgroups, including age, sex, and several comorbidity strata.

The result lines up with a longer trail of smaller signals. Earlier work, mostly on the now-discontinued live-attenuated shingles vaccine (Zostavax, withdrawn from the U.S. market in 2020), had already pointed in the same direction. The new analysis, summarized in lay terms by ScienceDaily and covered in MedPage Today, is the first to test the recombinant version at this scale in this specific population.

The critical read is that the design is still observational, even when dressed as a target trial emulation. Hayes and co-authors flag the same confounder that haunts every vaccine-and-chronic-disease study: healthier people are more likely to get vaccinated in the first place. A 24% drop in dementia diagnosis could be partly, mostly, or entirely the result of that selection, not a vaccine effect. As the science writer Skeptical Raptor summarized it, the signal is real and the headline number is right, but the causal story is not closed.

Two boundaries are worth pinning down before the 24% travels further. First, the cohort was older adults newly admitted to skilled nursing care, not the average 66-year-old living independently at home. Generalizing to a healthier, younger population is a leap the paper does not make. Second, the absolute-risk framing matters: a 24% relative reduction in a population whose baseline dementia risk is already elevated still leaves most vaccinated individuals with no change in outcome, and the "one case prevented per 17 vaccinated" figure is a model estimate from the study authors, not a clinical trial endpoint.

What would move this from association toward cause? A large, pragmatic randomized trial in older adults, with a pre-specified dementia endpoint, and follow-up long enough to capture late-onset cases. Until then, the practical read is narrower than the headline suggests: if an older adult is already eligible and considering Shingrix against shingles, a possible brain-health dividend is one more thing to ask a clinician about, not a reason to lead with the vaccine.