Older adults who received the recombinant shingles vaccine Shingrix after a stay in a skilled nursing facility had fewer recorded dementia diagnoses over four years than comparable unvaccinated adults in a large U.S. analysis. The result adds evidence to a growing, but still unsettled, line of research linking shingles vaccination with dementia outcomes.
The study, described by Brown University and published in Annals of Internal Medicine, examined 509,926 Medicare beneficiaries age 66 and older who entered more than 5,500 skilled nursing facilities from 2017 through 2022. It found an adjusted 24% lower risk of a dementia diagnosis among people who received at least one dose of recombinant zoster vaccine, the vaccine sold as Shingrix. Because vaccination was not randomly assigned, however, the findings show an association, not proof that the vaccine prevents dementia.
That caution is especially important because the vaccinated group was small: 8,843 people, or about 1.7% of the full study population. The researchers reported that vaccine recipients tended to be younger and healthier before vaccination than people who did not receive it. Statistical adjustment can account for many documented differences, but it cannot fully correct for characteristics absent from medical records or claims data.
What the study found
All participants had no prior dementia diagnosis and were considered eligible for shingles vaccination. Researchers linked Medicare claims with electronic health records and followed people for four years after the study’s defined starting point. By the end of follow-up, 18.8% of vaccinated participants and 24.6% of unvaccinated participants had a recorded dementia diagnosis.
The difference between those reported proportions is 5.8 percentage points. Put another way, the unadjusted figures show roughly 58 fewer diagnoses per 1,000 people in the vaccinated group over four years. The headline 24% figure is different: it is the study’s adjusted relative-risk estimate, not a 24-percentage-point decline in dementia diagnoses.
Kaley Hayes, a Brown University School of Public Health researcher who led the work, said the observed difference was roughly equivalent to one fewer dementia diagnosis for every 17 people vaccinated. That calculation describes the gap seen between the study groups; it should not be read as an established number of cases the vaccine would prevent in routine care.
The analysis used an approach called target-trial emulation. Rather than simply comparing anyone vaccinated with anyone unvaccinated, researchers specify rules intended to mimic parts of a hypothetical randomized trial: who would be eligible, when follow-up would begin, what counted as vaccination and what outcome would be measured. It can make observational evidence more rigorous than a conventional retrospective comparison, but it does not create the balance achieved when participants are randomly assigned to receive a vaccine or a placebo.
Why diagnosis data leave uncertainty
The outcome was a recorded dementia diagnosis, not a uniform research assessment of cognition performed on every participant. Claims and health-record data are valuable for studying large populations over years, but diagnoses can be affected by when a person is evaluated, which clinicians they see, their access to follow-up care and how thoroughly symptoms are documented.
People who are vaccinated may also differ from those who are not in ways related to health care use, family support, preventive care or frailty. The study adjusted for measured factors, and its setting offers a defined point at which to compare patients entering skilled nursing care. Still, differences that were unmeasured or imperfectly recorded could help explain some or all of the association.
The population also limits how broadly the findings can be applied. Participants were older adults who had recently entered skilled nursing facilities, a medically vulnerable group with health experiences that may not resemble those of all Americans eligible for Shingrix. The available description does not establish whether completing the usual two-dose vaccine series has a different association from receiving one dose, because the exposure was defined as at least one dose.
Nor does the research identify a biological mechanism. Shingles is caused by reactivation of varicella-zoster virus, and scientists have proposed several possible pathways by which infections, inflammation or immune responses could be connected to later neurological disease. Those hypotheses do not establish that vaccination changes the processes that lead to dementia.
How it fits with earlier vaccine research
Several previous studies have reported lower dementia rates after shingles vaccination, but they have not all studied the same vaccine or used the same design. That distinction is relevant because Shingrix is a recombinant vaccine and is now the only shingles vaccine marketed in the United States; much earlier research involved the older live zoster vaccine, which is no longer used in the country.
A separate U.S. observational analysis published in 2024 and available through PubMed Central compared recipients of recombinant and live shingles vaccines. It reported an association between recombinant vaccination and lower dementia risk, offering context for the new nursing-facility analysis but not independent confirmation of its precise 24% estimate.
Research in Wales took a different approach. A 2025 report from Stanford Medicine described a study that used an age-based eligibility cutoff for the older live vaccine, effectively comparing people born just before and after a date that determined eligibility. That quasi-experimental design was intended to reduce some of the selection differences that complicate ordinary vaccine comparisons, and it too found fewer dementia diagnoses among people eligible for vaccination.
Together, the studies make the question more compelling, but they do not settle it. Different vaccines, patient populations, health systems and methods can produce results that are not directly interchangeable. A randomized clinical trial designed to measure dementia outcomes would be needed to determine whether Shingrix itself reduces dementia incidence.
No change to the vaccine’s purpose
Shingrix is recommended to prevent shingles and related complications, not as a dementia-prevention treatment. The new findings do not change that indication or provide a reason for patients to use vaccination as a substitute for evaluation of memory symptoms or for established preventive care.
Brown University said the study authors disclosed funding from GlaxoSmithKline, the manufacturer of Shingrix. The authors said the company had no role in the study design, analysis or decision to publish. The university’s original announcement was published June 15, 2026; a later version carried by ScienceDaily does not alter the underlying study description.
