Herpes zoster vaccination and incident dementia in Canada: an analysis of natural experiments.
Level 2 - randomized trial
Natural experiment using regression discontinuity and synthetic control designs
PubMed 41579903 · doi:10.1016/S1474-4422(25)00455-7
What was done
Analyzed electronic health record data from the Canadian Primary Care Sentinel Surveillance Network (1,434 primary care providers) for individuals born between 1930 and 1960. The study used a regression discontinuity design around Ontario birth-date eligibility cutoffs (Jan 1, 1946 primary; Jan 1, 1945 secondary) for a publicly funded live attenuated herpes zoster vaccination programme. The authors also applied synthetic difference-in-differences and synthetic control methods comparing eligible Ontario cohorts with ineligible cohorts from other Canadian provinces lacking vaccination programmes. The primary outcome was new dementia diagnosis over 5.5 years of follow-up (records from Jan 1, 1990, to June 30, 2022).
What was found
Among 232,124 Ontario patients analyzed (125,719 [54.2%] female, 106,354 [45.8%] male), baseline health characteristics were similar across cutoffs except for vaccination rates. Vaccine eligibility at the Jan 1, 1946 cutoff decreased the probability of incident dementia diagnosis by an absolute difference of 2.0 percentage points (95% CI 0.4–3.5, p=0.012) over 5.5 years. The Jan 1, 1945 cutoff showed an identical reduction of 2.0 percentage points (0.2–3.8, p=0.025). New dementia diagnoses were also significantly less common among eligible Ontario cohorts compared to corresponding cohorts in other Canadian provinces without vaccination programmes.
Why it matters
Using policy cutoffs that minimize confounding common in observational studies, this paper strengthens evidence that live attenuated herpes zoster vaccination delays or prevents dementia onset in older adults.
Limits
Individual-level vaccination status was not directly linked in the primary EHR dataset; uptake was estimated from population surveys. Dementia ascertainment relied on primary care electronic records without standardized clinical re-evaluations. Follow-up was restricted to 5.5 years, evaluated only the live attenuated formulation rather than recombinant vaccines, and was conducted within the Canadian primary care context.
Cited by
- supports Natural experiments in Wales, Australia, and Canada found that eligibility for or receipt of the live attenuated shingles vaccine was associated with a decreased risk of dementia.