Association Between Herpes Zoster Vaccination and Dementia Risk: A Systematic Review and Meta-Analysis.

BACKGROUND/

OBJECTIVES

Dementia represents a major global health challenge, and preventive strategies remain limited. Observational studies have suggested a possible association between herpes zoster vaccination (HZV) and dementia risk, although the influence of vaccine type and dose regimen remains unclear.

METHODS

PubMed, Embase, and Web of Science were searched through January 2026. The primary random-effects meta-analysis included one adjusted estimate per independent study or data-source cluster. Quasi-experimental and correlated secondary estimates were reported separately, and alternative-estimate and leave-one-out analyses retained independent statistical units. The protocol was registered in PROSPERO (CRD420261326042).

RESULTS

Fourteen reports representing 13 studies were included. HZV was associated with a lower dementia risk in the primary analysis of five independent data sources (ratio estimate = 0.72; 95% CI: 0.65-0.81; I2 = 97.1%), with similar sensitivity results (0.70-0.76). Two regression-discontinuity studies reported eligibility-associated absolute reductions in dementia diagnosis of 1.3 and 1.8 percentage points. Two direct vaccine comparisons favored the recombinant zoster vaccine (RZV; Shingrix) over the live attenuated vaccine (ZVL; Zostavax) (RMTL ratio = 0.83; 95% CI: 0.80-0.87; RR = 0.82; 95% CI: 0.69-0.98). The only direct dose comparison favored two or more RZV doses over one dose (RR = 0.81; 95% CI: 0.74-0.88). Active-comparator associations were generally attenuated, whereas age- and subtype-specific findings varied. These secondary findings were exploratory.

CONCLUSIONS

HZV was associated with lower dementia risk in observational studies, with directionally consistent quasi-experimental findings. Direct evidence suggested potentially stronger inverse associations for RZV and complete vaccination, but was based on few independent data sources. Substantial heterogeneity and residual confounding preclude causal interpretation, and further prospective or randomized evidence is needed.