Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Apr 23.
Published in final edited form as: Lancet Neurol. 2026 Feb;25(2):170–180. doi: 10.1016/S1474-4422(25)00455-7

Herpes zoster vaccination and incident dementia in Canada: an analysis of natural experiments

Michael Pomirchy 1, Seunghun Chung 1, Christian Bommer 1, Stephenson Strobel 2, Pascal Geldsetzer 1,3,4,5,*
PMCID: PMC13101831  NIHMSID: NIHMS2152981  PMID: 41579903

Summary

Background:

Two natural experiment studies have found evidence that live attenuated herpes zoster vaccination prevents or delays dementia onset. We aimed to determine the effect of live attenuated herpes zoster vaccination on incident dementia diagnoses among people aged 70 years and older using a natural experiment in Ontario, Canada, and to triangulate these findings, using a second natural experiment in Ontario and a quasi-experimental approach that uses data from multiple Canadian provinces.

Methods:

Our analysis of natural experiments included people born in Canada between Jan 1, 1930, and Dec 31, 1960, who were registered with one of 1434 primary care providers in the Canadian Primary Care Sentinel Surveillance Network (CPCSSN) on Sept 15, 2016. We compared patients born immediately before versus immediately after Jan 1, 1946, in our primary analysis, and immediately before versus immediately after Jan 1, 1945, in our secondary analysis, as these thresholds determined eligibility for herpes zoster vaccination in Ontario. The key strength of this natural experiment is that these comparison groups are not expected to differ in their health characteristics and behaviours given that all that divides them is a small discrepancy in age. Dementia diagnosis was established using electronic health records data from Jan 1, 1990, to June 30, 2022, from the primary care practices. We used a population-representative survey of people aged 65 years or older in Ontario to measure herpes zoster vaccination uptake. Using regression discontinuity analysis, we estimated the difference in vaccination uptake and dementia diagnoses between individuals born immediately on either side of the eligibility thresholds for herpes zoster vaccination. Additionally, we used synthetic difference-in-differences and a synthetic control method to compare trends in dementia incidence (before versus after the start date of the herpes zoster vaccination programme) among birth cohorts in Ontario who were eligible for vaccination with the same birth cohorts (all of whom were ineligible for vaccination) in other provinces of Canada.

Findings:

We extracted data on 464,637 patients who were registered with a primary care provider in the CPCSSN as of Sept 15, 2016. Of 232,124 patients born in Ontario included in the analysis, 125,719 (54.2%) were female, 106,354 (45.8%) were male, and 51 (<0.5%) had missing information on sex. Patients born immediately before versus immediately after the two eligibility thresholds for herpes zoster vaccination did not differ in their health characteristics at the time of the start date of the vaccination programme, except for a large difference in their probability of receiving herpes zoster vaccination. Being born immediately before versus immediately after Jan 1, 1946, decreased the probability of receiving a new dementia diagnosis by an absolute difference of 2.0 percentage points (95% CI 0.4–3.5, p=0.012) over a 5.5-year follow-up. Using the Jan 1, 1945, threshold, dementia diagnoses were also reduced by 2.0 percentage points (0.2–3.8, p=0.025) over 5.5 years. After the start of the programme, new dementia diagnoses among the birth cohorts eligible for herpes zoster vaccination in Ontario were significantly less common than in the same birth cohorts in other Canadian provinces that did not have a herpes zoster vaccination programme.

Interpretation:

This analysis of natural experiments provides evidence, which is more likely to reflect a causal relationship than previous evidence from more standard observational data analyses, that herpes zoster vaccination prevents or delays incident dementia. Mechanistic research into this effect could provide insights into the pathophysiology of dementia and maintenance of neuroimmune health in older age.

Introduction

Microbes have long been suggested to have a role in the pathogenesis of dementia.1–3 Neurotropic herpesviruses have received particular attention in this regard because they preferentially target the nervous system,4 their reactivations become more common with age and can cause encephalitis,4,5 and they are highly prevalent.4 Herpesviruses can promote the accumulation of β-amyloid in mice and increase tau phosphorylation in human brain organoids.1,6,7 Targeting herpesviruses through vaccination might be a particularly effective approach because the immune system appears crucial in the development of dementia.8 Indeed, vaccines, especially those that are live-attenuated, appear to often have broader health benefits beyond their intended target.9–12 The only vaccination against a neurotropic herpesvirus that is in clinical use is herpes zoster vaccination.

Except for our natural experiments in Wales and Australia,13,14 and a study that compared the recombinant to the live-attenuated herpes zoster vaccine,15 studies on the relationship between herpes zoster vaccination and dementia have compared vaccine recipients with non-recipients.16–23 These studies are inherently limited in drawing causal conclusions because people who decide to be vaccinated differ from those who do not in a myriad of characteristics, many of which (e.g., dietary or physical activity behavior) are not captured in medical claims or electronic health record data.24 Nonetheless, these studies have consistently found a protective association between herpes zoster vaccination and dementia. Such protective associations, including dose-response associations,16,25,26 have also been reported for other vaccines (e.g., influenza),27–29 potentially suggesting a pathogen-independent immunomodulatory mechanism.30

In this study, we take a different approach by taking advantage of the quasi-randomised rollout of herpes zoster vaccination in Canada. Specifically, in Ontario (Canada’s most populous province31), individuals who had their 71st birthday after Jan 1, 2017, were eligible for publicly funded herpes zoster vaccination, whereas those who had their 71st birthday before this date were ineligible.32,33 This eligibility rule created two comparison groups born immediately on either side of the cutoff date who are not expected to differ in any other characteristics. We aimed to compare new dementia diagnoses between these two groups in Ontario and to compare dementia incidence between vaccine-eligible birth cohorts in Ontario with incidence in the same birth cohorts in provinces that did not have a herpes zoster vaccination programme.

Methods

Study design and participants

The live-attenuated single-dose herpes zoster vaccine (Zostavax [Merck]) was included on Ontario’s immunisation schedule on Sept 15, 2016, for adults between the ages of 65 and 70 years.32 In addition, from Sept 15, 2016, to Dec 31, 2016, all people born in 1945 were also given the opportunity to receive publicly funded herpes zoster vaccination.33 The date-of-birth eligibility thresholds of Ontario’s herpes zoster vaccination programme created three comparison groups of people: never eligible because they were born before Jan 1, 1945; eligible for only 3.5 months because they were born in 1945; and eligible for at least 1 year and 3.5 months because they were born between Jan 1, 1946, and Sept 15, 1951 (i.e., aged 65 to 70 years on Sept 15, 2016).33,34 More detail on Ontario’s herpes zoster vaccination program is provided in Text S1.

We used data from the Canadian Primary Care Sentinel Surveillance Network (CPCSSN), which is the only pan-Canadian electronic health record database for primary care.35 The CPCSSN consists of 14 primary care research networks, which recruit primary care practices and form a learning health-care system designed to generate new knowledge and improve the delivery and quality of care. The data from all patients in these primary care practices are available in the CPCSSN data except for those who opt out of contributing their data (79 patients at the time of our data extraction). The database includes the birth month and year for each patient and the entire patient record other than free-text notes and pdf documents, and includes information from secondary and tertiary care as entered by the primary care provider. We received CPCSSN data for Jan 1, 1990, to June 30, 2022. We obtained ethics approval for this research from the Stanford University Institutional Review Board, which approved the study on 9 June 2023 and considered it minimal risk (protocol number, 70277). This approval included a consent waiver because consent was not feasible to obtain from each patient, patients in the CPCSSN are informed that their deidentified data will be made available for research, and the research only used data without any direct identifiers.

We also used data from January 2 2019 to December 12 2020 from the Canadian Health Survey on Seniors (CHSS) to verify that there was an abrupt increase in the probability of receiving herpes zoster vaccination at the date-of-birth eligibility thresholds in Ontario. The CHSS is a cross-sectional population-representative survey for adults aged 65 years and older.36 We had access to month and year of birth and responses to the question: “As an adult, have you ever had the shingles vaccine, also known as herpes zoster vaccine?”. Additional detail on the CPCSSN and CHSS data are provided in Text S1.

In the CPCSSN data, we excluded patients who died before Sept 15, 2016 (n = 7,697) and those who had a missing date for their dementia diagnosis (n = 129). For our primary analyses, we additionally excluded individuals who received a diagnosis of dementia before Sept 15, 2016 (n = 10,570) and those who were coded as “inactive” before Sept 15, 2016 (n = 15,311) according to the primary care provider’s own criteria. In our primary analysis of comparing individuals across the January 1, 1946, date-of-birth eligibility threshold, we started our follow-up period on Jan 1, 2017, because this was the date on which the 3.5-month eligibility period for those born in 1945 ended. For our secondary analysis comparing individuals across the Jan 1, 1945, threshold, we started the follow-up period on the start date of the vaccination programme (Sept 15, 2016). In robustness checks, we started the follow-up period 3, 6, 9, and 12 months after Jan 1, 2017 (when analyzing the Jan 1, 1946, threshold) and Sept 15, 2016 (when analyzing the Jan 1, 1945, threshold). Follow-up for all analyses ended on June 30, 2022.

Exposure and outcomes

Our exposure was eligibility for publicly funded herpes zoster vaccination as determined by date of birth. The Jan 1, 1945, threshold divided those who were ineligible for life from those who were eligible for 3.5 months (from Sept 15, 2016, to Dec 31, 2016), and the Jan 1 1946 threshold divided those who were eligible for 3.5 months from those who were eligible for at least one year and 3.5 months. Our primary analyses focused on the comparison of individuals born immediately around the Jan 1, 1946, threshold because this threshold resulted in both a larger abrupt increase in vaccination uptake (in our analyses on vaccination uptake in the CHSS data) and a longer difference in the period of eligibility than the Jan 1, 1945, threshold. To triangulate our findings using a different date-of-birth eligibility threshold, we also analyzed the January 1, 1945, threshold in secondary analyses.

Our outcome of interest was new dementia diagnoses made during our 5.5-year follow-up period. Given the neuropathological overlap between dementia types and the difficulty in distinguishing dementia types clinically,37–39 and our reduced statistical power when studying less common outcomes, we defined dementia as dementia of any type or cause. We used the date of the first diagnosis of dementia in our data as the date on which the outcome occurred. We also used the first recorded date for all other disease diagnoses. The codes used to define dementia and all other diagnoses in the CPCSSN data are shown in Table S1.

Statistical analysis

Our analysis is based on the intuition that individuals born immediately before versus immediately after the date-of-birth eligibility threshold for vaccination are expected to be similar except for their eligibility for vaccination. Unlike when comparing individuals who opt to obtain vaccination with those who do not, we would expect our two comparison groups (individuals born immediately before versus immediately after the date-of-birth eligibility threshold) to be balanced in their observed and unobserved (e.g., dietary or physical activity) characteristics. Our analyses explicitly control for age, and we are holding patients’ date of birth constant when comparing eligible and ineligible patients. While we include some patients in our estimation who are years apart in age (although with a far lower weighting than those born closer to the threshold), as long as any potential confounding variables do not abruptly change at precisely the date-of-birth eligibility threshold for herpes zoster vaccination, our regression discontinuity approach yields unbiased causal effects.40,41

Regression discontinuity is a causal inference approach widely used in the social sciences in the setting of a threshold-based treatment assignment.42 Figure S1 describes the main analysis featuring a regression discontinuity design at two date-of-birth eligibility thresholds. More details on our regression discontinuity analyses, including robustness checks, are provided in Text S2 (Appendix pages 6–12). This also includes an extension of the regression discontinuity design (called comparative regression discontinuity [CRD]) and two types of analyses to ensure that our findings are robust to abrupt changes in follow-up times or competing risks at the date-of-birth eligibility threshold for herpes zoster vaccination.

As stated earlier, confounding variables can only introduce bias into our analysis if they change abruptly at the date-of-birth eligibility thresholds for herpes zoster vaccination.40,41 Such an abrupt change could occur if another intervention used the same date-of-birth threshold as an eligibility criterion as the herpes zoster vaccination program. We first investigated this possibility by carrying out baseline balance checks. Specifically, we applied the same regression discontinuity specification that we used in our primary analyses for new dementia diagnoses during the follow-up period to diagnoses made at any time before, instead of after, the starting date (Sept 15, 2016) of Ontario’s herpes zoster vaccination program. We also used the 15 most common diagnoses in the CPCSSN data (for our entire patient sample) for these analyses. Second, if another intervention used the same date-of-birth threshold as the herpes zoster vaccination programme and was not designed to primarily affect dementia risk, being born just after versus just before this threshold would likely have an effect on common disease diagnoses other than dementia. We, therefore, implemented the same regression discontinuity analysis as for our primary analysis for the occurrence of new diagnoses during our follow-up period of each of the 15 most common diagnoses in the CPCSSN data (Text S3). Third, if an annual intervention that influences dementia risk used Jan 1 as its date-of-birth eligibility criterion, then we would expect to observe effects of the Jan 1 threshold on new dementia diagnoses in birth years other than 1945 and 1946 (i.e., in birth years other than only the ones used by the herpes zoster vaccination programme as eligibility criteria). We, therefore, implemented the same regression discontinuity analysis as for the Jan 1 1946 threshold, but used Jan 1 of each of the five years prior to 1945 and two years subsequent to 1946. Fourth, if there was a nation-wide intervention that used the same date-of-birth cutoff as the herpes zoster vaccination programme in Ontario, then we would expect to observe effects on incident dementia at this cutoff in provinces other than Ontario. We, therefore, conducted the same regression discontinuity analyses as in Ontario amongst all provinces other than Ontario. More detail on these analyses is provided in Text S2 (Appendix pages 6–12).

To triangulate our findings, we compared, before versus after the start date of the herpes zoster vaccination programme in Ontario, dementia incidence among birth cohorts in Ontario who were eligible for vaccination with the same birth cohorts in other provinces of Canada where they were ineligible. More detail on these analyses is provided in Text S2 (Appendix pages 6–12). The regression equations for our analyses are provided in Text S3 (Appendix pages 12–13).

Results

Our dataset contained 464,637 patients born between Jan 1, 1930, and Dec 31, 1960, who were registered with one of the 1,434 primary care providers in the CPCSSN on Sept 15, 2016. 251,779 of these patients were registered with one of the 634 CPCSSN primary care providers in Ontario. Figure S2 presents a flowchart detailing the sampling and inclusion criteria for our primary analysis. Table 1 shows the sample characteristics for patients at primary care providers in Ontario (separately for all patients born between Jan 1, 1930 and Dec 31, 1960, as well as only those born within the 12 months on either side of the Jan 1, 1946 date-of-birth eligibility threshold used in our primary analysis) and patients at primary care providers in Canadian provinces other than Ontario. The sample characteristics of participants in the CHSS (the dataset we used to estimate the effect of herpes zoster vaccination eligibility on herpes zoster vaccination receipt) are shown in Table S2.

Table 1:

Sample characteristics at the time of the start date of the herpes zoster vaccination program.1,2,3,4,5,6

Canada except Ontario Ontario Ontario - born between January 1 1945 and December 31 1946
Total, n (%) Women, n (%) Men, n (%) Total, n (%) Women, n (%) Men, n (%) Total, n (%) Women, n (%) Men, n (%)
Women 86,692 (53.1) 125,719 (54.2) 8,488 (53.7)
Men 76,432 (46.8) 106,354 (45.8) 7,300 (46.2)
Rural 28,082 (17.2) 14,393 (16.6) 13,656 (17.9) 40,795 (17.6) 20,560 (16.4) 20,225 (19) 2,806 (17.8) 1,397 (16.5) 1,407 (19.3)
Clinical Diagnoses
Dementia 3,443 (2.1) 1,826 (2.1) 1,603 (2.1) 5,506 (2.3) 3,091 (2.4) 2,415 (2.2) 263 (1.6) 150 (1.7) 113 (1.5)
Dyslipidemia 67,184 (41.1) 33,047 (38.1) 34,124 (44.6) 119,169 (51.3) 60,680 (48.3) 58,479 (55.0) 8,928 (56.5) 4,619 (54.4) 4,308 (59.0)
Hypertension 55,377 (33.9) 28,568 (33.0) 26,794 (35.1) 86,416 (37.2) 45,726 (36.4) 40,684 (38.3) 6,850 (43.7) 3,576 (42.1) 3,273 (44.8)
Diabetes 23,703 (14.5) 10,963 (12.6) 12,732 (16.7) 39,883 (17.2) 18,888 (15.0) 20,991 (19.7) 3,218 (20.4) 1,503 (17.7) 1,714 (23.5)
Osteoarthritis 25,797 (15.8) 15,453 (17.8) 10,339 (13.5) 38,163 (16.4) 23,614 (18.8) 14,547 (13.7) 3,018 (19.1) 1,897 (22.3) 1,121 (15.4)
Chronic kidney disease 16,458 (10.1) 9,295 (10.7) 7,157 (9.4) 31,453 (13.6) 18,996 (15.1) 12,453 (11.7) 2,518 (15.9) 1,508 (17.8) 1,010 (13.8)
Depression 23,501 (14.4) 15,033 (17.3) 8,461 (11.1) 32,960 (14.2) 21,664 (17.2) 11,293 (10.6) 2,128 (13.5) 1,436 (16.9) 691 (9.5)
COPD 9,117 (5.6) 4,267 (4.9) 4,846 (6.3) 12,572 (5.4) 6,509 (5.2) 6,062 (5.7) 1,000 (6.3) 524 (6.2) 476 (6.5)
Chronic heart failure 8,146 (5.0) 3,772 (4.4) 4,372 (5.7) 12,854 (5.5) 6,093 (4.8) 6,759 (6.4) 983 (6.2) 458 (5.4) 525 (7.2)
Ischemic heart disease 5,657 (3.5) 1,864 (2.2) 3,790 (5) 7,136 (3.1) 2,124 (1.7) 5,012 (4.7) 588 (3.7) 169 (2) 419 (5.7)
Asthma 5,540 (3.4) 3,404 (3.9) 2,134 (2.8) 5,638 (2.4) 3,645 (2.9) 1,993 (1.9) 383 (2.4) 249 (2.9) 134 (1.8)
Hypothyroidism 3,047 (1.9) 2,353 (2.7) 694 (0.9) 2,789 (1.2) 2,234 (1.8) 555 (0.5) 200 (1.3) 160 (1.9) 40 (0.5)
Atrial fibrillation 4,431 (2.7) 1,799 (2.1) 2,632 (3.4) 2,268 (1.0) 884 (0.7) 1,384 (1.3) 176 (1.1) 55 (0.6) 121 (1.7)
Hyperplasia of prostate 3,652 (2.2) -- 3,646 (4.8) 1,052 (0.5) -- 1,050 (1) 100 (0.6) -- 100 (1.4)
Parkinson’s disease 558 (0.3) 223 (0.3) 335 (0.4) 868 (0.4) 372 (0.3) 496 (0.5) 80 (0.5) 34 (0.4) 46 (0.6)
Multiple sclerosis 659 (0.4) 487 (0.6) 172 (0.2) 568 (0.2) 424 (0.3) 144 (0.1) 44 (0.3) 27 (0.3) 17 (0.2)
Preventive Health Services
Antihypertensive use 51,697 (31.6) 26,107 (30.1) 25,579 (33.5) 82,408 (35.5) 42,370 (33.7) 40,032 (37.6) 6,569 (41.6) 3,320 (39.1) 3,247 (44.5)
Statin use 39,010 (23.9) 17,049 (19.7) 21,956 (28.7) 67,539 (29.1) 30,238 (24.1) 37,295 (35.1) 5,714 (36.2) 2,623 (30.9) 3,090 (42.3)
Open in a new tab

Abbreviations: COPD = Chronic Obstructive Pulmonary Disorder

1

We show sample characteristics for patients in Ontario born between January 1 1945 and December 31 1946 because this is the sample within the 12-month bandwidth used in our primary outcome analysis.

2

Information on patient sex is ascertained by providers.

3

There are race and ethnicity fields in electronic medical records (EMRs) for patients, but these are entirely unpopulated and therefore unusable for the purposes of our analysis.

4

The clinical diagnoses shown are the 15 most common diagnoses ever recorded in our data. All diagnoses refer to diagnoses made prior to the start date of the herpes zoster vaccination program in Ontario (i.e., prior to September 15 2016).

5

The preventive health services shown are measured prior to the start date of the herpes zoster vaccination program in Ontario (i.e., prior to September 15 2016).

6

All variables shown in the table, except for dementia, were calculated after excluding patients who received a diagnosis of dementia prior to the start date of the herpes zoster vaccination program (i.e., prior to September 15 2016) because this was the analysis sample for our primary analyses with dementia diagnosed during the follow-up period as outcome.

Using CHSS data and our regression discontinuity approach of comparing individuals born immediately before versus immediately after the given date-of-birth eligibility threshold, we estimated that being eligible for herpes zoster vaccination based on the Jan 1, 1946 threshold led to a 27.4 (95% CI: 0.0 – 54.7, p=0.049) percentage point increase in herpes zoster vaccination receipt (Figure S3). Using the Jan 1, 1945 threshold, the corresponding estimate was 13.9 (95% CI: −5.8 – 33.6, p=0.167) percentage points.

Between January 1 2017 and June 30 2022, 10,789 individuals at primary care providers in Ontario received a new diagnosis of dementia. When restricting our sample to patients in Ontario who were born within the 12-month bandwidth around the Jan 1, 1946 date-of-birth eligibility threshold, the number of new dementia diagnoses during the follow-up period were 751 (4.3% [381/8795)] eligible and 5.3% [370/6997] ineligible patients); for the Jan 1, 1945 threshold, the number of new diagnoses between September 15 2016 and June 30 2022 was 818 (5.5% [382/6997] eligible and 6.2% [436/6993] ineligible patients).

Being born immediately before versus immediately after the Jan 1, 1946 eligibility threshold for herpes zoster vaccination decreased the probability of receiving a new dementia diagnosis by an absolute difference of 2.0 (95% CI: 0.4 – 3.5, p=0.012) percentage points over the 5.5-year follow-up period (Figure 1). This effect was robust to using a local quadratic (instead of linear) polynomial, using uniform (instead of triangular) kernel weights, including patients with a diagnosis of dementia received prior to the start date of the vaccination program, including inactive patients, excluding primary care provider-level fixed effects, using the comparative regression discontinuity design, and across grace periods ranging from zero to 12 months (Figure 1). The point estimates also remained stable when using bandwidths ranging from 12 to six months (Figure 1). Similarly, when we controlled for baseline diagnoses for the top 15 most common clinical conditions in the data, herpes zoster vaccination eligibility reduced new diagnoses of dementia over our 5.5-year follow-up period by 1.9 (95% CI: 0.4 – 3.4; p = 0.016) percentage points. In an additional robustness check for which we defined the outcome as any new dementia diagnoses between two ages for all patients (instead of between Jan 1, 2017 and Jun 30, 2022) to more explicitly control for age, eligibility for herpes zoster vaccination at the Jan 1, 1946 date-of-birth threshold led to a 1.6 (95% CI: 0.3 – 3.0, p = 0.016) percentage point reduction in new dementia diagnoses between ages 70 and 75.5.

Figure 1: The effect of being eligible for HZ vaccination (based on being born immediately after versus immediately before January 1 1946) on new dementia diagnoses.1,2,3.

Figure 1:

Open in a new tab

Abbreviations: CI = Confidence Interval; Incl. = Including; Excl. = Excluding; FEs= Fixed Effects; CRD = Comparative Regression Discontinuity

1 The main specification used a local linear polynomial, triangular kernel weights, a grace period of zero months, and a 12-month bandwidth.

2 With “grace periods” we refer to time periods since January 1 2017 after which follow-up time is considered to begin.

3 Dots show the point estimate and horizontal bars the 95% confidence interval.

In our secondary analysis using the Jan 1, 1945 instead of the Jan 1, 1946 date-of-birth eligibility threshold, we estimated that being eligible for herpes zoster vaccination decreased the probability of a new dementia diagnosis by an absolute difference of 2.0 (95% CI: 0.2 – 3.8, p=0.025) percentage points over 5.5 years (Figure S4). This effect was robust across a variety of analytical specifications (Figure S4). Our comparative regression discontinuity analyses were also robust across specifications using either date-of-birth eligibility threshold (Figure S5). When we defined the outcome as new dementia diagnoses between two ages (71.0 to 76.75 years) instead of between the start date of the programme and the end date of our dataset, eligibility for herpes zoster vaccination at the Jan 1, 1945 date-of-birth threshold led to a 1.9 (95% CI: 0.3 – 3.6, p = 0.020) percentage point reduction in dementia diagnoses.

The effect of herpes zoster vaccination eligibility on new diagnoses of dementia was statistically significant among women but not among men (Figure 2). The test for interaction by gender, using the main specification (i.e., triangular kernel, local-linear polynomial, excluding people who had a diagnosis of dementia at baseline, excluding inactive patients, including providerlevel fixed effects, 12-month bandwidth, and no grace period) yielded a p-value of 0.218. Gender heterogeneity effects at the Jan 1, 1945 threshold are presented in Figure S6.

Figure 2: The effect of being eligible for HZ vaccination (based on being born immediately after versus immediately before January 1 1946) on new dementia diagnoses, separately for women and men.1,2,3.

Figure 2:

Open in a new tab

Abbreviations: CI = Confidence Interval; Incl. = Including; Excl. = Excluding; FEs = Fixed Effects; CRD = Comparative Regression Discontinuity; Int. = Interaction

1 The main specification used a local linear polynomial, triangular kernel weights, a grace period of zero months, and a 12-month bandwidth.

2 With “grace periods” we refer to time periods since January 1 2017 after which follow-up time is considered to begin.

3 Dots show the point estimate and horizontal bars the 95% confidence interval.

After adjusting for multiple hypothesis testing, there was no indication of any systematic differences between individuals born immediately before versus immediately after the Jan 1, 1946 threshold (Figure 3). This was also true when using the Jan 1, 1945 threshold (Figure S7). Similarly, we find no evidence of baseline imbalances when more explicitly controlling for the effect of age by measuring diagnoses in the same age range across patients, instead of between dates (Table S3). In addition, except for dementia, we found no significant effects from herpes zoster vaccination eligibility on any of the 15 most common diagnoses in our data nor for prescriptions of antihypertensive and statin medications (Figure S8). Neither did herpes zoster vaccination eligibility have an effect on educational attainment nor total household income in the CHSS (Figure S9). We also found this lack of an effect on diagnoses and medication prescriptions in the CPCSSN and socioeconomic variables in the CHSS when using the Jan 1, 1945 date-of-birth threshold to define eligibility (Figure S9, S10, and S11). We observed a significant effect on dementia incidence only for the Jan 1, 1946 and Jan 1, 1945 date-of-birth threshold (i.e., the two thresholds used by the herpes zoster vaccination programme) and not for the first day of other years (Figure S12). Finally, when performing a regression discontinuity analysis using patients from all provinces that did not publicly fund herpes zoster vaccination, with the same analytical specification as in Ontario, the estimate for new dementia diagnoses was −0.93 (95% CI: −3.12 — 1.26, p = 0.407) percentage points at the Jan 1, 1946 threshold and 0.9 (95% CI: −1.71 — 3.52, p = 0.498) percentage points at the Jan 1, 1945 threshold.

Figure 3: Baseline balance checks using the 15 most common diagnoses in the CPCSSN data as well as statin and antihypertensive medication use.1,2,3.

Figure 3:

Open in a new tab

Abbreviations: CI = Confidence Interval; COPD = Chronic Obstructive Pulmonary Disease

1 All analyses shown are for the January 1 1946 date-of-birth eligibility threshold.

2 Dots show the point estimate and horizontal bars the 95% confidence interval.

3 We present sharpened q-values (alongside p-values) that represent the false positive rate to adjust for multiple hypothesis testing.51

10,259 patients (4.4%) died in the follow-up period and 20,654 patients (8.9%) were censored because of a dementia diagnosis or death. When analyzing our data using an accelerated failure time-regression discontinuity (AFT-RD) model, we found that the acceleration factor was significantly greater than one for all bandwidths (in one-month increments) ranging from six to 12 months (Figure S13), indicating that those born immediately after (and, thus, eligible for herpes zoster vaccination) Jan 1, 1946 experienced on average a longer time without a dementia diagnosis over 5.5 years than those born immediately before Jan 1, 1946. This was also the case in our secondary analysis of using the Jan 1, 1945, instead of Jan 1, 1946, threshold (Figure S14). In addition, both the Kaplan-Meier survival curves as well as the cumulative incidence curves, plotted within a narrow bandwidth around the Jan 1, 1946, threshold, identified a significantly longer time without a dementia diagnosis for herpes zoster vaccination-eligible than ineligible patients (Figure S15). Again, the same held true when using the Jan 1, 1945, threshold instead (Figure S16). Figures S17 and S18 show the Kaplan-Meier and cumulative incidence curves separately for women and men. Figure S19 indicates that eligible and ineligible patients within a narrow bandwidth around both eligibility thresholds were similar at baseline in the 15 most common diagnoses in the CPCSSN data.

Both our synthetic difference-in-differences and synthetic control method approach achieved near-identical trends in dementia incidence for the time period before the start of the herpes zoster vaccination programme between the herpes zoster vaccine-eligible birth cohort in Ontario and the same birth cohort in other provinces (Figure S20, S21, and S22; Table S4). After the start of the herpes zoster vaccination programme, there was a significantly lower incidence of new dementia diagnoses among the herpes zoster vaccine-eligible birth cohort (born between Jan 1, 1945 and Dec 31, 1946) in Ontario than in the same birth cohort in other provinces (Figure 4). This was also the case when defining the eligible birth cohort as those born between Jan 1, 1945 and Dec 31, 1950. Although unable to provide meaningful p-values because of the limited number of provinces in Canada, the SCM approach yielded similar results to the SDID approach.

Figure 4: The effect of being eligible for HZ vaccination on dementia incidence between the HZ vaccine-eligible birth cohorts in Ontario compared to the same birth cohorts in other provinces, across follow-up periods in one-month increments.1,2.

Figure 4:

Figure 4:

Open in a new tab

Abbreviations: HZ = Herpes Zoster

1 This analysis used the Synthetic Difference-in-Differences method.

2 The “1945-1946 cohort” was defined as those born between January 1 1945 and December 31 1946. The “1945-1950 cohort” was defined as those born between January 1 1945 and December 31 1950.

Discussion

Being born immediately after versus immediately before the date-of-birth eligibility threshold for Ontario’s herpes zoster vaccination programme led to a large difference in herpes zoster vaccination receipt, and these groups of individuals, who are not expected to differ systematically, also had a large difference in the probability of receiving a new dementia diagnosis over a 5.5-year follow-up period. Components of our study that increase the credibility of our findings are i) the replication of the finding of a protective effect on dementia incidence at a second date-of-birth eligibility threshold for herpes zoster vaccination that was used in Ontario’s herpes zoster vaccination programme; ii) a reduction in dementia incidence after the start of Ontario’s herpes zoster vaccination programme among vaccine-eligible birth cohorts in Ontario compared to the same birth cohorts in other Canadian provinces (none of which provided publicly funded herpes zoster vaccination during the study period); and iii) that the only two January-1 date-of-birth thresholds that led to a reduction in new diagnoses of dementia were the two thresholds that were used by Ontario’s herpes zoster vaccination programme.

The key strength of this study is its quasi-randomized design. Except for two natural experiments by our group,13,14 published evidence has compared those who opt to be vaccinated with those who do not, whilst adjusting (using analysis methods such as multivariable regression or propensity score matching) for potential confounding variables that differ between vaccine recipients and non-recipients.15–23 The validity of this approach relies on having sufficient information available in electronic health record and medical claims data to fully control for these differences.43 Unfortunately, information on many important confounding variables, especially health-related behaviors, is not available in these data sources. Our approach instead compares groups of individuals born on either side of the date-of-birth eligibility threshold for herpes zoster vaccination, which are expected to be balanced in observed and unobserved potential confounding variables except for the large difference in the probability of receiving herpes zoster vaccination.

Because our approach focuses on comparing individuals born immediately on either side of the date-of-birth eligibility threshold for herpes zoster vaccination, the critical advantage of our study design is that any potential confounding variable would have to change abruptly at this threshold to bias our analysis.40,41 Such confounding could occur if another intervention or policy with a large protective effect on dementia incidence used the identical date-of-birth threshold as the herpes zoster vaccination programme in Ontario. However, our analyses suggest that this is unlikely to be a source of bias in our study. These findings include i) the lack of a difference across the date-of-birth eligibility threshold in common disease diagnoses other than dementia, nor in educational attainment or household income; ii) no difference in the incidence of dementia across the date-of-birth eligibility threshold prior to the start date of the herpes zoster vaccination programme; iii) the observation that the only Jan 1 date-of-birth thresholds that had an effect on new diagnoses of dementia were the two thresholds (Jan 1 1946 and Jan 1 1945) used by Ontario’s herpes zoster vaccination programme; and iv) the lack of an effect on dementia incidence at the same date-of-birth thresholds in Canadian provinces other than Ontario.

We believe that our findings warrant further research into the relationship between herpes zoster vaccination and dementia and neuroimmune health more generally.44 In addition to investing in a randomized trial, investments into the mechanism through which herpes zoster vaccination may affect dementia could yield important insights. One possible mechanism is an immunomodulatory pathway that is independent of the varicella zoster virus. The fact that studies have reported a protective association with dementia for vaccines other than those for herpes zoster given in older age adds plausibility to such a mechanism.28–30 A few such pathways have been described elsewhere.45 Another possible mechanism is through a reduction in subclinical and clinical reactivations of the varicella zoster virus. Research on such reactivations of the virus has already identified several potential pathways, including that they can result in neuroinflammation,46,47 amyloid deposition and tau-protein aggregation,48 cerebrovascular disease with a pattern similar to that seen in Alzheimer’s disease,46,47 and vasculopathy that can lead to long-term cognitive impairment.49 In addition, reactivations of the varicella zoster virus may result in reactivations of the herpes simplex virus in the brain,50 thus connecting the varicella zoster virus to the larger body of evidence on the herpes simplex virus as a causative factor in the development of dementia.1,2,7 How the herpes zoster vaccine may have benefits for neuroimmune health in older age more generally has been recently detailed elsewhere.44

Our study has limitations. First, we are unable to pinpoint the precise magnitude of our observed effects given the wide 95% confidence intervals around our estimates. The statistical power in our analysis is limited by the relatively short follow-up period and comparatively low incidence of dementia in individuals’ early 70s.37 Second, we were unable to determine the magnitude of the effect of herpes zoster vaccination receipt (as opposed to vaccination eligibility) on incident dementia because the patients in the learning health systems of the CPCSSN differed systematically from the population-representative sample of seniors in the CHSS data. Third, there is likely to be a degree of underascertainment and delay in diagnosing dementia in our electronic health record data, although this is unlikely to differ between individuals born immediately on either side of the date-of-birth eligibility threshold. More broadly, for any limitations in data quality (or events that could affect dementia diagnoses, such as the COVID-19 pandemic) to introduce bias into our analysis, they would have to change abruptly at the Jan 1, 1946 and Jan 1, 1945 date-of-birth thresholds.40,41 Fourth, our follow-up period was constrained to 5.5 years. As such, we are unable to determine longer-term effects of herpes zoster vaccination on dementia incidence. Fifth, because regression discontinuity assigns the greatest weight to those individuals born in closest proximity to the date-of-birth eligibility threshold, we are unable to determine the effect of herpes zoster vaccination in other age groups. Sixth, the primary care practices that are part of the CPCSSN network are not a random sample of all primary care practices in Ontario, but rather pertain to patients enrolled at primary care practices that chose to be part of a network that focusses on healthcare quality improvement and research. Seventh, because regression discontinuity is implemented using local regression,40,41 the effect estimates obtained in regression discontinuity are absolute effect estimates (in percentage points) as opposed to relative effects. Eighth, although our across-province comparison using SDID and SCM is robust to confounding from variables that change over time but do not differ in their trend between Ontario and other Canadian provinces (or changes that already began prior to the start of the follow-up period such that they are taken into account by our weighting algorithm), we cannot exclude potential confounding from variables that change differentially over time between Ontario and other Canadian provinces during the follow-up period. Lastly, given that Ontario’s herpes zoster vaccination programme began with funding the live-attenuated as opposed to the recombinant herpes zoster vaccine,32 our findings pertain to the live-attenuated vaccine only.

In conclusion, this study provides strong evidence of a protective effect of eligibility for herpes zoster vaccination on incident dementia. Our main analysis exploited a unique programme in Ontario whereby two date-of-birth eligibility thresholds created similar groups who likely only differed in their eligibility for a free herpes zoster vaccination and a few weeks in age. Using this design, we demonstrated that both date-of-birth eligibility thresholds reduced the probability of a new dementia diagnosis over a 5.5-year time period. These results suggest that herpes zoster vaccination can be an effective tool to prevent, or delay, dementia and have implications for research into the pathogenesis of dementia and potentially maintenance of neuroimmune health in older age more broadly.

Supplementary Material

Supplement

Research in context.

Evidence before this study:

We searched PubMed for articles published from database inception to Nov 15, 2025, with no language restrictions, using variations of the search terms “herpes zoster”, “vaccination”, and “dementia”. Except for two natural experiments by our group, all published studies that investigated an association between herpes zoster vaccination and dementia lack a causal interpretation. These studies compared vaccinated individuals with those who were unvaccinated while attempting to control for all dementia-related characteristics that differed between these groups. These studies therefore assumed that the databases used (i.e., medical claims or electronic health record data) contained sufficient information to fully control for all differences in health-related motivation and behaviour between vaccine recipients and nonrecipients, including factors such as physical activity, diet, and education, which are usually unavailable or poorly recorded in such data sources.

Added value of this study

This is the first natural experiment in Canada to investigate the effect of herpes zoster vaccination on dementia diagnoses and the first to triangulate its findings using a second natural experiment and different quasi-experimental approach. We compared patient groups that had different vaccination eligibility based on date of birth. Therefore, we avoided relying on the assumption that medical claims or electronic health record data contain sufficient information on health-related motivation and behaviours to fully adjust for all dementia-related differences between vaccine recipients and non-recipients. Instead, we assumed that, after flexibly controlling for age, two large patient groups whose only difference is a discrepancy by a few weeks are probably similar in observed and unobserved characteristics. Confounding in our analysis, therefore, only occurs if the confounding variable changes abruptly at the date of birthbased eligibility threshold for vaccination.

Implications of all the available evidence:

The only other natural experiment studies of the effect of herpes zoster vaccination on dementia incidence used data from Wales and Australia. However, these studies were unable to triangulate the findings using several date-of-birth-based eligibility thresholds or compare their findings to areas of the country that did not implement a herpes zoster vaccination programme. Our study provides additional evidence to suggest a protective cause-and-effect relationship between herpes zoster vaccination and incident dementia, as well as new research avenues for the pathogenesis of dementia and the maintenance of neuroimmune health in older age.

Acknowledgements

The authors are grateful for the continuous advice and support from the CPCSSN team who made this data available for research. The data that support the findings of this study are available from CPCSSN. This study was funded by grants to P.G. by the Stanford Center for Digital Health (no grant number), The Phil & Penny Knight Initiative for Brain Resilience at the Wu Tsai Neurosciences Institute, Stanford University (KPI-003), the National Institute on Aging (R01AG084535), National Institute of Allergy and Infectious Diseases (DP2AI171011), and Biohub, San Francisco (no grant number).

Funding:

National Institute on Aging, National Institute of Allergy and Infectious Diseases, Stanford Center for Digital Health, Stanford Knight Initiative for Brain Resilience, Biohub.

Role of the funding source

The funder of the study had no role in study design, data collection, data analysis, data interpretation, writing of the manuscript, or the decision to submit.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare no competing interests.

Data sharing

Researchers must request access to the data directly from CPCSSN. The authors have no permission to share the data. All codes to define variables are available in the Appendix. All statistical analysis code (in R and STATA) will be made available in a publicly accessible repository upon acceptance of the manuscript for publication.

References

  • 1.Moir RD, Lathe R, Tanzi RE. The antimicrobial protection hypothesis of Alzheimer’s disease. Alzheimers Dement. 2018. Dec;14(12):1602–14. [DOI] [PubMed] [Google Scholar]
  • 2.Devanand DP. Viral Hypothesis and Antiviral Treatment in Alzheimer’s Disease. Curr Neurol Neurosci Rep. 2018. Jul 14;18(9):55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Itzhaki RF, Lathe R, Balin BJ, Ball MJ, Bearer EL, Braak H, et al. Microbes and Alzheimer’s Disease. J Alzheimers Dis. 2016;51(4):979–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Arvin A, Campadelli-Fiume G, Mocarski E, et al. Human Herpesviruses: Biology, Therapy, and Immunoprophylaxis. Cambridge University Press, 2007. [PubMed] [Google Scholar]
  • 5.Matthews E, Beckham JD, Piquet AL, Tyler KL, Chauhan L, Pastula DM. Herpesvirus-Associated Encephalitis: an Update. Curr Trop Med Rep. 2022;9(3):92–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Eimer WA, Vijaya Kumar DK, Navalpur Shanmugam NK, Rodriguez AS, Mitchell T, Washicosky KJ, et al. Alzheimer’s Disease-Associated β-Amyloid Is Rapidly Seeded by Herpesviridae to Protect against Brain Infection. Neuron. 2018. Jul 11;99(1):56–63.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Hyde VR, Zhou C, Fernandez JR, Chatterjee K, Ramakrishna P, Lin A, et al. Antiherpetic tau preserves neurons via the cGAS-STING-TBK1 pathway in Alzheimer’s disease. Cell Rep. 2025. Jan 28;44(1):115109. [DOI] [PubMed] [Google Scholar]
  • 8.Bettcher BM, Tansey MG, Dorothée G, Heneka MT. Peripheral and central immune system crosstalk in Alzheimer disease - a research prospectus. Nat Rev Neurol. 2021. Nov;17(11):689–701. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Adhisivam B, Kamalarathnam C, Bhat BV, Jayaraman K, Namachivayam SP, Shann F, et al. Effect of BCG Danish and oral polio vaccine on neonatal mortality in newborn babies weighing less than 2000 g in India: multicentre open label randomised controlled trial (BLOW2). BMJ. 2025. Sept 22;390:e084745. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Benn CS, Fisker AB, Rieckmann A, Sørup S, Aaby P. Vaccinology: time to change the paradigm? Lancet Infect Dis. 2020. Oct;20(10):e274–83. [DOI] [PubMed] [Google Scholar]
  • 11.Benn CS, Amenyogbe N, Björkman A, Domínguez-Andrés J, Fish EN, Flanagan KL, et al. Implications of Non-Specific Effects for Testing, Approving, and Regulating Vaccines. Drug Saf. 2023. May;46(5):439–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Aaby P, Benn CS, Flanagan KL, Klein SL, Kollmann TR, Lynn DJ, et al. The non-specific and sex-differential effects of vaccines. Nat Rev Immunol. 2020. Aug;20(8):464–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Eyting M, Xie M, Michalik F, HeB S, Chung S, Geldsetzer P. A natural experiment on the effect of herpes zoster vaccination on dementia. Nature. 2025. May;641 (8062):438–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Pomirchy M, Bommer C, Pradella F, Michalik F, Peters R, Geldsetzer P. Herpes Zoster Vaccination and Dementia Occurrence. JAMA. 2025. Jun 17;333(23):2083–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Taquet M, Dercon Q, Todd JA, Harrison PJ. The recombinant shingles vaccine is associated with lower risk of dementia. Nat Med. 2024. Oct;30(10):2777–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Wiemken TL, Salas J, Morley JE, Hoft DF, Jacobs C, Scherrer JF. Comparison of rates of dementia among older adult recipients of two, one, or no vaccinations. Journal of the American Geriatrics Society. 2022;70(4):1157–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Lophatananon A, Carr M, Mcmillan B, Dobson C, Itzhaki R, Parisi R, et al. The association of herpes zoster and influenza vaccinations with the risk of developing dementia: a population-based cohort study within the UK Clinical Practice Research Datalink. BMC Public Health. 2023. Oct 2;23(1):1903. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Douros A, Ante Z, Suissa S, Brassard P. Common Vaccines and the Risk of Incident Dementia: A Population-based Cohort Study. J Infect Dis. 2023. May 29;227(11):1227–36. [DOI] [PubMed] [Google Scholar]
  • 19.Ukraintseva S, Yashkin AP, Akushevich I, Arbeev K, Duan H, Gorbunova G, et al. Associations of infections and vaccines with Alzheimer’s disease point to a role of compromised immunity rather than specific pathogen in AD. Exp Gerontol. 2024. Jun 1;190:112411. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Lehrer S, Rheinstein PH. Herpes Zoster Vaccination Reduces Risk of Dementia. In Vivo. 2021. Nov 3;35(6):3271–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Wilkinson T, Schnier C, Bush K, Rannikmäe K, Lyons RA, McTaggart S, et al. Drug prescriptions and dementia incidence: a medication-wide association study of 17000 dementia cases among half a million participants. J Epidemiol Community Health. 2022. Mar;76(3):223–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Tang E, Ray I, Arnold BF, Acharya NR. Recombinant zoster vaccine and the risk of dementia. Vaccine. 2025. Feb 6;46:126673. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Polisky V, Littmann M, Triastcyn A, Horn M, Georgiou A, Widenmaier R, et al. Varicella-zoster virus reactivation and the risk of dementia. Nat Med. 2025. Oct;1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Remschmidt C, Wichmann O, Harder T. Frequency and impact of confounding by indication and healthy vaccinee bias in observational studies assessing influenza vaccine effectiveness: a systematic review. BMC Infectious Diseases. 2015. Oct 17;15(1):429. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Wiemken TL, Salas J, Hoft DF, Jacobs C, Morley JE, Scherrer JF. Dementia risk following influenza vaccination in a large veteran cohort. Vaccine. 2021;39(39):5524–5531. [DOI] [PubMed] [Google Scholar]
  • 26.Wu SY, Liu JC, Liu SH. MP311 Influenza vaccination reduces dementia risk in chronic kidney disease patients: a population-based cohort study. Nephrol Dial Transplant. 2016. May 1;31(suppl_1):i442. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Wu X, Yang H, He S, Xia T, Chen D, Zhou Y, et al. Adult Vaccination as a Protective Factor for Dementia: A Meta-Analysis and Systematic Review of Population-Based Observational Studies. Front Immunol. 2022. May 3;13:872542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Yang WK, Shao SC, Liu CC, Chi CC. Influenza vaccination and risk of dementia: a systematic review and meta-analysis. Age Ageing. 2025. Jul 1;54(7):afaf169. [DOI] [PubMed] [Google Scholar]
  • 29.Veronese N, Demurtas J, Smith L, Michel JP, Barbagallo M, Bolzetta F, et al. Influenza vaccination reduces dementia risk: A systematic review and meta-analysis. Ageing Research Reviews. 2022. Jan 1;73:101534. [DOI] [PubMed] [Google Scholar]
  • 30.Bukhbinder AS, Ling Y, Harris K, Jiang X, Schulz PE. Do vaccinations influence the development of Alzheimer disease? Hum Vaccin Immunother. 2023. Aug 1;19(2):2216625. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Statistics Canada, Annual Demographic Estimates: Canada, Provinces and Territories, 2022. https://www150.statcan.gc.ca/n1/pub/91-215-x/2022001/sec1-eng.htm (accessed June 2 2025).
  • 32.Martins D, McCormack D, Tadrous M, Gomes T, Kwong JC, Mamdani MM, et al. Impact of a Publicly Funded Herpes Zoster Immunization Program on the Burden of Disease in Ontario, Canada: A Population-based Study. Clinical Infectious Diseases: An Official Publication of the Infectious Diseases Society of America. 2020. Jan 10;72(2):279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Government of Ontario, Ontario Making Shingles Vaccine Free for Seniors, news.ontario.ca. https://news.ontario.ca/en/release/41815/ontario-making-shingles-vaccine-free-for-seniors (accessed November 15, 2024).
  • 34.Shingles Vaccine Free for Ontario Seniors, news.ontario.ca. https://news.ontario.ca/en/release/43226/shingles-vaccine-free-for-ontario-seniors (accessed June 6, 2025).
  • 35.Canadian Primary Care Sentinel Surveillance Network (CPCSSN). https://cpcssn.ca/ (accessed November 16, 2024).
  • 36.Statistics Canada, Government of Canada, Canadian Health Survey on Seniors (CHSS), 2019. https://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=5267 (accessed June 6, 2025).
  • 37.2024 Alzheimer’s disease facts and figures. Alzheimers Dement. 2024. May;20(5):3708–821. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Boyle PA, Yu L, Leurgans SE, Wilson RS, Brookmeyer R, Schneider JA, et al. Attributable risk of Alzheimer’s dementia attributed to age-related neuropathologies. Ann Neurol. 2019. Jan;85(1):114–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Boyle PA, Yu L, Wilson RS, Leurgans SE, Schneider JA, Bennett DA. Person-specific contribution of neuropathologies to cognitive loss in old age. Ann Neurol. 2018. Jan;83(1):74–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Cattaneo MD, Keele L, Titiunik R. A guide to regression discontinuity designs in medical applications. Statistics in Medicine. 2023;42(24):4484–513. [DOI] [PubMed] [Google Scholar]
  • 41.Cattaneo MD, Idrobo N, Titiunik R. A Practical Introduction to Regression Discontinuity Designs: Foundations. Elements in Quantitative and Computational Methods for the Social Sciences, 2019. [Google Scholar]
  • 42.Villamizar-Villegas M, Pinzon-Puerto FA, Ruiz-Sanchez MA. A comprehensive history of regression discontinuity designs: An empirical survey of the last 60 years. Journal of Economic Surveys. 2022;36(4):1130–78. [Google Scholar]
  • 43.Hernan MA, Robins JM. Causal Inference: What If. Chapman & Hall/CRC, 2020. [Google Scholar]
  • 44.Huang X, Gu BJ. Shingles vaccination and neuroimmune vulnerability. Trends Neurosci. 2025. Sept;48(9):655–62. [DOI] [PubMed] [Google Scholar]
  • 45.Prosser LA, Harpaz R, Rose AM, Gebremariam A, Guo A, Ortega-Sanchez IR, et al. A Cost-Effectiveness Analysis of Vaccination for Prevention of Herpes Zoster and Related Complications: Input for National Recommendations. Ann Intern Med. 2019. Mar 19;170(6):380–8. [DOI] [PubMed] [Google Scholar]
  • 46.Nagel MA, Traktinskiy I, Stenmark KR, Frid MG, Choe A, Gilden D. Varicella-zoster virus vasculopathy: immune characteristics of virus-infected arteries. Neurology. 2013. Jan 1;80(1):62–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Nagel MA, Traktinskiy I, Azarkh Y, Kleinschmidt-DeMasters B, Hedley-Whyte T, Russman A, et al. Varicella zoster virus vasculopathy: analysis of virus-infected arteries. Neurology. 2011. Jul 26;77(4):364–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Bubak AN, Como CN, Hassell JE, Mescher T, Frietze SE, Niemeyer CS, et al. Targeted RNA Sequencing of VZV-Infected Brain Vascular Adventitial Fibroblasts Indicates That Amyloid May Be Involved in VZV Vasculopathy. Neurol Neuroimmunol Neuroinflamm. 2022. Jan;9(1):e1103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Silver B, Nagel MA, Mahalingam R, Cohrs R, Schmid DS, Gilden D. Varicella zoster virus vasculopathy: a treatable form of rapidly progressive multi-infarct dementia after 2 years’ duration. J Neurol Sci. 2012. Dec 15;323(1–2):245–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Cairns DM, Itzhaki RF, Kaplan DL. Potential Involvement of Varicella Zoster Virus in Alzheimer’s Disease via Reactivation of Quiescent Herpes Simplex Virus Type 1. J Alzheimers Dis. 2022;88(3):1189–200. [DOI] [PubMed] [Google Scholar]
  • 51.Anderson ML. Multiple Inference and Gender Differences in the Effects of Early Intervention: A Reevaluation of the Abecedarian, Perry Preschool, and Early Training Projects. Journal of the American Statistical Association, 2008. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement
NIHMS2152981-supplement-Supplement.docx (1.2MB, docx)

Data Availability Statement

Researchers must request access to the data directly from CPCSSN. The authors have no permission to share the data. All codes to define variables are available in the Appendix. All statistical analysis code (in R and STATA) will be made available in a publicly accessible repository upon acceptance of the manuscript for publication.

RESOURCES