Public health impact of herpes zoster vaccination on older adults in Singapore: a modeling study

ABSTRACT

In Singapore, population aging and rising life expectancy are increasing herpes zoster (HZ) burden, which may be reduced by vaccination. The present study modeled the public health impact of HZ vaccination in Singapore using ZOster ecoNomic Analysis (ZONA) model adapted with Singapore-specific key model inputs, where available. Base case analysis was conducted in adults ≥ 50 years of age (YOA), exploring three vaccination strategies (no vaccination, recombinant zoster vaccine [RZV], zoster vaccine live [ZVL]) under mass vaccination setting (30% coverage). Scenario and sensitivity analyses were performed. Out of 1.51 million adults in 2021 (base case population), 406,513 (27.0%) cases of HZ, 68,264 (4.5%) cases of post-herpetic neuralgia (PHN), and 54,949 (3.6%) cases of other complications were projected without vaccination. RZV was estimated to avoid 73,129 cases of HZ, 11,094 cases of PHN, and 9,205 cases of other complications over the subjects’ remaining lifetime; ZVL would avoid 17,565 cases of HZ, 2,781 cases of PHN, and 1,834 cases of other complications. The number needed to vaccinate to prevent one case of HZ/PHN was lower for RZV (7/41) than ZVL (26/163). Among all five age-stratified cohorts (50–59/60–64/65–69/70–79/≥80 YOA), RZV (versus no vaccination/ZVL) avoided the largest number of cases in the youngest cohort, 50–59 YOA. Results were robust under scenario and sensitivity analyses. Mass vaccination with RZV is expected to greatly reduce the public health burden of HZ among Singapore individuals ≥ 50 YOA. Findings support value assessment and decision-making regarding public health vaccination strategies for HZ prevention in Singapore.

Plain Language Summary

Risk of shingles (herpes zoster) increases with age, especially from 50 years. Shingles is a major public health concern in Singapore, given its rapidly aging population. Vaccination can prevent shingles and reduce its public health burden. Two shingles vaccines are available in Singapore: recombinant zoster vaccine (RZV) since 2021, zoster vaccine live (ZVL) since 2008. To understand the value of preventing shingles via vaccination, this study assessed the public health impact of shingles vaccination. Three vaccination strategies (no vaccination, vaccination with RZV, vaccination with ZVL) were compared in 1.51 million Singapore adults aged 50 years and above. Without vaccination, public health burden of shingles would be high; an estimated 406,513 (27.0%) would have shingles, 68,264 (4.5%) would have shingles-related long-term nerve pain, 54,949 (3.6%) would have other shingles-related complications, and 17,762 (1.2%) would be hospitalized due to shingles. Shingles vaccination could reduce this public health burden: RZV avoided 73,129 cases of shingles, 11,094 cases of shingles-related long-term nerve pain, 9,205 cases of other shingles-related complications, and 2,827 hospitalizations due to shingles, which was 4–6 times that avoided with ZVL (shingles: 17,565; shingles-related long-term nerve pain: 2,781; other shingles-related complications: 1,834; hospitalizations due to shingles: 484). Shingles vaccination for adults aged 50 years and above, especially early vaccination from 50–59 years, could reduce its public health burden more than vaccination at later ages and contribute toward healthy aging, preventive care, and the Healthier SG initiative. Results support local public health value assessments and decision-making for shingles prevention.

GRAPHICAL ABSTRACT

KEYWORDS:

• Aging population
• Asia-pacific
• herpes zoster
• mass vaccination
• older adults
• post-herpetic neuralgia
• public health impact
• recombinant zoster vaccine
• Singapore
• vaccination

Introduction

Herpes zoster (HZ) is a viral infection characterized by a painful, blistering rash typically distributed along 1–2 adjacent dermatome(s).^1–3 HZ occurs when latent varicella-zoster virus (VZV) is reactivated, brought about when VZV-specific cell-mediated immunity falls below a critical threshold due to age-related decline in immunity and/or immunosuppression.¹

HZ skin lesions usually subside within 2–4 weeks,⁴ however, patients are significantly impacted by acute pain from HZ, particularly in terms of quality of life.⁵ Furthermore, patients may develop complications associated with HZ.⁶ Post-herpetic neuralgia (PHN), a type of persistent neuropathic pain lasting ≥90 days after rash onset,⁷ is the most common HZ-related complication and can affect up to 30% of HZ patients globally.² The impact of PHN on patients’ lives may be seen across their physical, psychological, social, and functional domains of health.⁶ Other HZ-related complications include ocular and neurological complications, such as stroke, and may result in long-term physical impairment.^6–8-10 HZ and its complications may have spill-over effects on the patient’s family, caregivers, and society,¹¹ in the form of caregiver burden and public health burden due to increased healthcare utilization.

Globally, the burden of HZ is substantial, with approximately one in three individuals expected to develop HZ in their lifetime.³ The risk of HZ grows with age, particularly from 50 years of age (YOA).¹² For individuals who live to 85 YOA or later, the estimated lifetime risk of HZ increases from one in three to one in two.¹³ In Singapore, a recent study in the primary care setting reported that the mean age of HZ diagnosis in 2018–2020 was approximately 60 years and HZ was more prevalent among older adults ≥ 50 YOA (410–829 cases per 100,000 population) compared with younger adults 21–49 YOA (147–210 cases); expectedly, older age (≥50 years) was found to be a key risk factor for HZ.¹⁴ With the proportion of individuals aged ≥50 years in Singapore increasing from 29.8% in 2011 to 38.3% in 2022, the resident population is rapidly aging.¹⁵ Therefore, the burden of HZ and its complications is estimated to be significant and is expected to increase with population aging. Notably, the seroprevalence of VZV antibodies has been reported in approximately 88% of Singapore adults ≥ 25 YOA,¹⁶ suggesting that majority of adults are predisposed to developing HZ.

A single-center study of patients ≥ 50 YOA seeking medical attention for HZ or PHN locally (N = 347) found that up to over 80% of patients reported prodromal (e.g., pain, itch, and insomnia) and clinical symptoms (e.g., pain or discomfort, blistering rashes followed by scarring, and insomnia or sleep disturbance) of HZ.¹⁷ Pain was reported as the most unbearable HZ-associated symptom, with a large proportion of patients (55.6%) tapping into healthcare resources for analgesics.¹⁷ Generally, HZ may be treated with anti-virals and/or managed with analgesics¹⁸; in Singapore, traditional Chinese medicine (i.e., acupuncture) may also be considered for treatment.¹⁹ Specifically, anti-virals should be initiated within 72 hours of rash onset (i.e., acute HZ) for optimal efficacy^3,20; however, patients may be delayed in seeking medical attention and pharmacologic treatment,²¹ due in part to nonspecific symptoms at HZ onset,⁶ or potentially low awareness of HZ burden.²² Studies have shown that anti-virals and/or analgesics may be considered suboptimal in the management of HZ due to limited efficacy against PHN,^23–26 and minimal impact on patient quality of life.^6,26

Proactive prevention against HZ via vaccination would be an effective way to reduce the burden of HZ and its complications,^27–29 and therefore deliver preventive care and support healthy aging which are key objectives of Healthier SG.³⁰ The vaccines presently licensed for HZ prevention in adults ≥ 50 YOA in Singapore include the non-live, adjuvanted recombinant zoster vaccine (RZV),³¹ and zoster vaccine live (ZVL).³² RZV has demonstrated high and sustained vaccine efficacy estimates against HZ (89.8–97.2%) among adults ≥ 50 YOA in global clinical trials.^33,34 When comparing the efficacy of RZV versus ZVL indirectly across separate studies, RZV showed higher and longer-lasting estimates.^33–38

RZV and ZVL are available in Singapore through the private market and were registered for use in 2021 and 2008, respectively.³⁹ While the 2023 Society of Infectious Diseases Singapore guideline recommends HZ vaccination for immunocompetent adults ≥ 60 YOA or adults with prior HZ,⁴⁰ RZV and ZVL are not currently listed in the local National Adult Immunization Schedule.⁴¹

Public health impact (PHI) studies are helpful in understanding the value of vaccination in HZ prevention and its unmet public health need; such studies on RZV from other locales or regions such as the United States of America (USA), Canada, Germany, Japan, Hong Kong, and Beijing, China, have demonstrated superior PHI of RZV over no vaccination, ZVL (where evaluated), or both.^42–48 Therefore, this study aims to model the PHI of HZ vaccination in Singapore by comparing vaccination with RZV or ZVL versus no vaccination among adults ≥ 50 YOA.

Methods

Model overview

The ZOster ecoNomic Analysis (ZONA) model is a static multi-cohort Markov model developed in Microsoft Excel to simulate outcomes of HZ vaccination.^44,45 This study adapted the existing ZONA model to the Singapore setting through updating key model inputs, as well as undertaking relevant scenario and sensitivity analyses. Supplementary Figure S1 provides an overview of the ZONA model structure used.⁴⁴

The model has a cycle length of one year, following cohorts from the year of vaccination over their remaining lifetime. Transition probabilities between health states occur using an annual time step and probabilities of moving between health states are age-specific and to the extent possible, based on country-specific data. The model does not apply an age for the end of lifetime. Instead, age-specific all-cause mortality probabilities (Table 1) are progressively applied across the age cohorts over their remaining lifetime and the model cycles until no individuals are alive. All subjects remain in their initial cohort and all subsequent events are counted in that cohort only.

Table 1. Model inputs for demographic, epidemiology, and vaccination parameters.

	Age (years)	Base value	Range^a
			Lower limit	Upper limit
Demographics
Population size (n)^Citation15	50–59	584,188	N/A	N/A
	60–64	282,612	N/A	N/A
	65–69	235,391	N/A	N/A
	70–79	272,304	N/A	N/A
	≥80	131,313	N/A	N/A
Epidemiology^b
Incidence (probability 0–1)
HZ^c, Citation49	50–59	0.008	0.007	0.010
	60–64	0.011	0.009	0.013
	65–69	0.011	0.009	0.013
	70–79	0.012	0.010	0.015
	≥80	0.010	0.008	0.012
Proportion (%) of HZ cases with
PHN^d, Citation17,Citation50,Citation51	50–59	7.000	0.063	0.390
	60–64	14.450	0.063	0.390
	65–69	14.450	0.063	0.390
	70–79	14.450	0.063	0.390
	≥80	25.610	0.063	0.390
Ocular complications^e, Citation17,Citation52	50–59	2.870	1.000	11.000
	60–64	4.230	1.000	11.000
	65–69	4.230	1.000	11.000
	70–79	4.530	1.000	11.000
	≥80	6.910	1.000	11.000
Neurological complications^Citation52	50–59	2.230	0.600	3.860
	60–64	3.170	1.000	5.210
	65–69	3.170	1.000	5.210
	70–79	5.920	1.000	8.650
	≥80	4.880	1.000	7.570
Cutaneous complications^Citation52	50–59	1.590	0.210	2.980
	60–64	1.060	0.000	2.250
	65–69	1.060	0.000	2.250
	70–79	2.090	0.440	3.750
	≥80	2.850	0.770	4.920
Other non-pain complications^Citation52	50–59	1.590	0.210	2.980
	60–64	1.410	0.040	2.780
	65–69	1.410	0.040	2.780
	70–79	2.090	0.440	3.750
	≥80	2.850	0.770	4.920
Hospital resource utilization, mean number per case of HZ
Hospitalizations^Citation53	50–59	0.018	N/A	N/A
	60–64	0.018	N/A	N/A
	65–69	0.029	N/A	N/A
	70–79	0.048	N/A	N/A
	≥80	0.066	N/A	N/A
Mortality
All-cause mortality (probability 0–1)^f, Citation15	50–54	0.002	N/A	N/A
	55–59	0.004	N/A	N/A
	60–64	0.006	N/A	N/A
	65–69	0.010	N/A	N/A
	70–74	0.016	N/A	N/A
	75–79	0.028	N/A	N/A
	80–84	0.049	N/A	N/A
	85–89	0.089	N/A	N/A
	90–94	0.152	N/A	N/A
	95–99	0.167	N/A	N/A
	≥100	0.181	N/A	N/A
HZ-related mortality (proportion of HZ cases, %)^Citation54	50–54	0.001	N/A	N/A
	55–59	0.001	N/A	N/A
	60–64	0.002	N/A	N/A
	65–69	0.002	N/A	N/A
	70–74	0.006	N/A	N/A
	75–79	0.006	N/A	N/A
	80–84	0.024	N/A	N/A
	85–89	0.073	N/A	N/A
	90–94	0.073	N/A	N/A
	95–99	0.073	N/A	N/A
	≥100	0.073	N/A	N/A
Vaccination parameters
Initial vaccine efficacy (%) against HZ
RZV (one dose)^Citation33,Citation34	50–69	90.100	58.900	98.900
	≥70	69.500	24.900	89.100
RZV (two doses)^Citation33,Citation34	50–69	98.900	94.000	100.000
	≥70	95.400	89.700	100.000
ZVL^Citation35–37	50–59	69.800	54.100	80.600
	60–64	63.900	56.000	71.000
	65–69	63.900	56.000	71.000
	70–79	40.900	28.000	52.000
	≥80	18.300	0.000	48.000
Initial vaccine efficacy (%) against PHN
RZV (one dose)^Citation33,Citation34	50–69	90.100	58.900	98.900
	≥70	69.500	24.900	89.100
RZV (two doses)^Citation33,Citation34	50–69	98.900	94.000	100.000
	≥70	95.400	89.700	100.000
ZVL^Citation35–37	50–59	69.800	30.800	89.600
	60–64	65.700	25.400	84.200
	65–69	65.700	25.400	84.200
	70–79	73.400	51.600	85.800
	≥80	39.500	0.000	73.800
Annual waning of vaccine efficacy (%)
RZV (one dose)^g, ^Citation35–37	All ages, Years 1–4 post-vaccination	5.400	1.000	7.400
	All ages, Years ≥ 4 post-vaccination	5.100	3.600	6.900
RZV (two doses)^Citation45	50–69	1.500	0.000	3.400
	≥70	2.300	0.300	4.400
ZVL^Citation35–37	All ages, Years 1–4 post-vaccination	5.400	4.500	6.400
	All ages, Years ≥ 4 post-vaccination	5.100	4.100	6.000
Vaccine coverage (%)^Citation55,Citation56	All ages	30.000	N/A	N/A
Second-dose compliance (%)^Citation33,Citation34,Citation57	All ages	82.500	70.000	95.000

^aParameters indicated as ‘N/A’ were not varied in scenario or sensitivity analysis. ^bThe incidences of HZ and PHN were gender-weighted and combined. ^cThe incidences of initial HZ and recurrent HZ were assumed to be the same; the lower and upper bound values of the range of HZ incidence were calculated as a 20% reduction and increase from the base values. ^dThe proportions (%) of initial cases and recurrent cases of HZ with PHN were assumed to be the same, given that the incidence of initial HZ was assumed to be the same as recurrent HZ; the lower and upper bound values of the range of PHN incidence were obtained from Chen et al.¹⁷ (2015) and Goh et al.⁵¹ (1997), respectively. ^eThe upper bound value of the range of ocular complications (as a proportion of HZ cases) was obtained from Chen et al.¹⁷ (2015). ^fAll-cause mortality rates from SingStat¹⁵ were transformed to probabilities. The probability of all-cause mortality for the 95–99 YOA and ≥ 100 YOA age groups are based on an exponential smoothing forecast performed in Microsoft Excel. ^gNo clinical trials have been conducted to assess the vaccine efficacy of a single dose of RZV; the efficacy of one dose of RZV was thus derived based on a post-hoc analysis of patients from the ZOE-50 and ZOE-70 trials who had received one dose of RZV (including those who went on to receive the second dose and those who received only the first dose). HZ: herpes zoster; N/A: not applicable; PHN: post-herpetic neuralgia; RZV: recombinant zoster vaccine; YOA: years of age; ZVL: zoster vaccine live.

The base case target population was the 2021 population of adults ≥ 50 YOA in Singapore. The population was stratified into five age groups (50–59, 60–64, 65–69, 70–79, and ≥ 80 YOA), and the modeled impact of vaccination was analyzed for the entire study cohort ≥ 50 YOA (i.e., all five age groups combined) and for each individual age group.

The model explored three vaccination strategies: (1) no vaccination, (2) vaccination with RZV (two-dose), and (3) vaccination with ZVL (one-dose). Depending on the corresponding vaccination coverage for strategies (2) and (3), and the assumed second-dose compliance rate for strategy (2) only, individuals could be fully or partially compliant with the vaccine dosing schedule, or not vaccinated at all.

Adults ≥65 YOA in Singapore make up the Merdeka and Pioneer Generations eligible for additional government social support, including healthcare subsidies.^58,59 Hence, one scenario analysis was performed to assess the impact of vaccination with RZV or ZVL versus no vaccination, within an aggregated cohort of adults ≥ 65 YOA.

To assess the robustness of the base case results to uncertainties surrounding the model inputs, a deterministic one-way sensitivity analysis (DSA) was conducted on the number of HZ cases avoided with vaccination with RZV versus no vaccination in the mass vaccination setting (30% vaccination coverage).

Model inputs

The model inputs were categorized into (1) demographics, (2) epidemiology, and (3) vaccine efficacy, coverage, and second-dose compliance. Table 1 reports the complete list of model inputs obtained from Singapore-specific data sources where possible. Inputs for which Singapore-specific databases were unavailable were derived from other locales or regions (described below).

Demographics

The model applied population and mortality figures from the Singapore Department of Statistics 2021; the estimated resident population (including citizens and permanent residents) ≥50 YOA in Singapore in 2021 was 1.51 million.¹⁵ Age groups were aggregated to match the ZONA age cohorts, with gender data weighted and combined.

Epidemiology

HZ epidemiology in Asia-Pacific has generally been consistent with global data,^2,60 with higher incidence reported in Asia-Pacific.⁶¹ Therefore, relevant articles from a separate regional systematic literature review (SLR; Burden of HZ in selected locales in Asia-Pacific; unpublished) were identified to inform the epidemiological inputs in this study. The key epidemiological parameters of interest included age-stratified incidence of HZ, PHN, and other non-PHN complications (i.e., ocular, neurological, cutaneous, and other non-pain).

The selected studies and selection process for each epidemiological model input are as described below. Full details are reported in Appendix 1.

The epidemiological inputs for HZ and PHN were derived from Lin et al.⁴⁹ (2010) and Jih et al.⁵⁰ (2009), respectively, which reported on the burden of HZ and its epidemiology in Taiwan. Lin et al.⁴⁹ (2010) and Jih et al.⁵⁰ (2009) collected and reported data from the National Health Insurance Research Database (NHIRD), which is a population-level database in Taiwan and currently offers the largest possible sample size nationally. PHN epidemiological inputs were derived from Jih et al.⁵⁰ (2009) as it accounted for differences in the incidence of PHN across age groups. Gender-specific HZ and PHN incidence rates were weighted and combined according to each of the ZONA age groups. Sensitivity was tested at ± 20% of the base value for HZ incidence; Chen et al.¹⁷ (2015) and Goh et al.⁵¹ (1997) were used to determine the lower and upper bounds of PHN incidence, respectively, in the DSA.

The incidence of recurrent HZ was assumed to be the same as the incidence of initial HZ, as supported by published data.^62–64 As with HZ, the incidence of recurrent PHN was also assumed to be the same as that of its initial event.

For all non-PHN complications assessed (i.e., ocular, neurological, cutaneous, and other non-pain), a USA publication, Yawn et al.⁵² (2007), was used as a single data source to ensure consistency. Chen et al.¹⁷ (2015) was used to determine the upper bound for ocular complications in the DSA.

Separately, the probability for all HZ-related complications (including PHN) was conservatively assumed to be the same regardless of vaccination status. Although this assumption does not account for the potential impact of vaccination on the severity of HZ, a shorter duration and lower severity of pain have been shown in vaccinated versus non-vaccinated individuals.²⁹

HZ-related mortality data were sourced from a USA publication, Le et al.⁵⁴ (2015), as no regional or local publications identified from the literature search had reported these data. The use of a non-local source was based on the assumption that differences in HZ mortality across countries are negligible since HZ has usually been associated with morbidity rather than mortality, and HZ-associated deaths are expected to be low.^65,66 It was also assumed that no additional mortality would be attributable to PHN specifically, on top of that attributed to HZ.

The inputs for hospital resource utilization (mean number of hospitalizations per case of HZ, by age cohort) were derived based on the proportion of cases of HZ with HZ-related hospitalizations described by Lu et al.⁵³ (2018), a Taiwan cohort study reporting data from the NHIRD. As described above, the NHIRD may be considered a nationally representative data source.

Vaccine efficacy, coverage, and second-dose compliance

Model inputs for vaccine efficacy and its waning rates for RZV and ZVL have previously been described by Curran et al.^44,45 (2017, 2021) and remained unchanged during adaption of the ZONA model to the Singapore setting.

RZV efficacy against HZ and PHN was derived from the clinical studies ZOE-50 and ZOE-70,^33,34 and their ZOE-049 long-term follow-up (LTFU) extension study.⁶⁷ As RZV efficacy against PHN can be attributed to the vaccine’s effect against HZ and given the limited number of HZ cases in ZOE-50 and ZOE-70,⁶⁸ RZV efficacy against PHN was assumed the same as its efficacy against HZ. RZV efficacy was not adjusted for ethnicity since no variation was identified in an Asian subgroup analysis of global ZOE-50 and ZOE-70 trial data.⁶⁹ Overall, initial vaccine efficacy (i.e., take) was estimated at 98.9% (95% confidence interval [CI], 94.0–100%) and 95.4% (95% CI, 89.7–100%) for adults 50–69 YOA and ≥ 70 YOA, respectively.⁴⁵

RZV efficacy waning rates were extrapolated by applying linear approximation to annual vaccine efficacy estimates from ZOE-50, ZOE-70, and up to eight years of follow-up data from ZOE-049 LTFU.⁶⁷ Consequently, annual waning rates for RZV efficacy were estimated to be 1.5% (95% CI, 0.0–3.4%) and 2.3% (95% CI, 0.3–4.4%) for individuals 50–69 YOA and ≥ 70 YOA, respectively.⁴⁵

For ZVL, initial vaccine efficacy against HZ was derived from clinical data from the Shingles Prevention Study,³⁵ the Short-Term Persistence Substudy,³⁶ and the Long-Term Persistence Substudy.³⁷ Efficacy estimates were 69.8% (95% CI, 54.1–80.6%) for 50–59 YOA, 63.9% (95% CI, 56.0–71.0%) for 60–64 and 65–69 YOA, 40.9% (95% CI, 28.0–52.0%) for 70–79 YOA, and 18.3% (95% CI, 0.0–48.0%) for ≥ 80 YOA. These data were modeled to obtain ZVL efficacy waning rate estimates of 5.4% for the first four years of vaccination and 5.1% thereafter, for all age cohorts.

A 30% vaccination coverage was assumed and applied for vaccination with both RZV and ZVL under the base case government-subsidized mass vaccination setting. As there were no HZ vaccines recommended in the Subsidised Vaccine List or National Immunization Schedule in Singapore, there were no appropriate local vaccine coverage data to reference. The coverage applied was hence estimated based on the 2020 national vaccination target (30%) set and achieved for HZ in the USA.^55,56

RZV is a two-dose vaccine and its second dose is recommended for administration 2–6 months following the first dose.³¹ In this study, the second-dose compliance rate was assumed to be 82.5%, which represented an average of compliance in the real-world setting in the USA (approximately 70%)⁵⁷ and in pivotal clinical trials (approximately 95%).^33,34

Outcomes

The primary health outcomes of interest were the number of cases of HZ, PHN, other HZ-related complications (including ocular, neurological, cutaneous, and other non-pain), and HZ-related deaths. The secondary health outcome of interest was hospital resource utilization, measured by the number of cases of HZ-related hospitalizations. Outcomes of interest were explored under the mass vaccination setting (30% coverage).

The potential PHI of HZ vaccination was analyzed and presented for the base case population, with stratifications made for age (five age cohorts: 50–59, 60–64, 65–69, 70–79, and ≥ 80 YOA). Results were accumulated over the entire time horizon under each strategy of no vaccination, vaccination with RZV, or vaccination with ZVL. The incremental value per outcome was calculated to report the impact of vaccination with RZV or ZVL versus no vaccination.

The number needed to vaccinate (NNV) to avoid one case of HZ and PHN, respectively, was also calculated.

Compliance with ethics guidelines

The analyses performed were based on mathematical modeling with inputs extracted from published data or existing literature. No patients or prospective patient-level data were involved in the analyses.

Results

Disease burden of HZ in Singapore

Out of the total base case population of 1.51 million adults ≥ 50 YOA in Singapore, an estimated 406,513 (27.0%) cases of HZ, 68,264 (4.5%) cases of PHN, and 54,949 (3.6%) cases of other complications were projected and could occur in their remaining lifetime in the absence of HZ vaccination (Table 2).

Table 2. Base case results in the mass vaccination setting (30% coverage).

Outcome of interest	No vaccination (control)	RZV vaccination		ZVL vaccination		RZV versus ZVL
	Cases (n)	Cases (n)	Cases avoided (n)	Cases (n)	Cases avoided (n)	Cases avoided (n)
HZ	406,513	333,384	73,129	388,948	17,565	55,564
PHN	68,264	57,170	11,094	65,483	2,781	8,312
Other complications	54,949	45,745	9,205	53,116	1,834	7,371
Ocular	20,103	16,756	3,347	19,430	672	2,675
Neurological	18,303	15,186	3,117	17,696	607	2,510
Cutaneous	8,093	6,762	1,330	7,828	265	1,065
Other non-pain	8,451	7,041	1,410	8,162	289	1,121
Hospitalizations	17,762	14,935	2,827	17,278	484	2,343
HZ-related deaths	72	64	8	71	1	8

Data were rounded to the nearest whole and slight discrepancies may be present. The base case referred to the 2021 population of Singapore adults ≥ 50 YOA, comprising a total of 1.51 million adults, modeled from the year of vaccination over their remaining lifetime. With 30% coverage in the mass vaccination setting, the total vaccinated cohort of adults ≥ 50 YOA was 451,742. HZ: herpes zoster; PHN: post-herpetic neuralgia; RZV: recombinant zoster vaccine; YOA: years of age; ZVL: zoster vaccine live.

Base case results in the mass vaccination setting (30% coverage)

With 30% coverage in the mass vaccination setting, the total vaccinated cohort of adults ≥ 50 YOA was 451,742. Mass vaccination with RZV (versus no vaccination) was estimated to avoid 73,129 cases of HZ, 11,094 cases of PHN, and 9,205 cases of other complications (Table 2; Figure 1) in the base case population. These results showed that vaccination with RZV avoided 4.2, 4.0, and 5.0 times the number of cases of HZ, PHN, and other complications, respectively, compared with ZVL, which avoided 17,565 cases of HZ, 2,781 cases of PHN, and 1,834 cases of other complications.

Figure 1. Number of cases of (a) HZ, (b) PHN, (c) other complications avoided in the mass vaccination setting (30% coverage), by age cohort.

HZ: herpes zoster; PHN: post-herpetic neuralgia; RZV: recombinant zoster vaccine; ZVL: zoster vaccine live.

Compared with no vaccination, mass vaccination with RZV would also prevent 2,827 hospitalizations and 8 HZ-related deaths, while mass vaccination with ZVL would prevent 484 hospitalizations and 1 HZ-related death (Table 2).

In the overall cohort of adults ≥ 50 YOA, the NNV with RZV to prevent one case of HZ and one case of PHN was 7 and 41, respectively; and with ZVL, 26 and 163, respectively (Figure 2). Comparatively, the NNV with RZV to prevent one case of HZ and one case of PHN, respectively, was 3.7 and 4.0 times lower than with ZVL.

Figure 2. NNV to prevent one case of (a) HZ, (b) PHN, by age cohort.

HZ: herpes zoster; PHN: post-herpetic neuralgia; NNV: number needed to vaccinate; RZV: recombinant zoster vaccine; ZVL: zoster vaccine live.

Base case results in the mass vaccination setting (30% coverage), by age cohort

Among the five age cohorts, mass RZV vaccination (versus no vaccination) reduced the largest number of cases in the youngest cohort of 50–59 YOA, which contributed toward 47.0%, 39.3%, and 41.1% of the total number of cases of HZ, PHN, and other complications avoided, respectively, in the gross cohort of adults ≥ 50 YOA (Figure 1). The largest number of hospitalizations avoided with RZV vaccination was also observed for the 50–59 YOA cohort (Table 3). In general, the number of cases avoided gradually decreased with age in the older age cohorts (Figure 1; Table 3).

Table 3. Base case results for hospitalizations in the mass vaccination setting (30% coverage), by age cohort.

Outcome of interest	Age (years)	No vaccination (control)	RZV vaccination		ZVL vaccination		RZV versus ZVL
		Cases (n)	Cases (n)	Cases avoided (n)	Cases (n)	Cases avoided (n)	Cases avoided (n)
Hospitalizations	50–59	7,793	6,751	1,042	7,664	130	913
	60–64	3,462	2,890	572	3,358	104	468
	65–69	2,742	2,241	502	2,607	135	366
	70–79	2,889	2,363	526	2,790	100	426
	≥80	875	690	185	860	15	169
	Total ≥50	17,762	14,935	2,827	17,278	484	2,343

Data were rounded to the nearest whole and slight discrepancies may be present. The base case referred to the 2021 population of Singapore adults ≥ 50 YOA, comprising a total of 1.51 million adults, modeled from the year of vaccination over their remaining lifetime. With 30% coverage in the mass vaccination setting, the total vaccinated cohort of adults ≥ 50 YOA was 451,742. RZV: recombinant zoster vaccine; YOA: years of age; ZVL: zoster vaccine live.

Similarly, mass ZVL vaccination (versus no vaccination) also reduced the largest number of cases of HZ and other complications in the youngest age group of adults 50–59 YOA and the corresponding number of cases avoided gradually decreased with age (Figure 1a,c). The number of cases of PHN and hospitalizations avoided with ZVL demonstrated no apparent trend with age, with the 70–79 and 65–69 YOA cohorts showing the greatest number of cases of PHN and hospitalizations, respectively (Figure 1b; Table 3).

With RZV vaccination, the NNV to prevent one case of HZ was the lowest for the youngest age cohorts of 50–59 and 60–64 YOA, and increased with age up to the ≥ 80 YOA cohort, from 6 to 15 (Figure 2a). The NNV to prevent one case of PHN initially decreased from 41 in the 50–59 YOA cohort to 35 in the 60–64 YOA cohort, before increasing to 55 in the ≥ 80 YOA cohort (Figure 2b).

With ZVL vaccination, the NNV to prevent one case of HZ was the lowest in the 60–64 and 65–69 YOA cohorts, and increased substantially with age up to the ≥ 80 YOA cohort, from 25 to 170 (Figure 2a). The NNV to prevent one case of PHN was the lowest in the 70–79 YOA cohort (100), with no age-specific trend (Figure 2b).

Scenario analysis

Aggregated cohort of adults ≥ 65 YOA, under the mass vaccination setting (30% coverage)

Within the aggregated cohort of 639,008 adults ≥ 65 YOA, 283,835 (44.4%) cases of HZ, 55,078 (8.6%) cases of PHN, and 43,716 (6.8%) cases of other complications were estimated without HZ vaccination (Table 4).

Table 4. Scenario analysis results in the aggregated cohort of adults ≥ 65 YOA, under the mass vaccination setting (30% coverage).

Outcome of interest	No vaccination (control)	RZV vaccination		ZVL vaccination		RZV versus ZVL
	Cases (n)	Cases (n)	Cases avoided (n)	Cases (n)	Cases avoided (n)	Cases avoided (n)
HZ	283,835	228,168	55,667	269,720	14,115	41,552
PHN	55,078	44,932	10,146	51,400	3,678	6,468
Other complications	43,716	35,421	8,295	41,880	1,835	6,460
Ocular	15,768	12,823	2,945	15,131	637	2,309
Neurological	14,683	11,806	2,877	13,995	688	2,188
Cutaneous	6,580	5,358	1,222	6,335	246	976
Other non-pain	6,684	5,434	1,251	6,420	264	987
Hospitalizations	15,333	12,476	2,857	14,743	589	2,267
HZ-related deaths	74	63	11	73	1	10

Data were rounded to the nearest whole and slight discrepancies may be present. This scenario analysis was conducted in a subgroup of the 2021 population of Singapore adults ≥ 65 YOA, comprising a total of 639,008 adults, modeled from the year of vaccination over their remaining lifetime. With 30% coverage in the mass vaccination setting, the vaccinated cohort of adults ≥ 65 YOA was 191,702. HZ: herpes zoster; PHN: post-herpetic neuralgia; RZV: recombinant zoster vaccine; YOA: years of age; ZVL: zoster vaccine live.

With 30% coverage in the mass vaccination setting, the total vaccinated cohort of adults ≥ 65 YOA was 191,702. Mass vaccination with RZV would avoid 55,667 cases of HZ, 10,146 cases of PHN, and 8,295 cases of other complications (Table 4), which was 3.9, 2.8, and 4.5 times the number of cases avoided with ZVL, respectively. Mass vaccination with ZVL was estimated to prevent 14,115 cases of HZ, 3,678 cases of PHN, and 1,835 cases of other complications.

Mass vaccination with RZV would also avoid 2,857 hospitalizations and 11 HZ-related deaths, and mass vaccination with ZVL would avoid 589 hospitalizations and 1 HZ-related death (Table 4). The number of hospitalizations that would be avoided with mass vaccination (RZV or ZVL) is also presented by age cohort in Table 5.

Table 5. Scenario analysis results for hospitalizations in the aggregated cohort of adults ≥ 65 YOA, under the mass vaccination setting (30% coverage), by age cohort.

Outcome of interest	Age (years)	No vaccination (control)	RZV vaccination		ZVL vaccination		RZV versus ZVL
		Cases (n)	Cases (n)	Cases avoided (n)	Cases (n)	Cases avoided (n)	Cases avoided (n)
Hospitalizations	65–69	6,462	5,280	1,182	6,143	319	864
	70–79	6,809	5,569	1,239	6,574	235	1,005
	≥80	2,062	1,627	435	2,026	36	399
	Total ≥65	15,333	12,476	2,857	14,743	589	2,267

Data were rounded to the nearest whole and slight discrepancies may be present. This scenario analysis was conducted in a subgroup of the 2021 population of Singapore adults ≥ 65 YOA, comprising a total of 639,008 adults, modeled from the year of vaccination over their remaining lifetime. With 30% coverage in the mass vaccination setting, the vaccinated cohort of adults ≥ 65 YOA was 191,702. RZV: recombinant zoster vaccine; YOA: years of age; ZVL: zoster vaccine live.

Deterministic sensitivity analysis (DSA)

Results were robust under DSA (Figure 3). The parameters which were shown to have the greatest sensitivity on the number of HZ cases avoided with RZV included (in sequence): (1) annual waning of vaccine efficacy against HZ (post-second dose) for individuals 50–69 YOA, (2) annual incidence of initial HZ, (3) annual waning of vaccine efficacy against HZ (post-second dose) for individuals ≥ 70 YOA, (4) second-dose compliance, and (5) initial vaccine efficacy against HZ (two doses).

Figure 3. DSA on cases of HZ avoided with RZV vaccination versus no vaccination, in the mass vaccination setting (30% coverage).

Note: Number of HZ cases avoided with RZV vaccination (versus no vaccination) in the base case: 73,129. DSA: deterministic sensitivity analysis; HZ: herpes zoster; RZV: recombinant zoster vaccine; YOA: years of age.

Discussion

RZV has been registered and available in Singapore since 2021,^39,70 and this is the first study evaluating the PHI of the HZ vaccines currently available in Singapore (RZV and ZVL) compared with no vaccination in adults ≥ 50 YOA. The PHI of HZ vaccination in Singapore has previously only been partially explored for ZVL as public health outcomes within a cost-effectiveness analysis.⁷¹ The present study used Singapore-specific model inputs, where available, and predicted a high public health burden of HZ in Singapore which can be reduced with HZ vaccination.

As explored in the base case analysis, an approximate 27.0%, 4.5%, and 3.6% of 1.51 million adults ≥ 50 YOA in Singapore in 2021 could develop HZ, PHN, and other complications, respectively, and 1.2% could be hospitalized due to HZ in their remaining lifetime without HZ vaccination. These data are consistent with existing evidence of an estimated 30% lifetime risk of HZ both globally and in the Asia-Pacific region,^3,60 and suggest considerable public health burden of HZ in Singapore in the absence of vaccination. Similar findings were further reported in a recent population-based study highlighting the clinical burden and healthcare resource utilization of HZ, as well as the need for HZ vaccination among older adults locally.¹⁴

Vaccination can be an effective public health strategy for preventing HZ and its complications, and reducing their public health burden. Compared with no vaccination in the base case analysis, the number of cases of HZ, PHN, other complications, HZ-related deaths, and HZ-related hospitalizations was reduced with mass HZ vaccination in the overall ≥ 50 YOA cohort. Notably, vaccination with RZV and ZVL was projected to avoid 2,827 and 484 HZ-related hospitalizations, respectively.

As in Singapore, the high clinical burden of HZ is similarly documented worldwide.¹² For example, the USA set HZ vaccination as one of their public health priorities in 2020,^55,56 while other countries have gradually incorporated HZ vaccination into their respective National Immunization Programs (NIPs) targeting older adults (e.g., RZV in Australia,⁷² Germany,^73,74 the United Kingdom,⁷⁵ and New Zealand^76,77). In Germany, the decision to provide HZ vaccination via the NIP considered the increased risk of severe HZ and PHN in older adults, high RZV efficacy, and the anticipated period of protection provided by RZV⁷⁴; in Australia, listing RZV under the NIP considered its high vaccine efficacy demonstrated in clinical trials, and availability of high quality, long-term data in older adults.⁷²

Vaccination with RZV demonstrated a greater impact on public health outcomes among individuals ≥ 50 YOA versus vaccination with ZVL, where RZV avoided 4–5 times the number of cases of HZ, PHN, or other complications (Table 2). In particular, mass vaccination with RZV was estimated to avoid 73,129 HZ cases that would otherwise occur in the absence of vaccination, with an overall NNV of 7 to prevent one case. Even when the upper bound waning rate was explored in the DSA, a reduction of 56,238 cases of HZ was predicted with RZV versus no vaccination; this was more than three times the number of cases avoided with ZVL versus no vaccination. Vaccination with RZV was also more efficient than ZVL in that the NNV to prevent one case of HZ and PHN was 3–4 times lower for RZV (HZ: 7; PHN: 41) than ZVL (HZ: 26; PHN: 163) (Figure 2). This is due to the higher initial vaccine efficacy and sustained protection conferred by RZV versus ZVL over the projected period. These NNV results are comparable to those from PHI analyses of RZV in Beijing, and of RZV and ZVL in Hong Kong.^47,48 NNV was not previously reported in the published cost-effectiveness analysis of ZVL in Singapore.⁷¹

The target age group for HZ vaccination may also be important to consider. Early vaccination against HZ (i.e., from 50 YOA) was estimated to have the most substantial PHI. The base case results reported that the greatest percentage of cases avoided was seen in the youngest cohort (50–59 YOA) compared with the older cohorts for both RZV (outcomes of HZ, PHN, and other complications) and ZVL vaccination (outcomes of HZ and other complications only) (Figure 1). When vaccinated with RZV, the 50–59 YOA cohort accounted for almost half (47.0%) of the total number of cases of HZ avoided, and approximately two-fifths of the total number of cases of PHN (39.3%) and other complications (41.1%) avoided. These results are reasonable considering the 50–59 YOA vaccinated cohort is overall the largest and is expected to have a greater number of life years remaining to benefit from an extended duration of protection through RZV. With ZVL vaccination, the 70–79 YOA cohort reported the largest number of PHN cases avoided (Figure 1b), which may indicate its higher protection against PHN in individuals ≥ 70 YOA than younger individuals. Overall, these findings suggest that early HZ vaccination would result in the most substantial PHI.

Early vaccination against HZ is also supported by the NNV results (Figure 2). With RZV vaccination, the NNV to prevent one case of HZ was the smallest in the 50–59 and 60–64 YOA cohorts, and the NNV to prevent one case of PHN was the smallest in the 60–64 YOA cohort, compared with other age groups. The NNV with RZV to prevent HZ increased with age, from 6 in the 50–59 YOA cohort to 15 in the ≥ 80 YOA cohort, while the NNV to prevent PHN decreased from 41 in the 50–59 YOA cohort to 35 in the 60–64 YOA cohort and then increased to 55 in the ≥ 80 YOA cohort; these trends in the NNV with RZV in Singapore were similar to those reported for Hong Kong.⁴⁷ For HZ, the increase in NNV with age would mainly be due to the combined effect of lower initial vaccine efficacy, greater waning of vaccine efficacy, and shorter follow-up period in older adults. On the other hand, the drop in NNV for PHN in the 60–64 YOA cohort was due to the greater burden of PHN at 60 YOA compared to 50 YOA, while its subsequent increase with age was due to increased vaccine efficacy against PHN at 60 YOA compared to 70 YOA.

Given the retirement age of 63 and reemployment age of 68 in Singapore,^78,79 both set to further increase by 2030,⁷⁹ most local adults remain active contributors to the workforce in their 50s and 60s.^15,80 Considering the consequent aging workforce,⁸⁰ avoiding HZ and PHN in these individuals would hence beneficially reduce productivity loss to society. Furthermore, the long life expectancy of residents and the resultant aging population in Singapore⁸¹ raise concerns regarding the growing prevalence of age-related diseases or disabilities and its increasing pressure on the healthcare system. To circumvent this, the Healthier SG Initiative was enacted as a national action plan for healthcare transformation to focus on preventive care, encourage healthy aging, and slow the rate of increase of disease burden and healthcare expenditure.³⁰ To address these objectives of Healthier SG, early HZ vaccination could alleviate the disease burden of HZ, accrue its health benefits over a longer period of time, and free up healthcare expenditure and resources for the provision of medical services for other diseases or conditions.

The results of the base case analysis were based on a vaccination coverage rate for RZV of 30% and a second-dose compliance rate of 82.5%. In general, the vaccination coverage and second-dose compliance of HZ vaccination in Singapore is expected to be higher than observed elsewhere (e.g., >30% and > 70%, respectively, in the USA)^55–57 given the ease of accessibility to vaccines via healthcare providers and overall adherence to medical advice in Asian communities.^82,83 Additionally, published PHI models of RZV, in majority Chinese populations of Hong Kong and Beijing, also used similar second-dose compliance rates (80.0–82.5%).^47,48

Due to limited local epidemiological data, there was a need to apply inputs from neighboring locales or countries, which the authors understand is common practice and generally accepted by payers and prescribers locally. As described in the Methods, this study utilized updated demographic data of Singapore and locally relevant epidemiological data, based on a systematic selection of model inputs that included the conduct of a literature search and evaluation of the selected data source for each parameter. Notably, the Taiwan population-level database was mainly used as an alternative to local data sources and may be considered suitable based on shared population ethnicities (i.e., Chinese majority) with Singapore.

For vaccine-related parameters, model inputs of vaccine efficacy were based on data sources,^33,34,67 which align with the studies identified in a recent systematic review and meta-analysis assessing HZ vaccine efficacy in immunocompetent individuals.⁸⁴

Certain comorbidities (e.g., diabetes, chronic kidney disease, and cardiovascular diseases) may increase risk of HZ,^85–91 and hence its complications. In Singapore, rates of comorbidities are high among older adults, with more than 50% of the population developing at least one chronic disease by 50 YOA and a second or more comorbidity (i.e., multimorbidity) by 60 YOA.⁹² Among patients ≥ 50 YOA seeking medical attention for HZ or PHN in Singapore (N = 347), comorbidities were common and observed in up to 51.3% of patients and a large proportion of all patients were 50–60 YOA.¹⁷ The base case population in this study was adults ≥ 50 YOA which, when considering the prominent overlap of comorbidities and old age, would expectedly include those with comorbidities in addition to those without; the PHI of HZ vaccination was analyzed and reported in the overall base case population, and not specifically for the comorbidities subgroup. Future studies may hence explore the effects of HZ vaccination in a subset of patients with specific comorbidities. In particular, diabetes could be of interest as it raises the risk of HZ by 24–38%,^93,94 and its growing prevalence rate in Singapore is one of the fastest globally.⁹⁵ Such analyses may supplement the present PHI study regarding HZ vaccination and account for these specific patient groups in Singapore. A comprehensive value assessment of HZ vaccination and its unmet need locally would require understanding the PHI (described in this study) as well as the cost-effectiveness of HZ vaccination for older adults in Singapore. Since there are currently no available local cost-effectiveness data on RZV based on an economic SLR, a cost-effectiveness analysis of RZV may therefore be explored in the future to complement results from the present study and support discussions regarding HZ vaccination strategies for older adults in Singapore.

In conclusion, this analysis demonstrated that mass HZ vaccination with RZV among adults ≥ 50 YOA in Singapore has a superior PHI, compared to no vaccination. These results may be helpful in supporting value assessment and decision-making regarding public health vaccination strategies for HZ prevention in Singapore.

Authors’ contributions

Substantial contributions to study conception and design: HO, CT, CW, NG, and CN; substantial contributions to analysis and interpretation of the data: HO, CT, CW, NG, and CN; drafting the article or revising it critically for important intellectual content: HO, CT, CW, NG, and CN; final approval of the version of the article to be published: HO, CT, CW, NG, and CN.

Data sharing statement

All data generated or analyzed during this study are included in this published article or as supplementary information files.

Prior presentation

This manuscript is based on work that has been previously presented at the 7^th World One Health Congress, Singapore (November 7–11, 2022).

Acknowledgments

The authors acknowledge Desmond Curran, GSK, and Sumitra Shantakumar, GSK, for contributions to study concept and data interpretation, and Roeland Van Kerckhoven, GSK, for publication management. The authors also thank Costello Medical for editorial assistance and publication coordination, on behalf of GSK, and acknowledge Ren Ping Yong, Costello Medical, Singapore for medical writing and editorial assistance based on authors’ input and direction.

Disclosure statement

HO and CT: Nothing to disclose; CW: Employed by the GSK group of companies; NG and CN: Employed by and hold shares in the GSK group of companies.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2024.2348839

Additional information

Funding

This study was sponsored by GSK (Study identifier eTrack VEO-000294). Support for third-party writing assistance for this article, provided by Ren Ping Yong, Costello Medical, Singapore, was funded by GSK in accordance with Good Publication Practice (GPP) 2022 guidelines (https://www.ismpp.org/gpp-2022).