Routine varicella vaccination program and hospitalization for herpes zoster in Japan

ABSTRACT

Whether reducing exposure to varicella by the implementation of the routine varicella vaccination program for children leads to increased incidence of herpes zoster (HZ) remains controversial. The aim of the present study was to identify the trend in the hospitalization associated with HZ before and after the introduction of routine varicella vaccination by using nationally representative data from an inpatient database in Japan. Data were obtained on the number of inpatients hospitalized for HZ from the “Survey on the effect of the introduction of Diagnosis Procedure Combination (DPC) database” and the total population in Japan from the Population Estimates created by the former Statistics Bureau between fiscal years 2013 and 2018. The data from the DPC hospitals only and all hospitals in the survey were analyzed separately. The trends in the annual incidence of HZ hospitalization were identified. The trends in the annual hospitalization for HZ per 100,000 persons were then analyzed by age group (0–20, 21–40, 41–60, 61–79, and ≥80 years of age). The annual number of hospitalizations for HZ was approximately 20,000 in the DPC hospitals and 25,000 in all hospitals, showing no upward trend. The age-specific annual hospitalization rate for HZ did not increase in all the age groups. As age increased, the hospitalization rate also increased. This study presents no upward trend in the hospitalizations for HZ after the implementation of the routine varicella vaccination program in Japan.

KEYWORDS:

• Diagnosis procedure combination
• epidemiology
• exogenous boosting
• varicella

Introduction

The varicella-zoster virus (VZV) causes varicella (chickenpox) at the primary infection, establishing a lifelong latency in the sensory ganglia during varicella. Herpes zoster (HZ) results from the reactivation of the VZV,¹ which is characterized by a painful vesicular rash occurring unilaterally and most commonly restricted in one or two adjacent dermatomes. In rare cases, the rash can be more widespread and affect three or more dermatomes. Although most cases of HZ are mild and self-limiting, complications such as encephalitis, disseminated zoster, or Ramsay Hunt syndrome may occur in some cases. The annual global incidence and hospitalization rates of HZ range from 3 to 5 per 1,000 persons and from 2 to 25 per 100,000 persons, respectively.² One of the most serious complications of HZ is postherpetic neuralgia (PHN).³ PHN is neuropathic pain due to peripheral nerve damage, which causes burning pain and sensitivity to light touch that last ≥3 months after the rash has healed. Although common, HZ remains an important health issue as it can cause serious complications.

In Japan, the varicella vaccine was approved in 1986, but because the vaccine is available for voluntary vaccination, which was not recommended at that time, the vaccination rate was only approximately 50% before the routine vaccination program was started.⁴ Although the recommended number of vaccinations remains uncertain, the Japan Pediatric Society recommended two doses of the varicella vaccine as a voluntary vaccination in April 2011. Thereafter, the routine varicella vaccination program for children aged 12 to 36 months was started in October 2014, in which two doses of the vaccine are administered with at least a 3-month interval (6 to 12 months as a standard interval) between the first and second vaccinations.⁵ After the start of the routine vaccination, the number of reported varicella cases dramatically decreased by >70%.⁶

According to the proposed Hope–Simpson’s exogeneous boosting hypothesis, VZV-specific immunity is boosted through repeated re-exposure to circulating VZV, thereby inhibiting VZV reactivation.^1,7 If this hypothesis is correct, the introduction of the routine varicella vaccine to children may suppress the circulation of VZV, reduce the chance of re-exposure, and increase the incidence of HZ. However, no conclusion has been reached about this. A study that assessed for VZV-specific T-cell immunity and antibody titers in grandparents exposed to their grandchildren with varicella reported that exogeneous boosting only occurred in 17%–25% of them for <1 year.⁸ Furthermore, whether the introduction of the routine varicella vaccination program led to an increase in the incidence of HZ at the population level by reducing re-exposure to varicella remains unclear.^8–10 The aim of the present study was to identify the trend in hospitalizations associated with HZ before and after the introduction of the childhood routine varicella vaccination by using a nationally representative inpatient data in Japan.

Materials and methods

Study design

A repeated cross-sectional study was performed using published data of the “Survey on the effect of the introduction of Diagnosis Procedure Combination (DPC) database,” which was based on the Japanese DPC database. The “Survey on the effect of the introduction of DPC database” does not include individual patient data. Informed consent was waived owing to the anonymous nature of the data.

Data source

“Survey on the effect of the introduction of DPC database” was based on the Japanese DPC database. The DPC is a case-mixed, patient classification system linked to a lump-sum payment system for inpatients in acute care hospitals.¹¹ DPC hospital is a type of hospital that has participated in the DPC system. The term “all hospitals” referred to DPC hospitals plus hospitals preparing for adoption. The DPC hospital annually collects and reports on administrative claim data and detailed information obtained during patient discharge. This system was implemented in 2002. The DPC-related data from all participating institutions are gathered by The Ministry of Health, Labor and Welfare (MHLW) in Japan, covering 550,000 hospital beds and approximately 11 million discharged cases from 3,000 acute care hospitals annually.¹² In Japan, the coverage rate of acute inpatients registered in the DPC database was >50%.¹¹ Although the utilization of the DPC-related data is limited, the accumulated summaries of the data are published as “Survey on the effect of the introduction of DPC database,” which was published online by the MHLW. At the start of the study, we had data available through fiscal year (FY) 2018.

Participants and data extraction

The number of inpatient cases of HZ (DPC code: 080020xxxxxxxx) was extracted from the “Survey on the effect of the introduction of DPC database” between the FYs 2013 and 2018, which were downloaded from Excel files from the website (https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/0000049343.html).^13–18 This DPC code includes the following International Classification of Disease, Tenth Revision, Clinical Modification (ICD-10-CM) codes as the most resource-consuming diagnosis: B022, B023, B027, B028, and B029 (Table 1).

Table 1. The international classification of disease, tenth revision, clinical modification codes included the study

B02.2	Zoster with other nervous system involvement
	B02.21 Postherpetic geniculate ganglionitis
	B02.22 Postherpetic trigeminal neuralgia
	B02.23 Postherpetic polyneuropathy
	B02.24 Postherpetic myelitis
	B02.29 Other postherpetic nervous system involvement
B02.3	Zoster ocular disease
	B02.30 Zoster ocular disease, unspecified
	B02.31 Zoster conjunctivitis
	B02.32 Zoster iridocyclitis
	B02.33 Zoster keratitis
	B02.34 Zoster scleritis
	B02.39 Other herpes zoster eye disease
B02.7	Disseminated zoster
B02.8	Zoster with other complications
B02.9	Zoster without complications

The total population data in Japan were obtained between FYs 2013 and 2018 from the Population Estimates created by the former Statistics Bureau, which were also downloaded the Excel files from the website (https://www.stat.go.jp/data/jinsui/index.html).^19–24

Statistical analyses

All statistical analyses were performed with the R version 3.6.1 software. We computed descriptive statistics for the main items. The data from the DPC hospitals only and all hospitals included in the survey were analyzed separately. The trends in the annual incidence of HZ hospitalization were identified. A breakdown of the five age groups (0–20, 21–40, 41–60, 61–79, and ≥80 years of age) was also shown. The trends in the annual HZ hospitalization rates per 100,000 persons were then analyzed by age group. Furthermore, the ICD-10-CM codes were broken down by FY.

Results

Table 2 summarizes the main characteristics of the study. The proportion of females was slightly higher than that of males. Approximately 20,000 and 25,000 inpatients hospitalized for HZ per year were counted from the data collected from DPC hospitals only and all hospitals in the survey, respectively. Figure 1 illustrated the numbers of hospitalizations for HZ by FY. No upward trend was found in the number of hospitalizations for HZ from the DPC hospital-only data (Figure 1a).

Table 2. Main characteristics of the study

	Survey year
	2013	2014	2015	2016	2017	2018
Number of hospitals, n
DPC hospital only	1,497	1,584	1,580	1,666	1,663	1,730
All hospitals	1,804	2,942	3,191	3,501	3,701	4,764
Number of hospitalizations, n
DPC hospital only	20,591	20,240	19,847	19,627	19,601	19,652
All hospitals	21,723	22,291	23,395	23,045	23,222	23,252
Gender, female, %
DPC hospital only	55	55	54	55	55	54
All hospitals	55	56	55	56	56	56
Hospitalization rate per 100,000 persons, (95% CI)
DPC hospital only	16.2 (16.0, 16.4)	15.9 (15.7, 16.1)	15.6 (15.4, 15.8)	15.5 (15.2, 15.7)	15.5 (15.3, 15.7)	15.5 (15.3, 15.8)
All hospitals	17.1 (16.8, 17.3)	17.5 (17.3, 17.8)	18.4 (18.2, 18.6)	18.2 (17.9, 18.4)	18.3 (18.1, 18.6)	18.4 (18.2, 18.6)

CI, confidence interval; DPC, diagnosis procedure combination.

Figure 1. The number of hospitalizations for herpes zoster by age group from 2013 to 2018.

(a): DPC hospital-only data. (b): All-hospital data

The annual hospitalization per 100,000 persons in Japan was 17.8 based on the 6-year all-hospital data. Hospitalization rates per 100,000 persons in DPC hospitals alone before (2013–2014) and after (2015–2018) the introduction of routine vaccination program were 16.1 (95% confidence interval [CI], 15.9–16.2) and 15.5 (95% CI, 15.4–15.6), respectively. The hospitalizations per 100,000 by age group was 2.7 for the 0- to 20-year age group, 5.2 for the 21- to 40-year age group, 10.0 for the 41- to 60-year age group, 35.3 for the 61- to 79-year age group, and 63.2 for the ≥80-year age group, based on the 6-year all-hospital data. The trends of the annual hospitalization rates per 100,000 persons by age group are shown in Figure 2. The elderly had a higher annual rate of hospitalization for HZ. The percentage of people aged >61 years according to the all-hospital data was 75.8% (95% CI 75.6%–76.0%). No upward trend was observed in the annual HZ hospitalization rates for all the age groups based on the DPC hospital-only data (Figure 2a).

Figure 2. The rates of hospitalization for herpes zoster per 100,000 persons by age group from 2013 to 2018.

(a): DPC hospital-only data. (b): All-hospital data

The breakdown of the DPC codes is shown in Figure 3. The proportion of the B028 code (Zoster with other complications) has a slight upward trend, whereas that of the B029 code (Zoster without other complications) has a slight downward trend.

Figure 3. The proportions of the international classification of disease, tenth revision, clinical modification codes from 2013 to 2018.

(a): DPC hospital-only data. (b): All-hospital data

Discussion

This study suggests no upward trend in the hospitalizations associated with HZ after the implementation of the routine varicella vaccination program in Japan. The age-specific annual rate of hospitalization for HZ also did not increase in all age groups.

The routine varicella vaccination program had little effect on the hospitalizations for HZ. Although the incidence of childhood varicella decreased drastically in the FY after the start of the routine varicella vaccination,⁶ it was confirmed that the rate of hospitalization for HZ was unchanged. However, whether reducing exposure to varicella by the implementation of the routine varicella vaccination program leads to increased incidence of HZ remains controversial.^8–10 The findings of this study are consistent with those from previous epidemiological studies from the United States,^25–27 Canada,²⁸ and Australia²⁹ that suggested that the vaccination program had little impact on the increase in the incidence of HZ at the population level. On the contrary, the results of other studies supported the exogeneous boosting hypothesis.^30–32 Although many of the studies that examined at the individual level supported this hypothesis,¹⁰ the impact of exogeneous boosting might not be as large as previously assumed;⁷ therefore, it is possible that the impact of exogeneous boosting could not be confirmed by studies examined at the population level using real-world data.

The influence of the routine varicella vaccination program for children may differ between the community incidence of HZ not requiring hospitalization and that requiring hospitalization. The effectiveness of varicella vaccine has been suggested to vary between both groups.³³ A recent study from Japan reported an upward trend in the incidence of HZ between 2010 and 2017,³⁴ contrary to the present results. Studies from Australia^29,35 found that the number of HZ cases not requiring hospitalization increased, but the number of HZ hospitalizations did not increase. Accordingly, the decline in exposure to varicella may lead to an increase in the community incidence of HZ but may not change the incidence of HZ requiring hospitalization. HZ immunology and the mechanism of exogeneous boosting are complicated and not yet well understood. Further research is needed on the differences in the effects on the incidence rates of community HZ and HZ hospitalization.

The elderly accounted for a higher rate of hospitalization for HZ. The hospitalization rates for HZ in this study were consistent with those reported in previous studies.² Approximately 75% of the patients hospitalized for HZ were aged >61 years in this study. Another study from Japan reported that the number of HZ cases in the ≥60-year age group had increased from 1997 to 2006.³⁶ The high prevalence of HZ in the elderly could be due to the decreased cell-mediated immunity associated with puberty or high prevalence of underlying disease that may cause immunocompromise. In Japan, the use of a live attenuated HZ vaccine was approved in March 2016 for adults aged ≥50 years as an extended dose for the prevention of HZ. In January 2020, the adjuvant recombinant zoster vaccine was approved for the prevention of HZ. However, the HZ vaccination coverage is not yet expected to be widespread in Japan. HZ vaccination has been demonstrated to be effective for preventing hospitalizations for HZ in the United States³³ and Canada,³⁷ where routine varicella vaccination has been practiced. Widespread HZ vaccination for adults ≥50 years old is required for the prevention of HZ. As the HZ vaccination rates increased for adults, the HZ hospitalization rates in adults aged ≥50 years could decrease regardless of the effect of reduced exposure to varicella. If so, evaluating the exogenous boosting hypothesis could be difficult.

From this study, the reason for the increase in the percentage of “Zoster with other complications” and decrease in the proportion of the “Zoster without other complications” could not be determined. To the best of our knowledge, no study has evaluated the impact of routine varicella vaccination on the breakdown of HZ-associated hospitalizations. Hence, further research is needed.

The strength of this study was its use of nationwide inpatient data, which were published by the MHLW in Japan. The present study also had several limitations. First, certain medical information such as immunocompromise, diagnostic method, blood laboratory data, HZ vaccine uptake, and complications could not be reviewed. Second, because the study was conducted up to 4 years after the start of the routine varicella vaccination program, the study period may be too short to assess the influence of the routine varicella vaccination. The duration and magnitude of exogeneous boosting through re-exposure to varicella have been unclear. A study from the UK suggested a possible 33% reduction in the incidence of HZ at least for 2 years after the exposure of a child.³⁰ Therefore, follow-up research is needed to determine the long-term effects of varicella vaccination on hospitalization for HZ. Third, although the data used were based on large populations and one of the largest databases available in Japan, they do not cover all hospitals. Accordingly, the annual rate of hospitalization for HZ in this study is likely to be a small estimate.

In conclusion, this study presents no upward trend in the incidence of hospitalization for HZ after the implementation of the routine varicella vaccination program. Further studies are needed to identify the long-term effects of the routine varicella vaccination program in Japan.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors would like to thank Enago (www.enago.jp) for the English language review.