Authors: Jing Chen, Chin-Yen Ho, Yu-Kang Tu, Yi-Chih Lin, Yun Hsia, Ying-Chun Lin, Sumitra Shantakumar
Categories: Zoster, Herpes zoster, epidemiology, autoimmune diseases, immunocompromised, meta-analysis, Asia-Pacific
Source: Human Vaccines & Immunotherapeutics
Authors: Jing Chen, Chin-Yen Ho, Yu-Kang Tu, Yi-Chih Lin, Yun Hsia, Ying-Chun Lin, Sumitra Shantakumar
Asia-Pacific (APAC) faces an increasing burden of herpes zoster (HZ) over time. The risk of HZ and its complications are increased in immunocompromised (IC) patients and those with autoimmune diseases (AID). Our study aimed to synthesize evidence on the epidemiological burden of HZ and its complications among the general adult population and patients with IC/AID conditions in APAC. Following a systematic literature review, we performed meta-analyses for outcomes where ≥3 studies met the inclusion criteria. Of the 271 articles identified, 75 were included for meta-analysis. We found a high burden of HZ and its complications (i.e., postherpetic neuralgia, HZ ophthalmicus), particularly among individuals with IC/AID conditions in APAC. Patients with IC/AID conditions had significantly increased HZ risk and a higher proportion of HZ recurrence than the general adult population. These findings may inform clinical practice and public health decisions regarding HZ prevention, including HZ vaccination strategies, among the IC/AID population in APAC.
Herpes zoster (HZ), also known as shingles, is caused by the reactivation of varicella zoster virus, which could be attributed to age-related decline in immunity or immunosuppression.^1–3^ Approximately 30% of individuals experience HZ during their lifetime.^2,4^ HZ can cause complications which are challenging to treat, the most common and debilitating being postherpetic neuralgia (PHN), which occurs in 5–30% of patients with HZ.^1,3,5^ As such, HZ and its complications may result in reduced quality of life of individuals,^1,3,5^ while posing an economic burden on society and a strain on healthcare resources.^6^
The risks of HZ and HZ-related complications increase with age,^3,7^ and are elevated in patients with certain underlying conditions, particularly those who are immunocompromised (IC) or have autoimmune diseases (AID).^8,9^ Moreover, it has been estimated that over half of patients with HZ have moderate to very severe comorbidities, and comorbidities involving immune dysregulation (e.g., AID, malignancies) are particularly prevalent.^10^ In line with this finding, a recent global meta-analysis found that underlying conditions (particularly transplantation) were risk factors for HZ, with odds ratios ranging 1.17–4.51.^11^ Similarly, a systematic literature review (SLR) reported increased HZ incidence rates (IRs) among IC populations in Europe, particularly those with solid organ and stem cell transplants, cancer, and rheumatoid arthritis (RA).^12^
Within the overall HZ patient population, individuals with IC and AID conditions contribute substantially to the burden on healthcare systems, due to the increased risk of more severe, complicated disease manifestations (e.g., slower healing lesions, secondary infections), HZ-related hospitalization and/or specialized treatment regimens, and recurrent HZ.^13,14^ Given that HZ is a vaccine-preventable disease, proactive HZ vaccination could be effective in reducing disease burden,^15,16^ especially among at-risk patients.^17^
Presently, two types of vaccines are available for HZ prevention globally, including in the Asia-Pacific (APAC) the live attenuated zoster vaccine and the adjuvanted recombinant zoster vaccine (RZV).^4–6^ Unlike live attenuated vaccines which are contraindicated in immunosuppressed or immunodeficient individuals due to the risk of disease dissemination from uncontrolled replication of vaccine strain virus,^18–20^ RZV is non-live and does not carry the same risk in patients with IC and AID conditions.^21^ RZV is currently licensed for HZ prevention in at-risk individuals (including patients with IC and AID conditions) aged ≥18 years in APAC including Australia, Japan, South Korea, and Taiwan.^22–25^
In APAC, HZ incidence has been observed to increase by approximately 5% per year.^7^ Despite the high burden of HZ in APAC, data on the incidence and risk of HZ among patients with IC/AID conditions in this region have not been comprehensively consolidated. The synthesis of HZ epidemiological data in patients with IC/AID conditions in APAC could help support clinical decisions and the implementation of HZ preventative strategies in this population. A meta-analysis was hence conducted to assess the scale of HZ burden in the general adult population and patients with IC/AID conditions in APAC, with a focus on the incidence and recurrence of HZ, and the burden of HZ-related complications (i.e., PHN, HZ ophthalmicus [HZO], and disseminated HZ).
The primary objectives of this study were to estimate HZ IR in the general adult population and the IC/AID population, and the IR ratio (IRR) of HZ in the IC/AID population versus the non-IC/AID population. The secondary objective was to estimate the proportions of HZ patients with PHN and other HZ complications, while the tertiary objective was to estimate the proportion of patients with recurrent HZ.
An SLR with a search period of January 2000–April 2022 consolidated HZ epidemiology and disease burden data in adults aged ≥18 years in selected APAC countries/territories. English articles were identified through MEDLINE and Embase. Reference lists of identified articles were hand-searched to supplement the database searches. The full methodologic details of the SLR (including the search terms and full list of outcomes of interest) have been described in Chen et al. 2024.^4^ For the current meta-analysis, the literature search was extended to five Southeast Asian countries (Indonesia, Malaysia, the Philippines, Thailand, and Vietnam).
Briefly, the inclusion criteria 1) studies published in the English language, 2) studies providing HZ data for adults ≥18 years, 3) studies conducted in Australia, China, Hong Kong, Indonesia, Japan, Korea, Malaysia, New Zealand, the Philippines, Singapore, Taiwan, Thailand, and Vietnam, and 4) observational studies. Case reports, clinical trials, meta-analyses, reviews, and letters to the editor were excluded.
In this study, the general adult population collectively refers to 1) study cohorts which may include individuals with IC/AID conditions if this was not differentiated within the reference article (e.g., single-arm studies), and 2) non-IC/AID comparator groups (e.g., double- or multi-arm studies).
HZ IR was defined as the number of new cases per 1,000 person-years (PY). The proportion of HZ recurrence refers to the number of individuals experiencing a subsequent episode among those with a previous episode of HZ, in studies where both the first and subsequent episodes were observed.
Data extraction was independently undertaken by three reviewers. The following data were extracted from each (1) publication information including first author, year of publication, and title, (2) study characteristics including country/territory or region, study period, study design, age range, sample size, and IC/AID conditions, and (3) reported outcomes of interest and other variables which were necessary to calculate the outcomes of interest. Data related to different IC/AID conditions were treated as separate datasets, even if they were published in the same article/study.
Where the value for a parameter of interest was not explicitly reported, the parameter was estimated using available data on other variables and the following formulae. For outcomes such as IRs, there may be minor discrepancies with values from the original study due to the lack of specific covariate details.IR=numberofnewHZcasesPYofobservationinstudypopulation≈numberofnewHZcasesmeanormediantime∗sizeofstudypopulation
The quality of the included articles was assessed using the Joanna Briggs Institute (JBI) checklist, a critical appraisal tool for observational studies comprising nine items, to detect potential biases and measure the reliability of a study.^26^ For each included study, the number of items meeting the requirements of the checklist were summed up and divided by nine to form a percentage score.
A meta-analysis for an outcome was conducted if there were ≥3 studies that met the inclusion criteria. Results of the independent studies were combined to obtain a weighted average and its 95% confidence interval (CI).
Random effects meta-analysis was performed to account for between-study heterogeneity in the distribution of study weights across the observational studies included. Subgroup analyses on HZ incidence by specific IC/AID conditions, age, and country/territory were performed to mitigate potential heterogeneity.
A meta-regression was performed to test whether sex ratio was associated with HZ IRs in patients with IC/AID conditions and the general adult population. The expected proportions of females and males were assumed to be equal in the general adult population, and the proportion of females was centered at 50%.
All meta-analyses were conducted using the software R version 4.2.1, with two-tailed p < .05 considered statistically significant.
This study is a literature review and meta-analysis, which involves the analysis of previously published data, and does not require ethical review.
Of the 271 articles identified in the SLR, 75 met the inclusion criteria for this meta-analysis study (Figure 1). These articles included 19 studies from Korea, 18 from Japan, 18 from Taiwan, eight from China, six from Australia, two from Hong Kong, two from New Zealand, one from Thailand, and one international study which recruited patients from Taiwan, Korea, and Thailand (Supplementary Table). Figure 1.Flowchart detailing the meta-analysis article selection process. SLR: systematic literature review.
Of the included studies conducted in patients with IC/AID conditions, the reported conditions comprised hematological malignancy, inflammatory bowel disease (IBD), malignant lymphoma, multiple myeloma, non-Hodgkin lymphoma, oral cavity cancer, prostate cancer, psoriasis, RA, solid organ malignancy, systemic lupus erythematosus (SLE), and transplantation (i.e., liver transplantation, stem cell transplantation). Articles which reported IC conditions or AID as an entire class without further exploring specific diseases were also included in the meta-analysis.
Across the 75 articles included in this meta-analysis study, the average number of items meeting the requirements of the JBI critical appraisal checklist was 8.63, and the average score was 95.85%. The most commonly missed items included the study subjects and setting not being described in detail. The JBI checklist scores of each study are reported in the Supplementary Table.
A total of 28 IC/AID datasets from 24 studies were analyzed.^27–50^ The ages of individuals in these studies ranged from 18 to 82 years old, the average proportion of females was 49.07%, and the range of follow-up time was 0.5–8.0 years (195–608,964 PY, median 16,145 PY).
The reported disease/condition-specific HZ IRs ranged from 2.28/1,000 PY in RA to 95.20/1,000 PY in malignant lymphoma. HZ IRs were heterogeneous across patients with different IC/AID conditions as well as within each disease group. The pooled HZ IR was 16.53/1,000 PY (95% CI, 12.15–22.48) (Figure 2(a)). Figure 2.HZ incidence in patients with IC/AID conditions. (a) Forest plot of HZ IRs reported in studies included in the meta-analysis, in ascending order. Articles ^27–50^. (b) Pooled HZ IR by IC/AID condition, in descending order. Articles hematological malignancies,^30,32,36^ SLE,^28,32,48^ solid organ tumor,^32,33,44,50^ liver transplantation,^31,37,47^ RA,^29,32,35,38,40,46,48^ and psoriasis.^42,45,49^ (c) IC/AID condition-specific HZ IRs by age. Articles Chen et al. 2011,^28^ Chang et al. 2018,^27^ Min et al. 2021,^49^ Kao et al. 2021,^50^ and Tanaka et al. 2021.^46^ *Age-stratified HZ incidences rates for psoriasis have been calculated manually based on values reported in Min et al. 2021.^49^ AID: autoimmune diseases; CI: confidence interval; HSCT: hematopoietic stem cell transplantation; HZ: herpes zoster; IBD: inflammatory bowel disease; IC: immunocompromised; IR: incidence rate; PY: person-years; RA: rheumatoid arthritis; SLE: systemic lupus erythematosus; SOM: solid organ malignancies.
Pooled condition-specific HZ incidence was assessed for six IC/AID hematological malignancies,^30,32,36^ liver transplantation,^31,37,47^ psoriasis,^42,45,49^ RA,^29,32,35,38,40,46,48^ SLE,^28,32,48^ and solid organ tumor.^32,33,44,50^ Across the IC/AID conditions assessed, pooled HZ IR ranged from 7.42/1,000 PY (95% CI, 4.85–11.34) in psoriasis to 63.78/1,000 PY (95% CI, 46.09–88.26) in hematological malignancies (Figure 2(b)).
Five studies reported age-specific HZ IRs in patients with IBD, oral cavity cancer, psoriasis, RA, and SLE.^27,28,46,49,50^ In general, HZ IRs in patients with these IC/AID conditions increased with age (Figure 2(c)).
26 datasets from 23 studies reported the sex distribution of patients with IC/AID conditions.^27–47,49,50^ Seven datasets of patients with SLE or RA consisted of more than 70% females.^28,29,32,35,38,40,46^ Two studies on prostate cancer were conducted among male patients.^33,44^ The proportion of females was not significantly associated with HZ IR (p = .16).
A total of 33 studies which reported on HZ IRs in the general adult population were analyzed.^28,34,40–43,49–75^ The ages of individuals in these studies ranged from 18 to 103 years old, and the average proportion of females was 54.4%. The range of follow-up time was 1–10 years.
Within each study, the PY of observation ranged from 20,233 to nearly 40 million, with a median value of 500,758 PY. The reported HZ IRs in the general adult population ranged from 0.75/1,000 PY to 16.69/1,000 PY. The pooled IR was 6.49/1,000 PY (95% CI, 5.30–7.95) (Figure 3(a)). Figure 3. Forest plots of HZ IRs in the general adult population for all studies included in the meta-analysis (a) in ascending order of IR, (b) by age group <50 YOA and ≥50 YOA, and (c) by country/territory. Articles included in all three Australia,^60,61,63,64,68,75^ China,^41,57,58,73^ Japan,^34,40,66,67,69–71^ Korea,^49,51,53–56,74^ New Zealand,^65^ and Taiwan.^28,42,43,50,52,59,62,72^ CI: confidence interval; HZ: herpes zoster; IR: incidence rate; YOA: years of age.
Overall, HZ IRs in the general adult population were heterogeneous across age groups and countries/territories, as well as within each age group and country/territory.
27 datasets from 20 studies were analyzed for age-specific HZ IRs in the general adult population.^41,49–52,54–58,60–62,65,67–69,71–73^ Among adults aged ≥50 years, the reported HZ IRs ranged from 1.48/1,000 PY to 19.46/1,000 PY, and the pooled IR was 8.73/1,000 PY (95% CI, 6.66–11.45) (Figure 3(b)). Among adults aged <50 years, the reported HZ IRs ranged from 0.36/1,000 PY to 8.28/1,000 PY, and the pooled IR was 2.80 (95% CI, 1.43–5.49)/1,000 PY. Figure 3.(Continued).
22 datasets reported sex distribution in the general adult population.^28,34,40–43,49,50,53,55–60,63,64,66,67,69,70,72^ Five of these datasets were derived from comparative studies with matched sex ratios between IC/AID and non-IC/AID groups.^28,38,40,42,50^ Four of the five studies were excluded in the meta-regression analysis, as they had a very high (>75%) or low (<25%) proportion of females due to matching with IC/AID patient groups for sex.^28,38,40,50^ A separate dataset with an extremely low HZ IR (0.75/1,000 PY) was also excluded to avoid influencing the accuracy of the regression model and parameter estimates.^58^
In the meta-regression which included 17 datasets, the proportion of females was significantly associated with higher HZ IRs in the general adult population, with a regression coefficient of 0.05 (95% CI, 0.01–0.09; p = .015).
Across all the included APAC countries/territories, studies in China reported the lowest mean HZ IR in the general adult population (Figure 3(c)). There were two studies with extremely low IRs, including Li et al. 2022 which reported an HZ IR of 0.75/1,000 PY among adults aged ≥20 years in China,^58^ and Kang et al. 2008 which reported an HZ IR of 1.41/1,000 PY among healthy young males aged 19–24 years in Korea.^54^
Nine studies compared HZ occurrence in patients with IC/AID to that in the non-IC/AID population.^28,34,38,40–43,49,50^ These included five studies from Taiwan,^28,38,42,43,50^ two from Japan,^34,40^ and one study each from China^41^ and Korea.^49^ The number of participants in the included studies ranged from 6,262 to 2,778,476 individuals.
The pooled IRR was 1.98 (95% CI, 1.43–2.74; p < .01), indicating that patients with IC/AID conditions were more likely to develop HZ than the non-IC/AID population (Figure 4). Figure 4.Forest plot of IRRs for the IC/AID population versus the non-IC/AID population reported in studies included in the meta-analysis, in ascending order. Articles Chen at al. 2011,^28^ Imafuku et al. 2019,^34^ Kao et al. 2021,^50^ Liao et al. 2017,^38^ Min et al. 2021,^49^ Sakai et al. 2018,^40^ Sun et al. 2021,^41^ Tsai et al. 2015,^43^ and Tsai et al. 2017.^42^ AID: autoimmune diseases; CI: confidence interval; IBD: inflammatory bowel disease; IC: immunocompromised; IRR: incidence rate ratio; RA: rheumatoid arthritis; SLE: systemic lupus erythematosus.
The pooled proportions of PHN and HZO were higher in HZ patients with IC/AID conditions compared with HZ patients in the general adult population (Figure 5(a–d)). Figure 5. Forest plots of the proportion of patients with HZ-related complications reported in studies included in the meta-analysis. (a) Proportion of PHN in HZ patients with IC/AID conditions and (b) the general adult population. (c) Proportion of HZO in HZ patients with IC/AID conditions and (d) the general adult population. (e) Proportion of disseminated HZ in patients with IC/AID conditions. Articles PHN (IC/AID population),^31,34,38,41,76,85–87^ PHN (general adult population),^34,41,53,57,62,68,69,73,86–99^ HZO (IC/AID population),^34,59,85,87^ HZO (general adult population),^34,87,100–104^ and disseminated HZ (IC/AID population).^34,87,105^ AID: autoimmune diseases; CI: confidence interval; HZ: herpes zoster; HZO: herpes zoster ophthalmicus; IC: immunocompromised; PHN: postherpetic neuralgia; RA: rheumatoid arthritis; SLE: systemic lupus erythematosus.
In the IC/AID population, the pooled proportion of patients with disseminated HZ was estimated at 5.39% (95% CI, 0.88–26.71) (Figure 5(e)). Figure 5.(Continued).
Six datasets from four studies analyzed the proportion of HZ recurrence in the IC/AID population and the general adult population.^34,57,74,76^
Among patients with IC/AID conditions, the pooled proportion of HZ recurrence was estimated at 3.75% (95% CI, 1.72–7.98) (Figure 6(a)). In the general adult population, the incidence of HZ recurrence ranged 2.05–5.07/1,000 PY, and the pooled proportion of HZ recurrence was 3.16% (95% CI, 2.02–4.92) (Figure 6(b)). Figure 6.Forest plots of proportion of HZ recurrence reported in studies included in the meta-analysis, in ascending order among (a) patients with IC/AID conditions, and (b) the general adult population. Articles IC/AID population,^34,74,76^ general adult population.^34,57,74^ AID: autoimmune diseases; CI: confidence interval; HZ: herpes zoster; IC: immunocompromised; SLE: systemic lupus erythematosus.
As most studies did not report the intervals between the first and the second episode of HZ, the time between the two episodes could not be adjusted for in this analysis.
This study consolidated and synthesized current data on HZ epidemiology and burden among the IC/AID population and the general adult population in APAC.
Patients with IC/AID conditions were found to have a higher risk of developing HZ than the general adult population, despite the high heterogeneity observed across individual studies. Pooled HZ IR in the IC/AID population (16.53/1,000 PY) was two to three times higher than that of the general adult population (6.49/1,000 PY). From studies which included both the IC/AID and the non-IC/AID populations, the pooled IRR was 1.98 (95% CI, 1.43–2.74), indicating that patients with IC/AID conditions were almost twice as likely to develop HZ than adults without these conditions. The pooled proportion of HZ patients with PHN was also higher in the IC/AID population (21.10%) than the general adult population (13.99%).
Our findings are consistent with studies conducted outside of APAC. For example, a retrospective cohort study from England reported a higher HZ IR in patients with IC conditions (7.8/1,000 PY) compared with those without IC conditions (6.2/1,000 PY),^14^ while a population-based retrospective study from Spain reported up to twelve-fold high HZ IR among hematopoietic stem cell transplantation (HSCT) recipients compared with immunocompetent adults.^77^ Among patients with AID in the US, age-specific IRs have similarly been estimated to be approximately 1.5–2 times greater than corresponding rates in healthy adults, and this elevated HZ incidence was particularly pronounced in patients with SLE.^78^ Given the elevated risk of HZ in patients with IC/AID conditions, vaccination strategies and policies targeting this patient population may have a substantial public health impact.
The risk of HZ increased with age in the general population, with HZ IRs being higher in adults aged ≥50 years than those aged <50 years. As most studies included participants with a wide age range, more age-specific analyses could not be performed in the current meta-analysis. Nevertheless, the results from this analysis are consistent with a worldwide meta-analysis, which found that older age was associated with a greater risk of HZ in the general population (risk ratio, 1.65 [95% CI, 1.37–1.97]),^9^ and a separate global meta-analysis in the general population which observed that HZ incidence in the highest age category (≥85 years) was over two-fold that of the lowest age category (50–54 years).^79^ In our current study, due to the lack of relevant studies reporting on age-stratified IRs, potential age trends were not explored for all IC/AID conditions. The IC/AID conditions with age-stratified data available showed a general trend of HZ incidence increasing with age, and therefore could be an area for future studies.
We found that the proportion of females was positively associated with higher HZ incidence in studies involving the general adult population; prior global studies have demonstrated that female sex is a risk factor for HZ and its recurrence.^9,79–81^ However, we did not find a significant association of female proportion with HZ incidence in the IC/AID population. This may be due to factors such as a wide variation of sex ratios within studies of each IC/AID condition, study design heterogeneity, and insufficient studies assessed.
In studies of patients with IC/AID conditions, condition-specific HZ IRs varied widely. HZ IRs tended to be higher in patients with hematological malignancies and SLE than in those with other IC/AID conditions, possibly due to the severe immune dysregulation associated with the pathogenesis of these diseases.^82^ Our results are comparable to global meta-analyses, which found that the risk of HZ was elevated in patients with SLE, malignancy, and RA, more so than other conditions.^9,80^ Studies specific to Europe and the US also observed that patients with malignancy/cancer, RA, and HSCT had the highest rates of HZ.^12,83^ Nonetheless, it is important to consider the regional/locale-specific context (e.g., differences in treatment practices and prophylactic measures) when considering HZ prevention for patients with specific IC/AID conditions.
Regarding HZ-related complications, the proportions of HZ patients with PHN and HZO in the IC/AID population similarly tended to be higher than the general population. Although not statistically significant, possibly due to small sample sizes and large within-study variation, these results are comparable with a retrospective cohort study in England, in which patients with IC/AID conditions reported a higher risk of HZ-associated complications (e.g., PHN, ocular complications) versus those without IC/AID conditions.^14^
There was an overall lack of studies on HZ recurrence; where available, HZ recurrence varied widely across studies due to differences in study design, follow-up periods, and study populations. Nonetheless, the pooled proportion of patients with IC/AID conditions who had recurrent HZ was numerically higher than the general adult population. These findings are in line with global literature which report wide-ranging proportions of HZ recurrence in IC individuals (0.0–18.2%) and the general adult population (1.2–9.6%),^81^ and a population-based US study whereby IC patients had a 2.35 times higher risk of HZ recurrence than immunocompetent patients over an average follow-up of 7.3 years.^84^
A high level of heterogeneity was observed for most outcomes, indicating considerable variation in the outcomes of interest reported across studies. The small sample sizes in studies of patients with IC/AID conditions may partially explain the greater variation of HZ IRs reported in these studies than those reported in the general adult population. Factors such as study design/setting, sample size, data source, patient demographics (e.g., age, sex, comorbidity, medications used), and length of follow-up may also contribute to the heterogeneity observed.
Although subgroup analyses and meta-regression were performed, a considerable amount of the heterogeneity remained unexplained. As meta-regression can only include study-level covariates, future studies exploring more covariates or involving patient-level data may be required to explain the remaining heterogeneity.
All studies included in this meta-analysis were evaluated against the JBI checklist which assesses the risk of bias, with the majority of checklist items being present (≥7 out of 9 items).
Nevertheless, while our findings are consistent with previous research, this study had several limitations that need to be considered in the overall interpretation of the results. Firstly, for several IC/AID conditions (e.g., multiple myeloma, non-Hodgkin lymphoma), there were limited published studies available, and therefore these conditions could not be included in the meta-analysis. Of the included studies, study design heterogeneity limited the ability to compare studies, as discussed in the subsection above. Furthermore, the number of studies within individual IC/AID populations was also insufficient for subgroup analysis or meta-regression to explore sources of heterogeneity.
This meta-analysis establishes the substantial burden of HZ and its complications among patients with IC and AID conditions in APAC. Further research is needed to better understand the risk of HZ and its prevention in patients with IC and AID conditions, particularly for conditions and countries/territories with limited relevant studies available. Nevertheless, it is crucial to raise awareness of HZ and strengthen access to preventative strategies in these patient populations. Our results may also inform clinical practice and management, such as guiding the development of medical association guidelines and government policies related to public health programs (e.g., HZ immunization programs, public subsidies), to alleviate the burden of HZ on at-risk patient populations and healthcare systems within APAC.