Grading of Recommendations, Assessment, Development, and Evaluation (GRADE): Use of Recombinant Zoster Vaccine in Immunocompromised Adults Aged ≥19 Years

Introduction

Recombinant zoster vaccine (RZV) was licensed in the United States for prevention of herpes zoster for adults aged ≥50 years by the Food and Drug Administration (FDA) and recommended for immunocompetent adults aged ≥50 years by the Advisory Committee on Immunization Practices (ACIP) in 2017* (1). On July 23, 2021, the FDA expanded the indication for RZV to include adults aged ≥18 years who are or will be at increased risk for herpes zoster because of immunodeficiency or immunosuppression caused by known disease or therapy (2). During December 2017–October 2021, the ACIP Herpes Zoster Work Group (HZWG) participated in monthly or bimonthly teleconferences to review herpes zoster epidemiology and evidence for the efficacy and safety of RZV in immunocompromised adults.

The ACIP adopted a modified Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach in 2010 as the framework for evaluating the scientific evidence that informs recommendations for vaccine use (3). As part of the process employed by the ACIP, a systematic review and GRADE evaluation of the evidence for use of RZV in immunocompromised adults was conducted and presented to ACIP. All CDC staff and ACIP HZWG members involved in the GRADE analysis complied with ACIP standards regarding conflict of interest reporting.

The primary policy question was, “Should adults ≥19 years of age who are or will be immunodeficient or immunosuppressed because of disease or therapy be recommended to receive two doses of recombinant zoster vaccine for the prevention of herpes zoster and related complications?” (Table 1).

Methods

We conducted a systematic review of evidence on the efficacy and safety of a two-dose regimen^† of RZV in immunocompromised adults. We assessed outcomes and evaluated the quality of evidence using the GRADE approach. GRADE evidence type indicates the certainty in estimates from the available body of evidence. Evidence certainty ranges from type 1 (high certainty) to type 4 (very low certainty) (3).

Patient-important outcomes (including benefits and harms) for assessment (Tables 1 and 2) were selected by the HZWG during work group calls and via online surveys where members were asked to rate and rank the importance of relevant outcomes. The potential benefits pre-specified by the HZWG included prevention of herpes zoster (critical), prevention of postherpetic neuralgia (important), and prevention of herpes zoster-related hospitalization (important). The pre-specified harms included serious adverse events (SAEs) (critical), immune-mediated disease (important), graft-versus-host-disease (important), graft rejection (important), and reactogenicity grade ≥3 (important).

We identified studies in Medline, Embase, CINAHL, Scopus, Cochrane Library, and clinicaltrials.gov, written in any language, with the earliest search date restricted to the earliest article containing phase I/II data on safety and immunogenicity of RZV from October 1, 2012. The search end date was April 27, 2021. Two records that were published after this cutoff date were included in the review. Search terms included herpes-zoster, varicella-zoster, zoster, shingles, zona, ADJ5 vaccin*, ADJ5 subunit, HZ su, Shingrix, RZV, GSK 1437173A, and many variations on these terms (for a full list of search terms, see Appendix 3).

Articles were included if they provided data on vaccination with RZV, involved human subjects, and included immunocompromised adults (ages 18 years and older). Specifically, the following six categories of immunocompromise were considered: (1) hematopoietic cell transplant recipients, (2) patients with hematologic malignancies, (3) renal or other solid organ transplant recipients, (4) patients with solid tumor malignancies, (5) persons living with HIV, (6) and immunocompromised populations at increased risk for herpes zoster not covered in groups 1 through 5, such as patients with primary immunodeficiencies, patients with autoimmune and inflammatory conditions, and patients taking immunosuppressive medications or therapies. In addition, efforts were made to obtain unpublished and other relevant data by hand-searching reference lists, and consulting with the vaccine manufacturer and subject matter experts. Titles and abstracts were screened and full-texts were reviewed independently and in duplicate by two separate reviewers.

The quality of evidence from included studies were assessed using the GRADE approach. The GRADE assessment across the body of evidence for each outcome was presented in a respective evidence profile.

Results

The results of the GRADE assessment were presented to ACIP on September 29, 2021 (4).

After title and abstract screening of 2,408 records, 133 studies were identified as eligible for full-text review. Of these, 114 were excluded, 53 due to being a duplicate or results being published in another manuscript, 22 for the wrong patient population, 14 for the wrong intervention, 13 only contained an abstract, and 12 had an ongoing study. This yielded 19 studies included in the evidence synthesis. Ten of these studies were excluded from the GRADE analysis (Appendix 2): Nine due to no comparison group for analysis and one due to no presentation of counts for the outcome under consideration, prevention of herpes zoster, but rather just odds ratios. Data included in the GRADE analysis were reviewed from two phase I/II randomized-controlled trials (RCTs), one phase II/III RCT, three phase III RCTs, one pooled post hoc efficacy analysis of two phase III RCTs, and two cohort studies (5–13). Characteristics of the included studies are listed in Appendix 1.

Benefits

Risk for herpes zoster, postherpetic neuralgia, and hospitalizations attributed to herpes zoster were reduced among persons who received two doses of RZV vs. placebo. Five studies in four immunocompromised groups evaluated herpes zoster as an outcome. Estimates of vaccine efficacy (VE) came from three studies, with VE of 68.2% (95% Confidence Interval [CI]: 55.6%–77.5%) for autologous hematopoietic cell transplant recipients, and 87.2% (44.3%–98.6%) and 90.5% (73.5%–97.5%) in post hoc efficacy analyses for patients with hematologic malignancies and potential immune-mediated diseases, respectively (Table 3a). Subgroup analyses by age revealed similar efficacy. Non-randomized studies examining patients with autoimmune and immunocompromising conditions or inflammatory bowel disease reported VE estimates of 68% (95% CI: 62.3%–72.8%) for those with autoimmune conditions and 64.1% (95% CI: 57.2%–69.8%) for those with immunocompromising conditions. For patients with inflammatory bowel disease over age 60, hazard ratios were reported of 0.41 (95% CI: 0.19–0.87) for non-steroid users and 0.34 (95% CI: 0.05–2.44) for steroid users. Immunogenicity data from six RCTs showed humoral vaccine response rates ranging from 65.4% to 96.2% and cell-mediated responses ranging from 50% to 93%, compared with 0% to 4.2% and 0% to 16.7% for humoral and cell-mediated response among placebo recipients, respectively (Table 3b). Only one study provided data on postherpetic neuralgia and herpes zoster-related hospitalizations in autologous hematopoietic cell transplant recipients, showing a VE of 89% (95% CI: 22%–100%) for prevention of postherpetic neuralgia (Table 3d) and 85% (95% CI: 32%–97%) for prevention of herpes zoster-related hospitalization (Table 3e).

Harms

Serious adverse events were evaluated in seven studies in six immunocompromised groups (2,541 RZV recipients). Serious adverse events were common in both vaccine and placebo groups, and the risk was balanced across groups (Table 3c). Overall, risk ratios for SAEs ranged from 0.79 (95% CI: 0.60–1.05) to 1.99 (95% CI: 0.42–9.44). Serious adverse events attributed to vaccination were rare, ranging from 0% to 1.6% among RZV recipients and 0% to 0.76% among placebo recipients. Non-specified immune-mediated diseases were reported by 0% to 1.4% of RZV recipients, new onset immune-mediated diseases were reported by 1.6% to 3.0% of RZV recipients, and exacerbations of existing immune-mediated diseases were reported by 2.8% of RZV recipients. Overall, similar rates of immune-mediated diseases were reported in the placebo group, with risk ratios ranging from 0 to 1.63 for non-specified, from 0.68 to 2.0 for new-onset, and 0.98 (95% CI: 0.58–1.65) for exacerbations of immune-mediated diseases (Table 3f). One study in patients with hematologic malignancies reported on graft-versus-host-disease among hematopoietic cell transplant recipients and found 3/19 (15.8%) RZV recipients experienced graft-versus-host-disease, compared to 4/21 (19.0%) placebo recipients, for a calculated risk ratio of 0.83 (95% CI: 0.21–3.24) (Table 3g). One study among renal transplant patients reported on graft rejection, with 4 episodes of rejection reported in 132 RZV recipients (3.0%), compared with 7/132 (5.3%) in placebo recipients, for a calculated risk ratio of 0.57 (95% CI: 0.17, 1.91) (Table 3h). Grade 3 reactogenicity was increased among persons who received at least 2 doses of RZV rather than placebo (Table 3i). Solicited local and systemic reactions within 7 days after vaccination were frequent and mostly mild to moderate. Grade 3 local reactions were reported by 10.7% to 14.2% of RZV recipients across reviewed studies, with the most frequent reported symptom being injection-site pain. Grade 3 solicited systemic reactions were reported by 9.9% to 22.3% of RZV recipients and 6.0% to 15.5% of placebo recipients, with the most commonly reported symptoms being fatigue and myalgia.

GRADE Summary

The GRADE findings and evidence for each outcome of interest are summarized in Tables 4 and 5. The initial GRADE evidence level was type 1 (high) for each outcome because the body of evidence was from RCTs. In terms of benefits, the available data indicated that the vaccine was effective for preventing herpes zoster, with VE ranging from 68.2% to 90.5% among several immunocompromised groups. This body of evidence was downgraded for indirectness because, across the body of evidence, data did not include all relevant immunocompromised groups under consideration, resulting in an overall certainty of type 2 (moderate). The certainty in the estimate for prevention of postherpetic neuralgia was type 3 (low), downgraded for indirectness and imprecision. The certainty in the estimate for prevention of herpes zoster-related hospitalizations was also type 3 (low), downgraded for indirectness and imprecision.

The certainty in the estimate of the effect of SAEs was type 2 (moderate), downgraded for indirectness because, across the body of evidence, data did not include all relevant immunocompromised groups under consideration. The certainty in the estimate of effect for immune-mediated diseases was type 4 (very low), downgraded for inconsistency, indirectness, and imprecision. The certainty in the estimate of the effect of graft-versus-host-disease was also type 4 (very low), downgraded for serious concern of indirectness and downgraded twice for very serious concern of imprecision due to a very small sample size, with only 7 events observed in a sample of 40 patients. The certainty in the estimate of the effect of graft rejection was type 3 (low), with a downgrade for indirectness and imprecision. Finally, the certainty in the effect of reactogenicity was type 2 (moderate), with a downgrade only for indirectness.

References

1. Dooling KL, Guo A, Patel M, et al. Recommendations of the Advisory Committee on Immunization Practices for use of herpes zoster vaccines. MMWR Morb Mortal Wkly Rep 2018;67:103–108.
   http://dx.doi.org/10.15585/mmwr.mm6703a5
2. Food and Drug Administration. Shingrix [package insert], revised: 07/2021. Silver Spring, MD: US Department of Health and Human Services, Food and Drug Administration; 2021. https://www.fda.gov/media/108597/download
3. Advisory Committee on Immunization Practices. Advisory Committee on Immunization Practices (ACIP): GRADE (grading of recommendations, assessment, development and evaluation). Atlanta, GA: US Department of Health and Human Services, CDC; 2019. www.cdc.gov/vaccines/acip/recs/index.html
4. Advisory Committee on Immunization Practices. ACIP meeting information. Atlanta, GA: US Department of Health and Human Services, CDC; September 29, 2021. www.cdc.gov/vaccines/acip/meetings/index.html
5. Stadtmauer EA, Sullivan KM, Marty FM, et al. A phase 1/2 study of an adjuvanted varicella-zoster virus subunit vaccine in autologous hematopoietic cell transplant recipients. Blood 2014;124:2921–9. https://doi.org/10.1182/blood-2014-04-573048
6. Berkowitz EM, Moyle G, Stellbrink HJ, et al. Zoster-015 HZ/su Study Group. Safety and immunogenicity of an adjuvanted herpes zoster subunit candidate vaccine in HIV-infected adults: a phase 1/2a randomized, placebo-controlled study. J Infect Dis 2015;211:1279–87. https://doi.org/10.1093/infdis/jiu606
7. Bastidas A, de la Serna J, El Idrissi M, et al. ZOE-HSCT Study Group Collaborators. Effect of recombinant zoster vaccine on incidence of herpes zoster after autologous stem cell transplantation: a randomized clinical trial. JAMA 2019;322:123–33. doi: 10.1001/jama.2019.9053.
8. Dagnew AF, Ilhan O, Lee WS, et al. Zoster-039 Study Group. Immunogenicity and safety of the adjuvanted recombinant zoster vaccine in adults with haematological malignancies: a phase 3, randomised, clinical trial and post-hoc efficacy analysis. Lancet Infect Dis 2019;19:988–1000. https://doi.org/10.1016/S1473-3099(19)30163-X
9. Dagnew AF, Rausch D, Hervé C, et al. ZOE-50/70 Study Group. Efficacy and serious adverse events profile of the adjuvanted recombinant zoster vaccine in adults with pre-existing potential immune-mediated diseases: a pooled post hoc analysis on two parallel randomized trials. Rheumatology (Oxford) 2021;60:1226–33. https://doi.org/10.1093/rheumatology/keaa424
10. Vink P, Delgado Mingorance I, Maximiano Alonso C, et al. Zoster-028 Study Group. Immunogenicity and safety of the adjuvanted recombinant zoster vaccine in patients with solid tumors, vaccinated before or during chemotherapy: a randomized trial. Cancer 2019;125:1301–12. https://doi.org/10.1002/cncr.31909
11. Vink P, Ramon Torrell JM, Sanchez Fructuoso A, et al. Z-041 Study Group. Immunogenicity and safety of the adjuvanted recombinant zoster vaccine in chronically immunosuppressed adults following renal transplant: a phase 3, randomized clinical trial. Clin Infect Dis 2020;70:181–90. https://doi.org/10.1093/cid/ciz177
12. Izurieta HS, Wu X, Forshee R, et al. Recombinant zoster vaccine (Shingrix): real-world effectiveness in the first 2 years post-licensure. Clin Infect Dis 2021;73:941–48. https://doi.org/10.1093/cid/ciab125
13. Khan N, Wang L, Trivedi C, et al. Efficacy of recombinant zoster vaccine in patients with inflammatory bowel disease. Clin Gastroenterol Hepatol 2021:S1542-3565(21)00754-0. https://doi.org/10.1016/j.cgh.2021.07.023

* This recommendation became official CDC policy in January 2018.

† Some studies included 3 doses of RZV, as noted in the respective GRADE tables.

Table 1: Policy Question and PICO

1. The second RZV dose should typically be given 2–6 months after the first; for persons who are or will be immunodeficient or immunosuppressed and who would benefit from a shorter vaccination schedule, the second dose can be administered 1–2 months after the first.
2. HCT = Hematopoietic cell transplant
3. SOT = Solid organ transplant

Table 2: Outcomes and Rankings

^aThree options: 1. Critical; 2. Important but not critical; 3. Not important for decision making
^bSOT = Solid organ transplant

Appendix 1. Studies Included in the Review of Evidence

Abbreviations: AI = Autoimmune; CMS = Center for Medicare and Medicaid Services; FDA = Food and Drug Administration; GSK = GlaxoSmithKline; HCT = Hematopoietic Cell Transplant; HZ = Herpes Zoster; IBD = Inflammatory Bowel Disease; IC = Immunocompromised; pIMD = Potential Immune-Mediated Diseases; RCT = Randomized Control Trial; RZV = Recombinant Zoster Vaccine; SAEs = Serious Adverse Events

^a. This study included participants from the ZOE-50/70 trials with diagnosed autoimmune and inflammatory conditions not on immunosuppressive therapies. These participants may represent a less severe spectrum of autoimmune and inflammatory disease.

Table 3a: Summary of Studies Reporting Herpes Zoster (HZ)

Abbreviations: AI = Autoimmune; CI = Confidence Interval; HCT = Hematopoietic Cell Transplant; IC = Immunocompromised; NE = Not Evaluable; pIMD = Potential Immune-Mediated Disease; HR = Hazard Ratio; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine; VAHS = Veteran’s Affairs Health System; VE = Vaccine Efficacy

^a. A suspected case of HZ was defined as (1) a new rash characteristic of HZ (e.g., unilateral, dermatomal, and accompanied by pain broadly defined to include allodynia, pruritus, or other sensations) or a vesicular rash suggestive of varicella-zoster virus (VZV) infection regardless of distribution, with no alternative diagnosis or (2) clinical symptoms and/or signs suggestive of  VZV infection and specific laboratory findings, such as VZV–positive culture or immunohistological staining or RT-PCR assay in the absence of characteristic HZ rash. Participants with any suspicion of HZ were examined by investigators within 96 hours. Participants were followed up for at least 90 days after rash onset or until the rash resolved and the participant had been pain-free for 4 weeks. HZ episodes were confirmed using RT-PCR assay or, if not possible, by a blinded ascertainment committee.

^b. Data presented in the table above is from the modified total vaccination cohort (mTVC), which included all participants who received 2 doses RZV/Placebo and excludes participants who developed an episode of HZ <1 month after receiving the second dose. Sensitivity analyses in the total vaccinated cohort (TVC), containing all who received at least one dose of RZV/placebo, found 242 confirmed cases of HZ (70/922 in the vaccine group (incidence 34.7/1000 p-y) and 172/924 in the placebo group (incidence 95.6/1000 p-y)), resulting in an IRR of 0.36 (95% CI: 0.27-0.48). Among those aged 18-49 years (TVC), there were 13 cases/230 vaccinated participants (incidence: 25.2/1000 p-y), and 39 cases/229 participants in the placebo group (incidence: 84.9/1000 p-y). Among those aged ≥50 years (TVC), there were 57 cases/692 vaccinated participants (37.9/1000 p-y), and 133 cases/695 placebo recipients (incidence 99.3/1000 p-y). These results corresponded to an IRR of 0.30 (95% CI: 0.15-0.57) for the TVC aged 18-49 years, and an IRR of 0.38 (95% CI 0.28-0.52) for the TVC aged ≥50 years.

^c. Berkowitz, 2015 used a 3-dose regimen of HZ/su at months 0, 2, and 6. 74 participants received 1 dose, 72 received 2 doses, and 71 received 3 doses of HZ/su, and 49 received 1 dose placebo, 47 received 3 doses placebo. In the table above, the denominator was subset to those patients who received at least 2 doses HZ/su/Placebo.

^d. One case of HZ was reported during the study. This case occurred in the HZ/su group 83 days after the subject’s first and only vaccine dose, and thus is not reported in the table because the individual did not complete 2 doses of HZ/su (the pre-determined intervention criteria)

^e. A suspected case of HZ was defined as a new unilateral, dermatomal rash accompanied by pain (broadly defined to include allodynia, pruritus, or other sensations), or a vesicular rash suggestive of VZV infection regardless of the distribution, and no alternative diagnosis; or a clinical presentation and specific laboratory findings (VZV positive PCR, culture, or immunohistochemical staining) specific to VZV infection in the absence of characteristic HZ rash. HZ cases were confirmed by PCR on samples collected from lesions or by a HZ ascertainment committee (HZAC) if the case could not be confirmed or excluded by PCR. The HZAC comprised three physicians with expertise in infectious diseases and hematology or oncology who were masked to treatment assignments and PCR results. Each member classified the reviewed cases as HZ or non-HZ based on the available clinical information from the study site. A suspected case was only considered HZ if the HZAC members concurred unanimously.

^f. The authors assessed post hoc efficacy in the mTVC, which included all participants from the total vaccinated cohort, except those who did not receive the second dose or who developed a confirmed case of HZ before 30 days after dose two (n = 259 in the vaccine group and n = 256 in the placebo group, with results presented above).

^g. The analysis of vaccine efficacy in the subset of pooled ZOE-50/70 participants with pre-existing pIMDs at enrollment was performed in the modified total vaccinated cohort (mTVC), which included all participants who received both doses of study vaccine and did not develop a confirmed HZ episode before 30 days post-dose 2. For the efficacy assessment, only the first confirmed HZ episode in a participant was considered, and the follow-up period ceased at the time of the first occurrence.

^h. While the RCTs (ZOE 50/70) met low risk of bias criteria, given this analysis for this subgroup was post hoc and the analysis was not powered for this outcome nor able to address type 1 error, this analysis has moderate/high risk of bias.

^i. At the start of study, subjects were educated to recognize the signs and symptoms of HZ. They were to contact the study physician immediately for clinical evaluation and sample collection for VZV-specific PCR if they suspected HZ.

^j. At study months 0, 1, and 3, subjects received 3 doses of 50mg VZV gE combined with the liposome-based adjuvant AS01B; 3 doses of 50mg gE combined with the liposome-based adjuvant AS01E; 1 dose of saline followed by 2 doses of gE/AS01B; or 3 doses of saline. The now-licensed SHINGRIX vaccine is administered according to the third protocol: 2 doses of gE/AS01B. In the 3 doses gE/AS01E group, there were 2 cases of HZ among 29 participants (6.9%), however these results are not presented because this was not the adjuvant ultimately used in the SHINGRIX vaccine.

^k. Risk ratio was calculated, along with Wald confidence intervals, in R and SAS.

^l. Beneficiaries were classified into an autoimmune population if they consulted for any of the selected autoimmune conditions included at least twice in the 1 year prior to the index date. Autoimmune definitions were adapted from Cooper et al. 2009. These included patients diagnosed with any of the following conditions 365 days prior to index date: Graves’ disease, Hashimoto’s thyroiditis, MS, Myasthenia gravis, Polymyalgia rheumatica, Primary biliary cirrhosis, Psoriasis, Psoriatic arthritis, Rheumatoid arthritis, Scleroderma, Sjogren’s syndrome, Systemic lupus erythematosus, Ulcerative colitis (UC), Vitiligo. The authors used a previously published, independently validated algorithm presented by Greenberg and colleagues to determine categories of immunosuppression.

^m. HZ was defined in a claim in an outpatient or professional setting with ICD-10 diagnosis code for HZ (B020, B021, B022.x, B027-B029) in any position with a claim for HZ specific antiviral identified using national drug codes within Medicare part D claims within 7 days of diagnosis. As a sensitivity analysis, the authors defined cases using the same clinical definition without the antiviral prescription.

^n. Overall, this study presents with concerns of confounding, with no demographics or risk-factors presented for the immune-compromised and autoimmune populations. Additionally, given that it is a Medicare claims-based study, reliant on algorithmic determination of immunocompromised and autoimmune status, there remains significant risk of both confounding and information bias in interpreting the VE.

^o. HZ was determined by relevant ICD-9-CM (053.xx) and ICD-10-CM (B02.xx) codes for HZ and their complications. To ensure that the codes accurately reflected incident HZ cases, the clinical notes and laboratory data of each patient who had the code was individually reviewed by 2 members of the team.

^p. In the Cox regression model without any interaction, full dose of RZV was associated with lower risk of HZ compared with the unvaccinated group, after adjusting for other baseline and time-varying covariates. Specifically, in the 50 to 60-year group, the HR was 0 (95% CI, 0-0). The HR was 0.39 (95% CI,0.19-0.80) in the >60 group.

^q. This was a large cohort analysis, yet the VA patient population may not be generalizable to the general population. Coupled with the authors’ retrospective case ascertainment, this analysis has moderate/high risk of bias

Table 3b: Summary of Studies Reporting Herpes Zoster via Immunogenicity

Abbreviations: CI = Confidence Interval; gE = glycoprotein E; GMR = Geometric Mean Ratio; HCT = Hematopoietic Cell Transplant; IQR = Inter-Quartile Range; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine; SD = Standard Deviation;

^a. Humoral VRR was 45% 24 months after dose 2. The highest anti-gE geometric mean concentrations were recorded 1 month after dose 2 and remained above baseline 24 months after dose 2. Vaccine CMI response was 71% 24 months after dose 2. Frequency of gE-specific CD4[2+] T cells was highest 1 month after dose 2 and declined but remained above pre-vaccination levels through 24 months after dose 2. In the placebo group, humoral response was highest 24 months after dose 2 with no increase observed for anti-gE antibody concentrations at any postvaccination time. CMI was highest 24 months after dose 2 with no increase observed for CD4[2+] T-cell frequencies.

^b. Anti-gE GMCs increased significantly after 3rd dose and reached maximum 1 month after 3rd dose. One year after 3rd dose, levels remained above pre-vaccination concentrations, but decreased 60% between month 7 and 18. The proportions of subjects with anti-gE humoral vaccine responses were between 92.3% and 98.1% at the different timepoints. At month 7, frequency of gE-specific CD4 cells was higher following vaccinated group than saline (P < 0.0001). Frequency peaked 1 month after 2nd dose of RZV but did not increase further after the 3rd dose. GMR of gE-specific CD4 frequencies at 1 month after dose 1 vs.1 month after dose 2 was 1.04; comparable response profile for VZV-specific T cells. The proportion of respondents with vaccine associated CMI response was 40% 1 month after dose 1, 85.7% after dose 2, 90% after dose 3; 64.5% of subjects had CMI levels above response threshold 1 year after last dose. In the placebo group, the proportion of subjects with anti-gE humoral vaccine response was <2.8%, and the proportion of subjects with a vaccine associated CMI response was between 8.3% and 16.7%. No placebo recipients had CMI levels above the response threshold 1 year after the last dose.

^c. The pre-vaccination anti-gE antibody GMC was 964.0 mIU/mL (95% CI 814.5–1140.8) in the vaccine group and 13445.6 mIU/mL (10158.9–17795.6) at month 2, with 142 (65.4%) of 217 participants having a humoral vaccine response. In all participants, excluding those with non-Hodgkin B-cell lymphoma and chronic lymphocytic leukemia at month 2, 119 (80.4%) of 148 participants had a humoral vaccine response. The adjusted geometric mean anti-gE antibody concen­tration was 23132.9 mIU/mL(16642.8–32153.9). At month 13, the anti-gE antibody GMC was 5202.7 mIU/mL (95% CI 4074.8–6642.8) and 86 (52.1%) of 165 participants had a humoral vaccine response. The proportion of participants with a humoral vaccine response was higher in participants who were vaccinated after immunosuppressive cancer treatments compared with those who were vaccinated during immunosuppressive treatments. Median CD4[2+] T-cell frequencies before vaccination were 77.5 (IQR 1.0–191.4). At month 2, median CD4[2+] T-cell frequencies were 3081.9 (1766.2–7413.6), and 36 (83.7%) of 43 participants had a vaccine response.  At month 13, the median CD4[2+] T-cell frequency was 1006.7 (IQR 416.0–3284.5) and 22 (66.7%) of 33 participants had a CMI vaccine response. For the placebo group: pre-vaccination anti-gE antibody GMC was 883.7 mIU/mL (749.9–1041.4). At month 2, the anti-gE GMC was 832.0 mIU/mL (701.1–987.3), and 1 (0.5%) of 198 participants had a humoral vaccine response. Excluding those with non-Hodgkin B-cell lymphoma and chronic lymphocytic leukemia at month 2, 1 (0.8%) of 130 placebo recipients had a humoral vaccine response. The adjusted anti-gE GMC was 7776 mIU/mL (7028–8603). At month 13, 5 (3.6%) of 140 participants had a humoral vaccine response. Median CD4[2+] T-cell frequencies before vaccination were 101.2 (1.0–193.1) in the placebo group, and 99.1 (1.0–268.3) at month 2, and 3 (6.8%) of 44 participants had a vaccine response.

^d. Stadtmauer et al. reported immunogenicity of three dosing regimens: 2 and 3 doses of HZ/su adjuvanted with AS01B (the adjuvant used in RZV), and 1 regimen adjuvanted with AS01E, 3 doses. AS01E was not ultimately used in the RZV vaccine, and the table above only includes results from the 2-dose regimen, as that is the formulation currently used for SHINGRIX.

^e. 1 month after the final dose, anti-gE GMCs were higher in all gE/AS01 groups than in the saline group (P<0.0001). GMCs peaked at month 4 in the 3 gE/AS01 groups and were 25- to 55-fold higher than before vaccination. These GMCs remained high for at least 1 year after the last vaccination, although they decreased 29% to 46% from month 4 to month 15. Approximately one-third of subjects had an anti-gE response after the first dose of gE/AS01B or gE/AS01E. At month 4, response rates were similar for the 3 vaccine groups (72.4%-76.9%). Anti-gE GMCs after vaccination did not increase in subjects with non-Hodgkin B-cell lymphoma, although the GMCs increased for all other diagnoses. At month 4, gE-specific CMI, as measured by the CD4[2+] T-cell frequency after induction with gE, was significantly higher for all gE/AS01 groups than for saline (P<0.0001). The CD4[2+] T-cell frequencies in the gE/AS01B 3-dose and 2-dose groups were superior to that of the saline group. One month after the second dose of gE/AS01B, the CD4[2+] T-cell frequency was 2.3-fold higher after early vaccination (2 and 3 months post-HCT) than after delayed vaccination (3 and 5 months post-HCT). There were no differences in CMI responses according to underlying disease. At month 4, gE-specific CMI was significantly higher for all gE/AS01 groups than for saline (P<0.0001). Additionally, reported CMI GMRs for the 3-dose regimen are not reported in the table above, but are 74.41 (25.74-215.09) for 1 month after last vaccination, and 28.70 (10.92-75.46) for 12 months after last vaccination.

^f. Before vaccination, >99% of participants were seropositive for anti-gE antibody. At month 1, in pooled RZV recipients, the GMC was 24793.1 mIU/mL; both GMCs and humoral vaccine response rates (VRRs) were higher in RZV-PreChemo than in the RZV-OnChemo group.  In pooled RZV recipients, the month 2 GMC was 18291.7 mIU/mL and remained higher in the RZV-PreChemo group than in the RZV-OnChemo group. The humoral VRR to RZV was 93.8% (95% CI, 85.0%-98.3%) at month 2 in the RZV-PreChemo group. At month 13, the GMC was 4477.3 mIU/mL in pooled RZV recipients; both GMCs and humoral VRRs were similar in the RZV-PreChemo group and RZV-OnChemo group. In RZV recipients ≥50 years of age, GMCs tended to be higher at month 2, 6, and 13than in those 18-49 years of age, while humoral VRR tended to be higher only at month 6 and 13. Before vaccination, median CD4[2+] T-cell frequencies were 127.3 in RZV recipients. At month 2, the CMI VRR was 50.0% (95% CI, 28.2%-71.8%). In RZV recipients, the highest median CD4[2+] T-cell frequency was observed at month 2 (778.8), which persisted above baseline up to month 13 (332.9), a level similar to that observed at month 1 (391.9). CMI VRRs followed a similar pattern, with the highest VRR at month2. CD4[2+] T cell frequencies were in similar ranges in RZV recipients 18-49 and ≥50 years of age, whereas CMI VRRs tended to be higher in those 18-49 years of age at all postvaccination time points. For the placebo group, postvaccination anti-gE antibody GMCs and humoral VRRs remained near the pre-vaccination level. Additionally, the CMI GMR, not listed in the table above, was reported as 9.94 (3.63-27.19).

^g. The humoral VRR in the RZV group was 80.2% (95% CI: 71.9%–86.9%) at month 2. Anti-gE antibody GMCs increased from 1354.4 mIU/mL at pre-vaccination to 19163.8 mIU/mL at month 2 and persisted through month 13 at 8545.1 mIU/mL. No apparent differences in anti-gE antibody GMCs were seen linked to the types of maintenance immunosuppressive therapy. Humoral immune responses appeared higher in the 18–49 years of age cohort than in the ≥50 years of age cohort, but within the 18–49 years of age cohort, humoral immune responses similar for ages 18–29 and 30–49 years. The VRR for CMI responses was 71.4% (51.3%–86.8%) in the RZV group at month 2. At pre-vaccination, median CD4[2+] T-cell frequencies were 21.2 in the RZV group. Median frequencies increased to 2149.0 at month 2 and remained significantly elevated over baseline at month 13. At each time point, CMI responses appeared to be higher in the 18–49 years of age cohort. Within the 18–49 years of age cohort, CMI responses were similar for the age 18–29 and 30–49 years of age groups. For the placebo group: postvaccination anti-gE antibody GMCs, humoral VRRs, CD4[2+] T-cell frequencies, and CMI VRRs remained near the pre-vaccination level. At pre-vaccination, median CD4[2+] T-cell frequencies were 59.7 in the placebo group. The CMI GMR, not listed in the table above, was reported as17.26 (5.92-50.36).

Table 3c: Summary of Studies Reporting Serious Adverse Events (SAEs)^a

Abbreviations: CI = Confidence Interval; HCT = Hematopoietic Cell Transplant; pIMD = Potential Immune-Mediated Disease; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine; SD = Standard Deviation
*All risk ratios and Wald confidence intervals calculated using R and SAS.

^a. Death, life-threatening event, hospitalization, incapacity to perform normal life functions, medically important event, or congenital anomaly/birth defect. The figures reported in the above table represent individuals who reported at least one SAE during the entire study period or longest period of follow-up, with footnoted details for shorter periods (i.e. 30 days after last vaccination) of follow up and individuals who experienced >1 SAE.

^b. SAEs within 30 days after last vaccination occurred in 7% of both vaccine (n = 68/922) and placebo (n=66/924) recipients. The most frequent SAEs were neoplasms. Three participants in the vaccine group reported 5 serious adverse events (neutropenia, immune thrombocytopenic purpura, cutaneous vasculitis, arthralgia, and atrial fibrillation) that were considered related to vaccination; 4 participants in the placebo group reported 4 such events (constipation, HZ, disseminated cutaneous HZ, and skin eruption). Fatal SAEs occurred in 242 participants (vaccine group, n=118; placebo group, n=124) during the entire study and were mainly due to recurring malignancy and non–HZ–related infections.

^c. Within 30 days after last vaccination, 4 SAEs were reported (2 each in the HZ/su and saline groups), but none were considered related to vaccination. Over the study period, 7 SAEs were reported in 6 of 74 subjects (8.1%) in the HZ/su group, and 2 SAEs were reported in 2 of 49 (4.1%) in the saline group. None were considered related to vaccination. One subject withdrew because of 2 SAEs (portal hypertension and esophageal variceal hemorrhage) 97 days after the second HZ/su dose. No deaths were reported.

^d. Within 30 days of last vaccination, SAEs occurred in 17/283 RZV recipients (6.0%) and 29/279 placebo recipients (10.4%). Of these, none was considered related to vaccination. Within 6 months of the last vaccination, SAEs occurred in 50/283 RZV recipients (17.7%) and 60/279 placebo recipients (21.5%), none of which was considered related by the investigator in the vaccine group, and 1 which was considered related in the placebo group (1/279, 0.4%). For the full study period, the SAE counts are described in the table above, with 1 SAE considered related to vaccination in each of the RZV and placebo groups (1/283, 0.4%, 1/279, 0.4%, respectively). The incidence of SAEs was similar between the study groups during all time periods evaluated. The most frequent SAEs were febrile neutropenia (14 (4.9%) vaccine participants and 11 (3.9% placebo) and pneumonia (11 (3.9%) in each study group).  From first vaccination until study end, fatal SAEs were reported in 29 (10.2%) of 283 participants in the vaccine group and 37 (13.3%) of 279 participants in the placebo group.

^e. The occurrence of fatal SAEs during the entire study period was similar between the RZV (5.1%) and placebo (6.6%) groups and appeared to be similar to the overall pooled ZOE-50/70 population (RZV: 4.3%, placebo: 4.6%).

f. While the RCTs (ZOE 50/70) met low risk of bias criteria, given this analysis for this subgroup was post hoc and the analysis was not powered for this outcome nor able to address type 1 error, this analysis has moderate/high risk of bias.

^g. Through study month 15, 54 SAEs were reported in 33 subjects (6 in the gE/AS01B 3-dose group, 9 in the gE/AS01E 3-dose group, and 8 in the saline group). Only the SAEs reporting using AS01B adjuvant are included, as the AS01E adjuvant was ultimately not included in the RZV vaccine. One SAE was considered possibly related to vaccination by the investigator: pneumonia 105 days after a second dose of vaccine in the gE/AS01B 2-dose group. Nine subjects died during the study: 2 from recurrence of underlying malignancy and 2 from unknown reasons. No deaths were considered vaccination-related.

^h. SAEs reported from the first vaccination through 30 days after last vaccination occurred in 16/117 (13.7%) RZV recipients and 14/115 (12.2%) placebo recipients. SAEs reported from 30 days after last vaccination through the study end occurred in 30/117 (25.6%) RZV recipients and 31/115 (27.0%) placebo recipients. The most frequent SAEs reported were infections and infestations, malignant neoplasms, and neutropenia. Fatal SAEs occurred in 12 (10.3%) of RZV and 11 (9.6%) of placebo recipients. None of the SAEs were considered by the investigators to be causally vaccine related. In addition to the fatalities, 2 participants withdrew due to nonfatal SAEs: 1 RZV recipient experienced brain metastases and febrile neutropenia, and 1 placebo recipient had pulmonary metastases of prostatic cancer, anemia, bilateral hydronephrosis, and acute renal failure.

^i. From first vaccination through month 13, SAEs were reported by 26 (19.7%) RZV and 33 (25.0%) placebo recipients. Of these, 3 SAEs (febrile neutropenia, mucosal inflammation, and Burkitt lymphoma) were considered causally related to vaccination and were reported by 1 placebo recipient. From first vaccination to 30 days after last vaccination, there were 6 SAEs reported in 132 RZV recipients (4.5%) and 5 SAEs reported in 132 placebo recipients 93.8%). From 30 days after last vaccination to study end, there were 21 SAEs reported in the RZV group (15.9%), and 29 SAEs reported in the placebo group (22.0%).

Table 3d: Summary of Studies Reporting Postherpetic Neuralgia (PHN)^a

Abbreviations: CI = Confidence Interval; HCT = Hematopoietic Cell Transplant; P-Y = Person-Years; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine; VE = Vaccine Efficacy

^a. PHN = presence of HZ-associated ‘worst’ pain persisting or appearing more than 90 days after onset of the HZ rash; ‘Worst’ pain = HZ-associated pain rated as 3 or greater on the “worst pain” Zoster Brief Pain Inventory question.

^b. Results presented above for modified total vaccination cohort (mTVC), which included individuals who received 2 doses of RZV. Sensitivity analyses were conducted using the total vaccinated cohort (TVC; individuals receiving at least 1 vaccine), with the results are as follows: n/N intervention group (TVC) = 4/70, n/N comparison group (TVC) = 12/172. The reported incidence in the RZV group (TVC) = 1.9/1000 p-y, and reported incidence in the comparison group (TVC) was 5.8/1000 p-y. The VE for the TVC was 67%.

Table 3e: Summary of Studies Reporting Herpes Zoster-Related Hospitalization

Abbreviations: CI = Confidence Interval; HCT = Hematopoietic Cell Transplant; P-Y = Person-Years; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine; VE = Vaccine Efficacy

^a. Results in table above presented for modified total vaccination cohort (mTVC), individuals who received 2 doses RZV. The authors conducted sensitivity analyses on the total vaccinated cohort (TVC), evaluating all participants who received at least one dose of the vaccine. Results are as follows: n/N RZV group (TVC) = 4/70; n/N comparison group (TVC) = 20/172. Incidence RZV group (TVC) = 1.9/1000 p-y, incidence comparison group (TVC) = 9.7/1000 p-y.

^b. The authors reported a hazard ratio (HR) of 0.15 (95% CI, 0.03-0.68; P = .01), which was converted to VE using the formula (1-HR)*100%.

Table 3f: Summary of Studies Reporting Immune-Mediated Disease (IMD)^a

Abbreviations: CI = Confidence Interval; HCT = Hematopoietic Cell Transplant; NE = Not evaluable; pIMD = Potential Immune Mediated Disease; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine
*Risk ratios were calculated using Wald confidence intervals in SAS and R.

^a. Bastidas 2019, Dagnew 2019, and Vink 2019 did not specify the type of pIMD recorded. Dagnew 2021 recorded both exacerbations and new-onset immune-mediated disease, and Stadtmauer 2014 and Vink 2020 only reported new-onset immune-mediated disease.

^b. pIMDs were assessed within 365 days post-last vaccination in the total vaccinated cohort (TVC). In the RZV group, 3 pIMDs were deemed related to vaccination (0.3%), and in the placebo group, 0 were deemed related to vaccination. For the RZV group, MedDRA Primary System Organ class and preferred term included skin and subcutaneous tissue disorders (n=4), blood and lymphatic system disorders (n=3), musculoskeletal and connective tissue disorders (n=2), nervous system disorders (n=2), endocrine disorders (n=2), and infections and infestations (n=1). Psoriasis (n=2) was the most reported pIMD in the RZV group. In the placebo group, interstitial lung disease (n=2) was the most reported.

^c. pIMDs were assessed at three time periods: from first vaccination to 30 days after last vaccination, from first vaccination up to 6 months after last vaccination, and from first vaccination until study end. The table reflects pIMDs from first vaccination until study end. For the RZV group, through 30 days after last vaccination, 1/283 (0.4%) RZV recipients experienced a pIMD, and from first vaccination through 6-months after last vaccination, 3/283 (1.1%) recipients experienced a pIMD. For the placebo group, there were 0/279 (0%) pIMDs reported in the first 30 days after last vaccination, and 1/279 (0.4%) through 6 months after last vaccination. These diseases by MedDRA Preferred Term were autoimmune pancytopenia, gout, and erythema nodosum in the vaccine group, and autoimmune hemolytic anemia and Guillain Barré syndrome in the placebo group.

^d. Exacerbations of pre-existing pIMDs and new-onset pIMDs were assessed over the whole study period. Most frequent pre-existing pIMDs reported at enrolment were psoriasis, spondyloarthropathy, rheumatoid arthritis, and celiac disease; their prevalence was balanced between study groups.

^e. While the RCTs (ZOE 50/70) met low risk of bias criteria, given this analysis for this subgroup was post hoc and the analysis was not powered for this outcome nor able to address type 1 error, this analysis has moderate/high risk of bias.
^f. pIMDs were evaluated over the entire study period.

^g. pIMDs were evaluated at two time periods: from first vaccination through 30 days after last vaccination, and from 30 days after last vaccination through study end. For the period, there were 0 pIMDs reported in both the RZV and placebo groups. Through study end, there were 0/117 (0%) pIMDs reported in the RZV group, and 1/115 (0.9%) reported in the placebo group.

^h. Vink 2020 reported new-onset pIMDs from 30 days after last vaccination to the study end, as is reported in the table above. There were 0 pIMDs reported in either the RZV or placebo group within 0-30 days after last vaccination.

Table 3g: Summary of Studies Reporting Graft-Versus-Host-Disease

Abbreviations: CI = Confidence Interval; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine

*Risk ratios and Wald confidence intervals were calculated in R and SAS.

^a None of the events was considered related to vaccination by the investigator and most of the events were mild or moderate in intensity.

Table 3h: Summary of Studies Reporting Graft Rejection

Abbreviations: CI = Confidence Interval; HCT = Hematopoietic Cell Transplant; P-Y = Person-Years; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine; VE = Vaccine Efficacy

*Risk ratios and Wald confidence intervals were calculated in R and SAS.

^a. Renal allograft function and rejections were followed for up to 1 year, reported period for graft rejection is from 30 days after last vaccination until study end.

^b. No episodes of graft rejection were reported in the time interval between the first vaccination up to 30 days after last vaccination (0/132 in both RZV and comparator groups).

Table 3i: Summary of Studies Reporting Reactogenicity (Grade 3)^a

Abbreviations: CI = Confidence Interval; HCT = Hematopoietic Cell Transplant; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine; VE = Vaccine Efficacy

*Risk ratios and Wald confidence intervals were calculated in R and SAS.

^a. All trials solicited events through electronic diaries for 7 days following each dose. For local reactions, grade 3 was defined as: pain at injection site that prevents normal activity; redness > 10 cm; and swelling > 10 cm. For systemic reactions, grade 3 was defined as: fever >39.0°C; gastrointestinal symptoms, headache, fatigue, shivering, or myalgia that prevents normal activity.

^b. Adverse events were analyzed in the total vaccinated cohort (TVC) which included all participants who received at least 1 dose of vaccine or placebo.

^c. Overall local grade 3 reactions were noted in 128/901 RZV recipients (14.2%). Grade 3 pain was reported in 99/901 (11.0%) RZV recipients, redness in 28/901 (3.1%) RZV recipients, and swelling in 13/901 RZV recipients (1.4%).

^d. Overall local grade 3 reactions were noted in 3/892 placebo recipients (0.3%). Grade 3 pain was reported in 3/892 placebo recipients (0.3%), and grade 3 redness and swelling were reported in 0/892 placebo recipients (0%).

^e. Overall systemic grade 3 reactions were noted in 119/901 (13.2%) RZV recipients. Grade 3 fatigue was reported in 66.901 (7.3%), gastrointestinal symptoms in 18/901 (2.0%), headache in 26/901 (2.9%), myalgia in 56/901 (6.2%), shivering in 35/901 (3.9%), and fever in 3/901 (0.3%) RZV recipients.

^f. Overall systemic grade 3 reactions occurred in 54/892 (6.0%) placebo recipients, with grade 3 fatigue, GI symptoms, headache, myalgia, shivering, and fever occurring in 31/892 (3.5%), 17/892 (1.9%), 10/892 (1.1%), 19/892 (2.1%), 7/892 (0.8%), and 1/892 (0.1%) placebo recipients.

^g. Berkowitz, 2015 did not report overall local and systemic reactions, just symptom-specific. In the RZV group, grade 3 pain was reported in 12/73 (16.4%) RZV recipients, redness in 4/73 (5.5%), and swelling in 1/73 (1.4%).

^h. In the RZV group, grade 3 fatigue was reported in 12/73 (16.4%) RZV recipients, gastrointestinal in 2/73 (2.7%), headache in 6/73 (8.2%), myalgia in 10/73 (13.7%), shivering in 11/73 (15.1%), and fever in 0/73 (0%) RZV recipients.

^i. In the placebo group, grade 3 fatigue was reported in 4/48 (8.3%) recipients, gastrointestinal symptoms in 1/48 (2.1%), headache in 2/48 (4.2%), myalgia in 1/48 (2.1%), and shivering and fever in 0/48 (0%) placebo recipients.

^j. Stadtmauer, 2014 did not report overall local and systemic reactions, just symptom-specific. In the RZV group, two dosing regimens were considered: 3 doses of gE/AS01B, and 2 doses of gE/AS01B with one dose placebo. For the 3-dose regimen, grade 3 pain was reported in 3/30 (10%) recipients, and grade 3 wellness and swelling were reported in 0/30 (0%) recipients. For the 2-dose regimen, grade 3 pain was reported in 1/28 (3.6%) recipients, with grade 3 redness and swelling reported in 0/28 (0%) recipients.

^k. In the placebo group with 3 doses saline, no local reactions were reported (0 reports of pain, redness, and swelling among the 30 recipients).

^l. For the 3-dose regimen: grade 3 fatigue was reported in 4/30 (13.3%), gastrointestinal in 0/30 (0%), headache in 3/30 (10%), myalgia in 8/30 (26.7%), and fever in 1/30 (3.3%). For the 2-dose regimen, fatigue was reported in 1/28 (3.6%), gastrointestinal in 1/28 (3.6%), headache in 0/28 (0%), myalgia in 1/28 (3.6%), and fever in 0/28 (0%).

^m. In the placebo group with 3 doses saline, fatigue was reported in 1/30 (3.3%), and myalgia reported in 1/30 (3.3%) recipients. grade 3 gastrointestinal, headache, and fever were reported in 0/30 (0%) recipients.

^n. Local grade 3 reactions were reported in 13/112 (11.6%) RZV recipients. grade 3 pain was reported in 11/112 (9.8%), redness in 2/112 (1.8%), and swelling in 0/112 (0%) RZV recipients.

^o. Systemic grade 3 reactions were reported in 25/112 (22.3%) RZV recipients, with grade 3 fatigue reported in 16/112 (14.3%), gastrointestinal in 6/112 (5.4%), headache in 6/112 (5.4%), myalgia in 12/112 (10.7%), shivering in 6/112 (5.4%), and fever in 0/112 (0%).

^p. Systemic grade 3 reactions were reported in 17/110 (15.5%) placebo recipients, with grade 3 fatigue reported in 8/110 (7.3%), gastrointestinal in 7/110 (6.4%), headache in 3/110 (2.7%), myalgia in 4/110 (3.6%), shivering in 3/110 (2.7%), and fever in 0/110 (0%).

Table 4: Grade Summary of Findings Table

Abbreviations: AI = Autoimmune; CI = Confidence Interval; IC = Immunocompromised; IMD = Immune-Mediated Disease; HR = Hazard Ratio; HCT = Hematopoietic Cell Transplant; HZ = Herpes Zoster; RCT = Randomized Controlled Trial; RR = Risk Ratio; RZV = Recombinant Zoster Vaccine; VE = Vaccine Efficacy; VRR = Vaccine Response Rate

* RRs were calculated with Wald confidence intervals in R and in SAS.

** Incidence Rate Ratios (IRRs) were presented rather than VE, and VE was calculated using the formula VE = (100 * (1-IRR)).

^a. These 5 RCTs cover a wide range of populations: autologous HCT recipients, patients with HIV, patients with hematologic malignancies, and patients with pIMDs. Due to these different subgroups with different underlying conditions, this body of evidence covers some, but not all the populations being considered for the recommendation, resulting in a downgrade (-1) for indirectness, as these subpopulations may have different baseline risks, and the intervention may have a different effect in these subpopulations.

^b. Risk of bias related to blinding of participants and personnel may have been present. Although participants and study staff were blinded to intervention assignments, they may have inferred receipt of vaccine or placebo based on reactogenicity. This was deemed unlikely to overestimate efficacy or underestimate risk of serious adverse events, therefore the risk of bias was rated as not serious.

^c. Interpreting immunogenicity results for prevention of HZ faces a very serious (-2) downgrade for indirectness due to indirectness in two domains of the PICO question: population and outcome. For population, the included studies evaluate the immunogenicity of RZV in autologous HCT patients, patients with HIV, patients with hematological malignancies, patients with solid tumors, and renal transplant recipients. These different subgroups represent some, but not all the subgroups under consideration for the population for this recommendation. Additionally, there is inconsistency in using the proxy measure of immunogenicity to evaluate vaccine efficacy, or prevention of HZ, given that there are no established correlates of protection. Thus, this body of evidence is doubly downgraded.

^d. The two included cohort studies were both determined to have serious risk of bias, the first (using Medicare claims data) presents with significant confounding concerns, with no demographics or risk factors presented for the immune-compromised and autoimmune populations. The second was a VA-based study, which utilized retrospective case ascertainment and had a non-generalizable population, considered to be moderate/high risk of bias. However, this bias is already included in the initial downgrade of evidence to type 3 for non-randomized study data, thus there is no further need for a downgrade for risk of bias.

^e. The two cohort studies resulted in different VE estimates: with the first reporting VE in the autoimmune population of 68.0% (95% CI: 62.3% – 72.8%), and 64.1% (95% CI: 57.2% – 69.8%) in the immune-compromised population. The second study found no reported cases in the 50–60-year age group, corresponding to a VE of 100%, and in those aged >60 years, a HR of 0.41 (95% CI: 0.19-0.87) for those not using steroids, corresponding to an approximate VE of 59%, and for those using steroids: an HR of 0.34 (95% CI: 0.05-2.44), corresponding to a VE of 66%. While these results represent different age groups (65+ for the first study, and 50+ for the second study), the effect estimates are not very heterogeneous, and thus not downgraded for inconsistency.

^f. The two included cohort studies assessed incidence of HZ in different patient populations: the first in autoimmune/immunocompromised patients enrolled in Medicare, and the second in IBD patients in the VA Health System. These patient populations represent different subgroups of the overall population being considered for this recommendation and do not represent the entire range of conditions; thus the evidence is downgraded for indirectness (-1).

^g. Both studies reported a strong association, with the VE or 64.1% in IC patients and 68% in AI patients, and VE of 100% in 50–60-year-olds with IBD in the VA study, and VE of 59% in non-steroid users and 66% in steroid users over the age of 60. Thus, this body of evidence was upgraded (+1)  level due to strength of association.

^h. Across the 7 included RCTs (one of which was a pooled post hoc analysis of two RCTs (ZOE-50 and ZOE-70), among a subset of participants who reported at least one pIMD at enrollment), there are a wide range of included populations included: autologous HSCT recipients, patients with HIV, patients with hematologic malignancies, patients with pIMDs, patients with solid tumors receiving cytotoxic or immunosuppressive therapy, and renal transplant patients on daily immunosuppression. The wide variety of patient sub-populations results in a downgrade for indirectness (-1).

^i. This outcome has data from only a single study, and thus cannot downgrade for inconsistency.

^j. This outcome only has data in autologous HCT patients, which does not cover the entire recommended patient population for which a recommendation is being considered. Thus, this study is downgraded (-1) for indirectness.

^k. The total population among the RZV and placebo group in the modified total vaccinated cohort is 184, in which only 10 events of PHN were observed and 15 hospitalizations were observed. Given a corresponding relative risk reduction of 11% (VE 89%) with wide confidence intervals for PHN and 15% with wide confidence intervals for HZ-related hospitalization, this does not meet the threshold for the optimal information size and is thus downgraded (-1) for imprecision.

^l. These results were downgraded (-1) for inconsistency because the calculated risk ratio of pIMDs varied substantially by study: ranging from 0 to 2.00 for unspecified immune-mediated diseases, with additionally different values for exacerbations and new-onset pIMDs.

^m. These studies cover a wide range of populations: solid tumor patients, renal transplant recipients, hematologic malignancies, and patients with pIMDs originally enrolled into the ZOE-50 and ZOE-70 trials. Due to these different subgroups with different underlying conditions, this body of evidence covers some, but not all the populations being considered for the recommendation, resulting in a downgrade (-1) for indirectness

^n. These results were downgraded for imprecision (-1) due to wide confidence intervals crossing the threshold of both harm and benefit.

^o. This outcome only has data in patients with hematological malignancy, which does not cover the entire recommended patient population for which graft-versus-host-disease is a potential harm. Thus, this study is downgraded (-1) for indirectness.

^p. These results were doubly downgraded for imprecision (-2) due to not meeting the optimal information size, with only 7 total events occurring out of 40 total patients and wide confidence intervals crossing the harm/benefit threshold.

^q. This outcome only has data in renal transplant patients, which does not cover the entire patient population for which graft rejection is a potential harm (i.e. other solid organ transplants). Thus, this study is downgraded (-1) for indirectness.

^r. These results were downgraded for imprecision (-1) due to not meeting the optimal information size, and wide confidence intervals crossing the harm/benefit threshold.

^s. Solicited injection site reactions occurred in 14.2% of patients in Bastidas 2019, with solicited systemic reactions occurring in 13.2% of patients. The most common injection-site reaction was pain (11%) and the most common systemic reactions were fatigue (7.3%) and myalgia (6.2%). In Dagnew 2019, solicited injection site reactions occurred in 13.3% of patients, and systemic occurred in 15.5% of patients. In Stadtmauer 2014, local reactions occurred in 10% of patients, and systemic in 26.7% of patients (most commonly fatigue and myalgia). In Vink 2019, solicited local symptoms occurred in 11.6% of patients, most commonly pain (9.8%), and systemic symptoms in 22.3% of patients (most commonly fatigue (14.3%) and myalgia (10.7%). In Vink 2020, local grade 3 reactions occurred in 10.7% of patients, with systemic reactions occurring in 9.9% of patients. Overall, these results were quite consistent and thus were not downgraded for inconsistency, despite occurring across a broad set of population subgroups.

^t. These studies cover a wide range of populations: solid tumor patients, renal transplant recipients, hematologic malignancies, and autologous HCT recipients. Due to these different subgroups with different underlying conditions, this body of evidence covers some, but not all the populations being considered for the recommendation, resulting in a downgrade (-1) for indirectness.

Table 5: Summary of Evidence for Outcomes of Interest

^a. HCT = Hematopoietic cell transplant

^b. SOT = Solid organ transplant

Appendix 2. Studies Included in the Review of Evidence but not Included in GRADE Analysis

Abbreviations: BMS = Bristol Myers Squibb; CLL = Chronic Lymphoblastic Leukemia; HCT = Hematopoietic Cell Transplant; IBD = Inflammatory Bowel Disease; NIAID = National Institutes of Allergy and Infectious Disease; NIH = National Institutes of Health; NRES = Non-Randomized Experimental Study; RZV = Recombinant Zoster Vaccine

Appendix 3. Databases and Strategies Used for Systematic Review