Recombinant Zoster Vaccine Is Efficacious and Safe in Frail Individuals

Desmond Curran, PhD; Joon H. Kim, MD; Sean Matthews, MSc; Christophe Dessart, MSc; Myron J. Levin, MD; Lidia Oostvogels, MD; Megan E. Riley, PhD; Kenneth E. Schmader, MD; Anthony L. Cunningham, MD; Shelly A. McNeil, MD; Anne E. Schuind, MD; Melissa K. Andrew, MD


J Am Geriatr Soc. 2021;69(3):744-752. 

In This Article


Study Design and Participants

This ZOE-Frailty study (NCT03563183), was an international, observational, retrospective study designed to assess the baseline frailty status of participants in the ZOE-50 and the ZOE-70 studies.[15,16] These two parent phase III randomized, observer-blinded, placebo-controlled clinical trials were conducted concurrently at the same study sites using the same methods with participants aged ≥70 years randomly assigned to the ZOE-50 or ZOE-70 study. ZOE study participants belonging to sites willing to take part in the current study were included in the ZOE-Frailty study. In the ZOE studies, while patient reported outcomes (PRO) data were collected from all participants, encoding of PRO questionnaires (Short Form Survey-36 (SF-36), EuroQol-5 Dimension (EQ-5D)) was only performed for participants who developed a suspected HZ episode during the study. In the ZOE-Frailty study, we encoded the remaining PRO questionnaires. We linked this data with the data from the ZOE studies, which allowed us to assess baseline frailty status and perform the analysis of clinical outcomes as a function of frailty. The study was conducted in accordance with the principles of the Declaration of Helsinki and Good Clinical Practice guidelines.

In the parent studies, vaccine or placebo (0.9% saline solution) was administered (0.5 mL) intramuscularly at month 0 and month 2 with 1:1 randomization.[15,16]


The primary objective of the ZOE-Frailty study was to evaluate the baseline frailty status of participants in the parent ZOE-50 and ZOE-70 phase III trials. The secondary objectives included the evaluation of VE against HZ, VE against HZ burden of illness (BOI), humoral and cellular immunogenicity, vaccine reactogenicity, and safety by frailty status.


Frailty status was measured using the accumulation of deficits approach.[18–20] The different aspects of frailty composing the frailty index (FI) were assessed through the medical history and components of the SF-36 and EQ-5D questionnaires recorded before dose one, as previously validated.[21] The SF-36 is a multi-purpose health survey comprising 36 questions, including scales for physical functioning, role physical, bodily pain, general health, vitality, social function, role emotional, and mental health.[22] EQ-5D is a generic measure of health status that defines health in terms of mobility, self-care, usual activities, pain/discomfort, and anxiety/depression.[23] Further details on the deficits assessed and the scoring of the FI components are provided in the supplemental material (Supplementary File S1 and Supplementary Tables S1 and S2). Deficits were coded as 0 = absent to 1 = present. Each individual's deficits were summed to generate a total deficit score. The FI was then calculated by dividing by the number of possible deficits as follows: FI = (accumulation of deficits)/(41-nmissQoL), where nmissQoL was the number of missing components of the 29 items from the SF-36 and EQ-5D questionnaires. Each study participant was assigned to one of three subgroups based on the FI as follows: FI ≤0.08 is classified as non-frail; FI >0.08 to ≤0.25 is classified as pre-frail; FI >0.25 is classified as frail.[20] Participants with a missing FI were classified as unknown.

The Zoster Brief Pain Inventory (ZBPI) questionnaire severity of illness score was calculated as the area under the curve of the ZBPI worst pain score from day 0 until day 182.[24] The BOI was then estimated by aggregating the severity of illness scores over all the participants in a group and dividing by the total number of years of participant follow-up. Consequently, this composite measure took into account the incidence of HZ as well as the severity and duration of pain.[25] Details are presented elsewhere.[26]

The total vaccinated cohort (TVC) included all participants who received at least one dose of RZV or placebo. The primary cohort for efficacy analyses was the modified vaccinated cohort (mTVC), which excluded participants who did not receive the second dose or who had a confirmed HZ episode before 1 month post-dose two. VE against HZ was defined as 1 minus the ratio of HZ incidence of confirmed cases in the RZV group to that in the placebo group, multiplied by 100. The VE in reducing the BOI was similarly defined and calculated. All statistical tests were performed two tailed using a .05 significance level. All statistical analyses were performed with SAS software, version 4.7 (SAS Institute).

The analysis of humoral immunogenicity was performed based on the according-to-protocol cohort for immunogenicity, at each time point, including all participants who received both doses, met all the eligibility criteria, complied with the protocol, and had immunogenicity data available.[27] Serum anti-gE antibody concentrations were measured using a GSK in-house enzyme-linked immunosorbent assay (ELISA). gE-specific CMI responses were measured by flow cytometry to assess the frequency of CD4+ T cells expressing two or more of the following activation markers (hereafter termed CD42+): interferon-gamma (IFN-γ), interleukin-2 (IL-2), tumor necrosis factor-α (TNF-α), and CD40 ligand, following ex vivo stimulation with gE peptides. gE-specific CMI analysis was limited to a small subset of participants from the Czech Republic, Japan, and the United States (i.e., CMI subset, Supplementary Table S4). Details of this selection and of the immunologic assays are presented elsewhere.[27]

The humoral response threshold for the calculation of vaccine response rates (VRR) was defined as a fourfold or more increase in the anti-gE antibody concentration as compared to the pre-vaccination concentration (for initially seropositive participants) or as compared to the anti-gE antibody cut-off value for seropositivity (97 milli-International Units (mIU)/mL, for initially seronegative participants). The CMI-response threshold was defined as a twofold or more increase in the frequency of CD42+ T cells, as compared to pre-vaccination frequencies (for participants with pre-vaccination CD42+ T-cell frequencies above the cut-off of 320 positive cells per 10[6] CD4 T cells counted) or a twofold or more increase above the cut-off (for participants with pre-vaccination frequencies below the cut-off). Exact 95% CIs were computed at each time point for the percentage of humoral and CMI responders. Medians with interquartile ranges were calculated for CD42+ T-cell frequencies. The 95% CI for GMCs was computed by anti-log transformation of the 95% CI for the mean of log-transformed concentrations (which were calculated assuming that log-transformed values were normally distributed with unknown variance).

In the parent ZOE-50 and ZOE-70 studies, a randomly selected subgroup of age-stratified participants (i.e., TVC diary card subset) recorded injection-site reactions (pain, redness, and swelling) and systemic reactions (fatigue, fever, gastrointestinal symptoms, headache, myalgia, and shivering) on diary cards for 7 days after each injection. Depending on the severity, solicited adverse events (AEs) were graded from 0 to 3.[15,16] Unsolicited reports of AEs were recorded for 30 days after each dose for all participants. Serious AEs (SAEs) were recorded for all participants for 12 months after the second dose. Fatal AEs, vaccination-related SAEs and potential immune-mediated diseases (pIMDs) were recorded in all participants throughout the trial.