A Double-Blind, Randomized, Controlled, Multicenter Safety and Immunogenicity Study of a Refrigerator-Stable Formulation of Zostavax

Vaccine description.Zostavax (zoster vaccine live; Oka/Merck) is a single-dose, sterile, lyophilized, preservative-free, live attenuated virus vaccine (23, 27) manufactured by Merck & Co., Inc., West Point, PA. This study compared two Zostavax formulations: a refrigerator-stable formulation with phosphate-gelatin-sucrose-urea (PGSU) stabilizer and a frozen formulation with PGS stabilizer. The formulations were visually indistinct, supplied in identical glass vials, and shipped to the study sites frozen on dry ice. To maintain study blinding and generally equivalent potencies between the formulations throughout the duration of the study, both formulations were stored at −15°C (+5°F) or colder until they were reconstituted. Immediately preceding administration, both vaccines were reconstituted with 0.7 ml of sterile diluent. Each subject received a single ∼0.65-ml subcutaneous injection of either the refrigerated (PGSU) or the frozen (PGS) formulation, each at a potency of approximately 50,000 PFU/dose. Consistent with the intended storage and use of the PGSU- and PGS-containing formulations, the study vaccines are referred to henceforth as the Zostavax refrigerated and Zostavax frozen formulations, respectively.

Study design.This was a randomized, controlled, double-blind (with in-house blinding), multicenter study conducted in the United States to evaluate the safety, tolerability, and immunogenicity of the Zostavax refrigerated formulation performed between July and October 2005. Varicella history-positive adults were randomized in a 1:1 ratio to receive either the Zostavax refrigerated or the Zostavax frozen formulation. Enrollment was stratified in a 1:2 ratio by age (50 to 59 years and ≥60 years) in order to acquire safety and immunogenicity data with a broader age group of subjects. Vaccinations were administered on day 1. All subjects were monitored for exposure to varicella or HZ or the development of any varicella/varicella-like or HZ/HZ-like rashes, as well as any other adverse experiences (AEs). Injection-site reactions, rashes, and other AEs were recorded by each subject by the use of a vaccination report card (VRC) for 28 days postvaccination.

Blood samples were obtained from all subjects immediately before vaccination and at day 28 postvaccination and were assessed for VZV antibody titer by a glycoprotein enzyme-linked immunosorbent assay (gpELISA) (25, 35). The gpELISA is a validated, well-characterized assay that has successfully been used to compare formulations and to demonstrate lot consistency in pediatric varicella vaccine trials (4, 21, 25). In a substudy of the pivotal Shingles Prevention Study, the immune response elicited by a dose of Zostavax was evaluated by using the gpELISA and two cellular immunity assays (the gamma interferon enzyme-linked immunospot assay and responder cell frequency) (19, 23). The statistical analyses of the immunogenicity data in the substudy demonstrated that the results obtained by all three assays correlated with vaccine-induced protection against HZ; but the immune response measured by gpELISA, in terms of the postvaccination geometric mean titer (GMT) and the geometric mean rise (GMR) in the VZV antibody titer from the baseline to the period postvaccination, correlated best with protection against HZ (23). Therefore, this assay was chosen as the optimal tool for immunologic bridging within and between clinical studies of Zostavax, including a comparison of the VZV-specific immune responses to the two vaccine formulations evaluated in this clinical study.

Study population.Immunocompetent subjects 50 years of age and older with a history of varicella or residence in a country where VZV infection is endemic were eligible for the study. Subjects were excluded if they had a clinical history of hypersensitivity or anaphylactic reactions to gelatin or neomycin, used any form of nontopical antiviral therapy, had received a live vaccine within 4 weeks prior to the study dose or an inactivated vaccine within 1 week prior to the study dose, or another vaccination was planned before the subject was due to complete the study. Study exclusions also included a history of HZ, pregnancy, or breast-feeding; the plan to conceive within the duration of the study; known or suspected immune dysfunction; and alcohol or other substance abuse that might interfere with the evaluation required by the study.

Safety surveillance.The primary safety parameter of interest in the study was the frequency of vaccine-related serious AEs. All subjects were monitored for 28 days after vaccination for general safety and tolerability. During this period, subjects used a VRC to record AEs, such as fever, injection-site reactions, and systemic AEs. Each subject was asked to report immediately to the study site any varicella/varicella-like or HZ/HZ-like rashes and any serious AEs. The relationship between each AE and the study vaccination was assessed by the study investigator.

Immunogenicity measurements.The two coprimary hypotheses were that (i) the GMT of VZV antibodies at 28 days postvaccination in subjects who received the Zostavax refrigerated formulation would be noninferior to that in subjects who received the Zostavax frozen formulation and that (ii) the Zostavax refrigerated formulation would induce an acceptable rise in the VZV antibody titers from prevaccination to 28 days postvaccination. The coprimary endpoints were the GMTs of VZV antibodies in both vaccination groups at 28 days postvaccination and the GMR in VZV antibody titers from the baseline to the period postvaccination among the recipients of the Zostavax refrigerated formulation. The gpELISA methodology for measuring VZV antibody has been described elsewhere (25, 35).

Statistical methods.The primary and secondary immunogenicity analyses were based on the per protocol population. With 162 evaluable subjects (as planned) in each vaccination group and an estimated standard deviation of 1.1 for the natural log of the VZV gpELISA antibody titer, the overall power to claim that the study was a success with respect to the two primary immunogenicity hypotheses was 91.2% (91.3%·99.9%), assuming the independence of the two primary hypothesis tests. The first coprimary hypothesis was tested by estimating the GMT ratio (GMT of the Zostavax refrigerated formulation/GMT of the Zostavax frozen formulation) and its 95% confidence interval (CI) by using a longitudinal regression model with adjustment for the prevaccination antibody titer, age, and study center. The statistical criterion for success on the noninferiority hypothesis corresponded to a value for the lower bound of the two-sided 95% CI on the GMT ratio (GMT of the Zostavax refrigerated formulation/GMT of the Zostavax frozen formulation) of >0.67. The second coprimary hypothesis was tested by estimating the GMR in the Zostavax refrigerated group from prevaccination to day 28 postvaccination. The statistical criterion for success on the acceptability hypothesis corresponded to a value for the lower bound of the two-sided 95% CI on the GMR of >1.2-fold. The point estimate and its 95% CI were calculated by use of a t distribution. The secondary immunogenicity hypothesis was tested by estimating the GMR in the Zostavax frozen formulation group from prevaccination to day 28 postvaccination.

Subject characteristics and accounting.A total of 368 subjects were enrolled and randomized in the study, and 367 subjects were vaccinated; the data for 1 subject, who withdrew consent prior to the day 1 blood draw and who was not vaccinated, are not included in any of the following data analyses. Five additional subjects were screened but not randomized.

Table 1 provides details of the subject demographics across the two vaccination groups. The Zostavax refrigerated and Zostavax frozen groups were similar with regard to gender (female subjects represented 53% and 57% of the enrollment in the two groups, respectively) and mean age (63.4 and 63.2 years, respectively). In addition, the distributions of the subjects according to age (50 to 59, 60 to 69, and ≥70 years of age) and race were similar between the two groups.

Figure 1 illustrates an accounting of all subjects vaccinated in the study. A subject was considered to have completed the study if the subject received the study vaccine, completed the scheduled blood draws, and returned a completed VRC at the day 29 postvaccination visit. By the use of these criteria, 361 of the 368 subjects (98.1%) were considered to have completed the study. Among the seven subjects who discontinued the study, no safety follow-up was obtained for two subjects in each vaccination group. These subjects were excluded from all safety analyses. Two additional subjects withdrew their consent after vaccination. The numbers and percentages of subjects who discontinued the study and the reasons for discontinuation were generally similar between the two vaccination groups.

• Open in new tab
• Download powerpoint

FIG. 1.

Subject accounting.

Safety and tolerability.Table 2 provides a summary of the numbers and the percentages of clinical AEs reported from day 1 to day 28 postvaccination. One serious AE was reported in the study: a case of gastroenteritis in the Zostavax refrigerated group, which was determined by the investigator to be definitely not related to the study vaccine. In general, clinical AEs were reported at a lower rate by the recipients of the Zostavax refrigerated formulation than by the recipients of the Zostavax frozen formulation. The majority of both injection-site AEs and systemic AEs were rated by the subjects as mild in intensity. The most frequently reported injection-site AEs (≥10% in both vaccination groups) were erythema, pain, and swelling, all of which were reported at lower frequencies by subjects who received the Zostavax refrigerated formulation. The incidences of systemic clinical AEs were similar in both vaccination groups, with ∼6% determined to be vaccine related in either vaccination group. One non-injection-site varicella-like rash with three lesions was reported by one subject in the Zostavax refrigerated group. No subject discontinued the study due to an AE.

Immunogenicity.Table 3 presents the VZV antibody response for each age stratum and for the two strata combined, by vaccination group. For both age strata, the GMT increased substantially from prevaccination to day 28 postvaccination in both the Zostavax refrigerated and the Zostavax frozen formulation vaccination groups (GMRs, 2.6- and 2.9-fold for the two formulations, respectively, for the age groups combined). A statistical analysis for the comparison of the VZV antibody titers in both vaccination groups at day 28 postvaccination is as follows. By use of the GMT as the endpoint, the estimated GMTs for the Zostavax refrigerated group (n = 182) and the Zostavax frozen group (n = 185) were 727 and 834, respectively. The estimated GMT ratio (Zostavax refrigerated formulation/Zostavax frozen formulation) was 0.87 (95% CI, 0.71 to 1.07; P = 0.005), which meets the prespecified criterion that the VZV antibody response induced by the Zostavax refrigerated formulation is similar to that observed with the Zostavax frozen formulation. The estimated responses, GMT ratios, 95% CIs, and one-sided P value for the testing of noninferiority (GMT ratio [refrigerated formulation/frozen formulation], >0.67) were computed on the basis of a longitudinal regression model adjusted for prevaccination VZV antibody titers, age (years), and study center. The P value for the testing of the treatment-by-age-stratum interaction was 0.496. The P value for the testing of the treatment-by-center interaction was 0.656. A lower bound of the 95% CI on the difference (GMT ratio) that excludes a 1.5-fold decrease implies that the difference is statistically significantly less than the prespecified clinically relevant decrease of 1.5-fold and allows a conclusion of similarity (noninferiority). The P value for the testing of noninferiority (≤0.025) also supports this conclusion.

The results of the acceptability analysis performed on the basis of the VZV antibody responses in the Zostavax refrigerated and the Zostavax frozen vaccination groups at day 28 postvaccination are as follows. By use of the GMR as the endpoint, the estimated GMRs of the VZV antibody response for the Zostavax refrigerated group (n = 174) and the Zostavax frozen group (n = 177) were 2.6-fold (95% CI, 2.2- to 3.0-fold; P < 0.001) and 2.9-fold (95% CI, 2.4- to 3.4-fold; P < 0.001), respectively. The one-sided P value for the testing of the acceptability for GMR (>1.2-fold) was computed on the basis of the t test. The 95% CIs were also computed on the basis of the t distribution. Since the lower bound of the 95% CI was >1.2 in each case and the one-sided P value for the testing of the acceptability hypothesis (GMR, >1.2-fold) was <0.025 for both vaccination groups, the VZV antibody responses induced by both Zostavax formulations were found to be acceptable.

Figure 2 shows the reverse cumulative distribution of the VZV-specific antibody titer at day 1 and day 28 for each vaccination group. The data presented in Fig. 2 are consistent with the results of the noninferiority analysis described above, indicating that the Zostavax refrigerated formulation induced a response in VZV antibodies at 28 days postvaccination similar to that of the Zostavax frozen formulation.

• Open in new tab
• Download powerpoint

FIG. 2.

Reverse cumulative distribution of VZV-specific antibody titers at prevaccination and 4 weeks postvaccination by vaccination group.