SUMMARY
Background
rVSVΔG-ZEBOV-GP is the first approved vaccine with demonstrated clinical efficacy against Ebola virus disease. Although a seroprotective threshold has not been defined for those at occupational risk of exposure, the current vaccine strategy is to attain a sustained high level of antibody titers. The goal of this trial was to explore the effects of delayed boosting upon both the height and duration of antibody titers following primary immunization.
Methods
In this randomized, controlled, open-label phase 2 trial, we compared antibody titers at Month 36 in those who had received a homologous booster dose at Month 18 following primary immunization with those who had received no booster. During the period of 25 October 2016 through 29 January 2020, healthy adults ≥18 years old deemed at occupational risk of exposure to Ebola virus due to laboratory work, clinical duties, or travel to an active endemic region were recruited from four U.S. and one Canadian hospital clinics and received primary vaccination with 2×107 PFU/mL VSVΔG-ZEBOV-GP. Eighteen months later, individuals who consented and were still eligible were randomized 1:1 to receive either a homologous booster dose or no booster. Study visits for safety and serial blood collections for antibody titers were conducted on enrolled participants at Months 0, 1, 3, 6, 12, 18, 19, 24, 30, and 36. Through July 2021, a web-based application was used to perform randomization, including assignments with schedules for each of the five sites using mixed permuted blocks. The trial was not blinded to participants or site staff. The primary endpoint was a comparison of geometric mean titers (GMTs) of anti-Ebola virus glycoprotein immunoglobulin G antibody at Month 36 (i.e., 18 months after randomization) for all randomized participants who completed the 36 months of follow-up (primary analysis cohort). Investigators were aware of antibody titers from baseline (enrollment) through Month 18 but were blinded to summary data by randomization group after Month 18.
Trial registration number:
Findings
Of the 248 participants who enrolled and received their primary immunization, ultimately 114 proceeded to the randomization step at Month 18. The two randomization groups were balanced: 57 participants (median age 42, interquartile range [IQR] 35–50, 24 [42%] females) were randomized to the booster group and 57 (median age 42, IQR 36–51, 24 [42%] females) to the no-booster group. Of those randomized, 92 participants (45 in the booster group and 47 in the no-booster group) completed 36 months of follow-up. At 18 months after primary immunization, GMTs in the no-booster group increased from a baseline of 10 ELISA units (EU)/mL to 1,451 EU/mL; GMTs in the booster group increased from 9 to 1,769 EU/mL. At Month 19, GMTs were 31, 408 and 1,406 EU/mL for the booster and no-booster group; at Month 36, GMTs were 10,146 and 1,240 EU/mL for the booster and no-booster groups, respectively. Accordingly, the geometric mean ratio (GMR) of antibody titers had increased almost 21-fold more in the booster versus no-booster group at one month after booster administration (GMR=20·6; 95% confidence interval [CI]: 18·2–23·0; p<0·0001) and was still over seven-fold higher at Month 36 (GMR=7·8; 95% CI: 5·5–10·2; p<0·0001). Consistent with previous reports of this vaccine’s side effects, transient mono- or oligo-articular arthritis was diagnosed in 18 of 207 (9%) primary vaccination recipients; after randomization, arthritis was diagnosed in one participant in the no-booster group (2%). No new cases of arthritis developed after booster administration. Four serious adverse events (SAEs) occurred following randomization: one (2%; epistaxis) in the booster group and three (6%; gastrointestinal hemorrhage, prostate cancer, and tachyarrhythmia) in the no-booster group. None of the SAEs was judged attributable to the booster vaccination assignment.
Interpretation
In addition to no new safety concerns and in marked contrast to earlier trials evaluating short-term boosting, delaying a rVSVΔG-ZEBOV-GP booster until Month 18 resulted in an increase in GMT that remained several-fold above the no-booster group GMT for at least 18 months. These findings may have implications in the optimal timing of booster doses as pre-exposure prophylaxis in populations at ongoing risk for Ebola virus exposure.
INTRODUCTION
The first vaccine demonstrated to prevent human Ebola virus disease (EVD), rVSVΔG-ZEBOV-GP, received initial regulatory approval in the U.S. and Europe in 2019.1,2 rVSVΔG-ZEBOV-GP uses an attenuated replication-competent recombinant vesicular stomatitis Indiana virus (VSIV) that encodes the Ebola virus (EBOV) glycoprotein (GP) instead of its own glycoprotein (G).3 Using a single 2×107 PFU/mL dose, clinical efficacy of this vaccine was first documented in an open-label cluster-randomized ring vaccination trial performed in Guinea and Sierra Leone during the large-scale 2013–16 Western Africa EVD outbreak.4 Most other studies of rVSVΔG-ZEBOV-GP have focused on the humoral response to vaccination. For example, in a 2014 phase 1 dose-ranging study of this vaccine involving 78 healthy adults, participants who received a homologous booster dose 28 days after primary vaccination had a small increase in anti-EBOV GP antibody titers at Day 56 compared to those who did not receive a booster, but this effect was substantially diminished by 6 months.5 This outcome suggested that rechallenging a recently primed immune system with additional antigen may only result in limited gains of limited duration. Subsequently, in 2018, a randomized placebo-controlled EVD vaccine trial of rVSVΔG-ZEBOV-GP performed at six sites in Western Africa showed a similar decline of anti-EBOV GP antibody titers by 6 months following a booster given only 56 days after primary vaccination.6
To date, a seroprotective threshold for rVSVΔG-ZEBOV-GP has not been established in clinical studies, raising the question of what antibody titer after primary vaccination should be considered adequate for those at risk. A related concern of healthcare and laboratory staff at risk for occupational exposure to EBOV is whether the peak humoral response induced by a single dose of rVSVΔG-ZEBOV-GP vaccine given as pre-exposure prophylaxis (PREP) is durable and, if so, for how long. Adopting a different boosting strategy from that used in previous trials, we conducted a randomized clinical trial of an rVSVΔG-ZEBOV-GP prime–boost approach when boosting is delayed for 18 months. Here we report the results and assessments from this study, including immunogenicity, durability, and safety.
METHODS
Trial design and participants
“A Multicenter Study of the Immunogenicity of Recombinant Vesicular Stomatitis Vaccine for Ebola-Zaire (rVSVΔG-ZEBOV-GP) for Pre-Exposure Prophylaxis (PREP) in Individuals at Potential Occupational Risk for Ebola Virus Exposure (PREPARE)” was a phase 2 multicenter randomized clinical trial characterizing the effects of immunization with rVSVΔG-ZEBOV-GP (Merck & Co., Rahway, New Jersey) as PREP in healthy adults ≥18 years old deemed by site investigators to be at occupational risk for exposure to EBOV (see risk categories in Supplemental Material, page 3). All participants were immunized intramuscularly with 2×107 PFU/mL of rVSVΔG-ZEBOV-GP at study start (Month 0), followed by 1:1 randomization at Month 18 to receive either a homologous booster dose or no booster dose.
Eligibility criteria (complete listing in Supplemental Material, pages 58 and 84–85) for prior to primary vaccination (Month 0) and for prior to randomization to the booster or no-booster groups at Month 18 were similar. Briefly, adults ≥18 years old had to agree to provide signed informed consent, accept randomization to either study group, use effective contraception during the peri-immunization period, minimize body fluid exposure to household contacts for 14 days after immunization, and not receive another investigational vaccine during that period. Major exclusion criteria were also minimal (see Supplemental Material, page 59).
Prior to unblinding of the randomization groups, a decision was made to exclude 13 enrolled participants, 10 of whom were later randomized to the booster or no-booster group, who had self-reported that they had received an investigational EVD vaccine prior to enrollment in PREPARE (Supplemental Material, page 3).
Randomization and masking
For those eligible, randomization was performed by the Division of Biostatistics at the University of Minnesota Statistical Data Management Center with an interactive web-based application using a computer-generated random sequence (mixed permuted blocks) within strata defined by five clinical sites.
Randomization was stratified for the five participating sites: 1) Hope Clinic of Emory Vaccine Center, Emory University, Atlanta, GA, USA; 2) Public Health Agency of Canada, Children’s Hospital Research Institute, Winnipeg, MB, Canada; 3) National Institutes of Health Clinical Research Center, Bethesda, MD, USA; 4) University of Texas Medical Branch in Galveston, TX, USA; and 5) University of Nebraska Medical Center, Omaha, NE, USA. Randomization was carried out at Month 18 rather than at the time of the primary vaccination both to re-assess medical clearance to receive a second inoculation and to increase the likelihood that participants received their assigned intervention. Study participants and investigators at the clinical sites were not blinded to receipt of the booster (i.e., no placebo booster was used). Investigators were aware of antibody titers from baseline (enrollment) through Month 18 but were blinded to summary data by randomization group after Month 18.
Procedures
The study protocol was approved by each site’s institutional review board. After signed informed consent was obtained, demographics (including age, sex [male/female at birth], and occupational risk category) and medical histories were reviewed with participants, physical examinations performed, and blood collected for anti-EBOV GP antibody titer measurements. Assigned immunizations were all administered at each site, with the exception of participants in TX and NE who received all immunizations at the National Institutes of Health but had follow-up visits conducted locally. Targeted safety assessments and serial blood collections were repeated at 1, 3, 6, 12, 18, 19, 24, 30, and 36 months. Eighteen months after the primary vaccination, protocol requirements were re-reviewed with all participants and eligibility for randomization to receive the booster was re-assessed.
Primary and secondary outcomes
Anti-EBOV GP antibody measurement
Serial serum titers of immunoglobulin G (IgG) binding antibodies against EBOV GP were measured in batch at the U.S. National Institutes of Health (NIH) National Institute of Allergy and Infectious Diseases (NIAID) Integrated Research Facility at Fort Detrick (IRF-Frederick) in Frederick, MD, using the Filovirus Animal Nonclinical Group (FANG) enzyme-linked immunosorbent assay (ELISA) by staff blinded to the randomization assignment.7,8 The primary endpoint was initially defined as the difference at Month 36 in the number of participants with a ≥ four-fold increase in anti-EBOV GP antibody titers above baseline in each study group. However, early analyses of titers of pre-randomization antibody levels suggested that all participants would meet this endpoint, and the primary endpoint was revised by blinded investigators in September 2020 (Protocol Version 11·2) to be the comparison in geometric mean titers (GMTs) between the two randomized groups at Month 36.
Antibody binding kinetics to recombinant EBOV GP
As a secondary immunogenicity outcome to be performed in the first 25 participants in whom a full 36 months of samples were collected, antibody binding kinetics to recombinant EBOV GP were also assessed by surface plasmon resonance (SPR), blinded to the treatment assignment, as previously described (Supplemental Material, page 6).9 Unlike the FANG assay, assessment of SPR measures total combined antibody binding of all isotypes (IgM+IgG+IgA) present in sera. SPR was performed at the Division of Viral products, Center for Biologics Evaluation and Research (CBER), Food and Drug Administration (FDA), White Oak, MD.
Safety assessments
In addition to the two immunogenicity outcomes, safety outcomes, including serious adverse events, were collected throughout the follow-up period. After primary vaccinations and booster administrations, all recipients were observed for 30 minutes to assess pain and injection-site reactions. Briefly, the presence and severity of nine targeted symptoms and the presence of joint problems after vaccine administration reported by participants were recorded at the Month 1 visit and at the Month 19 visit for booster recipients (Supplemental Material, pages 3–5). A REDCap (Version 6.10.14) diary was used by participants as a mnemonic aid. Diagnoses of arthritis, other joint problems, serious adverse events, and targeted symptoms were collected and recorded at each scheduled visit through Month 36.
Statistical analysis
Sample size
The planned sample size for the primary endpoint at Month 36 was 200 randomized participants, 100 in the booster group and 100 in the no-booster group. The sample size provided 80% power at the 0·05 (two-sided) level of significance to detect a 50% higher anti-EBOV GP antibody titer level at 36 months in the booster compared to the no-booster group. By June 2022, however, it became clear that as a result of travel restrictions amid the coronavirus disease (COVID-19) pandemic and staff deployments, the sample size of 200 would not be achieved; only four participants were randomized from July 2021 through June 2022. Sample size was re-estimated by investigators who were blinded to treatment comparisons after Month 18 (Supplemental Material, pages 6–7). Anti-EBOV GP antibody measurements for both treatment groups were used to estimate the standard deviation of the Month 36 antibody titers. This standard deviation was 0·54 ELISA units (EU)/mL on the log10 scale after adjustment for the Month 18 (prior to booster randomization) titer level. With a standard deviation of 0·54, a difference between treatment groups of 0·278 on the log10 scale (1·9 fold) could be detected with 80% power at the 0·05 (2-sided) level of significance with 120 participants (60 per group) (Supplemental Material, page 6). Based on this re-estimate of sample size and in order to complete the study in a timely manner, it was decided to limit the primary analysis to those randomized through July 2021 and complete trial follow up by the end of 2022. This excluded 14 enrolled participants who had not reached the randomization stage by end of July 2021. Follow up was completed for the trial on 05 January 2023.
Primary analysis cohort
The analysis cohort for the primary endpoint included all participants who: 1) met the eligibility criteria, including not having received prior immunization with investigational EVD vaccines prior to enrollment, and consented to enrollment at Day 0 and to randomization at Month 18; 2) were randomized to the booster or no-booster group and received a booster if assigned; and 3) attended the Month 36 visit. Safety comparisons for the randomized groups included all participants who met conditions 1 and 2 (modified intention-to-treat [ITT] cohort).
Comparisons of booster and no-booster randomized groups
For the primary analysis cohort, GMTs of anti-EBOV GP IgG antibody and geometric mean ratios (GMRs, i.e., ratio of GMTs at each timepoint for the booster group versus the no-booster group) for the randomized participants were compared using analysis of covariance stratified by site and with the pre-randomization Month 18 antibody titer after log10 transformation as a covariate. This planned analysis for comparing the booster and no-booster groups was supplemented with a stratified analysis of variance which excluded the Month 18 antibody titer as a covariate (carried out because there was evidence of lack of parallelism of the regression lines for Month 36 level on Month 18 titer level for the booster and no-booster groups). The same methods were used to compare the booster and no-booster groups for antibody binding kinetics. Fold-change in titers at each timepoint for the booster and no-booster groups were compared using chi-squared tests. Additional descriptions of these analyses are given in the Supplemental Material, page 7.
Two-sided p-values and 95% confidence intervals (CIs) are cited. A two-sided p-value <0·05 was considered significant.
Protocol-defined subgroup analyses were performed by age, sex, occupational risk, and clinical site for the primary immunogenicity endpoint at Month 36. In addition, since the primary analysis indicated heterogeneity of regression slopes for Month 36 on the Month 18 titer level, a subgroup analysis by tertile of Month 18 titer was also carried out. For each subgroup, an analysis of variance model was expanded to include the subgrouping variable and an interaction term between booster assignment and the subgrouping variable. Subgroups analyses should be interpreted cautiously, since there was no adjustment for type 1 error, and interaction tests have low power. Chi-square tests were used to compare the booster and no-booster groups for targeted symptoms, joint problems, diagnoses of arthritis, and serious adverse events reported after Month 18.
Supplementary analyses
To assess possible differences in baseline characteristics, adverse events and titer levels after the primary vaccination in those randomized were compared to those who were not. Predictors of randomization among those enrolled receiving the primary vaccination were summarized with a logistic regression model.
To assess the potential impact of missing data after randomization on the estimated treatment difference at 36 months, analyses (using Student’s t-test and chi-square tests) were carried out comparing the characteristics of randomized participants who attended the Month 36 visit with those who did not. Predictors of missing the Month 36 visit were summarized with logistic regression models (see Supplemental Material page 7).
Statistical analyses were performed with the use of SAS software, version 9·4 (SAS Institute).
Role of the funding sources
Staff and investigators funded directly or indirectly by the cited government agencies (e.g., NIAID, FDA, etc.) or contractors were responsible for the study design and participated in the collection, analysis, and interpretation of data, and writing of the report. Merck & Co. supplied investigational vaccine for the trial but otherwise had no role in study design, data collection or analysis, or interpretation or reporting of outcomes.
RESULTS
During the period 25 October 2016 – 29 January 2020, a total of 221 eligible consenting participants were enrolled. After enrollment and receipt of the primary vaccination, four participants withdrew consent, nine declined follow up, and 54 were not screened for randomization at Month 18. A total of 154 participants attended the Month 18 visit. Of the 154 participants who were screened at the Month 18 randomization visit, 19 declined randomization but met all other eligibility criteria, four declined randomization and did not meet at least one other eligibility criterion, and 17 were willing to be randomized but did not meet at least one other eligibility criterion, leaving a total of 114 participants who were eligible and also agreed to randomization and assignment to the ITT cohort (57 to the booster group and 57 to the no-booster group) (Figure 1). Three participants randomized to the booster group did not receive the booster dose due to COVID-19 travel restrictions. At least one post-Month 18 measurement of antibody titers was conducted for 108 of the 114 randomized participants (modified ITT cohort). Ultimately, the primary analysis cohort included 92 participants: 45 in the booster group and 47 in the no-booster group.
Figure 1:
CONSORT diagram of the enrollment and screening, randomization, and follow-up.
Table 1 and Table S1 summarize characteristics at the time of enrollment (Day 0) for participants in the primary analysis cohort (92 participants) and their pre-study medical histories, respectively. The characteristics of the two groups were balanced at baseline. Demographic and occupational risk characteristics at enrollment for all 114 randomized participants (the ITT cohort) as a whole (Table S2) were similar to those in the primary analysis cohort. Table S3 shows baseline characteristics by site for the ITT cohort.
Table 1.
Demographics at the Time of Enrollment by Randomization Group: Primary Analysis Cohort
| Randomization Group | |||
|---|---|---|---|
| Site1 |
Booster
N (%) |
No booster
N (%) |
Total N (%) |
| National Institutes of Health (NIH) Clinical Center | 13 (29%) | 12 (26%) | 25 (27%) |
| Emory University Hospital | 17 (38%) | 20 (43%) | 37 (40%) |
| University of Texas Medical Branch | 1 (2%) | 1 (2%) | 2 (2%) |
| Children’s Hospital Research Institute of Manitoba | 14 (31%) | 14 (30%) | 28 (30%) |
| Age (years) | |||
| Mean (S.D.2) | 44 (10) | 44 (12) | 44 (12) |
| Median (IQR3) | 42 (37, 51) | 43 (36, 52) | 43 (36, 52) |
| Male | 26 (58%) | 27 (57%) | 53 (58%) |
| Female | 19 (42%) | 20 (43%) | 39 (42%) |
| Race | |||
| Asian | 6 (13%) | 1 (2%) | 7 (8%) |
| Black | 1 (2%) | 2 (4%) | 3 (3%) |
| Hispanic/Latino | 2 (4%) | 1 (2%) | 3 (3%) |
| White | 36 (80%) | 43 (91%) | 79 (86%) |
| Other | 0 (0%) | 0 (0%) | 0 (0%) |
| Occupational risk 4 | |||
| Will travel to country with potential risk of Ebola exposure | 31 (69%) | 28 (60%) | 59 (64%) |
| Laboratory worker who may have exposure to Ebola | 16 (36%) | 19 (40%) | 35 (38%) |
| Potential exposure to Ebola patient | 13 (29%) | 16 (34%) | 29 (32%) |
| Other5 | 0 (0%) | 1 (2%) | 1 (1%) |
| No. of participants | 45 | 47 | 92 |
Excludes three participants randomized at the University of Nebraska Medical Center, two who were randomized to the booster group who did not receive the booster, and one randomized to the no-booster group who did not have a Month 36 titer.
Standard deviation.
Interquartile range.
Participants could report more than one risk group.
Animal source exposure.
Table 2 summarizes adverse events, including serious adverse events (SAEs) and arthritis diagnoses by randomized group after each vaccination. Although judged unrelated to vaccination by site investigators, four participants had SAEs in the 18 months after the booster vaccination: one (epistaxis) in the booster group (2%) and three (gastrointestinal hemorrhage, prostate cancer, and tachyarrhythmia) in the no-booster group (6%). In the modified ITT cohort, only one (2%) participant in the booster group and two (4%) participants in the no-booster group experienced a grade-3 or -4 event (Month 19).
Table 2.
Targeted Symptoms and Adverse Events Occurring Between the Prime Vaccination And Month 1 and Between Randomization and Month 19 And Serious Adverse Events and Arthritis Prior to And After Randomization, by Randomization Group: Modified ITT Cohort
| Month 11 | Month 192 | ||||
|---|---|---|---|---|---|
| Booster3 n (%) |
No booster4 n (%) |
Booster3 n (%) |
No booster5 n (%) |
p-value6 | |
| No. attending visit | 52 | 55 | 52 | 50 | |
| Symptoms | |||||
| Fatigue7 | 42 (81%) | 47 (85%) | 35 (67%) | 3 (6%) | < .0001 |
| Myalgia8 | 44 (85%) | 45 (82%) | 30 (58%) | 1 (2%) | < .0001 |
| Headache9 | 42 (81%) | 40 (73%) | 22 (42%) | 5 (10%) | 0.00022 |
| Nausea/vomiting | 9 (17%) | 12 (22%) | 6 (12%) | 0 (0%) | 0.013 |
| Abnormal sweating | 11 (21%) | 17 (31%) | 0 (0%) | 0 (0%) | |
| Rash | 4 (8%) | 3 (5%) | 0 (0%) | 0 (0%) | |
| Mouth ulcers | 7 (13%) | 6 (11%) | 3 (6%) | 0 (0%) | 0.085 |
| Unexplained bleeding or bruising | 0 (0%) | 3 (5%) | 0 (0%) | 0 (0%) | |
| Joint pain | 21 (40%) | 19 (35%) | 9 (17%) | 4 (8%) | 0.16 |
| Joint problems10 | 21 (40%) | 19 (35%) | 10 (19%) | 4 (8%) | 0.10 |
| Grade 3 or 4 adverse events (symptoms or events not previously reported) 11 | 6 (12%) | 9 (16%) | 1 (2%) | 2 (4%) | 0.53 |
| Through Randomization | After Randomization | ||||
| Booster12 n (%) |
No booster13 n (%) |
Booster14 n (%) |
No booster15 n (%) |
||
| No. with at least one visit | 54 | 57 | 54 | 54 | |
| Serious adverse event 16 | 0 (0%) | 0 (0%) | 1 (2%) | 3 (6%) | |
| Arthritis | 5 (9%) | 2 (4%) | 0 (0%) | 1 (2%) | |
For symptoms and events occurring between the primary vaccination and Month 1.
For symptoms and events occurring between randomization and Month 19.
Includes 52 of 54 randomized participants who received the boost and who attended the indicated follow-up visit.
Includes 55 of 57 participants randomized to the no-boost group and who attended the indicated follow-up visit.
Includes 50 of 57 participants randomized to the no-boost group and who attended the indicated follow-up visit.
From chi-square test, for the difference between randomization groups in the number with each symptom or event at Month 19.
At Month 1, includes 6 in the booster group and 8 in the no booster group that experienced grade 3 fatigue.
At Month 19, includes 1 in the booster group and 1 in the no booster group that experienced grade 3 fatigue.
At Month 1, includes 3 in the booster group and 5 in the no booster group that experienced grade 3 myalgia.
At Month 19, includes 0 in the booster group and 1 in the no booster group that experienced grade 3 myalgia.
At Month 1, includes 4 in the booster group and 2 in the no booster group that experienced grade 3 headache.
At Month 19, includes 0 in the booster group and 1 in the no booster group that experienced grade 3 headache.
Pain/tenderness, swelling, stiffness, or redness/warmth.
For Month 1, 5 of 6 participants in the booster group and 6 of 9 participants in the no booster group had more than one event.
For Month 19, 0 of 1 participants in the booster group and 1 of 2 participants in the no booster group had more than one event.
At Month 19, one event in the no booster group was Grade 4 chronic eosinophilic pneumonia; all other events at either visit were Grade 3.
Includes all 54 randomized participants who received the boost and who attended at least one visit through Month 18.
Includes all 57 participants randomized to the no-boost group and who attended at least one visit through Month 18.
Includes 54 of 57 randomized participants who received the boost and who attended at least one visit after Month 18.
Includes 54 of 57 participants randomized to the no-boost group and who attended at least one visit after Month 18.
In the boost group, the event was epistaxis, and the event was judged to be unlikely related to study treatment.
In the no-boost group, the events were gastrointestinal hemmorrhage, prostate cancer, and tachyarrhythmia, each occurring in a different participant, and all events were judged to be not related to the study treatment.
During the 18 months after the primary vaccination, a diagnosis of transient arthritis was reported, with onset typically within 14 days after immunization, by seven (6·3%) participants, five (9%) of whom were later randomized to the booster group and two (4%) to the no-booster group. In contrast, no participants developed arthritis after the booster administration at Month 18. Fatigue, myalgia, headache, and nausea/vomiting were significantly more common in the booster group compared to the no-booster group at Month 19. Most symptoms after booster vaccination did not reach grade 3 or 4: exceptions were one (2%) participant in the booster group and two (4%) in the no-booster group. In the booster group, the percentage of participants reporting symptoms was greater after the primary vaccination than after booster administration (Table 2).
The median pain score was 5 (interquartile range 3–6) during the primary vaccination for participants who were ultimately randomized. The pain score declined to 0 (interquartile range 0–1) 30 minutes later. The median pain scores during and 30 minutes after booster administration were similar to those observed after primary vaccination—6 (interquartile range 4–7) and 0 (interquartile range 0–1), respectively (Table S4). The frequency of injection-site reactions after the primary and booster vaccinations was also similar. The most common reported injection-site reaction was pain (52 [46%] of participants after the primary vaccination and 26 [48%] after the booster administration).
Factors associated with not being randomized after the primary vaccination were assessed in univariate and multivariate analyses (Tables S5–S9). Pain levels following primary immunization and targeted symptoms, SAEs, and arthritis events between randomized and non-randomized participants are shown in Tables S6 and S7, respectively. In the multivariate analysis, non-white race, pain score 3 or higher within 30 minutes after prime vaccination, and development of arthritis or an SAE during the 18 months after primary vaccination were significantly associated with not being randomized (Table S9).
Primary endpoint
Follow up for the primary endpoint at 36 months was similar for the booster and no-booster groups. Among the 114 randomized participants (ITT cohort), 54 in both the booster and no-booster groups (95%) received their randomized assignment and were measured at least once for anti-EBOV GP IgG titers during the 18 months after randomization (Figure 1). Of these participants, 45 (83%) in the booster group and 47 (87%) in the no-booster group were measured for titers at Month 36.
In supplementary analyses that investigated factors associated with missing Month 36 titer measurements, only age emerged as a significant predictor (Tables S10–S13).
For the 92 participants in the primary analysis cohort anti-EBOV GP IgG titers increased 21-fold more in the booster group compared to the no-booster group one month after randomization (Month 19) (GMR=20·6; 95% CI: 18·2–23·0; p<0·0001) (Figure 2, Table 3). While the difference between the groups declined between Month 19 and Month 36, GMTs for the booster group remained much higher (10,146 EU/mL) than in the no-booster group (1,240 EU/mL) at Month 36 (GMR=7·8; 95% CI: 5·5–10·2; p<0·0001). For this analysis, which considered the Month 18 titer as a covariate, there was evidence of a different slope for the regression of 36 Month titer levels on Month 18 titer levels (p<0·0001) (Figure S1).
Figure 2.
Plot of geometric mean titers (EU/mL) of anti-EBOV GP IgG ELISA by randomization group and visit. Shown are the geometric mean antibody titers (in EU/mL, with 95% confidence intervals) of the primary analysis cohort over the 36 months of study participation. The no-booster group is depicted in blue and the booster group is shown in orange.
Table 3.
Geometric Mean Titer (EU/mL) of anti-Gp IgG ELISA through Month 36 By Randomization Group: Primary Analysis Cohort
| Randomization group | ||||||
|---|---|---|---|---|---|---|
| Booster | No booster | |||||
| Visit | n | Geometric mean (95% C.I)1 | n | Geometric mean (95% C.I)1 | Geometric ratio (95% CI)1,2 | p-value 2 |
| Baseline | 45 | 9 (6 – 16) | 47 | 10 (7 – 14) | ||
| Month 1 | 42 | 1949 (1502 – 2527) | 45 | 2034 (1588 – 2605) | ||
| Month 3 | 42 | 2423 (1883 – 3116) | 47 | 2480 (2067 – 2976) | ||
| Month 6 | 43 | 2022 (1561 – 2620) | 43 | 2036 (1559 – 2658) | ||
| Month 12 | 40 | 1865 (1306 – 2665) | 40 | 1527 (1140 – 2047) | ||
| Month 18 | 45 | 1769 (1348 – 2321) | 46 | 1451 (1118 – 1882) | ||
| Month 19 | 41 | 31408 (23181 – 42554) | 44 | 1406 (1078 – 1833) | 20.6 (18.2 – 23.0) | < 0.0001 |
| Month 24 | 38 | 17349 (13282 – 22661) | 38 | 1250 (927 – 1686) | 12.3 (9.9 – 14.7) | < 0.0001 |
| Month 30 | 43 | 11110 (8485 – 14547) | 44 | 1222 (927 – 1611) | 8.7 (6.3 – 11.1) | < 0.0001 |
| Month 36 (primary endpoint) | 45 | 10146 (7960 – 12933) | 47 | 1240 (984 – 1563) | 7.8 (5.5 – 10.2) | < 0.0001 |
Confidence Interval.
From the analysis of covariance, stratified by site and adjusted for the most recent log10 titer through Month 18.
One participant in the no boost group did not have a Month 18 titer, and so their Month 12 titer value was imputed for their Month 18 titer value.
Similar antibody differences between the booster and no-booster groups were found using all available data in the linear mixed-effects regression model for the 111 participants in the modified ITT cohort for whom at least one measurement of anti-EBOV GP IgG titers was conducted during the 18 months after randomization. GMR estimates at Month 19 and Month 36 were 22·3 (95% CI: 19·9–24·6; p<0·0001) and 7·8 (95% CI: 5·5–10·2; p<0·0001), respectively (Table S14 and Figure S2).
In analyses that excluded the Month 18 titer level as a covariate, GMR estimates were similar: GMR=8·3; 95% CI: 5·9–10·6; p<0·0001 for the analysis of variance model using Month 36 antibody levels and GMR=8·3; 95% CI: 6·1–11·4; p<0·0001 for the linear mixed-effects regression model, which utilized all of the follow-up titers after randomization to estimate the Month 36 GMR.
An analysis stratified by site that included a covariate corresponding to age at entry as well as clinical site yielded GMRs similar to those in Table S14 at Month 19 (GMR=21·9; 95% CI: 15·7–30·6; p<0·0001) and Month 36 (GMR=8·0; 95% CI: 5·9–10·8, p<0·0001).
Comparison of the randomized groups for the percentages with fold changes in titers, ranging from ≥1 to ≥32, were significant for all fold changes considered at each follow-up visit after 18 months (Table S15).
The superiority of the antibody response at Month 36 in the booster group was evident in each pre-specified subgroup (Table S16). Heterogeneity of response was evident by occupational risk (p=0·0009) and by tertile of the Month 18 titer (p<0·0001). For the latter subgroup, the treatment effect estimated by GMRs decreased with increasing Month 18 titer.
In a subsample of participants in the primary analysis cohort, one month after primary vaccination with rVSVΔG-ZEBOV-GP, participants’ samples reacted with recombinant EBOV GP by SPR, and peak titers were measured for 25 participants (Figure S3, Table S17). However, thereafter the mean serum antibody reactivity declined: 80% of individuals had substantially reduced EBOV GP-binding antibodies by Month 18. After the booster administration at Month 18, a strong anti-EBOV GP antibody response was detected in participants in the booster group at Month 19, whereas EBOV GP-binding antibodies in the no-booster group continued to decline from their Month 1 peak. While anti-EBOV GP antibody titers also declined in the booster group from Month 19 to Month 36, they remained significantly higher at Month 36 compared with the no-booster group (1,503 response units [RU]/mL versus 126 RU/mL; GMR=10·7; 95% CI: 7·1–14·3; p=0·00077).
DISCUSSION
This randomized clinical trial has shown that by delaying boosting until 18 months after a primary vaccination with rVSVΔG-ZEBOV-GP—i.e., well beyond the expected time needed for the immune response to fully mature and resume a resting state—a remarkable increase in GMTs is observed one month after the booster administration. While restricted to healthy U.S. participants in this trial, this increase is in the range of GMTs observed in some EVD survivors (median of 19,242 EU/mL, IQR 12,550–33,201).9 A gradual GMT decline in the booster group occurred from Month 19 to Month 36, but it nonetheless remained consistently several fold above the GMTs of participants in the no-booster group throughout that interval. Likewise, the second rVSVΔG-ZEBOV-GP inoculation at Month 18 strongly increased the broader anti-EBOV GP antibody response as measured by SPR.
In contrast, most previous EVD vaccine trials studying the potential benefits of boosting have focused on relatively short intervals between primary and secondary immunizations. Short intervals typically were chosen due to the exigency of trying to protect at-risk individuals during new EVD outbreaks, the increasing frequency of those outbreaks over the past 10 years, and the desire to induce high titers of antibodies as quickly as possible within a narrow window of time.5,6 Short-term increases in GMTs through boosting have been demonstrated in these trials, but they’ve tended to regress close to those of no-booster comparator GMTs within a 6-to-12-month time frame, likely narrowing the window during which the additional dose might be viewed as a useful adjunct. Indeed, SPR analysis also revealed that rVSVΔG-ZEBOV-GP vaccination generated anti-EBOV GP-binding antibodies that peaked after one month and declined sharply by Month 3, an effect also seen with short-term boosting.10 Because an optimal memory B-cell response to foreign antigen may take several months to mature, this regression may reflect the limited benefit of frontloading the immune system with repetitive antigen exposures in rapid succession before antibody maturation has been achieved, as was observed in prime–boost interval studies in earlier H5N1 influenza vaccine trials.11,12
In this study, the booster administration did not appear to confer any additional, and possibly even fewer, safety concerns compared to primary vaccination. Assuming that transient VSIV viremia after immunization is the probable cause of rVSVΔG-ZEBOV-GP’s short-term side effects, one explanation for the favorable safety profile could be that an anti-VSIV antibody response induced by the primary vaccination likely delimits the extent and duration of that viremia after rechallenge. We should note, however, that because primary vaccine recipients could self-select not to proceed to randomization for any reason, it is possible that those potentially at greatest risk for recurrence of adverse events might have been underrepresented in safety assessments beyond Month 18.
Strengths of our findings are: 1) the novel boosting strategy given 18 months after prime vaccination; 2) corroboration of the anti-EBOV GP IgG titer results in the FANG assay by broader subclass analyses possible through SPR; and 3) the 18-month follow-up period after the booster randomization. Weaknesses include: 1) failure to achieve the initial randomization target due to participants who did not proceed to the booster randomization 18 months after primary vaccination; and 2) missing some primary outcome data at 36 months (18 months after the booster randomization). COVID-19 pandemic restrictions on travel and adverse events after the primary vaccination were strongly related to low attendance at the Month 18 randomization visit. Albeit quite transient, 204 (92%) participants developed a pain score ≥3 during the primary vaccination, and 26 (12·6%) developed transient arthritis or an SAE during the 18-month follow-up period after primary vaccination. In general, the loss of participants between primary vaccination and randomization 18 months later resulted in a smaller sample size than planned and that, in turn, results in a lack of precision around the primary endpoint, for subgroups of the primary endpoint and for risk estimates of major safety outcomes.
While we cannot be certain that missing primary outcome titers at 36 months are not informative and lead to a biased estimate of the treatment difference as measured by the GMR at 36 months, it was reassuring that missing data did not vary by randomization group; the mixed-effects regression model led to similar GMRs, and GMRs did not vary significantly among the clinical sites. Age was positively associated with collection of Month 36 titers (older participants provided more data than younger participants), but adjusting for age had little impact on the GFR estimates.
Of note, these data and analyses focused only on the humoral response to vaccination and thus do not address the potential effects of a delayed booster on the cellular component of the immune response. Although peripheral blood mononuclear cells (PBMCs) could only be collected from a small number of participants and stored at one site for the 36 months of follow up, additional insight through in-depth characterization of B-cell responses in that subset of participants at key timepoints may be possible and, if so, will be reported in the future.
Finally, it should be noted that in some other viral diseases far less lethal than EVD for which live replicating viral vaccines are available, two doses of vaccine (e.g., measles, mumps, and rubella [MMR]) separated in time are generally recommended to ensure durable protection.13 However, it is equally important to note that the current licensed indication for rVSVΔG-ZEBOV-GP is for a single dose. To date, studies to define an actual seroprotective threshold for this vaccine in an at-risk population have not yielded clear results nor is the contribution of the cellular immune response to sustaining this protection well-characterized.14 Nevertheless, advisory groups making recommendations for how to provide optimal protection for prior vaccine recipients who remain occupationally at risk for potential future exposures to EVD may want to take note of the striking difference in both the elevation and duration of anti-EBOV antibody titers that results from adopting a delayed boosting strategy.
Supplementary Material
Research in context.
Evidence before this study
A PubMed search of all human, preclinical, and animal studies focusing on rVSVΔG-ZEBOV-GP experimental data published in English from database inception through May 2024 was performed using major search terms “vsv ebola vaccine”, “rVSVΔG-ZEBOV-GP”, “experimental ebola vaccines”, and “ebola prevention” singly or in combinations. From at least 9 studies or reviews of the rVSVΔG-ZEBOV-GP vaccine experience in nonhuman primates published since 2008 it can be concluded that primary vaccination administered as pre-exposure prophylaxis can provide substantial protection from an otherwise lethal dose of Ebola virus administered within weeks of vaccine administration. Subsequently, an open-label cluster-randomized ring vaccination trial performed in Western Africa during the 2013–16 Ebola virus disease (EVD) outbreak showed a high level of protection after primary vaccination when administered shortly after exposure to at-risk contacts, or contacts of contacts, of EVD cases. Antibody titers before or after vaccination were not measured in that study. A seroprotective threshold for this vaccine has not been established nor is it clear how long the protection afforded by primary vaccination may persist. According to published reports, prior attempts to induce a sustained elevated serologic response above post-immunization levels by administering a booster dose within 1–2 months after primary vaccination have demonstrated that the peak response after boosting is not durable and generally wanes to pre-booster levels in less than a year.
Added value of this study
In contrast to prior studies demonstrating the transient effects of short-term boosting, the results of this study demonstrate that, by delaying a homologous booster dose until after the human immune system’s response to the initial antigen challenge has had sufficient time to mature before re-challenge, a much higher and much more sustained level of anti-virus glycoprotein antibody production can result.
Implications of all the available evidence
The results of this study provide data that directly address the uncertainty of when to provide a booster dose to primary vaccine recipients who may face either ongoing or future occupational risk for exposure to Ebola virus in either the clinic or laboratory. If achieving an augmented antibody response is the goal, one clear implication is that these at-risk individuals should consider not delaying their primary vaccination until after a new Ebola virus disease outbreak is declared but rather receive primary vaccination well in advance of potential exposures and then, depending upon the interval since that vaccination, augment their antibody levels with a booster dose at or near the time of heightened exposure risk. In the meantime, further research to identify whether a discrete seroprotective threshold against Ebola virus infection can be identified in vaccine recipients, whether preliminarily in the nonhuman primate model or, ideally, in an outbreak setting, should be pursued when possible.
Acknowledgements
We thank Mark Mednikov and Elizabeth Elliott for assisting with study coordination, sample receipt and inventory management, sample testing, and data analysis and reporting at the National Institutes of Health (NIH) National Institute of Allergy and Infectious Diseases (NIAID) Division of Clinical Research (DCR) Integrated Research Facility at Fort Detrick (IRF-Frederick). We also thank Anya Crane and Jiro Wada (both with NIH NIAID DCR IRF-Frederick) for critically editing the manuscript and figure preparation, respectively.
Source of funding statements
This work was supported in part through the U.S. National Institutes of Health (NIH) National Institute of Allergy and Infectious Diseases (NIAID) under prime contract with Battelle Memorial Institute (HHSN272200700016I) and, subsequently, with Laulima Government Solutions, LLC (HHSN272201800013C). Tunnell Government Services (TGS) was a subcontractor of Battelle Memorial Institute and Laulima Government Solutions, LLC, under both contracts. This work was also supported in part with federal funds from the NIH National Cancer Institute (NCI), under Contract No. 75N910D00024 with Leidos Biomedical Research, Inc., as also supported by the Applied and Developmental Research Directorate, Frederick National Lab for Cancer Research (sponsored by NCI). The surface plasmon resonance (SPR) antibody characterization work described in this manuscript was supported by U.S. Defense Threat Reduction Agency (HDTRA1930447) funds to S.K. Funding was also provided by the Canadian Immunization Research Network through the Public Health Agency of Canada and the Canadian Institutes of Health Research.
The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the U.S. Department of Health and Human Services or of the institutions and companies affiliated with the authors, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
Funding
The Division of Intramural Research (DIR) and the Division of Clinical Research (DCR) of the National Institute of Allergy and Infectious Diseases (NIAID) at the U.S. National Institutes of Health (NIH)
Canadian Immunization Research Network through the Public Health Agency of Canada
Canadian Institutes of Health Research
U.S. Defense Threat Reduction Agency
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declaration of interests
The authors declare that no conflicts of interest exist.
Data sharing
The study protocol, informed consent document, and other study documents are available upon request. In addition to study materials available on ClinicalTrials.gov, access to de-identified data from qualified external researchers will be made available within one year after publication of the primary manuscript and upon completion of a signed Data Transfer Agreement with NIAID.
REFERENCES
- 1.U.S. Food & Drug Administration. First FDA-approved vaccine for the prevention of Ebola virus disease, marking a critical milestone in public health preparedness and response. https://www.fda.gov/news-events/press-announcements/first-fda-approved-vaccine-prevention-ebola-virus-disease-marking-critical-milestone-public-health 2019.
- 2.European Medicines Agency. First vaccine to protect against Ebola. https://www.ema.europa.eu/en/news/first-vaccine-protect-against-ebola. 2019.
- 3.Garbutt M, Liebscher R, Wahl-Jensen V, et al. Properties of replication-competent vesicular stomatitis virus vectors expressing glycoproteins of filoviruses and arenaviruses. J Virol 2004; 78(10): 5458–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Henao-Restrepo AM, Camacho A, Longini IM, et al. Efficacy and effectiveness of an rVSV-vectored vaccine in preventing Ebola virus disease: final results from the Guinea ring vaccination, open-label, cluster-randomised trial (Ebola ça Suffit!). Lancet 2017; 389(10068): 505–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Regules JA, Beigel JH, Paolino KM, et al. A recombinant vesicular stomatitis virus Ebola vaccine. N Engl J Med 2017; 376(4): 330–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.PREVAC Study Team, Kieh M, Richert L, et al. Randomized trial of vaccines for Zaire Ebola virus disease. N Engl J Med 2022; 387(26): 2411–24. [DOI] [PubMed] [Google Scholar]
- 7.Logue J, Tuznik K, Follmann D, et al. Use of the Filovirus Animal Non-Clinical Group (FANG) Ebola virus immuno-assay requires fewer study participants to power a study than the Alpha Diagnostic International assay. J Virol Methods 2018; 255: 84–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Anderson MS, Niemuth NA, Sabourin CL, Badorrek CS, Bounds CE, Rudge TL Jr. Interlaboratory comparison for the Filovirus Animal Nonclinical Group (FANG) anti-Ebola virus glycoprotein immunoglobulin G enzyme-linked immunosorbent assay. PLoS One 2020; 15(8): e0238196. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.PREVAIL III Study Group, Sneller MC, Reilly C, et al. A longitudinal study of Ebola sequelae in Liberia. N Engl J Med 2019; 380(10): 924–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Khurana S, Fuentes S, Coyle EM, Ravichandran S, Davey RT, Jr., Beigel JH. Human antibody repertoire after VSV-Ebola vaccination identifies novel targets and virus-neutralizing IgM antibodies. Nat Med 2016; 22(12): 1439–47. [DOI] [PubMed] [Google Scholar]
- 11.Khurana S, Wu J, Dimitrova M, et al. DNA priming prior to inactivated influenza A(H5N1) vaccination expands the antibody epitope repertoire and increases affinity maturation in a boost-interval-dependent manner in adults. J Infect Dis 2013; 208(3): 413–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ledgerwood JE, Zephir K, Hu Z, et al. Prime-boost interval matters: a randomized phase 1 study to identify the minimum interval necessary to observe the H5 DNA influenza vaccine priming effect. J Infect Dis 2013; 208(3): 418–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Centers for Disease Control and Prevention. Measles, mumps, and rubella (MMR) vaccination: what everyone should know. https://www.cdc.gov/vaccines/vpd/mmr/public/index.html. 2021.
- 14.Adriaensen W, Oostvogels S, Levy Y, Leigh B, Kavunga-Membo H, Watson-Jones D. Urgent considerations for booster vaccination strategies against Ebola virus disease. Lancet Infect Dis 2024. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The study protocol, informed consent document, and other study documents are available upon request. In addition to study materials available on ClinicalTrials.gov, access to de-identified data from qualified external researchers will be made available within one year after publication of the primary manuscript and upon completion of a signed Data Transfer Agreement with NIAID.