A Phase 3, randomized, non-inferiority study of a heterologous booster dose of SARS CoV-2 recombinant spike protein vaccine in adults

Due to waning immunity following primary immunization with COVID-19 vaccines, booster doses may be required. The present study assessed a heterologous booster of SII-NVX-CoV2373 (spike protein vaccine) in adults primed with viral vector and inactivated vaccines. In this Phase 3, observer-blind, randomized, active controlled study, a total of 372 adults primed with two doses of ChAdOx1 nCoV-19 (n = 186) or BBV152 (n = 186) at least six months ago, were randomized to receive a booster of SII-NVX-CoV2373 or control vaccine (homologous booster of ChAdOx1 nCoV-19 or BBV152). Anti-S IgG and neutralizing antibodies (nAbs) were assessed at days 1, 29, and 181. Non-inferiority (NI) of SII-NVX-CoV2373 to the control vaccine was assessed based on the ratio of geometric mean ELISA units (GMEU) of anti-S IgG and geometric mean titers (GMT) of nAbs (NI margin > 0.67) as well as seroresponse (≥ 2 fold-rise in titers) (NI margin −10%) at day 29. Safety was assessed throughout the study period. In both the ChAdOx1 nCoV-19 prime and BBV152 prime cohorts, 186 participants each received the study vaccines. In the ChAdOx1 nCoV-19 prime cohort, the GMEU ratio was 2.05 (95% CI 1.73, 2.43) and the GMT ratio was 1.89 (95% CI 1.55, 2.32) whereas the difference in the proportion of seroresponse was 49.32% (95% CI 36.49, 60.45) for anti-S IgG and 15% (95% CI 5.65, 25.05) for nAbs on day 29. In the BBV152 prime cohort, the GMEU ratio was 5.12 (95% CI 4.20, 6.24) and the GMT ratio was 4.80 (95% CI 3.76, 6.12) whereas the difference in the proportion of seroresponse was 74.08% (95% CI 63.24, 82.17) for anti-S IgG and 24.71% (95% CI 16.26, 34.62) for nAbs on day 29. The non-inferiority of SII-NVX-CoV2373 booster to the control vaccine for each prime cohort was met. SII-NVX-CoV2373 booster showed significantly higher immune responses than BBV152 homologous booster. On day 181, seroresponse rates were ≥ 70% in all the groups for both nAbs and anti-S IgG. Solicited adverse events reported were transient and mostly mild in severity in all the groups. No causally related SAE was reported. SII-NVX-CoV2373 as a heterologous booster induced non-inferior immune responses as compared to homologous boosters in adults primed with ChAdOx1 nCoV-19 and BBV152. SII-NVX-CoV2373 showed a numerically higher boosting effect than homologous boosters. The vaccine was also safe and well tolerated.


Study participants
The study participants were healthy adults aged ≥ 18 years who gave written informed consent and who had completed primary COVID-19 immunization schedule of two doses with either ChAdOx1 nCoV-19 or BBV152 at least 6 months ago.Participants with a history of laboratory confirmed COVID-19, history of allergic reactions after previous vaccinations, hypersensitivity to any component of study vaccines, any condition with impaired/ altered function of immune system, any clinically significant systemic disorder were excluded.Clinically wellcontrolled comorbidities were allowed.

Immunogenicity assessments
Anti-S IgG antibodies against prototype strain (Wuhan) were assessed using validated enzyme-linked immunosorbent assay (ELISA) and nAbs were assessed by validated microneutralization (MN) assay using wild type virus (Ancestral strain: SARS-CoV-2 hCoV-19/Australia/VIC01/2020, GenBank MT007544.1)with an inhibitory concentration of 50% (MN 50 ).Seroresponse was defined as 2 fold increase in antibody titers from baseline.Immunogenicity against VoC was assessed in a randomly selected subset of 46 participants for anti-S IgG against Omicron BA.1, and BA.5, and 50 participants for nAbs against Omicron B.1.1.529(BA.1) from each prime cohort (maintaining 1:1 allocation).In addition, in the same subset of 46 participants, immune responses against Wuhan strain, Omicron BA.1, and BA.5 variants were determined by validated hACE2 receptor binding inhibition assay, as an exploratory objective.CMI responses were measured by quantification of antigen specific T cells using an ex vivo interferon-γ enzyme-linked immune absorbent spot (ELISpot) assay in a randomly selected subset of 18 participants from each of the two prime cohorts maintaining 1:1 allocation.
The lower limit of quantification (LLOQ) for anti-S IgG ELISA assay was 200 ELISA units per mL (EU/mL), with titers below this level documented as 100 EU/mL.The LLOQ for MN assay was a titer of 20, with titers below this level documented as 10.The LLOQ for hACE2 receptor binding inhibition assay was a titer of 10, with titers below this level documented as 5. Details of assays are included in the Supplementary appendix.
Immunogenicity testing was performed in compliance with Good Clinical Laboratory Practice (GCLP) requirements at Novavax, USA (Anti-S IgG and hACE2 receptor binding inhibition assay), at 360biolabs, Australia (nAbs) and at NARI-ICMR, India (CMI).

Safety assessments
Solicited local and systemic adverse events (AEs) were actively collected for 7 days after vaccination using participant diary cards.The solicited local AEs included injection site pain, tenderness, erythema, swelling, and induration.The solicited systemic AEs included fever, headache, fatigue, malaise, arthralgia, myalgia, nausea, and vomiting.Unsolicited AEs were collected for 28 days after vaccination.Unsolicited AEs were any AEs reported by the participant, observed by the study staff during study visits or those identified during review of medical records or source documents.Serious adverse events (SAEs), and adverse events of special interest (AESIs) including potentially immune-mediated medical condition (PIMMCs) and AESIs relevant to COVID-19 were collected throughout the study.
The severity of all AEs was graded using the Division of AIDS (DAIDS) Table for Grading the Severity of Adult and Pediatric Adverse Events, corrected Version 2.1, July 2017.A subjective grading scale was used to grade the severity of all AEs that were not listed in the DAIDS Table .Safety was monitored during the study by on-site clinical staff.Periodic reviews of safety data was performed by the protocol safety review team (PSRT).In addition, an independent data safety monitoring board (DSMB) reviewed the safety data and provided oversight on the study.

Statistical considerations
Total 372 participants were randomized in the study (186 in each prime cohort).Assuming non-evaluable participants ≤ 15% (158 evaluable participants for each cohort), group sample sizes of 93 each in each vaccine group would provide 80% power to detect non-inferiority using a one-sided, two-sample t-test, with a true ratio of the means of 1.00, one-sided significance level (alpha) of 0.025 and coefficients of variation of 1.1.Non-inferiority of SII-NVX-CoV2373 booster over the control vaccine for each cohort was to be concluded separately if the lower limit of the two-sided 95% confidence interval (CI) for the geometric mean ELISA units (GMEU) ratio for anti-S IgG and geometric mean titers (GMT) ratio for nAbs between SII-NVX-CoV2373 and the control vaccine at day 29 was > 0.67 (non-inferiority margin).The same sample size provided 82% power to demonstrate non-inferiority in terms of difference in proportion of participants with seroresponse (anti-S IgG and nAbs) between SII-NVX-CoV2373 and the control vaccine at day 29, using a one-sided, two-sample z-test at alpha level of 0.025.Proportion of participants achieving seroresponse for control arm was assumed 95% and the margin of non-inferiority was −10%.The true difference of the proportion was assumed to be 0. Sample size calculations were performed using a non-inferiority test for the ratio of two means in PASS 15.0.7 Version software.
Analysis of Covariance (ANCOVA) was fitted to the log transformed anti-S IgG or nAbs for vaccine group, log baseline titer, age, sex, duration between first and second dose of primary immunization, duration between second dose to booster dose.Individual mean and 95% CI by treatment group from this model were used to estimate GMTs/GMEUs and geometric mean ratio (GMR) with 95% CI by back transforming to the original scale for non-inferiority comparison at day 29.The unadjusted GMTs/GMEUs were also summarized.The difference between the vaccine groups in the proportion of the participants with seroresponse along with two-sided 95% CIs were calculated using the Miettinen and Nurminen method.
Enrolled Population was all participants who provided written informed consent, regardless of screening, randomization and treatment status in the study.Safety Population included all participants who received the study vaccines.Per Protocol population consisted of all participants who received the study vaccine, provided an evaluable serum sample post vaccination for at least one assessment, had baseline (day 1) immunogenicity data available, excluding any data from time points following a SARS-CoV-2 infection or major protocol deviation before day 29.This was the primary population for immunogenicity analyses.Full Analysis Population comprised of all participants who received the study vaccine and provided an evaluable serum sample post vaccination for at least one assessment.
Continuous data was described using descriptive statistics (n, mean, standard deviation, median, minimum, and maximum).Categorical data was described using the participant count and percentage for each category.
All statistical analyses were performed using SAS® software version 9.4.

Ethical considerations
The

Disposition of participants
ChAdOx1 nCoV-19 prime cohort A total of 190 participants were enrolled and 187 were randomized.Of these, 186 received the study vaccine (92 SII-NVX-CoV2373 and 94 ChAdOx1 nCoV-19).Only a single participant withdrew the consent before day 29.The safety population comprised 186 participants, both full analysis and per protocol populations comprised 185 participants (Fig. 1).
The demographic and baseline characteristics between the groups were comparable.Mean age was about 36 years and 45% were males.About 15% participants had comorbidity.None of the participants was SARS CoV-2 RT-PCR positive.Median duration between two doses of prime series of ChAdOx1 nCoV-19 was 90 days and that between second dose of prime series and the booster dose was 270 days.The proportion of participants with seropositivity for anti-S IgG or nAbs at baseline was comparable between the groups (Table 1).

BBV152 prime cohort
A total of 190 participants were enrolled and 186 were randomized and received the study vaccine (92 SII-NVX-CoV2373 and 94 BBV152).Two participants withdrew consent before day 29.The safety population comprised 186 participants, both full analysis and per protocol populations comprised 184 participants (Fig. 1).
The demographic and baseline characteristics between the groups were comparable.Mean age was about 37 years and about 64% were males.About 15% participants had comorbidity.None of the participants was SARS CoV-2 RT-PCR positive.Median duration between two doses of prime series of BBV152 was 35 days and that between second dose of prime series and the booster dose was 298 days.The seropositivity for anti-S IgG and nAbs at baseline was comparable between the groups (Table 1).

ChAdOx1 nCoV-19 prime cohort
Anti-S IgG: Baseline GMEUs of anti-S IgG against prototype strain were comparable between the SII-NVX-CoV2373 and the ChAdOx1 nCoV-19 groups.At day 29, there was a 3.9 fold-rise (95% CI 3. 4  GMEU Geometric mean ELISA unit, GMT Geometric mean titer.For each study vaccine, the GMEU of Anti-S IgG and GMT of nAbs and 95% CI were calculated by transforming to the original scale of log 10 -transformed mean and its two-sided 95% CI limits at each visit.

Table 2. Non-inferiority of SII-NVX-CoV2373 heterologous booster to homologous booster in terms for
Anti-S IgG and neutralizing antibodies at Visit 2-Day 29 -Per Protocol Population.LS least squares, GMT geometric mean titer, GMR geometric mean ratio.[1] ANCOVA results, LS Mean and it's 95% CI values by treatment were used to generate the GMEU/GMT and 95% CI and the differences in LS Means and corresponding 95% CI limits were used to obtain GMEU/GMT Ratio and 95% CI using back transforming to the original scale.ANCOVA model includes vaccine group, log baseline titer, age, sex, duration between first and second dose of the prime vaccine, duration between second dose of the prime vaccine to booster dose of the study vaccine.The 95% CIs for seroresponse for each vaccine group were calculated by using the Clopper-Pearson method.Estimate of difference in proportion along with associated 95% CI was obtained using Miettinen and Nurminen method.Anti-S IgG GMEU [1]   S4).

SAEs and AESIs
Two SAEs were reported in the SII-NVX-CoV2373 group (BBV152 prime cohort) -lower respiratory tract infection and gastroenteritis with dehydration, both unrelated to the study vaccine.No AESI was reported.

Unsolicited AEs
ChAdOx1 nCoV-19 prime cohort: Only a single unsolicited AE (1.1%) was reported in the SII-NVX-CoV2373 group and 6 unsolicited AEs in 4 participants (4.3%) in the ChAdOx1 nCoV-19 group and none was treatmentrelated.All AEs were of mild severity (except for 1 event of joint injury in the SII-NVX-CoV2373 group of moderate severity) and resolved without any sequelae.

BBV152 prime cohort
A total of 4 unsolicited AEs in 4 participants (5.4%) were reported in the SII-NVX-CoV2373 group and 6 unsolicited AEs in 5 participants (5.3%) in the BBV152 group.None was treatment-related in the SII-NVX-CoV2373 group and 2 events in the BBV152 group were treatment-related (diarrhoea and injection site pain).All AEs were of mild to moderate severity (except for 1 severe event of hypertension in the BBV152 group) and resolved without any sequelae.
There were no AEs which led to study discontinuation.There was only one case of COVID-19 at 3 months after the booster in the BBV152 group which was mild in severity, and resolved without any sequelae.

BBV152 prime cohort
There were 84 solicited AEs in 34 participants (37%) in the SII-NVX-CoV2373 group and 87 solicited AEs in 43 participants (45.7%) in the BBV152 group.The most common local solicited AEs were injection site pain, tenderness, and swelling (Table 3).The common systemic solicited AEs were fatigue, headache, malaise, myalgia, arthralgia, and fever (Table 3).Almost all AEs (98%) were of mild severity and all resolved without any sequelae.

Discussion
This Phase 3 study was conducted to assess heterologous booster effect of SII-NVX-CoV2373 in 372 adults who were primed with ChAdOx1 nCoV-19 or BBV152.The heterologous booster dose of SII-NVX-CoV2373 was non-inferior to homologous booster dose of either ChAdOx1 nCoV-19 or BBV152 in terms of anti-S IgG and nAbs.Also, the SII-NVX-CoV2373 booster showed significantly higher immune responses than BBV152 homologous booster in terms of anti-S IgG, nAbs and hACE2 receptor binding inhibition antibodies against prototype / ancestral strain as well as Omicron variants.All three vaccines were found safe and well tolerated.infections.However, anti-S IgG against Omicron showed a decline in titers except in BBV152 group.The reason for these discordant findings is not known.
Neutralizing antibody titers are an accepted measure of immunity against COVID-19 while there is a strong correlation for nAb titers with anti-S IgG and hACE-2 binding inhibition among vaccinated individuals 34 .We also found the similar trend in our study.
It is interesting to note that the baseline titers of both anti-S IgG and nAbs in the ChAdOx1 nCoV-19 prime group were almost two times as compared to the BBV152 prime group even though the intervals since the last vaccination were quite similar.This has been reported earlier that ChAdOx1 nCoV-19 gave higher, more durable antibody response than BBV152 [35][36][37][38] .Moreover, the post-booster titers in the ChAdOx1 nCoV-19 prime-ChAdOx1 nCoV-19 boost participants were much higher than in the BBV152 prime-BBV152 boost participants.This has also been reported previously 39 .In our study, there were more males in the BBV152 prime cohort than in the ChAdOx1 nCoV-19 prime cohort which could be an incidental finding.However, we do not anticipate any impact of this finding on our results.
Effectiveness of ChAdOx1 nCoV-19 has also been higher than BBV152 [40][41][42][43] .The same effectiveness data for NVX-CoV2373 booster are not available because the vaccine was approved much later than these two vaccines.As a result, it was not used in the programme in India to a great extent.Higher levels of humoral immune markers are known to correlate with a reduced risk of symptomatic infection [44][45][46] .Considering this, the higher immune response seen in this study indicates that SII-NVX-CoV2373 booster may give similar or higher effectiveness than ChAdOx1 nCoV-19 and BBV152 booster.
Our study had a few limitations.Our study was unable to assess the immunogenicity among the elderly population with sufficient sample size, as most of this population had already received a booster dose prior to the commencement of the study.This exclusionary factor limited the elderly cohort, constituting only 5% of our study participants.This subset of participants, especially those with comorbidities, might exhibit immunosenescence, thereby adding complexity to our observations.The study did not assess BA.2, BA.3, BA.4, and Delta-variants.We evaluated only one booster dose whereas in some countries second and third boosters have also been introduced particularly for older adults and immunocompromised individuals 47 .This was because in India only one booster dose is recommended.The study was not designed to assess efficacy of the booster, however as mentioned above higher levels of all immune markers are known to correlate with a reduced risk of symptomatic infection [44][45][46] .Though the efficacy of SII-NVX-CoV2373 has been reported to be higher than ChAdOx1 nCoV-19 and BBV152, the present study was not designed as a superiority trial as this was not necessary for regulatory approval of a heterologous booster dose.In our study, we did not assess the homologous booster dose of SII-NVX-CoV2373 because SII-NVX-CoV2373 was not used for primary immunization in India.However, NVX-CoV2373 has already been evaluated as a homologous booster dose which showed that for both the prototype strain and all variants evaluated, immune responses following the booster were similar to or higher than those associated with high levels of efficacy in phase 3 studies of the vaccine 6 .
A study reported that the heterologous booster induced greater systemic reactogenicity than their homologous counterparts 13 .However, in our study heterologous booster induced numerically higher systemic reactions than homologous booster in ChAdOx1 nCoV-19 prime cohort, while homologous booster induced numerically higher local reactions than heterologous booster in BBV152 prime cohort, though our study was not powered to detect these differences.
In conclusion, a heterologous booster of SII-NVX-CoV2373 in ChAdOx1 nCoV-19 and BBV152-primed adults demonstrated a non-inferior booster response, which was numerically higher than that achieved with a homologous booster of ChAdOx1 nCoV-19 and BBV152.The vaccine also exhibited an acceptable safety profile.

Figure 2 .
Figure 2. Summary of Anti-S IgG and Neutralizing Antibodies (nAb) at each visit (Per Protocol Population).GMEU Geometric mean ELISA unit, GMT Geometric mean titer.For each study vaccine, the GMEU of Anti-S IgG and GMT of nAbs and 95% CI were calculated by transforming to the original scale of log 10 -transformed mean and its two-sided 95% CI limits at each visit.

Figure 3 .
Figure 3. Summary of Neutralizing Antibodies (nAb) against Omicron B.1.1.529at each visit (Per protocol population).GMT: Geometric Mean Titer.For each study vaccine, the GMT of Neutralizing Antibodies (nAb) and 95% CI were calculated by transforming to the original scale of log 10 -transformed mean and its two-sided 95% CI limits at each visit.
study was approved by the Indian regulatory authority and the Institutional Ethics Committees (IECs) of each of the eight participating study sites.
The IECs approving the study were Jamia Hamdard Institutional Ethics Committee, Delhi; Institutional Ethics Committee Bharati Vidyapeeth Deemed University, Pune; Ethics Committee Dr. D. Y. Patil Vidyapeeth, Pune; Institutional Ethics Committee-KIMS, Kalinga Institute of Medical Sciences, Bhubaneshwar; Noble Hospital Institutional Ethics Committee, Pune; Institutional Ethics Committee JSS Medical College, Mysuru; KEM Hospital Research Centre Ethics Committee, Pune; and Institutional Human Ethics Committee All India Institute of Medical Sciences, Gorakhpur.The study was conducted in compliance with Declaration of Helsinki (Revised Fortaleza, 2013) and the International Council for Harmonization (ICH) GCP Guideline, E6 R2 (2016).

Table 1 .
Demographics and baseline characteristics -safety population.