Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

HXP-GPOVac is the specific, vaccine candidate developed against the inactivated COVID-19 virus. In this manuscript the authors described the phase II trial of the Thailand produced HXP-GPOVac based on the overall strategy used for the Newcastle Disease Virus expressing a stabilized, six-proline "HexaPro" spike which was developed and easily produced in chicken eggs and demonstrated high neutralizing antibody responses.

 

The safety and immunogenicity of the HXP-GPOVac was compared to the Pfizer -BioNTech mRNA vaccine BNT162b2 in adults aged 18–75 years. In this 3:1 randomized, double-blind, active-controlled study which consisted of immunizing 300 participants to receive either HXP-HGPOVac or BNT162b2 on Days 1 and 29. 

The solicited adverse events (AEs) of the NDV-HXP-S. vaccine and SAEs, MAAEs, and AESI were evaluated through and through Day 197. The anti-spike IgG geometric mean concentrations (GMCs) vary significantly based on vaccination status, time since last dose, and the variant, typically measured in Binding Antibody Units per mL (BAU/mL). Both the HXP-GPOVac and the BNT162b2 vaccines were well tolerated. Most AEs were mild/moderate. No vaccine-related SAEs or AESIs. Three deaths were noted in the HXP-GPOVac group, and these were unrelated to vaccination.

This is a well-structured Phase II randomized, double-blind, active-controlled trial comparing the safety and immunogenicity study of a locally produced, inactivated NDV-based COVID-19 vaccine (HXP-GPOVac) against the mRNA vaccine BNT162b2 (Pfizer-BioNTech). An interesting finding was realized when individuals with both vaccination and natural infection, also known as hybrid immunity) showed the highest IgG levels (geometric mean: 525.4 BAU/mL, ranging from 15.5–6517.8 BAU/mL

In conclusion, the authors found that HXP-GPOVac was well tolerated and induced robust humoral and cellular immune responses, characterized by high seroconversion rates and balanced T-cell polarization. 

The manuscript is well written, and the results obtained indicate that further clinical development of the inactivated NDV-HXP-S vaccine platform are needed.

The following are some minor queries necessary for better clarification of this manuscript:

1. You defined the non-inferiority threshold as the 95% CI for NT50 GMT ratio exceeding 0.5. The observed ratio at Day 43 was 0.51 (0.39–0.67). The lower bound is 0.39, which is below 0.5. Were these thresholds chosen based on regulatory guidance (e.g., WHO, FDA)? If so, which specific guidance? And would your conclusion change if a different margin (e.g., 0.67) were used?
2. While evaluating the baseline SARS-CoV-2 serostatus were the low baseline titers (NT50 ~5.6), but was prior infection definitively excluded by both serology and PCR history due to undetected prior infection, which could affect immunogenicity comparisons.
3. Also, you reported19 lost to follow-up in the HXP-GPOVac group and 3 deaths. Did loss to follow-up differ by age or comorbidity status, and could this have biased the Day 197 immunogenicity estimates?
4. The vaccine uses aluminum hydroxide adjuvant. Did you evaluate any other adjuvants or dose levels (e.g., 5 μg or 20 μg) in earlier phases? If so, why was 10 μg selected as the final dose?

Author Response

Comments 1: You defined the non-inferiority threshold as the 95% CI for NT50 GMT ratio exceeding 0.5. The observed ratio at Day 43 was 0.51 (0.39–0.67). The lower bound is 0.39, which is below 0.5. Were these thresholds chosen based on regulatory guidance (e.g., WHO, FDA)? If so, which specific guidance? And would your conclusion change if a different margin (e.g., 0.67) were used?

Response 1: The non-inferiority margin of 0.5 was prespecified in the study protocol prior to trial initiation. We have revised the Statistical Analysis section to state this clearly and have added explicit null and alternative hypotheses: "The null hypothesis was that the NT50 GMT ratio (HXP-GPOVac/BNT162b2) was less than 0.5; the alternative hypothesis was that the ratio was at least 0.5." The conclusion does not change under a 0.67 margin, the lower bound of 0.39 falls below both thresholds. This is now stated explicitly in the manuscript.

 

Comments 2: While evaluating the baseline SARS-CoV-2 serostatus were the low baseline titers (NT50 ~5.6), but was prior infection definitively excluded by both serology and PCR history due to undetected prior infection, which could affect immunogenicity comparisons.

Response 2: We have added explicit clarification to the Statistical Analysis section: "At enrollment, prior SARS-CoV-2 infection was excluded through a positive COVID-19 test history and a negative Rapid Antigen Test confirmed by RT-PCR; baseline serological testing was not used as an eligibility criterion." The low baseline NT50 GMTs (5.62 and 6.45) are consistent with a predominantly seronegative population. Participants with seropositive baseline titers were excluded from the per-protocol GMT analysis per prespecified criteria, and this is now explicitly stated.

 

Comments 3: Also, you reported 19 lost to follow-up in the HXP-GPOVac group and 3 deaths. Did loss to follow-up differ by age or comorbidity status, and could this have biased the Day 197 immunogenicity estimates?

Response 3: Age-stratified or comorbidity-stratified analysis of loss to follow-up was not prespecified and is not available. We have added the following to the Limitations paragraph: "Loss to follow-up at Day 197 was 8.5% in the HXP-GPOVac group and 6.6% in the BNT162b2 group; differential attrition by age stratum or comorbidity cannot be fully excluded as a potential source of bias in the Day 197 immunogenicity estimates." All immunogenicity analyses were performed as complete-case analyses on observed data, and this is now explicitly stated.

 

Comments 4: The vaccine uses aluminum hydroxide adjuvant. Did you evaluate any other adjuvants or dose levels (e.g., 5 μg or 20 μg) in earlier phases? If so, why was 10 μg selected as the final dose?

Response 4: The 10 µg antigen dose was selected based on results from the preceding phase I trial (Pitisuttithum et al., EClinicalMedicine 2022; reference [11]), which evaluated doses of 1, 3, and 10 µg without adjuvant and 1 and 3 µg with CpG 1018 adjuvant. The 10 µg aluminium hydroxide formulation demonstrated the most favourable immunogenicity and tolerability profile and was carried forward to this phase II trial. This is described in the Introduction and Vaccines and Administration sections.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The article demonstrates a randomized, double-blind phase II clinical trial evaluating the safety and its immunogenicity of a locally provided inactivated NDV-HXP-S COVID-19 vaccine in comparison with BNT162b2 in Thai adults. The research resolves an important public health and vaccine-accessibility field, especially for low- and middle-income countries looking for scalable domestic vaccine manufacturing platforms. In addition, the paper is seemeingly well organized, and the clinical trial design is suitable. The inclusion of both humoral and exploratory cellular immunity assessments strengthens the scientific value of the work.
However, several important issues should be addressed before the manuscript can be considered for publication.
Major Comments:
The conclusions appear somewhat overstated relative to the presented immunogenicity data. Although HXP-GPOVac induced measurable humoral and cellular immune responses, the neutralizing antibody responses remained substantially lower than those induced by BNT162b2, and the prespecified non-inferiority criterion was not met. The conclusions needs revision to provide a balanced interpretation of the outcomes.
More details in terms of the reported critical side effects and deaths are required. Although the investigators and DSMB considered these events unrelated to vaccination, the manuscript should provide more detailed clinical descriptions, timing relative to vaccination, participant comorbidities, and rationale supporting causality assessment.
The manuscript lacks detailed laboratory safety findings despite stating that hematology and serum chemistry analyses were performed. The authors are encouraged to include a summary of laboratory safety outcomes. preferably in supplementary tables, including any clinically significant abnormalities.
The cellular immunity analysis remains relatively limited. The ELISpot findings are informative; however, the discussion should more clearly acknowledge the exploratory nature of these analyses and the absence of deeper T-cell characterization such as CD4/CD8 phenotyping or functional profiling.
The research can benefit from having more clarification on terms of the pseudovirus neutralization assay, particularly whether variant-specific neutralization against currently circulating SARS-CoV-2 variants was estimated.
The statistical analysis section should further justify the absence of multiplicity adjustment given the number of immunogenicity endpoints and subgroup analyses performed.
The discussion section should include a clearer limitations paragraph addressing:
the relatively small sample size,
male predominance in the study population,
limited ethnic diversity,
lack of efficacy endpoints,
and the absence of variant-specific immunogenicity analyses.
Minor Comments:
Figure 1 could be improved by providing more detailed reasons for participant exclusion and discontinuation.
Many tables are information-dense and may need formatting improvements for better readability.
The authors are encouraged to make sure of consistent terminology throughout the manuscript for seroconversion/seroresponse definitions.
The discussion section has to be contextualize the observed antibody waning at Day 197 related to published data for inactivated and mRNA COVID-19 vaccines.
Overall, the study is scientifically relevant and potentially valuable, particularly from the perspective of regional vaccine manufacturing and pandemic preparedness. However, the manuscript requires major revision to strengthen data interpretation, improve safety reporting transparency, and better contextualize the immunogenicity findings.

Author Response

Comments 1: The conclusions appear somewhat overstated relative to the presented immunogenicity data. Although HXP-GPOVac induced measurable humoral and cellular immune responses, the neutralizing antibody responses remained substantially lower than those induced by BNT162b2, and the prespecified non-inferiority criterion was not met. The conclusions needs revision to provide a balanced interpretation of the outcomes.

Response 1: Agreed. Conclusions have been rebalanced throughout. The abstract conclusions now read: "HXP-GPOVac was well tolerated and induced substantial humoral and cellular immune responses, with high seroconversion rates and balanced T-cell polarization. Although absolute antibody levels were lower than those induced by BNT162b2 and the prespecified non-inferiority criterion was not met, these findings support continued evaluation of the inactivated NDV-HXP-S vaccine platform." The Discussion opening and Summary paragraph have been revised to explicitly acknowledge the NI failure, the lower absolute antibody levels relative to BNT162b2, and the phase II, non-efficacy context of the trial.

 

Comments 2: More details in terms of the reported critical side effects and deaths are required. Although the investigators and DSMB considered these events unrelated to vaccination, the manuscript should provide more detailed clinical descriptions, timing relative to vaccination, participant comorbidities, and rationale supporting causality assessment.

Response 2: Done. Individual clinical narratives have been added to the Results section for all three deaths: "Three deaths occurred among HXP-GPOVac recipients during follow-up, all occurring more than 100 days after the second dose. Two participants (aged 49 and 55 years) experienced unattended deaths at Day 104 and Day 163 post-second dose, respectively; no preceding clinical events were recorded. One participant (aged 75 years) experienced subarachnoid hemorrhage, infective endocarditis, and bacterial pneumonia beginning approximately 126 days after the second dose, resulting in death on Day 128; unstable supraventricular tachycardia also occurred and resolved. The temporal separation from vaccination, absence of biological plausibility, and the nature of the events were the primary bases for causality assessments. All deaths were assessed by investigators and the independent DSMB as unrelated to study vaccination."

 

Comments 3: The manuscript lacks detailed laboratory safety findings despite stating that hematology and serum chemistry analyses were performed. The authors are encouraged to include a summary of laboratory safety outcomes, preferably in supplementary tables, including any clinically significant abnormalities.

Response 3: Done. Two new supplementary tables have been added: Table S5 (hematology: hemoglobin, WBC, platelets) and Table S6 (serum chemistry: creatinine, ALT, AST, total bilirubin), both by treatment group, visit (Baseline, Day 8, Day 36), and DAIDS toxicity grade. A summary paragraph has been added to the Results section: "Clinical laboratory assessments performed at baseline, Day 8, and Day 36 revealed no clinically significant changes from baseline. Abnormal values were predominantly Grade 1 or Grade 2, with rare isolated Grade 3 findings assessed as not clinically significant and not related to vaccination. Detailed results are provided in Supplementary Tables S5 and S6."

 

Comments 4: The cellular immunity analysis remains relatively limited. The ELISpot findings are informative; however, the discussion should more clearly acknowledge the exploratory nature of these analyses and the absence of deeper T-cell characterization such as CD4/CD8 phenotyping or functional profiling.

Response 4: Done. The Discussion now states: "Given the exploratory nature of the cell-mediated immunity substudy, the limited sample size, and the absence of CD4/CD8 phenotyping or functional profiling, these findings should be interpreted descriptively and do not support formal conclusions regarding the magnitude or quality of T-cell responses." This is also acknowledged in the expanded Limitations paragraph.

 

Comments 5: The research can benefit from having more clarification on terms of the pseudovirus neutralization assay, particularly whether variant-specific neutralization against currently circulating SARS-CoV-2 variants was estimated.

Response 5: Variant-specific neutralization was not assessed; all assays used an ancestral SARS-CoV-2 strain pseudovirus. This has been explicitly added to the Limitations paragraph: "Immunogenicity assays were based on the ancestral SARS-CoV-2 strain, and variant-specific neutralization was not assessed, limiting extrapolation to protection against currently circulating variants."

 

Comments 6: The statistical analysis section should further justify the absence of multiplicity adjustment given the number of immunogenicity endpoints and subgroup analyses performed.

Response 6: Done. The following has been added to the Statistical Analysis section: "Given the phase II nature of the study and the prespecified exploratory character of secondary and subgroup analyses, formal multiplicity adjustment was not applied, consistent with the descriptive, phase II nature of this study."

 

Comments 7: The discussion section should include a clearer limitations paragraph addressing: the relatively small sample size, male predominance in the study population, limited ethnic diversity, lack of efficacy endpoints, and the absence of variant-specific immunogenicity analyses.

Response 7: Done. The Limitations paragraph has been substantially expanded to address all points: phase II design and absence of efficacy endpoints; ancestral strain-based assays; exploratory CMI design and absence of CD4/CD8 phenotyping; male predominance (75.7%); single-site, single-ethnicity design; small BNT162b2 comparator group (n = 76); potential for differential attrition bias at Day 197; and exclusion of immunocompromised participants.

 

Comments 8: Figure 1 could be improved by providing more detailed reasons for participant exclusion and discontinuation.

Response 8: Figure 1 has been comprehensively revised. The updated CONSORT diagram now shows side branches from the vertical connecting line between the first and second dose boxes indicating reasons for not receiving the second dose (9 HXP-GPOVac and 7 BNT162b2 participants missed the second dose visit). The per-protocol immunogenicity population is shown for all three key timepoints (D29, D43 [primary], and D197) for both arms. Reasons for not completing the study are also shown.

 

Comments 9: Many tables are information-dense and may need formatting improvements for better readability. The authors are encouraged to make sure of consistent terminology throughout the manuscript for seroconversion/seroresponse definitions.

Response 9: Done. The term "seroconversion" is now used consistently throughout the manuscript; "seroresponse" has been removed. Table formatting has been reviewed.

 

Comments 10: The discussion section has to be contextualize the observed antibody waning at Day 197 related to published data for inactivated and mRNA COVID-19 vaccines.

Response 10: Done. The Discussion now contextualizes waning against published data for both platforms, with reference [24] (Widge et al., N Engl J Med 2021) cited: "Antibody levels declined by six months after the second dose in both groups, consistent with the waning observed following primary COVID-19 vaccination with inactivated and mRNA platforms."

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

See attachment.

Comments for author File: Comments.pdf

Author Response

Comments 1: Please revise the manuscript title to include the design as follows: Safety and immunogenicity of a locally produced inactivated NDV-HXP-S COVID-19 vaccine (HXP-GPOVac) compared with BNT162b2: a phase II randomized, controlled, double-blind noninferiority trial in Thai adults.

Response 1: Done. The title has been updated exactly as suggested.

 

Comments 2: Suggested Abstract Addition (Methods): This was a phase II, randomized, double-blind, active-controlled non-inferiority trial... Neutralizing antibody response was assessed for non-inferiority compared to BNT162b2 based on a GMT ratio margin of 0.5.

Response 2: Done. The abstract Methods section now includes: "The primary immunogenicity analysis assessed non-inferiority of HXP-GPOVac compared with BNT162b2 based on the NT50 geometric mean titer ratio, with a prespecified non-inferiority margin of 0.5."

 

Comments 3: The analysis plan should clearly state the null and alternative hypotheses for the non-inferiority test. The justification provided (consistency with immunobridging principles used in regulatory guidance [24]) is a strong starting point. However, providing more specific context on why the 0.5 margin was chosen for this specific vaccine platform and population would further strengthen the manuscript's rigor.

Response 3: Done. The null and alternative hypotheses are now explicitly stated: "The null hypothesis was that the NT50 GMT ratio (HXP-GPOVac/BNT162b2) was less than 0.5; the alternative hypothesis was that the ratio was at least 0.5." The margin of 0.5 was prespecified in the study protocol prior to trial initiation and this is now stated directly. The conclusion is unchanged whether 0.5 or the more commonly applied 0.67 immunobridging threshold is applied, the lower bound of 0.39 falls below both thresholds.

 

Comments 4: Suggested Revision for the Results Section of the Abstract: At Day 43 (14 days after the second dose), NT50 GMTs were 505 in the HXP-GPOVac group and 993 in the BNT162b2 group. While seroconversion rates at Day 43 exceeded 95% for both vaccines, the prespecified non-inferiority criterion for the NT50 GMT ratio (lower bound of 95% CI > 0.5) was not met.

Response 4: Done. The abstract Results now reads: "The NT50 GMT ratio (HXP-GPOVac/BNT162b2) at Day 43 was 0.51 (95% CI: 0.39–0.67); the lower bound did not exceed the prespecified non-inferiority margin of 0.5, and non-inferiority was not established. Seroconversion rates at Day 43 were 97.6% for HXP-GPOVac and 97.1% for BNT162b2 (neutralizing antibody) and 98.6% and 97.1%, respectively (anti-spike IgG)."

 

Comments 5: Results: Report the GMT Ratio (HXP-GPOVac/BNT162b2) and its 95% CI at Day 43. Report the two-sided 95% CI for this ratio and mention that the lower bound of the 95% CI was required to exceed 0.5 to satisfy the non-inferiority criterion.

Response 5: Done. The following has been added to the Immunogenicity Results section: "The NT50 GMT ratio (HXP-GPOVac/BNT162b2) at Day 43 was 0.51 (95% CI: 0.39–0.67); the lower bound of 0.39 did not exceed the prespecified non-inferiority margin of 0.5, and non-inferiority was not established." The ratio is also presented in Table 4.

 

Comments 6: The text mentions block randomization but does not specify the block size or whether the block sizes were fixed or variable. It does not specify the software or mechanism used, nor the mechanism used to conceal the sequence until the moment of assignment.

Response 6: Done. The Randomization and Blinding section now reads: "The randomization schedule was generated by an unblinded statistician at the Center of Excellence for Biomedical and Public Health Informatics (BIOPHICS), Mahidol University, using computer software with block randomization stratified by age group. Block sizes were 8 for the 18–59 year stratum and 12 for the 60–75 year stratum. Allocation concealment was maintained through a centrally held randomization schedule accessed sequentially by unblinded study staff; study vaccines were prepared and administered in identical-appearing syringes to preserve blinding of participants, investigators, and outcome assessors."

 

Comments 7: Solicited AEs within 7 days of vaccination were predominantly mild. After the first dose, one or more solicited AEs were reported by 22.8% (95% CI: 17.4–28.8) of HXP-GPOVac recipients and 44.7% (95% CI: 33.3–56.6) of BNT162b2 recipients.

Response 7: Done. The abstract Results now includes: "one or more solicited local or systemic AEs were reported by 23.7% (95% CI: 18.3–29.8) of HXP-GPOVac recipients and 44.7% (95% CI: 33.3–56.6) of BNT162b2 recipients after the first dose." (The proportion of 23.7% is consistent with Table 2.)

 

Comments 8: Define the Full Analysis Set: Explicitly state if the FAS followed the ITT principle.

Response 8: Done. The following has been added to the Statistical Analysis section: "The Full Analysis Set followed the intention-to-treat principle, comprising all randomized participants analyzed according to their assigned treatment group."

 

Comments 9: Ensure Figure 1 (Participant Disposition) clearly shows how many participants were excluded from the PP population and for what specific reasons to satisfy ITT/PP reporting requirements.

Response 9: Figure 1 has been comprehensively revised based on the clinical study report. The updated CONSORT diagram shows side branches from the vertical connecting line between the first and second dose boxes indicating the number and reason for participants not receiving the second dose: 9 HXP-GPOVac and 7 BNT162b2 participants missed the second dose visit. The per-protocol immunogenicity populations are explicitly labelled for all three key timepoints: D29 (HXP-GPOVac n=215, BNT162b2 n=69), D43 primary (HXP-GPOVac n=211, BNT162b2 n=69), and D197 (HXP-GPOVac n=193, BNT162b2 n=66). No post-randomisation protocol violations resulting in per-protocol exclusions were identified in either arm.

 

Comments 10: The manuscript should be updated to explicitly state the assumptions and methods used to handle missing data. For continuous immunogenicity outcomes, mixed models for repeated measures (MMRM) were employed to account for missing observations without explicit imputation.

Response 10: Mixed models for repeated measures (MMRM) were not used in this study. All immunogenicity analyses were performed as complete-case analyses on observed data at each timepoint, without formal imputation, consistent with the prespecified statistical analysis plan. The following has been added to the Statistical Analysis section: "All immunogenicity analyses were performed on observed data without formal imputation. Participants with missing data at a given time point were excluded from analyses at that time point. The proportion of participants with missing immunogenicity data was small, and no systematic differential patterns of missingness were identified between treatment groups. Analyses assumed data were missing at random (MAR)."

 

Comments 11: Transitioning Figure 2 from a bar graph to a longitudinal line graph would significantly improve the clarity and impact of the manuscript.

Response 11: We have converted the y-axes of both panels to a logarithmic scale as recommended, substantially improving the visual informativeness of Figure 2. We have retained the bar graph format because the study has only four discrete sampling timepoints with no data between visits; connecting points with a line would imply a continuous trajectory not supported by the data. This approach is also consistent with the format used in the directly comparable phase I [11] and phase II Vietnam [13] NDV-HXP-S publications.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors  positively addressed most comments. The manuscript is suitable for publication in the present form.