Development and Validation of ELISA for In Vitro Diagnosis of SARS-CoV-2 Infection
Round 1
Reviewer 1 Report
The title it is clear, but the word High-performance for the test is not accurate since the reported sensibility ranges only from (53% to 63%). Moreover, the title does not provide information about a validation test study
Line 44: add a parentheses
Line 48: Incorrect date.
Line 49 should include a specific and clear source and specify whether the 0.5–3% refers to CFR or IFR. The range is broad and context-dependent (e.g., age, comorbidities, variant, healthcare access).
Lines: 68-73 (2.1 Study Design): The sample calculation as described is methodologically flawed; it corresponds to a prevalence estimate, not a diagnostic validation. Calculating sample sizes separately for sensitivity and specificity is recommended, based on expected values ​​and desired precision.
Lines 71-73: should add a few lines explaining why there are more subjects than the number in the sample calculation.
Lines 89-92(2.2 antigen selection): It would be useful to more clearly justify the choice of S1 and S2 as diagnostic antigens, as well as the technical reasons for their expression in E. coli and insect cells, respectively. References are missing
Lines 94-106 (2.3 S1 and S2 ELISA) : Although the ELISA protocol is well described, should add key details to ensure complete reproducibility, especially regarding the blocking step (volume, specific conditions, or product code, if applicable), even if the manufacturer's instructions were followed. References are missing.
Line 118 (2.4 Statistical analysis): would be advisable to provide a more solid justification for the choice of the fixed factor used in the CO calculation for ELISA-S2, as well as to include information on the distribution of negative data used in said estimate.
No statistical analyzes were included to formally compare groups. We suggest including comparative statistical analyzes between groups (such as proportions tests, McNemar tests, or DeLong tests for ROC curves), as well as describing the distribution of the data.
Furthermore, categorical variables (sex, age group, or disease severity) were not analyzed. And there is no mention of those that fell into the grey range or how they were treated in the analysis.
Should made a explicitly discuss about the low sensitivity of tests in the early stages of the disease, and discuss the aspects suggested in the data analysis.
Although the discussion compares the results of the ELISA-S1/S2 with some commercial assays, the discussion lacks direct references or comparative data to support its claims. It is important to compare with other validated tests to contextualize the test's performance to strengthen the validity of the conclusions.
Lines 23, 224, 248: clarify the use of the terms ‘cost-effective assay’ and ‘low-cost ELISA’ since the study does not perform a formal cost-effectiveness analysis, but is only a validation of the assay.
Line 35: rewrite the abstract paragraph, as the conclusion is inconsistent with the study's objectives, which, although not explicitly stated in the text, correspond to a diagnostic validation and not a cost analysis.
Lines 251–257 (Conclusion): We suggest reviewing and rewriting this section, as the absence of a formal cost-effectiveness analysis limits the robustness of the claims made. Clarify this issue to avoid conclusions that are not fully supported by the evidence presented.
Although the manuscript mentions informed consent and approval by an ethics committee, it does not clearly specify the origin of the samples, who was responsible for their collection, or whether they were stored in an accredited biobank.
It is suggested to perform an external validation with samples from other cohorts to strengthen the applicability and robustness of the test.
The writing is acceptable, although sometimes excessively technical.
Author Response
Question 1. The title is clear, but the word High-performance for the test is not accurate since the reported sensibility ranges only from (53% to 63%). Moreover, the title does not provide information about a validation test study.
Reply: We thank the reviewer for the insightful comments on our title. We acknowledge the concern regarding "High-performance" given the reported sensitivity, but we emphasize that this term reflects the overall diagnostic utility, particularly the S2-ELISA's exceptional specificity of 99.7% and high Diagnostic Odds Ratio of 590.6, which significantly outperform commercial assays and demonstrate minimal cross-reactivity, crucial for serological diagnostics in resource-limited settings. Our manuscript explicitly describes the study as the "development and validation" of ELISA assays in its Abstract and Introduction. To address the reviewer's suggestions and enhance clarity, we propose amending the title to:
Lines 2-3: "Development and Validation of ELISA for In Vitro Diagnosis of SARS-CoV-2 Infection".
Question 2. Line 44: add a parentheses.
Reply (line 47): Thank you for your valuable input. We confirm that the requested change for line 44 has been implemented, adding parentheses as you suggested.
Question 3. Line 48: Incorrect date.
Reply: Thank you for highlighting the incorrect date on Line 48 of the manuscript. We have reviewed the sentence and confirm that it contained an error. The corrected version now reads:
Lines 48-51: “In response to its swift global dissemination, the World Health Organization declared COVID-19 a Public Health Emergency of International Concern on January 30, 2020, and subsequently classified it as a pandemic on March 11 of the same year.”
Question 4. Line 49 should include a specific and clear source and specify whether the 0.5–3% refers to CFR or IFR. The range is broad and context-dependent (e.g., age, comorbidities, variant, healthcare access).
Reply: Thank you very much for your precise and important comments regarding Line 49. The mortality range of 0.5% to 3% cited in the original text is based on estimates of the Case Fatality Rate (CFR) derived from global case and death reports available through the WHO COVID-19 dashboard [1]. This range reflects the variation observed across different populations and contexts during the period relevant to our study (March to October 2020), and not the Infection Fatality Rate (IFR), which would generally be lower. To address your concern and avoid ambiguity, we have revised the sentence as follows:
Lines 51-54: “With case fatality rates varying greatly according to age, comorbidities and geograph-ical location, SARS-CoV-2 had infected over 777 million individuals and caused more than 7 million deaths by April 2025, with new cases continuing to be reported world-wide [1].”
Question 5. Lines: 68-73 (2.1 Study Design): The sample calculation as described is methodologically flawed; it corresponds to a prevalence estimate, not a diagnostic validation. Calculating sample sizes separately for sensitivity and specificity is recommended, based on expected values ​​and desired precision.
Reply: We agree with the reviewer regarding sample sizing. This was addressed and the sample size estimation was done according to expected performance and precision:
Lines 72-76: “Sample size was calculated using the Buderer’s formula [9], ensuring statistical robustness with at least 173 positive and 62 negative samples (considering an expected sensitivity of 90%, expected specificity of 85%, 95% confidence interval (95% CI), with precision within 10%, and an expected disease prevalence of 20%). In order to ensure result robustness a total, 691 serum samples were analyzed.”
Question 6. Lines 71-73: should add a few lines explaining why there are more subjects than the number in the sample calculation.
Reply: Thank you for this pertinent comment regarding lines 71-73. We appreciate the opportunity to clarify this point. Our study ultimately included a significantly larger and more comprehensive set of samples, totaling 691 serum samples (354 from RT-qPCR-positive hospitalized patients and 337 pre-pandemic blood donor samples). This decision was made to enhance the statistical power, representativeness, and overall robustness of our findings, allowing for more detailed analyses such as the time-stratified sensitivity evaluation and the comprehensive cross-reactivity assessment, which benefit from a larger dataset. Therefore, the number of subjects analyzed exceeded the calculated minimum to strengthen the study's conclusions. We have included the following sentence to clarify the choice of using a larger number of samples:
Lines 75-76: “In order to ensure result robustness a total, 691 serum samples were analyzed.”
Question 7. Lines 89-92(2.2 antigen selection): It would be useful to more clearly justify the choice of S1 and S2 as diagnostic antigens, as well as the technical reasons for their expression in E. coli and insect cells, respectively. References are missing.
Reply: Thank you very much for your insightful comments regarding Section 2.2 (“Antigen Selection,” lines 89–92). The selection of the S1 and S2 subunits of the SARS-CoV-2 spike protein was based on their critical role in viral entry and their strong immunogenicity, which makes them prime targets for antibody detection. The S1 subunit mediates receptor binding, while the S2 subunit facilitates membrane fusion, both essential for infection. Antibodies against these subunits, particularly IgG, are frequently generated during and after infection and have shown potential for sustained detection, making them suitable for serodiagnosis and seroprevalence studies. Several studies have demonstrated that ELISAs based on spike subunits (particularly RBD and S2) often display higher specificity and longer-lasting responses compared to nucleocapsid-based assays [14–17]. Regarding antigen production, the truncated S1 antigen (ref. DAGC157) was expressed in Escherichia coli, a cost-effective system appropriate for simpler protein constructs that do not require complex post-translational modifications. In contrast, the S2 subunit (ref. DAGC153), covering residues Ser686–Pro1213, was expressed in insect cells using a baculovirus system. This eukaryotic platform allows proper protein folding, disulfide bond formation, and glycosylation, features essential for maintaining the conformational epitopes of the more complex S2 structure, thereby preserving its immunoreactivity. We have added the following sentence to Section 2.2 to reflect these justifications:
Lines 102-106: “The selection of S1 and S2 as target antigens was based on their essential roles in viral entry and their strong immunogenic potential. Expression systems were chosen ac-cording to the structural complexity of each subunit: S1, with fewer post-translational modifications, was suitably expressed in E. coli, while the more structurally intricate S2 subunit required expression in eukaryotic insect cells."
Question 8. Lines 94-106 (2.3 S1 and S2 ELISA): Although the ELISA protocol is well described, it should add key details to ensure complete reproducibility, especially regarding the blocking step (volume, specific conditions, or product code, if applicable), even if the manufacturer's instructions were followed. References are missing.
Reply: Thank you very much for your feedback regarding the description of the ELISA protocol. You are absolutely correct that the inclusion of additional information is essential to ensure full reproducibility of the immunoassay. In response, we have carefully revised Section 2.3 (lines 94–106) to incorporate key details of the protocol, particularly regarding the blocking step.
Lines 108-124: “Optimal dilutions for antigen coating, serum samples, and enzyme-conjugated antibodies were determined using checkerboard titration. Flat-bottom 96-well poly-styrene microplates (Corning® Costar, Glendale, AZ, USA) were coated overnight at 4°C with 200 ng/well of S1 or 100 ng/well of S2 antigens in 100 µL carbonate-bicarbonate buffer (pH 9.6). Plates were blocked with 100 µL WellChampion™ synthetic blocking buffer (Ken-En-Tec Diagnostics A/S, Taastrup, Denmark) according to the manufacturer's instructions for 90 minutes. Serum samples were diluted 1:25 (S1-ELISA) or 1:200 (S2-ELISA) in 100 µL PBS containing 0.05% Tween-20 (PBS-T, pH 7.4) and incubated at 37°C for 1 hour. After washing five times with 250 µL PBS-T, 100 µL horseradish peroxidase (HRP)-conjugated goat anti-human antibodies (Bio-Manguinhos, Fiocruz/RJ, Brazil) were added at a 1:40,000 PBS-T and incubated for 30 minutes at room temperature. New washing cycle five times with 250 µL PBS-T was repeated. Color development was achieved using 100 µL TMB Plus substrate (Ken-En-Tec Diagnostics A/S). After a 15-minute incubation at room temperature in the dark, the reactions were stopped with 50 ?l 0.3 M H2SO4, and absorbance was read at 450 nm using a microplate reader (SPECTRAmax 340PC®, San José, CA, USA).“
Question 9. Line 118 (2.4 Statistical analysis): it would be advisable to provide a more solid justification for the choice of the fixed factor used in the CO calculation for ELISA-S2, as well as to include information on the distribution of negative data used in said estimate.
Reply: We analyzed a set of formulas and chose the one that closest resembled ROC cutoff, considering that the optimal cut-off point on the ROC curve is the point that maximizes the test's accuracy. Within our set of formulas, we chose a formula for the S2 antigen that reliably maintained the test’s accuracy; however, we couldn’t determine a formula that reliably maintained the test’s accuracy for the S1 antigen. Therefore, the ROC-based cutoff was employed for the S1 ELISA.
Question 10. No statistical analysis was included to formally compare groups. We suggest including comparative statistical analyses between groups (such as proportions tests, McNemar tests, or DeLong tests for ROC curves), as well as describing the distribution of the data.
Reply: Thank you for this valuable suggestion. We agree that including statistical significance testing when comparing the performance of our in-house assays to commercial kits will significantly strengthen the robustness of our findings. To address this, we will perform the McNemar test to assess the statistical significance of the differences in diagnostic performance between our S1-ELISA and S2-ELISA and the commercial assays. The McNemar test is particularly suitable for comparing the agreement and disagreement between two diagnostic tests applied to the same set of samples, which aligns perfectly with our study design. We have inserted new information into several parts of the manuscript:
Lines 149-150: 2.4 Statistical Analysis: "Differences between paired proportions were evaluated using McNemar’s test, with p < 0.05 indicating statistical significance."
Lines 180-182: Results: "No statistically significant difference in diagnostic performance was observed between the S1-ELISA and the S2-ELISA (p = 0.1078; OR = 1.42; 95% CI: 0.93–2.17)."
Lines 196-202: Results: "McNemar’s test was used to compare performance. The S1-IgG ELISA showed no significant difference compared to the GOLD ELISA (p = 0.5962; OR = 1.19; 95% CI: 0.69–2.09), but significantly outperformed EURO IgG (p = 0.0001; OR = 0.32; 95% CI: 0.20–0.52) and EURO IgM assays (p = 0.0015; OR = 0.51; 95% CI: 0.33–0.78). The S2-IgG ELISA demonstrated statistically better performance than all commercial assays: GOLD ELISA (p = 0.0043; OR = 0.40; 95% CI: 0.20–0.76), EURO IgG (p = 0.0001; OR = 0.11; 95% CI: 0.05–0.22), and EURO IgM (p = 0.0001; OR = 0.23; 95% CI: 0.13–0.38)."
Question 11. Furthermore, categorical variables (sex, age group, or disease severity) were not analyzed.
Reply: Thank you for bringing this important point to our attention. We have added a new statement to the Results section to address this issue:
Lines 159-167: "The study included 691 samples. Among the 128 individuals infected with SARS- CoV-2, the median age was 45 years (IQR: 32.5–66.0), and the female-to-male ratio was 0.85:1. Of these, 82 individuals (64%) were admitted to an intensive care unit (ICU). Clinical outcomes were available for 122 individuals: 18 (14.8%) died, while 104 (85.2%) recovered and were discharged. The most commonly reported symptoms included fever in 57.1% (64/112), cough in 70.8% (75/106), and dyspnea in 58.7% (64/109). No age information was available for the SARS-CoV-2-negative group (n = 354), which had a female-to-male ratio of 1:1.6. All blood donors were in residents of the state of Bahia, Brazil."
Question 12. Should made an explicitly discuss about the low sensitivity of tests in the early stages of the disease, and discuss the aspects suggested in the data analysis.
Reply: Thank you for this important suggestion. Serological assays detect the adaptive immune response, which takes time to develop the following infection. Consequently, antibodies are typically not detectable in the very early stages of SARS-CoV-2 infection, often before or shortly after symptom onset and while viral RNA is still high. This biological reality means that serological tests are not suitable for diagnosing acute, early infections and should not be used as a substitute for molecular tests (like RT-qPCR) in this context. Our data analysis directly supports this point. As shown in Table 2, the sensitivity of all immunoassays was significantly lower in samples collected after 15 days post-symptom onset compared to samples collected later in the course of the disease. The sensitivity progressively increased with increasing days post-symptom onset, reaching optimal levels in samples collected 15 days or more after symptom onset. This time-dependent increase in sensitivity is a key aspect suggested by our data analysis and is consistent with the known kinetics of antibody responses to SARS-CoV-2. In response to your suggestion, we have included a new sentence in the discussion section, as follows:
Lines 246-249: “Temporal analysis confirmed a common serological pattern: all assays showed limited sensitivity during the early phase (0–7 days post-symptom onset), followed by increased sensitivity peaking between 15–21 days. This trend, consistent with prior reports [13], reflects the dynamics of the humoral immune response.”
Question 13. Although the discussion compares the results of the ELISA-S1/S2 with some commercial assays, the discussion lacks direct references or comparative data to support its claims. It is important to compare with other validated tests to contextualize the test's performance to strengthen the validity of the conclusions.
Reply: Thank you for your valuable comment regarding the comparison of our in-house ELISAs with commercial assays in the Discussion section. We agree that contextualizing our test's performance in relation to other validated assays is important to strengthen the validity of our conclusions. As you rightly noted, including additional comparative data or references to a wider range of commercial tests would enhance the discussion. However, during the period in which this study was conducted (March to October 2020), only a limited number of commercial serological assays were widely available and routinely used in Brazil. Specifically, the three commercial kits included in our analysis (GOLD ELISA COVID-19 IgG+IgM, Anti-SARS-CoV-2 NCP ELISA IgG, and Anti-SARS-CoV-2 NCP ELISA IgM) were among the most accessible and relevant diagnostic tools within the national context at the time. For this reason, we deliberately focused our comparative analysis on these assays, as they represented the primary reference tests in Brazil during the early stages of the pandemic. Our goal was to offer a performance comparison that would be directly relevant to the diagnostic landscape in which the S2-ELISA is intended to be implemented, particularly in resource-limited settings. While we acknowledge that comparisons with a broader range of internationally validated tests could offer additional insight, we believe that our focus on locally available and widely used assays provides a practical and contextually meaningful benchmark. The performance metrics presented in Table 1 and discussed in Section 4, especially regarding specificity and diagnostic odds ratio, demonstrate the competitive performance of the S2-ELISA relative to these commercial options. We appreciate your suggestion and have clarified this rationale in the revised Discussion section to make our reasoning more explicit. In response to your suggestion, we have included a new paragraph in the discussion section, as follows:
Lines 267-273: “It is important to contextualize these findings within the diagnostic landscape in Brazil between March and October 2020, when the comparator assays were in wide-spread clinical use. Although additional international commercial kits could of-fer further comparisons, our study focused on assays that were available and relevant within the Brazilian public health context. This pragmatic approach strengthens the relevance of our findings for low-resource and decentralized settings, where access, affordability, and simplicity are key operational priorities.”
Question 14. Lines 23, 224, 248: clarify the use of the terms ‘cost-effective assay’ and ‘low-cost ELISA’ since the study does not perform a formal cost-effectiveness analysis, but is only a validation of the assay.
Reply: Thank you for pointing out the need to clarify our use of the terms “cost-effective assay” and “low-cost ELISA” in lines 23, 224, and 248. We appreciate this observation, as it is essential that our wording accurately reflects the scope and limitations of the study. You are correct in noting that this work focuses on the development and diagnostic validation of the S1 and S2 ELISAs and does not include a formal cost-effectiveness analysis. The terms “cost-effective” and “low-cost” were originally used to highlight the potential operational advantages of producing in-house assays using recombinant antigens, which generally incur lower production costs than commercial kits, especially in the context of large-scale implementation in resource-limited settings. Nonetheless, to prevent misinterpretation and ensure alignment with the study's objectives, we have revised the language in lines 23, 224, and 248. The updated text now avoids definitive claims and instead emphasizes the potential for cost-effectiveness or describes the assays as accessible or suitable for resource-constrained settings, based on their simplified design and robust diagnostic performance. These changes ensure the terminology remains consistent with the evidence presented in the manuscript.
Question 15. Line 35: rewrite the abstract paragraph, as the conclusion is inconsistent with the study's objectives, which, although not explicitly stated in the text, correspond to a diagnostic validation and not a cost analysis.
Reply: Thank you for pointing out the need to revise the conclusion paragraph of the Abstract to ensure consistency with the study’s primary objective, which focuses on the diagnostic validation of the S2-ELISA. We agree that the previous phrasing may have inadvertently suggested that a formal cost analysis was performed, which was not within the scope of this work. In response to your suggestion, we have revised the final paragraph of the Abstract as follows:
Lines 35-40: “Conclusions: The S2-ELISA offers a robust, highly specific, and operationally simple tool for serological detection of SARS-CoV-2 infection. Its strong diagnostic performance and accessibility make it well suited for implementation in diverse epidemiological settings, particularly where molecular testing is limited. The development of affordable, validated serological assays such as this is critical for strengthening surveillance, under-standing transmission dynamics, and informing public health responses.”
Question 16. Lines 251–257 (Conclusion): We suggest reviewing and rewriting this section, as the absence of a formal cost-effectiveness analysis limits the robustness of the claims made. Clarify this issue to avoid conclusions that are not fully supported by the evidence presented.
Reply: Thank you for your valuable feedback regarding the Conclusion section (lines 251–257) and for highlighting the importance of aligning our claims with the evidence presented. We fully agree that a formal cost-effectiveness analysis was not performed in this study, and we appreciate the opportunity to clarify this point. Our reference to the S2-ELISA being "cost-effective" or representing a "viable alternative for expanding COVID-19 diagnostics, especially in diverse epidemiological settings with limited resources," was based on the assay's characteristics, namely, the use of a single recombinant antigen (S2) and the simplicity of the ELISA format. These features generally contribute to lower production and implementation costs when compared to commercial assays that rely on multiple antigens or more complex detection systems. In response to your suggestion, we have revised the Conclusion section (lines 251–257) to clarify that the assay shows potential for being a cost-effective alternative, based on its design and comparative diagnostic performance (see Table 1 and related discussion), rather than making a definitive statement unsupported by formal economic analysis. The revised version now more accurately reflects the scope and limitations of the current study:
Lines 297-304: “Taking together, our findings demonstrate that the S2 subunit is a promising alternative to traditional RBD- or N-based targets for serological testing. While RBD-based assays may offer marginally higher sensitivity when both IgG and IgM are detected, their complexity and cost can limit scalability. In contrast, the S2-ELISA delivers excellent diagnostic performance, minimal cross-reactivity, and practical advantages stemming from its simple, single-antigen design. These features make it a compelling option for expanding COVID-19 diagnostic capacity in decentralized and resource-constrained settings, and a cost-effective tool for large-scale serological surveillance.”
Question 17. Although the manuscript mentions informed consent and approval by an ethics committee, it does not clearly specify the origin of the samples, who was responsible for their collection, or whether they were stored in an accredited biobank.
Reply: Thank you for raising this important point regarding sample origin, collection, and storage. We appreciate the opportunity to provide further clarification on these aspects, which are indeed crucial for transparency and reproducibility. The pre-pandemic serum samples were obtained from the Bahia State Blood Bank Foundation (HEMOBA) in Salvador, Brazil. These samples were collected by HEMOBA as part of their routine blood donation procedures in 2016 and 2017, well before the emergence of SARS-CoV-2. As they were collected under blood bank protocols, individual RT-qPCR confirmation was not possible due to donor anonymity, as noted in the study context. The SARS-CoV-2-positive serum samples were collected from hospitalized patients confirmed by RT-qPCR at Aliança D’Or Hospital and Aeroporto Hospital in Salvador and surrounding areas between March and October 2020. The collection of these samples was a collaborative effort involving the clinical teams at these hospitals and the research team from the Gonçalo Moniz Institute (Fiocruz-BA), as reflected in the author affiliations. Following collection, all samples were transported to and stored at the laboratories of the Gonçalo Moniz Institute (Fiocruz-BA) under appropriate conditions for serological analysis as per standard laboratory procedures. While HEMOBA operates as a recognized blood bank with established storage protocols, the samples utilized specifically for this research were processed and maintained within the research facilities at Fiocruz-BA for the duration of the study. As detailed in the manuscript, the study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board for Human Research at the Gonçalo Moniz Institute (IRB/IGM/Fiocruz-BA), and informed consent was obtained from all subjects involved in the study. In response to your suggestion, we have added the following sentence at the 2.1 Study Design and Sample Collection section:
Lines 78-83: “Peripheral blood (9 mL) was collected using clot-activator tubes with gel separators (Vacuplast, Brazil) in collaboration with hospital clinical teams. After a 30-minute clotting period at room temperature, samples were centrifuged at 3,000 × g for 10 minutes. Serum was aliquoted into cryotubes, transported to the Gonçalo Moniz Insti-tute (Fiocruz-BA), and stored at −20 °C until analysis.”
Question 18. It is suggested to perform an external validation with samples from other cohorts to strengthen the applicability and robustness of the test.
Reply: Thank you for this important suggestion. We fully agree that performing an external validation of the SI-ELISA and S2-ELISA using samples from other cohorts is a crucial next step. Our current study provides an initial validation within a specific cohort. However, to truly strengthen the applicability and demonstrate the robustness of the test across different populations, geographical regions, and varying epidemiological contexts, external validation is indispensable. We consider this external validation a high priority for future work. Such studies will provide valuable data on the assay's performance in diverse settings, which is essential for assessing its generalizability and informing its potential for wider implementation in public health strategies. In response to your suggestion, we have added the following sentence at the end of the limitations paragraph in the Discussion section:
Lines 293-295: “Future work will include external validation using independent sample cohorts to assess generalizability and reinforce the assay’s applicability across different epidemiological settings.”
Question 19. The writing is acceptable, although sometimes excessively technical.
Reply: Thank you for your feedback regarding the writing style. We understand the need for clarity and accessibility in scientific communication, while also maintaining the necessary precision for describing complex methods and results. We have reviewed the manuscript with this specific comment in mind and have made efforts to simplify sentence structures and clarify terminology where possible, without compromising the scientific accuracy.
Author Response File: Author Response.pdf
Reviewer 2 Report
The manuscript entitled "High-Performance of ELISA for In Vitro Diagnosis of SARS-2 CoV-2 Infection" is a good piece of work toward development of diagnostics. However, there are some comments need to be addressed.
It was a pleasure to review the article entitled “High-Performance of ELISA for In Vitro Diagnosis of SARS-2 CoV-2 Infection “ which describe diagnostic test development, optimization, and validation of high performance, low-cost ELISA assays using recombinant antigens for the detection of SARS-CoV-2-specific IgG antibodies.
I was in the first line for SARS-2 CoV-2 testing during pandemics, and I understand the importance of fast and accurate method of detection.
I have some comments that would need to be addressed.
Abstract
Correct the number 338 line no. 26 in abstract, if I am not wrong it must be 337 to match the total.
Introduction
Line no. 58 “labor-intensive, time-consuming,” I am not sure I can agree with this about the PCR which is used globally in the SARS-2 CoV-2 testing.
Line no. 63 “low-cost ELISA assays” , did you considered the cost comparison between the developed ELISA assay and commercial ELISA assay?
Materials and Methods
Line no. 78 “All COVID-19-positive cases were confirmed by RT-qPCR” , what sample were considered for PCR?
Line no. 84 “88 pre-pandemic samples” what was the time duration of these samples? Was these samples confirmed by RT-qPCR?
What was the total time required for the developed ELISA assay (S2 antigen) and commercial ELISA assay?
Figure1, potentially eligible candidates, What was eligible criterion.
Results
Table 1, full form for the sensitivity, specificity, accuracy may be considered.
Table 2, Line no. 179 “symptom onset (PSO)” a word post is missing here.
EURO IgG 2 , ELISA-S1 IgG , ELISA-S2 IgG , Sensitivity (%) is decreasing after 22 days PSO , what could be te possible factor.
Discussion
Line no. 187, “with all 86 samples testing negative,” if I am not wrong it must be 88.
Line no. 204, “with 0 false positives among 86 pre-pandemic” if I am not wrong it must be 88.
Author Response
Dear Reviewer, thank you very much for your consideration. Please find our point-by-point replies below:
Abstract. Correct the number 338-line no. 26 in abstract, if I am not wrong it must be 337 to match the total.
Reply: Thank you for this careful observation. You are absolutely correct. We appreciate you pointing out the discrepancy in the abstract (line 26). The number of pre-pandemic blood donor samples should indeed be 337, to align with the total sample count and the details provided in the Methods section. We have promptly corrected this to "337" (line 26) in the revised manuscript.
Introduction. Line no. 58 “labor-intensive, time-consuming,” I am not sure I can agree with this about the PCR which is used globally in the SARS-2 CoV-2 testing.
Reply: Thank you for raising this important point regarding the characterization of PCR in the Introduction (line 58). We appreciate your perspective. We fully agree that RT-qPCR has been, and continues to be, a globally indispensable tool for SARS-CoV-2 testing and remains the gold standard for direct viral detection. Our intention with the statement "labor-intensive, time-consuming" was not to diminish the critical role or efficacy of RT-qPCR, but rather to highlight its operational demands in a comparative context, particularly when considering its implementation in resource-limited settings—a key focus articulated in our abstract and introduction. As stated in the Abstract: "The ongoing global health challenges posed by SARS-CoV-2 necessitate reliable and accessible diagnostic tools, especially in resource-limited settings where RT-qPCR may be impractical." And further elaborated in the Introduction: "However, these tests are labor-intensive, time-consuming, and require specialized personnel and costly infrastructure. Moreover, RT-qPCR sensitivity can decline over time due to reduced viral load, potentially leading to false-negative results [6]. Serological assays, such as ELISAs, offer operational advantages by being more affordable, faster, and less reliant on complex laboratory infrastructure [7,8]." Our aim was to emphasize the operational advantages of serological assays like ELISA, which are generally more affordable, faster, and require less complex infrastructure and specialized personnel compared to RT-qPCR, making them particularly valuable for expanding diagnostic capabilities in diverse epidemiological settings with limited resources. To ensure this nuance is clearer, we can rephrase the sentence slightly to emphasize the comparative and contextual nature of the statement. For instance, we could modify it to:
Lines 61-63: “However, RT-qPCR is labor-intensive, time-consuming, and dependent on specialized personnel and infrastructure—limitations that are particularly restrictive in re-source-limited settings.”
Introduction. Line no. 63 “low-cost ELISA assays”, did you consider the cost comparison between the developed ELISA assay and commercial ELISA assay?
Reply: Thank you for this important question regarding the characterization of our developed ELISA assays as "low-cost" in the Introduction (line 63). Yes, a comprehensive cost comparison between our developed ELISA assays and commercial ELISA assays was indeed conducted. This detailed economic evaluation was a significant component of the doctoral study of one of the authors. However, the specific data and in-depth analysis of this cost comparison were not included in the scope of the present manuscript, which primarily focuses on the diagnostic performance and validation of the assays. To ensure clarity and avoid implying a detailed cost analysis within this paper, we will rephrase the statement in line 63. We intend to align it more closely with the conclusion presented in the abstract, which describes the S2-ELISA as a "cost-effective" alternative, reflecting its potential for broader accessibility rather than a direct, quantified cost comparison within the manuscript itself.
Materials and Methods. Line no. 78 “All COVID-19-positive cases were confirmed by RT-qPCR”, what sample were considered for PCR?
Reply: Thank you for this excellent question. The RT-qPCR confirmations for the COVID-19-positive cases, which originated from 354 samples collected from 128 hospitalized individuals at Aliança and Aeroporto Hospitals between March and October 2020, were performed as part of their routine diagnostic workup. These tests were processed by a central laboratory in Bahia (State Laboratory). Therefore, all patients included in our study had their COVID-19 status molecularly confirmed prior to the collection of serum samples for our serological assays.
Materials and Methods. Line no. 84 “88 pre-pandemic samples” what was the time duration of these samples? Were these samples confirmed by RT-qPCR?
Reply: Thank you for your inquiry regarding the 88 pre-pandemic samples mentioned in line 84. These samples were collected between 2013 and 2017. They were obtained from the blood bank in Pernambuco and Bahia, Brazil, as well as the State laboratory of Pernambuco. Due to the nature of their collection and the ethical considerations involved, direct access to these anonymous blood donors for RT-qPCR confirmation was not feasible. These samples served as a crucial control group to assess potential cross-reactivity of our developed assays with antibodies from other infectious diseases, as they predated the SARS-CoV-2 pandemic.
Materials and Methods. What was the total time required for the developed ELISA assay (S2 antigen) and commercial ELISA assay?
Reply: Thank you for your question regarding the time required to perform the developed S2-ELISA and a commercial ELISA assay. The total time required to perform both the in-house S2-ELISA and a typical commercial ELISA assay is quite similar, generally taking approximately 2 hours from sample preparation to result reading. The primary difference in procedure lies in the initial step. For the in-house ELISA, an additional step is required to sensitize microplates with the antigen. This plate coating process typically takes no more than 30 minutes of hands-on time, although incubation periods may vary depending on the specific protocol. Once the plates are coated (which can be done in advance), the subsequent steps of sample incubation, washing, conjugate incubation, and substrate development are comparable in duration to those of commercial kits.
Materials and Methods. Figure1, potentially eligible candidates. What was eligible criterion.
Reply: As described in detail within the Materials and Methods section, specifically Section 2.1, the study population was drawn from distinct groups to ensure a robust evaluation of our developed assays. The potentially eligible candidates for the SARS-CoV-2 positive cohort were individuals who were hospitalized due to COVID-19 in Salvador, Brazil, and surrounding areas between March and October 2020. A key criterion for their inclusion was a confirmed diagnosis of SARS-CoV-2 infection by RT-qPCR, along with the presence of documented clinical symptoms. This ensured that the positive samples used for validation were from individuals with a verified infection. For the negative control group, potentially eligible candidates were healthy individuals whose samples were collected prior to the emergence of SARS-CoV-2, specifically before December 2019. These samples were obtained from a blood bank and served as a baseline to assess the specificity of the assays in a population unexposed to the virus. Additionally, a panel of pre-pandemic samples from individuals diagnosed with other infectious and parasitic diseases was included to evaluate potential cross-reactivity, although the primary "potentially eligible candidates" in the flowchart context refer mainly to the main positive and negative cohorts. Therefore, the eligibility criteria were designed to select appropriate samples representing confirmed SARS-CoV-2 infections and relevant control populations for the validation of the serological assays.
Results. Table 1, full form for the sensitivity, specificity, accuracy may be considered.
Reply: Thank you for your valuable comment regarding Table 1 in the Results section. We appreciate you pointing out the potential for improving clarity by providing the full forms of the abbreviations used. We agree that explicitly stating the full terms for Sensitivity (Sen), Specificity (Spe), and Accuracy (Acc) will enhance the readability and understanding of the table for all readers. In the revised version of the manuscript, we will incorporate the full forms of these metrics. This will likely be done by adding a footnote to the table or by including the full terms alongside the abbreviations within the table header or caption, ensuring that the meaning of each metric is immediately clear.
Results. Table 2, Line no. 179 “symptom onset (PSO)” a word post is missing here.
Reply: Thank you for carefully reviewing Table 2 and pointing out the omission in the caption on line 179. You are absolutely correct; the word "post" is indeed missing before "symptom onset (PSO)". This was an oversight during the preparation of the manuscript. We appreciate you identifying this error, and we will ensure that the caption for Table 2 is corrected in the revised manuscript to read "Sensitivity (%) of serological assays stratified by time since post-symptom onset (PSO)".
Results. EURO IgG 2, ELISA-S1 IgG, ELISA-S2 IgG, Sensitivity (%) is decreasing after 22 days PSO, what could be the possible factor.
Reply: The observed decrease in sensitivity for the EURO IgG 2, ELISA-S1 IgG, and ELISA-S2 IgG assays beyond 22 days post-symptom onset is likely influenced by several factors. Primarily, it can be attributed to the natural waning of SARS-CoV-2 specific IgG antibody titers over time in some individuals, potentially causing levels to fall below the assay's detection threshold or cut-off value. Individual variations in the magnitude and persistence of the antibody response, influenced by factors such as disease severity and immune status, also contribute to this variability. Furthermore, changes in antibody affinity or quality over time might affect their binding efficiency to the recombinant antigens used in the assays. It is also worth considering that the sample size for the later time point group is smaller, which could impact the precision of the sensitivity estimate for that specific window. Despite this decline from the peak, the sensitivity in the later phase remains valuable for detecting past infections in retrospective studies.
Discussion. Line no. 187, “with all 86 samples testing negative,” if I am not wrong it must be 88.
Reply: You are absolutely correct in your observation. Upon reviewing the Materials and Methods section (Section 2.1), we confirm that the panel used to assess potential cross-reactivity consisted of 88 pre-pandemic samples from individuals diagnosed with other infectious and parasitic diseases. The mention of "86 samples" on line 187 in the Discussion is indeed a typographical error. We will promptly correct this in the revised version of the manuscript to accurately reflect that the cross-reactivity panel comprised 88 samples.
Discussion. Line no. 204, “with 0 false positives among 86 pre-pandemic” if I am not wrong it must be 88.
Reply: You are absolutely correct in your assessment. As we discussed regarding the previous comment, the panel specifically evaluated cross-reactivity consisting of 88 pre-pandemic samples from individuals with other infectious diseases, as detailed in the Materials and Methods (Section 2.1). The mention of "86 pre-pandemic samples" on line 204 is indeed another instance of the same typographical error.
Author Response File: Author Response.pdf
Reviewer 3 Report
The author has presented results on the development, optimization, and validation of two in-house ELISA assays (S1-ELISA and S2-ELISA) for detecting anti-SARS-CoV-2 IgG antibodies, showing promising performance compared to commercial kits. Addressing the following concerns can improve the paper:
- Authors need to clarify the novelty and contribution of this work compared to the many previously published ELISA-based SARS-CoV-2 diagnostics, emphasizing how it advances the current field.
- It is needed to expand the limitations section to discuss the use of only hospitalized patient samples and the lack of evaluation in mild or asymptomatic cases, which affects generalizability.
- Authors should include statistical significance testing when comparing in-house assays to commercial kits and provide interpretation of why differences in performance were observed.
- It will be beneficial to discuss operational aspects such as assay reproducibility across batches, scalability, and steps required for regulatory approval, to strengthen the translational impact.
- English language of the paper can be improved.
I have provided the information in major comments.
Author Response
Dear Reviewer, thank you very much for your consideration. Please find our point-by-point replies below:
Question 1. Authors need to clarify the novelty and contribution of this work compared to the many previously published ELISA-based SARS-CoV-2 diagnostics, emphasizing how it advances the current field.
Reply: Thank you for this important question. We appreciate the opportunity to clarify how our study advances the current field. While numerous serological assays for SARS-CoV-2 have been developed, there remains a critical need for reliable, accessible, and cost-effective diagnostic tools, particularly in resource-limited settings where gold-standard RT-qPCR methods can be impractical due to infrastructure and cost constraints. Our study directly addresses this gap by focusing on the development and validation of two cost-effective in-house ELISA assays utilizing recombinant S1 and S2 antigens. The primary novelty and contribution of our work lies in the detailed evaluation and demonstration of the superior diagnostic performance of the S2-based ELISA, particularly its exceptionally high specificity (99.7%) and minimal cross-reactivity compared to the S1-based assay and several commercially available kits. As highlighted in our results and discussion, while other antigens like N or S1 are commonly used, assays targeting these can sometimes exhibit reduced specificity or increased cross-reactivity in real-world conditions. Our findings strongly suggest that the S2 subunit is a highly suitable antigenic target for serodiagnosis, offering a robust performance profile. Furthermore, we emphasize the practical implications of our S2-ELISA. Its development as an in-house assay, coupled with its demonstrated high performance, particularly in specificity and diagnostic odds ratio, presents a viable alternative to more complex or expensive commercial options. The operational simplicity and potential for cost-effectiveness of a single-antigen, IgG-only S2-ELISA make it particularly advantageous for expanding COVID-19 diagnostics in diverse epidemiological settings with limited resources.
Question 2. It is needed to expand the limitations section to discuss the use of only hospitalized patient samples and the lack of evaluation in mild or asymptomatic cases, which affects generalizability.
Reply: Thank you for this important comment regarding the limitations of our study, specifically concerning the sample population used for validation. We agree that this is a crucial point to address for a comprehensive understanding of our findings. We acknowledge that our sample panel for evaluating diagnostic performance consisted predominantly of hospitalized individuals who experienced moderate to severe COVID-19. This is indeed a limitation, as the magnitude and kinetics of antibody responses can vary depending on disease severity. Individuals with mild or asymptomatic infections may develop lower antibody titters or have different seroconversion patterns compared to those with more severe disease. Consequently, the sensitivity estimates reported in our study, which are based on this cohort of hospitalized patients, may not be directly generalizable to populations with a higher proportion of mild or asymptomatic cases. Evaluating the performance of the S2-ELISA in these less severe presentations is essential for a complete picture of its diagnostic utility across the full spectrum of SARS-CoV-2 infection. The limitations section has expanded in the revised manuscript to clearly articulate this point. We have explained that while our study provides a solid validation in a cohort of hospitalized patients, further studies are necessary to assess the assay's performance in individuals with mild or asymptomatic infections to fully understand its generalizability to broader community settings. Therefore, we have included the following sentence in the limitations section:
Lines 275-280: “This study has limitations. First, the sensitivity estimates were predominantly from hospitalized individuals with moderate to severe COVID-19. As disease severity is known to influence the magnitude and kinetics of antibody responses, our results may overestimate assay sensitivity in mild or asymptomatic cases. Individuals with milder illness may produce lower antibody titers or exhibit delayed seroconversion. Therefore, further studies evaluating S2-ELISA performance in community-based co-horts are essential for assessing its full diagnostic utility.”
Question 3. Authors should include statistical significance testing when comparing in-house assays to commercial kits and provide interpretation of why differences in performance were observed.
Reply: Thank you for this valuable suggestion. We agree that including statistical significance testing when comparing the performance of our in-house assays to commercial kits will significantly strengthen the robustness of our findings. To address this, we will perform the McNemar test to assess the statistical significance of the differences in diagnostic performance between our S1-ELISA and S2-ELISA and the commercial assays. The McNemar test is particularly suitable for comparing the agreement and disagreement between two diagnostic tests applied to the same set of samples, which aligns perfectly with our study design. We have inserted new information into several parts of the manuscript:
Lines 149-150: 2.4 Statistical Analysis: "Differences between paired proportions were evaluated using McNemar’s test, with p < 0.05 indicating statistical significance."
Lines 180-182: Results: "No statistically significant difference in diagnostic performance was observed between the S1-ELISA and the S2-ELISA (p = 0.1078; OR = 1.42; 95% CI: 0.93–2.17)."
Lines 196-202: Results: "McNemar’s test was used to compare performance. The S1-IgG ELISA showed no significant difference compared to the GOLD ELISA (p = 0.5962; OR = 1.19; 95% CI: 0.69–2.09), but significantly outperformed EURO IgG (p = 0.0001; OR = 0.32; 95% CI: 0.20–0.52) and EURO IgM assays (p = 0.0015; OR = 0.51; 95% CI: 0.33–0.78). The S2-IgG ELISA demonstrated statistically better performance than all commercial assays: GOLD ELISA (p = 0.0043; OR = 0.40; 95% CI: 0.20–0.76), EURO IgG (p = 0.0001; OR = 0.11; 95% CI: 0.05–0.22), and EURO IgM (p = 0.0001; OR = 0.23; 95% CI: 0.13–0.38)."
Question 4. It will be beneficial to discuss operational aspects such as assay reproducibility across batches, scalability, and steps required for regulatory approval, to strengthen the translational impact.
Reply: Thank you for this valuable observation. We agree that discussing operational aspects and the path to regulatory approval is crucial for highlighting the translational impact of a diagnostic assay. Our current study provides initial validation of the S1-, S2-ELISA in a specific cohort, demonstrating promising performance characteristics. However, as you rightly point out, before proceeding to detailed discussions of regulatory requirements, it is essential to conduct further comprehensive evaluations of the assay in diverse clinical settings and patient populations, including mild and asymptomatic cases, and across different geographical locations. These subsequent studies are necessary to fully assess the assay's reproducibility, scalability in real-world conditions, and overall performance across the full spectrum of SARS-CoV-2 infection. Demonstrating consistent performance in these varied scenarios is a critical prerequisite that will significantly strengthen the translational potential of the S1-, S2-ELISA and inform the specific steps required for formal regulatory submission.
Question 5. English language of the paper can be improved.
Reply: Thank you for your comment regarding the English language of the manuscript. We appreciate this feedback. We wish to inform you that the manuscript had previously undergone professional editing by a native English speaker specializing in scientific manuscripts to ensure clarity and accuracy. However, considering your observation, we have conducted a further thorough review of the text to identify and address any remaining points for potential language improvement. If, after this additional review, there are still specific areas where the language could be enhanced, we would be very grateful if you could kindly indicate these directly in the manuscript to help us make the necessary revisions.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
The authors have improved the paper.
The authors have addressed the concerns and improved the paper.