Comparison of SARS-CoV-2 Antigen Tests in Asymptomatic Testing of Passengers at German Airports under Time Constraints: Application of Three Different Antigen Test Formats

Round 1

Reviewer 1 Report

The Authors present a manuscript “Comparison of SARS-CoV-2 antigen tests in asymptomatic testing of passengers at German airports under time constraints: application of three different antigen test formats”

The study design is in part predicated on test kit and circumstances of testing and represents an incredible amount of work by the authors with data collected from over 35,000 antigen tests. With such a body of work, careful consideration must be given to elaborate on all parts of the study for clarity to the reader. As such there are some difficulties reading the manuscript for clarity, correctness and flow.

Part 1 Validation results in asymptomatic travellers

For Part 1(3.1) of the study, more information is required in the M&M regarding study design. ie where the PCR-positive samples came from, how they were selected, if/how they have been stored, where the antigen tests were performed, by who.

Table 1 requires information on the number of specimens in each group. ie how many of the 29 Roc-Ag-tested specimens had a CT >30, <30 or <26 cycles?

Discrepancy in stated true positive rate 29% (line 101) and 27.6% (table 1). Likewise false negatives 21/29 is not 71% as stated in the test (line 101)(would be 72%).

Lines 105-106 the CT-value ranges should reflect those described in table 1 ie >30 in table 1 >31 in text. This has lead to a discrepancy in false negative rates reported.

Lines 106-107 implies antigen test was performed on PCR negative swabs. This data should be summarised in table 1. Was this only for Roc-Ag. Was the specificity assessment performed on Q-Ag and FRB-Ag?

Line 122 consistency is required when reported limit of detection. Ie ~100,000 virus particles stated when previously used CT.

Line 129 consistency is required when reported limit of detection. Ie ~100,000 virus particles stated when previously used CT.

Lines 120-124, number of samples that had low virus particle content should be stated.

Part 2 (3.2)

For Part 2 of the study, M&M require more information about the study design.

            Where were the swabs tested

            Ethics

            Who undertook the testing

            Logistics involved

            Participants

Lines 154-155 discrepancy in stated FP rate compared with table 2

Part 3 – Normalization

This part of the manuscript is not written clearly. It appears that the authors wish to illustrate the percentage of Ag positive results that fall below a certain Ct threshold for the Quidell-Ag test and the Fujirebio-Ag test to highlight their respective performance differences. While the graphs (figure 3 ) look OK, I suspect this data would better be represented by table and numbers to illustrate the range in values /% differences. It is unclear how the conclusion for the final sentence (Line 182-184) is derived.

Discussion

Findings in the discussion are appropriate and fit well with the overall themes of the manuscript.

For the reader some mention of testing logistics would be helpful – around staffing required, issues, risk and biosafety/waste disposal.

Grammatical

Line 100, “summed up” should be changed to an alternative ie “accounted for”

Author Response

The Authors present a manuscript “Comparison of SARS-CoV-2 antigen tests in asymptomatic testing of passengers at German airports under time constraints: application of three different antigen test formats”

The study design is in part predicated on test kit and circumstances of testing and represents an incredible amount of work by the authors with data collected from over 35,000 antigen tests. With such a body of work, careful consideration must be given to elaborate on all parts of the study for clarity to the reader. As such there are some difficulties reading the manuscript for clarity, correctness, and flow.

Part 1 Validation results in asymptomatic travellers

For Part 1(3.1) of the study, more information is required in the M&M regarding study design. ie where the PCR-positive samples came from, how they were selected, if/how they have been stored, where the antigen tests were performed, by who.

Response: We agree with the reviewer and have included the required information in the M&M section (lane 61-64).

Table 1 requires information on the number of specimens in each group. ie how many of the 29 Roc-Ag-tested specimens had a CT >30, <30 or <26 cycles?

Response: We agree with the reviewer’s comment and have included 2 more tables including all relevant information. Table 1A) Results showing all ct values of the RT-PCR results 1B) Positive and negative AG results grouped according to ct values ≤26, ≤30 and ˃30 cycles 1C) Summary of Table 1B given in percent (formerly Table 1).

Discrepancy in stated true positive rate 29% (line 101) and 27.6% (table 1). Likewise false negatives 21/29 is not 71% as stated in the test (line 101)(would be 72%).

Response: We have now corrected the numbers in the text and also in Table 1C.

Lines 105-106 the CT-value ranges should reflect those described in table 1 ie >30 in table 1 >31 in text. This has led to a discrepancy in false negative rates reported.

Response: We agree with the reviewer and have now corrected the numbers.

Lines 106-107 implies antigen test was performed on PCR negative swabs. This data should be summarised in table 1. Was this only for Roc-Ag. Was the specificity assessment performed on Q-Ag and FRB-Ag?

Response: This has not been done consistently and not for Q-Ag and FRB-Ag.

Line 122 consistency is required when reported limit of detection. Ie ~100,000 virus particles stated when previously used CT.

Response: We agree with the reviewer and have adjusted the statement accordingly.

Line 129 consistency is required when reported limit of detection. Ie ~100,000 virus particles stated when previously used CT.

Response: As above, we have adjusted the statement using CT.

Lines 120-124, number of samples that had low virus particle content should be stated.

Response: We agree with the reviewer and have included additional information in the expanded Figure 1A-C.

Part 2 (3.2)

For Part 2 of the study, M&M require more information about the study design.

            Where were the swabs tested

            Ethics

            Who undertook the testing

            Logistics involved

            Participants

Response: A brief description of the logistics, staff and lab layout has been added in the M&M section (line 61-64) including a reference to ethics (line 78-80).

Part 3 – Normalization

This part of the manuscript is not written clearly. It appears that the authors wish to illustrate the percentage of Ag positive results that fall below a certain Ct threshold for the Quidell-Ag test and the Fujirebio-Ag test to highlight their respective performance differences. While the graphs (figure 3 ) look OK, I suspect this data would better be represented by table and numbers to illustrate the range in values /% differences. It is unclear how the conclusion for the final sentence (Line 182-184) is derived.

Response: The normalization has been used by Corman et al. (reference 4) to illustrate the unseen/missing results at higher ct. Nevertheless, we would prefer to have this complex figure retained.

Discussion
Findings in the discussion are appropriate and fit well with the overall themes of the manuscript.

Response: Thank you.

For the reader some mention of testing logistics would be helpful – around staffing required, issues

Response: We have outlined this briefly in the M&M section.

Grammatical
Line 100, “summed up” should be changed to an alternative ie “accounted for”

Response: We agree with the reviewer and have changed the sentence as suggested.

Reviewer 2 Report

Summary:

Pradas et al. manuscript provides experimental comparison and efficiency of antigen based test against SARS-CoV-2. Authors have compared the sensitivity of colorimetric (Roche), fluorometric (Quidel Sofia 2) and instrument-based chemiluminescent (Fuji-23 rebio Lumipulse® G) tests with last two yielding more than 90% sensitivity along with false positive in travelers at airport.  

Comment: 

It is well designed study with sufficient cohort size to support their findings.

Author Response

Summary:

Pradas et al. manuscript provides experimental comparison and efficiency of antigen based test against SARS-CoV-2. Authors have compared the sensitivity of colorimetric (Roche), fluorometric (Quidel Sofia 2) and instrument-based chemiluminescent (Fuji-23 rebio Lumipulse® G) tests with last two yielding more than 90% sensitivity along with false positive in travelers at airport.  

Comment:

It is well designed study with sufficient cohort size to support their findings.

Response: Thank you.