Review 1: "Performance of Screening for SARS-CoV-2 using Rapid Antigen Tests to Detect Incidence of Symptomatic and Asymptomatic SARS-CoV-2 Infection: findings from the Test Us at Home Prospective Cohort Study"

Gannon C Mak

doi:doi:10.1162/2e3983f5.27d71f76

Performance of Screening for SARS-CoV-2 using Rapid Antigen Tests to Detect Incidence of Symptomatic and Asymptomatic SARS-CoV-2 Infection: findings from the Test Us at Home prospective cohort study

by Apurv Soni, Carly Herbert, Honghuang Lin, Caitlin Pretz, Pamela Stamegna, Taylor Orwig, Colton Wright, Seanan Tarrant, Stephanie Behar, Thejas Suvarna, Summer Schrader, Emma Harman, Chris Nowak, Vik Kheterpal, Lokinendi V Rao, Lisa Cashman, Elizabeth Orvek, Didem Ayturk, Peter Lazar, Ziyue Wang, Bruce Barton, Chad J. Achenbach, Robert L. Murphy, Matthew Robinson, Yuka Manabe, Biqi Wang, Shishir Pandey, Andres Colubri, Laurel O’Connor, Stephenie C. Lemon, Nisha Fahey, Katherine L Luzuriaga, Nathaniel Hafer, William Heetderks, John Broach, and David D McManus

Show Description
dx.doi.org

Description

AbstractBackgroundPerformance of Rapid Antigen Tests for SARS-CoV-2 (Ag-RDT) varies over the course of an infection, and their performance is not well established among asymptomatic individuals.ObjectiveEvaluate performance of Ag-RDT for detection of SARS-CoV-2 in relation to onset of infection for symptomatic and asymptomatic participants.Design, Setting, and ParticipantsProspective cohort study conducted from October 2021 to February 2022 among participants > 2 years-old from across the US who enrolled using a smartphone app. During each testing encounter, participants self-collected one nasal swab and performed Ag-RDT at home; at-least fifteen minutes later, a second nasal swab was self-collected and shipped for SARS-CoV-2 RT-PCR at a central lab. Both nasal swabs were collected 7 times at 48-hour intervals (over approximately 14 days) followed by an extra nasal swab collection with home Ag-RDT test 48-hours after their last PCR sample. Each participant was assigned to one of the three emergency use authorized (EUA) Ag-RDT tests used in this study. This analysis was limited to participants who were asymptomatic and tested negative by antigen and molecular test on their first day of study participation.ExposureSARS-CoV-2 positivity was determined by testing a single home-collected anterior nasal sample with three FDA EUA molecular tests, where 2 out 3 positive test results were needed to determine a SARS-CoV-2 positive result. Onset of infection was defined as day on which the molecular PCR comparator result was positive for the first time.Main Outcomes and MeasuresSensitivity of Ag-RDT was measured based on testing once (same-day), twice (at 48-hours) and thrice (at 96 hours). Analysis was repeated for different Days Post Index PCR Positivity (DPIPP) and stratified based on symptom-status on a given DPIPP.ResultsA total of 7,361 participants enrolled in the study and 5,609 were eligible for this analysis. Among 154 eligible participants who tested positive for SARS-CoV-2 infection based on RT-PCR, 97 were asymptomatic and 57 had symptoms at onset of infection (DPIPP 0). Serial testing with Ag-RDT twice over 48-hours resulted in an aggregated sensitivity of 93.4% (95% CI: 89.1-96.1%) among symptomatic participants on DPIPP 0-6. Among the 97 people who were asymptomatic at the onset of infection, 19 were singleton RT-PCR positive, i.e., their positive test was preceded and followed by a negative RT-PCR test within 48-hours. Excluding these singleton positives, aggregated sensitivity on DPIPP 0-6 for two-time serial-testing among asymptomatic participants was lower 62.7% (54.7-70.0%) but improved to 79.0% (71.0-85.3%) with serial testing three times at 48-hour interval.DiscussionPerformance of Ag-RDT within first week of infection was optimized when asymptomatic participants tested three-times at 48-hour intervals and when symptomatic participants tested two-times separated by 48-hours.Key pointsQuestionWhat is the performance of serial rapid antigen testing (Ag-RDT) in the first week of infection among symptomatic and asymptomatic SARS-CoV-2 infections?FindingsSerial testing with Ag-RDT two-times separated by 48-hours resulted in detection of more than 90% of SARS-CoV-2 infections when symptomatic participants began testing within first week from onset of molecular positivity; participants who were asymptomatic when they began testing within the first-week of molecular positivity observed a sensitivity of 79.0% when they performed three rapid antigen-tests, 48 hours apart.MeaningTo optimize detection of SARS-CoV-2 infection with home antigen tests, people suspected to be infected with SARS-CoV-2 virus should test twice at least 48-hours apart if they are symptomatic and three times at 48-hour intervals if they do not have symptoms (asymptomatic).Key definitionsComparator positivecomposite definition of molecular positivity if majority of molecular assays were positiveDays Past Index Comparator Positive (DPIPP)Number of calendar-days past the day when first Comparator positive was observedOnset of InfectionDPIPP 0, when first Comparator positive was observedSymptomatic and Asymptomatic CasesBased on presence or absence of self-reported symptoms on the day of testing. Sensitivity was measured for Symptomatic and Asymptomatic cases on DPIPP 0-10First week of InfectionDPIPP 0 - 6

RR:C19 Evidence Scale rating by reviewer:

Reliable. The main study claims are generally justified by its methods and data. The results and conclusions are likely to be similar to the hypothetical ideal study. There are some minor caveats or limitations, but they would/do not change the major claims of the study. The study provides sufficient strength of evidence on its own that its main claims should be considered actionable, with some room for future revision.

***************************************

Review:

The manuscript written by Soni et al. described a study of assessing the performance of antigen detection rapid diagnostic tests (Ag-RDTs) to detect COVID-19 individuals. This study focused on the usage of Ag-RDTs to detect asymptomatic COVID-19 individuals. As the disease progressed, individuals may become symptomatic. Serial testing of Ag-RDTs demonstrated in the study could be useful for detecting these two groups of patients particularly for the asymptomatic group. Overall, the manuscript was clear and well written. The works were scientifically sound. The research gaps were clear, the study was well designed and performed. The results obtained correlated well with the objectives of the study. However, I do have several comments for the current version so that the quality of the manuscript could be improved further. The authors are recommended to define the time points for a participant to involve over the 15- day testing period. Instead of referring the study design to reference 6, it will be clear if the ‘Study Day’ were briefly mentioned in the method section. For example, ‘Study Day 0’ was defined as the start of the study, participants included in the analysis were met for the following three conditions: (1) asymptomatic, (2) tested negative by Ag-RDT, (3) testing negative by RT-PCR. Both Ag-RDT and RT-PCR tests were performed between Study Day 1 and Study Day 13 on 48 hours interval whereas Ag-RDT was only performed on Study Day 15.

Among the 154 COVID-19 individuals enrolled, it is expected that the number of individuals for DPIPP 0 varied between Study Day 1 and Study Day 13. It means that some individuals were tested RT-PCR positive early on the study, in the middle of the study, or late on the study. The authors should list out the number of participants for DPIPP 0 between Study Day 1 and Study Day 13. In addition, if the individuals were tested RT-PCR positive early or in the middle of study, the authors should list out the number of individuals if they were consecutive tested RT-PCR positive for the subsequent time points. The authors mentioned that there were 20 individuals who were tested RT-PCR positive for only one time point. There are three possible scenarios: (1) individuals positive early on the study, it is possible that the individuals did not realize they were infected but they were in the last phase of the disease, the vial load for these cases were low which could only be detected by RT-PCR but not by Ag-RDTs, (2) individuals test positive later on the study, it is possible that the individuals were infected during the study and were in the early phase of the disease, the vial load for these cases were low which could only be detected by RT-PCR but not by Ag-RDTs, (3) individuals positive in the middle of study, unknown reasons, maybe due to the sampling error, contamination, etc, the RT-PCR results were not reproducible for the subsequent time points. It should be the rationale for the authors to perform the analysis after excluding this group of individuals. The authors should discuss/justify this separate analysis rather than just saying further investigations should be performed to understand the clinical significance. Otherwise, readers will not realize the aims of this separate analysis.

The results generated in the study could help to provide the guidance of using Ag-RDTs during the first week of infection as defined by RT-PCR. The research question set by the authors was thoroughly addressed. However, there was a lack of in-depth results analysis and comprehensive discussion of this issue. The authors should guide the readers to realize how the aims of the study were achieved. Figure 2 and Supp Table 2 showed that there were advantages to perform serial Ag-RDTs in both symptomatic and asymptomatic individuals on or before DPIPP6. However, it was not worth doing serial Ag-RDTs as the % dropped after DPIPP6 which might be due to the viral load dropped in the late phase of the disease. When comparing with symptomatic and asymptomatic individuals to perform serial Ag-RDTs on or before DPIPP6, asymptomatic individuals were particularly benefited. When analyzing the DPIPP 0-6, it was more cost effective to perform serial Ag-RDTs as the % increased double from 34.4% for doing one Ag-RDT to 68.5% for doing three Ag-RDTs. The effect was less pronounced for the symptomatic individuals, the % was slightly increased from 82.5% to 94.5% for doing one and three Ag-RDTs respectively.

The strength of the study was the clear separation of symptomatic group and asymptomatic group of individuals. Although the viral load (e.g. Ct values of the RT-PCR) data were not analyzed, sensitive of Ag-RDTs correlated with viral load of the samples. Solving this issue should not be a concern since this topic has been thoroughly studied and published by different research groups. The authors could cite some articles to support this claim. Then, the findings of the study could indirectly demonstrate the difference of viral load dynamics between these two groups individuals. The authors are recommended to briefly touch about this issue in the discussion section.

Finally, the limitations of the study should be discussed/addressed. For example, the authors should discuss the sample sizes of COVID-19 individuals enrolled, lineages of the viruses detected (i.e. Delta and Omicron) and prevalence of the disease during the study period, the concerns of self-performed Ag-RDTs, self-collected specimens for RT-PCR, self-reported symptoms, etc. The readers should be well-alerted to those limitations so that they may apply the findings in their own settings. I hope my comments may be useful for the authors if they are planning to publish the manuscript or editors for providing them publishing options. Indirectly, wide variety of disciplines such as media, public, policy makers, academic/scientific communities are benefited. Subsequently, different research groups or policy makers may take reference to the findings for performing further studies or fine tuning the COVID-19 control policies.