Performance of Rapid Antigen Tests to Detect Symptomatic and Asymptomatic SARS-CoV-2 Infection : A Prospective Cohort Study

Abstract

Background: The performance of rapid antigen tests (Ag-RDTs) for screening asymptomatic and symptomatic persons for SARS-CoV-2 is not well established.

Objective: To evaluate the performance of Ag-RDTs for detection of SARS-CoV-2 among symptomatic and asymptomatic participants.

Design: This prospective cohort study enrolled participants between October 2021 and January 2022. Participants completed Ag-RDTs and reverse transcriptase polymerase chain reaction (RT-PCR) testing for SARS-CoV-2 every 48 hours for 15 days.

Setting: Participants were enrolled digitally throughout the mainland United States. They self-collected anterior nasal swabs for Ag-RDTs and RT-PCR testing. Nasal swabs for RT-PCR were shipped to a central laboratory, whereas Ag-RDTs were done at home.

Participants: Of 7361 participants in the study, 5353 who were asymptomatic and negative for SARS-CoV-2 on study day 1 were eligible. In total, 154 participants had at least 1 positive RT-PCR result.

Measurements: The sensitivity of Ag-RDTs was measured on the basis of testing once (same-day), twice (after 48 hours), and thrice (after a total of 96 hours). The analysis was repeated for different days past index PCR positivity (DPIPPs) to approximate real-world scenarios where testing initiation may not always coincide with DPIPP 0. Results were stratified by symptom status.

Results: Among 154 participants who tested positive for SARS-CoV-2, 97 were asymptomatic and 57 had symptoms at infection onset. Serial testing with Ag-RDTs twice 48 hours apart resulted in an aggregated sensitivity of 93.4% (95% CI, 90.4% to 95.9%) among symptomatic participants on DPIPPs 0 to 6. When singleton positive results were excluded, the aggregated sensitivity on DPIPPs 0 to 6 for 2-time serial testing among asymptomatic participants was lower at 62.7% (CI, 57.0% to 70.5%), but it improved to 79.0% (CI, 70.1% to 87.4%) with testing 3 times at 48-hour intervals.

Limitation: Participants tested every 48 hours; therefore, these data cannot support conclusions about serial testing intervals shorter than 48 hours.

Conclusion: The performance of Ag-RDTs was optimized when asymptomatic participants tested 3 times at 48-hour intervals and when symptomatic participants tested 2 times separated by 48 hours.

Primary funding source: National Institutes of Health RADx Tech program.

Visual Abstract.. Performance of Rapid Antigen Tests to Detect Symptomatic and Asymptomatic SARS-CoV-2 Infection

This prospective study evaluated the performance of rapid antigen tests for detection of SARS-CoV-2 among symptomatic and asymptomatic participants. Participants who were asymptomatic and negative for SARS-CoV-2 on study day 1 completed rapid antigen tests and RT-PCR testing for SARS-CoV-2 every 48 hours for 15 days.

Figure 1.. Study flow diagram.

In the Test Us At Home study to calculate performance of Ag-RDTs for detection of SARS-CoV-2, a total of 7361 participants enrolled in the study, and 154 were eligible for the analysis and tested positive for SARS-CoV-2 by RT-PCR during the study period (97 asymptomatic and 57 symptomatic on day of index comparator positive result). Ag-RDT = rapid antigen test; OTC = over-the-counter; RT-PCR = reverse transcriptase polymerase chain reaction. * Participants replaced their assigned Ag-RDTs with commercially obtained Ag-RDTs. † Dates of RT-PCR testing could not be verified based on triangulation of self-reported, shipping, and resulting data.

Figure 2.. Performance of Ag-RDTs for detection of SARS-CoV-2 in relation to first day of molecular positivity.

Sensitivity (rapid antigen positivity / comparator positivity) was calculated for single-day testing, 2-time serial testing at 48-h intervals, and 3-time serial testing at 48-h intervals for symptomatic and asymptomatic persons based on day and patterns of positivity. Calculations for sensitivity were repeated with testing starting on different days past index positivity on an RT-PCR test. Error bars represent 95% CIs. Performance of Ag-RDTs on day of infection onset was higher among symptomatic participants (59.6%) than among asymptomatic participants (9.3%). Serial testing with 2 Ag-RDTs 48 h apart and 3 Ag-RDTs 48 h apart improved the performance of Ag-RDTs within the first week of infection. Excluding participants with singleton RT-PCR–positive results improved the sensitivity of Ag-RDTs among asymptomatic participants. Ag-RDT = rapid antigen test; DPIPP = day past index polymerase chain reaction positivity; RT-PCR = reverse transcriptase polymerase chain reaction.

Figure 3.. Ct values, by DPIPP and symptom status.

Error bars represent 95% CIs. Symptomatic participants who tested positive by reverse transcriptase polymerase chain reaction (RT-PCR) had significantly lower Ct values on average than asymptomatic participants at DPIPPs 0 and 2. On DPIPP 0, >75% of asymptomatic persons had a Ct value ≥30, whereas <33% of symptomatic persons had a Ct value ≥30. Symptomatic participants had a lower proportion of persons with Ct values ≥30 at all DPIPPs. Ct = cycle threshold; DPIPP = day past index polymerase chain reaction positivity.

Figure 4.. Predicted probability of Ag-RDT positivity, by symptom status and serial testing schedule.

A mixed-effects logistic regression model was used to predict the probability of Ag-RDT positivity based on Ct value and symptom status. Error bars represent 95% CIs. The sensitivity was lowest among asymptomatic participants who performed a single test for all Ct values >20. Two-time serial testing among symptomatic persons and 3-time serial testing among both asymptomatic and symptomatic persons demonstrated sensitivity >80% for Ct values <32. Ag-RDT = rapid antigen test; Ct = cycle threshold; RT-PCR = reverse transcriptase polymerase chain reaction.