FLC | CDC finds COVID-19 antigen tests less accurate in real world than in EUA data

CDC finds COVID-19 antigen tests less accurate in real world than in EUA data

January 6, 2021

A recent report from the Centers for Disease Control and Prevention (CDC) and two universities in Wisconsin found that antigen tests in a real-world setting were less accurate than the current gold standard for screening asymptomatic people for COVID-19, and also less accurate than in the trial on which emergency use authorization (EUA) was based. However, the report concluded antigen tests may still be useful in rapidly identifying cases and controlling outbreaks in settings such as colleges and universities, which is particularly important when more accurate tests are not readily available or have a longer turnaround time for results.

The findings were reported in the January 1 issue of the CDC's Morbidity and Mortality Weekly Report.

This investigation evaluated performance of the Sofia SARS Antigen Fluorescent Immunoassay (FIA) compared with real-time reverse transcription–polymerase chain reaction (RT-PCR) for SARS-CoV-2 detection among asymptomatic (871 people) and symptomatic (227 people) individuals at the University of Wisconsin-Oshkosh and the University of Wisconsin-Madison. From September 28 to October 9, a total of 1,098 paired nasal swabs were tested using the Sofia SARS Antigen FIA and real-time RT-PCR. Virus culture was attempted on all antigen-positive or real-time RT-PCR–positive specimens.

The Sofia SARS FIA received an EUA from the Food and Drug Administration on May 8, 2020, for use in symptomatic people within five days of symptom onset. In the new investigation, among people reporting COVID-19–compatible symptoms at specimen collection, the test was less accurate (sensitivity = 80%; specificity = 99%) than reported in the EUA decision (sensitivity = 97%; specificity = 100%). Two of eight specimens from symptomatic people that had false-negative antigen test results were positive by viral culture, indicating that potentially infectious persons might not be detected by antigen testing. To reduce the impact of false-negative antigen test results, confirmatory testing with an FDA-authorized nucleic acid amplification test (NAAT), such as RT-PCR, should be considered following negative antigen test results in symptomatic individuals.

Among asymptomatic participants, antigen test sensitivity was 41.2%, specificity was 98.4%, and positive predictive value (PPV) in this population was 33.3%. This low PPV was observed despite a relatively high prevalence of SARS-CoV-2 in this population (5.2% prevalence overall; 2% among asymptomatic persons), suggesting that PPV could be even lower when using this antigen test among populations with lower expected SARS-CoV-2 prevalence. To account for false-positive results when using antigen tests for asymptomatic screening, confirmatory NAAT testing should be considered following positive antigen test results in asymptomatic persons, particularly when pretest probability of SARS-CoV-2 infection is low. The negative predictive value (NPV) of antigen testing among asymptomatic participants was 99%, and virus was not cultured from asymptomatic participants with antigen-negative results, indicating that asymptomatic persons with negative antigen results are unlikely to be infected with SARS-CoV-2 and would not require confirmatory NAAT.

Despite reduced sensitivity compared with real-time RT-PCR, the use of antigen tests for serial testing in these settings, particularly when RT-PCR tests are not available or have a prolonged turnaround time, might still allow rapid identification of infectious persons and control of outbreaks, the CDC wrote. However, antigen-based testing strategies should account for the lower sensitivity and lower PPV when used for asymptomatic screening by considering confirmatory testing with an FDA-authorized NAAT, such as RT-PCR, after a positive antigen test result in an asymptomatic person. Confirmatory testing should also be considered following a negative antigen test result in a person experiencing COVID-19–compatible symptoms.