Point-of-Care Instrument-Read Rapid Antigen SARS-CoV-2 and Influenza A/B Test

For SARS-CoV-2, the device was compared with RT-PCR on about 300 nasal swabs, about 60 of them RT-PCR positive. PPA was about 86% and NPA about 99%, with a handful of false negatives and very few false positives. Among the concordant SARS-CoV-2 results, the device reported a few influenza false positives, including one alongside a SARS-CoV-2 true positive.

Influenza performance was supported by an agreement study rather than a new prospective influenza trial. On about 100 frozen remnant nasopharyngeal swab specimens (roughly equal numbers of influenza A positives, influenza B positives and negatives) tested randomized and blinded on the combination assay and the previously cleared influenza A+B assay, positive and negative agreement were 100% for both influenza A and influenza B. The labeling states these results were also 100% concordant with historical reference RT-PCR results.

FDA issued the Emergency Use Authorization for use in CLIA-certified laboratories meeting requirements for moderate, high or waived complexity tests, including at the point of care in settings operating under a CLIA Certificate of Waiver, Compliance or Accreditation. FDA later granted a labeling revision responding to the Repeat Testing Revision Letter. The authorized use is now testing at least twice over three days with at least 48 hours between tests.

The SARS-CoV-2 PPA of about 86% meets the 80% minimum sensitivity in FDA's antigen template, and the roughly 60 RT-PCR positives are about twice the template's floor of 30. That is why the confidence interval is narrower than in many early antigen EUAs. The template text we reviewed does not specify a confidence-interval lower bound for antigen tests. NPA of about 99% rests on about 200 RT-PCR negatives, with very few false positives.

Two features shape how far the numbers generalize. First, all clinical testing used frozen specimens, and the labeling warns that performance may differ with fresh samples. Second, the influenza agreement data compare the new assay with its cleared predecessor on banked remnant specimens. They show the modification did not change influenza calls, but they are not a fresh prospective estimate of influenza sensitivity against a molecular reference.

The device is intended for the simultaneous qualitative detection and differentiation of SARS-CoV-2, influenza A and influenza B antigens directly from anterior nasal swabs. The intended patients are individuals a healthcare provider suspects of a viral respiratory infection consistent with COVID-19, within the first six days of symptom onset. The authorization covered CLIA-certified laboratories meeting the requirements for moderate, high or waived complexity tests, including point-of-care settings operating under a CLIA Certificate of Waiver, Compliance or Accreditation. FDA's authorization letter explains the clinical need: COVID-19 and influenza can look alike, and a test that differentiates them was needed during a flu season that coincided with the pandemic.

The device added SARS-CoV-2 reagents to a previously cleared influenza A/B antigen assay, with minor modifications. The SARS-CoV-2 claim therefore needed its own prospective clinical evidence against an authorized molecular comparator, in symptomatic patients within the six-day window and from the claimed anterior nasal specimen. Separately, the influenza results had to be shown to remain concordant with the cleared assay.

The study had to produce enough RT-PCR-positive symptomatic subjects to estimate agreement, and it had to do so in the early months of the pandemic. Testing capacity was strained at that time, and the clinical sample would have to be collected and stored before testing.

SARS-CoV-2 antigen tests reached the U.S. market mainly through Emergency Use Authorizations under section 564 of the Federal Food, Drug, and Cosmetic Act. These followed the HHS Secretary's determination of a public health emergency and the declaration that circumstances justified emergency use of COVID-19 diagnostics. An EUA is not a clearance: FDA authorizes use when it is reasonable to believe the product may be effective and its known and potential benefits outweigh its risks. The authorization comes with Conditions of Authorization that can require post-authorization studies and labeling updates.

FDA set out its clinical expectations in a series of downloadable EUA templates, revised several times during 2020 and 2021. The antigen template we reviewed recommends prospective, blinded, randomized clinical agreement studies with at least 30 positive and 30 negative natural clinical samples from the intended-use population, with days since symptom onset recorded for every subject. The comparator should be one of the more sensitive authorized RT-PCR assays using chemical lysis and solid-phase extraction, and a discordant-analysis plan should be fixed in advance. Candidate tests should show a minimum sensitivity of 80% or more for each sample type. For a point-of-care claim, the whole clinical evaluation should run at CLIA-waived sites with four to six non-laboratory-trained operators using only the quick reference instructions, supplemented by blinded near-LoD contrived samples and flex studies. For multi-analyte tests that add SARS-CoV-2 to a previously FDA-cleared influenza assay, the template asks for confirmation that the cleared analytes' LoD is unchanged, testing of at least 10 positive and 10 negative clinical samples for each previously cleared analyte, and a competitive inhibition study. All SARS-CoV-2 studies are still expected as for a standalone test.

Serial testing changed what an antigen authorization means. FDA later revised the authorized uses of all authorized SARS-CoV-2 antigen tests to require repeat testing. Tests authorized for symptomatic people are now for use at least twice over three days with at least 48 hours between tests. Tests authorized for asymptomatic people are now for use at least three times over five days. FDA based this on data showing that repeat testing after a negative antigen result increases the chance of an accurate result. Antigen tests are less sensitive than laboratory RT-PCR, so single-test sensitivity near the 80% expectation and strong dependence on days since symptom onset are what make the category hard. On the influenza side, cleared rapid influenza antigen tests are regulated under 21 CFR 866.3328, whose special controls set minimum clinical performance criteria and require annual strain-reactivity testing.

• Collecting both swabs at once by the dual nares method gave the device and comparator arms the same anatomical source and timing, which FDA's template stresses to reduce discordance from biological variability.
• Freezing within 30 minutes and testing in batches by site allowed blinded, sequential testing of a defined date-range cohort. The labeling discloses the tradeoff: clinical performance was evaluated with frozen samples and may differ with fresh samples.
• Restricting enrollment to symptomatic patients within six days of onset matched the requested indication, the population in which antigen sensitivity is highest.
• Several geographically diverse collection areas reduced the risk that one local outbreak or patient mix would drive the estimate. This matters for a combined respiratory test whose SARS-CoV-2 and influenza prevalence vary by region and week.