Point-of-Care Molecular SARS-CoV-2 Test

In the contrived-sample evaluation, agreement with expected results was 100% at every concentration and for negatives. In the retrospective study of about 50 specimens, PPA was about 95% and NPA 100%. The single discordant specimen, negative by the candidate test and positive by the comparator, was negative on re-test with the comparator and on a second EUA RT-PCR test.

In the prospective study at CLIA-waived sites, PPA was about 90% and NPA was 100%. SARS-CoV-2 was not detected by a second FDA-cleared molecular assay in one of the few false negative specimens.

FDA first issued the Emergency Use Authorization, then amended and reissued it several times. The final reissuance added new clinical performance data and a presumptive-negative claim to the intended use. Negative results are to be treated as presumptive and, if inconsistent with clinical signs and symptoms or necessary for patient management, confirmed with a different authorized or cleared molecular test in a CLIA-certified laboratory that meets the requirements for high or moderate complexity testing. The test was authorized for point-of-care use in patient care settings operating under a CLIA Certificate of Waiver, Certificate of Compliance or Certificate of Accreditation. The EUA was later revoked.

The contrived-sample results meet the policy's acceptance criteria of 95% agreement at 1x-2x LoD and 100% elsewhere. The prospective study is read differently. FDA's molecular EUA template, in its current form, recommends at least 95% positive and negative agreement for a molecular point-of-care test, but says positive agreement of at least 80% may be considered with limitations such as treating negatives as presumptive. The prospective PPA of about 90% falls between those two figures. The authorized labeling matches that framework: the reissuance added a presumptive-negative claim based on newly provided clinical performance data, and negatives are to be confirmed with a laboratory molecular test when clinically indicated. NPA of 100% meets the 95% recommendation.

With only about 40 comparator positives, the PPA interval is wide. Each additional false negative would have moved the point estimate by more than two percentage points, which is why the template's minimum of 30 positives is a floor, not a target. The discordant result (one of the few false negatives was negative on a second FDA-cleared assay) suggests at least one false negative may have been a comparator-only detection. Following the template, it was not used to adjust PPA. The retrospective PPA of about 95% came from diluted laboratory specimens and is not directly comparable to the direct-swab prospective result.

The test detects SARS-CoV-2 nucleic acid visually on a lateral-flow strip after RT-PCR on a small dock. As finally authorized, it was intended for clinician-collected anterior nasal or nasal mid-turbinate swabs, or clinician-instructed anterior nasal swabs self-collected on site, from individuals suspected of COVID-19 by their healthcare provider. Testing was authorized in CLIA-certified laboratories and at the point of care, in patient care settings operating under a CLIA Certificate of Waiver, Certificate of Compliance or Certificate of Accreditation.

The original authorization covered combined throat and nasal swabs, in high- and moderate-complexity laboratories and in patient care settings using the dock. Its intended use changed several times over the following years: to nasal swab alone, then adding mid-turbinate swabs, then on-site self-collection, then 'anterior nasal' wording. The clinical evidence behind the test grew over the same period.

Showing performance in waived point-of-care settings required evidence from where the test would actually be used. That meant CLIA-waived sites, non-laboratory operators and anterior nasal swabs from symptomatic or exposed patients, compared against a laboratory molecular method run on a paired swab in viral transport media.

Emergency Use Authorization for COVID-19 diagnostics rests on section 564 of the Federal Food, Drug, and Cosmetic Act and the HHS determination of a public health emergency (85 FR 7316). Under an EUA, FDA does not decide substantial equivalence. It assesses whether, based on the totality of scientific evidence, it is reasonable to believe the test may be effective and its known and potential benefits outweigh its known and potential risks. FDA's first COVID-19 test policy covered high-complexity laboratories. A revision extended it to commercial manufacturers distributing test kits to laboratories or to healthcare workers for point-of-care testing.

Clinical evidence expectations in those first weeks were minimal because positive patient specimens were scarce. The policy recommended that, in the absence of known positive samples, developers test at least 30 contrived reactive and 30 non-reactive specimens, with twenty spiked at 1x-2x the LoD and the rest spanning the assay range. The acceptance criteria were 95% agreement at 1x-2x LoD and 100% at all other concentrations and for negatives.

As natural specimens became available, FDA's molecular EUA template raised the bar. In its current form, the template states that contrived samples are inadequate to support clinical performance and recommends at least 30 positive and 30 negative natural clinical samples, with a pre-planned discordant analysis whose results are not included in PPA/NPA. For a point-of-care claim, it asks for a clinical evaluation at non-laboratory sites with intended healthcare operators, at least 30 prospectively collected positives and 30 negatives, testing for at least two weeks, near-LoD samples tested by minimally trained operators, and flex studies. A molecular POC test should show positive and negative agreement of at least 95%, although positive agreement of at least 80% 'may be considered with appropriate limitations', for example treating negatives as presumptive.

• The contrived panel of about 30 positives (most near the LoD) and about 30 negatives followed the structure of FDA's policy almost exactly. That was the fastest defensible route to an authorization when positive patient specimens were scarce.
• Diluting retrospective transport-media specimens in the test buffer was needed because viral transport media can inhibit the assay. It allowed archived positives to be used, at the cost of making the retrospective data less representative of direct-swab testing.
• Running the prospective study at CLIA-waived sites with non-laboratory personnel produced evidence in the setting and in the hands named in the point-of-care authorization, instead of extrapolating from laboratory data.
• Pairing an anterior nasal swab in the test's own buffer with a second swab in transport media for the comparator kept the candidate test in its intended workflow while giving the laboratory method a standard specimen.
• Testing all discrepant specimens on a second FDA-cleared molecular assay matches the template's recommendation of a pre-planned discordant analysis reported alongside, not inside, the PPA/NPA calculation.