Point-of-Care Molecular Strep A Test

Against blood agar culture, sensitivity was about 96% and specificity about 97%. Positive and negative predictive values were about 90% and 99%. Culture prevalence was about 20%. In discordant testing, the second molecular assay detected Strep A in about half of the culture false positives and did not detect it in some of the culture false negatives.

Against the FDA-cleared molecular assay, positive percent agreement was about 93% and negative percent agreement about 99%. The second molecular assay did not detect Strep A in more than half of the false negatives, nor in the single false positive.

In the near cut-off study, combined agreement was about 98% for low positives, about 97% for moderate positives and about 97% for negatives. Intended users' blinded control swabs agreed in about 98% of positives and about 98% of negatives. The initial invalid/error rate in the clinical study was about 8%: internal-control failures (about 5%) and dock system errors (about 2%). A small number of results were unresolved after one repeat, and fewer after a third test.

FDA found the device substantially equivalent under the dual-track review, and the test was granted a CLIA waiver for use in CLIA-waived settings such as physician offices and urgent care clinics.

21 CFR 866.2680 publishes no numeric acceptance thresholds, so the results are read against the comparators themselves. FDA's decision summary described agreement of more than 95% as acceptable for the reproducibility and blinded control-swab results, and all three panel members and both control types exceeded that. Against culture, the false positives are where the reference method matters: the second molecular assay detected Strep A in about half of them, consistent with an amplification test finding organisms that culture missed. Against the molecular comparator, the test produced only one false positive, and more than half of its false negatives were not confirmed by the second molecular assay.

Prevalence affects the predictive values. At about 20% culture prevalence, PPV against culture was about 90%, while NPV was about 99%. In a lower-prevalence population, PPV would fall even with unchanged sensitivity and specificity. Site-level estimates show why pooled analysis is used: one small site had only a couple of culture positives and a low sensitivity, while the largest sites, with about 150 tested samples each, gave sensitivities from about 91% to 100%. Small-site estimates are not interpretable alone.

The test was submitted as a combined 510(k) and CLIA waiver package for dual review. It detects Streptococcus pyogenes nucleic acid in throat swabs from patients with signs and symptoms of pharyngitis, for use in clinical laboratories and CLIA-waived sites. The labeling requires that all negative results be confirmed by bacterial culture, and calls for culture follow-up if symptoms persist or during an outbreak of acute rheumatic fever.

The clinical evidence had to show performance on fresh throat swabs tested the same day by typical intended operators (nurses, physician assistants and medical assistants) at physician office laboratories, urgent care centers and outpatient clinics. The results had to be compared both with blood agar culture, the well-accepted reference method the special controls call for, and with an FDA-cleared molecular assay.

For the waiver, the package also had to show reproducible detection at and just above the limit of detection by non-laboratory operators at CLIA-waived sites, and acceptable invalid and error rates for a dock-based system used in busy clinics.

Nucleic acid tests for Streptococcus species are classified under 21 CFR 866.2680, Streptococcus spp. nucleic acid-based assay, as Class II devices with special controls codified in 2017. The special controls are written into the regulation rather than a separate guidance. A premarket notification must include analytical studies (limit of detection, reactivity, inclusivity, precision, reproducibility, interference, cross-reactivity, carry-over and cross-contamination) and a clinical study, performed in a population consistent with the intended use, that compares the device with well-accepted reference methods. Labeling must address the need for culture confirmation of negative specimens as appropriate, explain how results are interpreted and the acceptance criteria, and, where appropriate, the submission must describe an end-user training program.

The regulation sets no numeric sensitivity or specificity thresholds. The hard part of the category is the reference standard. Blood agar culture is the traditional reference for Group A Streptococcus pharyngitis, but an amplification test can detect organisms that culture misses. Some 'false positives' against culture may therefore be true detections, and a molecular comparator alongside culture helps separate assay error from reference-method limits.

For point-of-care use, the test also needed a CLIA waiver. FDA's guidances on CLIA waiver applications and on Dual 510(k) and CLIA Waiver by Application studies recommend testing at a minimum of three sites representative of the intended waived settings, with 1-3 untrained operators per site and at least nine untrained operators across sites in the comparison study, and testing built into the daily workflow. They also recommend reproducibility at a minimum of three of the same sites with the same number of untrained operators at each, using a panel that includes a true negative, a sample close to the LoD and a moderate positive.

• Using both blood agar culture and an FDA-cleared molecular assay met the special controls' call for well-accepted reference methods, and let reviewers see whether disagreements with culture reflected culture's lower sensitivity or real assay error.
• Same-day testing by nurses, physician assistants and medical assistants at 8–10 physician office, urgent care and outpatient sites matches the guidance recommendation for representative waived sites and untrained operators working within daily clinical workflow.
• Refrigerated shipment of the comparator swab within 24 hours protected culture recovery, which degrades with delay and would otherwise inflate apparent false positives for the investigational test.
• Testing discrepants plus a random 10% of concordant specimens on a second molecular assay checked the concordant results too, not only the disagreements.
• The reproducibility panel used a low positive at ~1X LoD and a moderate positive at ~2X LoD, run by non-laboratory operators at waived sites concurrently with the clinical study. This is the panel structure the 2020 dual-submission guidance describes for qualitative tests with an analytical cutoff.