Post-Disease Divergence in SARS-CoV-2 RNA Detection between Nasopharyngeal, Anterior Nares and Saliva/Oral Fluid Specimens - Significant Implications for Policy & Public Health

Abstract Background Patients have been shown to shed SARS-CoV-2 viral RNA in nasopharyngeal (NP) specimens for over 100 days after resolution of clinical disease (1, 2). How this relates to anterior nares and oral fluid specimens has not previously been investigated. Methods We prospectively collected oral fluid, anterior nares, NP swab and serum specimens from 1,326 individuals at 2 “drive-through” testing locations. The Curative SARS-CoV-2 Assay (Curative Assay)(3) on oral fluid and anterior nares specimens was compared to the EURORealTime SARS-CoV-2 Assay (EuroRT Assay)(4) on anterior nares and NP specimens. Viral culture and IgG serology were used to assess infectious potential and stage of disease. Additionally we investigated differences in viral RNA detection between specimen types, both early (< 21 days) and late (> 21 days) in SARS-CoV-2 infection, by using an employee surveillance program with daily SARS-CoV-2 testing to precisely determine infection date, even without symptoms. We prospectively collected oral fluid, anterior nares and NP swab specimens from 165 subjects with early infections and 22 subjects with late infections. Specimens were tested using the Curative Assay with the “high-sensitivity” Hologic Aptima SARS-CoV-2 Assay (Hologic Assay)(5) on an NP swab used as the comparator. Late infection specimens were also tested with EuroRT and Zymo Quick SARS-CoV-2 rRT-PCR Kit (Zymo) (6) Assays. Results The “drive-through” study showed similar sensitivities of oral fluid and anterior nares specimens on the Curative Assay to anterior nares specimens tested with the EuroRT Assay. However NP specimens tested with the same EuroRT assay produced 20-30% more positives. Incorporating viral culture and serology data to exclude NP RT-PCR positives that are not infectious or late in the course of disease showed a Positive Percent Agreement (PPA) for of 98.2% and 96.2% and Negative Percent Agreement (NPA) of 97.6% and 98.1% for anterior nares and oral fluid specimens, respectively. Within 21 days of infection, the Curative Assay showed a PPA and NPA of 100% and 100%, respectively for oral fluid; of 100% and 99% respectively for anterior nares; and of 98.2% and 99.0%, respectively in nasopharyngeal specimens compared to an NP specimen on the Hologic Assay. 29 positives were asymptomatic and showed 100% PPA and 100% NPA for all specimen types. After 21 days from infection onset, significant divergence between NP and other specimen types occurred on all 4 assays. Out of 22 paired sample sets, 18, 13, 8 and 4 NP specimens were positive on the Curative, Zymo, Hologic and EuroRT assays, respectively, compared to only 3, 2, 0 and 1 positive anterior nares specimens. Only one oral fluid sample was positive in both the Curative and Zymo assays. Conclusions We used a unique population to show significant divergence between NP specimens and anterior nares or oral fluid specimens >21 days from SARS-CoV-2 infection, which appears to be biological variation and is independent of assay used. This has significant public health implications for the use of NP specimens in community testing programs and policy implications for evaluation of novel specimen types and tests where the use of NP swabs as a comparator may say more about the study population than the assay or specimen type to be evaluated and may unnecessarily limit access to testing.