Abstract
Introduction/Objective
SARS-CoV-2 is a pandemic that has required mobilization to meet urgent needs. In this mobilization, emergency use authorizations (EUA) have been issued by the FDA to expedite the deployment of these tests. This has led to a situation whereby sensitivity has not been rigorously studied for any of the assays with EUAs. Estimates can be extrapolated from the limited samples documented by the company in their instructions for use (IFU). Although the nationwide shortage of testing reagents prevent parallel testing of multiple platforms on all specimens, observations of repeat specimens at the Veteran Affairs Medical Center (VAMC) provides the first study in the literature of more complete data for SARS-CoV-2 nucleic acid (RT-PCR) assay on sensitivity on the Abbott (Abbott Park Ill) and Cepheid (Sunnyvale CA) assays.
Methods
A retrospective search was performed for all test results for SARS-CoV-2 by RT-PCR from 3/1/2020 to 4/14/2020 at Corporal Michael J. Crescenz Medical Center, in order to evaluate the sensitivity on Abbott m2000 and Cepheid platforms. Results across multiple reference laboratories and in-house testing platforms were collated in a table with all patients clinically requiring repeat testing recorded.
Results
114/863 patients had repeat testing. The tests were performed initially by outside reference laboratories (25 patients), on the Abbott m2000 (63 patients), and Cepheid Infinity (26 patients). 15/114 (13%) had discordant results on repeat testing. This included 1 test initially done by a reference laboratory. 8 days after the initial result from the reference lab, a positive for the same patient was identified on the Abbott platform. 11 initial Abbott results were discordant on further repeat testing on two platforms - Abbott (6 patients) and Cepheid (5 patients) 1-6 days later. In addition, 3 initial Cepheid were discordant on further repeat testing by the same Cepheid platform (1-16 days later).
Conclusion
While the instructions for use for both platforms suggest 100% sensitivity and specificity (due to the 100% positive and negative percent agreement in limited specimens), the true sensitivity is less than 100%, particularly early in the course of the infection. In our study, the positive percent agreement (surrogate for sensitivity) was 83% for initial Abbott tests, 88% for initial Cepheid tests, and 95% by Reference laboratory platform.
Multiple reference consistency check for LAAS: a novel position domain approach