Abstract
BACKGROUND: Coronavirus disease 2019 (COVID-19), which is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has become a global pandemic. Our laboratory initially used a two-step molecular assay, first reported by Corman et al, for SARS-CoV-2 identification (the Taiwan Center for Disease Control [T-CDC] method). As rapid and accurate diagnosis of COVID-19 is required to control the spread of this infectious disease, the current study evaluated three commercially available assays, including the TaqPath COVID-19 Combo kit, the cobas SARS-CoV-2 test, and the Rendu 2019-nCoV Assay kit, to establish diagnostic algorithms for clinical laboratories. METHODS: A total of 790 clinical specimens, including nasopharyngeal swabs, throat swabs, sputum, saliva, stool, endotracheal aspirate, and serum were obtained from patients who were suspected or already confirmed to have COVID-19 at the Taipei Veterans General Hospital from February to May 2020. These specimens were tested for SARS-CoV-2 using the different assays and the performance variance between the assays was analyzed. RESULTS: Of the assays we evaluated, the T-CDC method and the TaqPath COVID-19 Combo kit require lots of hands-on practical laboratory work, while the cobas SARS-CoV-2 test and the Rendu 2019-nCoV Assay kit are fully automated detection systems. The T-CDC method and the TaqPath COVID-19 Combo kit showed similar detection sensitivity; however, the T-CDC method frequently delivered false-positive signals for envelope (E) and/or RNA-dependent RNA polymerase (RdRP) gene detection, thus increasing the risk of reporting false-positive results. A manual test-based testing strategy combining the T-CDC method and the TaqPath COVID-19 Combo kit was developed, which demonstrated excellent concordance rates (>99%) with the cobas and Rendu automatic systems. There were a few cases showing discrepant results, which may be due to the varied detection sensitivities as well as targets among the different platforms. Moreover, the concordance rate between the cobas and Rendu assays was 100%. CONCLUSION: Based on our evaluation, two SARS-CoV-2 diagnostic algorithms, one focusing on the manual assays and the other on the automatic platforms, were proposed. Our results provide valuable information that allows clinical laboratories to implement optimal diagnostic strategies for SARS-CoV-2 testing based on their clinical needs, such as test volume, turn-around time, and staff/resource limitations.