Specificity testing by point prevalence as a simple assessment strategy using the Roche Elecsys® Anti-SARS-CoV-2 immunoassay.

Abstract Background The detection of antibodies against the novel coronavirus SARS-CoV-2 is mandatory for the diagnosis, retrospective assessment of disease progression, and correct evaluation of the current infection situation in the population. Many of these assays have been launched by various manufacturers. Unfortunately, the new FDA emergency use regulations have resulted in a situation where laboratories have to perform their own validation studies. Numerous of these labs do not have the biobank needed to carry the studies out. Methods In this paper, we introduce a method that allows institutions to quickly perform a verifiation study in a low-prevalence infection situation. As proof of concept, we used the Roche Elecsys Anti-SARS-CoV-2 ECLIA assay and an SAP-based hospital information system. The Shenzhen YHLO Biotech IgM and IgG assay targetting other surface patterns, was used as confirmatory test. Results The Roche assay demonstrated a limit of detection of 0.069 COI (cutoff index) and successfully passed the performance validation according to CLSI EP15-A3. The study population of 627 in-patients exhibited an age median of 64 years and approximately 13% of the group were under intensive care at the respective time point. All included patients were tested negative for SARS-CoV-2 infection by qRT-PCR (cobas 6800®, Roche, Mannheim, Germany). Only one false positive result was obtained, resulting in a specificity for the Roche Elecsys Anti-SARS-CoV-2 test of 99.84% and a negative predictive value of 99.98%. Conclusions The anonymized use of residual material enables quick evaluation of anti-SARS-CoV-2 immunoassays, as shown in this work with the Roche Elecsys assay. Comparison of the control population with economic data makes it possible to validate the sampling set and therefore to determine diagnostic specificity. Using the chosen approach, it was shown that the Roche test achieved very good results in terms of diagnostic specificity, reproducibility and limit of detection.