Use of serology immunoassays for predicting SARS-CoV-2 infection: a serology-based diagnostic algorithm

Abstract

Background. Detection of viral RNA by nucleic acid amplification testing (NAAT) remains the gold standard for diagnosis of SARS-CoV-2 infection but is limited by high cost and other factors. Whether serology-based assays can be effectively incorporated into a diagnostic algorithm remains to be determined. Herein we describe the development of a serology-based testing algorithm for SARS-CoV-2 infection. Patients and Methods. Between July 2020 and February 2021, we included symptomatic unvaccinated patients evaluated in the Emergency Department of our institution for suspected SARS-CoV-2. All patients had testing by real-time Reverse Transcription Polymerase Chain Reaction. The performance characteristics of five commercial enzymatic serology assays testing for different antibody isotypes were evaluated in a derivation cohort and the assay with the best performance was further tested on a validation cohort. Optimal cut-off points were determined using receiver operating characteristic (ROC) curves and further tested using logistic regression. Results. The derivation and validations cohorts included 72 and 319 patients, respectively. Based on its initial performance, the Elecsys Anti-SARS-CoV-2 assay (Roche Diagnostics) was further tested in the validation cohort. Using ROC curve analysis, we estimated the diagnostic performance for different cut-off points assuming a prevalence of positive tests of 5%. At any given cut-off point the NPV was over 97%. Discussion. This study suggests that an initial diagnostic strategy using the Elecsys Anti-SARS-CoV-2 serology test in symptomatic unvaccinated patients could help to rule out an acute SARS-CoV2 infection and potentially lead to appropriately tailored infection control measures or rational guidance for further testing with a potential cost reduction and increased availability