Assessment of serological assays for identifying high titer convalescent plasma

ABSTRACT

The COVID-19 pandemic has been accompanied by the largest mobilization of therapeutic convalescent plasma (CCP) in over a century. Initial identification of high titer units was based on dose-response data using the Ortho VITROS IgG assay. The proliferation of SARS-CoV-2 serological assays and non-uniform application has led to uncertainty about their interrelationships. The purpose of this study was to establish correlations and analogous cutoffs between commercially available serological tests (Ortho, Abbott, Roche), a spike ELISA, and a virus neutralization assay using convalescent plasma from a cohort of 79 donors from April 2020. Relationships relative to FDA-approved cutoffs under the CCP EUA were identified by linear regression and receiver operator characteristic curves. Relative to the Ortho VITROS assay, the r2 of the Abbott, Roche, the anti-Spike ELISA and the neutralizing assay were 0.58, 0.5, 0.82, and 0.44, respectively. The best correlative index for establishing high-titer units was 3.82 S/C for the Abbott, 10.89 COI for the Roche, 11,202 for the anti-Spike ELISA, and 1200 by the neutralization assay. The overall agreement using derived cutoffs compared to the CCP EUA Ortho VITROS cutoff of 9.5 was 92.4% for Abbott, 84.8% for Roche, 87.3% for the anti-S ELISA and 78.5% for the neutralization assay. Assays based on antibodies against the nucleoprotein (Roche, Abbott) and neutralizing antibody tests were positively associated with the Ortho assay, although their ability to distinguish FDA high-titer specimens was imperfect. The resulting relationships help reconcile results from the large body of serological data generated during the COVID-19 pandemic.