Antibody Responses to SARS-CoV-2 Infection—Comparative Determination of Seroprevalence in Two High-Throughput Assays versus a Sensitive Spike Protein ELISA

Abstract

Robust assay development for SARS-CoV-2 serological testing requires assessment of asymptomatic and non-hospitalised individuals to determine if assays are sensitive to mild antibody responses. Our study evaluated the performance characteristics of two high-throughput SARS-CoV-2 IgG nucleocapsid assays (Abbott Architect and Roche) and The Binding Site (TBS) Anti-Spike IgG/A/M ELISA kit in samples from healthcare workers (HCWs). The 252 samples were collected from multi-site NHS trusts and analysed for SARS-CoV-2 serology. Assay performance was evaluated between these three platforms and ROC curves were used to redefine the Abbott threshold. Concordance between Abbott and TBS was 66%. Any discrepant results were analysed using Roche, which showed 100% concordance with TBS. Analysis conducted in HCWs within 58 days post-PCR result demonstrated 100% sensitivity for both Abbott and Roche. Longitudinal analysis for >100 days post-PCR led to sensitivity of 77.2% and 100% for Abbott and Roche, respectively. A redefined Abbott threshold (0.64) increased sensitivity to 90%, producing results comparable to TBS and Roche. The manufacturer’s threshold set by Abbott contributes to lower sensitivity and elevated false-negative occurrences. Abbott performance improved upon re-optimisation of the cut-off threshold. Our findings provided evidence that TBS can be used as bespoke alternative for SARS-CoV-2 serology analysis where high-throughput platforms are not feasible on site.