Antibody response against SARS-CoV-2 spike protein and nucleoprotein evaluated by four automated immunoassays and three ELISAs

Abstract

Objectives: The aim was to determine the antibody response against SARS-CoV-2 spike protein and nucleoprotein using four automated immunoassays and three ELISAs for the detection of total Ig antibodies (Roche) or IgG (Abbott, Diasorin, Snibe, Euroimmun, Mikrogen) in COVID-19 patients.

Methods: Sensitivity and dynamic trend to seropositivity were evaluated in 233 samples from 114 patients with moderate, severe or critical COVID-19 confirmed with PCR on nasopharyngeal swab. Specificity was evaluated in 113 samples collected before January 2020, including 24 samples from patients with non-SARS coronavirus infection.

Results: Sensitivity for all assays was 100% (95% confidence interval 83.7-100) 3 weeks after onset of symptoms. Specificity varied between 94.7% (88.7-97.8) and 100% (96.1-100). Calculated at the cut-offs that corresponded to a specificity of 95% and 97.5%, Roche had the highest sensitivity (85.0% (79.8-89.0) and 81.1% (76.6-85.7), p < 0.05 except vs. Abbott). Seroconversion occurred on average 2 days earlier for Roche total Ig anti-N and the three IgG anti-N assays (Abbott, Mikrogen, Euroimmun) than for the two IgG anti-S assays (Diasorin, Euroimmun) (≥50% seroconversion day 9-10 vs. day 11-12 and p < 0.05 for percent seropositive patients day 9-10 to 17-18). There was no significant difference in the IgG antibody time to seroconversion between critical and non-critical patients.

Discussion: Seroconversion occurred within 3 weeks after onset of symptoms with all assays and on average 2 days earlier for assays detecting IgG or total Ig anti-N than for IgG anti-S. The specificity of assays detecting anti-N was comparable to anti-S and excellent in a challenging control population.

Keywords: COVID-19; Coronavirus; Diagnosis; ELISA; Immunoassay; Nucleocapsid protein; SARS-CoV-2; Sensitivity and specificity; Seroconversion; Spike glycoprotein.

Fig. 1

Diagnostic performance of the different assays. (A) ROC curve (samples used to calculate sensitivity and specificity, n = 346). (B) Dynamic trend to seropositivity in 222 samples from 106 patients with COVID-19. Of note, the average time to seroconversion lags behind the true time of seroconversion by a couple of days since patients were not tested daily and a patient is only considered to have seroconverted after the first positive result.

Fig. 2

Antibody response to SARS-CoV-2 N-antigen and S-antigen. (A) Dynamic trend to seropositivity for Roche total Ig, the 3 IgG anti-N assays (Abbott, Mikrogen, Euro NCP) and IgG anti-S assays (Diasorin, Euro S1). ^†p <0.05 for IgG anti-S1 assays vs. IgG anti-N assays. ^∗p < 0.05 for anti-S assays vs. IgG anti-N assays and Roche total Ig anti-N. (B) Dynamic trend to seropositivity for IgG anti-N and IgG anti-S assays in critical and non-critical patients.

Fig. 3

Evolution of the antibody levels (P25, median, P75) with the different assays. Results of the individual assays were normalized by dividing the result by the cut-off proposed by the manufacturer (lowest cut-off in case the manufacturer defines an equivocal zone).