Serological Testing for COVID-19 (SARS-CoV-2): Initial Assessment of Various Approaches in a Large Mixed-Method Diagnostic Accuracy Study

Background: Significant hope and expectation has been placed on serological immunoassays to test for previous COVID-19 infection, to determine immunity against SARS-CoV2 and to assess the response to vaccination. To date, however the diagnostic utility of such immunoassays remains unclear. In a mixed-design evaluation study, we compared the performance of four serological immunoassays targeting various proteins of the SARS-CoV-2 virus: (a) the receptor binding domain (RBD; ELISA, IgG, IgM), (b) the nucleocapsid (N; ELISA, IgG, IgM), (c) the S1 domain of the spike protein (S1; ELISA, IgG) and (d) the spike protein (S; lateral flow immunoassay, IgG, IgM). Methods: Patients admitted between March and April 2020 to Inselspital University Hospital with confirmed COVID-19 infection were prospectively followed alongside medical staff in a cross-sectional design. Biobank samples from the previous winter period 2018/2019 were additionally tested. An in-house enzyme-linked immunosorbent assay utilising recombinant RBD from the SARS-CoV-2 spike protein was developed and compared to three commercially available serological assays. The primary reference standard was defined as a positive RT-PCR (nasopharyngeal swab). Findings: One-hundred and fifty-nine individuals were included, comprising 25 patients (with 187 data points), 102 medical staff, and 32 biobank samples. Prevalence was 30.8% (n=49). Time to IgG seroconversion occurred between day 2 and day 21 after the start of symptoms and between day 0 and 21 after a positive RT-PCR. Regarding IgG, the area under the receiver operating characteristic curve was 0.97 for RBD (95% confidence interval 0.94 to 1.00), 0.97 for S1 (0.95, 1.00), and 0.94 for N ELISA (0.88, 0.99). Corresponding sensitivity was 79.6% (RBD; 66.4, 88.5), 87.8% (S1; 75.8, 94.3), 67.4% (95% CI 52.5, 80.1), and 71.4% (S LFI; 58.7, 82.1). Specificity was 98.2% (RBD; 93.6, 99.7), 99.1 (S1; 95.0, 100.0), 95.5% (N ELISA 89.7, 98.5), and 95.8% (S LFI). Interpretation: Diagnostic performance of serological immunoassays targeting RBD, S1, N and S appears be adequate for identification of patients with previous COVID-19 infection. This observation should be confirmed by full diagnostic accuracy studies to also derive the precise time-point of seroconversion and presence and duration of immunity. Funding Statement: No specific funding was obtained for the purpose of this study.Declaration of Interests: All authors declare that there is no conflict of interests.Ethics Approval Statement: The study protocol was approved by the appropriate ethics committee and the authorities of the University Hospital and conducted in accordance with the Declaration of Helsinki.