Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spurred the Coronavirus 2019 (COVID-19) global pandemic, causing clinical laboratories to rapidly implement tests for SARS-CoV-2. This testing rush led to the release of an abundance of SARS-CoV-2 serologic assays with questionable performance during the early months of the pandemic. These assays were distributed without requirement for prior review by the Food and Drug Administration (FDA) Emergency Use Authorization (EUA) pathway. Studies confirmed several assays had poor sensitivity and specificity, thus clinical laboratories had to exercise caution regarding which assays to adopt [1], [2]. Given these issues, the University of Minnesota (UMN) developed an in-house serological SARS-CoV-2 assay in April [3]. In May, the FDA required all manufacturers to submit EUA applications for serologic testing which led to market removal of poorly performing tests and release of FDA-EUA approved assays. Therefore, we verified performance of three FDA-EUA approved commercial serological SARS-CoV-2 assays to provide high-throughput capability (>10,000 tests/day), and compared them to our UMN method. Two assays detect “total” spike protein receptor binding domain (RBD) antibodies (UMN and Siemens ADVIA Centaur SARS-CoV‑2 Total assay), and two detect nucleocapsid protein antibodies (Abbott SARS-CoV-2 IgG and Roche Cobas Anti-SARS-CoV-2). Commercial assays were performed according to manufacturer’s instructions and the UMN assay was performed as previously described [3].
Results from 56 remnant serum specimens (N = 56 unique patients) are summarized in Table 1 . 28 patients had COVID-19 infection confirmed by a polymerase chain reaction (PCR) assay. Negative patient specimens were either collected pre-pandemic (N = 25), had a negative PCR result (N = 1), or were tested for antibodies only (N = 2). Serum was collected an average of 16.7 (±8.3) days after COVID-19 symptom onset (range: 6–35 days) and 9.7 (±6.0) days after a positive PCR test (range: 2–22 days). In 22 of the 28 PCR-positive patients, antibodies were detected with all four assays. Two serum specimens from PCR-positive patients (collected 14 and 15 days post-symptom onset) were negative with all four tests. Additionally, one serum specimen (collected 6 days post-symptom onset) was positive for spike protein RBD antibodies (UMN/Siemens) and negative for nucleocapsid protein antibodies (Abbott/Roche). Another specimen (collected 7 days post-symptom onset) tested positive for nucleocapsid protein antibodies (Abbott/Roche) and negative for spike protein RBD antibodies (UMN/Siemens). In both cases, immature SARS-CoV-2 immune response could be the issue [3], [4]. Two separate specimens tested positive with three of four assays, with one negative on the Siemens assay and the other negative on the Abbott assay.
Table 1.
Sensitivity and specificity summary of SARS-CoV-2 antibody measurement procedures.
| Measurement Procedure (Antigen, Ab* Isotype) |
Positive Antibody Results/ Confirmed Positive PCR |
Negative Antibody Results/ Prepandemic or Negative PCR |
Sensitivity (95% CI) |
Specificity (95% CI) |
PPV* | NPV* |
|---|---|---|---|---|---|---|
| UMN (Spike RBD, Total Ab) | 25/28 | 28/28 | 89.3% (71.8–97.7%) | 100% (87.7–100%) | 100% | 90.3% |
| Siemens (Spike RBD, Total Ab) | 24/28 | 28/28 | 85.7% (67.3–96.0%) | 100% (87.7–100%) | 100% | 87.5% |
| Roche (Nucleocapsid, Total Ab) | 25/28 | 28/28 | 89.3% (71.8–97.7%) | 100% (87.7–100%) | 100% | 90.3% |
| Abbott (Nucleocapsid, IgG) | 24/28 | 28/28 | 85.7% (67.3–96.0%) | 100% (87.7–100%) | 100% | 87.5% |
95% Confidence Intervals = 95% CI, were calculated using the Exact (Clopper-Pearson) formula.
Antibody = Ab, PPV = Positive Predictive Value, NPV = Negative Predictive Value.
In summary, all four SARS-CoV-2 serologic assays provided clinical sensitivities >85%. The specificities and positive predictive values were 100% and negative predictive values were all >87%. Reasons for discordance between assays are in need of further study, but likely derive from various factors including time from symptom onset or PCR positive result to collection, immune response variability, assay-specific cutoffs, or analytical performance. With the caveat of a small sample size (N = 56 patients), this study suggests these four assays provide utility in detecting SARS-CoV-2 antibodies. These tests may provide critical data on COVID-19 prevalence to support pandemic countermeasure decision-making.
References
- 1.J.D. Whitman, J. Hiatt, C.T. Mowery, B.R. Shy, R. Yu, T.N. Yamamoto, et al. Test performance evaluation of SARS-CoV-2 serological assays. [Preprint] medRxiv. 17 May 2020.
- 2.Hansen S.U.K. hits hurdle in COVID-19 serology testing as assays don’t meet performance criteria. BioCentury. 2020;7 [Google Scholar]
- 3.S.N. Thomas, G. Altawallbeh, C. Zaun, K. Pape, J.M. Peters, et al. Initial determination of COVID-19 seroprevalence among outpatients and healthcare workers in Minnesota using a novel SARS-CoV-2 total antibody ELISA. Preprints with The Lancet. 2020. [DOI] [PMC free article] [PubMed]
- 4.M.S. Tang, K.G. Hock, N.M. Logsdon, J.E. Hayes, A.M. Gronowski, N.W. Anderson, C.W. Farnsworth, Clinical performance of two SARS-CoV-2 serologic assays. Clin. Chem. hvaa120, 13 May 2020; Online ahead of print. [DOI] [PMC free article] [PubMed]