In response to the current COVID-19 pandemic, multiple companies marketed serological tests. Rigorous, independent and comparative performances of these assays on defined clinical specimens are needed.
In a first preliminary phase, we investigated 16 IgG, IgM, IgA and pan Ig serological ELISA using a panel of 180 sera, comprising 97 sera from patients with a positive RT-PCR, and 83 negative sera sampled before November 1, 2019. In a second phase and to complete the evaluation on the full panel (100 positive and 300 negative), tests that passed pre-defined exclusion criteria of 90% sensitivity and 97% specificity were further evaluated on 220 additional sera chosen to assess possible cross-reactivity with other human viral infections.
Among the 16 tests evaluated in the preliminary phase, two were excluded due to insufficient sensitivity at 15 days post-symptom onset and one was excluded due to poor specificity. Of the 13 tests evaluated using the full panel comprised of a diverse pool of sera including those reactive against known respiratory viruses, no systematic cross-reactivity was observed. However, heterogeneities across tests were found. Consistent with kinetics of antibody expression, maximal sensitivity was found two weeks post-symptom onset.
In this independent evaluation, we compared the performance of 16 SARS-CoV-2 serological tests using well-characterized sera and found 13 tests with more than 90% sensitivity at 15 days post-symptom onset and 97% specificity across a diverse range of negative samples.