ABSTRACT
Background Serological testing for SARS-CoV-2 is integral for understanding prevalence of disease, tracking of infections, confirming humoral response to vaccines, and determining timing and efficacy of boosters. The study objective was to compare the specificity of serology assays in emergency department populations across the United States in 2019 (pre-pandemic) and early 2020, incorporating an automated confirmatory assay. Methods Patient specimens (n = 1954) were from 4 regions in the United States: New York, NY;Milwaukee, WI;Miami, FL;and Los Angeles, CA. Specimens were tested with SARS-CoV-2 anti-spike receptor-binding domain assays: SARS-CoV-2 IgG on the Abbott Alinity i (AdviseDx SARS-Cov-2 IgG II) and Beckman Coulter Access 2 (SARS-CoV-2 IgG II), and SARS-CoV-2 IgM on the Abbott Alinity i (AdviseDx SARS-CoV-2 IgM). Reactive samples were tested with a research use only angiotensin-converting enzyme 2 binding inhibition assay (Abbott ARCHITECT) for confirmation of SARS-CoV-2 neutralizing antibodies. Assay specificity was determined and comparisons performed with Fisher’s exact test. Results Overall SARS-CoV-2 IgG specificity was 99.28% (95% confidence interval, 98.80%–99.61%), 99.39% (98.93%–99.68%), and 99.44% (98.99%–99.72%) for SARS-CoV-2 IgG by Abbott and Beckman, and SARS-CoV-2 IgM, respectively. Overall agreement for the two IgG assays was 99.28% (range for the 4 sites: 98.21% to 100%). There were no specificity differences between assays or sites. Conclusions The specificity of the serological assays evaluated in a large, diverse emergency department population was >99% and did not vary by geographical site. A confirmatory algorithm with an automated pseudo-neutralization assay allowed testing on the same specimen while reducing the false positivity rate and increasing the value of serology screening methods.