Head-to-Head Comparison of Two SARS-CoV-2 Serology Assays

Link to article at PubMed

J Appl Lab Med. 2020 Jul 27:jfaa125. doi: 10.1093/jalm/jfaa125. Online ahead of print.


BACKGROUND: While molecular techniques remain the gold standard for diagnosis of acute SARS-CoV-2 infection, serological tests have the unique potential to ascertain how much of the population has been exposed to the COVID-19 pathogen. There have been limited published studies to date documenting the performance of SARS-CoV-2 antibody assays.

METHODS: We compared the DiaSorin Liaison SARS-CoV-2 S1/S2 IgG and Roche Diagnostics Elecsys Anti-SARS-CoV-2 assays using 228 samples spanning patients with positive PCR for SARS-CoV-2, patients with compatible symptoms but negative PCR, pre-COVID specimens, and potential cross-reactives.

RESULTS: Both assays detected antibodies in 18/19 samples collected at least one week after a positive PCR result. Neither method consistently detected antibodies in specimens collected within one week of a positive PCR result (sensitivity < 50%), but antibodies were detected by only Roche in four samples in this time frame. Using 139 pre-COVID and 35 PCR-negative samples, the Roche and DiaSorin assays demonstrated specificities of 100.0% and 98.9%, respectively. Neither assay demonstrated cross-reactivity from other coronaviruses (229E, HKU1, NL63, OC43), respiratory pathogens (adenovirus, metapneumovirus, rhinovirus/enterovirus), or antibodies to other viruses (HIV, EBV, CMV, HBV, HCV, HAV).

DISCUSSION: Overall, the qualitative interpretations afforded by the Roche and DiaSorin assays agreed for 99% of samples evaluated. Minor discrepancies in sensitivity and specificity were observed between methods, with the differences in specificity more clinically significant for our low-prevalence population. For the DiaSorin assay, all disagreements with the Roche assay occurred in samples with quantitative signals near the cut-off determining positivity.

PMID:32717056 | DOI:10.1093/jalm/jfaa125

Leave a Reply

Your email address will not be published.