The levels of the antibody response against SARS-CoV-2 varies widely between individuals, which together with the decline of antibody responses over time, complicates the correct classification of seropositivity using conventional assay cut-offs. All subjects in a cohort of SARS-CoV-2 PCR+ individuals representing different disease severity categories (n=105), and a group of PCR+ hospital staff (n=33), developed IgG against pre-fusion-stabilized spike (S) trimers and 97% did against the receptor-binding domain (RBD). The levels differed by several orders of magnitude and associated with disease phenotype. Concomitant analysis of a cohort of healthy blood donors and pregnant women (n=1,000), representing individuals who had undergone milder infections, demonstrated highly variable IgG titers, including several that scored between the classical 3SD and 6SD cut-offs. Since the correct classification of seropositivity is critical for epidemiological estimates, we trained probabilistic algorithms to assign likelihood of past infection using anti-S and -RBD IgG data from PCR+ individuals and a large cohort of historical negative controls (n=595). Applied to blood donors and pregnant women, this probabilistic approach provided a more accurate way to interpret antibody titers spread over a large continuum offering a probability-based diagnosis. The methods described here are directly applicable to serological measurements following natural infection and vaccination.

Download data

• Downloaded 1,525 times
• Download rankings, all-time:
  • Site-wide: 10,413
  • In allergy and immunology: 30
• Year to date:
  • Site-wide: 4,877
• Since beginning of last month:
  • Site-wide: 4,877

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide