Serological reconstruction of COVID-19 epidemics through analysis of antibody kinetics to SARS-CoV-2 proteins

Pelleau, S.; Woudenberg, T.; Rosado, J.; Donnadieu, F.; Garcia, L.; Obadia, T.; Gardais, S.; Elgharbawy, Y.; Velay, A.; Gonzalez, M.; Nizou, J.-Y.; Khelil, N.; Zannis, K.; Cockram, C.; Merkling, S.; Meola, A.; Kerneis, S.; Terrier, B.; de Seze, J.; Planas, D.; Schwartz, O.; Dejardin, F.; Petres, S.; von Platen, C.; Arowas, L.; Perrin de Facci, L.; Duffy, D.; Ni Cheallaigh, C.; Conlon, N.; Townsend, L.; Auerswald, H.; Backovic, M.; Hoen, B.; Fontanet, A.; Mueller, I.; Fafi-Kremer, S.; Bruel, T.; White, M. T.

Infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) induces a complex antibody response that varies by orders of magnitude between individuals and over time. Waning antibody levels lead to reduced sensitivity of serological diagnostic tests over time. This undermines the utility of serological surveillance as the SARS-CoV-2 pandemic progresses into its second year. Here we develop a multiplex serological test for measuring antibodies of three isotypes (IgG, IgM, IgA) to five SARS-CoV-2 antigens (Spike (S), receptor binding domain (RBD), Nucleocapsid (N), Spike subunit 2, Membrane-Envelope fusion) and the Spike proteins of four seasonal coronaviruses. We measure antibody responses in several cohorts of French and Irish hospitalized patients and healthcare workers followed for up to eleven months after symptom onset. The data are analysed with a mathematical model of antibody kinetics to quantify the duration of antibody responses accounting for inter-individual variation. One year after symptoms, we estimate that 36% (95% range: 11%, 94%) of anti-S IgG remains, 31% (9%, 89%) anti-RBD IgG remains, and 7% (1%, 31%) anti-N IgG remains. Antibodies of the IgM isotype waned more rapidly, with 9% (2%, 32%) anti-RBD IgM remaining after one year. Antibodies of the IgA isotype also waned rapidly, with 10% (3%, 38%) anti-RBD IgA remaining after one year. Quantitative measurements of antibody responses were used to train machine learning algorithms for classification of previous infection and estimation of time since infection. The resulting diagnostic test classified previous infections with 99% specificity and 98% (95% confidence interval: 94%, 99%) sensitivity, with no evidence for declining sensitivity over the time scale considered. The diagnostic test also provided accurate classification of time since infection into intervals of 0 - 3 months, 3 - 6 months, and 6 - 12 months. Finally, we present a computational method for serological reconstruction of past SARS-CoV-2 transmission using the data from this test when applied to samples from a single cross-sectional sero-prevalence survey.