Antibody Persistence and Risk of COVID-19 Infection: Insights from Modeling

Abstract

Background: In this post hoc exploratory study of the APHP-COVIBOOST trial (NCT05124171), we used statistical modeling to describe the evolution of neutralizing antibody (nAb) titers over time, asses its impact on SARS-CoV-2 infection, and explore potential differences between three booster vaccine formulations (D614, B.1.351, and BNT162b2). Methods: Antibody titers were measured for 208 adult participants at day 28, 3 months, and 6 months using a microneutralization assay against different Omicron subvariants. We developed four specific Bayesian statistical models based on a core model, accounting for vaccine-specific antibody decline, boosting of nAb for breakthrough infection, and risk of infection according to nAb levels. The model findings were cross-verified using another validated microneutralization assay. Results: The decrease in nAb titers was significantly lower for the B.1.351 vaccine than for the other booster formulations. An inverse relationship was found between risk of infection upon exposure and nAb levels. With Omicron BA.1 data, these results translated into a positive relative vaccine efficacy against any infection over 6 months for the B.1.351 vaccine compared to the BNT162b2 (31%) and D614 (21%) vaccines. Conclusions: Risk of infection decreased with increasing nAb titers for all vaccines. Statistical models predicted significantly better antibody persistence for the B.1.351 booster formulation compared to the other evaluated vaccines.

Keywords

• antibody persistence
• statistical modeling
• COVID-19
• correlates of protection
• neutralizing antibody titer