A statistical immune correlates of protection model for predicting efficacy from neutralizing antibody titers to establish immunobridging of monoclonal antibodies for prevention of COVID-19
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Neutralizing antibody titers are recognized as an acceptable surrogate efficacy endpoint for immunobridging next-generation monoclonal antibodies (mAbs) to those with demonstrated clinical efficacy for the prevention of COVID-19. However, titers measured at early time points after dosing overestimate the titer levels required for clinical protection, but long-term efficacy data is limited due to continued evolution of SARS-CoV-2 and loss of activity of previously effective mAbs against emerging variants. To address these challenges, we set out to develop a predictive tool for efficacy using neutralizing titers to establish immunobridging for mAbs for prevention of COVID-19.
Methods
Drug concentration and clinical efficacy data collected over 12 months following administration of pemivibart in the phase 3 CANOPY trial for prevention of COVID-19 were used to develop a Cox proportional hazards model with time-varying covariate as a statistical immune correlates of protection (CoP) model. The time-varying covariate was estimated serum virus neutralizing antibody (sVNA) activity, which changes over time in response to the emergence and prevalence of different SARS-CoV-2 variants. sVNA was estimated at 2-week intervals by integrating drug concentration data with weighted average IC 50 (half-maximal inhibitory concentration) values of the circulating variant population. This model was used to predict clinical efficacy in both immunocompromised (IC) and non-immunocompromised (non-IC) populations based on sVNA titers.
Results
Efficacy increased with antibody titer in a non-linear manner, with smaller incremental gains at higher concentrations. Predicted efficacy was lower at all titer levels in the IC cohort. Model-derived estimates aligned well with observed infection outcomes and external analyses, supporting model validity. Based on the model, a sVNA titer of 1:500 predicted an estimated 50% efficacy in IC and 70% efficacy in non-IC populations, where efficacy is defined as the relative reduction in the risk of COVID-19 infection.
Conclusion
The Cox model provides a foundation for evaluation of suitable neutralizing titer targets to guide dosing, predict estimated clinical benefit, and support immunobridging in both IC and non-IC populations.
