A predictive model for rapid assessment of protective efficacy against emerging SARS-CoV-2 variants

An effective predictive model of protection would be very helpful to provide a timely and reliable evaluation of the vaccine induced protection against corresponding to rapidly emerging evolving SARS-CoV-2 variants. By integrating the validated "neutralizing antibody-vaccine efficacy" and "modified genetic distance-vaccine efficacy" models, we developed a composite sieve analysis framework (the "neutralizing antibody-genetic distance-vaccine efficacy" model) to predict the protective efficacy of COVID-19 vaccine regimens, particularly for different heterologous prime-boost COVID-19 vaccination regimens. Data for the model building were extracted from 23 published studies. Leave-one-out method was used to validate the model. Model validation demonstrated that the composite framework achieved high predictive accuracy, with concordance correlation coefficients of 0.95 (95% CI: 0.82-0.98) for the "neutralizing antibody-vaccine efficacy" submodel and 0.93 (95% CI: 0.49-0.99) for the "modified genetic distance-vaccine efficacy" submodel. Most prediction errors were within 5% and 10%, respectively. By applying this framework with neutralizing antibody data and SARS-CoV-2 variant sequencing data, we predicted the protective efficacy of different heterologous prime-boost regimens. The regimen of two-dose CoronaVac plus one-dose aerosolized Ad5-nCoV was estimated to confer 95.40% (95% CI: 92.67-98.13%) protection against symptomatic infection with wild-type SARS-CoV-2 at day 28 post-boost, 79.56% (95% CI: 53.31-100.00%) against the Delta variant, and 68.21% (95% CI: 42.53-93.89%) against Omicron BA.5.2.20. Predicted efficacy against other Omicron sublineages was generally below 50%, with near-zero efficacy for KP.2, KP.3 and XDV.1. Compared with this regimen, two-dose CoronaVac plus one-dose intramuscular Ad5-nCoV booster or three-dose CoronaVac yielded consistently lower predicted efficacy across all variants. This study offers a generalizable approach for rapidly evaluating the efficacy of COVID-19 vaccines against emerging variants, providing timely evidence to guide vaccine deployment in future outbreaks.