Defining the ever-evolving correlates of protection against symptomatic infection is critical in light of the widespread deployment of mRNA vaccines in children. Since immune maturation and exposure histories differ between children and adults, immune kinetics and correlates of protection identified in adult cohorts may not directly generalize to pediatric vaccine recipients. A prospective cohort of 5–12-year-old children (N = 70) was monitored longitudinally for 12 months after SARS-CoV-2 mRNA (BNT162b2) vaccination. Four immunological biomarkers were assessed: anti-Spike immunoglobulin G (anti-S IgG), neutralizing antibodies (nAbs), Spike-specific memory B cells (S+ MBCs), and S-reactive T cell response. We utilized mathematical models to reconstruct time-varying biomarker trajectories, while accounting for multiple immunity-conferring events (i.e., primary vaccination, breakthrough infection, and booster vaccination). Using the biomarker levels as time-varying covariates in survival analyses, we evaluated the circulating correlates of protection against symptomatic breakthrough SARS-CoV-2 infection across three post-primary vaccination periods (30 days to 3 months, 3 to 6 months, and 6 to 12 months), as an interpretable summary of early, intermediate, and late phases. We also repeated this evaluation in a global analysis without considering specific time windows. Serological markers were best described by the exponential model, while S+ MBC and T cell responses followed the inactivation model. While early protection was associated with antibodies, multivariable analyses identified T cell responses as the primary correlate of protection from 3 months onward, and this association was strongest in the presence of hybrid immunity. Under our prespecified threshold for the T cell response, the model-projected protection duration under hybrid immunity was estimated to be approximately 500 days after primary vaccination; however, empirical validation in independent cohorts will be required. Importantly, we obtained similar conclusions on the identity and direction of the correlates of protection in several sensitivity analyses adjusting for model structure, accounting for age and sex, and including asymptomatic infections. Our findings highlight the time-dependent nature of correlates of protection, emphasizing the need for employing mathematical models to reinforce the accuracy and reliability of clinical analysis so as to prevent overreliance on single-timepoint measurements of immune parameters.
Liao et al. (Wed,) studied this question.