Comparison of performance of varicella vaccines via infectious disease modeling

Background: Population-level infectious disease models for varicella require vaccine parameters, namely 'take' and 'duration of protection' (defined here as vaccine performance), to quantify the impact of vaccination. Current published models for varicella use vaccine parameters derived from various methodologies which does not allow for the direct comparison of different vaccines.

Methods: We estimated take and duration of protection using deterministic compartmental models to simulate clinical trials of one- or two-dose varicella vaccination using Varivax® (V-MSD) and Varilrix® (V-GSK). We fit different models to clinical trial data on breakthrough infections and evaluated their respective goodness-of-fit using the Akaike Information Criterion (AIC).

Results: Based upon the clinical trial data, we estimated that 90.3% (95% CI: 87.8-92.9%) of the cohort gained permanent protection from breakthrough varicella after the first dose of V-MSD compared to 61.7% (95% CI: 58.2-65.3%) with the first dose of V-GSK. We further estimated that a total of 97.0% (95% CI: 95.2-98.8%) and 93.8% (95% CI: 92.2-95.4%) of the cohort were permanently protected after two-doses of V-MSD and V-GSK, respectively. According to the AIC, our new model (V-MSD AIC = 92.7; V-GSK AIC = 170.3) provided a better fit than an existing model (V-MSD AIC = 108.9; V-GSK AIC = 216.1).

Conclusions: The model developed fits the long-term clinical trial data on breakthrough infections for both V-MSD and V-GSK, thus, allowing for the direct comparison of vaccine performance. We estimated that a single dose of V-MSD was more likely to provide permanent protection than a single dose of V-GSK, while the protection offered by two doses was similar for both vaccines.