Background: The soon to come the availability of a combined MMR-varicella vaccine has re-stimulated the debate around universal infant vaccination against varicella. In France, the incidence of varicella is estimated at about 700,000 cases per year, with approximately 3500 hospitalisations and 15-25 deaths, the latter mainly occurring in those over 15years. Vaccination would certainly decrease the overall incidence of the disease but concerns about vaccination leading to a shift in the average age at infection followed by an increase in incidence of severe cases and congenital varicella, still remain. In order to provide support for decision-making, a dynamic mathematical model of varicella virus transmission was used to predict the effect of different vaccination strategies and coverages on the epidemiology of varicella and zoster.
Methods: A deterministic realistic age-structured model was adapted to the French situation. Epidemiological parameters were estimated from literature or surveillance data. Various vaccine coverages and vaccination strategies were investigated. A sensitivity analysis of varicella incidence predictions was performed to test the impact of changes in the vaccine parameters and age-specific mixing patterns.
Results: The model confirms that the overall incidence and morbidity of varicella would likely be reduced by mass vaccination of 12-month-old children. Whatever the coverage and the vaccine strategy, the vaccination will cause a shift in age distribution with, for vaccination coverage up to at least 80% in the base-case analysis, an increased morbidity among adults and pregnant women. However, the total number of deaths and hospitalisations from varicella is predicted to remain below that expected without vaccination. The model is very sensitive to the matrix of contacts used and to the parameters describing vaccine effectiveness. Zoster incidence will increase over a number of decades followed by a decline to below prevaccination levels.
Conclusion: Mass varicella vaccination, in France, will result in an overall reduction of varicella incidence but will cause a shift in age distribution with an increase in adult cases. Due to the uncertainties in key parameters values, the exact magnitude of this shift is difficult to assess.