BACKGROUND: Varicella vaccination confers high and long-lasting protection against chickenpox and induces robust immune responses, but an absolute correlate of protection (CoP) against varicella has not been established. This study models the relationship between varicella humoral response and protection against varicella. METHODS: This was a post-hoc analysis of data from a Phase IIIb, multicenter, randomized trial (NCT00226499) conducted in ten varicella-endemic European countries. Healthy children aged 12-22 months were randomized 3:3:1 to receive one dose of measles-mumps-rubella and one dose of varicella vaccine (one-dose group) or two doses of measles-mumps-rubella-varicella vaccine (two-dose group) or two doses of measles-mumps-rubella vaccine (control group) six weeks apart. The study remained observer-blind until completion, except in countries with obligatory additional immunizations. The objective was to correlate varicella-specific antibody concentrations with protection against varicella and probability of varicella breakthrough, using Cox proportional hazards and Dunning and accelerated failure time statistical models. The analysis was guided by the Prentice framework to explore a CoP against varicella. RESULTS: The trial included 5803 participants, 5289 in the efficacy (2266: one-dose group, 2279: two-dose group and 744: control group) and 5235 (2248, 2245 and 742 in the same groups) in the immunogenicity cohort. The trial ended in 2016 with a median follow-up time of 9.8 years. Six weeks after vaccination with one- or two-dose varicella-containing vaccine, more than 93.0% of vaccinees were seropositive for varicella-specific antibodies. Estimated vaccine efficacy correlated positively with antibody concentrations. The fourth Prentice CoP criterion was not met, due to predicted positive vaccine efficacy in seronegative participants. Further modelling showed decreased probability of moderate to severe varicella breakthrough with increasing varicella-specific antibody concentrations (ten-year probability <0.1 for antibody concentrations ≥2-fold above the seropositivity cut-off). CONCLUSIONS: Varicella-specific antibody concentrations are a good predictor of protection, given their inverse correlation with varicella occurrence. CLINICAL TRIAL: NCT00226499.

Charles University Faculty of Medicine Department of Social Medicine Simkova street 870 500 03 Hradec Kralove Czechia

Clinic of Children's Diseases Institute of Clinical Medicine Faculty of Medicine Vilnius University M K Čiurlionio g 21 27 03101 Vilnius Lithuania

Department of Preventive Medicine University of Medical Sciences ul Święcickiego 6 60 781 Poznań Poland

GSK 14200 Shady Grove Road Rockville MD 20850 USA

GSK Avenue Fleming 20 1300 Wavre Belgium

GSK Rue de l'Institut 89 1330 Rixensart Belgium

Pediatric Clinic Pietro Barilla Children's Hospital Department of Medicine and Surgery University of Parma Via Gramsci 14 43126 Parma Italy

References provided by Crossref.org