Clinically- versus serologically-identified varicella: A hidden infection burden. A ten-year follow-up from a randomized study in varicella-endemic countries

Abstract

Varicella-zoster virus (VZV) infections cause a substantial disease burden, which is underestimated due to incomplete reporting data and lack of serological surveillance. In this post-hoc analysis of a randomized, Phase IIIb clinical trial (NCT00226499) with a ten-year follow-up period, we report anti-VZV antibody levels and persistence in non-vaccinated children, as a varicella infection estimate in ten European countries with endemic varicella. The present analysis specifically focuses on clinical and serological data from the control group, which included 827 healthy participants aged 12–22 months who received two doses of measles-mumps-rubella (MMR) vaccine. The per-protocol cohort included 744 children for whom varicella occurrence was evaluated by clinical definitions, epidemiological links and PCR test outcomes. Anti-VZV antibody levels were assessed by ELISA. The primary objective of this analysis was to correlate varicella occurrence with anti-VZV antibody levels. Varicella was confirmed in 47% of MMR recipients. Among participants without reported varicella, the percentage of anti-VZV seropositive children increased to 75% and average anti-VZV antibody concentrations increased to 250 mIU/mL at year ten after vaccination, suggesting infection or exposure. An eight-fold increase in anti-VZV antibody concentrations between two consecutive visits, which is also observed after confirmed varicella, was detected in 37% of these participants during the follow-up period. About one-third of children not vaccinated against varicella and not diagnosed with varicella developed an anti-VZV immune response, suggesting subclinical varicella occurrence. Longitudinal studies combining serology and disease incidence are necessary to reliably estimate total varicella burden of infection.