Background: The mechanism for reactivation of varicella zoster virus as herpes zoster (HZ) is not well understood. One hypothesis postulates that re-exposure to circulating wild-type varicella can boost individual immunity and prevent reactivation (Exogenous Boosting, EB). The validity of this hypothesis has been debated and the evidence to evaluate it is limited. Dynamic transmission model outcomes of impact and cost-effectiveness of universal varicella vaccination (UVV) are sensitive to EB characterization and assumptions, occasionally leading to conclusions that UVV programs may not be cost-effective and could lead to temporary increases in HZ incidence. The goal of this study was to use data from 20 years of UVV in the United States from 1996-2016 to evaluate whether the hypothesized increases in HZ incidence have been realized.
Methods: This is a retrospective study of de-identified administrative claims data from the US MarketScan® databases between 1991-2016. The incidence of HZ was analyzed by calendar year and age category using interrupted time series (ITS) analysis implemented through a negative binomial generalized linear regression model over three time periods: pre-UVV (1991-1995); 1 dose UVV (1996-2006); and 2 dose UVV (2007-2016). The ITS approach (Bernal et al, IJE, 2017) is an effective way to evaluate the impact of public health interventions implemented at specific time points.
Results: HZ incidence in the pre-UVV period increased at annual rates between 3.67%-12.38%, with the highest increases in the 0-17 and 65+ age groups. The rate of HZ increase was lower in the 1 dose UVV period compared to the pre-UVV period for all age groups except for minor increases in the 18-35 (0.52%) and 55-65 (0.14%) groups. During the 2 dose UVV period, the rate of increase in HZ was lower in all groups than in the pre-UVV period, with the largest reductions in the 0-17 (-22.58%), 65+(-10.68%), and 18-<35 (-3.57%) age groups.
Conclusion: This evaluation of the impact of UVV on rates of change in HZ does not support the hypothesis of an increase in HZ incidence due to UVV. While overall HZ incidence rates have been increasing year on year, the rate of that increase has been declining in both UVV periods. Our findings have implications on the assumptions used in economic evaluations of UVV programmes.
Merck & Co., Inc.:
Y. T. Chen, Merck & Co., Inc.: Employee and Shareholder , Salary .
W. Ou, Merck & Co., Inc.: Employee and Shareholder , Salary .