Modeling and predicting the long-term effects of various strategies and objectives of varicella-zoster vaccination campaigns

Point-Counterpoint: The Hope-Simpson Hypothesis and Its Implications Regarding an Effect of Routine Varicella Vaccination on Herpes Zoster Incidence

Some 50 years ago, Edgar Hope-Simpson published his hypothesis regarding the interactions between varicella and herpes zoster. As part of this hypothesis, Hope-Simpson postulated that reactivation of varicella zoster virus (VZV) was under immunological control, and that this immunological control could be boosted "endogenously" due to reactivation of latent VZV, and "exogenously" due to exposure to varicella. This hypothesis has important policy implications and remains a source of debate today; namely, does reducing VZV circulation through effective pediatric varicella vaccination programs lead to unintended increases in herpes zoster (HZ) incidence? This article provides 2 very different perspectives on this issue. The first perspective (Rafael Harpaz: Evidence Against an Effect) highlights the empiric experience of the United States, with its population of >300 million, a highly effective national varicella vaccination program lasting >20 years, and with several credible sources of data regarding HZ incidence. The US data have shown an increase in HZ incidence that preceded the availability of varicella vaccination by decades; indeed, HZ rates appear to have plateaued among older adults since varicella vaccination was introduced. Furthermore, HZ rates are not different in states having higher vs lower preschool varicella vaccination rates. The second perspective (Albert J. van Hoek: Evidence for an Effect) cites data that persons with close exposure to children appear to be at lower risk of HZ before universal VZV vaccination, but not so thereafter. Due to historic demographic changes, exogenous boosting could play a role in explaining the observed increase in HZ before varicella vaccination. Thus, it might be difficult to separate declines in exogenous boosting due to demographic changes from those caused by the varicella vaccination program. Additional data will be needed to conclusively rule out an impact of varicella vaccination on HZ.

Do varicella vaccination programs change the epidemiology of herpes zoster? A comprehensive review, with focus on the United States

Introduction: Policy-makers in many countries have been wary of introducing varicella vaccination programs because of concerns that reduced exposures to varicella-zoster virus could increase herpes zoster (HZ) incidence. The U.S. introduced varicella vaccination in 1996 and has empiric evidence regarding this concern. Areas covered: This comprehensive review provides background emphasizing the epidemiology of varicella and of HZ in the U.S. before and after the introduction of their respective vaccines. The epidemiology is complex, and interpretation is complicated by methodologic challenges, by unexplained increases in age-specific HZ incidence that preceded varicella vaccination, and by introduction of vaccines for prevention of HZ. Nonetheless, observations from studies using different platforms and designs have yielded consistent findings, suggesting they are robust. Expert opinion: There has been no evidence that the U.S. varicella vaccination program increased HZ incidence in the general adult population over baseline trends. Furthermore, HZ incidence in children is declining. The U.S. experience can inform the development of new generations of models to predict HZ trends. More importantly, it provides reassurance for countries considering varicella vaccination that an effective program can reduce varicella morbidity and mortality while reducing the likelihood of HZ among children, and potentially, over time, across the entire population.

Evaluation of the effect of chickenpox vaccination on shingles epidemiology using agent-based modeling

BACKGROUND

Biological interactions between varicella (chickenpox) and herpes zoster (shingles), two diseases caused by the varicella zoster virus (VZV), continue to be debated including the potential effect on shingles cases following the introduction of universal childhood chickenpox vaccination programs. We investigated how chickenpox vaccination in Alberta impacts the incidence and age-distribution of shingles over 75 years post-vaccination, taking into consideration a variety of plausible theories of waning and boosting of immunity.

METHODS

We developed an agent-based model representing VZV disease, transmission, vaccination states and coverage, waning and boosting of immunity in a stylized geographic area, utilizing a distance-based network. We derived parameters from literature, including modeling, epidemiological, and immunology studies. We calibrated our model to the age-specific incidence of shingles and chickenpox prior to vaccination to derive optimal combinations of duration of boosting (DoB) and waning of immunity. We conducted paired simulations with and without implementing chickenpox vaccination. We computed the count and cumulative incidence rate of shingles cases at 10, 25, 50, and 75 years intervals, following introduction of vaccination, and compared the difference between runs with vaccination and without vaccination using the Mann-Whitney U-test to determine statistical significance. We carried out sensitivity analyses by increasing and lowering vaccination coverage and removing biological effect of boosting.

RESULTS

Chickenpox vaccination led to a decrease in chickenpox cases. The cumulative incidence of chickenpox had dropped from 1,254 cases per 100,000 person-years pre chickenpox vaccination to 193 cases per 100,000 person-years 10 years after the vaccine implementation. We observed an increase in the all-ages shingles cumulative incidence at 10 and 25 years post chickenpox vaccination and mixed cumulative incidence change at 50 and 75 years post-vaccination. The magnitude of change was sensitive to DoB and ranged from an increase of 22-100 per 100,000 person-years at 10 years post-vaccination for two and seven years of boosting respectively (p < 0.001). At 75 years post-vaccination, cumulative incidence ranged from a decline of 70 to an increase of 71 per 100,000 person-years for two and seven years of boosting respectively (p < 0.001). Sensitivity analyses had a minimal impact on our inferences except for removing the effect of boosting.

DISCUSSION

Our model demonstrates that over the longer time period, there will be a reduction in shingles incidence driven by the depletion of the source of shingles reactivation; however in the short to medium term some age cohorts may experience an increase in shingles incidence. Our model offers a platform to further explore the relationship between chickenpox and shingles, including analyzing the impact of different chickenpox vaccination schedules and cost-effectiveness studies.