Elsevier

Vaccine

Volume 18, Issue 25, 15 June 2000, Pages 2775-2778
Vaccine

Analysis of varicella vaccine breakthrough rates: implications for the effectiveness of immunisation programmes

https://doi.org/10.1016/S0264-410X(00)00100-6Get rights and content

Abstract

The objective of this study was to quantify key parameters describing varicella zoster virus (VZV) vaccine efficacy. To do so a mathematical model was developed to represent breakthrough cases as a function of time after vaccination in vaccine efficacy trials. Efficacy parameter sets were identified by fitting the predicted annual number of breakthrough infections with that observed in three clinical trials chosen to represent the plausible range of vaccine efficacy. Results suggest that varicella vaccination seems to result in a high proportion of individuals who are initially totally protected (97% for the base-case). However, individuals lose full protection relatively rapidly (3% per year for the base-case). Once total protection has waned individuals have a high probability of developing a breakthrough infection if exposed to varicella (73% of the probability in unvaccinated susceptibles for the base-case). Results also highlight that vaccine efficacy parameters should be estimated concurrently to take into account dependencies between parameters.

Introduction

In clinical trials vaccine efficacy is measured as the relative risk of infection in the vaccinated group compared with unvaccinated individuals [1]. In general, vaccine efficacy is comprised of three components [2]: the proportion of individuals who acquire protection; the degree to which they are protected; and the duration of that protection. It can be difficult to estimate the contribution of each effect to the overall vaccine efficacy measured in controlled trials, because these often have a short time span relative to the period over which the vaccine is intended to offer protection (often life-time). Clearly, however, it is critical to have good estimates of these parameters to allow accurate long-term predictions of the impact of vaccination programs to be made.

Halloran et al. attempted to estimate vaccine efficacy parameters for varicella zoster (VZV) vaccine by reviewing the literature available at the time and employing a Delphi process [3]. These parameter values have been used in subsequent studies (such as Coudeville et al. [4]). Since many countries are in the process of evaluating the cost-effectiveness of varicella vaccination it is now timely to re-examine these estimates and amend them if necessary.

In this paper, we re-evaluate Halloran et al.’s estimates in the light of new data. Additionally, Halloran et al. used the standard approach of estimating each parameter separately. This approach can result in a poor fit to the data when all the parameters are combined even though each individual estimate seems plausible. Our approach differs in that we estimate the parameters simultaneously from the observed breakthrough rates (varicella infection in seroconverted vaccinees [3]) over time.

Section snippets

Selection of studies

Most of the published studies of vaccine immunogenicity and efficacy have involved the Oka/Merck product (Varivax®, Merck Research Laboratories). Published data using the other manufacturer’s vaccines were too few to include in the analysis.

The pre-licensure studies of Varivax® involved four different lots prepared in 1982, 1984, 1987 and 1991 [5], [6], [7], [8], [9], [10], [11], [12], [13], [14]. These lots differed in the relative amounts of both live virus and non-viable viral antigen as

Results

Fig. 2 shows the data and the best fit model for the best, worst and base-cases (best fit parameter values are given in Table 1). For comparative purposes the annual proportion of the cohort with breakthrough infections as predicted using Halloran et al.’s base-case and worst-case (“low-efficacy”) parameter combinations (see Table 1) are also shown in Fig. 2. It is clear that none of Halloran et al.’s parameter sets fit the data: even the least optimistic vaccine assumptions give far fewer

Discussion

We have updated Halloran et al.’s vaccine efficacy parameters. Our estimates yield higher breakthrough rates, which give a better fit to the observed data. Indeed, the annual number of breakthrough cases with Halloran et al.’s worst-case scenario is lower than that of the lowest observed breakthrough data (lot 1982, with passive reporting). Thus studies using Halloran et al.’s parameter sets may be overestimating the predicted effectiveness of varicella immunisation.

This study also shows that

Acknowledgements

We wish to acknowledge Merck Research Laboratories for generously providing us with previously unpublished data.

References (15)

There are more references available in the full text version of this article.

Cited by (47)

  • The cost-effectiveness of varicella and combined varicella and herpes zoster vaccination programmes in the United Kingdom

    2012, Vaccine
    Citation Excerpt :

    Varicella vaccination was introduced into the United States in the mid 1990s [1–3]. However, other countries have been slower to introduce varicella prevention programmes, partly as a result of concerns that infant vaccination may increase the occurrence of adult varicella (which tends to be more serious) [4,5], or may increase the incidence of herpes zoster [6–9], and partly because data from the clinical trials [10] and subsequent surveillance [3,11,12] suggest that vaccination may not provide solid, lasting protection for many vaccinees. Previous modelling and economic analyses (e.g. [13]) have addressed some, or all of these concerns (see [14] and [15] for reviews), and shown that an increase in herpes zoster (in particular) can have a profound effect on the estimated cost-effectiveness of infant varicella vaccination [13].

  • Modelling the impact of a combined varicella and zoster vaccination programme on the epidemiology of varicella zoster virus in England

    2011, Vaccine
    Citation Excerpt :

    With the licensure of a vaccine to prevent HZ [13,14], any increase in HZ following childhood vaccination could be offset (at least in part) by vaccination of the elderly (60+ years) against Herpes Zoster. There are other concerns with varicella vaccination, which include the potential increase in adult varicella (which tends to be more serious than childhood infection) that may occur following mass childhood vaccination [11,15,16], and concerns regarding the efficacy of a single dose of the vaccine at preventing varicella [17–19]. This latter concern has lead to recommendations for two-dose policies in childhood [3,20], which has an obvious impact on the economic attractiveness of the programme.

  • Modeling the impact of one- and two-dose varicella vaccination on the epidemiology of varicella and zoster

    2010, Vaccine
    Citation Excerpt :

    This assumption results in a lower overall rate of zoster in vaccinees given that many will not develop breakthrough varicella. Using methods similar to those described in Brisson et al. [40], vaccine efficacy parameter values were estimated by fitting simultaneously: (1) the incidence of varicella amongst individuals receiving 1 and 2 doses in the Kuter et al. [5] trial, and (2) the proportion of mild, moderate and severe varicella among vaccinated individuals [5,21] (see Appendix A for model fit). Five different vaccine efficacy model structures were investigated, by setting parameters such as the proportion of primary failures (F) or the degree of protection in vaccinated susceptibles (1-b) to 0.

  • Economic evaluation of varicella vaccination in Italian children and adolescents according to different intervention strategies: The burden of uncomplicated hospitalised cases

    2008, Vaccine
    Citation Excerpt :

    The waning rate per year (i.e. the proportion of people initially completely protected by the vaccine who will become only partially protected over time) was set to 3% and reflects a worst case scenario [5–7,25,26]. The relative susceptibility (i.e. the likelihood of a vaccinated individual to contract varicella relative to an unvaccinated individual following a single exposure to the virus) was set at 73%, which can also be considered a worst case scenario [5,7,25,26]. The relative infectiousness (i.e. the likelihood of a vaccinated patient contracting varicella to transmit the infection to an unvaccinated patient) was assumed to be 40% [15].

  • VACCINES

    2008, Pharmacology and Therapeutics: Principles to Practice
View all citing articles on Scopus
View full text