Several varicella vaccines are available worldwide. Countries with a varicella vaccination program use 1- or 2-dose schedules.
We examined postlicensure estimates of varicella vaccine effectiveness (VE) among healthy children.
Systematic review and descriptive and meta-analysis of Medline, Embase, Cochrane libraries, and CINAHL databases for reports published during 1995–2014.
Publications that reported original data on dose-specific varicella VE among immunocompetent children.
We used random effects meta-analysis models to obtain pooled one dose VE estimates by disease severity (all varicella and moderate/severe varicella). Within each severity category, we assessed pooled VE by vaccine and by study design. We used descriptive statistics to summarize 1-dose VE against severe disease. For 2-dose VE, we calculated pooled estimates against all varicella and by study design.
The pooled 1-dose VE was 81% (95% confidence interval [CI]: 78%–84%) against all varicella and 98% (95% CI: 97%–99%) against moderate/severe varicella with no significant association between VE and vaccine type or study design (P > .1). For 1 dose, median VE for prevention of severe disease was 100% (mean = 99.4%). The pooled 2-dose VE against all varicella was 92% (95% CI: 88%–95%), with similar estimates by study design.
VE was assessed primarily during outbreak investigations and using clinically diagnosed varicella.
One dose of varicella vaccine was moderately effective in preventing all varicella and highly effective in preventing moderate/severe varicella, with no differences by vaccine. The second dose adds improved protection against all varicella.
Marin et al report a VE of 81% (95% C.I.: 78-84) for the one-dose varicella vaccination protocol. [1] This figure is biased high and declines rapidly as the vaccine becomes widespread and exogenous boosting becomes rare. Many of the clinical trials and studies that reported VE were conducted within the first few years of the start of varicella vaccination—during a time period when vaccinees were additionally boosted by exogenous exposures to those shedding wild-type (or natural) varicella-zoster virus during annual outbreaks. Annual VE, derived from secondary family attack rate (SFAR) data among contacts aged <20 years reporting to the Antelope Valley Varicella Active Surveillance Project (VASP), demonstrated an annual increase from 87% in 1997 to 96% in 1999 (the last year that varicella displayed its characteristic seasonality), then declined to 85% and 74% in 2000 and 2001, respectively when exogenous boosting substantially decreased. [2]
In the conclusion, the Meta-analysis by Marin et al states that several studies reported a lower risk for herpes zoster among varicella vaccinated children “and a decline in herpes zoster incidence among cohorts targeted for varicella vaccination.” [1] While it is true that vaccinated individuals experience a lower risk for herpes zoster, the later part of the statement regarding a “decline” in HZ incidence among cohorts targeted for varicella vaccination is patently false. There are two confounders in the cited studies that contribute to this erroneous conclusion and a consideration of these confounders helps to provide an explanation why the VASP study [3] (referenced in the meta-analysis [1]) reports that the 2000-2006 HZ incidence increased by 63% among 10- to 19-year olds “without a conclusive explanation for the increase” and HZ incidence decreased by 32% from 98.3/100,000 person-years (p-y) in 2006 to 66.7/100,000 p-y in 2010, with “substantial fluctuation in annual HZ rates.”
The authors of both the Meta-Analysis [1] and supporting reference [3] have erroneously assumed that the HZ cases reported to the VASP represent 100% reporting completeness. However, using capture-recapture with two ascertainment sources (schools and physicians), it was shown that varicella cases among 2- to 18-year-olds were under-reported by approximately 45%. [4] Likewise, it can be shown that VASP also experienced approximately 50% under-reporting of HZ cases [2], leading to the Marin et al study [3] reporting incidence rates that are one-half the actual rates. HZ incidence rates that have not been ascertainment corrected simply reflect the incidence of reported HZ cases to the VASP and not the HZ incidence rate in the community. It is invalid to compare the VASP-reported HZ rates to those rates reported by other studies that possess much higher case ascertainment. [5]
Additionally, the >10- to 19-year old age category consists of three different cohorts with widely differing HZ incidence rates. Marin et al [3] considers only the mean HZ incidence rate for each age category [1]. However those in the >10- to 19-year old age category consist of three widely different cohorts: (1) those still susceptible to varicella and never vaccinated (i.e., 0 cases/100,000 p-y); (2) those that have had a prior history of wild-type varicella who exhibit increasing HZ incidence rates from approximately 120 cases/100,000 p-y to 500 cases/100,000 p-y (in the absence of exogenous boosting); and (3) those vaccinated who exhibit an HZ incidence rate less than 120 cases/100,000 p-y.
In summary, unless HZ incidence rates are ascertainment corrected [5], such rates will erroneously be reported as “lower” than other studies. [1] Also, reporting the mean HZ incidence of a bimodal distribution masks the widely differing incidence rates among those vaccinated and those with a prior history of varicella. Further, this invalid mean masks the significant effects of exogenous boosting. [7]
Varicella vaccination innoculates children with the Oka-strain VZV. When these children are exposed to natural varicella or herpes zoster in adults, they may additionally harbor the natural VZV strain. Both strains are subject to reactivation as HZ. This is another confounder in the reporting of HZ incidence rates. Health officials initially believed that only a single dose of varicella vaccine would provide long-term protection and have negligible impact on the incidence of HZ. These assumptions are incorrect and have led to a continual cycle of treatment and disease. The shingles vaccine now provides the boosting to postpone or suppress the reactivation of HZ in adults aged 60 years and older—a substitute for the exogenous boosting that was obtained at no cost and prevalent in the pre-varicella vaccination era. [5]
[1] Marin M, Marti M, Kambhampati A, Jeram SM, Seward JF. Global varicella vaccine effectiveness: A meta-analysis. Pediatrics Feb. 16, 2016; DOI: 10.1542/peds.2015-3741.
[2] Goldman GS. Universal Varicella Vaccination: Efficacy Trends and Effect on Herpes Zoster. Int J Toxicol 2006 Sep-Oct; 25(5):313-317.
[3] Marin M. Civen R, Zhang J, et al. Update on incidence of herpes zoster among children and adolescents following implementation of varicella vaccination, Antelope Valley, CA. 2000-2010. Presented at IDweek 2015, October 7-11, 2015; San Diego, CA.
[4] Seward JF, Watson BM, Peterson CL, Mascola L, Pelosi JW, Zhang JX, et al. Varicella disease after introduction of varicella vaccine in the United States, 1995-2000. JAMA 2002; 287(5):606-611.
[5] Hook EB, Regal RR. The value of capture-recapture methods even for apparent exhaustive surveys: the need for adjustment for source of ascertainment intersection in attempted complete prevalence studies. Am J Epidemiol 1992; 135:1060-1067.
[6] Goldman GS, King PG. Review of the United States universal varicella vaccination program: Herpes-zoster incidence rates, cost effectiveness, and vaccine efficacy based primarily on the Antelope Valley Varicella Active Surveillance Project data. Vaccine 2013; 31(13):1680-1694.
[7] Guzzetta G, Poletti P, Del Vava E, et al. Hope-Simpson’s progressive immunity hypothesis as a possible explanation for herpes –zoster incidence data. Am J Epidemiol 2013; 77(10):1134-1142.