Varicella-zoster virus seroprevalence in children and adolescents in the pre-varicella vaccine era, Germany

Varicella-zoster virus (VZV) is a ubiquitous human herpesvirus. Primary infection with VZV results in varicella (chickenpox), which mostly affects children and is regarded as a generally benign illness [1]. However, varicella can also lead to serious complications resulting in hospitalization or even death [2]. Particularly in adults or in immunocompromised patients, varicella can take a severe course. After primary infection, immunity against varicella is considered to protect life-long. However, VZV becomes latent in the dorsal root ganglia; and later in life, VZV reactivation can cause herpes zoster (shingles).

In 2004, the German Standing Committee on Vaccination (STIKO) recommended universal varicella vaccination for children with a single dose to be given at the age of 11–14 months [3]. In 2009, a two-dose schedule was endorsed with a second dose at the age of 15–23 months [4]. In the birth cohorts of 2004 to 2009, varicella vaccination coverage increased in 24-month-old children from 43% to 87% for the first dose and from 1% to 64% for the second dose [5]. With the increase in coverage, a significant decrease in varicella cases was observed in a nationwide physician-based sentinel surveillance system: from April 2005 to March 2012, an 85% decline in varicella cases was reported by sentinel physicians. Furthermore, from April 2005 to March 2009, complications or severe courses of varicella declined by 81% [6, 7]. Based on sentinel and coverage data, vaccine effectiveness was estimated to be 87% after one vaccine dose and 97% after two vaccine doses [8].

For the evaluation of the current universal varicella vaccination strategy, disease burden data before and after vaccine introduction as well as data on vaccination uptake are essential [9]. Seroepidemiological data before and after vaccine introduction serve as an additional pillar in the evaluation of the vaccination strategy and can guide recommendations on catch-up vaccination activities. Although enzyme-linked immunosorbent assays (ELISAs) are the most commonly used test for measuring immunity against VZV (anti-VZV IgG antibody levels), the fluorescent antibody to membrane antigen test (FAMA) is considered to be the “gold standard” assay for determining immunity against VZV, however, FAMA is more labor-intensive and time-consuming [10]. Studies directly comparing these two test methods are rare.

The objectives of this study were: (1) to representatively describe the age-specific prevalence of anti-VZV IgG antibodies for children and adolescents in Germany (1–17 years of age) in the pre-varicella vaccine era, (2) to identify social factors that are associated with the acquisition of anti-VZV IgG antibodies by naturally acquired varicella, and (3) to compare the commercially available standard ELISA versus an in-house FAMA with respect to its suitability as a test for use in population-level serosurveys.

Here, we report for the first time representative population-based VZV seroprevalence data for children aged 1 to 17 years in the pre-varicella vaccine era in Germany. VZV seropositivity increased with age, with 60% of children testing positive at 4 years of age and >90% testing positive at 9 years of age. This gradual increase in VZV seropositivity by age is comparable with the VZV seroprevalence data of German children <18 years of age that were reported by Wutzler et al. in 2001 [13]. For this earlier seroprevalence study, two serum banks were used with residual serum samples that had originally been collected for routine laboratory diagnostics or within the German National Health Interview and Examination Survey between 1995 and 1999. They found that VZV seropositivity was 62% in the 4- to 5-year old children and >90% in the 10- to 11-year old children. Thus, the seroprevalence data of the 1995–1999 study are similar to ours, and we can assume that there was no significant change in the VZV seroprevalence in children in Germany for the time period between 1995 and 1999 and 2003–2006. Also similar to our results, the majority of varicella infection occurred in early childhood in Northern European countries without universal varicella vaccination [16]. Interestingly, children in the Netherlands acquired anti-VZV IgG antibodies, on average, even earlier in childhood than did children in Germany or in Nordic countries, and some have proposed that the relatively high population density in the Netherlands might explain this observation [17]. There are contradicting reports from the Netherlands concerning whether or not VZV seropositivity is associated with day-care attendance [17, 18]. In our study, we found an association between VZV seropositivity and an early (<3 years of age) start of day care attendance. However, this finding may be explained by the low rate of young children (0–3 years of age) who attend day care facilities at all in Germany. In 2006, around 12% of children who were <3 years of age visited a day care center in Germany [19]; in contrast, around 50% of children in the Netherlands who were <3 years of age attended day care facilities in 2010 [20]. Since we can assume that in the absence of varicella vaccination the force of varicella infection was high in day care facilities, we were able to discriminate in regard to VZV seropositivity between children with an early start of day care attendance and children who stayed at home. Additionally, we showed that siblings living in the same household acted as a risk factor for acquiring varicella: for young children, older siblings in the same household were a risk factor for VZV seropositivity, and, for older children, their younger siblings played this role. Thus, in the absence of universal varicella vaccination in children, our data confirm that children are the major force of varicella infection.

There are concerns about an upward shift in varicella infection age following the introduction of universal varicella vaccination in young children [21]. Our study adds to the efforts of monitoring the impact of varicella vaccination in Germany, in part to address these concerns. Despite these worries, this phenomenon has not been observed in varicella surveillance in Germany, which is a physician-based sentinel system established in 2005 shortly after the initiation of universal childhood varicella vaccination [22]. Similarly, in the United States, where universal varicella vaccination in children was implemented in 1995, no shift in varicella cases towards older age groups has been observed as of today [23]. This present serosurvey will serve as a baseline assessment, and a follow-up survey (KiGGS wave 2) which will allow a comparison between more recent results and those of this present study, is currently underway and will recruit children from 2014 to 2017.

In our survey, we detected subgroups within the general population that have an increased risk of being susceptible to varicella infection if they are not reached by an ad-hoc catch-up vaccination; children 7–17 years of age with a migration background and children without siblings were less likely to be anti-VZV IgG-positive. Since no concerted catch-up campaign has been implemented among older children or adolescents following the introduction of universal childhood vaccination in toddlers, pediatricians and general practitioners should be aware of this fact and provide individual varicella catch-up vaccinations to these risk groups if an individual lacks a history of chickenpox and has no documented varicella vaccination. STIKO recommends individual catch-up vaccination to all children up to their 18th birthday and to all VZV-seronegative women of childbearing age.

We compared the detection of anti-VZV IgG by a commercially available ELISA with that by an in-house FAMA, which can be regarded as the gold standard [24]. We observed that 45.8% of samples with equivocal ELISA results were assessed as positive by FAMA. Additionally, 2.6% and 9.0%, respectively (depending on the age of the tested children), of ELISA-negative serum samples were determined to be positive by FAMA. As the gold standard assay, FAMA has a higher sensitivity and a lower detection limit than an ELISA [25]. The EUROIMMUN Anti-VZV (IgG) ELISA was reported to have a significantly lower sensitivity than FAMA, especially for the detection of anti-VZV IgG after varicella vaccination has been performed [14]. Interestingly, when serum samples that produced a negative or equivocal ELISA result were tested by FAMA, we more often observed a positive FAMA result in samples from varicella-vaccinated children than in samples from varicella-unvaccinated ones (Table 3). Although there was only a small number of varicella-vaccinated children in the study sample, our findings might be important for the monitoring of VZV seroprevalence in a varicella-vaccinated population. Because it is possible that anti-VZV IgG levels in varicella-vaccinated populations are overall lower, which would lead to a higher frequency of equivocal or even negative ELISA results, retesting these sera by FAMA might become much more relevant. Unfortunately, FAMA is a labor-intensive and time-consuming method that is not amenable to automation and requires experience for an appropriate interpretation of its results. Currently, FAMA tests are only performed in laboratories that are highly specialized in VZV diagnostics and research, and they are not commercially available. Therefore, it will be difficult to conduct FAMA for the detection of anti-VZV IgG on a routine basis.

Our study has several limitations. For the weighted seroprevalence rate calculation for varicella-unvaccinated children 1–17 years of age, we excluded 1156 study subjects with at least one reported varicella vaccination or with no vaccination card presented. Furthermore, most of the information on migration status, family members, and SES were self-reported or reported by the parents. However, this limitation would only influence the results of the risk factor analysis, not the VZV seroprevalence data.

In our study, we determined the age-specific VZV seroprevalence among children and adolescents in the pre-varicella vaccine era in Germany. We confirmed that, in the absence of universal varicella vaccination, VZV seropositivity increased with age, especially in the first 6 years of life, which suggests that the major force of varicella transmission occurred in young children. We identified children without siblings in the same household and children with a migration background as being more susceptible to contracting varicella, even at higher ages (7–17 years old). Therefore, varicella-unvaccinated children or adolescents should be targeted for catch-up activities if they lack a history of chickenpox.

We also compared the test qualities of a commercially available ELISA with those of a gold standard test (FAMA). A high number of serum samples with a negative or equivocal result in the ELISA produced a positive FAMA result, especially those from varicella-vaccinated children. Our findings suggest that the commercially available ELISA can be used for the monitoring the VZV seroepidemiology in populations with or without an established universal varicella vaccination, but that FAMA testing is superior in situations with equivocal ELISA results or in ELISA-negative samples from individuals who were vaccinated against varicella in the past. This study provides baseline data on VZV seroepidemiology in children and adolescents for monitoring the impact of the recently implemented varicella vaccination strategy in Germany. Moreover, these data act as input for mathematical modelling of the varicella and herpes zoster-attributable disease burden and of future epidemiological effects of the varicella-vaccination [26].