Abstract
Varicella is a highly contagious disease caused by primary infection with varicella zoster virus (VZV). VZV infection, as well as varicella vaccination, induces VZV-specific antibody and T-cell-mediated immunity, essential for recovery. The immune responses developed contribute to protection following re-exposure to VZV. When cell-mediated immunity declines, as occurs with aging or immunosuppression, reactivation of VZV leads to herpes zoster (HZ). It has been almost 20 years since universal varicella vaccination has been implemented in many areas around the globe and this has resulted in a significant reduction of varicella-associated disease burden. Successes are reviewed here, whilst emphasis is put on the challenges ahead. Most countries that have not implemented routine childhood varicella vaccination have chosen to vaccinate high-risk groups alone. The main reasons for not introducing universal vaccination are discussed, including fear of age shift of peak incidence age and of HZ incidence increase. Possible reasons for not observing the predicted increase in HZ incidence are explored. The advantages and disadvantages of universal vs targeted vaccination as well as different vaccination schedules are discussed.
Keywords
Introduction
Varicella zoster virus
Varicella zoster virus (VZV) is an alpha herpesvirus that infects exclusively humans [Hambleton and Gershon, 2005]. Each VZV virion consists of four parts: (1) the core (made of a linear double-stranded DNA); (2) the capsid; (3) the tegument (that surrounds the capsid); and (4) the envelope (that surrounds the tegument and incorporates the major viral glycoproteins [gps]) [Arvin, 1996]. VZV’s virion is spherical with a diameter of 180–200 nm while the core consists of a single linear double-stranded DNA 125 kb long. VZV is the smallest of the human herpes viruses and its envelope is made out of gps of about 8 nm long. During lytic infection the virus produces at least 12 gps which are expressed both on the surface of the virions and the envelope of infected human cells [Gershon and Gershon, 2010].
VZV associated diseases
VZV DNA is present in respiratory secretions (upper respiratory) and fluid secretions from skin vesicles. It is transmitted through direct contact with the skin lesions, through respiratory secretions or through inhalation of the airborne particles of the virus.
Primary infection is manifested as varicella (or chickenpox) and leads to lifelong latent infection of sensory ganglion neurons. Reactivation of the latent infection causes HZ (HZ, shingles) [Hambleton and Gershon, 2005].
Varicella (or chickenpox) is a highly contagious disease in children with an estimated household secondary attack rate of 90% [Arvin, 1996; Ross, 1962]. During the incubation period (10–21 days), the virus initially replicates in the upper respiratory tract and through a primary subclinical viremia, the virus spreads to the reticuloendothelial system (liver, spleen) and other organs. Following additional viral replication, a second viremic phase ensues and prodromal clinical symptoms (fever, malaise) appear followed by the eruption of the typical rash. The rash consists of pruritic, maculopapules, vesicles and crusted lesions in varying stages of evolution. They quickly progress from one stage to the next. In the beginning, each skin vesicle develops on an erythematous base, then progresses into a pustule and then into a crusted pustule. The distribution of the lesions is mainly central. It affects the trunk and the face and spreads rapidly to the rest of the body. New crops of vesicles are generated for 3–7 days from virus colonies in the peripheral blood mononuclear cells. The average number of lesions is 300 (10–1500).
Importantly, during the late phase of the incubation period, the virus is transported to the respiratory system, spreading to susceptible contacts, before the characteristic rash appears.
Although varicella is a generally benign disease, even uncomplicated cases experience significant discomfort due to itching. Complications are not uncommon and mainly involve secondary skin and soft tissue infections (streptococcal and staphylococcal) and central nervous system involvement including either viral encephalitis or post-infectious cerebellitis [Arvin, 1996]. However, it should be noted that most patients suffering from complications were previously healthy individuals. Moreover, among immunocompetent patients there are not identifiable risk factors for complications [Boelle and Hanslik, 2002; Bonhoeffer et al. 2005; Galil et al. 2002a]. Moreover, varicella infection in immunocompromised patients and healthy adults is associated with increased morbidity and mortality [Arvin, 1996]. Congenital infection can result in fetal varicella syndrome in up to 2% of cases, if the mother develops varicella during weeks 8–20 of gestation. The associated pathology may involve the skin, limbs, central nervous system and eyes. If maternal varicella onset coincides with the perinatal period, neonatal varicella follows which is very severe since no maternal antibodies are produced to be transferred vertically and to protect the newborn. If the infection left untreated, the mortality rate can be high at up to 30% [Enders et al. 1994; Paryani and Arvin, 1986].
VZV infection leads to lifetime immunity against the virus. Both cellular and humoral immunity contribute for the protection against a second infection, but cell-mediated immunity plays the predominant role. Patients with impaired cellular immunity are at greater risk of getting sick with varicella and reactivating VZV as herpes zoster (HZ) [Gershon and Gershon, 2010].
Reactivation of latent VZV infection can cause HZ also known as shingles. HZ is a unilateral vesicular lesion with dermatomal distribution. The most common dermatomes affected are thoracic and lumbar. HZ is more common in adults than children and its clinical manifestations differ between these age groups. In children, unlike adults, local pain, hyperesthesia and pruritus are not common. The lesions are mild and they appear as erythematous maculopapular lesions that rapidly evolve into vesicles. These vesicles coalesce to form bullous formations. The disease lasts up to 15 days and it might take over a month for the skin to completely heal. Common complications of HZ are post-herpetic neuralgia, iridocyclitis, secondary glaucoma, meningoencephalitis and encephalitis [Arvin, 1996].
Aim of the review
The best way to prevent varicella infection and effectively decrease the disease and associated economic burden is to provide primary prevention. Following that principle, a live attenuated varicella vaccine has been developed. Universal vaccination was first introduced in the United States in 1995 [ACIP, 1996; Marin et al. 2007]. The same year, the World Health Organization (WHO) adopted the mass vaccination against varicella [WHO, 1998]. Up to now only few countries around the world have implemented universal vaccination namely Australia, Canada, Germany, Qatar, Republic of Korea, Saudi Arabia, Taiwan, Uruguay, Italy (Sicily only) and Spain (Madrid only) [Bonanni et al. 2009]. This is mainly attributable to the preference of many countries’ public health authorities to recommend targeted varicella vaccination of the high-risk population and/or different prioritization with the availability of few new vaccines over the past decade (human papillomavirus [HPV] vaccine, rotavirus vaccine, etc.). Furthermore, concerns have been raised including the fear of increase of HZ incidence, the fear of age shift of the disease towards older age groups, the possible difficulties in achieving high coverage rates and last but not least the economic burden of universal vaccination implementation [Bonanni et al. 2009].
The aim of this review is to discuss the successes achieved by universal varicella vaccination and to address challenges needed to be tackled in the future.
Varicella vaccine
The live attenuated varicella vaccine was first developed in Japan by Takahashi [Takahashi et al. 1974]. Initially, the vaccine was solely used to protect high-risk leukemic children [Gershon et al. 1984; Takahashi et al. 1985]. In 1989, the vaccine was first introduced to healthy children in Japan and Korea and in 1995 the US Food and Drug Administration (FDA) approved the vaccine for children aged at least 12 months with a negative varicella history [Hambleton and Gershon, 2005].
There are two available live attenuated virus vaccines, Varilrix (GlaxoSmithKline Biologicals S.A., Rixensart, Belgium) and Varivax (Sanofi Pasteur Limited, Lyon, France). Both vaccines contain the Japanese varicella viral strain, Oka, and are safe and highly immunogenic [Arvin, 1996; Marin et al. 2007].
The adverse effects of the vaccine are few and are benign in nature [Arvin, 1996; Plotkin et al. 1985]. About 20% of the vaccinated population may develop mild pain, redness or swelling in the injection site, while 3–5% of the vaccinated population may present a vesicular rash either at the site of injection or more generalized. Up to 15% of recipients may suffer from fever [Arvin, 1996; Marin et al. 2007; Plotkin et al. 1985]. Serious complications of the vaccine are rarely reported (2.6/100,000 doses) [Chaves et al. 2008; Wagenpfeil et al. 2004]. Transmission of Oka strain varicella from a vaccinated child is a very rare phenomenon, since only 3 cases have been reported after 16 million doses of the vaccine had been administered [Hambleton and Gershon, 2005].
The live attenuated varicella vaccine poses a threat to immunocompromised patients. However, the vaccine can be safely administered to leukemic children under remission and HIV-infected children without severe immunodeficiency [Gershon et al. 1984, 2009]. Rarely, severe generalized varicella infection due to the Oka strain post-immunization has been reported in children vaccinated before the diagnosis of severe immunodeficiency [Chaves et al. 2008; Levin et al. 2003; Levy et al. 2003].
The Oka vaccine strain is capable of causing latent infection and it has been clearly shown that, as wild virus, it remains in the dorsal root ganglia [Chen et al. 2003; Gershon and Gershon, 2010]. Importantly, however, the risk of developing zoster among the vaccinated population is significantly lower when compared with children post-natural varicella [Baxter et al. 2013; Civen et al. 2009; Tseng et al. 2009]. It has been hypothesized that this is due to the lower viral loads induced [Gershon and Gershon, 2010].
Data from pre-licensure clinical trials have indicated that 95% of children develop protective antibodies following one dose of varicella vaccine [Provost et al. 1991]. Based on these studies one dose of varicella vaccine was initially proposed for the vaccination of healthy children <12 years of age [CDC, 1995]. However, as described below, outbreaks of varicella in schools due to breakthrough infection among immunized children were described and post-licensure studies estimated an overall effectiveness of 85% [Vazquez et al. 2004]. In 2006, both the CDC and ACIP recommended the administration of two doses of varicella vaccine to healthy children [ACIP, 2008; Marin et al. 2007]. A second dose of varicella vaccine elicits a robust anamnestic response in individuals partially primed by a single dose, with increases in both humoral and cellular immunity [Nader et al. 1995; Ngai et al. 1996; Watson et al. 1995a]. The implementation of two-dose vaccination has significantly decreased the incidence of breakthrough disease [Marin et al. 2007]. In addition, varicella vaccination may be administered as a tetravalent vaccine in combination with measles–mumps–rubella (MMR) vaccine. The measles–mumps–rubella–varicella (MMRV) vaccine produced by Merk & Co., contains higher dose of the Oka strain compared with the monovalent vaccines and was licensed in the USA in 2005. However, in the USA, post-licensure data however in the USA, indicated that the rate of febrile seizures doubled 7–10 days post-administration of the quadrivalent vaccine to 12–23-month-old children [CDC, 2008]. It was calculated that post-MMRV, one additional febrile seizure may occur for every 2300 doses given than with monovalent vaccines [Klein et al. 2010]. In Europe, MMR-V vaccine produced by GSK was initially introduced in Germany in 2006 [Siedler and Arndt, 2010]. Most recently, in July 2012, STIKO, the German National Committee on Immunization Practices, also updated their recommendations, expressing the preference of separate vaccination for the first dose due to the slight increased risk for febrile seizures 5–12 days after application of the combined MMRV vaccine compared to the simultaneous vaccination with a varicella and MMR vaccine (see http://www.rki.de/EN/Content/Prevention/Vaccination/recommandations/STIKO_Recommendations_2012_en.pdf?__blob=publicationFile, accessed 20 July 2013). However, for the second dose, MMRV is preferred [Jacobsen et al. 2009].
Pre-vaccine epidemiology
Incidence
Chickenpox is a very common disease with estimated 60 million new cases annually worldwide [Plotkin et al. 1985]. In the USA, in the pre-vaccine era, 4 million cases occurred annually (15–16/100,000) with the birth rate adding up to the same number [Marin et al. 2007]. In Germany, with an annual birth cohort of 800,000, new varicella cases reached 760,000 cases/year [Wagenpfeil et al. 2004]. Chickenpox is a disease of childhood affecting mainly children up to 12 years of age. Peak incidence age used to be between 5 and 9 years old. However, over the past 50 years, the increased use of daycare centers and childcare facilities decreased the peak incidence age to 1–4 years. Moreover, an increase in varicella incidence among infants was observed [Bonhoeffer et al. 2005]. Seroprevalence data indicate that 95% of young adults 20–29 years old are immune [Aebi et al. 2001; Nardone et al. 2007].
Climate factors and varicella incidence
The epidemiology of varicella infection is remarkably different between temperate and tropical regions [Liyanage et al. 2007; Tseng et al. 2005]. It is well known that the disease incidence is significantly lower among children living in tropical areas [Pollock and Golding, 1993]. Therefore, while varicella is a disease of pre-school and school-age children during winter and spring in temperate regions, in the tropics it often affects older age groups all year around [Garnett et al. 1993; Lee, 1998; Lolekha et al. 2001]. In addition, in tropical regions the seroprevalence rate of VZV is higher in urban than rural populations while this has not been observed in areas with temperate climate [Lolekha et al. 2001; Mandal et al. 1998]. More recently, seroepidemiologic data from Europe has revealed that the seropositivity rate of VZV is lower in subjects living in the Southern Mediterranean countries possibly linked to their climate [Katsafadou et al. 2008]. Although, some have supported that humidity negatively affects VZV transmission [Garnett et al. 1993], other factors such as population density, nursery attendance and socioeconomic development also may influence the epidemiology of VZV [Yawn et al. 1997]. Although the association between climate and varicella epidemiology has been mainly studied in tropical regions, recently it was shown that even in temperate regions the recent climate changes due to the ozone phenomenon and following global warming may have an impact on varicella incidence. More specifically, in Greece over the last two decades (1982–2003) it was noted that the frequency of hospitalized varicella cases was positively associated with wind speed, inversely associated with air temperature but not correlated with humidity [Critselis et al. 2012a].
Hospitalization
Although varicella is considered a mild disease, it may cause severe complications, leading to hospitalization in about 2–6% of infected subjects [Bonanni et al. 2009]. However, it should be noted that most hospitalized individuals due to VZV infection have no history of underlying disease [Liese et al. 2008]. In the pre-vaccine era in USA, hospitalization rates due to chickenpox ranged between 2.3 and 6.3/100,000 population [Galil et al. 2002a]. Children younger than 5 years of age accounted for more than 40% of varicella hospitalizations while infants and adults aged >20 years had 6- and 13-fold higher risk, respectively, for hospitalization [Davis et al. 2004; Galil et al. 2002a]. Hospitalization incidence in Europe ranged from 1.3 to 4.5/100,000 varicella cases and was higher in children younger than 16 years old (12.9–28/100,000) [Bonanni et al. 2009]. The average duration of hospitalization ranges between 3 and 8 days.
Morbidity and mortality
The most common complications that contribute to varicella-associated mortality include pneumonia, central nervous system involvement (encephalitis, meningoencephalitis), secondary bacterial systematic infections (mainly streptococcal and staphylococcal), hemorrhagic conditions and cardiovascular involvement (myocarditis). In the USA, during a period of 25 years (1970–1994) the average number of annual deaths due to varicella was 105 (rate 0.04/100,000) [Meyer et al. 2000]. Increased mortality rates are observed in adults older than 20 years old with a 25-fold higher death risk compared with children 1–4 years old (case-fatality rate: 21.3 and 0.8/100,000 cases, respectively) [Choo et al. 1995]. In Europe, the overall annual death rate is below 0.05/100,000 cases but again adults have significant higher mortality when compared to children [Boelle and Hanslik, 2002; Bonanni et al. 2009]. In England and Wales the rate is 0.04–0.05/100,000 per year [Rawson et al. 2001], in France 0.04/100,000 per year [Deguen et al. 1998; Dubos et al. 2007], in Germany 0.08/100,000 annually [Ziebold et al. 2001].
VZV vaccine successes
Varicella vaccine effectiveness in the USA
In the USA, universal varicella vaccination was first implemented in 1995. Initially one dose of varicella vaccine was recommended for all toddlers 12–18 months of age, while catch-up vaccination for all susceptible children <12 years of age was also recommended [ACIP, 1996]. Vaccination coverage increased in children under 19–35 months old from <30% in 1997 to >80% in 2006 [Marin et al. 2007]. Importantly, studies indicated that the varicella vaccine coverage did not differ by ethnicity or race [Luman et al. 2006]. Although prospective data on varicella incidence was not available since varicella infection was not reportable, significant reduction of varicella incidence was noted in two regions with active surveillance. Compared with 1995 (pre-vaccination era) a decrease of 77% and 90% were noted by 2000 and 2005, respectively [Guris et al. 2008; Seward et al. 2002]. Although the largest decrease was observed among children 1–9 years of age, other age groups were also protected by herd immunity and varicella rates decreased in adults as well [Marin et al. 2008; Seward et al. 2002]. Importantly, a decrease by almost 90% in the incidence of varicella was noted in infants, a particularly vulnerable group ineligible for vaccination [Chaves et al. 2011]. Moreover, during the first decade after varicella implementation a significant decrease in varicella associated morbidity, hospitalization and mortality were noted [Davis et al. 2004; Marin et al. 2011; Nguyen et al. 2005; Shah et al. 2010; Zhou et al. 2005]. Ambulatory visits due to varicella significantly decreased by 66% from 106.6 to 36.4/100,000 population and as expected this was mostly evident in the age group mainly affected, i.e. young children 1–4 years old [Shah et al. 2010]. Moreover, hospitalization rates decreased significantly [Dubos et al. 2007; Reynolds et al. 2008; Zhou et al. 2005]. As hospitalizations rates decreased to a greater extent among children when compared with adults, consequently, an increase in the proportion of adult cases was observed [Zhou et al. 2005]. The overall medical expenditure associated with varicella infection was also significantly decreased [Marin et al. 2008].
Despite these successes, outbreaks of varicella continued to occur despite the high vaccination coverage achieved. This was mainly due to breakthrough disease defined as a usually mild form of varicella occurring in a child with a documented vaccination against varicella at least 42 days prior to the eruption of the characteristic rash [Chaves et al. 2007]. However, it soon became clear that children with breakthrough disease were contagious and could transmit the virus to susceptible peers. Outbreaks of breakthrough varicella among vaccinated children in daycare centers and schools were reported in several US states since the introduction of the one-dose universal vaccination program. One-dose varicella vaccine effectiveness during these outbreaks was as low as 44%, but ranged from 80% to 89% [Marin et al. 2008]. This was not significant different from that described in non-placebo-controlled pre-licensure trials while a review of post-licensure studies of vaccine effectiveness found a median vaccine effectiveness of 84.5% for preventing all disease [Seward et al. 2008]. Based on clinical presentation it is difficult to distinguish between primary and secondary vaccine failure. Primary vaccine failure represents failure to mount a protective immune response after vaccination, whereas secondary vaccine failure is the gradual waning of immunity over time [Bonanni et al. 2013]. There have been many studies trying to distinguish whether breakthrough disease is due to primary or secondary vaccine failure as described more in detail below. In addition, several studies tried to identify risk factors associated with varicella vaccine effectiveness during outbreaks [Marin et al. 2008]. It has been supported that vaccination at a younger age, time since vaccination, vaccination within 28 days post-MMR and a history of asthma, eczema or steroid intake may be associated with varicella vaccine failure [Bayer et al. 2007; Black et al. 2008; Galil et al. 2002b; Marin et al. 2008; Silber et al. 2007; Verstraeten et al. 2003]. Moreover, commercially available most serologic tests lack sufficient sensitivity and specificity while among tests currently in use, only FAMA, a labor intensive test that cannot be automated, has been shown to correlate well with protective antibodies from both natural and vaccine-induced immunity [Breuer et al. 2008]. Over the first decade of varicella vaccine implementation, data from active surveillance revealed that the proportion of cases that occurred in vaccinated children significantly increased reaching 60% of cases in 2004. In addition, the age distribution of varicella cases was shifted to older children, since the peak age moved from 3–6 years old to older age groups. More specifically, it was noted that the peak age was between 6–9 years and 9–12 years for vaccinated and unvaccinated children, respectively [Chaves et al. 2007].
Therefore a two-dose childhood varicella vaccination schedule was recommended in 2006 by the Advisory Committee on Immunization Practices and the American Academy of Pediatrics [AAP, 2007; Marin et al. 2007]. The recommendations included a first dose between 12 and 15 months of age and a second dose to be administered between 4 and 6 years of age. Moreover, a second dose was recommended to be given as a catch-up dose to all those (children, adolescents, and adults) who were previously vaccinated with a single dose. The recommendation of a second dose was supported by previous data indicating that the second dose of varicella vaccine produces an improved humoral and cellular immune response that correlates with improved protection against disease [Kuter et al. 2004; Watson et al. 1995a, 1995b]. More importantly, a post-licensure clinical trial estimated that the risk for breakthrough disease was 3.3-fold lower among those children receiving two doses of varicella vaccine [Kuter et al. 2004]. Finally, most recently a case control study estimated the two doses vaccine effectiveness as 98.3% while the matched odds ratio for two doses versus one dose of the vaccine was 0.053 [Shapiro et al. 2011].
Varicella vaccine effectiveness in areas outside USA
In 1998, the WHO recommended that routine childhood varicella vaccination should be considered in countries where the disease is considered an important public health and socioeconomic problem. The WHO stressed that in areas where vaccination is implemented it will be necessary to achieve and maintain a high vaccine coverage in order to avoid disease shift to older age groups [WHO, 1998]. Universal varicella vaccination has been implemented by a few more countries since then including Australia, Canada, Germany, Greece, Qatar, Republic of Korea, Saudi Arabia, Taiwan, Uruguay and parts of Italy (Sicily) and Spain (Autonomous Community of Madrid) [Bonanni et al. 2009]. To date there has been evidence showing a significant decline in the varicella incidence from a few of those countries. In Australia, where the vaccine has been available since 1999 but only recently became publically funded (2005) data has indicated a 7% decline in varicella associated hospitalization between 2000 and 2007, which was even higher among children <5 years of age [Carville et al. 2010]. Active surveillance data from Canada has shown that hospitalization rates have decreased by 80–88% depending on the length of the universal vaccination program [Tan et al. 2012]. In Germany varicella vaccination of all toddlers was implemented in 2004 while active surveillance was started in 2005. Sentinel data from April 2005 to March 2009 showed a reduction of 55% of varicella cases in all ages, 63% in the age group 0–4 years and 38% in 5–9 year olds [Siedler and Arndt, 2010]. Moreover, using a multivariate time-series regression model, the effectiveness of one-dose vaccine administered during the second year of life was estimated as 83.2% (95% CI 80.2–85.7) [Hohle et al. 2011]. In Greece, data from a tertiary pediatric hospital in Metropolitan Athens indicated that the incidence of varicella associated ambulatory visits significantly decreased by 73.6% during the post-licensure period [Critselis et al. 2012b]. Notably, however, the incidence remained unchanged in children >10 years of age as well as in non-Greek and Roma children. These observations require attention to ensure that high vaccination coverage is achieved in susceptible adolescents and underserved children. On the other hand, in Sicily, a two-cohort universal vaccination program was implemented aiming to obtain high vaccine coverage in toddlers in their second year of life as well as susceptible adolescents [Giammanco et al. 2009]. Moreover, a surveillance study involving family pediatricians showed a significant decline in incidence rates from 2004 to 2009 involving all age groups. Finally, data from Uruguay have shown that 6 years after the introduction of varicella vaccine in routine immunization at the age of 12 months, hospitalization rates and ambulatory visits were decreased by 81% and 87%, respectively [Quian et al. 2008].
Challenges in varicella vaccine
Universal VZV vaccination has been delayed in many countries throughout the world. This is due to several reasons, including fear for an upward shift in the peak of age of the disease and fear of an increase in HZ incidence due to elimination of the circulation of the virus. Only few countries have mandatory reporting of varicella cases making comparisons between studies difficult, due to differences in surveillance practices and nations’ policies. Moreover as a result, cost-effectiveness studies evaluating universal childhood varicella vaccination programs are hard to interpret. Major concerns and challenges regarding varicella vaccination are further analyzed in the following paragraphs.
Fear of age shift of varicella cases to older age groups
Recent data from countries where varicella vaccination has been implemented indicate that the burden of varicella (varicella-related morbidity and mortality) has decreased substantially after the introduction of the vaccination programs almost in all age groups. As the highest decline of the incidence is recorded mostly in children <10 years of age, adolescents and young adults may be more susceptible to the disease while the possibility of outbreaks rises [Marin et al. 2008]. Over the past few years, data from the USA suggest an upward shift in the age distribution of varicella. Data from the 10-year varicella surveillance in Antelope Valley indicate that a shift in peak varicella incidence occurred over time from 3–6 year olds (in 1995) to 9–11 year olds (in 2004) [Chaves et al. 2007; Guris et al. 2008; Miller et al. 1993; Seward et al. 2002; Sloan and Burlison, 1992]. Consequently, the proportion of complications may increases as clinical presentation is more severe in this age group. However, it should be noted and stressed that to date the overall rates of adolescents and adults are declining [Seward et al. 2002; Chaves et al. 2007]. Surveillance data from Germany reveal reduced VZV incidence in all age groups [Siedler and Arndt, 2010]. Post-vaccination data from Uruguay reveal no change in age epidemiology [Quian et al. 2008]. Data from Slovenia, where vaccination coverage is low, have shown a downward shift in the age of varicella infection towards pre-school age with highest incidence at 3 years old [Socan and Blasko, 2007].
Data from Mediterranean countries reveal a similar picture of epidemiological alterations [Bonanni et al. 2009; Pinot de Moira and Nardone, 2005]. More specifically, data from Greece indicate high susceptibility in adolescence which may lead to an upwards shift in the future [Katsafadou et al. 2009]. Epidemiological data from Spain are confusing due to the high number of immigrants in the population, revealing a group with high susceptibility in childhood and another with greater susceptibility in adulthood [Valerio et al. 2009]. Finally, results after 5 years of universal vaccination in Sicily, one of the three Italian region which have introduced vaccination programs, do not suggest any shift in the epidemiology of varicella [Giammanco et al. 2009].
It is generally believed that adolescents and adults may be protected by decreasing circulating VZV in early childhood [Giammanco et al. 2009]. In countries, where less than 90% of children are covered by universal vaccination, VZV infection is not completely eradicated and therefore peak incidence age may increase from childhood to young adults which may suffer from serious complications [Doerr, 2013]. This expected shift is not a cause for concern, as long as the overall rates in adults are declining [Marin et al. 2008; Seward et al. 2002]. In countries that introduce universal varicella vaccination, continuous surveillance will be necessary to assess the need for catch-up vaccination in specific populations at risk.
Fear of HZ incidence increase among naturally infected individuals
HZ occurs in 10–20% of immune subjects and mainly affects older age groups, with the majority of cases involving individuals >50 years of age [Gnann and Whitley, 2002]. The major risk factor for HZ is the reduced T-cell-mediated immunity to VZV. Varicella infection as well as the administration of a varicella vaccine generates VZV-specific humoral and cellular immunity which protects from following re-exposure to VZV [Weinberg et al. 2010]. When these responses decline, as in the elderly and immunosuppressed individuals, reactivation of VZV by means of HZ is possible. These essential immune responses are boosted by the VZV vaccine developed to prevent HZ [Weinberg and Levin, 2010].
Re-exposure to VZV through contact with an infected person possibly protects latently infected individuals against HZ [Brisson and Edmunds, 2002; Jumaan et al. 2005]. Therefore, in areas with universal varicella vaccination, there is a concern that HZ incidence will increase among those with pre-existing natural immunity since their opportunity for re-exposure will significantly diminish [Brisson and Edmunds, 2002]. However, a rise in HZ incidence has been observed in countries in the absence of a varicella vaccination program (Canada and the United Kingdom), while in the USA, the age-specific HZ incidence has remained stable despite vaccine-associated decrease of varicella and does not differ between states with high and low varicella coverage [Brisson et al. 2001; Jumaan et al. 2005; Leung et al. 2011; Reynolds et al. 2008]. In addition, in two recent retrospective case-control studies, HZ incidence was not associated with varicella contacts [Rimland and Moanna, 2010; Yih et al. 2005].
Recently, Brisson and colleagues proposed a mathematical model in order to predict the impact of a two-dose varicella vaccine program on the incidence of HZ. The model showed that a two-dose of varicella live attenuated vaccine program will initially lead to an increase (~30%) of the incidence of HZ. Nevertheless, a reduction will follow as the proportion of individuals with a history of VZV infection will be limited [Brisson et al. 2010].
However, over the past decade it has become evident that although some studies have indicated a slight increase of HZ incidence in areas where universal varicella vaccination has been implemented, this issue still remains controversial. In addition, as mentioned above, it has been shown that HZ incidence has increased in many areas in the absence of universal varicella vaccination. The main conclusions from studies examining HZ incidence in countries where universal vaccination has been implemented are summarized in Table 1.
HZ incidence changes in countries where universal vaccination has been implemented.
HZ, herpes zoster, URV, universal vaccination.
It has become evident that no systematic significant incidence of HZ has been observed. Exposure to wild VZV by travelling to areas where varicella vaccine has not been implemented as well as importation through immigration may play a minor role.
Importantly, however, it has now become evident that patients with HZ are an important source of transmission of wild varicella virus in the community [Bloch and Johnson, 2012]. It is well known that HZ patients, even those with localized lesions in covered body areas, may transmit VZV to their contacts, either through direct contact or via aerosolized virus [Josephson and Gombert, 1988; Lopez et al. 2008; Schmid and Jumaan, 2010; Suzuki et al. 2004]. A recent retrospective analysis of data collected from active varicella and HZ surveillance reported from schools and daycare facilities in Philadelphia over a 7-year period clearly indicated that patients with HZ transmit VZV to their contacts, contributing significantly to varicella morbidity [Viner et al. 2012]. Interestingly, it was documented that 15% of varicella and 9% of HZ cases transmitted VZV infection to a susceptible individual in their environment resulting in varicella cases. The severity of these varicella cases was similar for those exposed to HZ and those exposed to varicella. Furthermore, recent studies have clearly indicated that most patients with HZ have viremia which may persist for months, while VZV DNA viral load is associated with longer duration of symptoms and risk factors for PHN [Quinlivan et al. 2007, 2011]. Moreover, VZV DNA has been detected in the saliva of 100%, 52% and 59% of patients with acute HZ, Ramsey Hunt syndrome and zoster sine herpete, respectively [Furuta et al. 2001; Nagel et al. 2011]. Finally, prospective molecular epidemiologic data from UK, using genotyping methods on VZV DNA isolated from saliva in children with varicella, provided further evidence supporting the hypothesis that HZ patients are an important source of VZV transmission in the community. Clade 5 viruses (non-European virus) were significantly more prevalent among nonwhite children than white children with varicella possibly reflecting population mixing patterns, with susceptible children more likely to be exposed in their community, household, or daycare to others of the same ethnicity and their indigenous viruses [Quinlivan et al. 2013].
In addition, the role of endogenous boosting might have been underestimated. Although, subclinical reactivation of VZV has been postulated since Hope–Simpsons hypothesis, recent data has indeed revealed that this phenomenon is common and potentially provides significant protection against HZ. Detection of VZV DNA in blood has been well described. Cell-associated and whole-blood VZV subclinical viremia have been detected in 9% of healthy blood donors, 17% of immunocompromised patients and 27% of healthy elderly subjects [Devlin et al. 1992; Kimura et al. 2000; Ljungman et al. 1986; Schunemann et al. 1998; Wilson et al. 1992]. Moreover, salivary shedding of VZV DNA has been detected in a third of healthy astronauts during and after space flights supporting the hypothesis that asymptomatic VZV reactivation occurs more often than considered previously, and that it may occur in response to stress [Mehta et al. 2004]. More importantly, viable virus has been detected indicating that subclinical reactivation not only provides boosting of cell-mediated immunity in the individual, but also that viral transmission to their contacts may follow [Cohrs et al. 2008]. Most recently, in a small prospective study, VZV reactivation was observed in 17% of children hospitalized in ICU. Among those severely stressed children, only those post natural varicella infection reactivated, while none of the varicella-vaccinated children or healthy controls had detectable VZV DNA in any blood or saliva samples examined [Papaevangelou et al. 2013].
In conclusion, these two phenomena, namely varicella virus transmission from HZ cases and endogenous boosting, had not been included in the mathematic models exploring HZ incidence in the era of universal vaccination and might explain the absence of the predicted (up to 30%) HZ increase [Brisson et al. 2000, 2010].
Cost-effectiveness studies evaluating the introduction of varicella vaccine
Few economic analyses conducted in the USA, Australia and Europe have shown varicella vaccination programs to be cost-effective from a healthcare perspective and even cost saving from a societal perspective [Beutels et al. 1996; Bonanni et al. 2008; Coudeville et al. 1999; Diez Domingo et al. 1999; Scuffham et al. 1999; Thiry et al. 2004]. Most of these studies, however, performed the economic evaluation using a one-dose varicella vaccine immunization scheme without taking into account breakthrough varicella infection or potential increase of zoster incidence [Rozenbaum et al. 2008]. It is clear that if the doubling of cost by the two-dose schedule is not compensated for by a reduction in the cost of vaccine failures, it will negatively impact the vaccination program’s cost-effectiveness. Indeed, Zhou and colleagues have shown that the two-dose regimen is cost-effective when compared with no vaccination but not when compared with the one-dose regimen [Zhou et al. 2008]. Moreover, it is important to understand that there is a considerable heterogeneity in the methodologies used in different studies which have an impact on their results [Soarez et al. 2009]. More specifically, the type of model used (static versus dynamic), the perspective examined (whether society’s perspective is added to that of providers), the inclusion of the herd immunity effect and vaccination coverage and the different amounts used for vaccine price and parental wages also had a significant impact on the computed results. Finally, other than the perspective adopted, other factors such as public opinion and public health priorities may also have an influence on the interpretation of such studies.
Universal versus high-risk population vaccination
As mentioned previously, many European public health authorities have elected to implement targeted varicella vaccination of high-risk populations, mainly focusing on susceptible adolescents and household contacts of immunocompromised patients [Bonanni et al. 2009]. This strategy prevents varicella in older age groups where morbidity and mortality are significantly higher and in childbearing women, decreasing the risk of congenital and perinatal varicella [Sengupta et al. 2008]. Furthermore, minimizes alleged risk of future peak age incidence shift to older age and addresses fear of HZ incidence increase. However, targeted population vaccination has failed multiple times in the recent past, as for example in HBV and rubella vaccination [Mongua-Rodriguez et al. 2013; Van Damme et al. 1997]. This is mainly due to difficulties in reaching populations at risk. Specifically, when discussing varicella vaccination one has to face the challenge of reaching adolescents since it has been well documented that this is a group with poor preventive care [McCauley et al. 2008; Wong et al. 2013]. Therefore, this strategy remains an option solely in countries with very well-organized catch-up vaccination programs or school vaccination. This is particularly problematic in areas such as Southern Europe where a higher percentage of adolescents are susceptible [Gabutti et al. 2001; Katsafadou et al. 2008]. Moreover, this strategy may not be as effective in preventing infection in infants, a particularly vulnerable group [Boelle and Hanslik, 2002].
Dose interval in universal varicella vaccination
In most countries recommending varicella vaccination, the first dose (either using the monovalent vaccine or MMRV) is administered at 12–15 months of age [Bonanni et al. 2009]. The second dose, however, may be administered either during the second year of life, at least 1 month after the first dose (Germany) or at the age of 4–6 years of age (USA and most other countries). Both schedules have advantages and disadvantages. The main advantage of administering both doses during the second year of life is the increased vaccine coverage and the reduced risk of breakthrough disease [Bonanni et al. 2009]. This is more apparent if one postulates that breakthrough disease is rather due to primary vaccine failure (= no take) than secondary vaccine failure (= waning of immunity). This question was addressed in a recently published review by Bonanni and colleagues [Bonanni et al. 2013]. A relatively high rate of primary vaccine failure amongst recipients of one-dose varicella vaccine and limited convincing evidence of secondary vaccine failure was concluded. This supports the administration of a second dose of varicella vaccine with a short interval aiming to reduce the number of subjects remaining susceptible to varicella. This is most relevant in countries that plan to or have recently introduced varicella vaccination where wild virus is still prevalent and therefore preventing breakthrough disease is especially important. Prospective data collected from Northern California (Kaiser Permanente) clearly indicates that breakthrough varicella rates were higher over the first 4 years post-varicella implementation and were reduced thereafter even before the introduction of the second vaccine dose [Baxter et al. 2013].
Conversely, when the period between the two doses increases to 2–6 years, breakthrough cases of varicella can be expected, especially during the first years post-universal vaccination implementation. This could not only result in outbreaks in daycare centers and schools but also in transmission to susceptible adolescents and adults. Furthermore, once breakthrough disease occurs the subject might lose the advantage of low HZ risk later in life. However, consideration of potential disturbance of vaccine and associated routine doctors’ visits schedule is very important when planning the introduction of new vaccines in immunization schedules.
Conclusion
Varicella-associated morbidity and mortality can be effectively prevented by the implementation of universal varicella vaccination. Targeted vaccination of susceptible high-risk patients and/or their household contacts, recommended in many European countries, is an alternative strategy which necessitates a well-organized healthcare system. Although a slight increase in the peak incidence age of varicella has been observed post-universal varicella implementation, the incidence rates among adolescents and adults have decreased when compared with the pre-vaccination era. To avoid an age shift to older age groups it is necessary to ensure catch-up vaccination and maintain high vaccine coverage. The predicted increase of HZ incidence among naturally immune adults following varicella vaccination has not been observed, possibly because travelling, contact with HZ patients and subclinical reactivation offer greater boosting effects on the cell-mediated immunity than expected.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflicts of interest
The authors declare no conflicts of interest in preparing this article.