Varicella vaccination - the global experience

ABSTRACT

Introduction: Varicella, although a frequently benign childhood disease, nevertheless represents a considerable health burden. WHO recommends including varicella vaccines in universal routine vaccination programs, and maintaining coverage >80%. Many countries have successfully introduced varicella vaccination and have benefited from lower disease burden, but many others have not adopted the vaccine. Reasons include cost commitment for a ‘mild childhood disease’ or concerns that vaccination will shift varicella to older age groups or increase herpes zoster incidence.

Areas covered: This literature review summarizes the effectiveness and epidemiological impact of varicella immunization programs.

Expert commentary: Varicella vaccines are immunogenic with acceptable safety profiles. One and two dose schedules are highly effective against varicella and large reductions in disease incidence, particularly moderate-severe disease, have been widely reported. There is currently no evidence to suggest that the introduction of varicella vaccination results in a shift of varicella disease burden to older age groups. Although epidemiological studies have shown an increased incidence of herpes zoster since the vaccines were launched, there are many other contributing factors, and indeed, this secular trend was evident before their introduction. In conclusion, varicella vaccination easily fits into existing immunization programs and significantly reduces the often underestimated burden of varicella.

KEYWORDS:

• Varicella
• herpes zoster
• vaccines
• effectiveness
• immunization

1. Introduction

Varicella-zoster virus (VZV) causes both varicella (also known as chickenpox) and herpes zoster (HZ, also known as shingles). Varicella is a common childhood disease, which usually confers lifetime immunity, whereas HZ arises when dormant VZV in the nerve ganglia reactivates in previously infected individuals [1]. HZ usually occurs later in life, with ≥95% immunocompetent individuals over 50 years of age being seropositive for VZV and at risk of developing HZ [2]. The lifetime risk for HZ is around 32% [1].

Varicella is not a universally notifiable disease, but standardized annual incidence rates from 300–1291 per 100,000 population have been reported in Europe [3]. Although varicella is mainly benign in children, serious complications can develop [4], and in 2014, the World Health Organization (WHO) estimated that approximately 4.2 million severe complications leading to hospitalization and 4200 related deaths occur globally each year [5]. In the pre-vaccination era, approximately 30.9 per 100,000 varicella cases were hospitalized in the United States [6], and 0.41 cases per million population were fatal [7]. However, post-licensure, varicella-related hospitalizations decreased to 14.5 per 100,000 cases [6], and deaths to 0.05 per million population [7]. Varicella is usually more severe in adults, with those aged ≥45 years having 4–50 times greater risk of hospitalization and 174-fold higher risk of dying than individuals aged 5–14 years [8]. Nevertheless, it should be noted that the burden of hospitalization is highest in immunocompetent or previously healthy individuals [9,10].

The main burden of varicella disease is economic due to the high number of cases and the need for parents and caregivers to look after their children. Noncomplicated cases tend to last for up to 2 weeks [11], during which time affected children will not be able to attend day care or school. It has recently been reported that in Sweden, one in four parents needs to take time off from work to care for children with varicella [12]. The indirect costs associated with parents taking time off from work make a significant but potentially underestimated contribution to the economic impact of VZV infection [13–16].

Varicella vaccines are highly effective in reducing the global incidence and burden of the disease [17]. The vaccine, as a frozen formulation, was licensed for use in 1984 and was the first commercially available varicella vaccine. It subsequently became the first refrigerator-stable varicella vaccine, its development commenced in 1991, and it has been licensed for use since 1994 [18]. Although not universally adopted, WHO recommends that in countries where varicella is an important public health burden, varicella vaccination should be introduced into their routine immunization programs [17].

Varicella dosing recommendations can include one or two doses, separated by a long or short dosing interval. The most common schedule comprises a first dose at 12–18 months followed, if adopted, by a second dose at between 4 and 6 years of age. Alternatively, the second dose can be administered in children below 4 years of age, provided that 3 or more months have elapsed since the first dose [19,20]. Although it is assumed that a shorter interval may be optimal in terms of epidemiologic impact, pragmatically some countries retain longer intervals to better fit with their childhood vaccination programs [19].

A single-dose schedule is effective at controlling severe disease, but varicella breakthrough still occurs [21–24]. Alternatively, the addition of a second dose provides protection against all severities [25]. The economics of implementing one- or two-dose schedules have been widely debated [20,26,27], but national choices will ultimately depend on whether their priorities are varicella elimination or prevention of severe disease. Interestingly, a recent modeling study from Italy demonstrated that out of coverage, efficacy, number of doses, or dosing interval, high coverage is the critical success factor [28]. Indeed, WHO has already recommended that vaccine coverage should be maintained above 80% [17].

The varicella vaccine can be administered as a monovalent vaccine (e.g. Varilrix; GSK, Belgium or Varivax; Merck & Co. Inc., USA) [29,30] or combined with the measles, mumps, and rubella vaccine as a quadrivalent vaccine (MMRV; e.g. Priorix-Tetra; GSK, Belgium or ProQuad; Merck & Co. Inc., USA) [31,32]. The immunogenicity and safety of both the varicella monovalent vaccine and MMRV are well established and have been extensively reviewed [18,33]. As there is no accepted correlate of immunity for varicella, efficacy data provide more clinically relevant information than immunologic data and the efficacy of varicella-containing vaccines in preventing VZV infections has been widely studied and reviewed [34]. High levels of long-term protection have been observed after both single and two-dose schedules [35,36].

In December 2014, varicella vaccines were recommended in 33 predominantly higher socioeconomic status countries (Figure 1), implying that despite established effectiveness, many countries still do not routinely vaccinate children against VZV. Reasons for low adoption could include cost of implementation for a ‘mild childhood disease,’ or fears that vaccination may shift the disease to older individuals in whom the disease is more severe or may increase the incidence of HZ [3]. We therefore undertook this review to ascertain the effectiveness of the varicella vaccine and its impact upon disease-associated morbidity and mortality, as well as determining whether there is any published evidence to support either an age-shift in varicella incidence or an increased incidence of HZ.

Figure 1. World map representing different national universal routine vaccination (URV) schedules against varicella (national-level guidelines are represented, unless specific region data was publically available). *In Cyprus, varicella vaccination is administered universally in the private sector. †Varicella URV is recommended in Hong Kong, but not yet implemented [7,37,42,70,72,115,116].

2. Evidence for varicella vaccination

2.1. Effectiveness

Vaccine effectiveness (VE), defined as the measure of protection attributable to a vaccine administered under field conditions to a given population [43], provides an estimate of the effect of vaccines in real-world settings. The effectiveness of the varicella vaccine has been assessed in outbreak, case-control and longitudinal, database, observational, and modeling studies, of which outbreak studies are the most numerous.

Table 1 presents the results from individual studies showing the VE of the varicella vaccine. As can be seen, VE is influenced by a number of factors including the number of administered doses, disease severity, and age at which the vaccine is administered [25,44]. VE for one dose of varicella vaccine against any disease ranged from 55% to 87%, while after two doses, the VE ranged between 84% and 98% [23,44–46]. VE was higher against moderate or severe disease, ranging from 70% to 98% after one dose and 94% to 98% after two doses [23,47]. Two studies from Israel [48] and Korea [49] recorded much lower VE – in case of Israel, this was most probably due to the very low coverage (37%) [48] and in the case of Korea was due to the ineffectiveness of one of the administered vaccines [49]. With these exceptions, the ranges shown are consistent with previous reviews [50,51]. Additionally, in the recent meta-analysis from Marin et al., VE against all varicella was estimated as 81% after one and 92% after two vaccine doses and as 98% after one dose for moderate/severe varicella [25]. Further, the incremental VE of two doses over one has recently been calculated as 84.6% [52].

Table 1. Effectiveness of varicella vaccines.

Outbreak investigations
						VE (%) (CI)
Country [Reference]	Setting	Vaccine	N vaccinated cases	Vaccinated population	Vaccine doses	Any disease	Moderate/severe disease	Vaccine coverage (%)
Germany [23]	Prospective, 7 day-care centers	Varilrix	19	77	1	56 (29–72)	86 (62–95)	62 (49–81)
		Priorix-Tetra	7	38	1	55 (8–78)	70 (4–90)
			2	56	2	91 (65–98)	94 (54–99)
		Varivax	4	48	1	86 (56–96)	96 (67–99)
Italy [117]	Day-care center/elementary school	NR	17	210 (Day care)	1	60 (48–70)	–	86.6
				358 (School)	1	69 (51–88)	–	51.9
Italy [24]	Preschool	NR	7	55	1	82.4	–	53.9
USA [45]	Prospective elementary and affected schools	NR	16	109	1	81 (67–89)	–	32.7
			9	220	2	95 (89–97)	–	66.1
			13	85	1	80 (64–89)	–	31.4
			22	181	2	84 (74–90)	–	66.8
Taiwan [118]	Elementary school	NR	10	321	NR	69–100	86	71.2
Israel [48]	Day-care centers	NR	37	89	NR	20 (0–40)	93 (75–98)	37
Case-control studies
						VE (%) (CI)
Country [Reference]	Setting	Vaccine	N cases	N matched controls	Vaccine doses	Any disease	Moderate/Severe disease	Vaccine coverage (%)
Germany [47]	Pediatric practices	Any	432	432	1	86 (77–92)	98 (91–99)	13.2 (Cases); 45.1 (controls)
		OKA/GSK	35	63	1	72 (49–84)	95 (78–99)
China [53]	General community	Any	1000	1000	1	84 (77–89)	–	19
		Varilrix			1	86 (73–93)	–
China [54]	Schools and day-care centers	Any	180	679	1	83 (71–90)	–	10.0 (Cases); 34.5 (controls)
USA: New Haven, Connecticut [119]	Pediatric practices	Any	202	389	1	85 (78–90)	97 (93–99)	23 (Cases); 61 (controls)
USA: New Haven, Connecticut [44] Note that this study expands upon results from the same source as Ref. [119] above	Pediatric practices	Any	339	669	1	87 (81–91)	–	36 (Cases); 70 (controls)
USA: New Haven, Connecticut [46]	Pediatric practices	Any	66	117	1	86 (−45–99)	–	93.0 (Cases); 83.6 (controls)
Note that this study expands upon results from Refs. [119] and [44] above			0	22	2	98 (84–100)	–	0 (Cases) 15.7 (controls)
USA: Antelope Valley CA, Philadelphia [120]	Immunization registries	Any	32 (1–3 years) 93 (≥4 years)	103 (1–3 years) 305 (≥4 years)	1	76 (39–90)	78 (13–95)
					2	94 (76–98)	98 (83–100)	–
Korea [49]	National Notifiable Disease Surveillance	1 MAV 3 OKA	537	537	1	13 (−17–36)	–	–
Longitudinal, database, observational, and modeling studies
						VE (%) (CI)
Country [Reference]	Setting	Vaccine	N cases	Sample size	Vaccine doses	Any disease	Moderate/Severe disease	Vaccine coverage (%)
Turkey [21]	Well-child clinic and private pediatrician	Various	466	1683	1	62 (57–66)	–	60.1
		No vaccine	723	1119	0		–
Germany [121]	Primary care modeling	Various	679	NR	1	83 (80–86)	–	78
Germany [52]	Linear modeling based on Sentinel data	Various	8153	31,288	1	87 (85–88)	–	28.8
					2	97 (97–98)		58.8

CI: confidence intervals; N: number; NR: not reported; VE: vaccine effectiveness.

Several studies have been undertaken to assess whether VE wanes over time [22,54,55]. In the longest study to date, VE showed no reduction against any severity varicella disease for up to 14 years [55]. These findings were reinforced by other works [22,54], and although a generally accepted correlate of protection has not yet been identified for varicella, these studies suggest that VE mirrors antibody persistence [53–55].

Despite the effectiveness of the vaccine, low-level breakthrough varicella does however occur. Individual studies have reported incidences ranging from 8% to 32% after single-dose varicella vaccine [21–24], and 4% after two doses [23].

2.2. Impact of varicella vaccination on varicella incidence, morbidity, and mortality

The impact of vaccination is expressed as the proportionate reduction in disease burden, comparing incidences and mortality rates in the same population between the pre-vaccine era and after vaccine implementation [43]. Table 2 shows the difference in varicella incidence rates occurring before vaccination was implemented to that after one and two-dose schedules were introduced. All studies have shown impressive reductions in disease incidence compared with the pre-vaccination era. After one-dose programs, reductions up to 74% have been recorded [56]. Whereas reductions exceeding 90% have been recorded after two-dose schedules [37,57,58].

Table 2. Decrease in incidence rates following the implementation of varicella vaccination.

Country [Reference]	Data source	Year introduced	Doses compared	Years compared	Change
Germany [59]	Sentinel data	2004 (Dose 1) 2009 (Dose 2)	0 vs. 1	2009 vs. 2005	↓55%
Munich, Germany [60]	Annual parent surveys. Monthly pediatric practice surveillance, pediatric hospital databases		0 vs. 1	2008–9 vs. 2006–7	4.0 vs. 6.6 cases/1000
			0 vs. 2	2010–11 vs. 2006–7	2.2 vs. 6.6 cases/1000
			1 vs. 2	2010–11 vs. 2008–9	2.2 vs. 4.0 cases/1000
Italy [61]	Regional surveillance	2005	0 vs. 1	2008 vs. 2000–7	↓51%
Italy; Sicily [57]	National Surveillance System	2003 (Dose 1) 2010 (Dose 2)	0 vs. 2	2012 vs. 2003	↓95%
Italy; Tuscany [38]	Regionally notified varicella cases	2008	0 vs. 1	2009–12 vs. 2004–7	0.89 vs. 2.3 cases/1000
Italy; Apulia, Sicily, Veneto [122]	Statutory notification, National Hospital Discharge, mortality data	2003 (Dose 1) 2010 (Dose 2)	0 vs. 2	2010 vs. 2001–9	103 vs. 151 cases/100,000
Spain; Navarre [37]	Regional surveillance data	2007	0 vs. 2	2012 vs. 2006	1.0 vs. 50 cases/1000 (↓98.1%)
USA; Connecticut [62]	Varicella surveillance data	1995 (Dose 1) 2006 (Dose 2)	1 vs. 2	2014 vs. 2009	5.1 vs. 13.8/100,000 (p < .001)
USA [58]	Kaiser group	1995 (Dose 1) 2006 (Dose 2)	0 vs. 2	2009 vs. 1995	1.3 vs. 25.8/1000 person-years (↓90–95%)
USA [123]	Population-based active surveillance	1995 (Dose 1) 2006 (Dose 2)	0 vs. 2	2013–14 vs. 2005–6	3.9 vs. 25.4/100,000 population (↓85%)
USA, Washington State [63]	Group Health Cooperative	1995 (Dose 1)	0 vs. 1	2002 vs. 1992	2.63 (1995); 2.29 (1998); 0.92 (2002) cases/1000
Taiwan [64]	National Health Insurance Claims Database	2004 (Mass vaccination)	0 vs. 1	2008 vs. 2000	2.23 vs. 8.28 cases/1000
Taiwan [65]	National Health Insurance Claims Database	2004 (Mass vaccination)	0 vs. 1	2004–8 vs. 2000–3	Non-specified significant reduction
Costa Rica [56]	National surveillance data	2007	0 vs. 1	2015 vs. 2008	↓73.8% (↓79% <5 years)

Many studies have also shown that vaccination is associated with a significant decrease in varicella-related hospitalization rates (Table 3; ranging from 23% to 93% over a 4–14-year time period) [6,37,60,61,64–69]. The highest reductions were observed in individuals below 15 years old [6,37,60,61,65–67] and specifically in the youngest children [38,62,69]. Some studies found a relatively small decrease in varicella hospitalization rates [6,60,61,66,67,70], possibly due to reduced vaccination coverage and shorter study periods. Recently, Mota et al., who studied VZV-related hospitalizations and mortality in Brazil from 1996 to 2011, showed that average annual mortality rates for varicella in Brazil before vaccine implementation were 0.88/100,000 in infants under 1 year and 0.40/100,000 in children aged 1–4 years [71].

Table 3. Decrease in hospitalization rates following the implementation of varicella vaccination.

Country [Reference]	Data source	Year introduced	Doses compared	Years compared	Change
Germany [69]	National hospital discharge data	2004 (1 dose) 2006 (2 doses)	0 vs. 2	2005–12 vs. 1995–2003	1.9 vs. 3.3/100,000
Germany; Munich [60]	Annual parent surveys. Monthly pediatric practice surveillance, pediatric hospital databases	2004	0 vs. 1	2009 vs. 2005	4.3 vs. 7.6/100,000 (↓43%)
Germany; Bavaria [124]	Pediatric hospital discharge data	2004	0 vs. 1	2011 vs. 2005	↓60%
Italy [61]	Regional surveillance	2005	0 vs. 1	2007–8 vs. 2000	↓53%
Italy; Tuscany [38]	Regionally notified varicella cases	2008	0 vs. 1	2009–12 vs. 2004–7	2.2 vs. 4.1/100,000
Italy; Sicily [57]	National Surveillance System	2003	0 vs. 2	2012 vs. 2003	0.8 vs. 4.8/100,000
Italy [72]	Aggregate regional data	2003–13	0 vs. 2	2012 vs. 2004	↓75%
Spain; Navarre [37]	Regional surveillance data	2007	0 vs. 2	2012 vs. 2006	↓89%
Spain [70]	Hospital discharge data	2006 (2 doses)	0 vs. 2	2010 vs. 2005	37 vs. 47/100,000 (Children < 5 years)
Israel; Tel-Aviv [125]	Retrospective hospital chart review	2008	0 vs. 2	2009–12 vs. 2004–8	↓63% (↓75% in 1–6 years)
Canada [73]	Hospital discharge database	2000–6 (1 dose) 2010 (2 doses)	0 vs. 1	2010 vs. 1990	>70%↓ (All ages less than 40)
Australia; Victoria [66]	Medical consulting data	2005	0 vs. 1	2007 vs. 2000	↓23% (Annual 7% reduction)
Australia [74]	National hospital morbidity database	2005	0 vs. 1	2006–10 vs. 1998–9	↓52.7%
USA [58]	Kaiser group	1995 (Dose 1) 2006 (Dose 2)	0 vs. 2	2009 vs. 1994	0.25 vs. 2.13/100,000 (↓90%)
USA [68]	Truven claims data	1995 (Dose 1) 2006 (Dose 2)	0 vs. 2	2012 vs. pre-vaccination	↓93% (↓38% vs. one dose period)
USA; Connecticut [126]	Hospital discharge data	1995 (Dose 1) 2006 (Dose 2)	0 vs. 1	2001–5 vs. 1994–5	↓83% (p < 0.001)
Taiwan [65]	National Health Insurance Claims Database	2004	0 vs. 1	2004–8 vs. 2000–3	Non-specified significant reduction
Costa Rica [56]	ational surveillance data	2007	0 vs. 1	2015 vs. 2008	↓85.9% (↓87% (<5 years))

In the United States, few years after the implementation of the varicella vaccination program, significant reductions in varicella-related deaths, compared with the 5 years preceding the vaccination program, were demonstrated (92% in children 1–4 years, and 74–89% in infants <1 year and persons 5–49 years) [75].

2.3. Shift in varicella to older age groups

Data from the United States suggested an upward shift in the age distribution of varicella, as a result of childhood vaccination programs [76]. For example, surveillance data from Antelope Valley indicated a shift in varicella incidence peaks, from 3 to 6-year olds (in 1995) to 9–11-year olds (in 2004) [77]. Mathematical models predicted that the age shift occurred if coverage rates fell below 80–85% [78]. Such observations have prompted WHO to recommend that coverage rates above 80% should be achieved and maintained [79]. However, the number of varicella cases and varicella-related hospitalizations in the whole population fall after vaccine introduction; there does not appear to be an age shift. Furthermore, recent surveillance data from different countries have shown a reduction in the VZV incidence in all age groups [80] or under the age of 40 [74], suggestive of a herd effect.

When introducing a vaccine for routine childhood vaccination, there may be immunity gaps in older individuals, necessitating a catchup program. Some countries, such as Australia, have therefore implemented varicella vaccination of older individuals to prevent any potential shift to older age groups, despite available evidence suggesting that varicella rates still decrease in unvaccinated groups [6,59,67,73,74,77]. A two-dose schedule is recommended for adolescents and adults, as clinical trials have indicated a low response rate after single-dose varicella vaccination in these age groups [81,82].

2.4. Varicella vaccination and the incidence of HZ

In 2000, a model by Brisson et al. theoretically linked the implementation of universal varicella vaccination in children to an increased incidence of HZ, in the short and medium term following vaccination, in older populations [83]. In the long term, however, a decreasing incidence of HZ is expected to occur, assuming that vaccinated individuals are less likely to develop zoster when compared to naturally infected individuals [73]. The theory behind this model is that exogenous boosting by VZV exposure is needed to maintain cell-mediated immunity above a threshold and reduce the risk of developing HZ [84–86]. Further models have calculated that a temporary increase in HZ incidence, as a result of varicella vaccination, could be anticipated over the next 50–70 years [87,88]. As a potential increase in HZ can have implications on acceptability to a population and also on cost calculations [89,90], such observations can cause a delay or rejection of the varicella vaccination into national programs.

Epidemiological studies on the long-term trend of HZ show that the incidence of HZ has increased more than 4 times over the last six decades among all age groups and both sexes [91–93]. Although some studies show an increase in the HZ incidence after the introduction of varicella immunization program [64,69,94–97], others have shown no increase [63] and there is no concrete evidence to attribute this trend directly to varicella vaccination (Table 4) [70,98–101]. Otherwise, there is some evidence that children vaccinated for varicella have lower risk of developing HZ than those with history of varicella [102,103]. A review by Ogunjimi et al. concluded that although exogenous boosting plays a role in HZ incidence, its magnitude has yet to be accurately determined [104]. For example, increased oral corticosteroid use [105], chronic comorbid conditions [106–110], stress [107,110], and an increasing elderly population [111,112] all have an impact on the incidence of HZ. In addition, endogenous boosting, i.e. the subclinical reactivation of the latent VZV due to internal factors, can also play a role in boosting the anti-varicella immune response and hence changes in the incidence of HZ [113]. More research into the pathophysiology of HZ is warranted, particularly with reference to the endogenous and exogenous boosting hypotheses.

Table 4. HZ incidence and varicella vaccination.

Country [Reference]	Study design	Sample	Age at vaccination	Year introduced	Years compared	Change in incidence
Spain [70]	Hospital discharge data	27,236 HZ discharges	Dose 1 at 15–18 months/susceptible adolescents	2006	2010 vs. 2005	10.9 vs. 9.7/100,000 (>84 years: 98 vs. 70) Hospitalization rates
Canada; Ontario [101]	Retrospective analysis	13.2 million	Dose 1 at 12–15 months Dose 2 at 4–6 years	2005^a	2005–9 vs. 1992–8	303 vs. 309/100,000
Australia; New-South Wales [99]	Hospitalizations, antiviral prescriptions; presentations		Dose 1 at 18 months Catch-up: 12–13 years	2005	2006–7 vs. 1998–9	Annual age-specific hospitalization rates: 3.7%↓ (20–39 years); ↑1.6% (40–59 years)
Australia [94]	Retrospective analysis of management encounters	1,078,671 Encounters	Dose 1 at 18 months Catchup: 10–13 years	2005^a	2000–9 vs. 1998–9	2.01 vs. 0.58/1000 (↑55%)
Australia [74]	National hospital morbidity database	300,000	Dose 1 at 18 months Catchup: 12–13 years	2005	1998–2010	10.4 vs. 9.9/100,000 Hospitalization rates
Canada; British Columbia [127]	Population-DataBC database	238,295 cases	Dose 1 at 12–15 months Dose 2 at 4–6 years	2000–6 (1 dose) 2010 (2 doses)	2012 vs. 1997	4.5 vs. 3.2/1000
USA [102]	Population-based active surveillance	350,000	Dose 1 at 12–15 months Dose 2 at 4–6 years	1995 (Dose 1) 2006 (Dose 2)	2006 vs. 2000	<10 years: 33 vs. 75/100,000 (55%↓) 10–19 years: 97 vs. 60 (63%↑)
USA [100]	Medical claims data	3.5 million (1993) 10.9 million (2006)	Dose 1 at 12–15 months Dose 2 at 4–6 years	1995 (Dose 1) 2006 (Dose 2)	2006 vs. 1993	4.4 vs. 1.7/1000
USA [95]	HZ hospitalization data		Dose 1 at 12–15 months Dose 2 at 4–6 years	1995 (Dose 1) 2006 (Dose 2)	2004 vs. pre-2002	2.5/10,000 vs. not specified (significant↑) Hospitalization rates
USA [98]	Population-based active surveillance		Dose 1 at 12–15 months Dose 2 at 4–6 years	1995 (Dose 1) 2006 (Dose 2)	2007–10 vs. 2000–6	12.8 vs. 42/100,000 (<10 years) 78 vs. 68 (10–19 years)
USA, Washington State [63]	Group Health Cooperative			1995 (Dose 1)	2000 vs. 1992	3.47 vs. 4.05/1000 person years
USA; Minnesota [91]	Population-based cohort study	8017 Cases	Dose 1 at 12–15 months Dose 2 at 4–6 years	1995 (Dose 1) 2006 (Dose 2)	2000–7 vs. 1945–9	3.15 vs. 0.76/1000 person-years
USA; Massachusetts [97]	Retrospective study		Dose 1 at 12–15 months Dose 2 at 4–6 years	1995 (Dose 1) 2006 (Dose 2)	2003 vs. 1999	5.25 vs. 2.77/1000 (90%↑)
Taiwan [64]	National Health Insurance Claims Database		Dose 1 > 12 months	2004 (Mass vaccination)	2008 vs. 2000	6.89 vs. 4.45/1000
Taiwan [96]	Insurance claims data study	1 million	Dose 1 > 12 months	2004 (Mass vaccination)	2009 vs. 2000	6.24 vs. 4.04/1000 person-years
Germany [69]	National hospital discharge data		Doses 1 and 2 at <24 months	2004 (1 Dose) 2009 (2 Doses)	2005–12 vs. 1995–2003	16.8 vs. 8.8/100,000

^aPublically funded.

HZ: herpes zoster.

A recent model considered three main outcomes after varicella vaccination in relation to HZ development: progressive accumulation of immunity following repeated VZV exposure, partial VZV protection that wanes over time, and full but temporary HZ immunity. The authors concluded that better understanding of the processes is therefore needed [114]. If routine infant varicella vaccination causes an albeit small increase in the incidence of HZ, there is a potential ethical dilemma whereby varicella vaccination although protecting one population (children) might have a deleterious effect on older individuals [39]. Other workers have proposed a more pragmatic approach in which zoster vaccination is used to supplement the varicella program and prevent HZ in older adults [40,41]. Nevertheless, long-term data within general populations are needed to determine the potential direct impact of universal varicella vaccination on HZ incidence.

3. Conclusions

Varicella poses a significant public health concern in children and can be prevented with effective varicella vaccination programs. The balance of evidence shows that one dose of varicella vaccine provides high protection against moderate-to-severe varicella but two doses are required for optimal protection against all varicella disease, to limit transmission and to reduce the risk of breakthrough cases and outbreaks. In countries where routine universal vaccination has been implemented, real-world effectiveness and impact studies show significant reduction in the incidence and disease burden of varicella without predicted rises in adult varicella and HZ.

4. Expert commentary

VZV is a highly contagious virus, infecting nearly the whole population. Over 90% of infected individuals subsequently develop varicella, and though the disease is generally mild, serious complication may occur. Indeed, WHO estimates that approximately 4.2 million severe complications leading to hospitalization and 4200 related deaths occur globally each year. However, even mild disease has a significant societal impact, with parents and caregivers having to take time off work to look after infected individuals.

There is robust evidence in the literature showing that varicella vaccines are safe and effective in preventing morbidity and mortality associated with the disease. However, despite the impressive VE, not all countries recommend routine varicella vaccination. In fact, recommendations currently only exist in 33 countries. Nevertheless, where implemented, real-world data have shown impressive reductions in disease incidence compared with the pre-vaccination era. After one-dose programs, reductions up to 74% have been recorded and after two-dose schedules, reductions exceeding 90% have been observed. Not surprisingly, vaccination programs have also been associated with a decrease in varicella-related hospitalization rates and death.

It has been suggested that childhood vaccination programs might result in an upward shift in the age distribution of varicella. However, to avoid this scenario, WHO recommends coverage rates above 80%. These measures as well as wider implementation of vaccination programs should see the effective reduction of this ‘mild,’ but potentially ‘serious,’ and frequently burdensome disease.

5. Five-year view

Over the next 5 years, more real-world data, particularly on long-term protection after one- and two-dose vaccination programs, will be available. These data will be supported by economic studies showing how vaccination can reduce the societal and economic burden of the disease. Better knowledge regarding varicella epidemiology under different coverage levels will emerge, and the impact of varicella vaccination upon the incidence of HZ will be more fully understood. Overall, more countries will have introduced universal varicella vaccination programs, as a result of clinical, real-world, and economic evidence.

Key issues

• VZV is a highly contagious virus infecting nearly all individuals

• Varicella is generally a mild disease but with potential serious complications and a high societal burden

• Varicella vaccines proved to be safe and effective in preventing the morbidity and mortality associated with the disease

• Vaccine recommendations currently exist in 33 countries

• Where implemented, real world data have shown impressive reductions in disease incidence compared with the pre-vaccination era, as well as fewer hospitalizations and deaths

Declaration of interest

P Wutzler reports personal fees from the GSK group of companies and Sanofi Pasteur MSD for participating in advisory boards and payment for lecture fees. P Bonanni reports grants and personal fees from Pfizer, GSK groups of companies, Novartis and Sanofi Pasteur MSD. A Gershon reports service contracts (molecular VZV diagnosis for vaccine safety) for ad hoc consulting on VZV by Merck & Co, for serving as a chair on the data and safety monitoring board for the VZV vaccine by the GSK group of companies and receiving NIH research funding. M Safadi reports grants and personal fees from the GSK group of companies, Novartis Vaccines, Sanofi Pasteur MSD and Pfizer for research and speakers’ honoraria. G Casabona is an employee of the GSK group of companies and holds shares in the GSK group of companies. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Acknowledgments

The authors acknowledge Preethi Govindarajan (GSK, Bangalore, India) and Julia Donnelly (Freelancer for XPE Pharma and Science, Wavre, Belgium c/o GSK) for writing support; and Ashmita Ravishankar (GSK, Bangalore, India) and Adrian Kremer (XPE Pharma and Science, Wavre, Belgium c/o GSK) for editorial support and publication management. The authors would also like to thank Dr. Iris Depaz, Dr. Nicolas Praet, Dr. Christophe Tournay and Dr. Jennifer Cnops for their critical input in the paper.

Trademark

Varilrix and Priorix-Tetra are trademarks of the GSK group of companies.

Varivax and ProQuad are trademarks of Merck & Co. Inc., USA.

Additional information

Funding

GlaxoSmithKline S.A. sponsored and funded this study and was involved in all stages of the study conduct and analysis and also took charge of all costs associated with the manuscript.