Assessing the potential effects and cost-effectiveness of programmatic herpes zoster vaccination of elderly in the Netherlands

Available vaccines and indications

Only one vaccine (ZOSTAVAX^®; SP-MSD) has been registered for the prevention of HZ. This live attenuated vaccine is manufactured by the same process as the chicken pox vaccine VARIVAX^® but has a higher viral load per dose [2]. The vaccine has been registered in the EU as a single dose vaccine for the prevention of HZ and PHN among people aged 50 years or older.

Vaccine efficacy

Natural protection against HZ may occur by exogenous boosting (due to circulating VZV in the population) or endogenous boosting (through subclinical reactivation of latent VZV). Although the mechanism of latency is not fully understood, there is strong evidence that the risk of developing HZ is linked to a decline in VZV-specific cell-mediated immunity (CMI) [1, 4]. The functional mechanism of the vaccine is to boost this specific CMI [2].

The efficacy of the vaccine was assessed in a large randomized placebo-controlled trial. There was a reduction of 51.3% in the incidence of HZ, 61.1% in the burden of illness (BOI) and 66.5% in the incidence of PHN [5]. The vaccine appeared less effective in the older age group (70+ years) compared to the younger age group (60-69 years) (Figure 1), indicating that the effect of vaccination is age dependent [5]. The long term efficacy of the vaccine is unknown (mean follow-up duration so far was three years), but the immunity seems to decrease over time after vaccination [2].

Contra-indications and adverse events following vaccination

Since the vaccine consists of live-attenuated virus, it should not be used in immunocompromised people, people with active untreated tuberculosis or in pregnant women [2].

Adverse events at the injection site occurred more frequently in the vaccine group (48.3%) compared to the placebo group (16.6%), but most of them were mild. Furthermore, vaccine-related systemic adverse events occurred more frequently in the vaccine group than in the placebo group (6.3% vs 4.9%) [5].

Factors affecting successful implementation

So far, influenza vaccination is the only generally advised vaccination for elderly in the Netherlands. The general practitioner (GP) invites all people aged 60 years or older annually for this vaccination, which has a high coverage (in 2008/2009 76.9%) [6]. HZ and influenza vaccine given concomitantly are well tolerated [7]. Furthermore, antibody responses were similar compared to sequential vaccination. A recent study, however, showed that among community-dwelling elderly to whom both influenza and HZ vaccination were offered within an existing influenza vaccination program, only 39% accepted HZ vaccination, whereas 76% accepted influenza vaccination [8]. Determinants of non-compliance with additional HZ vaccination were: perceived lack of recommendation by the GP, unwillingness to comply with the doctor's advice, perception of low risk of contracting HZ, perception of short pain duration of HZ and the opinion that vaccinations weaken one's natural defenses [8]. Other studies also found that a recommendation by the GP is a major determining factor of accepting vaccination in this age group [9, 10]. An international survey pointed out that the understanding of the risk of developing HZ, its symptoms, complications and treatment among adults ≥55 years of age is very limited [11]. Moreover, in the United States the lack of patient awareness and physician recommendation were pointed to be barriers to HZ vaccine uptake [10].

Herd immunity

The transmission of VZV resulting from patients with HZ is very low in comparison to varicella [1]: therefore no herd immunity effects are to be expected. Reaching a high vaccination coverage is therefore not important, unlike for most other vaccinations. HZ vaccination will only give benefit on individual level.

Pathogenicity

VZV-seroprevalence in the Netherlands approaches 100% from seven years onwards [12]. In HIV-infected persons the risk of HZ and its recurrence is increased (12-17 fold) [1]. Intercurrent infection with viruses that can alter CMI responses (such as Epstein-Barr virus and cytomegalovirus) also influences the risk of developing HZ [4].

Infectiveness and transmissibility

HZ is not transmitted directly; it is a reactivation of VZV that remains latent in sensory nerve ganglia after primary VZV-infection. The herpes lesions are contagious for non-immune persons (until the lesions have crusted) and can lead to varicella [1]. Subclinical reactivation of the VZV virus is possible but the frequency of occurrence is unknown [4]. In immunocompetent individuals, the frequency of recurrent HZ is low (1.7-5.2%) [13].

Antigenic variation

The VZV genome is extremely stable. So far, seven distinct genotypes of the wild-type VZV have been distinguished with a different geographic distribution, but all belong to the same serotype. No evidence for recombination among wild-type VZV-strains has yet been found [14]. Although recombination events could theoretically alter the virulence of circulating VZV strains [15], the impact of such events would probably be very small. Ecological consequences after implementation of vaccination are not expected. VZV is an exclusively human pathogen. Both the vaccine strain and the wild-type VZV establish a latent infection. Furthermore, interaction or competition with other alpha-herpes viruses like HSV-1 and HSV-2 has not been described for VZV [16].

Risk factors for herpes zoster

It is estimated that 23-30% of the population in Europe will develop HZ during their lifetime; approximately 50% of all people reaching the age of 85 years will have experienced HZ [13]. Prior infection with VZV, either with wild-type or vaccine virus, is a prerequisite for developing HZ. The vaccine virus may have less opportunity to reactivate than does wild-type VZV [4]. The vaccine virus usually does not cause viremia or skin infection, factors that are both likely to enhance the development of HZ [17].

The incidence of HZ increases with age, which is attributed to the natural process of age-related immunosenescence. Furthermore, the incidence is higher among people with immunity attenuating diseases or medication [1, 4, 18]. Other possible risk factors that have been suggested are physical trauma at the involved dermatome, psychological stress, changes in mental health, depression, white race and intercurrent infection with viruses that can alter CMI responses [1, 4, 18]. Some studies show also higher incidence rates in women, even after correction for higher average age and health care seeking behavior [18, 19]. VZV-infection in utero or shortly after birth has been found to be a risk factor for (childhood) HZ [1, 4, 18]. PHN is more likely to occur in older HZ patients and in HZ patients with severe pain or rash during the acute phase [4, 18, 20].

Consequences of herpes zoster

HZ begins with a prodrome, during which abnormal skin sensations and pain of varying severity are the most common symptoms, followed by a vesicular rash. This rash is typically unilateral, does not cross the mid-line, normally involves a single dermatome, is usually accompanied by acute pain and lasts for 7-10 days or longer. PHN, a persistent pain after resolution of the rash, is the most important complication of HZ and can last for several years [1, 4]. Therapeutical options for HZ and PHN are scarce. About half of the patients with PHN will benefit from therapy with only partial relief [4]. The quality of life during HZ is influenced by the severity and duration of the acute and chronic pain that can affect physical, psychological, social and functional domains [1, 4].

Alternative preventive measures

There are no direct alternatives to prevent HZ. Childhood vaccination against varicella might reduce the HZ incidence on the long term, because the vaccine strain is less likely to cause HZ than the wild-type. However, reduced VZV transmission due to varicella vaccination will diminish exogenous exposure (boosting), which might lead to an increase in the incidence of HZ in the mid-term (the first 30-50 years) [21]. Studies monitoring the incidence of HZ in the US, where universal vaccination against varicella was introduced in 1995, have shown inconsistent findings at this point. Two studies did not show an increase in overall incidence [22, 23], whereas three others demonstrated a rise [24–26].

GP consultations, hospitalizations and deaths

Most HZ patients will consult their GP as it is a painful condition. Age-specific incidence rates for the period 2002-2007 were derived from the Netherlands Information Network of General Practice (LINH) [27]. A correction was made for false positives (10%; 7.9-12.4%[28]) and immunocompromised people (5%[28]), since both groups will not benefit from vaccination. A linear regression was plotted on the HZ incidence of the separate years 2002-2007.

Hospitalization data (ICD-9 code 053) were taken from the National Medical Register (LMR) for the period 2000-2007. Only admissions with HZ as main diagnosis were included because these admissions represent cases that are preventable by vaccination. The incidence of clinical admissions was rather stable in the period 2000-2007. However, the incidence of admissions for one day decreased from 7.5 per 100000 in 2002 to 4.0 per 100000 in 2007 [29]. Therefore, an alternative scenario was included in which the daytime hospital visits were excluded. The distribution used in the probabilistic sensitivity analysis is listed in Additional file 1.

Mortality data (ICD-10 code B02 and G530) for the period 2000-2007 were derived from Statistics Netherlands (CBS). Only deaths with HZ as primary cause of death were included in the base case scenario. An alternative scenario without prevention of death was also included since it is likely that death is not caused directly by HZ.

Pain, incidence of PHN and quality-adjusted life-year (QALY) loss

The duration of pain by severity and age, and subsequently the QALY loss due to HZ, was estimated by Van Hoek et al [28] and applied to the Dutch situation. For clarity, this does include PHN which was defined as the presence of clinical relevant pain after three months. In the model QALY loss after onset was modeled based on the duration spent in clinical relevant or mild pain [28] instead of using a fixed percentage developing PHN.

Direct costs of disease

The major costs involved in HZ are the prescription of antivirals and repetitive GP visits for PHN patients. In the PINE study, detailed information on GP consultations, medication and additional use of health services due to HZ was available for the first 6 months of the study (Additional file 2). Based on those findings the average total costs per patient of GP consultations and drug use is €72.05 (€66.90 - €77.20) in case of HZ and €101.10 (€81.72 - €120.70) for PHN based on the first 6 months of the study. Because the duration of PHN can be longer, these costs were doubled: €201.91 (€163.30 - €241.15). Confidence intervals of the mean price (95%) were acquired by bootstrapping.

Indirect costs of disease

Indirect costs were considered for estimated work loss till the age of 65. Data on work loss due to HZ is scarce and the participation in the workforce is not high in the age group 60+. A questionnaire among 65 HZ patients in the UK [32] showed that 29 patients were employed, with an average working loss of 10.1 days (SD of mean 1.82). According to CBS, participation in the work force (in 2006) was only 20.8% in the age group 60-65. The number of hours of labour per week is 32 or 6.4 per day with a payment of €24.10 per hour. With a correction for participation in the workforce this is an average of €32.04 lost per day or €324 for the total work loss for someone in the age group 60-65.

Cost of the vaccine and the vaccination program

Because the HZ vaccine is not yet available in the Netherlands, the Dutch price is unknown. The official retail price of the HZ vaccine in the US is $153.93 or €110 (Pack 10-Vial; January 2009). However, in case of introduction in the NIP, the CDC price of $107.67 or approximately €77 (January 2009) seems more applicable. In the sensitivity analyses the effect of lower vaccine prices was calculated.

Based on experience with the introduction of the pneumococcal vaccine in the Netherlands in 2006, the once-only costs (not included in the cost-effectiveness model) are estimated to be €0.3 million and include costs for education of GPs, developing information material (invitation letter, flyer, publicity campaign, website), adjustment of software for registration and monitoring, and administration. In case of implementing HZ vaccination within the current influenza vaccination program and assuming a vaccination coverage of 75%, the estimated yearly administration costs range from ~€14.7 million for vaccinating people at the age of 60 to ~€4.9 million for vaccination at the age of 80. This includes compensating vaccination personnel (€4.80 per application, this is half the influenza tariff) and coordination costs (€1.65 per application). In the sensitivity analyses the effect of higher applications costs (€9.60 instead of €4.80 per vaccination) was calculated. Monitoring of adverse events can be included in the already existing passive surveillance system, for which the total costs are estimated to be €0.4 million per year. Vaccine effectiveness, reflected by the reduction of the incidence of HZ, PHN and related hospitalizations, could be monitored using GP and hospitalization statistics.