Abstract

Background

Herpes zoster (HZ), caused by reactivation of latent varicella zoster virus present in the dorsal root ganglia, is an acute disease characterised by a vesicular rash. The lifetime risk of developing HZ is about 20-30% [1-3]. However, the risk rises markedly with age, approximately doubling for each decade after 50 years of age [4]. This age-related increase in HZ, which can occur decades after primary varicella zoster infection, is due to a varicella zoster virus-specific decline in cell-mediated immunity with increasing age [5].

While pain associated with the acute infection tends to resolve with the rash, post-herpetic neuralgia (PHN), an intractable pain in the dermatome affected by HZ, remains a common complication. Precise definitions of PHN vary, but it is estimated to affect 10-20% of all patients with HZ aged >50 years [6-8] and up to 30% of those aged ≥80 years [6]. Pain associated with acute HZ infection and more especially PHN causes distress and adversely impacts patients' quality of life [9-14]. Moreover, the management of HZ and particularly PHN, the most common complication in the elderly, is often suboptimal [15,16].

After Japan, Italy has the highest proportion of elderly people in its population [17], and yet data on the epidemiology of HZ and PHN in Italy and the associated costs are limited. An observational study carried out a decade ago in Italy noted an incidence of 4.14 per 1000 population (aged ≥15 years) per year. In that study, over 75% of all HZ cases occurred in those aged ≥50 years, with about 50% occurring in the ≥65-year age group [18]. A more recent prospective Italian study found an annual incidence of 1.59 per 1000 population (aged ≥15 years), and calculated a mean cost for each HZ case treated in the outpatient setting of €360.60 and for each hospitalised case of €4082.59 [19].

The present study was designed to provide additional data on the incidence of HZ and to determine the proportion of HZ cases developing PHN in Italy. The study also sought to determine the associated medical resource utilisation and costs, and to estimate the total economic burden of HZ and PHN in Italy. Both parts of the study focused on the immunocompetent population aged ≥50 years, who would be eligible to receive a preventive vaccine that will be available soon in the European Union.

Methods

Results

Epidemiological analysis

HZ incidence

In total, 5675 incident cases of HZ were documented over the 3-year study period, with 1843, 1898 and 1934 cases recorded in 2003, 2004 and 2005, respectively. Patients aged ≥50 years accounted for 4119 (72.6%) incident cases, of which 3620 (87.9%) occurred in immunocompetent patients. Among men and women aged ≥50 years, this corresponded to an incidence of 6.65 per 1000 person-years (95% CI: 6.35-6.97), or 6.31 per 1000 person-years (95% CI: 6.01-6.62) in the immunocompetent population. In comparison, the incidence was 4.31 per 1000 person-years (95% CI: 4.11-4.52) for the adult population as a whole, or 4.07 per 1000 person-years (95% CI: 3.88-4.27) for the immunocompetent adult population. Incidence increased markedly with age, with the peak occurring in those aged 75-79 years (Figure ​(Figure1).1). HZ was more common among women than men in the total population, with an incidence of 4.75 (95% CI: 4.47-5.03) compared with 3.82 (95% CI: 3.62-4.03) per 1000 person-years, respectively.

Open in a separate window

Figure 1

Incidence of HZ in the general adult (≥15 years) and immunocompetent adult populations in Italy.

In the multivariate analysis, classical risk factors (female sex, age ≥55 years, and being immunologically compromised) were found to be independently associated with an increased risk of acquiring HZ infection.

Projected number of HZ cases per year

With over 22 million people aged ≥50 years living in Italy in 2005 [22], 139,435 new cases of HZ can be expected to occur annually in immunocompetent people aged ≥50 years. A total of 157,100 cases would be expected in all people aged ≥50 years, which represents 72.6% of the 216,391 cases expected for the adult population as a whole. Italian primary-care physicians, who have an average patient registry of 1283, can expect to see between 5 and 6 new cases of HZ per year.

Proportion of patients with HZ who developed PHN

In total, 452 cases of PHN1 were documented in the adult population over the 3-year study period, of which 350 (77.4%) cases also met criteria for PHN3. Thus, 8.0% (452/5675) (95% CI: 7.0-8.9) of the adult population with HZ experienced PHN lasting at least 1 month and 6.2% (350/5675) (95% CI: 5.3-7.0) experienced PHN for at least 3 months. The proportion of patients with HZ who developed PHN was highest in those aged ≥50 years (Figure ​(Figure2),2), with these older patients accounting for 91.2% (412/452) of all cases of PHN1 and 90.9% (318/350) of all cases of PHN3. The PHN1 proportion and PHN3 proportion were similar for all HZ patients aged ≥50 years (10.0% [412/4119] [95% CI: 8.8-11.2] and 7.7% [318/4119] [95% CI: 6.6-8.8], respectively) and the immunocompetent HZ patients aged ≥50 years (9.4% [342/3620] [95% CI: 8.2-10.7] and 7.2% [261/3620] [95% CI: 6.2-8.2], respectively).

Open in a separate window

Figure 2

Proportion of PHN in the general adult (≥15 years) and immunocompetent adult populations in Italy.

PHN was more common in women than in men with HZ. In the adult population as a whole, the proportion of women vs. men with HZ who developed PHN was 8.9% (294/3296) (95% CI: 7.7-10.2) vs. 6.6% (158/2379) (95% CI: 5.5-7.7) for PHN1 and 6.9% (228/3296) (95% CI: 5.8-8.0) vs. 5.1% (122/2379) (95% CI: 4.1-6.1) for PHN3. Over 65% of cases of both PHN1 and PHN3 in primary care were documented in female patients.

In the multivariate analysis, classical risk factors (female sex, increasing age [>60 years], and being immunologically compromised) were found to be independently associated with an increased risk of developing PHN.

Hospitalisation for HZ and PHN

Over the 3-year study period, the average incidences of hospitalisation for primary diagnoses of HZ and PHN were 10.34 and 1.89 per 100,000 person-years, respectively, in immunocompetent patients aged ≥50 years. The total incidence rose to 20.31 per 100,000 person-years (17.45 for HZ and 2.86 for PHN) in this population when both primary and secondary diagnoses on SDOs were combined. Corresponding figures for the adult immunocompetent population as a whole were 5.55 per 100,000 person-years for a primary diagnosis (4.61 for HZ and 0.94 for PHN) and 9.60 per 100,000 person-years for primary and secondary diagnoses combined (7.94 for HZ and 1.66 for PHN). Within the older population, rates of hospitalisation increased progressively for each decade up to 85 years (Figure ​(Figure3).3). Based on these figures, HZ or PHN as a primary diagnosis is responsible for 2829 hospitalisations annually in Italy among the immunocompetent adult population, of which 2611 (92.3%) admissions are in those aged ≥50 years. When primary and secondary diagnoses are combined, about 1.3% of HZ cases in Italy result in hospitalisation, rising to almost 2% where patients develop PHN. Overall, 16.9% of hospitalisations related to HZ were due specifically to PHN.

Open in a separate window

Figure 3

Age-related incidence of hospitalisation for HZ and PHN in immunocompetent adults in Italy.

Discussion

This retrospective, population-based study represents the largest investigation to date of the epidemiology and economic impact of HZ and PHN in Italy. Drawing on a primary-care database of over 450,000 patient records, together with data from national hospital-discharge records covering four major regions in Italy, our study showed that the incidence of HZ in immunocompetent individuals aged ≥50 years (6.31 per 1000 person-years), the primary focus of the study, was comparable with that reported in other studies [8,26], including an earlier Italian study [18]. Furthermore, the incidence reported here was consistent with recently published figures, for the population aged ≥50 years, from the UK (5.23 per 1000 person-years) [6] and Australia (9.7 per 1000 person-years) [7] despite some differences in study design. However, our overall HZ incidence is higher than the one recently published in Italy (4.31 vs. 1.59 per 1000 person-years) [19]. In common with these studies, HZ incidence increased markedly with age up to 80 years. In Italy, most patients with HZ consult their primary-care physician [18] and so our data, retrieved from the SIMG database, can be considered a reliable and robust source for estimates of HZ incidence.

HZ incidence was more common in women than men, a finding supported by several other studies [6,26-28], albeit not all [4,29]. In addition to older age and immune status, female sex was an independent risk factor for HZ, and indeed, for PHN as well. It is known that immunosuppression predisposes patients to increased risk of HZ and its complications [30]. It has been suggested that women may have a different immune response to latent viral infection and that this may account for their increased risk compared with men [27].

PHN, which in our study was identified by ICD9-CM codes or prescriptions for specific neuropathic pain medications, affected 9.4% of immunocompetent patients aged ≥50 years, of whom about 77% continued to have pain 3 months after HZ onset. These rates are consistent with those reported for several other community-based studies [31-33]. However, they are considerably lower than rates reported in clinical trials [34-36], in other various studies [8,26] and, interestingly, in a recent prospective study conducted in Piemonte in Italy [19] and the recent UK General Practice Research Database (UKGPRD) analysis [6], where 17.4% and 19.5% of patients developed PHN1, respectively.

Certain limitations of this study may have resulted in an underestimation of the proportion of PHN. In Italy, primary-care physicians manage most cases of HZ but only about a quarter of related cases of PHN because most cases of PHN are managed by specialists. Thus, this study did not capture cases of PHN that were not seen by a primary-care physician or referred to hospital. This demonstrates that use of the SIMG database to evaluate PHN in Italy has limits as it excludes patients who may have approached specialists directly. In addition, estimating the number of PHN cases retrospectively on the basis of drug prescriptions for pain relief rather than on the basis of pain severity may have led to an underestimation of the proportion of HZ cases that developed PHN because less serious cases which were not treated with drugs and serious cases where patients refused drug treatment may not have been detected. While our estimates of HZ incidence are robust, those for the proportion of patients with HZ who developed PHN are less reliable and are probably underestimates. Indeed, it is arguable that no study has produced reliable estimates of the proportion of HZ cases that develop PHN, and this is exemplified by the variation seen between studies [6,19,26,34,35]. Furthermore, in retrospective studies based on data retrieved from large databases, inaccuracies and, therefore, underestimations due to incompleteness of data cannot be excluded. Prospective studies, with patient follow-up, would provide a more reliable approach to future estimates of the proportion of patients with HZ developing PHN.

Some patients with HZ and PHN require hospital treatment, including inpatient care. Overall, about 1.3% of HZ cases required inpatient care in this study, increasing to almost 2% where only PHN cases were considered. At 9.60 per 100,000 person-years, the incidence of hospitalisations for HZ or PHN was higher in our study than rates reported in other European countries or North America [37,38] despite the exclusion in our study of immunocompromised patients who constitute those at greater risk of complications of HZ. In the recent prospective study in Piemonte, the annual incidence of hospitalisations for HZ was 0.12 per 1000 population [19]. Similar to the Piemonte study, age was an important predictor of likely hospitalisation for HZ or PHN in our study, with rates of hospitalisation increasing progressively for each decade up to 85 years of age, an observation that has also been reported previously [7]. Our data show a striking age-related increase in hospitalisation due to both HZ and PHN whether as a primary or secondary diagnosis.

We also estimated the economic impact of HZ and its complications in terms of direct healthcare costs as well as estimating likely indirect costs. As this study has demonstrated, HZ and PHN, the most common complication of HZ, impose a significant economic burden on the Italian population, with total costs for patients aged ≥50 years exceeding €41 million a year. At €28.2 million, direct medical costs accounted for just over two-thirds of total costs. Indirect costs associated with lost productivity were responsible for almost a third of total costs in the older population and reflect the debilitating nature of HZ and PHN among older people of working age. Although the recent prospective study in Piemonte did not extrapolate data to obtain total direct costs for the Italian population, at €136.06 per observed outpatient case of HZ, €360.60 per case treated in the outpatient setting, and €4082.59 for each hospitalised case (€2499.67 where HZ was the principal diagnosis) direct costs per case were broadly similar to our estimates [19].

Studies have consistently shown that HZ and especially PHN significantly impact patients' quality of daily life [9-14] and, moreover, are difficult to treat. Therapy relies on antiviral drugs to address the underlying cause of HZ and analgesia for pain relief in patients who develop PHN. As this study has shown, medications were responsible for over 80% of all direct outpatient costs, a finding supported by other studies [6,39]. This primary cost determinant may vary widely between countries. In the present study 78% of patients received antivirals, similar to the proportion reported for a recent study in Australia, whereas less than 25% of patients with HZ in The Netherlands receive antiviral therapy [7,40]. If administered within 72 hours of the onset of rash, antivirals have been shown to promote rash healing and reduce the intensity and pain of acute HZ [35,41]. However, a recent Cochrane Review found no evidence for an impact of antiviral therapy on PHN, with their efficacy for prevention of PHN considered modest at best [42].

Previously, it has been shown that patients requiring hospitalisation are likely to have more severe HZ or PHN, and thus, hospital-discharge data provide a measure of more serious cases [2]. In the present study, direct costs associated with inpatient care for an HZ case were more than 20 times the cost for outpatient care. In comparison, direct costs associated with inpatient care for a PHN episode were more than five times the cost for outpatient care. This illustrates the substantial costs associated with the management of more serious cases of HZ and PHN, where patients may require hospital care.

It is difficult to make direct comparisons between the present study and studies carried out in other countries to assess direct and indirect costs associated with HZ and PHN because of differences in healthcare practices and pricing. Nonetheless, as others have concluded, the economic and social costs of managing HZ and its complications represent an important burden on healthcare services and society [43], a view with which we would concur.

Our study had a number of limitations in addition to those already discussed. The higher percentage of home consultations in the very elderly, which may not have been recorded in the SIMG database, may have led to underestimation of incidence in the very old. Hospital-discharge data do not state unequivocally that HZ or PHN was the initial reason for hospitalisation, which may have affected the estimates of disease burden. As GP prescriptions could not be linked to specific PHN diagnostic codes and the drugs may have been prescribed for conditions other than PHN, outpatient costs for PHN may have been overestimated. Indirect costs were based solely on productivity losses due to absence from work and were derived from expert opinion, an approach that may limit the reliability of indirect cost estimates. Experts agreed that HZ patients will usually loose between 7 and 10 working days, while in case of serious pain 10-15 working days are reasonably lost due to PHN. Accordingly, the conservative assumption of 7 and 10 working days lost due to HZ and PHN, respectively, was made in the estimate of indirect costs. However, all focus group specialists stressed the extreme variability of the non-acute phase. Finally, when combining the results from the epidemiological and economic parts of the study it was not possible to obtain a measure of the reliability of estimates of the total cost of HZ and PHN (covariance between the different components of the total variability was unknown).

Conclusions

In conclusion, this study has demonstrated that HZ and related complications such as PHN impose a considerable disease and economic burden in Italy. While our estimates of HZ incidence in the Italian population can be considered robust, those for the proportion of patients with HZ who develop PHN are likely to be underestimates and would best be assessed in prospective studies. As HZ and PHN disproportionately affect the elderly, without intervention the burden to patients and society is likely to grow as the proportion of elderly in the Italian population continues to increase.

Competing interests

LN provided consultancy advice to Sanofi Pasteur MSD to assist with developing the study. AG has provided consultancy advice to Sanofi Pasteur MSD. AV has provided consultancy advice to Sanofi Pasteur MSD and Novartis, and has received honoraria for lectures from these companies and from GlaxoSmithKline and Menarini. He has also received honoraria for preparing educational materials for Menarini. LEG, MM and PP are consultants for Informa, Rome, who coordinated this study. MB, LS and RM are employees of Sanofi Pasteur MSD, a provider of a herpes zoster vaccine approved in the European Union.

Authors' contributions

LEG participated in the study design, analysis of data and drafting of the study report and helped to draft the manuscript. MM and PP participated in the study design and analysis of data. LN, AG and AV validated the protocol, design and final results of the study. MB, LS and RM participated in the study design, coordinated the study and validated the study results and study report. All authors read and approved the final draft manuscript.

Pre-publication history

The pre-publication history for this paper can be accessed here:

http://www.biomedcentral.com/1471-2334/10/230/prepub

References