Hospitalization due to varicella in the Netherlands

The varicella zoster virus (VZV) causes varicella (primary infection) as well as herpes zoster (reactivation of latent VZV in sensory nerve ganglia). Childhood varicella usually results in mild to moderate illness but serious complications like central nervous system involvement, pneumonia, secondary invasive bacterial infections and death may occur [1]. Since a life attenuated vaccine to prevent varicella has become available, many countries consider including it in their National Immunization Program. The United States was the first country that introduced universal childhood varicella vaccination. Their vaccination program started in 1995 (vaccine coverage among children aged 19 to 35 months increased nationally from 27% in 1997 to 89% in 2006) and has reduced overall disease incidence by 57% to 90%, hospitalizations by 75% to 88%, deaths by >74%, and direct inpatient and outpatient medical expenditures by 74% [2]. So far, in Europe varicella is listed as routine childhood vaccination in Germany and Greece. In some other European countries like Austria, Belgium, Cyprus, France, Italy, Poland, Spain, Sweden and Switzerland, routine childhood vaccination is only offered in some regions, only in the private sector or only to high risk groups and/or susceptible adolescents [3, 4].

In the Netherlands [5], the number of varicella related consultations of a general practitioner (GP), hospital admissions and/or deaths per 100,000 inhabitants is low compared with other countries (pre-vaccine area), such as the United States [2], England and Wales [6‐8] and Germany [9]. In England and Wales, where the health care system is most comparable to the Dutch situation, there were 507 GP-consults in the period 2001-2007 [6], 5.8 hospital admissions (based on data for England only) in 2000/2001-2008/2009 [7] and 0.038 deaths in 2000-2008 [8] per 100,000 inhabitants annually, which is roughly twice as much as the 238 GP-consults, 1.6 hospital admissions and 0.018 deaths per 100,000 inhabitants in the Netherlands in the period 2000-2008. It is not clear to what extent these differences can be ascribed to regional differences in the epidemiology of VZV, considering the fact that the mean age of infection in the Netherlands was lower than in England and Wales according to seroprevalence studies [10]. Other reasons might be differences in national surveillance systems, the health care system or health care seeking behavior. If these differences could be attributed to differences in the health care system or health care seeking behavior, which is in general more conservative in the Netherlands, there would be a possibility that Dutch hospitalizations due to varicella concern more severe cases than in other countries. Therefore, a medical record research was conducted among hospitalized patients with diagnosis varicella. This article presents the results of this study.

Data on hospital admissions due to main and/or side diagnosis varicella (ICD-9 code 052) in the period 2003-2006 were obtained from the National Medical Register from research institute Prismant [11]. Clinical admissions were included only (admissions for one day or less were excluded). To get more detailed insight into the severity and complications of varicella among hospitalized patients, a medical record research was conducted. An acknowledged Medical Ethical Committee of the University Medical Centre Utrecht approved of the study that was in accordance with the European statements for good clinical practice, which includes the provisions of the Declaration of Helsinki of 1989. We invited 23 out of 91 Dutch hospitals (situation November 2009), including one of the eight academic hospitals, to participate in this study; five of these hospitals had more than 750 hospital beds in 2008. All the hospitals consented in participation and patients were sent an information letter asking permission to include their personal medical record data in the study. Patients that could not be reached because they had moved or died, patients that did not give (active) permission and patients whose medical record could not be traced were excluded from the study. For the remaining patients, trained and authorized medical students collected additional information on comorbidity, primary reason of admission, complications and rest symptoms of varicella out of the medical record using a standardized form, under supervision of a research physician. Afterwards, all complications were divided in relatively moderate or relatively severe complications by a medical doctor. For all deaths, including deceased patients that could not be included in the full research, additional information was collected on the cause of death and the time interval between the admission due to varicella and death. Varicella-related death was defined as death within one month after varicella related hospital admission, unless another cause of death than varicella was appointed.

The medical record research showed that varicella complications occurred in the majority (76%) of hospitalized patients; 37% of these patients had at least one relatively severe complication. The most frequently mentioned complication that came forward, bacterial super infections of skin lesions, usually can be treated effectively by antibiotics. (Imminent) dehydration, the second most frequently mentioned complication, can normally be treated effectively by administering a rehydration solution. As expected due to the small study size, we did not find some unusual complications of varicella known from the literature such as Reye's syndrome, toxic shock syndrome, neonatal varicella and proliferative glomerulonephritis. Remarkably, there were no differences in the occurrence of complications between hospitalized patients with and patients without an underlying chronic condition. In 3% of the patients in this study serious rest symptoms were reported. Based on the cause of death and/or time between admission and death, none of the deaths were considered to be varicella related. This also counts for 7 of the 8 deceased patients that could not be included in the study. For one deceased patient for whom both date and cause of death were unknown, we cannot totally exclude the possibility that death was varicella related. However, this patient was admitted because of moderate complications (infected skin lesions and dehydration) that are unlikely reasons for case fatality. At this moment we have therefore no reason to assume that the lower number of admitted varicella patients in the Netherlands concerned more severe or other presenting cases than in other countries.

It is not easy to make a comparison of the severity of varicella complications among hospitalized patients between different countries. In a German study [9], where the incidence of varicella-related hospital admissions was estimated to be approximately three times higher than in the Netherlands (Germany 19.7 per 100,000 in children ≤16 years of age, the Netherlands 6.8 per 100,000 children ≤15 years of age), the proportion of certain complications among hospitalized varicella cases was rather comparable with the proportions found in our study. In this German study they found a febrile convulsion in 7.6% (Netherlands 8.7% in the same age-group ≤16 years of age), meningo-encephalitis in 5.7% (Netherlands 1.6%), cerebral vasculitis/infarct in 0.7% (Netherlands 0.4%), staphylococcal scalded skin syndrome in 0.7% (Netherlands 0.4%), coagulation disorder in 5.6% (Netherlands 2.8%) and sepsis in 2.3% (Netherlands 2.0%) of the hospitalized varicella patients. The reason behind the difference between both countries in the proportion of patients with meningo-encephalitis and coagulation disorder to a lesser extent is not clear. A possible explanation could be the difference in median age of the study population (for Germany 3.3 years and for the Netherlands 1 year) because the severity of varicella complications is known to be related to increasing age. Another possibility could be the differences in data collection methods between the German study (nationwide, prospective, hospitals were asked each month to send a reporting card to report all patients with a varicella related condition, including postinfectious varicella complications, and for these patients additional information was collected) and the Dutch study (not nationwide, retrospective, based on available information in the medical record of patients).

A limitation of this study is that the selection of patients was based on registration of varicella (as main or side diagnosis) in the National Medical Register. In a considerable part of patients, varicella was incorrectly registered as side diagnosis instead of main diagnosis. Furthermore, a considerable part of the patients were incorrectly registered in the National Medical Register as 'varicella without complication' or as 'varicella with unspecified complication', sometimes despite the fact that additional codes were registered in the National Medical Register that were very likely complications caused by varicella. We cannot exclude the possibility that varicella might not have been registered at all in another unknown number of patients. This means that there is a chance of underestimation of the number of hospitalizations due to varicella. Future research should therefore focus on following up patients that consult their GP due to varicella, to see how often they are referred to a hospital.

Only a part of all Dutch hospitals participated in the study, representative for 21.7% of all the admissions with varicella in the Netherlands in 2003-2006, which limits our study sample size. Furthermore, the records of 3% of all patients with varicella diagnosis could not be included due to lack of permission or missing records in the hospital. However, the study population was representative for all hospitalized patients with varicella diagnosis in the Netherlands in 2003-2006, with the exception of the type of hospital, but within the study population both median duration of admission and the proportion with at least one complication in the academic hospital did not differ significantly from other hospitals.

This study showed that a considerable part of the varicella complications among hospitalized cases was rather moderate and can normally be treated effectively, although in a third of the hospitalized cases with complications, severe complications occurred. Furthermore, in spite of the lower incidence of hospital admissions due to varicella, there is no reason to assume that the severity of varicella among hospitalized patients in the Netherlands differs from other countries. These data are relevant in the decision-making process regarding whether or not to introduce routine varicella vaccination in the Netherlands.