Pediatric Hematology and Oncology

ISSN: 0888-0018 (Print) 1521-0669 (Online) Journal homepage: http://www.tandfonline.com/loi/ipho20

Varicella in pediatric oncology patients in the post-

vaccine era—Analysis of routine hospital data
from Bavaria (Germany), 2005–2011

Andrea Streng, Verena Wiegering & Johannes G. Liese

To cite this article: Andrea Streng, Verena Wiegering & Johannes G. Liese (2016) Varicella in
pediatric oncology patients in the post-vaccine era—Analysis of routine hospital data from
Bavaria (Germany), 2005–2011, Pediatric Hematology and Oncology, 33:7-8, 468-479, DOI:
10.1080/08880018.2016.1245805

To link to this article: http://dx.doi.org/10.1080/08880018.2016.1245805

Published online: 14 Dec 2016.

Submit your article to this journal

Article views: 6

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ipho20

Download by: [Fudan University] Date: 02 January 2017, At: 14:55

PEDIATRIC HEMATOLOGY AND ONCOLOGY
, VOL. , NOS. –, –
http://dx.doi.org/./..

Varicella in pediatric oncology patients in the post-vaccine

era—Analysis of routine hospital data from Bavaria (Germany),
–
Andrea Streng, Verena Wiegering, and Johannes G. Liese
Department of Paediatrics, University of Würzburg, Würzburg, Germany

ABSTRACT ARTICLE HISTORY

Varicella in oncology patients can result in serious complications. We Received  September 
analyzed trends in hospitalization rates and characteristics of pediatric Accepted  October 
oncology and non-oncology patients hospitalized with varicella dur-
KEYWORDS
ing the first 7 years after introduction of routine varicella vaccination. Children; hospitalization;
Our data included children <17 years of age with an International Clas- oncology; surveillance;
sification of Diseases and Related Health Problems, 10th Revision (ICD-10) varicella zoster virus
main or secondary discharge diagnosis of varicella identified by annual
database queries in 22–29 pediatric hospitals in Bavaria (Germany)
in 2005–2011. Of a total of 1,245 varicella-associated hospitalizations,
42 children (median age 4 years, interquartile range 3–5) had an underly-
ing malignancy (67% with acute lymphoblastic leukemia). Overall, addi-
tional diagnoses potentially associated with varicella were reported less
often in oncology than in non-oncology varicella patients (62% vs. 77%,
p = 0.041), suggesting earlier hospitalization of high-risk patients. Acute
hematological diagnoses (29% vs. 3%, p < 0.001) and coinfections (inva-
sive 12% vs. 2%, p = 0.001; noninvasive 19% vs. 8%, p = 0.019) were more
frequent, whereas neurological (5% vs. 19%, p = 0.023) and upper respi-
ratory tract diagnoses (2% vs. 16%, p = 0.014) were less frequent in oncol-
ogy compared to non-oncology varicella patients. Oncology varicella
patients showed a longer hospital stay (median 5 vs. 3 days, p < 0.001).
Hospitalization rates in non-oncology varicella patients declined con-
stantly since 2006, from 114.8 (2006) to 30.5 (2011) per 1,000 pediatric beds.
The rates of varicella-associated hospitalizations in oncology patients
indicated an overall decreasing trend (3.8, 1.9, 4.6, 3.5, 0.4, 2.1 and 0.6
cases per 1,000 pediatric beds in 2005–2011). Thus, pediatric oncology
patients potentially profit from herd protection effects, resulting from
increasing vaccine coverage in the general population.

Introduction
Varicella (chickenpox) is primarily a childhood illness, usually with an uncomplicated course
of disease in otherwise healthy children [1]. However, varicella-associated complications with
hospitalization do occur, with incidence estimates ranging from 13 to 28 per 100,000 children
[1,2], Such complications are observed both in immunocompetent and immunocompromised
patients [3,4]. Varicella infection in oncology patients immunocompromised by their dis-
ease or therapy is potentially life-threatening and may result in serious complications such as

CONTACT Dr. rer. nat. Andrea Streng Streng_A@ukw.de Department of Paediatrics, University of Würzburg, Josef-
Schneider-Str. ,  Würzburg, Germany.
©  Taylor & Francis Group, LLC
 PEDIATRIC HEMATOLOGY AND ONCOLOGY 469

varicella pneumonia, encephalitis, hepatitis, coagulopathies, and bacterial superinfections

[3–5]. However, available data on this specific patient group are limited and restricted to
case reports and results from few, mainly monocentric studies [3,5–15]. Children with acute
leukemia, the most common form of pediatric malignancy, are reported to be at a higher risk
of varicella pneumonia and fatal varicella compared to children with other types of cancer
[5,6,16].
In Germany, universal, health insurance financed one-dose varicella vaccination for all
children aged 11–14 months was recommended in July 2004, [17] and the recommendation
was extended to a second dose for children aged 15–23 months in July 2009 [18]. In 2014,
nationwide varicella vaccination coverage in children aged 5–6 years had reached 86% for the
first dose and 82% for the second dose [19]. The introduction of universal varicella vaccina-
tion resulted in an overall decline of varicella-associated complications and hospitalizations in
Germany [20,21]. Although vaccination with the live attenuated varicella zoster virus is usu-
ally contraindicated for children in immunosuppressed condition, such as oncology patients,
these risk group patients might profit from herd protection effects as a consequence of increas-
ing levels of vaccination coverage in the general population [4,22].
In the study presented here, we compared patient characteristics and additional diagnoses
in oncology and non-oncology pediatric patients hospitalized with varicella and analyzed
varicella hospitalization rates in these two patient groups over a 7-year period.

Methods

Study setting
Data collection was carried out at pediatric hospitals from January 2005 to December 2011
within the framework of the “Bavarian Varicella Surveillance Project ”[23]. The pediatric hos-
pitals in Bavaria (37 hospitals with 3054 beds in 2005, 36 hospitals with 3044 beds since 2007)
covered an annual study population of an average 2,069,000 children <17 years of age in
the study period 2005–2011 [24]. The hospitals were invited to participate in annual, stan-
dardized database queries of routine hospital data, based on codes according to the Interna-
tional Classification of Diseases and Related Health Problems, 10th Revision (ICD-10). Dur-
ing the observation period, the “Population-based Cancer Registry Bavaria” reported annu-
ally between 231 and 279 new cases of pediatric oncology patients <15 years of age for
Bavaria [25].

Data collection
Participating hospitals provided basic demographic data, month and year of hospital admis-
sion, duration of hospital stay, all ICD-10 codes, and procedural codes for each patient fulfill-
ing the following inclusion criteria: at least one ICD-10 code for varicella (B01) as main diag-
nosis or as any secondary diagnosis at discharge; age below 17 years at hospital admission;
hospital stay within the time period 2005–2011; and duration of hospital stay at least 1 day
[23]. The subgroup of oncology patients among all reported varicella patients was identified
by checking for an additional ICD-10 code for malignancy (C00–C97) as main diagnosis or as
any secondary diagnosis at discharge. Hospital name, year of admission, sex, age, residential
code, and ICD-10 diagnoses indicating underlying chronic conditions were used to identify
double data sets/patients with multiple admissions. In cases with multiple admissions, only
the first admission of the patient was included in the analysis. Based on additionally reported
470 A. STRENG ET AL.

ICD-10 diagnoses, patients were allocated to predefined categories of complications consid-

ered to be potentially varicella-associated [3].

Statistical analyses
Data were submitted by the hospitals as Microsoft Excel files and transferred to SPSS version
23.0 (SPSS Inc., Chicago, Illinois) for statistical analysis. In view of the annual variation in
the number and size of participating pediatric hospitals, the annual numbers of oncology and
non-oncology varicella patients are reported as averages per 1,000 hospital beds. All other
data are reported as numbers and percentages, or as median with interquartile range (IQR).
Patient characteristics were compared between subgroups using Pearson’s chi-square test or
Fisher’s exact test for categorical data, and the Mann-Whitney U test for continuous data,
with a predefined level of significance of 0.05 (two-sided). Annual estimates for the rates of
oncology and non-oncology children hospitalized with varicella in Bavaria were calculated
by extrapolating the number of varicella patients observed per hospital bed in the subgroup
of participating hospitals to the total number of pediatric hospitals beds in the study area [3].

Ethical considerations and data protection

Only anonymized data were collected. The study was approved by the Bavarian Data Protec-
tion Office (Munich, Germany) and the Ethics Committees of the Medical Faculties at the
Universities of Munich and Würzburg.

Results

Participating hospitals and patient reports

The number of pediatric hospitals participating per study year ranged between 22 and 29 (61–
78% of all 36–37 eligible hospitals). From January 2005 to December 2011, the participating
hospitals reported a total of 1,263 hospitalized children with varicella (median 36, IQR 20–51).
Median age was 3 years (IQR 1–5); 55% were male children. The 1,263 varicella-associated
hospitalizations reported included 42 oncology and 1,203 non-oncology patients; 18 patients
were excluded as likely cases of multiple admissions/double reports. All 42 cases of varicella
hospitalizations in oncology patients were reported by a subgroup of 12 hospitals, with 36
(86%) patients reported by seven large hospitals with pediatric oncology departments.

Patient characteristics of oncology and non-oncology varicella patients

The 42 hospitalized oncology patients with varicella (median age 4 years, IQR 2.8–5.3; 57%
male children) included 28 patients (67%) with acute lymphoblastic leukemia (ALL) as the
most frequent malignancy. Nine patients had a solid tumor (21%), including malignancies
of the brain (3 patients), kidney and/or lung (2), peripheral abdominal nerves (2), bone (1),
and soft tissue of the head (1). Other types of leukemia/lymphoma were reported for five
patients (12%), including chronic (2 patients) or unspecified (1) lymphatic leukemia, acute
myeloid leukemia (1), and diffuse non-Hodgkin’s lymphoma (1). The majority (23; 55%) of
these 42 hospital admissions occurred during the months of January to May. Demographic
characteristics, underlying conditions, hospital stay, and outcome are summarized in Table 1.
Additional underlying conditions were reported for nine oncology patients (21%). For a total
 PEDIATRIC HEMATOLOGY AND ONCOLOGY 471

Table . Characteristics of oncology and non-oncology children with varicella-associated hospitalization.

Hospitalized
Hospitalized oncology non-oncology varicella
varicella patients N = § patients N = §§
n (%), or median (IQR) n (%), or median (IQR) p-value#

Demographic data
Male sex  (.)  (.) .
Age (years)  (–)  (–) .
Age group .
– years  (.) (.)
– years  (.)  (.)
– years  (.)  (.)
Underlying chronic condition##
At least one underlying  (.)  (.) <0.001
chronic condition∗
Immunocompromised  (.)  (.) <0.001
condition∗
Cardiac/pulmonary system  (.)  (.) .
Developmental disorder,  (.)  (.) .
pre-term birth
Metabolic disorder  (.)  (.) .
Congenital malformation  (.)  (.) 0.034
Neurological disorder  (.)  (.) .
Ear/nose/throat/eye  (.)  (.) .
Renal disease/uropathy  (.)  (.) .
Syndromal disease  (.)  (.) .
Other comorbidity  (.)  (.) .
Hospital stay, outcome
Length of hospital stay (days)  (–)  (–) <0.001
Death  (.)  (.) .

Data from pediatric hospitals in Bavaria, –.

§  of  data sets excluded due to double reporting/multiple admission.
§§  of  data sets excluded due to double reporting/multiple admission.
# Chi² or Fisher’s exact test, as appropriate, for categorical data; Mann-Whitney U test for continuous data.
## Multiple categorizations possible.
∗ Pre-defined condition for oncology patients.

of 17 (41%) patients, treatment with chemotherapy, irradiation, or instillation of cytotoxic

material or immunomodulators was documented, and two patients (5%) had received a stem
cell transplant. Six patients (14%) were reported to be in remission.
A comparison of hospitalized varicella patients with and without oncological conditions
(Table 1) showed no significant difference in age in both groups (median 4 vs. 3 years, p =
0.071) and largely similar frequencies regarding the occurrence of non-oncological underly-
ing conditions. Oncology varicella patients showed a longer hospital stay than non-oncology
patients (median 5 vs. 3 days; p < 0.001). No fatalities were observed among oncology vari-
cella patients, whereas two fatalities occurred in non-oncology varicella patients (in a child
below 1 year of age due to varicella pneumonia, and in a 3-year old, immunocompromised
child with DiGeorge syndrome due to cardiac arrest).

Additional ICD-10 diagnoses in oncology and non-oncology varicella patients

Additional ICD-10 diagnoses in oncology patients potentially indicating complications asso-
ciated with varicella disease are summarized in Table 2. Twenty-six patients (62%) showed
at least one additional diagnosis. The most frequent diagnoses were hematological complica-
tions (12 patients, 29%; including 11 patients with thrombocytopenia), gastrointestinal com-
plications (8 patients, 19%; including diarrhea/gastroenteritis, obstipation, mucositis, and
472 A. STRENG ET AL.

Table . Additional acute diagnoses (potentially associated with varicella disease) in oncology and non-
oncology children with varicella-associated hospitalization.
Hospitalized oncology Hospitalized non-oncology
varicella patients N = § varicella patients N = §§
n (%), or median (IQR) n (%), or median (IQR) p-value#

Patients with at least one 0.041

additional acute diagnosis
No  (.)  (.)
Yes  (.)  (.)
Duration of hospital stay
In patients without additional  (–)  (–) 0.007
acute diagnoses
In patients with additional  (–)  (–) <0.001
acute diagnoses
Additional acute diagnoses
(categorized)##
Hematological  (.)  (.) <0.001
Gastrointestinal tract  (.)  (.) .
Coinfections (excluding severe  (.)  (.) 0.019
invasive bacterial infections
or skin infections)
Severe invasive bacterial  (.)  (.) 0.001
infection
Skin  (.)  (.) .
Neurological  (.)  (.) 0.023
Lower respiratory tract  (.)  (.) .
Coagulation and sequelae  (.)  (.) .
Upper respiratory tract, ENT,  (.)  (.) 0.014
eye
Reactive arthritis  (.)  (.) .
Other acute diagnoses  (.)  (.) .

Data from pediatric hospitals in Bavaria, –.

§  of  data sets excluded due to double reporting/multiple admission.
§§  of  datasets excluded due to double reporting/multiple admission.
# Fisher’s exact test.
## Multiple categorizations possible.

dehydration), noninvasive coinfections with other pathogens (8 patients; 19%; coinfection

with adenovirus (2), norovirus (1), Pseudomonas sp. (1), Haemophilus influenzae (1), Staphy-
lococcus aureus (resistant) (1), and two cases of unspecified coinfections), and invasive bac-
terial coinfection (5 patients (12%); one with sepsis due to S. aureus, four with unspecified
sepsis). Four patients (10%) had skin complications, and two (5%) patients each had lower
respiratory tract complications (varicella pneumonia, acute respiratory insufficiency), neu-
rological complications (cerebellitis, drug-induced polyneuropathy/myopathy), and coagula-
tion disorders. For 13 patients (31%), other diagnoses were reported, including eight patients
with drug-induced agranulocytosis/neutropenia; further complications were drug-induced
Cushing syndrome, feeding problems, cardiac arrhythmia, pericardial effusion, splenomegaly,
and conjunctivitis.
A comparison of varicella hospitalization patients with and without an oncological con-
dition (Table 2) showed that additional diagnoses were in general less frequently reported in
oncology varicella patients (62% vs. 77%, p = 0.041). However, oncology varicella patients
were more often reported to have had hematological complications (29% vs. 3%) and invasive
bacterial (12% vs. 2%) or noninvasive (19% vs. 8%) coinfections. Neurological diagnoses and
upper respiratory tract/ENT/eye diagnoses, by contrast, were reported less often for oncology
than for non-oncology varicella patients (5% vs. 19% and 2% vs. 16%, respectively).
Duration of hospital stay was longer in oncology than in non-oncology varicella patients,
in patients without additional diagnoses (median 5 vs. 2 days, p = 0.007) as well as in patients
 PEDIATRIC HEMATOLOGY AND ONCOLOGY 473

Table . Oncology patients with varicella-associated hospitalization, with or without additional diagnoses
(potentially associated with varicella disease).
Oncology varicella patients Oncology varicella patients
with additional acute without additional acute
diagnoses N =  diagnoses N = 
n (%), or median (IQR) n (%), or median (IQR) p-value#

Demographic data
Male sex  (.)  (.) .
Age (years)  (–)  (–) .
Age group .
– years  (.)  (.)
– years  (.)  (.)
– years  (.)  (.)
Underlying chronic condition#
Acute lymphoblastic leukemia  (.)  (.) .
Oncology status “complete  (.)  (.) .
remission”
 (.)  (.) .
Chemotherapy/irradiation/instillation
of cytotoxic material or
immunomodulators
Implanted catheter/vascular  (.)  (.) .
access
Additional underlying chronic  (.)  (.) .
disease
Hospital stay
Length of hospital stay (days)  (–)  (–) 0.035

Data from pediatric hospitals in Bavaria, –.

# Fisher’s exact test or Mann-Whitney U test, as appropriate.

with such diagnoses (median 7 vs. 4 days, p < 0.001; Table 2). Oncology varicella patients with
lower respiratory tract, gastrointestinal tract, or “other” additional acute diagnoses showed a
considerably longer duration of hospital stay compared to non-oncology varicella patients
with such diagnoses (median 43 vs. 5 days, p < 0.001; median 12 vs. 3 days, p = 0.004; and
median 14 vs. 4 days, p < 0.001, respectively).

Oncology varicella patients with and without additional diagnoses

Within the group of oncology varicella patients, children with (n = 26) and without (n = 16)
additional diagnoses did not differ regarding their underlying conditions (Table 3). However,
oncology varicella patients with additional diagnoses revealed a significantly longer duration
of hospital stay (median 7 vs. 5 days, p = 0.035) than patients without such diagnoses. Fur-
thermore, they showed a trend to be more often male children (69% vs. 38%; p = 0.059) and
to receive more often treatment involving the use of a permanent central intravenous catheter
or vascular access device than those without such diagnoses (69% vs. 38%; p = 0.059).

Change in the hospitalization rate of oncology and non-oncology varicella patients

over time
From 2005 to 2011, between 1 and 12 pediatric oncology patients hospitalized with varicella
were reported annually from the participating hospitals, corresponding to an estimated rate
of <1 and 5 patients per 1,000 hospital beds in Bavaria (Table 4). Although the number of
oncology patients hospitalized with varicella showed considerable variations between study
474 A. STRENG ET AL.

Table . Hospitalized pediatric oncology and non-oncology varicella patients (absolute numbers and rates
per , hospital beds, stratified by year.
Study year
      

Oncology patients
hospitalized with
varicella (n = )
Absolute number of       
reported cases in
participating
hospitals
Absolute number of . . . . . . .
estimated cases in
all hospitals in
Bavaria
Rate per , hospital . . . . . . .
beds
Non-oncology patients
hospitalized with
varicella (n = )
Absolute number of       
cases reported from
participating
hospitals
Absolute number of . . . . . . .
estimated cases in
all hospitals in
Bavaria
Rate per , hospital . . . . . . .
beds

A total of – hospitals participated per year. Estimates of cases for all – pediatric hospitals in Bavaria were calculated by
extrapolation from the numbers per pediatric bed in participating hospitals to all pediatric beds. Data from pediatric
hospitals in Bavaria, –.

years, the data indicated a decreasing varicella hospitalization rate between the first 3 and the
last 3 years of observation (Table 4).
In comparison, the number of patients reported with non-oncology varicella hospitaliza-
tions decreased from 239 in 2005 (maximum 303 in 2006) to 53 in 2011. In this group, the
estimated rate of varicella-associated hospitalizations declined constantly since 2006, from
115 varicella patients per 1,000 hospital beds in 2006 to 31 in 2011 (Table 4).

Discussion
The results of this multicenter study from Germany showed that pediatric oncology and non-
oncology patients hospitalized with varicella were largely comparable with regard to demo-
graphic characteristics and the presence of most non-oncological underlying chronic condi-
tions. However, they differed regarding the frequency and the types of additional diagnoses
potentially associated with varicella disease.
Overall, additional diagnoses indicating varicella-associated complications were reported
less often in oncology than in non-oncology varicella patients (62% vs. 77%, p = 0.041). Simi-
larly, we had found a non-elevated rate of varicella-associated complications in oncology com-
pared to non-oncology varicella patients in a previous single-center study from Germany [5].
These observations suggest that the very vulnerable oncology patients may be hospitalized
early in the course of varicella disease for antiviral treatment, whereas non-oncology chil-
dren are hospitalized usually after developing relevant complications of varicella disease [3]. A
 PEDIATRIC HEMATOLOGY AND ONCOLOGY 475

similar difference in the hospitalization pattern was recently also reported from immunocom-
promised (mainly oncological) and immunocompetent pediatric patients with reactivation of
latent varicella zoster virus infection (herpes zoster) [26].
In the present study, the evaluation of additional diagnoses indicating varicella-associated
complications showed that oncology varicella patients were affected more often by hematolog-
ical diagnoses (29% vs. 3%) and suffered more frequently from invasive (12% vs. 2%) and non-
invasive (19% vs. 8%) coinfections. In contrast, they were less often diagnosed with neurolog-
ical (5% vs. 19%) and upper respiratory tract complications (2% vs. 16%) than non-oncology
varicella patients. In line with the results presented here, previous comparisons of immuno-
compromised (including oncological) and immunocompetent varicella patients had observed
a higher frequency of hematological [27] and invasive bacterial coinfections [3] and a lower
frequency of neurological complications [3,27] in immunocompromised varicella patients.
These previous studies also reported a significantly higher frequency of lower respiratory
tract complications [3,27] in immunocompromised patients, whereas in the present study,
such complications were reported with a similar frequency in oncology and non-oncology
varicella patients.
In our study, oncology varicella patients stayed for a longer period in the hospital (by a
median of 2–3 days) than non-oncology varicella patients; this difference was observed both
for patients without and with additional diagnoses. It is known that the duration of varicella
in immunocompromised patients is usually longer than that in non-immunocompromised
patients. Therefore, it is likely that even in the absence of additional complications, oncology
varicella patients remain under vigilance or are treated for a longer period than non-oncology
varicella patients.
Within the group of oncology varicella patients, children with additional diagnoses indicat-
ing varicella-associated complications required a significantly longer hospital stay than those
without such complications (median 7 vs. 5 days), and the occurrence of complications tended
to be associated with male sex (p = 0.059). These results are similar to those reported for a
study recently carried out in Pakistan; [14] however, in contrast to that study, we did not
observe a difference in age between oncology varicella patients with and without complica-
tions. In our study, complications in oncology patients also tended to be associated with the
presence of implanted central intravenous catheters (p = 0.059). Central venous catheters
used in the standard management of patients with acute leukemia are known to be a poten-
tial source of bacterial infection in neutropenic patients [28] and might explain—besides the
immunosuppression itself—the higher rate of invasive bacterial infections in oncology vari-
cella patients.
Since universal varicella vaccination was recommended in Germany, vaccination coverage
in children has increased constantly but with some regional differences [19,29]. Surveillance
in Bavaria during the study period showed an increase in first-dose coverage rates in children
up to 3 years of age from 38% in 2006 to 51–53% in 2007–2009, and to 66–68% in 2010–
2011 [23]. Coverage for the second dose increased from 29% (2009) to 52% (2010) and to
59% (2011) [23]. Corresponding to this increase in coverage, the annually estimated num-
ber of hospitalized non-oncology varicella patients in Bavaria in the present study showed a
continuous decrease after the year 2006. Such a constant trend was not apparent in oncol-
ogy patients hospitalized with varicella; their estimated rates showed considerable variations
between study years (between a maximum of 4.6 per 1,000 hospital beds in 2007 and a min-
imum of 0.4 in 2009). Nevertheless, they also indicated a certain trend toward a reduction
in numbers, with the two highest rates reported during the first 3 years and the two lowest
476 A. STRENG ET AL.

rates during the last 3 years of observation. A decline in the prevalence of pediatric oncol-
ogy patients in Bavaria seems unlikely to be the cause, as the incidence of childhood cancer
cases in Bavaria remained fairly stable throughout the study period. Our data therefore give
some indication that oncology patients may profit from herd protection effects by rapidly
increasing coverage of varicella vaccination in the age-related population. This indirect form
of protection is of particular importance as intensive chemotherapy in pediatric patients with
ALL or other forms of cancer leads to long-term impairment of humoral immunity, and vacci-
nation with the live attenuated varicella virus is usually contraindicated in immunocompro-
mised patients [16,30]. The options of post-chemotherapy varicella vaccination in children
with ALL in remission or of vaccination during maintenance therapy are the subject of con-
troversial discussion [4,31–34] and therefore not regularly done.
In the course of the observation period 2005–2011, younger children were increasingly
covered by the recommendation for varicella vaccination [23]. Hence, another potential factor
contributing to the decrease in the number of oncology patients hospitalized with varicella is
the possibility that some children may have been vaccinated before the onset of cancer [35].
However, our study did not permit us to distinguish between individual and herd protection
effects, as information on the varicella vaccination status of the patients is not available in
ICD-10 diagnoses.
The study presented here, which used routine, ICD-10-based data derived from hospital
databases, has various other limitations [3]. Available medical information on the patients was
based solely on ICD-10 diagnoses, and only limited information on procedural codes was
available. It is likely that the procedural coding available may have been restricted to codes
relevant for hospital reimbursement. Information on antiviral treatment was not available
through the coding system; neither was information on steroid therapy, which is known to
have an impact on the severity of varicella in patients with ALL [11]. ICD-10 misclassifica-
tions or incomplete documentation may have occurred in some patients. It is possible that not
all cases of multiple admissions or transfer of a patient to another hospital were detected in
the data, as data protection restrictions prevented the use of unique patient identifiers. Fur-
thermore, the number per 1,000 hospital beds of oncology patients hospitalized with varicella
showed strong variations between study years, as is to be expected in view of the small abso-
lute numbers per year. It cannot, therefore, be ruled out that the indications of a decrease in
the number of varicella-associated hospital admissions in oncology patients we observed in
our study were a result of random variation.

Conclusion
ICD-10 data and procedural codes from routine documentation databases of hospitals
can be a useful tool in the surveillance of oncology varicella patients. Compared to non-
oncology patients, oncology varicella patients had less frequently diagnoses indicating poten-
tial varicella-associated complications. This suggests a tendency toward earlier, preemp-
tive hospitalization in these immunocompromised children. Varicella in oncology patients
resulted in a longer hospital stay and was more often associated with hematological diag-
noses and coinfections and less often with neurological complications than in non-oncology
varicella patients. The low rates of varicella-associated hospitalizations observed in pediatric
oncology patients during the last three study years suggest that these patients may profit from
herd protection effects, resulting from the increasing varicella vaccination coverage in the
general population.
 PEDIATRIC HEMATOLOGY AND ONCOLOGY 477

Acknowledgments
The authors thank Karin Seeger for the valuable comments on the manuscript, and Sabrina Hanke and
Christina Bungartz for their support in data management. They also thank the following hospitals for
participation in the study: Kreisklinik Altötting, Zentrum für Kinder-und Jugendmedizin (Altötting);
Klinikum St. Marien Amberg, Klinik für Kinder und Jugendliche (Amberg); Klinikum Aschaffenburg,
Klinik für Kinder-und Jugendmedizin (Aschaffenburg); Klinikum Augsburg, Klinik für Kinder und
Jugendliche (Augsburg); Klinik Josefinum, Kinderkrankenhaus (Augsburg); Klinikum Sozialstiftung
Bamberg, Kinderklinik (Bamberg); Klinikum Bayreuth GmbH, Kinderklinik (Bayreuth); Regiomed-
Kliniken gGmbH, Klinik für Kinder-und Jugendmedizin (Coburg); DONAUISAR Klinikum
Deggendorf, Kinder-und Jugendmedizin (Deggendorf); Universitätsklinikum Erlangen, Kinder-
und Jugendklinik (Erlangen); Klinikum Fürth, Klinik für Kinder und Jugendliche (Fürth);
Klinikum Garmisch-Partenkirchen GmbH, Fachabteilung Kinder-und Jugendmedizin (Garmisch-
Partenkirchen); Sana Klinikum Hof GmbH, Kinderklinik (Hof); Klinikum Kaufbeuren, Abteilung
Pädiatrie (Kaufbeuren); Klinikum Kempten-Oberallgäu, Kinderklinik (Kempten); Klinikum
Landsberg am Lech, Abteilung für Kinder-und Jugendmedizin (Landsberg); Kinderkranken-
haus St. Marien, Abteilung für Kinder-und Jugendmedizin (Landshut); Städtisches Klinikum
München GmbH, Klinikum Harlaching, Klinik für Kinder-und Jugendmedizin (München);
Ludwig-Maximilians-Universität München, Dr von Haunersches Kinderspital (München); Städti
sches Klinikum München GmbH, Klinikum Schwabing, Klinik für Kinder-und Jugendmedizin
(München); Klinikum Dritter Orden, Klinik für Kinder-und Jugendmedizin (München); Kliniken
St. Elisabeth, Abteilung für Kinder-und Jugendmedizin (Neuburg a.d. Donau); Klinik Haller-
wiese, Cnopf ’sche Kinderklinik (Nürnberg); Klinikum Nürnberg, Klinik für Neugeborene, Kinder
und Jugendliche (Nürnberg); Kinderklinik Dritter Orden (Passau); Krankenhaus Barmherzige
Brüder Regensburg, Klinik für Kinder und Jugendmedizin (Regensburg); RoMed Klinikum
Rosenheim GmbH, Kinderklinik (Rosenheim); Leopoldina-Krankenhaus der Stadt Schweinfurt
GmbH, Klinik für Kinder und Jugendliche (Schweinfurt); Klinikum Starnberg, Klinik für Kinder-
und Jugendmedizin (Starnberg); Kliniken Südostbayern AG, Klinikum Traunstein, Klinik für
Kinder-und Jugendmedizin (Traunstein); Kliniken Nordoberpfalz AG, Klinikum Weiden, Klinik
für Kinderheilkunde und Jugendmedizin (Weiden); Universitätsklinikum Würzburg, Kinderklinik
und Poliklinik (Würzburg); and Missionsärztliche Klinik gGmbH, Kinderklinik (Würzburg).

Declaration of interest
The Bavarian Varicella Surveillance Project was financially supported by an unrestricted research grant
from GlaxoSmithKline (GSK) Vaccines (Rixensart, Belgium) from 2006 to 2011. AS and JGL received
further research grants from both manufacturers of varicella vaccine, GSK and Sanofi Pasteur MSD. AS
and JGL received support for conference attendance, and AS honoraria for expert meetings from GSK
and Sanofi Pasteur MSD. VW declares no conflict of interest.

References
[1] Heininger U, Seward JF. Varicella. Lancet. 2006;368:1365–1376.
[2] Papaloukas O, Giannouli G, Papaevangelou V. Successes and challenges in varicella vaccine. Ther
Adv Vaccines. 2014;2(2):39–55.
[3] Liese JG, Grote V, Rosenfeld E, et al. The burden of varicella complications before the intro-
duction of routine varicella vaccination in Germany. Pediatr Infect Dis J. 2008;27(2):119–
124.
[4] Roderick M, Finn A, Ramanan AV. Chickenpox in the immunocompromised child. Arch Dis
Child. 2012;97:587–589.
[5] Wiegering V, Schick J, Beer M, et al. Varicella-zoster virus infections in immunocompromised
patients—a single centre 6-years analysis. BMC Pediatrics. 2011;11:31.
[6] Feldman S, Hughes WT, Daniel CB. Varicella in children with cancer: seventy-seven cases. Pedi-
atrics. 1975;56:388–397.
478 A. STRENG ET AL.

[7] Feldman S, Loft L. Varicella in children with cancer: impact of antiviral therapy and prophylaxis.
Pediatrics. 1987;80:465–472.
[8] Ho CML, Khuzaiah R, Yasmin AM. Varicella in children with haematological malignancy—
outcome of treatment and prevention. Med J Malaysia. 1994;49(1):29–35.
[9] Rowland P, Wald ER, Mirro JR Jr, et al. Progressive varicella presenting with pain and minimal
skin involvement in children with acute lymphoblastic leukemia. J Clin Oncol. 1995;13(7):1697–
1703.
[10] Folatre I, Zolezzi P, Schmidt D, Marín F, Täger M. Infections caused by Varicella Zoster
virus in children with cancer aged less than 15 years old [Article in Spanish]. Rev Med Chil.
2003;131(7):759–764.
[11] Hill G, Chauvenet AR, Lovato J, McLean TW. Recent steroid therapy increases severity of varicella
infections in children with acute lymphoblastic leukemia. Pediatrics. 2005;116:e525–e529.
[12] Sørensen GV, Helgestad J, Rosthøi S. Varicella-associated morbidity in children undergo-
ing chemotherapy for acute lymphoblastic leukaemia [Article in Danish]. Ugeskr Laeger.
2009;171(46):3354–3359.
[13] Escaño-Gallardo ET, Bravo LC. Varicella in immunocompromised children at the Philippine Gen-
eral Hospital: a six-year review. PIDSP Journal. 2011;12(1):27–39.
[14] Alam MM, Qamar FN, Khan ZW, Kumar V, Mushtaq N, Fadoo Z. Risk factors for complicated
varicella infection in pediatric oncology patients at a tertiary health care facility in Pakistan. J
Infect Dev Ctries. 2014;8(2):215–220.
[15] Noronha V, Ostwal V, Ramaswamy A, et al. Chicken pox infection in patients undergoing
chemotherapy: a retrospective analysis from a tertiary care center in India. J Infect Public Health.
2016, Mar 9. [Epub ahead of print]
[16] Patel SR, Bate J, Maple PAC, Brown K, Breuer J, Heath P. Varicella zoster immune status in children
treated for acute leukemia. Pediatr Blood Cancer. 2014;61:2077–2079.
[17] Robert Koch-Institut. Empfehlungen der Ständigen Impfkommission (STIKO) am Robert Koch
Institut/Stand: Juli 2004. Epidemiol Bull. 2004;30:235–250.
[18] Robert Koch-Institut. Empfehlungen der Ständigen Impfkommission (STIKO) am Robert Koch
Institut/Stand: Juli 2009. Epidemiol Bull. 2009,30:279–298.
[19] Robert Koch-Institut. Impfquoten bei der Schuleingangsuntersuchung in Deutschland. Epidemiol
Bull. 2016,16:129–133.
[20] Spackova M, Muehlen M, Siedler A. Complications of varicella after implementation of routine
childhood varicella vaccination in Germany. Pediatr Inf Dis J. 2010;29(9):884–886.
[21] Siedler A, Dettmann M. Hospitalization with varicella and shingles before and after introduction
of childhood varicella vaccination in Germany. Hum Vaccines Immunother. 2014;10(12):3594–
3600.
[22] Gershon AA, LaRussa P, Steinberg S. The varicella vaccine. Clinical trials in immunocompro-
mised individuals. Infect Dis Clin North Am. 1996;10:583–594.
[23] Streng A, Grote V, Carr D, Hagemann C, Liese JG. Varicella routine vaccination and the effects
on varicella epidemiology—results from the Bavarian Varicella Surveillance Project (BaVariPro),
2006–2011. BMC Infect Dis. 2013;13:303.
[24] Bayerisches Landesamt für Statistik und Datenverarbeitung. Altersstruktur der Bevölkerung Bay-
erns. Available at: https://www.statistik.bayern.de/statistik/zensus/00843.php. Accessed August
30, 2016.
[25] Bevölkerungsbezogenes Krebsregister Bayern [Population-based Cancer Registry Bavaria]. Avail-
able at: http://www.krebsregister-bayern.de. Accessed August 30, 2016.
[26] Kuchar E, Szenborn L, Lis I, Jaroszewska A, Czeladzka J. Clinical presentation of herpes zoster
in immunocompetent and immunocompromised hospitalized children treated with acyclovir. J
Pediatr Hematol Oncol. 2016;38(5):394–397.
[27] Peterson CL, Mascola L, Chao SM, et al. Children hospitalized for varicella: a prevaccine review.
J Pediatr. 1996;129(4):529–536.
[28] Karthaus M, Doellmann T, Klimasch T, Krauter J, Heil G, Ganser A. Central venous catheter
infections in patients with acute leukemia. Chemotherapy. 2002;48(3):154–157.
[29] Helmuth IG, Poulsen A, Suppli CH, Mølbak K. Varicella in Europe—a review of the epidemiology
and experience with vaccination. Vaccine. 2015;33:2406–2413.
 PEDIATRIC HEMATOLOGY AND ONCOLOGY 479

[30] Bochennek K, Allwinn R, Langer R, et al. Differential loss of humoral immunity against measles,
mumps, rubella and varicella-zoster virus in children treated for cancer. Vaccine. 2014;32:3357–
3361.
[31] Levin MJ. Varicella vaccination of immunocompromised children. J Infect Dis. 2008;197 (Suppl
2):S200–S206.
[32] Caniza MA, Hunger SP, Schrauder A, et al. The controversy of varicella vaccination in children
with acute lymphoblastic leukemia. Pediatr Blood Cancer. 2012;58:12–16.
[33] Van de Wetering MD, Vossen MT, Jansen MH, Caron HN, Kuijpers TW. Varicella vaccination in
pediatric oncology patients without interruption of chemotherapy. J Clin Virol. 2016;75:47–52.
[34] Smedegaard LM, Poulsen A, Kristensen IA, Rosthøj S, Schmiegelow K, Nygaard U. Varicella vac-
cination in children with leukemia without interruption of maintenance therapy—a Danish expe-
rience. Pediatr Infect Dis J. 2016;35:e348–e352. Jun 10 [Epub ahead of print]
[35] Fisher JPH, Bate J, Hambleton S. Preventing varicella in children with malignancies: what is the
evidence? Curr Opin Infect Dis. 2011;24:203–211.