Pediatr Blood Cancer 2011;56:10921097

Invasive Fungal Infections in Pediatric Oncology

Meirav Mor, MD,1,2 Gil Gilad, MD,2,3 Liora Kornreich, MD,2,4 Salvador Fisher, MD,2,3
Isaac Yaniv, MD,2,3 and Itzhak Levy, MD1,2*
Background. Data on the epidemiology and outcome of invasive fungal infections in children with cancer are limited. The aim
of the study was to delineate the epidemiologic, clinical features,
risk factors, and outcome of invasive fungal infections in this population. Procedure. The medical records of all children with malignancies diagnosed with an invasive fungal infection in 19982006
at a tertiary pediatric medical center were reviewed for demographic, clinical, and laboratory data. Invasive fungal infection
was diagnosed according to the latest EORTC/MSG criteria. Results.
Of the 1,047 children hospitalized in the hematology/oncology
department during the study period, 75 (7.2%) were diagnosed with
a proven (n 16, 21.3%), probable (n 18, 24%), or possible
(n 41, 54.7%) invasive fungal infection. Fifteen (20%) had candidemia (non-albicans in 60%), and 60 (80%) had a mold infection
(non-Aspergillus in 55%). Crude mortality was 21.7%. The most

Key words:

common underlying diseases were myeloid leukemia (n 26, 34.7%)

and acute lymphoblastic leukemia (n 24, 32%). Compared to
other malignancies, acute myeloid leukemia was signicantly
associated with the development of invasive fungal infections. Profound neutropenia and high treatment intensity were present in
89% and 73% of patients with IFI respectively. Conclusions. The
current mortality rates of invasive fungal infection in children with
cancer are lower than previously reported in children and adults.
However, the proportion of non-albicans candidemia is increasing,
and non-Aspergillus molds are emerging as important pathogens,
which may have important implications for prophylaxis and empiric
therapy. Improved prevention, early detection, and advanced treatment strategies are needed to improve the outcome. Pediatr Blood
Cancer 2011;56:10921097. 2011 Wiley-Liss, Inc.

children; fungal infections; hematology/oncology; invasive

INTRODUCTION
Invasive fungal infections (IFIs) are a major cause of morbidity and mortality in patients with hematologic malignancies. The
incidence in adults may be as high as 30%, and the mortality, up
to 50% [13]. In children, the true incidence and outcome are
more difcult to assess because of the growing trend toward more
invasive chemotherapeutic regimens, the development and introduction of new antifungal drugs and prophylactic antifungal strategies, and differences in design and patient populations among the
various studies. In general, incidence rates appear to be lower than
in adults. However, relative to the many reports describing different aspects of IFIs in adults, the data in children, particularly for
the last 10 years, are limited. The information that is available
pertains mostly to the neonatal age group or to Candida bloodstream infections [4,5].
In 2005, Rosen et al. [6] described the 11-year experience of a
single center in the treatment of fungal infections in children with
cancer. They reported an incidence rate of 4.9% and a mortality
rate of 59%. However, their cohort included patients with mucosal
infections as well as IFIs, and they did not dene IFIs according
to international guidelines. Castagnola et al. [7], in a 2-year prospective multicenter study of IFIs diagnosed according to the
criteria of the European Organization for Research and Treatment
of Cancer (EORTC), documented a mortality rate of 28% in
pediatric patients with cancer and after hematopoietic stem-cell
transplantation (HSCT). More recently, Kobayashi et al. [8] in a
single-center pediatric study from Japan, reported a 6.9% incidence of IFIs, and a 48.2% related mortality rate, but they also
included patients with underlying metabolic abnormalities and
immunodeciencies. Although there have been several reports
on the epidemiology, risk factors, and outcome of pediatric IFIs,
all were single-center studies [9], limited to patients with leukemia or after HSCT [1012], or to patients with Aspergillus infections [13,14], Candida infections [15,16], or bloodstream
infections [17]. Some were done before the newer antifungal
agents and diagnostics became available, and some were based
on the earlier EORTC denitions.

2011 Wiley-Liss, Inc.

DOI 10.1002/pbc.23005
Published online 11 February 2011 in Wiley Online Library
(wileyonlinelibrary.com).

Factors that appear to be associated with IFIs in both children

and adults in this severely immune-suppressed population are the
underlying malignancy (mainly acute leukemia), presence of profound and long-lasting neutropenia, high intensity of the chemotherapeutic regimen, HSCT, and previous antibiotic therapy [3,9,18].
The purpose of the present study was to further delineate the
current clinical and epidemiologic features of IFIs in patients
hospitalized in a pediatric hematology/oncology department.

PATIENTS AND METHODS

This retrospective study included all hematology/oncology
patients and recipients of the hematopoietic stem cell transplants
aged 020 years who were diagnosed with a proven, probable, or
possible IFI at the Schneider Childrens Medical Center of Israel,
a tertiary university-afliated hospital, between 1998 and 2006.
The patients were identied by the diagnoses list at discharge, the
hospital pharmacy list of hemato-oncology patients prescribed
antifungal medications, and records of the hospitals microbiology
laboratory.
The medical, microbiological, and imaging records of the
patients who met the inclusion criteria were reviewed for the
following variables: Demographic and clinical data: Age and
1

Infectious Diseases Unit, Schneider Childrens Medical Center of

Israel, Petah Tikva, Israel; 2Sackler Faculty of Medicine, Tel Aviv
University, Tel Aviv, Israel; 3Department of Pediatric Hematology/
Oncology, Schneider Childrens Medical Center of Israel, Petah
Tikva, Israel; 4Department of Imaging, Schneider Childrens Medical
Center of Israel, Petah Tikva, Israel
Grant sponsor: Gilead Sciences Research Foundation, Foster City,
CA.
Conict of interest: Nothing to declare.
*Correspondence to: Itzhak Levy, MD, Infectious Diseases Unit,
Schneider Childrens Medical Center of Israel, 14 Kaplan Street,
Petah Tikva 49202, Israel. E-mail: [email protected]
Received 26 May 2010; Accepted 3 December 2010

Invasive Fungal Infections in Pediatric Oncology

gender, type, and stage of underlying disease at diagnosis; history
of HSCT; predisposing factors such as radiation therapy; cancer
therapy protocol; intensity of chemotherapy administered (according to an international consensus) [19]; time from the end of the
last chemotherapy course to the rst symptoms of IFI; corticosteroid use 14 days prior to IFI onset; presence and type of central
venous catheter (CVC); presence of mucositis; administration of
total parenteral nutrition (TPN) within 14 days of onset of the
fungal infection; use and type of antibiotics within 14 days of
onset of the fungal infection; use of granulocyte-colony stimulating factor (GCSF) within 14 days of onset of the fungal infection;
type and duration of symptoms and signs of the fungal infection.
Laboratory data: Complete blood count; serum electrolytes
included blood and urine cultures; fungus detection tests included
direct stains, cultures, and fungal polymerase chain reaction
(PCR) studies of bronchoalveolar lavage (BAL) uid, sinus aspirate, and samples from other sites (e.g., knee joint effusion).
Radiological data: X-ray, computed tomography (CT), and ultrasound studies were included. For the purpose of this study, all
radiological data were reviewed by a blinded pediatric radiologist
(L.K.) and classied according to the revised EORTC denitions
for IFI [20]. Treatment and outcome of IFI: Antifungal prophylaxis,
empiric therapy, and denitive therapy with one or a combination
of antifungal drugs, use of GCSF surgery and 3-month mortality.
Our departmental policy stipulates that patients after HSCT
receive prophylactic antifungal treatment with uconazole for the
rst 100 postoperative days, and patients with acute myeloid
leukemia (AML) receive prophylactic itraconazole during the
initial phase of induction and consolidation therapy. Other
patients on the oncology service do not receive antifungal prophylaxis. Febrile neutropenia (neutrophil count <500/mm3) is
empirically treated with piperacillintazobactam and amikacin;
when the fever persists for more than 4 days and no source is
identied, amphotericin B or other empiric antifungal therapy is
administered. Blood cultures are obtained daily during periods of
fever and neutropenia, and urine cultures are obtained at admission. Galactomannan assay is not routinely as a marker of invasive
aspergillosis due to its limitations in children treated with piperacillintazobactam. Radiologic evaluation is performed after 47
days of fever, including CT of the paranasal sinuses, chest, and
abdomen, in addition to cardiac echocardiogram and fundoscopic
eye examination. If a focus of suspected fungal infection is
observed, further evaluation is performed with functional endoscopic sinus surgery (FESS) of the paranasal sinuses or BAL uid
analysis for lung lesions. The samples are processed in the microbiology laboratory by direct stains, fungal cultures, and PCR tests.
If the source of the infection is not found and either the fever
persists for 2 weeks or the patients condition deteriorates, CT
evaluation is repeated.
The primary empiric antifungal treatment in our department is
amphotericin B, except in cases of highly suspected aspergillosis,
for which the rst-line therapy was changed from amphotericin B
to voriconazole in 2003. The standard approach to treatment of
invasive candidiasis is amphotericin B administration unless the
isolate is uconazole-susceptible. Otherwise, either a polyene or
an echinocandin is used. The antifungal agent used is sometimes
switched during the course of the illness if the patient is intolerant
(severe systemic reaction, renal or electrolyte disorders) and in
culture-positive cases, according to the drug susceptibilities of the
specic pathogen isolated.
Pediatr Blood Cancer DOI 10.1002/pbc

1093

In all cases, IFI was dened according to the latest version

(2008) of the EORTC/Mycosis Study Group (MSG) Consensus
[20]. Proven IFI was diagnosed by a positive fungal culture from a
normally sterile site. Probable IFI was diagnosed on the basis of a
combination of host factors, clinical and radiological features, and
mycological evidence, such as positive fungal culture or microscopy of BAL uid or sinus aspirate. Possible IFI was diagnosed
when the clinical and imaging ndings and host factors were
consistent with IFI but there was no mycological support. The
EORTC does not accept PCR results as mycological evidence of
IFI; therefore, we did not include them in the level of proof
determination. Patients were excluded from the study if data on the
fungal infection were insufcient or a fungal infection was ruled out.
Intensity of treatment (ITR) was dened according to an international consensus described by Werba et al. [19]. Yeast and mold
infections were analyzed separately. We also performed separate
analyses for patients after HSCT. Crude and attributable mortality
were calculated. The crude mortality was based on the rates of
death in the 3 months following diagnosis of IFI. Attributable
mortality was determined according to clinical parameters and
the treating physicians assessment of the cause of death. The
study was approved by the hospital review board.

Statistical Analysis
All epidemiological data are descriptive. The proportion of
patients with IFI and 95% condence intervals were calculated
for selected types of malignancies as well as for autologous and
allogeneic bone marrow transplant recipients. Exact mid-p condence limits were computed for proportions based on less than 50
cases [21].

RESULTS
Between 1998 and 2006, 1,047 patients (56.8% male) were
treated in the pediatric hemato/oncology department of our center.
The main diagnoses were solid tumors (609 patients, 58.2%),
acute lymphocytic leukemia (ALL) (205 patients, 19.6%), AML
(66 patients, 6.3%), and lymphoma (120 patients, 11.5%). A total
of 331 patients underwent HSCT: allogeneic in 201 (60%) and
autologous in 130 (40%).
Seventy-ve of the 1,047 patients (7.2%) were diagnosed with
an IFI: proven in 16 (21.3%), probable in 18 (24%), and possible
in 41 (54.7%). Fifteen patients (20%) had a yeast infection (1.4%
of all the patients with cancer) and 60 (80%) had a mold infection
(5.7% of all the patients with cancer). Thirty patients were diagnosed during the induction phase of chemotherapy, and 26
patients during cancer relapse or recurrence. The demographic
and clinical characteristics of the patients with IFIs are shown
in Table I. All patients with a yeast infection had a proven
diagnosis (positive blood cultures). Among the patients with a
mold infection, 1 (1.7%) had a proven diagnosis, 18 (30%) a
probable diagnosis, and 41(68.3%) a possible diagnosis of IFI.
The incidence of IFI by according underlying disease is shown
in Table II. Rates were highest in the patients with hematologic
malignancies, especially AML (9 patients with proven/probable
IFI, 13.6%; 95% CI 0.0580.217) and ALL (12 patients with
proven/probable IFI, 5.9%; 95% CI 0.0320.097), and lower in
the patients with solid tumors (10 patients with proven/probable
IFI, 1.6%; 95% CI: 0.0080.029).

1094

Mor et al.

TABLE I. Characteristics of 75 Children Diagnosed With an

Invasive Fungal Infection
Mold infection
No. of patients
Age at diagnosis of IFI
<6 years
612 years
>12 years
Gender (% male)
Baseline diagnosis
AML
ALL
Lymphoma
Aplastic anemia
Solid tumor
HSCT recipients
Predisposing factors
Steroid treatment
ANC < 500/ml
Mucositis
TPN in prior week
ITR level 4
Fever
Primary sites
Lungs
Sinus, nose, palate
Blood
Spleen
Kidney
Liver
Endocarditis
Joint
Eyes
Level of proof
Proven
Probable
Possible
Mortality
Crude
Attributable

Yeast infection

Characteristics

15

60

60%
0%
40%
80

28.3%
33.3%
38.3%
53.3

4 (26.7%)
1 (6.7%)
0
1 (6.7%)
7 (46.7%)
5 (33.3%)

20 (33.3%)
23 38.3%)
3 (5%)
4 (6.7%)
9 (15%)
19 (31.7%)

5 (33.3%)
12 (80%)
2 (13.3%)
3 (20%)
9 (60%)
15 (100%)

34 (56.7%)
55 (91.7%)
27 (45%)
11 (18.3%)
46 (76.7%)
55 (91.7%)

0
0
15 (100%)
1 (6.7%)
1 (6.7%)
1 (6.7%)
1 (6.7%)
1 (6.7%)
1(6.7%)

42 (70%)
22 (36.7%)
0
4 (6.7%)
1 (1.7%)
1 (1.7%)
0
1 (1.7%)

15 (100%)

1 (1.7%)
18 (30%)
41 (68.3%)

3 (20%)
3 (20%)

13 (21.7%)
6 (10%)

IFI, invasive fungal infection; AML, acute myeloid leukemia; ALL,

acute lymphoid leukemia; ANC, absolute neutrophil count; TPN, total
parenteral nutrition; ITR, intensity of treatment rating.

The classication of IFIs, causative fungi, and sites of infection are shown in Table III. Most of the patients with mold
infections had pulmonary and/or paranasal sinus involvement at
presentation, whereas all of the patients with yeast infections had
bloodstream infection.

Yeast Infections
The main underlying hemato-oncologic diagnoses in the
patients with yeast infections were AML (n 4, 26.7%) and
neuroblastoma (n 4, 26.7%). The most common predisposing
factors were severe neutropenia (80%), presence of a CVC (90%),
high ITR score (60%), and corticosteroid treatment (33%). In all
cases, fever was a presenting symptom of fungal infection.
Candida albicans was isolated in six patients and other Candida species in nine. All patients had a bloodstream infection.
Other sites of involvement are shown in Table III. After the
diagnosis of IFI was made, CVCs were promptly removed.
In 11 patients (73.3%), antifungal treatment was started only
after the pathogen was isolated in blood culture. The other four
patients (26.7%) received empiric treatment according to the
departmental protocol for patients with prolonged fever and neutropenia. The rst drug administered was amphotericin B in 13
patients (86.7%), uconazole in 1, and caspofungin in 1. In 10
patients (67.7%), the treatment was later changed according to the
drug susceptibility or drug toxicity; in two patients, a second drug
was added to amphotericin B because of an insufcient response
to treatment. Three patients (20%) died within 3 months of onset
of infection. All deaths were attributable to the fungal disease.

Yeast Infections After HSCT

Five patients had a yeast infection after HSCT (1.5%): within
3 months of transplantation in three patients (60%) and later in
two (40%). Two of the patients underwent allogeneic HSCT (1%)
and three autologous HSCT (2.3%). All patients in this subgroup
were male; 4 (80%) were age 6 years or younger. One patient
(20%) died within 3 months of acquiring the yeast infection.

Mold Infections
The diagnosis of proven mold infection (1 patient) was based
on Fusarium growth in knee joint uid. The diagnosis of probable
infection was based on positive fungal culture results (18
patients). The main underlying diseases were ALL (23 patients,
38.3%) and AML (21 patients, 35%). Among all the patients
admitted to our department during the study period, the incidence

TABLE II. Incidence of Invasive Fungal Infections by Underlying Malignancy and Level of Proof
Underlying malignancies

Patients with cancer, n (%)

Patients with IFI, n (%)
Proven or probable, n
Possible, n

Myeloid
leukemia

Lymphoid
leukemia

66 (6.3%)
26 (39.4%)
9
17

205 (19.6%)
24 (11.2%)
12
12

BMT

Lymphoma

Solid
tumors

Histiocytosis

Total

Auto

Allo

120 (11.5%)
3 (2.5%)
1
2

609 (58.2%)
16 (2.6%)
10
6

47 (4.5%)
1 (2.1%)
1
0

1047 (100%)
70a
33
37

201 (60%)
13 (6.5%)
8
5

130 (40%)
10 (7.7%)
4
6

BMT, bone marrow transplant; Allo, allogeneic; Auto, autologous; IFI, invasive fungal infection. aFive additional patients had aplastic anemia
and were not included in this analysis.
Pediatr Blood Cancer DOI 10.1002/pbc

Invasive Fungal Infections in Pediatric Oncology

1095

TABLE III. Classication, Etiology, and Sites of Fungal Infection

Etiology

Number of isolates

Proven yeast/mold infections (n 16)

Yeasts
Candida spp.
C. albicans
C. parapsilosis
C. krusei
C. tropicalis
C. glabrata
Molds
Fusarium sp.
Probable mold infections (n 18)a
Aspergillus spp.

1
6

Blood (fungemia) 1
Blood (fungemia) 5
Spleen, liver, kidney, eyes (with fungemia) 1
Endocarditis (with fungemia) 1
Blood (fungemia) 1
Blood (fungemia) 4
Knee joint (with fungemia) 1
Blood (fungemia) 1

Joints, skin 1

Lung 2
Paranasal sinuses 1
Paranasal sinuses 2
Lung 1
Paranasal sinuses, lung 1
Lung, spleen, liver 1
Paranasal sinuses 1
Lung 1
Paranasal sinuses 4
Lung, paranasal sinuses 1
Paranasal sinuses 1
Paranasal sinuses 2
Lung, paranasal sinuses 2
Paranasal sinuses 1

1
1
5

A. flavus

A. niger

A. fumigatus
Exserohilum
Mucor
Absidia
Fusarium spp.

1
4
1
1
4

Dematiceous fungi
Possible mold infections (n 41)

Site of infection

Paranasal sinuses 6
Lung 27
Lung, paranasal sinuses 3
Lung, liver 1
Lung, brain 1
Lung, kidney 1
Lung, spleen 1
Lung, kidney, spleen 1
a

Three patients had more than one isolated pathogen.

of mold infections was much higher in those with leukemia (AML

21/66, 31.8%; ALL 23/205, 11.2%) than in those with solid
tumors (9/609, 1.5%).
Fifty-ve patients (91.7%) presented with fever. The main
primary sites of involvement were the lungs (42 patients, 70%)
and the sinuses, nose, and palate (22 patients, 36.7%) (Table I).
Some patients had more than one site of involvement at the time
of diagnosis (Table III). The main predisposing factors for mold
infection were corticosteroid therapy at the time of infection (34
patients, 57%), severe neutropenia (55 patients, 91.7%), and
severe mucositis (27 patients, 45%). Forty-six patients (76.7%)
had an ITR of 4 (highest intensity) (Table I).
Cultures were positive for mold in 18 patients (30%). The
main pathogens were Aspergillus (10 patients), Exserohilum (4
patients), Fusarium (5 patients), and Zygomyces (2 patients), isolated from cultures of BAL uid (5 patients), sinus and nasal
lesions (12 patients), and knee joint uid (1 patient).
Until 2003, 46 patients (76.7%) were treated initially with
amphotericin B. Thereafter, voriconazole became the drug of
choice for presumed or highly suspected invasive aspergillosis
based on typical CT ndings, and seven patients (11.7%) received
Pediatr Blood Cancer DOI 10.1002/pbc

voriconazole as primary treatment. Additional rst-line drugs

were liposomal amphotericin B (5 patients), amphotericin B colloidal dispersion (1 patient), and itraconazole (1 patient). In 32
cases (53.3%), the primary drug was later changed because of
toxicity (17 patients, 28.3%) or culture results (13 patients,
21.7%); in two patients, it was changed because of failure to
control the infectious process. Another antifungal agent was
added to the primary drug in 12 cases, and a third antifungal
agent was prescribed in 13, due to clinical failure (n 7),
toxicity (n 3), or organism drug-susceptibility ndings
(n 3). Thirteen patients died of mold infections, of whom seven
had a probable diagnosis and six, a possible diagnosis. The crude
mortality rate within 3 months of diagnosis was 21.7%. The
attributable mortality rate was 10%.

Mold Infections After HSCT

Of the 19 patients after HSCT in this group (5.7%), 12 underwent allogeneic transplantation (6% of this group), and 7, autologous HSCT (5.3% of this group). Seven (36.8%) had a probable
diagnosis and 12 (63.2%), a possible diagnosis; 7 (36.8%)

1096

Mor et al.

acquired the mold infection within 3 months of transplantation

and 12 (63.2%) did so later. There was no difference between the
patients after HSCT and the other patients with mold infection in
demographic or clinical parameters. However, their mortality rate
within 3 months of infection was signicantly higher: 42.1%
versus 21.7%. Their attributable death rate was 26.3%.

DISCUSSION
The present study revealed a 7.2% incidence of IFI over a 9year period in a hematology/oncology department of a major
pediatric tertiary medical center. The rate of invasive yeast infections was 1.4%, and of mold infections, 5.7%. The main underlying diseases were AML (24/75, 32%) and ALL (24/75, 32%).
Profound neutropenia and high ITR were the most common predisposing factors, found in 80% and 60%, respectively, of patients
with yeast infection and 91.7% and 76.7%, respectively, of
patients with mold infection. Epidemiologically, there is a trend
toward an increase in non-albicans etiologies (60% of isolates) in
cases of invasive candidiasis and an emergence of non-Aspergillus
mold infections (55% of isolates). The crude and attributable
mortality rates were lower than previously described (21% and
12%, respectively).
It is difcult to compare the incidence of IFI among studies
owing to differences in inclusion and exclusion criteria and in
denitions. Other recent studies in children with cancer and after
HSCT combined yeast and mold infections. However, we believe
they warrant separate analyses because they differ in incidence,
organs involved, outcome, and prophylactic and denitive treatment. They also differ in the degree of diagnostic accuracy: Yeast
infections are usually proven by isolation of the culprit pathogen
in cultures from blood or other normally sterile sites, whereas in
mold infections, isolation of the pathogen occurs less often, and
the diagnosis is proven or probable in only 3050% of cases.
The 7.2% incidence of IFIs in the pediatric hematology-oncology patients in our study is close to the peak of 7.8% reported in
the last years of the 11-year study (19912001) of Rosen et al. [6]
in a similar population. However, they identied Candida as the
pathogen in 69% of cases, whereas in our study, it was responsible
for only 20% of cases. Interestingly, Rosen et al. [6] noted a
temporal trend of a decrease in Candida infections concomitant
with an increase in Aspergillus infections. This may have been
reected in our study (19982006) which overlapped the later part
of theirs.
Castagnola et al. [7], in a 2-year study published in 2006,
prospectively evaluated 96 fungal infections dened according
to the EORTC denitions. Their rate of proven infections was
very high (44%), including 27% fungemias and 17% with deep
organ involvement, mainly the lung. Most of the blood infections
were due to yeasts, and two were due to unidentied lamentous
fungi. Their rate of probable IFIs was 18%, and of possible IFIs,
38%. Our respective rates of probable and possible IFIs were 24%
and 55%.
Both Rosen et al. [6] and Castagnola et al. [7] noted a change
over time in the Candida species isolated, from C. albicans to
other species [6,7,15]. In our study, six of the Candida isolates
(40%) were C. albicans and nine (60%) were non-albicans,
mainly C. tropicalis (33.3% of Candida isolates).
The main underlying diseases in our study were AML and
ALL, which were signicantly associated with an increased risk
Pediatr Blood Cancer DOI 10.1002/pbc

for fungal infections. These data are in agreement with previous

reports [9,13,14].
Severe and prolonged neutropenia is a well-known risk factor
for fungal infections. It was noted in 62% of the patients in the
study of Rosen et al. [6] and 77% of patients in the study of
Castagnola et al. [7]. In our study, the rate of neutropenia at the
time of diagnosis of fungal infection was even higher (67/75,
89.3%). This nding is in line with the high ITR: 93.3% (56/
60) of our patients with mold infection and 86.7% (13/15) with
yeast infection had ITR levels of 3 or 4.
The pathogens causing mold infections in the earlier studies
were mainly Aspergillus spp. and Zygomyces [68]. In our study,
there were 10 cases of Aspergillus infection and 2 of Zygomyces
infection, in addition to infections with Fusarium spp. (n 5),
Exserohilum (n 4), and dematiaceous fungi (n 1). We
believe our ndings better represent the diverse etiologies of
invasive mold infections and the current emergence of non-Aspergillus mold infections in this pediatric immune-suppressed population. A recent increase in the incidence of Zygomyces infection
was noted by Pongas et al. [22] who related it to an increased use
of antifungal prophylaxis.
Castagnola et al. [7] found that most IFIs in hematopoietic
stem-cell transplant recipients occurred shortly after transplantation: 54% of the episodes before day 30 and 27% after
day 100 [7]. In our study, 60% of the invasive Candida infections
occurred within 3 months of transplantation (most of them within
30 days). In patients with invasive mold infection, however,
63.2% of patients were diagnosed more than 3 months after
HSCT. This corresponds to the bimodal distribution of IFIs after
HSCT described in adults (i.e., Candida infections earlier and
Aspergillus later) [23]. A recent multicenter study of fungal infections in hematopoietic stem-cell transplant recipients further
delineates this bimodal distribution [24]. In that study, the median
time of onset of candidiasis and aspergillosis after HSCT was 61
and 99 days, respectively. Aspergillus infections (43%) were more
prevalent than Candida (28%) infections [24], which also corresponds to our ndings.
In the present study, mortality due to Candida infection was
20% (3/15), and to mold infection, 21.7% (13/60). In the proven/
probable mold infection group, the mortality was higher (7/19,
36.8%). These rates are lower than those reported by Kobayashi
et al. [8] in children (48%) but similar to the rate reported more
recently by Castagnola et al. (28%) [7]. They are also lower than
the mortality rates for adults with IFI, which may reach 60% [3].
This may be attributable to greater effectiveness of the newer
antifungal agents and early empiric treatment for febrile neutropenic patients.
Our study has several limitations. The retrospective singlecenter design precludes generalization of the ndings to other
areas and populations. However, this limitation is partly mitigated
by the similarity of treatment protocols and intensity, both major
risk factors for IFI, in many countries. Moreover, this study was
conducted over a 9-year period, during which new antifungal
drugs were added to our therapeutic armamentarium, and prophylactic protocols for high-risk patients were adopted. The present
study was not intended to compare outcomes according to treatment protocols, so no conclusions pertaining to this issue were
reached.
To conclude, IFIs continue to be a major cause of morbidity
and mortality in children with cancer. Yeast and mold infections

Invasive Fungal Infections in Pediatric Oncology

differ signicantly in incidence, organs involved, outcome, and
treatment. Non-albicans Candida species and non-Aspergillus
mold species are emerging as important causes of IFI in this
population. This may have important implications for empiric
and prophylactic treatment. Mortality due to IFIs appears to be
lower than previously reported in children and adults. Larger,
prospective, multicenter studies using uniform denitions,
inclusion criteria, and prophylaxis and treatment protocols are
needed to develop better strategies for prevention, early detection,
and treatment of IFIs in this patient population.

ACKNOWLEDGMENT
The study was supported by an unrestricted research grant
from Gilead Sciences Research Foundation, Foster City, CA.

REFERENCES
1. Slobbe L, Polinder S, Doorduijn JK, et al. Outcome and medical
costs of patients with invasive aspergillosis and acute myelogenous leukemia-myelodysplastic syndrome treated with intensive
chemotherapy: An observational study. Clin Infect Dis 2008;47:
15071512.
2. Viudes A, Peman J, Canton E, et al. Candidemia at a tertiarycare hospital: Epidemiology, treatment, clinical outcome and risk
factors for death. Eur J Clin Microbiol Infect Dis 2002;21:767
774.
3. Nivoix Y, Velten M, Letscher-Bru V, et al. Factors associated
with overall and attributable mortality in invasive aspergillosis.
Clin Infect Dis 2008;47:11761184.
4. Zaoutis TE, Argon J, Chu J, et al. The epidemiology and attributable outcomes of candidemia in adults and children hospitalized
in the United States: A propensity analysis. Clin Infect Dis 2005;
41:12321239.
5. Pappas PG, Pappas PG, Rex JH, et al. NIAID Mycoses Study
Group. A prospective observational study of candidemia: Epidemiology, therapy, and inuences on mortality in hospitalized
adult and pediatric patients. Clin Infect Dis 2003;37:634643.
6. Rosen GP, Nielsen K, Glenn S, et al. Invasive fungal infections
in pediatric oncology patients. 11-year experience at a single
institution. J Pediatr Hematol Oncol 2005;27:135140.
7. Castagnola E, Cesaro S, Giacchino M, et al. Fungal infections in
children with cancer. A prospective, multicenter surveillance
study. Pediatr Infect Dis J 2006;25:634639.
8. Kobayashi R, Kaneda M, Sato T, et al. The clinical feature of
invasive fungal infection in pediatric patients with hematologic
and malignant diseases. A 10-year analysis at a single institution
at Japan. J Pediatr Hematol Oncol 2008;30:886890.
9. Abbassi S, Shenep J, Hughes W. Aspergillosis in children with
cancer: A 34-year experience. Clinical Infect Sis 1999;29:1210
1219.

Pediatr Blood Cancer DOI 10.1002/pbc

1097

10. Kaya Z, Gursel T, Kocak U, et al. Invasive fungal infections in

pediatric leukemia patients receiving uconazole prophylaxis.
Pediatr Blood Cancer 2009;52:470475.
11. Hale KA, Shaw PJ, Dalla-Pozza L, et al. Epidemiology of paediatric invasive fungal infections and a case-control study of risk
factors in acute leukaemia or post stem cell transplant. Br J
Haematol 2010;149:263272.
12. Sung L, Lange B, Gerbing R, et al. Microbiologically documented infections and infections-related mortality in children
with acute myeloid leukemia. Blood 2007;110:35323539.
13. Burgos A, Zaoutis T, Dvorak C, et al. Pediatric invasive aspergillosis: A multicenter prospective analysis of 139 contemporary
cases. Pediatrics 2008;101:e1286e1294.
14. Zaoutis T, Heydon K, Chu J, et al. Epidemiology, outcomes and
costs of invasive aspergillosis in immunocompromised children
in the United States, 2000. Pediatrics 2006;117:e711e715.
15. Zaoutis TE, Greves HM, Lautenbach E, et al. Risk factors for
disseminated candidiasis in children with candidemia. Pediatr
Infect Dis 2004;23:635641.
16. Velasco E, Bigni R. A prospective cohort study evaluating the
prognostic impact of clinical characteristics and comorbid conditions of hospitalized adult and pediatric cancer patients with
candidemia. Eur J Clin Microbiol Infect Dis 2008;27:10711078.
17. Krupova Y, Sejnova D, Dzatkova J, et al. Prospective study on
fungemia in childrem with cancer: Analysis of 35 cases and
comparison with 130 fungemias in adults. Support Care Cancer
2000;8:427430.
18. Dornbush HJ, Manzoni P, Roilides E, et al. Invasive fungal
infections in children. Pediatr Infect Dis 2009;28:734737.
19. Werba BE, Hobbie W, Kazak AE, et al. Classifying the intensity
of pediatric cancer treatment protocols: The intensity of treatment rating scale 2.0 (ITR-2). Pediatr Blood Cancer 2007;48:
673677.
20. De Paw B, Walsh TJ, Peter Donnelly J, et al. Revised denitions
of invasive fungal disease from the European Organization for
Research and Treatment of Cancer/Invasive Fungal Infections
Cooperative Group and The National Institute of Allergy and
Infectious Diseases Mycoses Study Group (EORTC/MSG) consensus group. Clin Infect Dis 2008;46:18131821.
21. Williams DA. Tests for differences between several small proportions. Appl Stat 1988;37:421434.
22. Pongas GN, Lewis RE, Samonis G, et al. Voriconazole-associated zygomycosis: A signicant consequence of evolving antifungal prophylaxis and immunosuppression practices. Clin
Microbiol Infect 2009;15:9397.
23. Marr KA. Fungal infections in hematopoietic stem cell transplant
recipients. Med Mycol 2008;46:293302.
24. Kontoyiannis DP, Marr KA, Park BJ, et al. Prospective surveillance for invasive fungal infections in hematopoietic stem cell
transplant recipients, 20012006: Overview of the transplantassociated infection surveillance network (TRANSNET) database. Clin Infect Dis 2010;50:10911100.