ORIGINAL ARTICLE MYCOLOGY

Diagnosis of chronic disseminated candidosis from liver biopsies

by a novel PCR in patients with haematological malignancies

M. Fleischhacker1, S. Schulz1, K. Johrens2, M. von Lilienfeld-Toal3, T. Held4, E. Fietze5, C. Schewe2, I. Petersen6 and
M. Ruhnke1
1) Medizinische Klinik und Poliklinik II, Charite Campus Mitte der Humboldt-Universitat zu Berlin, Berlin, 2) Institut fur Pathologie, Charite Campus Mitte,
Universitatsmedizin Berlin, Berlin, 3) Medizinische Klinik und Poliklinik III, Uniklinikum Bonn, Bonn, 4) Klinik fur Hamatologie, Onkologie und Tumorimmuno-
logie HELIOS Klinikum Berlin-Buch Klinikum Buch Klinik fur Hamatologie Onkologie und Tumorimmonologie, Schwanebecker Chaussee, Berlin,
5) Gemeinschaftspraxis fur Pathologie am Bundeswehrkrankenhaus, Scharnhorststr, Berlin and 6) Institut fur Pathologie, Universitatsklinikum Jena,
Ziegelmuhlenweg, Jena, Germany

Abstract

Hepatic Candida infection (HCI; known as chronic disseminated candidosis or CDC) is a distinct form of disseminated Candida infection
with predominant involvement of the liver. Diagnosis of HCI is usually made on clinical suspicion together with multiple lesions in liver
on ultrasound (US), CT and/or MRI scan. Fungal elements may not always be visible in liver tissue and mycological culture is frequently
negative, making the evidence for proven fungal disease difcult. We studied a novel commercially available low-cost and density-array
(LCD) chip technique for a molecular diagnosis of HCI. This is a two-step procedure with PCR amplication after DNA extraction fol-
lowed by hybridization on a small chip provided by the manufacturer (Fungi 2.1, Chipron GmbH). The analysis of DNA from 45 fungal
control strains showed an excellent specicity and sensitivity. The DNA from 11 liver biopsies of patients with haematological malignan-
cies suffering from CDC was analysed on the LCD chip and overall 11 fungal pathogens could be detected in eight liver biopsies, sup-
porting the clinical diagnosis of HCI/CDC. Analysis of liver biopsies from controls was negative for fungal DNA in all samples studied.
In conclusion, the novel LCD chip technique examined in our study was able to detect fungal pathogens in liver biopsies from patients
with haematological malignancies and suspected HCI/CDC but was negative in control biopsies.

Keywords: Candida, candidosis, CDC, diagnosis, fungal, haematological malignancies, liver, molecular, PCR
Original Submission: 12 July 2011; Revised Submission: 25 October 2011; Accepted: 30 October 2011
Editor: E. Roilides
Article published online: 3 November 2011
Clin Microbiol Infect 2012; 18: 10101016
10.1111/j.1469-0691.2011.03713.x
nated Candida infection. When both liver and spleen are
Corresponding author: M. Ruhnke, Medizinische Klinik und
involved, the term hepatosplenic candidosis (HSC) is used. HCI
Poliklinik II, Charite Campus Mitte der Humboldt-Universitat zu
Berlin, Chariteplatz 1, 10117 Berlin, Germany occurring in immunocompromised patients is typically associ-
E-mail: markus.ruhnke@charite.de ated with disseminated infection involving multiple organs.
Data were presented in part at the 50th Interscience Conference on Since rst reports in the early 1960s, HCI is recognized
Antimicrobial Agents and Chemotherapy (ICAAC), Boston, 12-15 increasingly as a complication of patients with acute leukaemia
September 2010.
and other haematological malignancies treated with chemo-
therapy [1]. Although more than 40 years have passed since
the rst patient was described, the optimal diagnosis, as well
as management of this condition, is not well established. Diag-
Introduction
nosis of HCI is usually made on clinical suspicion (after recov-
ery from prolonged granulocytopenia with persistent fever,
Hepatic Candida infection (HCI), also known as chronic anorexia and elevation of alkaline phosphatase) together with
disseminated candidosis (CDC), is a distinct form of dissemi- multiple lesions (hypodensities) in liver on ultrasound, com-

2011 The Authors

Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases
CMI Fleischhacker et al.
puter tomography (CT) scan and/or magnetic resonance imag-
ing (MRI) [2]. Among the imaging modalities, MRI has the high-
est diagnostic level as a non-invasive tool for the detection of
CDC/HCI but will not lead to proven diagnosis [3]. Proven
HCI requires a positive histology plus cultural evidence for
fungi from a (sterile obtained) liver biopsy [4].
However, a liver biopsy may not always establish the de-
nite diagnosis or some patients may be ineligible for the pro-
cedure. In addition, fungal elements may not always be
visible in liver tissue and mycological culture is frequently
negative. This makes the evidence for proven fungal disease
difcult [57]. Fungi are usually microscopically visible in
organ biopsies, but it has been often observed that yeast
forms and pseudohyphae are seen only in the central area of
the abscess. The abscess may be even missed if the liver
biopsy is not taken targeted by laparoscopy [8].
The role of other non-cultural diagnostic techniques (e.g.
PCR) for detection of fungi in tissue samples has been stud-
ied primarily for mould infection in the lungs but not for
yeast infection in the liver [912]. In this study, we used a
novel commercially available low-cost and density-array
(LCD) chip technique for studying the role of this molecular
tool in diagnosing HCI.
Materials and Methods
Retrospectively from 2001 to 2010, we identied 12 patients
from hospital charts who had a liver biopsy performed because
of suspected HCI. In 11 patients, liver biopsies were taken
guided by ultrasound. One patient who died underwent an
autopsy. One patient (patient 11) nally was diagnosed as hav-
ing bacterial liver abscesses and served as a control patient. The
remaining 11 patients (seven female/four male) had a mean age
of 48 years (range 1881 years) and received chemotherapy
for haematological malignancies (nine acute leukaemias (eight
acute myeloid leukaemia (AML)/one acute lymphocytic leukae-
mia (ALL)), one chronic myeloic leukaemia (CML) and one
non-Hodgkin lymphoma (NHL). HCI was clinically suggested by
the appearance of new multiple hypodense lesions in liver (and
spleen) assessed by radiological examination (CT, ultrasound
and/or MRI) and clinical signs and symptoms suggestive of HCI/
CDC. Further clinical information is given in Table 3.
In addition, 18 liver biopsies from patients without signs
and symptoms of HCI/CDC who underwent diagnostic liver
biopsy (e.g. after liver transplantation) were analysed and
served as controls.
Biopsies were cut into three parts. Two parts were trans-
ferred, each in 2 mL isotonic sodium chloride solution, for
microbiological culture and molecular diagnostics. Samples for
Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 10101016
Molecular diagnosis of hepatic candidosis 1011
molecular diagnostics were stored at )80C until used. The
third part was xed in neutral buffered formalin and embed-
ded in parafn for histological work-up. Histological features
that are regarded as suggestive of hepatosplenic candidosis
were analysed, including changes in hepatic architecture, pres-
ence of granulomas, portal and parenchymal inammation,
ductal proliferation, sinusoidal dilatation, necrosis, brosis and
fatty change [13,14]. Inammatory foci containing fungal
organisms were documented. In addition, biopsies were analy-
sed microbiologically using standard culture techniques.
DNA isolation from parafn-embedded tissue and reference
strains
Parafn was removed by extraction (twice) with xylene and
centrifugation. The tissue was rehydrated by incubation in etha-
nol (100%, 96%, 75%). The DNA was isolated with the Fast-
DNA Spin Kit (MP Biomedicals, Illkirch Cedex, France)
according to the instructions of the supplier with minor modi-
cations. The samples were homogenized in the FastPrep
instrument three times for 30 s at a speed setting of 55 with
cooling intervals of 1 min on ice. Finally, the DNA was eluted
from the binding matrix with 100 lL DES (part of the Fast-
DNA Spin Kit) and stored at )20C until used.
From 45 fungal control strains, DNA was isolated using a
standard procedure as described earlier using zymolase diges-
tion following phenol-chloroform extractions and ethanol pre-
cipitation and the DNA was stored at )20C until used [15].
PCR, gelelectrophoresis and chip hybridization
After DNA extraction, the DNA is amplied by PCR with
primers supplied with the kit and hybridized onto the LCD
chip. The DNA was amplied in a total volume of 25 lL
using Taq polymerase (Platinum Taq, Invitrogen GmbH,
Karlsruhe, Germany) according to the instructions of the
manufacturer of the LCD chip (Fungi 2.1; Chipron GmbH,
Berlin, Germany). The cycling conditions were set to 3 min
at 95C (initial denaturation), 30 s at 94C, 45 s at 56C,
45 s at 72C (3545 repetitions for amplication), 3 min at
72C for nal extension, and cooling at 4C, and a MJ
Research PTC-100 cycler was used. The hands-on time for
16 samples is <2 h. An aliquot of 7 lL of the PCR products
was run on a 2% agarose gel. In a second step, DNA was
spotted manually on LCD chips and hybridization was per-
formed according to the instruction manual. According to
the manufacturer, the array is able to discriminate between
25 different fungal species or species clusters, such as
Candida albicans, Candida glabrata, Candida tropicalis, Candida
parapsilosis, Candida krusei, Candida dubliniensis, Candida guillier-
mondii, Candida pelliculosa, Candida lusitaniae, Candida lambica,
Candida kefyr, Aspergillus niger complex, Aspergillus fumigatus,
2011 The Authors
1012 Clinical Microbiology and Infection, Volume 18 Number 10, October 2012 CMI

Aspergillus avus, Aspergillus terreus, Aspergillus nidulans, Mucor
spp., Rhizomucor pusillus, Rhizomucor oryzae/Rhizomucor ar-
rhizus, Rhizomucor azygosporus/Rhizomucor microspores, Rhizo-
mucor stolonifer, Cryptococcus neoformans, Paecilomyces variotii,
Scedosporium prolicans and Lichtheimia (Absidia) corymbifera.
For scanning and nal analysis a combination of a transmis-
sion light scanner and image analysis software supplied by
the manufacturer was used.
Results
Identication of reference strains according to the prede-
ned array format was correct in all isolates tested
(Table 1). In addition, various other fungal strains that should
not be detected by the predened array format were nega-
tive by chip hybridization (Table 2).
All 12 patients suffered from active haematological malig-
nancies (nine AML/one ALL, one NHL, one allogeneic bone
marrow transplantation for CML) and had radiological signs
highly suggestive of HCI showing multiple new lesions in the
liver by abdominal ultrasound, which was conrmed by
biphasic spiral liver computed tomography CT (ten patients)
or MRI (two patients) (Table 3).
Hepatic lesions showed the well-known abscess-like pat-
tern previously described [16]. Many patients had a history of
a previous fungal disease, mostly proven/probable invasive pul-
monary aspergillosis (IPA), in six individuals, and possible IPA,
in one patient. Hepatic lesions developed while these patients
were treated for pulmonary aspergillosis with broad-spectrum

TABLE 1. Identication of refer-

Control strain PCR result Chip hybridization Identication
ence strains according to a prede-
Lichtheimia (Absidia) corymbifera (IP 1280.81) Strong band Abidia corymbifera Correct ned array format
Aspergillus avus (ATCC 9643) Strong band A. avus Correct
Aspergillus fumigatus (DSM 819) Strong band A. fumigatus Correct
Aspergillus fumigatus (DSM 819) Strong band A. fumigatus Correct
Aspergillus nidulans (RKI 491/06) Strong band A. nidulans Correct
Aspergillus niger (clinical isolate) Strong band A. niger Correct
Aspergillus terreus (RKI 83/06) Strong band A. terreus Correct
Aspergillus terreus (RKI 533/02) Strong band A. terreus Correct
Aspergillus versicolor (RKI 826/02) Strong band A. versicolor Correct
Candida albicans (ATCC 90028) Strong band C. albicans Correct
Candida dubliniensis (CBS 8500) Strong band C. dubliniensis Correct
Candida glabrata (ATCC 90030) Strong band C. glabrata Correct
Candida guilliermondii (ATCC 90877) Strong band Pichia guilliermondii Correct
Candida kefyr (DSM 11954) Strong band Kluyveromyces marxianus Correct
Candida krusei (DSM 11956) Very weak band Issatchenkia orientalis Correct
Candida parapsilosis (ATCC 22019) Strong band C. parapsilosis Correct
Candida orthopsilosis (GHP 3540) Strong band C. parapsilosis Correct
Candida metapsilosis (GHP 3541) Strong band C. parapsilosis Correct
Candida parapsilosis (DSM 11955) Strong band C. parapsilosis Correct
Candida tropicalis (ATCC 28707) Strong band C. tropicalis Correct
Candida tropicalis (ATCC 750) Strong band C. tropicalis Correct
Cryptococcus neoformans (ATCC 62066) Strong band Crytococcus neoformans Correct
Mucor rouxii (IP 1248.80) Strong band Mucor Correct
Paecilomyces variotii (RKI 498/00) Strong band Paecilomyces variotii Correct
Rhizopus microsporus var. rhizopodi for. (IP 676.72) Strong band Rhizopus Correct
Rhizopus microsporus (IP 1123.75) Strong band Rhizopus Correct
Rhizomucor pusillus (IP 1127.75) Weak band Rhizomucor Correct
Rhizopus oryzae (IP 1443.83) Strong band Rhizopus Correct

TABLE 2. Control strains not

Chip
Control strain PCR result hybridization Identication detected by chip hybridization

Aspergillus alliaceusa Strong band No signal Not depicted on chip

Aureobasidium pullulansa Strong band No signal Not depicted on chip
Arthrobotrys oligosporaa Strong band No signal Not depicted on chip
Candida allociferiia Strong band No signal Not depicted on chip
Candida lipolyticab Strong band No signal Not depicted on chip
Cuninghamella bertholletiae (IP 1886.89) Strong band No signal Not depicted on chip
Fusarium oxysporum (RKI 592/07) Strong band No signal Not depicted on chip
Geotrichum candidum (clinical isolate) Weak band No signal Not depicted on chip
Penicillium purpurogenuma Strong band No signal Not depicted on chip
Pseudallescheria boydii (clinical isolate) Weak band No signal Not depicted on chip
Saccharomyces cerevisiae (ATCC 9763) Strong band No signal Not depicted on chip
Scopulariopsis brevicaulisa Strong band No signal Not depicted on chip
Syncephalastrum racemosum (IP 1541.84) Strong band No signal Not depicted on chip
Trichosporon asahii (GHP 3537)b Strong band No signal Not depicted on chip
Tritiratium oryzaea Strong band No signal Not depicted on chip
Ulcocladiuma Strong band No signal Not depicted on chip
a
Courtesy of H. Hof.
b
Courtesy of G. Haase.

2011 The Authors

Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 10101016
 CMI

TABLE 3. Clinical data from 12 liver biopsies obtained from patients with HCI (CDC) in correlation with PCR results

Antifungal
Age Fungus in Radiology therapy
(years)/ biopsy (PAS/ (CT/abdominal PCR from liver before
Patient sex Diagnosis Sample Histology Grocott stain) Fungal culture ultrasound) Clinical diagnosis tissue biopsy (days)

1 48/f AML Liver biopsy Granuloma, necrosis, scars Negative Negative Multiple lesions Proven IPA (lungs) Candida albicans 90
liver Possible HCI
2 81/m B-NHL Autopsy Multiple white lesions on n.d. at autopsy Candida albicans + Single hepatic lesions Proven IPA (lungs) Aspergillus fumigatus 6
liver surface; malignant Aspergillus spp. Possible HCI Candida albicans
lymphoma in liver
3 56/m CML/allo-BMT Liver biopsy Necrosis Negative Negative Multiple lesions Possible HCI Candida albicans 78
liver/spleen
4 19/m AML Liver biopsy Granuloma, cholostatic Negative Negative Multiple lesions Possible HCI Candida dubliniensis 420
inammation liver/spleen
5 28/f ALL Liver biopsy Granuloma, brosis Yeasts Negative Multiple lesions Proven HCI Candida albicans 21
liver
6 46/f AML Liver biopsy Fibrosis, scars, necrosis, Yeasts n.d. Multiple lesions Proven HCI Candida lipolytica* 60
fatty changes liver
7 57/f AML Liver biopsy Fibrosis, fatty changes Negative n.d. Multiple lesions liver Possible HCI Candida tropicalis, 92
Candida glabrata
8 44/f AML Liver biopsy Fibrosis Negative Negative Multiple lesions Probable IPA (lung) Candida albicans 35
Liver Possible HCI
Fleischhacker et al.

9 56/f AML Liver biopsy Fibrosis, necrosis, fatty Negative Negative Multiple lesions Probable IPA (lung) Negative 45
changes liver/spleen Possible HCI
10 26/f AML Liver biopsy Fibrosis, fatty changes Negative Negative Multiple lesions Probable IPA (lung) Candida lipolytica* 14
liver/spleen Possible HCI
11* 56/m AML Liver biopsy Granuloma, brosis, Negative Negative Multiple lesions Probable IPA (lung) Negative 21
purulent inammation (growth S. aureus) liver and suspected Possible HCI
Liver cirrhosis cirrhosis
12 71/m AML Liver biopsy Fibrosis, fatty changes Negative Negative Multiple lesions Possible IPA (lung) Candida albicans 90
liver/spleen Possible HCI

AML, acute myeloid leukaemia; ALL, acute lymphocytic leukaemia; B-NHL, B-cell non-Hodgkin lymphoma; CML, chronic myeloid leukaemia; allo-BMT, allogeneic bone marrow transplantation; HCI, hepatic Candida infection; IPA, invasive
pulmonary aspergillosis; *, control patient; n.d., not detected.
Molecular diagnosis of hepatic candidosis
1013

2011 The Authors

Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 10101016
1014 Clinical Microbiology and Infection, Volume 18 Number 10, October 2012
antifungal therapy (conventional amphotericin B, liposomal
amphotericin or voriconazole). All patients but one under-
went CT or ultrasound-guided liver biopsy. One individual
who died had an autopsy performed. Detection of yeasts with
pseudohyphae in liver tissue by periodic acid-Schiff reaction
staining succeeded in two patients (patients 5 and 6). Fungal
culture was negative in all but one biopsy (patient 2). In one
patient (patient 2), CT was performed 8 weeks before death
from autopsy-proven disseminated aspergillosis, showing
hepatic lesions of unclear aetiology. This patient suffered from
aggressive non-Hodgkin lymphoma and at autopsy clinically
silent recurrence of malignant lymphoma in the liver was
detected without further manifestations at other body sites.
This patient died from fulminant sepsis-like syndrome. Micro-
biological culture of the liver biopsy revealed growth of yeasts
and lamentous fungi (6 days of treatment with amphotericin
B before death). All other patients had received either empiri-
cal or prolonged pre-emptive antifungal therapy before the
biopsy was performed (14420 days, mean 87 days). In one
patient (patient 11), Staphyloccocus aureus was cultured from
the liver biopsy, but no fungi were detected and PCR was neg-
ative. This patient was regarded as having bacterial liver
abscesses and not HCI (control patient). Histological examina-
tion in this patient revealed clinically unsuspected liver cirrho-
sis together with granuloma, brosis and purulent
inammation, suggestive of bacterial infection.
With the LCD chip, overall 11 fungal pathogens could be
detected in eight liver biopsies, with six Candida albicans and
one Candida dubliniensis as single pathogens, and two mixed
infections, one with Candida albicans plus Aspergillus fumigatus
and one with Candida glabrata plus Candida tropicalis, respec-
tively. In two liver samples, PCR revealed a positive ampli-
cation product specic for a fungal pathogen. However, using
the LCD chip array a specic fungal pathogen could not be
identied. After additional sequencing of the PCR product,
Candida lipolytica was identied in two samples. This pathogen
will not be detected by the array and would be missed in tis-
sue samples where Candida lipolytica is the infecting pathogen.
We also analysed the DNA isolated from 18 liver biopsies
that had been taken from patients without evidence for HCI.
All 18 samples gave no signal after hybridization onto the
chip (data not shown) and were negative when examined
with a pan-fungal real-time assay.
Discussion
Chronic disseminated candidosis is a potentially life-threaten-
ing invasive fungal disease in patients with haematological
malignancies which requires systemic antifungal therapy.
2011 The Authors
Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 10101016
CMI
Diagnosis is primarily established on clinical ndings in immu-
nocompromised patients (predominantly with acute leukae-
mia) with persistent fever not responsive to conventional
antibiotics, together with clinical ndings, such as gastrointes-
tinal symptoms with hepatomegaly and laboratory signs (e.g.
elevated levels of alkaline phosphatase). Fever typically pre-
sents as recurrent fever that occurs after neutrophil recov-
ery and HCI is diagnosed when typical lesions in liver (and
sometimes spleen and/or other organs) can be seen on com-
puted tomography, ultrasound and/or magnetic resonance
imaging [3,5,17,18]. In the very rst reports on disseminated
Candida infections, which were called moniliasis until the
1960s, fungal infections were found predominantly in the
lungs, kidneys, heart and gastrointestinal tract but less fre-
quently in the liver [1]. Prevalence of hepatic involvement in
disseminated Candida infection varied depending on the publi-
cation but a mean prevalence of 14% has been suggested in
earlier reports [19]. Soon, it became clear that focal hepatic
Candida infection is a distinct clinical variant of disseminated
Candida infection (chronic disseminated candidosis, CDC) in
immunocompromised patients, mostly patients who have
received intense cytotoxic chemotherapy for acute leukaemia
[2,3,57,1922].
Lesions associated with HCI were described as granulo-
mas, microabscesses, centrilobular congestion, haemorrhagic
necrosis, bile stasis, inammatory parenchymal aggregates,
free yeasts in sinusoids and/or fatty changes [13,19].
With regard to the denite diagnosis of HCI, the rst
denition of the diagnosis of invasive fungal diseases (IFD) in
haematopoietic stem cell transplant recipients was estab-
lished for research purposes. Patients with typical lesions for
HCI on CT or ultrasound were regarded as probable IFD
without any need for mycological support. In the revised def-
initions, such cases are now classied as possible IFD [4]. In
earlier reports, when patients may or may not have received
Amphotericin B for suspected HCI, it has been suggested
that liver biopsies usually revealed yeast and/or hyphal forms,
but mycological cultures were frequently negative [5,19].
However, identiable Candida organisms may not be found in
all tissue specimens and only indirect signs of HCI may be
seen at histopathological examination [13].
The majority of cases in our study had a possible diagnosis
of HCI because fungal elements were not visible in the
biopsy but radiology was highly suggestive of fungal liver dis-
ease. Interestingly, many of the patients studied here had a
previous history of proven/probable invasive fungal disease in
the lungs (possible/probable aspergillosis) and have received
systemic antifungal therapy. Only in two patients, could
yeasts be detected in the liver sample. However, hepatic
Candida infection could not be proven by a positive culture
CMI Fleischhacker et al.
of the fungal organism in any of the cases. As outlined above,
all patients had received a substantial amount of systemic
antifungal therapy with Candida-active antifungal agents
before the liver biopsy was taken. This underlines the diag-
nostic dilemma when a tissue sample is taken under ongoing
antifungal therapy. It has been earlier discussed that laparo-
scopic liver biopsies may have higher yield in the diagnosis of
HCI because the procedure allows better sampling of hepatic
focal lesions [6,23]. Laparoscopy is nowadays rarely per-
formed in patients with haematological malignancies to estab-
lish the diagnosis of HCI because radiological diagnosis with
ultrasound, CT and/or MRI has markedly improved
[1618,2427]. In our patients, it cannot be excluded that a
laparoscopy-guided liver biopsy might have had a higher diag-
nostic yield. Anttila et al. reported that the diagnosis of HCI
was established signicantly more often when the biopsy was
targeted at white nodules than when targeted randomly or
at scars [6]. However, in this study many liver biopsies tar-
geted at white nodules had negative results at histological
and/or cytological examination as well.
Molecular techniques may improve diagnosis of invasive
fungal disease. Molecular methods for the identication of
Aspergillus spp and other pathogenic fungi in parafn wax
embedded tissues have been established [9,12]. In general, in
these assays DNA was amplied by PCR using pan-fungal
probes, and detected by Southern blot hybridization with a
probe specic for Aspergillus fumigatus and other fungi [9,12].
In our study, the specicity and sensitivity for detecting
fungal DNA from a large collection of control and reference
strains was good. All fungal isolates that are supposed to be
detectable by the LCD chip could be correctly identied
using reference strains. Furthermore, DNA from a large col-
lection of additional fungal control strains that should not be
identied could be amplied by PCR but not the nal hybrid-
ization step on the LCD chip. These additional fungal strains
may cause invasive fungal disease in immunocompromised
patients but would not be identied by the array, which lim-
its this method. Of interest, we found fungal DNA in two
liver samples from patients with possible CDC that could
not be further identied after nal chip hybridization. After
additional sequencing of the PCR product, Candida lipolytica
was identied. This rare Candida species (as well as, for
example, Trichosporon spp.) is not within the detection scope
of the array and may limit the test method. However, the
most common Candida and Aspergillus spp. as well as some
other pathogenic fungi could be identied correctly.
In a study on respiratory tract biopsies, two semi-nested
PCR assays for Aspergillus spp. and Zygomycetes offered a
reliable aetiological diagnosis that was superior to culture in
Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 10101016
Molecular diagnosis of hepatic candidosis 1015
patients with proven invasive mould infection [28]. Recently,
improvement has been achieved by establishing DNA micro-
arrays to detect and identify DNA from fungal pathogens,
mostly Candida and Aspergillus species [29,30]. In contrast to
the assay studied here, these microarray systems are not
commercially available but may detect up to 14 different fun-
gal pathogens from various samples (blood, bronchoalveolar
lavage and tissue) in high-risk patients [30].
In conclusion, the novel LCD chip technique examined in
our study was able to detect fungal pathogens in the majority
of liver biopsies from patients with suspected hepatic fungal
infection but was negative in control biopsies. Together with
the clinical information, detection of fungal DNA in liver
biopsies may help to improve establishing the diagnosis of
CDC even during antifungal therapy.
Acknowledgements
We thank Dr Kathrin Tintelnot (Robert-Koch-Institute/RKI
isolates, Berlin), Professor Herbert Hof (University Hospital
Mannheim), Professor Gerhard Haase (University Hospital
Aachen/GHP isolates) and Professor Olivier Lortholary
(Institute Pasteur/IP isolates, Paris, France) for providing
strains. We thank Dr Volker Heiser (Chipron GmbH, Berlin,
Germany) and Clarissa Radecke for technical assistance.
Transparency Declaration
All authors have substantially contributed to the manuscript
and can take public responsibility for the content. M. Ruhnke
has received research grants from Deutsche Krebshilfe/Ger-
man Cancer Aid, Pzer and Merck/MSD, served as a consul-
tant and at the speakers bureau of Astellas, Basilea, Essex/
Schering-Plough, Gilead Sciences, Janssen, Merck/MSD,
Novartis, Pzer, Pliva (all not related to the manuscript). M.
von Lilienfeld-Toal has received research grants from Merck/
MSD and served as a consultant and at the speakers bureau
of Merck/MSD and Wyeth. M. Fleischhacker, S. Schulz, K.
Johrens, T. Held, E. Fietze, C. Schewe, I. Petersen have noth-
ing to declare.
References
1. Louria DB, Stiff DD, Bennett B. Disseminated moniliasis in the adult.
Medicine (Baltimore) 1962; 41: 307338.
2011 The Authors
1016 Clinical Microbiology and Infection, Volume 18 Number 10, October 2012
2. Kontoyiannis DP, Luna MA, Samuels BI, Bodey GP. Hepatosplenic
candidiasis. A manifestation of chronic disseminated candidiasis. Infect
Dis Clin North Am 2000; 14: 721739.
3. Masood A, Sallah S. Chronic disseminated candidiasis in patients with
acute leukemia: emphasis on diagnostic denition and treatment. Leuk
Res 2005; 29: 493501.
4. dePauw B, Walsh TJ, Donnelly JP 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.
5. Tashjian LS, Abramson JS, Peacock JE Jr. Focal hepatic candidiasis: a
distinct clinical variant of candidiasis in immunocompromised patients.
Rev Infect Dis 1984; 6: 689703.
6. Anttila VJ, Elonen E, Nordling S, Sivonen A, Ruutu T, Ruutu P. Hepa-
tosplenic candidiasis in patients with acute leukemia: incidence and
prognostic implications. Clin Infect Dis 1997; 24: 375380.
7. Thaler M, Pastakia B, Shawker TH, OLeary T, Pizzo PA. Hepatic can-
didiasis in cancer patients: the evolving picture of the syndrome. Ann
Intern Med 1988; 108: 88100.
8. Anttila VJ, Farkkila M, Jansson SE et al. Diagnostic laparoscopy in
patients with acute leukemia and suspected hepatic candidiasis. Eur J
Clin Microbiol Infect Dis 1997; 16: 637643.
9. Kappe R, Okeke CN, Fauser C, Maiwald M, Sonntag HG. Molecular
probes for the detection of pathogenic fungi in the presence of
human tissue. J Med Microbiol 1998; 47: 811820.
10. Rickerts V, Just-Nubling G, Konrad F et al. Diagnosis of invasive
aspergillosis and mucormycosis in immunocompromised patients by
seminested PCR assay of tissue samples. Eur J Clin Microbiol Infect Dis
2006; 25: 813.
11. Hayden RT, Isotalo PA, Parrett T et al. In situ hybridization for the
differentiation of Aspergillus, Fusarium, and Pseudallescheria species
in tissue section. Diagn Mol Pathol 2003; 12: 2126.
12. Paterson PJ, Seaton S, McLaughlin J, Kibbler CC. Development of
molecular methods for the identication of aspergillus and emerging
moulds in parafn wax embedded tissue sections. Mol Pathol 2003;
56: 368370.
13. Johnson TL, Barnett JL, Appelman HD, Nostrant T. Candida hepatitis.
Histopathologic diagnosis. Am J Surg Pathol 1988; 12: 716720.
14. Procop GW, Wilson M. Infectious disease pathology. Clin Infect Dis
2001; 32: 15891601.
15. Schulz B, Weber K, Radecke C, Scheer C, Ruhnke M. Effect of differ-
ent sample volumes on the DNA extraction of Aspergillus fumigatus
from whole blood. Clin Microbiol Infect 2009.
16. Rudolph J, Rodenwaldt J, Ruhnke M, Hamm B, Kopka L. Unusual
enhancement pattern of liver lesions in hepatosplenic candidiasis. Acta
Radiol 2004; 45: 499503.
2011 The Authors
Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI, 18, 10101016
CMI
17. Shirkhoda A, Lopez-Berestein G, Holbert JM, Luna MA. Hepatosplen-
ic fungal infection: CT and pathologic evaluation after treatment with
liposomal amphotericin B. Radiology 1986; 159: 349353.
18. Semelka RC, Shoenut JP, Greenberg HM, Bow EJ. Detection of acute
and treated lesions of hepatosplenic candidiasis: comparison of
dynamic contrast-enhanced CT and MR imaging. J Magn Reson Imaging
1992; 2: 341345.
19. Lewis JH, Patel HR, Zimmerman HJ. The spectrum of hepatic candidi-
asis. Hepatology 1982; 2: 479487.
20. Haron E, Feld R, Tuffnell P, Patterson B, Hasselback R, Matlow A.
Hepatic candidiasis: an increasing problem in immunocompromised
patients. Am J Med 1987; 83: 1726.
21. Woolley I, Curtis D, Szer J et al. High dose cytosine arabinoside is a
major risk factor for the development of hepatosplenic candidiasis in
patients with leukemia. Leuk Lymphoma 1997; 6: 469474.
22. Chen CY, Chen YC, Tang JL et al. Hepatosplenic fungal infection in
patients with acute leukemia in Taiwan: incidence, treatment, and
prognosis. Ann Hematol 2003; 82: 9397.
23. Phillips EH, Carroll BJ, Chandra M et al. Laparoscopic-guided biopsy
for diagnosis of hepatic candidiasis. J Laparoendosc Surg 1992; 2: 33
38.
24. Karthaus M, Huebner G, Elser C, Geissler RG, Heil G, Ganser A.
Early detection of chronic disseminated Candida infection in leukemia
patients with febrile neutropenia: value of computer-assisted serial
ultrasound documentation. Ann Hematol 1998; 2: 4145.
25. Karthaus M, Huebner G, Geissler RG, Heil G, Ganser A. Hepatic
lesions of chronic disseminated systemic candidiasis in leukemia
patients may become visible during neutropenia: value of serial ultra-
sound examinations. Blood 1998; 91: 30873089.
26. Shirkhoda A. CT ndings in hepatosplenic and renal candidiasis.
J Comput Assist Tomogr 1987; 11: 795798.
27. Mortele KJ, Segatto E, Ros PR. The infected liver: radiologic-patho-
logic correlation. RadioGraphics 2004; 24: 937955.
28. Rickerts V, Mousset S, Lambrecht E et al. Comparison of histopatho-
logical analysis, culture, and polymerase chain reaction assays to
detect invasive mold infections from biopsy specimens. Clin Infect Dis
2007; 44: 10781083.
29. Leinberger DM, Schumacher U, Autenrieth IB, Bachmann TT. Devel-
opment of a DNA microarray for detection and identication of fun-
gal pathogens involved in invasive mycoses. J Clin Microbiol 2005; 43:
49434953.
30. Spiess B, Seifarth W, Hummel M et al. DNA microarray-based detec-
tion and identication of fungal pathogens in clinical samples from
neutropenic patients. J Clin Microbiol 2007; 45: 37433753.