Content uploaded by Tania C Sorrell
Author content
All content in this area was uploaded by Tania C Sorrell on Aug 07, 2016
Content may be subject to copyright.
Candidaemia in adult cancer patients: risks for fluconazole-resistant
isolates and death
Monica A. Slavin1*, Tania C. Sorrell2, Deborah Marriott3, Karin A. Thursky 1, Quoc Nguyen 3, David H. Ellis 4,
C. Orla Morrissey5and Sharon C.-A. Chen 2on behalf of the Australian Candidemia Study, Australasian Society
for Infectious Diseases†
1
Peter MacCallum Cancer Centre and the Centre for Research Excellence in Infectious Diseases, Royal Melbourne Hospital, Parkville,
VIC 3050, Australia;
2
Centre for Infectious Diseases and Microbiology and the University of Sydney, Camperdown, NSW 2006, Australia;
3
Department of Microbiology, St Vincent’s Hospital, Darlinghurst, NSW 2010, Australia;
4
The Mycology Unit, Women’s and Children’s
Hospital, Adelaide, SA 5006, Australia;
5
Infectious Diseases Unit, Department of Medicine, Alfred Hospital and Monash University,
Prahran, VIC 3004, Australia
*Corresponding author. Department of Infectious Diseases, Peter MacCallum Cancer Centre, St Andrew’s Place, East Melbourne, Victoria, Australia,
3002. Tel: þ61-396561707; Fax: þ61-396561185; E-mail: monica.slavin@petermac.org
†Members are listed in the Acknowledgements section.
Received 15 October 2009; returned 13 November 2009; revised 2 February 2010; accepted 2 February 2010
Background: Candidaemia in cancer patients is associated with increasing fluconazole resistance. Models for
predicting such isolates and their clinical impact are required.
Methods: Clinical, treatment and outcome data from a population-based candidaemia survey (2001–2004)
were collected at 5 and 30 days after diagnosis. Speciation and antifungal susceptibility testing was performed.
Results: There were 138 candidaemia episodes (33% Candida albicans) in adults with haematological malig-
nancies and 150 (51% C. albicans) in adults with solid organ malignancies. Thirty-nine isolates had fluconazole
MICs of 64 mg/L and 40 had MICs of 16 –32 mg/L (predominantly Candida glabrata and Candida krusei). By
multivariate analysis, triazole therapy, gastrointestinal tract (GIT) surgery in the 30 days before candidaemia
and age .65 years were predictive of fluconazole-resistant candidaemia. Thirty day crude mortality was
40% in haematology patients and 45% in oncology patients. Fluconazole-resistant isolates were associated
with increased risk of mortality by univariate (P¼0.04) and Kaplan–Meier survival analyses. By Cox proportional
hazards modelling, the strongest predictors of mortality at onset of candidaemia were invasive ventilation,
elevated creatinine, intensive care unit (ICU) admission and receipt of systemic triazoles or corticosteroids in
the previous 30 days. Removal of a central venous access device (CVAD) at or within 5 days of onset was associ-
ated with decreased mortality.
Conclusions: Risk factors for fluconazole-resistant candidaemia in adults with cancer include fluconazole/tria-
zole exposure and GIT surgery. ICU admission, invasive ventilation, renal impairment, age .65 years and prior
exposure to corticosteroids and triazoles are risk factors for death. CVAD removal reduced mortality. These find-
ings should be integrated into surveillance and treatment algorithms.
Keywords: mortality, Candida, triazoles, susceptibility
Introduction
Candidaemia is a complication of cancer and its treatment, and
results in mortality rates of 30%–50%, prolonged hospital stay
and substantial healthcare costs.
1–4
Recent studies of candidae-
mia in patients with haematological malignancy describe a shift
in epidemiology towards non-albicans Candida spp., particularly
Candida glabrata and Candida krusei.
4–7
This shift is clinically rel-
evant as these species are frequently fluconazole resistant.
8,9
There are fewer comprehensive data for patients with solid
tumours, although it appears that the species distribution is
more consistent with that of general hospital patients.
5,10
Such
trends have important implications for the selection of antifungal
therapy. Timely and appropriate antifungal therapy reduces mor-
tality
11,12
whereas fluconazole resistance has been associated
with increased mortality in general hospital patients with candi-
daemia.
13,14
Whether these factors are as important in cancer
patients is unknown.
15
#The Author 2010. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
For Permissions, please e-mail: journals.permissions@oxfordjournals.org
J Antimicrob Chemother 2010; 65: 1042–1051
doi:10.1093/jac/dkq053 Advance publication 4 March 2010
1042
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
We therefore analysed the data from a contemporary, nation-
wide study of candidaemia in all adult patients with cancer to
determine: first, the clinical risk factors for candidaemia due to
a fluconazole-resistant isolate at presentation; and, secondly,
whether bloodstream infection (BSI) with fluconazole-resistant
Candida adversely affects clinical outcome and/or is associated
with cross-resistance to other triazoles. These data are needed
to inform strategies for prevention, risk predictive models and
treatment guidelines.
Methods
Study design and data collection
Cases of candidaemia were prospectively identified by blood culture sur-
veillance at public and private microbiology laboratories over a 3 year
period (August 2001–July 2004) as previously described.
16
Approval for
the study was obtained from Human Ethics Review Committees at parti-
cipating institutions. Information collected from each case included age,
gender, co-morbidities and risk factors over the 30 days prior to onset of
candidaemia [recent surgery, central venous access devices (CVADs),
neutropenia, hyperalimentation, and administration of corticosteroids
and antimicrobial or systemic antifungal agents], clinical signs of
sepsis,
17
results of diagnostic studies, antifungal therapy and outcome.
Data were collected at 5 and 30 days following the date of the initial
positive blood culture.
Definitions
A case was the incident isolation of any Candida species from blood
during the study period. Adults were those aged .14 years. Relapses
were positive blood cultures within 30 days of the first positive blood
culture after an initial clinical and microbiological response. Polycandidal
candidaemia was an episode caused by more than one Candida species.
Haematological malignancy was classified according to the International
Statistical Classification of Diseases and Related Health Problems, 10th
revision, Australian modification.
18
Solid tumours were not further classi-
fied. Patients undergoing stem cell transplantation (SCT) were grouped by
their underlying condition: haematological malignancy or solid tumour.
Neutropenia was an absolute neutrophil count (ANC) of ,1.010
9
/L
and impaired renal function was a serum creatinine .0.17 mmol/L.
Sepsis and severe sepsis were according to international definitions.
17
Treatment was the antifungal drug received on the day of or after the
date the blood culture positive for Candida was drawn. Death was attrib-
uted to candidaemia, or not, by the treating physician, and overall mor-
tality determined at day 30 after diagnosis.
Microbiological methods
Candida isolates were speciated using standard phenotypic methods.
19,20
Candida albicans and Candidadubliniensis were differentiatedby PCR finger-
printing.
21
Isolates were forwarded to a reference laboratory (Women’s
and Children’s Hospital, Adelaide) for susceptibility testing as follows: for
amphotericin B, fluconazole, itraconazoleand voriconazole, the Sensititre
w
YeastOne
w
YO3 microdilution test panel (Trek Diagnostic Systems, Dutec
Diagnostics, Australia) was used. Posaconazole pure substance was
supplied by Schering-Plough (Schering-Plough Pty Ltd, North Ryde,
New South Wales, Australia). Tests were performed according to the
CLSI (formerly NCCLS) broth microdilution M27-A2 protocol
22
and
MICs were defined using CLSI methodology.
23
Resistance was defined
as: MIC64 mg/L [and susceptible dose-dependent (S-DD) as an MIC of
16–32 mg/L] for fluconazole; MIC1 mg/L for itraconazole; and
MIC4 mg/L for voriconazole.
21
For posaconazole, MICs of 2 mg/L were
interpreted as non-susceptible, although no breakpoints have been
defined.
23
Candida parapsilosis ATCC 22019 and C. krusei ATCC 6258 were
used as quality control strains.
22
Statistical analysis
Data were analysed using Stata software, version 9 (Statacorp, Texas).
Continuous variables were compared using Student’s t-test and pro-
portions compared using the x
2
or Fisher’s exact test. A Pvalue of
,0.05 was considered to be significant. Univariate analyses were per-
formed to identify risk factors associated with overall mortality and iso-
lation of a species with fluconazole resistance or S-DD. Candidate
variables with a univariate significance of P,0.15 or those found to be
significant in other publications were entered into the model for multi-
variate analysis using stepwise logistic regression. No adjustments were
made for multiple comparisons. The Kaplan–Meier method was used
to examine the probability of survival with fluconazole resistance. The
log rank test was used to test the null hypothesis between the two
groups. Cox proportional hazards regression with the day of onset of can-
didaemia as failure was also applied.
Results
Incidence and patient demographics
A total of 1095 incident cases of candidaemia were identified;
288 (26%) episodes occurred in 288 adults with cancer—138
with haematological malignancies (haematology patients) and
150 with solid organ malignancies (oncology patients). Demo-
graphic and clinical data were available for all 288 episodes,
and clinical and microbiological outcome data for 262/288
(91%) and 235/288 (82%), respectively. All percentages are of
number of episodes with data available. Species identity was
established for 280 (97%) isolates. There were no relapses and
69/288 episodes were breakthrough cases after prior antifungal
therapy. The median age of haematology and oncology patients
was 57 and 64 years, respectively (Table 1).
Co-morbidities and established risk factors
Haematology patients were more likely to have received che-
motherapy and corticosteroids, to have been neutropenic and to
have a CVAD in place at the time of candidaemia, but were less
likely to have undergone major surgery or to have received hyper-
alimentation than oncology patients. CVADs were removed more
often in oncology patients than in haematology patients.
Fifty-eight of 127 (45.7%) haematology patients had received
a systemic antifungal agent compared with 11/150 (7.3%)
(P,0.001) oncology patients (Table 1). Forty-four of 58 haema-
tology patients had received fluconazole prior to candidaemia
(73% for prophylaxis, 27% for treatment/empirical therapy).
The median daily dose of fluconazole was 200 mg. One or
more other antifungals was given in 18 haematology patients:
amphotericin B (n¼13); itraconazole (n¼9); voriconazole
(n¼5); posaconazole (n¼2); and caspofungin (n¼2). Eight of
the 11 oncology patients who received a systemic antifungal
agent prior to candidaemia received fluconazole.
Candida species and antifungal susceptibility (Figure 1)
C. albicans was the most common species in both haematology
patients (32.7% of episodes) and oncology patients (51% of
Fluconazole-resistant Candida in cancer
1043
JA
C
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
episodes). In haematology patients, C. albicans was followed in
frequency by C. parapsilosis (n¼27 episodes, 19.5%), C. krusei
(n¼23, 16.6%) and C. glabrata (n¼17, 12.3%), whilst in oncol-
ogy patients the second most common species was C. glabrata
(n¼29, 19.3%) and then C. parapsilosis (n¼24, 16%). C. krusei
candidaemia occurred in only one oncology patient. Additionally,
three haematology and four oncology patients had polycandidal
candidaemia, which included C. krusei in two and three of these
episodes, respectively.
No C. albicans isolate was resistant to triazoles. There were 39
fluconazole-resistant isolates (26 C. krusei,12C. glabrata and 1
C. parapsilosis) and 40 that tested S-DD (32 C. glabrata, 2 each of
C. krusei,C. dubliniensis and C. parapsilosis, and 1 each of
Candida guilliermondii and Candida rugosa). The one (of 60)
fluconazole-resistant C. parapsilosis isolate was cross-resistant
to itraconazole but not to voriconazole or posaconazole. Thirty-six
of 44 (82%) C. glabrata isolates that tested fluconazole resistant or
S-DD were cross-resistant to itraconazole and 4 (9%) were also
resistant to voriconazole. Twenty-eight of the 29 C. krusei isolates
which were tested were resistant (92%) or S-DD (8%) to flucona-
zole but none was cross-resistant to itraconazole or voriconazole.
Of four isolates each of C. parapsilosis,C. glabrata and C. krusei
tested for posaconazole susceptibility, none was resistant.
Candida species and relationship to prior antifungal
therapy
C. albicans caused 21% of episodes in the 58 haematology
patients receiving antifungal agents prior to onset of candidae-
mia and 41.5% in those who were not (P,0.01). For C. glabrata
the proportions were 16.1% and 7.3%, respectively (P¼0.02), for
C. krusei the proportions were 29.0% and 4.9%, respectively
Table 1. Demographics: adults with candidaemia in Australia 2001– 2004
Patient characteristic Haematological malignancy, N¼138 Solid tumour, N¼150
Median age (range) 57 (15–83) 64 (28–88)
Gender male, n(%) 77 (55.8) 87 (58.0)
Co-morbidities
diabetes, n(%) 13 (8.2) 23 (14.6)
ANC ,0.510
9
/L, n(%) 89 (64.4) 13 (8.7)
ANC ,0.110
9
/L, n(%) 79 (59.9) 7 (5.0)
Median duration of neutropenia (days) 10 5
Corticosteroid use, n(%) 43 (34.4) 108 (76.1)
Cytotoxic chemotherapy, n(%) 80 (71.4) 28 (25)
Presence of sepsis at diagnosis, n(%) 103 (83.1) 124 (87.9)
Diagnosis in ICU, n(%) 19 (13.8) 38 (25.3)
CVAD in place at diagnosis, n(%) 123 (97.6) 112 (77.2)
Median duration of CVAD insertion, days (range) 17 (2–625) 11 (1–446)
Use of antimicrobials, n(%) 121 (87.7) 125 (83.3)
3 antibiotic classes 51 (42.1) 47 (37.6)
4 antibiotic classes 66 (47.9) 78 (62.4)
Concomitant bacteraemia, n(%) 14 (10.1) 23 (15.3)
Systemic antifungals, n(%) 58 (45.7) 11 (7.7)
Surgery, n(%) 14 (10.1) 74 (49.3)
Surgical site infection, n(%) 4 (2.9) 41 (27.3)
TPN, n(%) 36 (26.1) 59 (39.3)
Mortality
died due to candidaemia, n(%) 17 (13.2) 19 (12.1)
30 day mortality, n(%) 44 (40) 61 (45)
ANC, absolute neutrophil count; ICU, intensive care unit; CVAD, central venous access device; TPN, total parenteral
nutrition.
Demographics were determined in the 30 days before onset of candidaemia unless otherwise specified.
Slavin et al.
1044
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
(P,0.001), and for C. parapsilosis the proportions were 29.0%
and 19.5%, respectively (P¼0.04). Duration of prior fluconazole
also affected the species recovered. Of those receiving ,6 days
of fluconazole before onset of candidaemia, 56% of episodes
were due to C. albicans and 11% to C. krusei, compared with
14% and 44%, respectively, for patients receiving at least
6 days of fluconazole before candidaemia. With other antifungal
agents, duration of prior treatment (,6 days versus 6 days) did
not affect the proportion of episodes due to C. albicans (25% and
22%, respectively). All 13 candidaemia episodes in patients
receiving fluconazole for .14 days before candidaemia were
due to a species other than C. albicans. Only 11 oncology
patients received prior systemic antifungals (8 were prescribed
fluconazole) at onset of candidaemia and there was no effect
of antifungal therapy on the causative species (data not
shown, P¼0.15).
Risk factors for infection with fluconazole-resistant
isolates
By univariate analysis, variables associated with the recovery of a
fluconazole-resistant Candida isolate included duration of neu-
tropenia (P¼0.02), prior receipt of any systemic antifungal drug
including a triazole and fluconazole (both P,0.001), acquisition
in the state of Victoria (P¼0.02) and prior gastrointestinal tract
(GIT) surgery (P¼0.04) (Table 2). Multivariate analysis revealed
that receipt of triazoles was the strongest independent predictor
for a fluconazole-resistant isolate [odds ratio (OR) 2.9, 95% con-
fidence interval (CI) 1.33–6.1; P¼0.007] followed by acquisition
of candidaemia in the state of Victoria (OR 2.0, 95% CI 1.1–
3.5; P¼0.02), recent GIT surgery (OR 2.0, 95% CI 1.1 – 4.2;
P¼0.05) and use of fluconazole within 14 days prior to onset
of candidaemia (OR 3.4, 95% CI 1.0–11.3; P¼0.05). There was
a trend towards significance for recovery of a fluconazole-
resistant Candida and age .65 years (OR 1.7, 95% CI 0.9 –3.1;
P¼0.09).
Mortality and risks for death
Forty-four (40%) haematology patients and 61 (45%) oncology
patients died within 30 days of onset of candidaemia, with
68% and 61% of these, respectively, within the first 10 days.
Variables associated with 30 day mortality by univariate analysis
are summarized in Table 3. The risk of death was reduced in
patients who had their CVAD removed within 5 days of onset of
candidaemia (P¼0.01) and in those treated with fluconazole
after diagnosis (P¼0.004). Table 4shows the risk factors for
death at day 30 in the Cox proportional hazards model, which
were: invasive ventilation at onset of candidaemia, receipt of a
systemic triazole in the 30 days prior to candidaemia, corticos-
teroids in the 30 days prior to candidaemia and onset of candi-
daemia in the intensive care unit (ICU). Removal of a CVAD
within 5 days of onset of candidaemia was protective, and this
is also shown in Figure 2(a) where the Kaplan–Meier survival
curve shows improved survival with catheter removal, log rank
P¼0.0096.
Antifungal treatment and outcome associated with
a fluconazole-resistant isolate
Ninety-five percent of haematology patients and 86% of oncol-
ogy patients with candidaemia were treated with an antifungal
after diagnosis (P¼0.01). Fluconazole was used in 22% of
Haem
Solid
Haem
Solid
Haem
Solid
Haem
Solid
Haem
Solid
Haem
Solid
Haem
Solid
Haem
Solid
020
MIC < 16 MIC 16–32 MIC ≥ 64 (mg/L)
40 60 80
%
C. albicans
C. dubliniensis
C. glabrata
C. krusei
C. parapsilopsis
C. tropicalis
Polycandidal inf.
Other Candida spp.*
Figure 1. Species responsible for candidaemia and fluconazole susceptibility in adults with haematological malignancy (Haem) or solid tumours
(Solid), Australia 2001– 2004. Polycandidal inf., infection with more than one Candida species; MIC, MIC of fluconazole. *Other Candida spp
included three each of C. guilliermondii,C. kefyr and C. lusitanae, two of C. famata and one each of C. rugosa and C. sake.
Fluconazole-resistant Candida in cancer
1045
JA
C
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
haematology patients and in 64% of oncology patients. The
median time to initiation of therapy was 1 day after the onset
of candidaemia (range: day 0 to 12 days after blood was
drawn) for haematology patients and 2 days (range day 0 to
43 days after blood was drawn) for oncology patients. Of the
39 patients with a fluconazole-resistant isolate, 19 were initially
treated with fluconazole and 6 (32%) died by day 30. A further
15 received other antifungal agents (9 amphotericin B-based
products, 3 caspofungin, 2 itraconazole and 1 voriconazole)
and 8/15 (53%) died by day 30. Outcome data were not available
for two patients, and three others received no antifungal
treatment.
Kaplan– Meier survival analysis showed that survival was
reduced in patients with candidaemia caused by a fluconazole-
resistant or S-DD isolate (Figure 2b), log rank test P¼0.0082.
These patients were alsomore likely toremain blood culture positive
at day 30 (P¼0.01) and there was a trend to poorer clinical resol-
ution (P¼0.07) and more deaths at day 30 (P¼0.05) (Table 5).
Discussion
This population-based description of candidaemia in cancer
patients is the first to identify independent risk factors for isolation
of fluconazole-resistant isolates in this population. Analysis of
comprehensive data from multiple institutions throughout
Australia revealed that most fluconazole-resistant or S-DD
Candida BSIs were due to either C. krusei or C. glabrata (91%). Clini-
cal risk factors identified for isolation of a fluconazole-resistant
isolate included prior triazole (all) as well as fluconazole receipt,
age .65 years and, for the first time, prior GIT surgery. Further-
more, clinical outcomes and survival were less favourable in
patients infected with fluconazole-resistant isolates. Although
the association between C. glabrata and C. krusei and prior fluco-
nazole exposure has been reported in cancer patients,
4,5
the
association between fluconazole receipt and infection with a
fluconazole-resistant or S-DD C. glabrata and other species such
as C. dubliniensis,C. parapsilosis,C. rugosa and C. guilliermondii
has not.
There were significant epidemiological differences between
haematological patients and those with solid organ cancer. In
both groups, C. albicans was susceptible to fluconazole but
C. albicans caused a substantially smaller proportion of episodes
in patients with haematological malignancy compared with
those with solid organ cancer (33% versus 64%). Conversely,
C. glabrata (32% of which were fluconazole susceptible) and
C. krusei (all fluconazole resistant) caused 11% and 15% of epi-
sodes in haematology patients, respectively and 20% and 0.7%
of episodes in oncology patients, respectively. Most of this differ-
ence is attributable to prior use of fluconazole which was ident-
ified in almost half of haematology patients but only 5% of
oncology patients. Not only was receipt of prior fluconazole (or
triazoles in general, but not other antifungal classes) a risk
factor for infection with a fluconazole-resistant Candida in hae-
matology patients, but the risk was increased significantly with
prior fluconazole therapy of .14 days duration (OR 3.35,
P¼0.05).
The identification of C. glabrata as the second most common
cause of candidaemia in oncology patients is of interest. Its
recent emergence was reported in a population-based study
that included patients without malignancy
24
and a series in
cancer patients.
5
In both studies C. glabrata infection was associ-
ated with fluconazole prophylaxis.
5,24
Although only a minority
(8 of 150) of oncology patients in our study had received fluco-
nazole, age .65 years and prior GIT surgery predisposed to
recovery of fluconazole-resistant isolates. The association of
C. glabrata candidaemia with older age was observed in an
Australian multicentre study of candidaemia in general hospital
Table 2. Risk factors for candidaemia with a fluconazole-resistant or
susceptible dose-dependent isolate in adults with cancer (univariate
analysis)
Variable OR 95% CI Pvalue
Male 0.78 0.46– 1.32 0.36
Age .65 years 1.54 0.91– 2.63 0.11
Haematological malignancy 1.60 0.95– 2.69 0.08
Solid organ malignancy 0.74 0.44– 1.25 0.26
Cytotoxic chemotherapy 1.50 0.89– 2.56 0.13
Neutropenia at diagnosis 1.25 0.73– 2.15 0.42
Days of neutropenia (as logarithm) 1.79 1.09– 2.94 0.02
ANC ,0.510
9
/L 1.31 0.76– 2.23 0.33
Systemic antifungal drugs 2.99 1.68– 5.36 0.00
Systemic triazoles 3.30 1.80– 6.08 0.00
fluconazole 2.96 1.57– 5.59 0.001
.14 days of fluconazole 3.27 0.95– 11.24 0.06
itraconazole 3.52 0.92– 13.5 0.06
voriconazole 4.16 0.68 – 25.3 0.12
posaconazole 2.71 0.17 – 43.93 0.48
Conventional amphotericin B 1.36 0.24– 7.55 0.73
Candidaemia acquired in Victoria 1.87 1.10– 3.12 0.02
Gastrointestinal surgery 1.91 1.02– 3.54 0.04
Surgery 1.44 0.82– 2.52 0.20
Ventilated at diagnosis of candidaemia 1.42 0.74– 2.73 0.29
Corticosteroid use 1.39 0.92– 2.35 0.22
Heparin use 1.38 0.67– 2.82 0.37
TPN 1.32 0.76 – 2.31 0.32
Community-acquired candidaemia 0.61 0.14– 2.61 0.51
Diagnosis of candidaemia in ICU 1.31 0.79– 2.47 0.39
Use of antimicrobial agents 1.79 0.65– 4.90 0.26
Antimicrobial-impregnated CVAD 0.89 0.24– 3.39 0.87
OR, odds ratio; CI, confidence interval; ANC, absolute neutrophil count;
TPN, total parenteral nutrition; ICU, intensive care unit; CVAD, central
venous access device.
Risk factors were determined in the 30 days prior to the diagnosis of can-
didaemia unless otherwise specified.
Slavin et al.
1046
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
Table 3. Univariate analysis for 30 day mortality in adult cancer patients with candidaemia
Variable
a
OR 95% CI Pvalue Proportional hazards 95% CI Pvalue
Male 1.55 0.91– 2.64 0.11 1.29 0.86 – 1.95 0.22
Age .65 years 2.17 1.27 –3.69 0.00 1.24 0.83–1.85 0.28
Haematological malignancy 1.05 0.63 – 1.77 0.84 1.20 0.91 – 1.79 0.38
allogeneic SCT 1.32 0.42 – 4.16 0.64 1.47 0.64 – 3.37 0.36
autologous SCT 0.22 0.03 – 1.78 0.16 1.74 0.43 – 7.09 0.44
leukaemia 0.84 0.46– 1.56 0.59 0.88 0.54 – 1.42 0.27
lymphoma 1.53 0.80–2.94 0.20 1.66 1.02– 2.67 0.042
Solid organ malignancy 0.95 0.57 – 1.60 0.85 0.82 0.55–1.24 0.35
Community-acquired candidaemia 0.83 0.83–0.16 4.36 0.75 0.18– 3.03 0.69
Neutropenia at diagnosis 1.05 0.60 – 1.84 0.86 1.12 0.73–1.74 0.59
ANC ,0.510
9
/L 2.47 1.27– 4.81 0.01 1.23 0.81 – 1.89 0.32
Cytotoxic chemotherapy 0.92 0.54– 1.57 0.77 0.92 0.61 – 1.40 0.71
Corticosteroid use 1.71 1.02– 2.89 0.04 1.51 1.01–2.26 0.04
Prior systemic antifungal drugs 1.43 0.80 – 2.55 0.23 1.37 0.88 – 2.12 0.17
prior fluconazole 0.67 0.23 – 1.94 0.46 1.15 0.68 – 1.95 0.60
Candida parapsilosis 0.52 0.25–1.11 0.09 0.69 0.37– 1.29 0.25
Candida glabrata 1.45 0.74 – 2.85 0.28 1.33 0.83 – 2.15 0.24
Candida krusei 1.31 0.52– 3.29 0.57 1.38 0.67 – 2.87 0.38
Candida albicans 1.10 0.65– 1.85 0.73 0.91 0.61 – 1.36 0.65
Fluconazole resistance (MIC.64 mg/L) 1.77 1.00 –3.14 0.05 1.55 1.02– 2.38 0.04
Surgery 0.90 0.52– 1.57 0.71 0.75 0.49 – 1.15 0.19
Gastrointestinal surgery 1.65 0.62 – 4.35 0.31 0.85 0.52–1.39 0.52
TPN 1.54 0.90– 2.63 0.11 1.12 0.74 – 1.68 0.58
Antibiotics 4.06 1.18– 13.93 0.03 1.42 0.62– 3.26 0.4
Sepsis at diagnosis 1.20 0.57– 2.56 0.63 0.79 0.44 – 1.43 0.45
Onset of candidaemia in ICU 5.50 2.91–10.37 0.00 2.07 1.36–3.13 0.001
Serum creatinine .0.17 mmol/L 1.82 1.02–3.27 0.04 1.84 0.89– 3.81 0.1
Ventilated at diagnosis of candidaemia 5.47 2.70 – 11.09 0.00 2.47 1.64 – 3.14 ,0.001
CVAD removed after diagnosis 0.43 0.22 – 0.84 0.01 0.52 0.35 – 0.77 0.001
Treatment given after diagnosis
b
amphotericin B (n¼66) 0.62 0.33– 1.17 0.15 0.86 0.54 – 1.39 0.54
lipid amphotericin B (n¼48) 1.05 0.54– 2.04 0.88 0.82 0.49 – 1.37 0.45
fluconazole (n¼190) 0.44 0.25– 0.77 0.004 0.59 0.39–0.89 0.013
itraconazole (n¼9) 1.04 0.25– 4.27 0.95 1.29 0.41 – 4.13 0.66
voriconazole (n¼24) 0.52 0.18– 1.44 0.21 0.54 0.24 – 1.23 0.14
caspofungin (n¼31) 1.37 0.62 – 2.96 0.43 1.35 0.77–2.35 0.29
combination (n¼99) 0.79 0.52 – 1.19 0.26
Time to starting antifungal treatment (days) 1.00 0.99 – 1.00 0.73
OR, odds ratio; CI, confidence interval; SCT, stem cell transplant; ANC, absolute neutrophil count; ICU, intensive care unit; CVAD, central venous access
device.
a
Variables were determined in the 30 days prior to candidaemia unless otherwise specified.
b
Treatments given after diagnosis do not equal the number of patients as the number receiving each agent includes those receiving combination
therapy.
Fluconazole-resistant Candida in cancer
1047
JA
C
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
patients and in a large European study,
1,16
and with GIT surgery
within a non-neutropenic intensive care cohort of patients.
25
Komshian et al.
26
further reported C. glabrata candidaemia to
be associated with surgery in patients with solid tumours.
Antifungal treatment guidelines universally recommend an
agent other than a triazole as initial therapy in a patient who
has received fluconazole prophylaxis.
27 –29
Our data suggest
that recent GIT surgery and age .65 years should also be
explored as indications for a non-triazole-based empirical treat-
ment regimen of candidaemia in cancer patients (Table 2).
The over-representation of fluconazole-resistant isolates of
C. glabrata and C. krusei causing candidaemia in one large hospi-
tal in Victoria was responsible for the appearance of this state as
an independent risk factor for fluconazole-resistant isolates,
emphasizing the importance of monitoring local hospital epide-
miology to guide treatment. Whilst there was no difference in
antifungal policy at this hospital, its case mix may differ from
that of other hospitals surveyed as it is a referral centre for
burns, trauma, critical care and organ transplantation.
Key findings from our study included a delay in clearance of
blood cultures, a more severe clinical course and reduced survival
in patients with fluconazole-resistant candidaemia (Table 5and
Figure 2b). Two recent studies of candidaemia, in mostly
non-neutropenic patients, reported an association between
fluconazole resistance (best expressed in AUC/MIC
24
or weight-
normalized daily fluconazole dose/MIC
24
) and patient mortality
in patients treated for the first 3 days or more with flucona-
zole.
11,12
Neither included many patients with cancer
13,14
so
the generalizability of these findings is uncertain. Our data
support an impact of fluconazole resistance in cancer patients.
They are also consistent with results from an experimental
model of disseminated candidiasis where the absence of neutro-
penia and timely initiation of active antifungal agents were the
important determinants of outcome.
30
Few modifiable predictors of mortality were identified. It is
likely that the prompt commencement of appropriate antifungal
treatment in cancer patients will also improve clinical outcomes.
Studies in general hospital patients, most of whom did not
receive antifungal prophylaxis, showed a mortality benefit
when antifungal therapy was commenced within 12 – 24 h of
onset of candidaemia.
11,12
A limitation of our study was the
inability to analyse the effect on mortality of initiating antifungal
therapy within 12 h of onset of candidaemia, as only the date of
commencement of antifungals was recorded in relation to the
date the positive blood culture was drawn. However, the
median day of starting antifungal treatment in haematology
patients was day 1 after onset of candidaemia, and was day 2
Survivor function for removal of CVAD
Survivor function for fluconazole resistance
1.00
(a)
(b)
0.75
0.50
0.25
0.00
01020
30
Analysis time
Days
Days
CVAD not removed after diagnosis
CVAD removed after diagnosis
FLC susceptible FLC resistant
1.00
0.75
0.50
0.25
0.00
0102030
Figure 2. (a) Survival of cancer patients with candidaemia according to
whether the central venous access device (CVAD) was removed within
5 days of first positive culture or not. (b) Survival estimates for adult
cancer patients with candidaemia according to the presence of a
fluconazole-susceptible isolate (‘FLC susceptible’) or a fluconazole-
resistant or susceptible dose-dependent isolate (grouped as ‘FLC
resistant’)
Table 4. Cox proportional hazards regression predicting 30 day mortality in adult cancer patients with candidaemia
Variable Cox proportional hazards 95% CI Pvalue
Voriconazole given after diagnosis 0.42 0.18– 0.99 0.045
Removal of CVAD after diagnosis 0.42 0.28–0.64 ,0.001
Fluconazole given after diagnosis 0.58 0.39–0.90 0.015
Use of corticosteroids within 30 days of onset 1.55 1.02 – 2.40 0.048
Onset in ICU 1.76 0.98 – 3.15 0.055
Prior systemic triazole use 1.80 1.06 – 3.06 0.031
Ventilation at onset of candidaemia 2.26 1.28 – 4.00 0.005
Serum creatinine .17 mmol/L at onset of candidaemia 2.62 1.19–5.74 0.016
CVAD, central venous access device; ICU, intensive care unit; CI, confidence interval.
Slavin et al.
1048
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
after onset in oncology patients, compared with a median time
to initiation of 24–48 h in the published studies.
11,12
Another potentially modifiable factor is CVAD removal, which
was the only independent predictor of improved survival in our
study, although it was performed less frequently in haematology
patients than in oncology patients. Whilst another population-
based study also showed a survival benefit with CVAD
removal,
24
a prospective cohort study in cancer patients did
not.
31
The issue of CVAD removal in neutropenic haematology
patients with candidaemia remains controversial. It is possible
that candidaemia originates from the GIT rather than the
CVAD;
32
further, removal of CVADs may be impractical in severely
ill patients and in those with significant intravenous access pro-
blems. Survival benefit was suggested in three studies,
32 –34
one of which comprised both solid tumour and haematology
patients, almost half of whom were neutropenic.
32
In the
latter study the greatest clinical benefit was seen at day 0 of
candidaemia, persisted to day 4, but was only modest when
compared with the much greater impact of visceral dissemina-
tion, APACHE (Acute Physiology and Chronic Health Evaluation)
score and neutrophil recovery.
32
Notably, only 18% of patients
received antifungal prophylaxis and only 15% of episodes were
due to C. glabrata or C. krusei.
32
Since antifungal prophylaxis is
now common in haematology patients, a reappraisal of the
value of CVAD removal in managing candidaemia is required. It
is possible that the epidemiological shift away from C. albicans
to other species brings with it an increased propensity to form
biofilms.
35
In the interim, removal of the CVAD should be con-
sidered for both haematology and oncology patients.
27
Other predictors of mortality were largely unmodifiable and
included onset of candidaemia in the ICU, invasive ventilation
at onset, elevated serum creatinine and prior corticosteroid
and triazole receipt. These variables are largely markers of under-
lying co-morbidity and disease severity, captured in some series
by APACHE II scores, although these have not been validated for
use outside of the ICU setting.
36
In conclusion, this nationwide, contemporary survey of candi-
daemia in adult cancer patients highlights differences in epide-
miology between patients with haematological malignancies
and solid organ cancers, risk factors for infection with a
fluconazole-resistant isolate and the impact of fluconazole
resistance on clinical outcome. Removal of the CVAD was the
only readily modifiable predictor of mortality. These findings
warrant evaluation in other populations and may form the
basis of more targeted treatment algorithms for cancer patients.
Acknowledgements
This work was presented in part at the Forty-ninth Interscience
Conference on Antimicrobial Agents and Chemotherapy, San Francisco,
CA, USA, 2009 (Abstract M-1010).
The Australian Candidemia Study consists of the following members
and acknowledges all infectious diseases physicians, clinical
microbiologists and hospital scientists that have contributed to the study:
Queensland: Cairns Base Hospital (J. McBride); Calboolture Hospital
(C. Coulter); Mater Adult Hospital (J. McCormack and K. Walmsley);
Princess Alexandra Hospital (D. Looke, B. Johnson, G. Nimmo and
G. Playford); Queensland Medical Laboratories (D. Drummond);
Rockhampton Hospital (E. Preston); Royal Brisbane Hospital (A. Allworth
and J. Faoagali); Sullivan and Nicolaides Pathology (J. Botes and
J. Robson); Townsville Hospital (R. Norton); and The Prince Charles
Hospital (C. Coulter).
New South Wales: Albury Base Hospital (D. Robb); Concord Hospital
(T. Gottlieb); Douglass Hanly Moir Pathology (I. Chambers); Gosford
Hospital (D. DeWit); Hunter Area Pathology Service (J. Ferguson and
L. Tierney); Liverpool Hospital (F. Jozwiak and R. Munro); Manning Base
Hospital (R. Pickles); Mayne Health (J. Holland); Narrabri District Hospital
(F. Groenwald); New Children’s Hospital (K. Hale); Orange Base Hospital
(R. Vaz); Prince of Wales Hospital (R. Hardiman and C. Baleriola); Royal
North Shore Hospital (R. Pritchard and K. Weeks); Royal Prince Alfred
Hospital (R. Benn and N. Adams); St George Hospital (R. Lawrence and
P. Taylor); St Vincent’s Private and St Vincent’s Public Hospital
(J. Harkness, D. Marriott and Q. Nguyen); Sydney Children’s Hospital
(P. Palasanthrian); Sydney Adventist Hospital (R. Grant); Westmead
Hospital (S. Chen, C. Halliday, O. C. Lee and T. Sorrell); and Wollongong
Hospital (P. Newton and N. Dennis).
Victoria: Alfred Hospital (C. Franklin, O. Morrisey, M. Slavin and
D. Spelman); Austin and Repatriation Hospital (B. Speed); Bendigo
Health Care Group (J. Hellsten and R. Russell); Melbourne Pathology
(S. Coloe); Melbourne Private Hospital (A. Sherman); Monash Medical
Centre (T. Korman); PathCare Consulting Pathologists (S. Graves); Peter
MacCallum Cancer Institute (M. Slavin and M. Huysmans); and Royal
Melbourne Hospital (M. Slavin and A. Sherman).
South Australia: Flinders Medical Centre (D. Gordon); Royal Adelaide
Hospital (K. Rowlands, D. Shaw and W. Ferguson); and Women’s and
Children’s Hospital (D. Ellis, R. Handke and S. Davis).
Western Australia: Fremantle Hospital (M. Beaman and J. McCarthy);
Royal Perth Hospital (C. Heath); and Sir Charles Gairdner Hospital
(S. Altmann, I. Arthur and D. Speers).
Tasmania: Launceston General (E. Cox); and Royal Hobart Hospital
(L. Cooley and A. McGregor).
Northern Territory: Royal Darwin Hospital (B. Currie, G. Lum and D. Fisher).
Australian Capital Territory: The Canberra Hospital (P. Collignon and
A. Watson).
Table 5. Outcomes for adult patients with candidaemia with fully fluconazole-susceptible isolates (no fluconazole resistance)
compared with those with fluconazole-resistant or susceptible dose-dependent isolates (fluconazole resistance present)
Outcome variable No fluconazole resistance Fluconazole resistance present Pvalue
Presence of sepsis at day 5 n(%) 71/151 (47.0) 31/54 (57.4) 0.21
Clinical resolution at day 30 n(%) 53/191 (27.8) 12/69 (17.4) 0.068
Persistently positive blood cultures at day 30 n(%) 3/109 (2.8) 6/39 (15.4) 0.011
Mortality at day 5 n(%) 19/193 (9.8) 10/69 (14.5) 0.37
Mortality at day 30 n(%) 56/193 (29.0) 40/69 (58.0) 0.053
Relapses after day 5 n(%) 17/162 (10.5) 9/49 (18.4) 0.145
Median length of stay (days) 30 33
Fluconazole-resistant Candida in cancer
1049
JA
C
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
Funding
The work was supported by an unrestricted educational grant from Pfizer.
Pfizer had no role in study design, analysis or writing of the manuscript.
Transparency declarations
M. A. S., T. C. S. and D. H. E. serve/have served on Advisory Boards for
Pfizer, Merck, Gilead and Schering-Plough and have received investigator
initiated grants from Pfizer, Merck and Gilead Sciences. D. M. has been a
consultant to Schering-Plough, Pfizer and Merck, Australia. C. O. M. has
been a consultant to, received grant support from and given lectures
for Gilead Sciences, Pfizer, Merck, Schering-Plough and Orphan
Australia. S. C.-A. C. has been a consultant to Pfizer Australia and
Gilead Sciences. Other authors: none to declare.
Author contributions
M. A. S. was part of the study design and management group, analysed
data and led the writing of the manuscript. T. C. S. was part of the study
design and management group, and contributed to writing of the
manuscript. D. M. was part of the study design and management
group, and contributed to writing of the manuscript. K. A. T. performed
the statistical analysis and contributed to writing of the
manuscript. Q. N. was the study coordinator and data manager, designed
the database and contributed to statistical analysis. D. H. E. was part of
the study design and management group, performed the antifungal
susceptibility testing and contributed to writing of the
manuscript. C. O. M. contributed to the data analysis and writing of the
manuscript. S. C.-A. C. was part of the study design and management
group, and contributed to writing of the manuscript.
References
1Tortorano AM, Kibbler C, Penman J et al. Candidemia in Europe:
epidemiology and resistance. Int J Antimicrob Agents 2006; 27: 359– 66.
2Wey SB, Mori M, Pfaller MA et al. Hospital-acquired candidemia. The
attributable mortality and excess length of stay. Arch Intern Med 1988;
148: 2642– 5.
3Zaoutis 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: 1232 –9.
4Sipsas N, Lewis RE, Tarrand J et al. Candidemia in patients with
hematologic malignancies in the era of new antifungal agents (2001–
2007). Cancer 2009; 115: 4745– 52.
5Hachem R, Hanna H, Kontoyiannis D et al. The changing epidemiology
of invasive candidiasis: Candida glabrata and Candida krusei as the
leading causes of candidemia in hematologic malignancy. Cancer 2008;
112: 2493– 9.
6Pagano L, Caira M, Candoni A et al. The epidemiology of fungal
infections in patients with hematologic malignancies: the SEIFEM-2004
study. Haematologica 2006; 91: 1068– 75.
7Marr KA, Seidel K, Slavin MA et al. Prolonged fluconazole prophylaxis is
associated with persistent protection against candidiasis-related death in
allogeneic marrow transplant recipients: long-term follow-up of a
randomized, placebo-controlled trial. Blood 2000; 96: 2055–61.
8Pfaller MA, Diekema DJ. Epidemiology of invasive candidiasis: a
persistent public health problem. Clin Microbiol Rev 2007; 20: 133–63.
9Antoniadou A, Torres HA, Lewis RE et al. Candidemia in a tertiary care
cancer center: in vitro susceptibility and its association with outcome of
initial antifungal therapy. Medicine (Baltimore) 2003; 82: 309– 21.
10 Sabino R, Verissimo C, Brandao J et al. Epidemiology of candidemia in
oncology patients: a 6-year survey in a Portuguese central hospital. Med
Mycol 2009; 5: 1– 10.
11 Garey KW, Rege M, Pai MP et al.Time to initiation of fluconazole
therapy impacts mortality in patients with candidemia: a
multi-institutional study. Clin Infect Dis 2006; 43: 25– 31.
12 Morrell M, Fraser VJ, Kollef MH. Delaying the empiric treatment of
Candida bloodstream infection until positive blood culture results are
obtained: a potential risk factor for hospital mortality. Antimicrob
Agents Chemother 2005; 49: 3640–5.
13 Pai MP, Turpin RS, Garey KW. Association of fluconazole area under
the concentration-time curve/MIC and dose/MIC ratios with mortality
in nonneutropenic patients with candidemia. Antimicrob Agents
Chemother 2007; 51:35–9.
14 Baddley JW, Patel M, Bhavnani SM et al. Association of fluconazole
pharmacodynamics with mortality in patients with candidemia.
Antimicrob Agents Chemother 2008; 52: 3022– 8.
15 Riddell JT, Kauffman CA. The evolution of resistant Candida species in
cancer centers: implications for treatment and prophylaxis. Cancer 2008;
112: 2334– 7.
16 Chen S, Slavin M, Nguyen Q et al. Active surveillance for candidemia,
Australia. Emerg Infect Dis 2006; 12: 1508– 16.
17 Levy MM, Fink MP, Marshall JC et al. 2001 SCCM/ESICM/ACCP/ATS/SIS
International Sepsis Definitions Conference. Crit Care Med 2003; 31:
1250– 6.
18 International Statistical Classification of Disease and Related Health
Problems, 10th Revision, Australian Modification (ICD-10-AM). Sydney,
Australia: University of Sydney, 1998.
19 Warren N, Hazen K. Candida,Cryptococcus, and other yeasts of
medical importance. In: Murray PR, Baron EJ, Pfaller MA et al., eds.
Manual of Clinical Microbiology. Washington, DC: ASM Press, 1999;
1184– 99.
20 Kurtzman CP, Fell JW. The Yeasts, A Taxonomic Study. 4th edn.
Amsterdam, The Netherlands: Elsevier Science B.V., 1998.
21 Meyer W, Maszewska K, Sorrell TC. PCR fingerprinting: a convenient
molecular tool to distinguish between Candida dubliniensis and Candida
albicans.Med Mycol 2001; 39: 185– 93.
22 National Committee for Clinical Laboratory Standards. Reference
Method for Broth Dilution Antifungal Susceptibility Testing of Yeasts—
Second Edition. Approved Standard M27-A2. NCCLS, Wayne, PA, USA, 2002.
23 Clinical and Laboratory Standards Institute. Reference Method for
Broth Dilution and Antifungal Susceptibility Testing of Yeasts: Third
Informational Supplement M27-S3. CLSI, Wayne, PA, USA, 2008.
24 Almirante B, Rodriguez D, Park BJ et al. Epidemiology and predictors of
mortality in cases of Candida bloodstream infection: results from
population-based surveillance, Barcelona, Spain, from 2002 to 2003.
J Clin Microbiol 2005; 43: 1829– 35.
25 Playford EG, Marriott D, Nguyen Q et al. Candidemia in
nonneutropenic critically ill patients: risk factors for non-albicans
Candida spp. Crit Care Med 2008; 36: 2034–9.
26 Komshian SV, Uwaydah AK, Sobel JD et al. Fungemia caused by
Candida species and Torulopsis glabrata in the hospitalized patient:
frequency, characteristics, and evaluation of factors influencing
outcome. Rev Infect Dis 1989; 11: 379– 90.
27 Pappas PG, Kauffman CA, Andes D et al. Clinical practice guidelines for
the management of candidiasis: 2009 update by the Infectious Diseases
Society of America. Clin Infect Dis 2009; 48: 503 – 35.
28 Thursky KA, Playford EG, Seymour JF et al. Recommendations for the
treatment of established fungal infections. Intern Med J 2008; 38:
496–520.
Slavin et al.
1050
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
29 Herbrecht R, Fluckiger U, Gachot B et al. Treatment of invasive Candida
and invasive Aspergillus infections in adult haematological patients. EJC
Supplements 2007; S5: 49– 50.
30 Hope WW, Drusano GL, Moore CB et al. Effect of neutropenia and
treatment delay on the response to antifungal agents in experimental
disseminated candidiasis. Antimicrob AgentsChemother 2007; 51: 285 –95.
31 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: 1071– 8.
32 Anaissie EJ, RexJH, Uzun O et al. Predictorsof adverse outcomein cancer
patients with candidemia. Am J Med 1998; 104: 238–45.
33 Vigouroux S, Morin O, Moreau P et al. Candidemia in patients with
hematologic malignancies: analysis of 7 years’ experience in a single
center. Haematologica 2006; 91: 717– 8.
34 Liu CY, Huang LJ, Wang WS et al. Candidemia in cancer patients:
impact of early removal of non-tunnelled central venous catheters on
outcome. J Infect 2009; 58: 154– 60.
35 Silva S, Henriques M, Martins A et al. Biofilms of non-Candida albicans
Candida species: quantification, structure and matrix composition. Med
Mycol 2009; 47: 681– 9.
36 Knaus WA, Draper EA, Wagner DP et al. APACHE II: a
severity of disease classification system. Crit Care Med 1985; 13:
818–29.
Fluconazole-resistant Candida in cancer
1051
JA
C
by guest on August 7, 2016http://jac.oxfordjournals.org/Downloaded from
