Fungal sepsis: optimizing antifungal therapy in the critical care setting.

Abstract

Invasive fungal infections (IFI) and fungal sepsis in the intensive care unit are increasing and are associated with considerable morbidity and mortality. In this setting, IFI are predominantly caused by Candida species. Outcomes continue to be suboptimal; however, there are a few key clinician modifiable factors. PK-PD studies with the approved antifungal agents have provided guidance on the dosing strategy that predicts improved outcome. In addition, time to therapy is a critical element. Therefore early recognition through improved risk factor analysis and diagnostics will be key developments. Source control for infected devices and endophthalmitis must be considered.

Full text

Page 1

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Page 2

Author's personal copy
Fungal Sepsis:
Optimizing
Antifungal Therapy
in the Critical Care
Setting
Alexander Lepak, MDa, David Andes, MDb,c,*
Invasive fungal infections (IFI) and fungal sepsis in the intensive care unit (ICU) are
increasing and are associated with considerable morbidity and mortality. In this
setting, IFI are predominantly caused by Candida species. Currently, candidemia
represents the fourth most common health care–associated blood stream infection.1–3
With increasingly immunocompromised patient populations, other fungal species
such as Aspergillus species, Pneumocystis jiroveci, Cryptococcus, Zygomycetes,
Fusarium species, and Scedosporium species have emerged.4–9However, this review
focuses on invasive candidiasis (IC).
Multiple retrospective studies have examined the crude mortality in patients with
candidemia and identified rates ranging from 46% to 75%.3In many instances, this
is partly caused by severe underlying comorbidities. Carefully matched, retrospective
cohort studies have been undertaken to estimate mortality attributable to candidemia
and report rates ranging from 10% to 49%.10–15Resource use associated with this
infection is also significant. Estimates from numerous studies suggest the added
hospital cost is as much as $40,000 per case.10–12,16–20Overall attributable costs
are difficult to calculate with precision, but have been estimated to be close to 1 billion
dollars in the United States annually.21
aUniversity of Wisconsin, MFCB, Room 5218, 1685 Highland Avenue, Madison, WI 53705-2281,
USA
bDepartment of Medicine, University of Wisconsin, MFCB, Room 5211, 1685 Highland Avenue,
Madison, WI 53705-2281, USA
cDepartment of Microbiology and Immunology, University of Wisconsin, MFCB, Room 5211,
1685 Highland Avenue, Madison, WI 53705-2281, USA
* Corresponding author. Department of Microbiology and Immunology, University of
Wisconsin, MFCB, Room 5211, 1685 Highland Avenue, Madison, WI 53705-2281.
E-mail address: dra@medicine.wisc.edu
KEYWORDS
? Invasive candidiasis ? Pharmacokinetics-pharmacodynamics
?Therapy ?Source control
Crit Care Clin 27 (2011) 123–147
doi:10.1016/j.ccc.2010.11.001
0749-0704/11/$ – see front matter. Published by Elsevier Inc.
criticalcare.theclinics.com

Page 3

Author's personal copy
Understanding the management factors that affect outcome in this disease state is
thus increasingly important for the critical care physician. The care strategies of
demonstrated relevance include prompt suspicion of IC, rapid evaluation, diagnosis
and initiation of antimicrobial therapy. In addition, the optimal agent to target these
pathogens and the proper dosing regimen must be chosen, and source control should
be considered. These care themes are similar to those of demonstrated importance
for bacterial sepsis, yet many features are unique to IC. These issues are examined
in this review.
EPIDEMIOLOGY
TheincidenceandepidemiologyofICintheICUhasundergoneconsiderablechangein
thepast3decades.Thisevolutionhashadasubstantialeffectonthetherapeutictarget
and thus empirical treatment of IC in the ICU. The incidence of Candida blood stream
infection (BSI) rose sharply in the 1980s with a more than 5-fold increase compared
withstudiesadecadeearlier.22Thistrendcontinuedwithagreaterthan200%increase
from 1979 to 2000.23The shift in the epidemiology of this health care–associated
infection has continued in the current decade, with rates now estimated to be between
25 and 30 per 100,000 persons.16,24In addition to an absolute increase in disease
incidence,therehasalsobeenachangeinthespeciesresponsiblefortheseinfections.
At least 17 Candida species have been reported to cause IC in humans, but 5 species
(C albicans, C glabrata, C parapsilosis, C tropicalis, and C krusei) represent more than
90%. C albicans has historically been the predominant pathogen in IC with rates of
80% or higher in the 1980s.3,25Presently, C albicans accounts for less than 50% of
all BSIs caused by the Candida genus.3,26–30The predominant non-albicans species
in the United States is C glabrata, with an estimated frequency between 20% and
25%.3In contrast, other countries have noted dramatic increases in C parapsilosis
and C tropicalis.3,31As C glabrata often exhibits reduced susceptibility to triazoles
andCparapsilosis hasreduced susceptibilitytoechinocandins, knowledge ofthe local
epidemiology is imperative for selection of appropriate empirical therapy.
TIME TO THERAPY
The importance of prompt identification of IC through a combination of risk factor
analysis and diagnostic assays has been demonstrated to be a key factor affecting
sepsis-related mortality.32–41Studies on patients with IC have similarly shown
excessive rates of inappropriate initial therapy and even higher mortality than
infections caused by bacterial pathogens in the ICU setting.10,14,33,34,42–47One of the
earliest studies to examine the effect of appropriate antimicrobial therapy in the ICU
observed a 33% lower mortality in patients receiving adequate therapy defined by in
vitro antimicrobial susceptibility testing.33Inappropriate treatment was most common
for Enterococcus spp; however, inadequate coverage for IC was the second most
common error and was associated with the highest mortality. Another retrospective
cohortstudybyKumarandcolleagues34demonstratedthatadministrationofadequate
antimicrobial therapy within the first hour of documented hypotension was associated
withasurvivalrateof79.9%,andeach1-hourdelayinantimicrobial therapywaslinked
to a 7.6% per hour decrease in survival. Subgroup analysis identified a significant
relationship between hospital survival and duration of time between onset of sepsis
and drug therapy for each microbial genus including Candida species.
Additional studies investigating the effect of time to initial appropriate therapy have
specifically targeted IC. Blot and colleagues42reported 78% mortality in patients with
IC when therapy was delayed more than 48 hours from onset of candidemia; in
Lepak & Andes
124

Page 4

Author's personal copy
contrast the mortality was 44% in those who had adequate initial therapy. Another
single-center retrospective cohort of 157 consecutive patients with a Candida BSI
showed a trend toward improved survival when appropriate antifungal therapy was
administered within 12 hours of blood culture collection (mortality 11.1% vs 33.1%;
P 5 .169).45In multivariate analysis the odds of death was 2-fold higher if adequate
therapy was delayed more than 12hours (P 5 .018). Garey and colleagues46examined
time to antifungal therapy in a multicenter, retrospective cohort of 230 patients with
a Candida BSI who were prescribed fluconazole. Mortality was lowest (15.4%) if flu-
conazole therapy was started on the same day that the culture was performed and
increased to 23.7% if fluconazole therapy was started on day 1, 36.4% on day 2,
and 41.4% if it was started day 3 (P 5 .0009). Multivariate analysis in this study also
provided a strong association between delay in therapy and mortality. In a recent
publication, Patel and colleagues48performed a retrospective review of cases of
Candida sepsis (positive blood culture within 72 hours of refractory shock) from
a single center from 2003 to 2007. Using classification and regression tree analysis
(CART), they found patients who received early, appropriate antifungal therapy (within
15 hours of collecting the first positive blood culture) had improved survival. Another
recent publication by Taur and colleagues49in a cohort of patients with cancer found
similar results. These data strongly support early appropriate antifungal therapy and
demonstrate a need for improved disease identification.
EARLY DIAGNOSTICS
Two areas of intense investigation designed to improve early IC identification include
novel laboratory assays and risk factor–based disease prediction. Unfortunately, IC
lacks specific and objective clinical findings.3,50–53In addition, the gold standard
diagnostic test for IC has been isolation of the organism in blood culture, however,
blood culture for Candida is insensitive.52–55For example, modern blood culture
systems are estimated to detect only 50% to 67% of cases of IC. Furthermore,
detection of candidemia by blood culture often takes more than 24 hours. For certain
species, such as C glabrata, the time to positive culture can be even longer.56–60In
a study by Fernandez and colleagues,56the mean time to yeast detection in blood
cultures for C albicans was 35.3 ? 18.1 hours, whereas that of C glabrata was
80.0 ? 22.4 hours. Mean time to final identification to species level for C albicans
was 85.8 ? 30.9 hours compared with 154 ? 43.8 hours for C glabrata. In this study,
the time to appropriate therapy for C albicans isolates was 43.3 ? 27.6 hours
compared with 98.1 ? 38.3 hours for C glabrata isolates. Thus, waiting for positive
blood culture results and potentially susceptibility testing leads to a significant delay
in appropriate therapy and in turn higher mortality.
The relatively slow growth of these organisms in culture systems has led to the
development of several non–culture-based diagnostic studies to detect fungal cell
wall components, antigens, or nucleic acids secreted into the blood. Approved
serologic tests in the United States or Europe include mannan antibody/antigen
detection (Platelia) and b-1,3-D-glucan (Glucatell, Fungitell). In Europe, a nucleic acid
identification system (SeptiFast) is approved for detection in blood of certain bacterial
and fungal pathogens including the 5 most commonly encountered Candida species.
A retrospective study evaluated mannan antigen or mannan antibody detection in
the sera of patients with culture proven candidemia.61In 36 of 43 (84%) patients with
cultureprovencandidemiaatleast1ofthe2serologictestswaspositive.Thesensitivities
were40%and98%andthespecificitieswere53%and94%formannanemiaorantibody
detection, respectively. These values reached 80% and 93% when the results of both
Fungal Sepsis
125

Page 5

Author's personal copy
tests were combined. A follow-up study of patients with candidemia revealed that 33 of
45 (73%) patients had positive serology using the Platelia assay for mannan antigen or
antimannan antibodies at least 2 days before having a positive blood culture, with
a mean positive serologic test 6 days before positive blood culture.62b-1,
3-D-Glucan detection has also been evaluated in critically ill patients, with sensitivity
and specificity for diagnosis of IC estimated to be 69.9% and 87.1%, respectively.63
Several nucleic acid detection platforms have been evaluated for detection of IC. For
example, McMullan and colleagues prospectively evaluated 3 Taqman-based nested
polymerase chain reaction (PCR) assays for the detection of candidemia in critical ill
patients. Twenty-three of 157 patients included in the study had proven IC. The
estimated clinical sensitivity, specificity, and positive and negative predictive values of
the assays in this trial were 90.9%, 100%, 100%, and 99.8%, respectively.64Nucleic
acid detection in this study identified fungal sepsis as soon as 6 hours within onset of
symptoms. SeptiFast, which is able to detect DNA of 20 common bacteria and fungi,
maybeahelpfuladjuncttobloodculturesforfastidiousorganismsandmaybeadvanta-
geous in settings where patients are either already on antibiotics or have been
pretreated.65,66An important limitation in evaluating these newer diagnostic tests is the
lack a reference gold standard (blood culture) with high specificity and sensitivity. For
example,inameta-analysisofstudiesusingPCRtodiagnoseIC,thesensitivityprogres-
sively decreased as the reference standard went from patients with culture proven can-
didemia to those with proven/probable IC, and even lower when compared with those
with proven, probable, or possible IC.67At present, there is significant heterogeneity in
assay characteristics including choice of sample, primer, nesting strategy, and nucleic
acid extraction. Although promising, there is need for further prospective validation of
these diagnostic tools and wider availability, as currently only a few academic centers
or reference laboratories have the capability to offer these advanced diagnostic tech-
niques.Untilconfirmatoryresultsareavailablefromongoingstudies,routineuseofthese
new diagnostic assays will remain investigative. If the empirical trial use of the b-1,
3-D-glucan test is shown to be useful, it is reasonable to expect this assay could be
adapted for use in most tertiary care centers in the near future.
EMPIRICAL AND PREEMPTIVE STRATEGIES BASED ON RISK IDENTIFICATION
The importance of prompt antifungal treatment and lack of sensitivity and timeliness of
diagnostic assays has led to the development of empirical and preemptive treatment
strategies.Theinitiationofantifungaltreatmentinthesecasesisbasedonacompilation
of host risk factors for IC. Multiple studies have evaluated risk factors for IC.51,68–86
Variables that have been associated with IC include high APACHE II score, exposure
tobroadspectrumantibiotics,cancerchemotherapy,evidenceofmucosalcolonization
by Candida spp, pancreatitis, indwelling vascular catheters (especially central venous
catheter [CVC]), administration of total parenteral nutrition, neutropenia, immune
suppression therapy, prior surgery (especially gastrointestinal surgery), renal failure
or hemodialysis, and prolonged ICU stay (especially in the surgical ICU). These risk
factors by themselves impart limited specificity as they are present in many critically
ill patients.
Attempts to enhance the predictive value of these variables have involved
examining a composite of risk factors.74,78,85,87–91A laboratory marker used in several
risk stratification schemes is culture isolation of Candida species from multiple
nonblood sites. The predictive value of this laboratory marker by itself has been
explored by Pittet and colleagues,87who examined the value of identifying Candida
colonization at multiple nonblood sites, termed the Candida colonization index (CI).
Lepak & Andes
126