candida colonization index and subsequent didier pittet ... · candida spp. on mucosal surfaces...
TRANSCRIPT
Philippe EggimannDidier Pittet
Candida colonization index and subsequentinfection in critically ill surgical patients:20 years later
Received: 20 March 2014Accepted: 23 May 2014Published online: 17 June 2014� The Author(s) 2014. This article ispublished with open access atSpringerlink.com
P. EggimannAdult Critical Care Medicine and BurnUnit, Centre Hospitalier UniversitaireVaudois (CHUV), Lausanne, Switzerlande-mail: [email protected]
D. Pittet ())Faculty of Medicine, Infection ControlProgram and WHO Collaborating Center onPatient Safety, University of GenevaHospitals (HUG), Rue Gabrielle-Perret-Gentil 4, 1205 Geneva, Switzerlande-mail: [email protected].: ?41 22 372 9844
Abstract Introduction: For dec-ades, clinicians dealing withimmunocompromised and critically illpatients have perceived a linkbetween Candida colonization andsubsequent infection. However, thepathophysiological progression fromcolonization to infection was clearlyestablished only through the formaldescription of the colonization index(CI) in critically ill patients. Unfortu-nately, the literature reflects intenseconfusion about the pathophysiologyof invasive candidiasis and specificassociated risk factors. Meth-ods: We review the contribution ofthe CI in the field of Candida infectionand its development in the 20 yearsfollowing its original description in1994. The development of the CIenabled an improved understanding ofthe pathogenesis of invasive candidi-asis and the use of targeted empiricalantifungal therapy in subgroups ofpatients at increased risk for infection.
Results: The recognition of specificcharacteristics among underlyingconditions, such as neutropenia, solidorgan transplantation, and surgicaland nonsurgical critical illness, hasenabled the description of distinctepidemiological patterns in thedevelopment of invasive candidiasis.Conclusions: Despite its limitedbedside practicality and before con-firmation of potentially more accuratepredictors, such as specific biomark-ers, the CI remains an important wayto characterize the dynamics of colo-nization, which increases early inpatients who develop invasivecandidiasis.
Keywords Candida albicans �Candidemia � Invasive candidiasis �Colonization � Colonization index �Empirical treatment �Nosocomial infections
Introduction
Candida spp. colonization occurs in up to 80 % of criticallyill patients after 1 week in intensive care [1–3]. This very highproportion contrasts strongly with the low rate of invasivecandidiasis in less than 10 % of them [4, 5]. Despite recentadvances in microbiological techniques, early diagnosis ofinvasive candidiasis remains problematic and microbiologi-cal documentation often occurs late in the course of infection[6–8]. This explains its high crude and attributable mortality,i.e., in the range reported for septic shock [9–11].
Early empirical treatment of severe candidiasis hasimproved survival, but is responsible for the overuse ofantifungals [12–14]. This overuse not only contributes toa huge financial burden, but has also promoted a shift toCandida species with reduced susceptibility to antifungalagents [15, 16]. Unfortunately, recent guidelines resultingfrom expert consensus failed to provide high-level rec-ommendations about empirical antifungal treatment or toclarify the nature of such treatment strategies [8, 17, 18].Despite limited levels of evidence, empirical treatmentcurrently relies on the positive predictive value of risk
Intensive Care Med (2014) 40:1429–1448DOI 10.1007/s00134-014-3355-z MY PAPER 20 YEARS LATER
assessment strategies, such as the colonization index (CI),Candida score, and predictive rules based on combina-tions of risk factors [19–21].
Improved knowledge of the pathophysiological spec-ificities of invasive candidiasis should promote better useof clinical tools in the evaluation of patients who couldtruly benefit from early antifungal therapy [22]. In con-trast to most bacterial infections, invasive candidiasis ischaracterized by a 7–10-day delay between exposure torisk factors and infection development [23–25]. This timewindow provides a unique opportunity for the imple-mentation of a structured approach to the rigorousselection of patients likely to benefit from early empiricalantifungal treatment [19, 26].
The CI is the most widely studied clinical tool for theearly risk assessment of invasive candidiasis among at-risk patients. The primary objective of this review is tohighlight its role in this setting. The use of colonizationdynamics, as assessed by the CI, for the early detection ofpatients likely to benefit from early antifungal treatment isalso explored in comparison with other tools currentlyavailable in clinical practice. Finally, we also propose aresearch agenda for the next decade.
Pathophysiology of invasive candidiasis
Nosocomial exogenous transmission of Candida has beenwell described [27]. Of note, both endogenous andexogenous colonization can co-exist in the clinical setting[1]. However, carefully designed studies using genotypingof Candida strains confirmed that endogenous coloniza-tion is responsible for the large majority of severecandidiasis [28, 29].
The epidemiology of invasive candidiasis has someimportant differences in immunocompromised and criti-cally ill patients (Fig. 1).
Neutropenic and bone marrow transplant recipients
Neutrophils are essential in the defense against invasivecandidiasis [30]. Their prolonged absence in neutropenicpatients or those with functional impairment after bonemarrow transplantation, combined with the selectivepressure of frequent and repetitive exposure to antibacte-rial agents, plays a major role in the development ofinvasive candidiasis in these patients. A high density ofCandida spp. on mucosal surfaces injured by chemother-apy develops progressively and is responsible for fungaltranslocation with a high risk of candidemia. The use ofsystematic antifungal prophylaxis in the past three decadesexplains the epidemiology of Candida infection, typi-cally characterized by breakthrough invasive candidiasis[31, 32]. Infection is almost always caused by a Candidastrain resistant to the antifungal agent used [15, 33].
Solid organ transplant recipients
Anti-rejection therapy increases the risk for Aspergillusand other filamentous fungal infections in solid organtransplant recipients, except among those receiving smallbowel and liver transplants, in whom additional specificsurgical factors further modify the epidemiology of dis-ease [34, 35]. Prolonged and intense impairment ofcellular T cell-mediated immune response may contributeto the high proportion of cutaneous and mucosal candi-diasis [36]. This factor has justified the use of systematicantifungal prophylaxis in these patients in whom, asamong neutropenic patients and bone marrow transplantrecipients, breakthrough Candida spp. infection developswith strains resistant to the agent used [18].
Nonsurgical critically ill patients
In nonsurgical critically ill patients, the continuous andprolonged support of failing organs and the selectivepressure of broad-spectrum antibiotics constitute key riskfactors for invasive candidiasis. Support of organ failurerequires the use of numerous devices, such as intravas-cular catheters, endotracheal tubes, naso- and oro-gastrictubes, and Foley catheters, which are frequently colonizedby Candida spp. as a result of the high affinity of theirbiofilm [1]. These specificities may explain progressivecolonization in a high proportion of patients after pro-longed stay in the intensive care unit (ICU) [6, 37, 38].They may also explain a higher proportion of catheter-related infections in the absence of severe immuneimpairment [39]. In this context, the role of exposure tosteroids largely used to improve the speed of recoveryfrom circulatory failure remains to be determined.
Limited data are available on the usefulness of pro-phylaxis and recent guidelines did not includerecommendation for antifungal prophylaxis in nonsurgi-cal critically ill patients [17, 18, 40].
Critically ill patients recovering from abdominalsurgery
Additional factors should be considered in patients afterabdominal surgery [41]. Any perforation or opening of thedigestive tract results in contamination of the peritoneumby bowel flora. In most cases, surgical cleaning of theabdominal cavity, eventually combined with antibiotics, issufficient to allow full recovery [42]. Except for patientspresenting with nosocomial peritonitis and some of thosepresenting with septic shock and multiple organ failure,the identification of Candida spp. has no clinical signifi-cance under these conditions [22]. Recurrent peritonitisfollowing anastomotic leakage or persistent inflammationmay be required for the progression of Candida spp.
1430
Neutropenic patients Solid organ transplant recipients
Candida spp.Gram positiveGram negative
Red cell
Macrophage
Critically ill non-surgical patientsCritically ill surgical patients
Candida spp.Gram positiveGram negative
Red cell
Macrophage
Lymphocyte
Neutrophil
Fig. 1 Pathophysiology of invasive candidiasis in different patient populations. Detailed mechanisms are summarized in thecorresponding sections of the text
1431
colonization to invasive candidiasis [43]. However, thedevelopment of invasive candidiasis requires concomitantexposure to additional factors, such as an increasingamount of Candida spp. in a non-functioning bowel,prolonged antibiotic treatment, and/or requirement fororgan support. In contrast to nonsurgical patients, theseparticularities explain why a high proportion of invasivecandidiasis cases do not manifest with candidemia anddevelop only late in the course of the disease [44].
Non-candidemic invasive candidiasis includes intra-abdominal abscess and peritonitis defined by one of thefollowing culture results from specimens obtained at sur-gery: (1) monomicrobial growth of Candida spp.; (2) anyamount of Candida spp. growth within a mixed-floraabscess; or (3) moderate or heavy growth of Candida spp. inmixed-flora peritonitis treated with appropriate antibacte-rial therapy according to susceptibility testing [43–46].
Two small prospective studies, including one placebo-controlled, suggested that antifungal prophylaxis inpatients presenting with anastomotic leakage afterabdominal surgery may prevent the development ofinvasive candidiasis [18, 45, 46].
In patients with severe acute necrotizing pancreatitis,invasive candidiasis develops with similar pathophysio-logical characteristics. Progressive colonization of thebowel within the first 2 weeks of the disease results intranslocation into necrotic tissues, and fungal infectionshave been documented in up to 10 % of patients notexposed to antibiotics.
Critically ill patients with Candida isolatedfrom the respiratory tract
Candida spp. colonization of the airway is frequentlyreported in mechanically ventilated critically ill patients,and its clinical significance is difficult to evaluate. Can-dida has a low affinity for alveolar pneumocytes andhistologically documented pneumonia has been rarelyreported. Hematogenous dissemination in the context ofcandidemia [1] may be responsible for multiple pulmon-ary abscesses and should be viewed as a distinct entity.Hence the existence of true candidal pneumonia isdoubtful and recovery of Candida spp. from the respira-tory tract should generally be considered as colonizationand does not justify antifungal therapy [17–19].
Candida colonization
The distinction between patients with Candida spp. col-onization who did not require treatment and those likelyto develop invasive infection was identified as a difficultclinical challenge in the field of infectious diseases morethan 40 years ago [47]. After the description in the early
1970s of Candida infection in surgical patients, mani-festing as ‘‘disseminated candidiasis’’ and characterizedby persistent fungemia and multiple disseminated mic-roabscesses on autopsy, several authors reported that itmight be preceded by fungal colonization. Klein andWatanakunakorn [48] reported the presence of visceralmicroabscesses in up to 30 % of patients with Candidaspp. colonization and recommended treatment of all suchpatients with critical illness or persistent candidemia.
Solomkin et al. confirmed this concept in a detailedreview of 63 surgical patients with postoperative candi-demia. Among 51 patients who developed candidemia as alate complication of intra-abdominal infection, adequateantifungal treatment ([3 mg/kg amphotericin B) resultedin survival of 10/15 patients compared with 6/36 untreatedpatients. Autopsies revealed visceral Candida spp. mic-roabscesses in 7/20 patients [42, 49]. Moreover, a digestivefungal source was identified in nearly all patients, andcandidemia was considered to originate from intestinalCandida spp. growth in the presence of mucosal loss.
Calandra et al. provided further evidence of the clinicalsignificance of Candida colonization of the peritoneum insurgical patients [39]. In a series of 49 patients withspontaneous perforation (n = 28) or surgical opening ofthe digestive tract (n = 21), 19 (39 %) patients developedinvasive candidiasis, including 7 cases of intra-abdominalabscesses and 12 cases of peritonitis [43]. Invasive can-didiasis developed significantly more frequently in patientswho had undergone surgery for acute pancreatitis than inthose with gastrointestinal perforations (90 vs. 32 %,p = 0.005) or other disorders, such as diseases of the bil-iary tract or colon (90 vs. 17 %, p = 0.003). Invasivecandidiasis was associated significantly with an initiallyhigh or increasing amount of Candida spp., as assessed bysemiquantitative cultures. Compared with uninfectedpatients, Candida spp. showed light, moderate, and heavygrowth from the first positive specimen in 26 (87 %) vs. 9(47 %), 4 (13 %) vs. 6 (32 %), and 0 vs. 4 (21 %) infectedpatients, respectively (p = 0.005). Moreover, the amountof Candida spp. cultured from subsequent specimensincreased in 15 (79 %) infected patients and only 2 (7 %)uninfected patients (p \ 0.001) [39].
Candida colonization index: the missing link
In the early 1990s, the origin of Candida spp. infectionwas highly debated [50]. Whether the infection wasexogenous or endogenous in origin, and whether patientscould become colonized with strains from the surroundingenvironment or with their ‘‘own’’ strains of Candida spp.was uncertain. We tested the hypothesis that yeast colo-nization would precede infection in critically ill patients[23]. We determined the genotypic characteristics of 322Candida spp. collected prospectively from 29 critically ill
1432
patients who developed significant colonization underroutine culture surveillance two to three times weeklyover a 6-month period. These patients belonged to acohort of 650 surgical ICU patients. Significant coloni-zation was defined as the presence of Candida spp. inthree or more samples taken from one or more sites on atleast two consecutive screening days. Candida spp.strains isolated from an individual patient had an identicalgenetic pattern, even when isolated from different bodysites, and this pattern remained the same over a prolongedperiod of up to 140 days. No horizontal transmissioncould be demonstrated during the study period. Invasivecandidiasis, including 8 cases of candidemia, occurred in11/29 (38 %) patients. All patients who developedinfection had previous colonization by strains with iden-tical genetic patterns. Our intensive surveillance failed todocument cross-transmission of Candida spp. betweencolonized and non-colonized patients.
Colonization index
The development of the CI has been viewed as a majorconceptual advance in the characterization of supportingthe progression from colonization to infection in surgicalpatients.
The Candida CI is defined as the ratio of the number ofdistinct non-blood body sites colonized by Candida spp. to
the total number of body sites cultured [24]. With theexception of blood cultures, samples collected from bodysites other than those routinely screened are also consideredin the CI (Table 1). Only strains of Candida spp. with thesame genetic identity are considered in calculating the CI.
All 29 included patients were heavily colonized withCandida spp., but with no significant difference in thereason for ICU admission and comorbidities betweeninfected and uninfected patients. The Acute Physiologyand Chronic Health Evaluation II (APACHE II) score andduration of antibiotic exposure before colonization werehigher among the 11 patients who ultimately developedinvasive candidiasis. In these patients, colonization pre-ceded infection by an average of 25 days (range 6–70).The 18 patients who did not develop infection were col-onized for a mean of 29 days (range 5–140). A total of153 [5.3 ± 1 (range 3–8) per patient] distinct body siteswere tested, and this number did not differ significantlybetween the two groups of patients.
The CI was calculated daily, and the highest valuesobtained before invasive candidiasis development werecompared with those recorded for patients who did notdevelop candidiasis. The average CI differed significantlybetween colonized and infected patients (0.47 vs. 0.70,p \ 0.01). All patients who ultimately developed infec-tion reached the threshold value of 0.5 before infection(Fig. 2). Importantly, infected patients reached thethreshold CI value an average of 6 days (range 2–21)
Table 1 Description of the colonization index
Index Definition Formula
Colonization index (CI) Ratio of the number of distinct non-blood body sitescolonized by Candida spp. to the total number ofbody sites cultured. With the exception of bloodcultures, samples collected from body sites otherthan those routinely screened are also considered.Only strains of Candida spp. with identicalelectrophoretic karyotypes are considered whencalculating the CI
CI¼ Number of sites colonizedNumber of sites cultured
Corrected colonizationindex (CCI)
Product of the CI and the ratio of the number ofdistinct body sites showing heavy growth (???from semi-quantitative culture or B105 in urine orgastric juice) to the total number of body sites withCandida spp. heavy growth
CCI ¼ Number of sites colonized
Number of sites cultured
� Number of sites with heavy growth
Total positive sites
Use of the CIa In patients perceived to be at risk of developing aninvasive candidiasis: twice weekly surveillanceculture of the following sitesb:
- Oropharynx swab or tracheal secretions- Gastric fluid- Perinea swab or stool sample- Urine sample- Surgical wound swab or drained abdominal fluids- Catheter insertion sites
a Identification of Candida species at the genetic level has onlybeen performed in the original study. However, different species ofCandida may contribute to the burden of fungi which will eventuallyresult in symptomatic infection. Accordingly, we suggest to con-sider all species of Candida identified for the calculation of the CI
b Except for specific invasive candidiasis discussed in the sectionon pathophysiology of non-candidemic invasive candidiasis, Can-dida spp. isolated from these sites should be considered ascolonization
1433
before candidiasis. All but one patient reached thisthreshold at least 3 days before the time of infection. Thesensitivity, specificity, and positive and negative predic-tive values of the CI were determined to be 100, 69, 66,and 100 %, respectively.
Corrected colonization index
In a post hoc analysis, we developed the corrected colo-nization index (CCI) by taking into account the amount ofCandida spp. recovered by semiquantitative cultures [24].The CCI is the product of the CI and the ratio of thenumber of distinct body sites showing heavy growth(??? or at most 105 in urine or gastric juice) to the totalof distinct body sites with Candida spp. growth (Table 1).
The mean CCI differed significantly between colo-nized (0.16) and infected (0.56) patients (p \ 0.01). CCIswere less than 0.35 in all colonized patients and at least0.4 in all infected patients (p B 0.001). The sensitivity,specificity, and positive and negative predictive valueswere all 100 %. In a multiple logistic regression analysis,only the APACHE II score [odds ratio (OR), 1.03; 95 %confidence interval, 1.01–1.05 per point; p = 0.007] andCCI (OR, 4.01; 95 % confidence interval, 2.16–7.45;p \ 0.001) independently predicted the development ofinvasive candidiasis.
Overview of the usefulness of the colonization index
Since the publication of the paper describing the CI in1994 in Annals of Surgery, many centers have used the CI
or a methodology derived from its original description toassess the dynamics of Candida colonization in differentsubgroups of critically ill patients at risk of invasivecandidiasis. Unfortunately, these data have not been val-idated in large multicenter trials.
Several studies have indirectly suggested the validityand potential usefulness of the CI, but almost exclusivelyin surgical patients (Table 2). This index has been used tocharacterize colonization dynamics [37, 38, 51–55],assess the significance of candiduria in critically illpatients [38, 56–59], and evaluate the impact of anti-fungal prophylaxis in subgroups of patients at risk ofcandidiasis [40, 46, 60–63]. At least four studies havereported the use of the CI to guide empirical antifungaltreatment. Although major methodological flaws maypreclude the validation of their findings, this applicationreduced the incidence of ICU-acquired invasive candidi-asis in all four studies (Table 3) [57, 64–66].
Dubau et al. [57] prospectively monitored CIs weeklyin 89/669 consecutive patients staying in the ICU formore than 7 days after digestive surgery who had pro-tein C levels greater than 100 mg/l or received antibiotictreatment for more than 7 days. All patients developinga CI above 0.5 were empirically treated with antifungalsuntil it decreased below 0.5. The proportion of patientswith CIs greater than 0.5 increased from 6 to 25, 40, 55,and 59 % after 1, 2, 3, 4, and more than 4 weeks,respectively. Only one of the 35 patients with CIsgreater than 0.5 who were empirically treated withantifungal medications developed candidiasis; the degreeof colonization decreased rapidly in the other 34patients.
In a before/after trial, Piarroux et al. [64] prospectivelyused weekly CIs and CCIs to assess the intensity ofCandida spp. colonization in 478 surgical patients stayingin the ICU. Patients with CCIs greater than 0.4 receivedempirical antifungal treatment. Compared with an his-torical cohort of 455 control subjects, the incidence ofinvasive candidiasis was lower among these patients (7.0vs. 3.8 %; p = 0.03). Moreover, this strategy completelyprevented the development of ICU-acquired invasivecandidiasis. The proportions of patients treated empiri-cally were 87 % of those with CCIs greater than 0.4(some of them died before the CCI could be calculated),18 % of those with CCIs less than 0.4 and CIs greaterthan 0.5, and 2 % of those with CCIs less than 0.4. A totalof 25 % of patients received empirical antifungal treat-ment. The authors estimated that 160 completemycological screenings and 10 preemptive antifungaltreatments were needed to prevent one proven ICU-acquired invasive candidiasis. Despite the absence ofrandomization, this study is the most valuable ever per-formed using the CI to guide empirical antifungaltreatment. However, the intense laboratory work requiredto perform CI in all ICU patients and the high proportionof patient treated may explain why this practice is not
Days
Colonization intex
0.8
0.6
0.4
0.2
Fig. 2 Colonization index. This index is defined as the ratio of thenumber of distinct non-blood body sites with Candida spp.colonization to the total number of distinct body sites tested. Itwas recorded for each patient from the first day of colonizationuntil discharge from the intensive care unit (uninfected patients) orsevere candidiasis development (infected patients). Black circlesrepresent patients who developed severe candidiasis, white circlesrepresent patients who remained colonized. Reproduced withpermission from Pittet et al. [24]
1434
Tab
le2
Rep
ort
edex
per
ience
wit
hth
euse
of
the
colo
niz
atio
nin
dex
(CI)
Ref
eren
ceT
ype
of
pat
ient;
study
des
ign
Num
ber
of
pat
ients
Num
ber
of
invas
ive
candid
iasi
s
case
s
CI
monit
ori
ng
Num
ber
of
surv
eill
ance
cult
ure
s
Mai
nC
Ifi
ndin
gs
Auth
ors
’co
ncl
usi
on
Gen
eral
studie
s(n
=7)
Tra
net
al.
[51
]C
ardia
csu
rger
y;
pro
spec
tive
81
7P
reoper
ativ
ely,
day
6or
at
dis
char
ge
773
(mea
n,
8.8
/
pat
ient)
;116
(16.3
%)
posi
tive
CI:
in57
%dev
elopin
gIC
,
23
%unin
fect
ed;
CI
asso
ciat
edw
ith
fem
ale
sex
(45
vs.
21
%),
cath
eter
dura
tion
(13
vs.
11
day
s),
length
of
ICU
stay
(5.7
vs.
4.9
day
s)
CI
did
not
chan
ge
in43
%
dev
elopin
gIC
,72
%
unin
fect
ed
‘‘T
his
study
show
edth
atth
ree
epid
emio
logic
alfa
ctors
—
dura
tion
of
the
intr
avas
cula
r
cath
eter
izat
ion
and
med
ical
dev
ices
,le
ngth
of
stay
inth
e
SIC
Uan
dth
ese
xof
the
pat
ient—
wer
eas
soci
ated
in
elec
tive
card
iovas
cula
r
surg
ical
pat
ients
wit
han
incr
ease
of
the
Candid
asp
p.
colo
niz
atio
nin
dex
,w
hic
h
hel
ps
toid
enti
fyhig
h-r
isk
pat
ients
for
Candid
a
infe
ctio
ns’
’
Yaz
dan
par
ast
etal
.[5
2]
Car
dia
csu
rger
y;
pro
spec
tive
131:
coro
nar
yar
tery
bypas
sgra
ftin
g
wit
h(n
=72)
and
wit
hout
(n=
59)
extr
acorp
ore
al
circ
ula
tion
0P
reoper
ativ
ely,
post
oper
ativ
ely
(day
not
spec
ified
)
147
Inpat
ients
requir
ing
extr
acorp
ore
alci
rcula
tion,
:C
Ias
soci
ated
wit
hm
ore
anti
bio
tic
use
(1.5
0±
0.8
3
vs.
1.0
8±
0.4
0;
dif
fere
nt
agen
ts;
p=
0.0
055)
‘‘E
pid
emio
logic
aldat
a
obta
ined
from
this
study
show
that
coro
nar
yar
tery
bypas
sgra
ftin
gw
ith
extr
acorp
ore
alci
rcula
tion
pro
cedure
isas
soci
ated
wit
h
anin
crea
sein
the
use
of
anti
bio
tics
and
subse
quen
tly
ahig
her
risk
of
Candid
a
colo
niz
atio
n-i
nfe
ctio
n’’
Char
les
etal
.
[37
]
Mix
edIC
U,
pat
ients
wit
h
candid
emia
;
retr
osp
ecti
ve
51:
20
MIC
U,
32
SIC
U(p
\0.0
5)
51
Dai
lyfo
r7
day
s
bef
ore
candid
emia
Not
spec
ified
(mea
n,
3.2
/pat
ient
bef
ore
candid
emia
)
Mea
nC
Iin
46
asse
ssed
:
0.5
6±
0.3
9
CI
C0.5
0in
21
(45.6
%)
MIC
U0.7
4±
0.3
1vs.
SIC
U
0.4
5±
0.4
0(p
=0.0
1)
Mea
nC
Iin
surv
ival
0.4
6vs.
dea
th0.6
3(p
=0.0
4)
Bet
ter
ICU
outc
om
eas
soci
ated
wit
hpri
or
surg
ery
[HR
,0.2
5
(0.0
9–0.6
7)]
,an
tifu
ngal
trea
tmen
t[H
R,
0.1
1
(0.0
4–0.3
0)]
,ab
sence
of
neu
tropen
ia[H
R,
0.1
0
(0.0
2–0.4
5)]
‘‘C
andid
emia
occ
urr
ence
is
asso
ciat
edw
ith
ahig
h
mort
alit
yra
team
ong
crit
ical
lyil
lpat
ients
.
Dif
fere
nce
sin
under
lyin
g
condit
ions
could
acco
unt
for
the
poore
routc
om
eof
the
med
ical
pat
ients
.S
cree
nin
g
for
fungal
colo
niz
atio
nco
uld
allo
wid
enti
fica
tion
of
such
hig
h-r
isk
pat
ients
and,
in
turn
,im
pro
ve
outc
om
e’’
1435
Ta
ble
2co
nti
nu
ed
Ref
eren
ceT
ype
of
pat
ient;
study
des
ign
Num
ber
of
pat
ients
Num
ber
of
invas
ive
candid
iasi
s
case
s
CI
monit
ori
ng
Num
ber
of
surv
eill
ance
cult
ure
s
Mai
nC
Ifi
ndin
gs
Auth
ors
’co
ncl
usi
on
Char
les
etal
.
[38]
Med
ical
ICU
,st
ay
[7
day
s;
pro
spec
tive
92
1(s
epti
csh
ock
wit
hty
pic
al
mac
ulo
pap
ula
r
rash
,posi
tive
skin
bio
psy
)
Wee
kly
(mea
n,
3.2
/
pat
ient)
1,6
96
(mea
n,
18.4
/pat
ient)
CI
C0.5
0in
36
(39.1
%),
alm
ost
all
wit
hdet
ecta
ble
fungal
colo
niz
atio
non
ICU
adm
issi
on
Pre
dic
tors
of
CI
C0.5
:
Candid
aco
loniz
atio
nat
adm
issi
on
[OR
18.8
0
(5.2
1–67.7
9)]
,[2
day
s
bla
dder
cath
eter
izat
ion
[OR
,
10.4
4(1
.61–67.8
5)]
CI:
asso
ciat
edw
ith
day
son
bro
ad-s
pec
trum
anti
bio
tics
[b=
0.0
1(0
.01–0.0
2)]
,
hem
atolo
gic
alm
alig
nan
cy
[b=
0.4
1(0
.09–0.7
3)]
,
candid
uri
a[b
=0.2
(0.0
9–0.3
1)]
Em
pir
ical
anti
fungal
trea
tmen
t
pre
scri
bed
in14/3
6(3
9%
)
wit
hC
IC
0.5
,7/5
6(1
3%
)
wit
hC
I\
0.5
(p=
0.0
07)
‘‘C
andid
asp
p.
mult
iple
-sit
e
colo
niz
atio
nis
freq
uen
tly
reac
hed
among
the
crit
ical
ly
ill
med
ical
pat
ients
.B
road
spec
trum
anti
bio
tic
ther
apy
was
found
topro
mote
fungal
gro
wth
inpat
ients
wit
hpri
or
colo
niz
atio
n.
Sin
cem
ost
of
the
invas
ive
candid
iasi
sin
the
ICU
sett
ing
are
thought
tobe
subse
quen
tto
colo
niz
atio
nin
hig
h-r
isk
pat
ients
,re
duci
ng
anti
bio
tic
use
could
be
use
ful
in
pre
ven
ting
fungal
infe
ctio
ns’
’
Elo
yet
al.
[53
]M
ixed
ICU
,ri
sk
fact
ors
(anti
bio
tics
or
hosp
ital
stay
[8
day
s,
neu
tropen
ia);
pro
spec
tive
75:
46
MIC
U,
29
SIC
U
4M
ICU
(2
resp
irat
ory
trac
t,2
uri
nar
y);
5
SIC
U(4
per
itonit
is,
1
uri
nar
y)
At
adm
issi
on,
wee
kly
901;
469
(52
%)
posi
tive
Acc
ura
cyof
CI[
0.5
in
pre
dic
ting
ICin
MIC
U:
sens,
75
%;
spec
,35
%;
PP
V,
10
%;
NP
V,
94
%
InS
ICU
:se
ns,
100
%;
spec
,
50
%;
PP
V,
29
%;
NP
V,
100
%
‘‘S
erolo
gic
alte
sts
fail
edto
dif
fere
nti
ate
infe
cted
from
non-i
nfe
cted
pat
ients
.T
he
Pit
tet’
sC
Iid
enti
fied
infe
cted
surg
ical
pat
ients
(Fis
her
exac
tte
st,
0.0
52),
whic
har
e
inth
epopula
tion
wit
h
CI[
0.5
’’
Agval
d-O
hm
an
etal
.[5
4]
Mix
edIC
U,
stay
[7
day
s;
pro
spec
tive
59
10
At
adm
issi
on,
then
wee
kly
401;
149
(37
%)
posi
tive
32
(54
%)
rece
ived
anti
fungal
s:
10
wit
hIC
[mea
nC
I,0.7
;8
had
dig
esti
ve
surg
ery,
7had
CIs
[0.5
(\0.8
in6)]
;22
wit
hout
IC(m
ean
CI,
0.2
6;
7
had
dig
esti
ve
surg
ery,
9had
CIs
[0.5
)
17/2
5(6
8%
)w
ith
CIs
C0.5
on
day
7re
ceiv
ed
anti
fungal
s
Pre
dic
tors
of
IC:
CI[
0.5
(OR
,
19.1
[2.4
–435])
,dig
esti
ve
surg
ery
[OR
,60
(2.4
–
infi
nit
y)]
‘‘H
igh
colo
niz
atio
nin
dex
and
rece
nt
exte
nsi
ve
gas
tro-
abdom
inal
surg
ery
wer
e
signifi
cantl
yco
rrel
ated
wit
h
IC.
The
resu
lts
indic
ate
that
ICU
pat
ients
expose
dto
exte
nsi
ve
gas
tro-a
bdom
inal
surg
ery
would
ben
efit
from
earl
yan
tifu
ngal
pro
phyla
xis
’’
1436
Ta
ble
2co
nti
nu
ed
Ref
eren
ceT
ype
of
pat
ient;
study
des
ign
Num
ber
of
pat
ients
Num
ber
of
invas
ive
candid
iasi
s
case
s
CI
monit
ori
ng
Num
ber
of
surv
eill
ance
cult
ure
s
Mai
nC
Ifi
ndin
gs
Auth
ors
’co
ncl
usi
on
Mas
sou
etal
.
[55
]
Med
ical
ICU
,st
ay
[2
day
san
dat
leas
tone
risk
fact
or;
pro
spec
tive
100
15
At
adm
issi
on,
then
wee
kly
816
(incl
udin
g143
blo
od
cult
ure
s)
CI
C0.5
in53
(53
%),
pre
dic
ted
only
by
cort
icost
eroid
s[O
R,
5.1
(1.0
2–25.2
)]
Acc
ura
cyof
CI[
0.5
in
pre
dic
ting
IC:
sens,
93
%;
spec
,48
%;
PP
V,
26
%;
NP
V,
98
%
Only
neu
tropen
iapre
dic
ted
invas
ive
candid
iasi
s[O
R,
18.3
(2.9
–114)]
‘‘C
Ihas
the
advan
tage
to
pro
vid
equan
tifi
eddat
aof
the
pat
ient’
ssi
tuat
ion
in
rela
tion
toth
eco
loniz
atio
n.
But,
itis
n’t
hel
pfu
lw
ith
pat
ients
hav
ing
anin
vas
ive
candid
iasi
sin
med
ical
inte
nsi
ve
care
unit
’’
Candid
uri
ast
udie
s(n
=5)
Chab
asse
[56
]15
mix
edIC
Us,
one
maj
or
or
two
min
or
risk
fact
ors
and
candid
uri
a;
pro
spec
tive
135
0A
tti
me
of
candid
uri
aan
d/
or
candid
emia
Not
avai
lable
Can
did
uri
a:\
10
3,
56
(42
%);
10
3–10
4,
21
(16
%);
[10
4,
56
(42
%)
CI[
0.5
in36/7
6te
sted
(65
%
wit
hca
ndid
uri
a[
10
4,
31
%
wit
hca
ndid
uri
a\
10
4;
p=
0.0
03)
‘‘Q
uan
tifi
cati
on
of
candid
uri
a
could
be
use
ful
tose
lect
pat
ients
athig
hri
skfo
r
dis
sem
inat
edca
ndid
iasi
s’’
Dubau
etal
.
[57
]
Surg
ical
ICU
,st
ay
or
anti
bio
tics
[7
day
sor
post
oper
ativ
e
fist
ula
and
CR
P[
100
mg/
ml;
pro
spec
tive
89
1/3
5em
pir
ical
ly
trea
ted
wit
h
CIs
C0.5
;0
wit
h
CIs
\0.5
;22
candid
uri
a
On
incl
usi
on,
then
wee
kly
2,2
38
Abse
nce
of
candid
uri
a:
CI
=0.3
–0.4
7(p
=0.0
08)
Pre
sence
of
candid
uri
a:
CI
=0.5
7–0.8
7
(p=
0.0
001)
‘‘T
he
pre
sence
of
aca
ndid
uri
a
was
signifi
cantl
yas
soci
ated
wit
han
incr
ease
din
vas
ive
candid
iasi
s’’
Sel
lam
iet
al.
[58
]
Mix
edIC
U,
stay
[3
day
s;
pro
spec
tive
162
6;
56
candid
uri
aO
nin
clusi
on,
then
wee
kly
Not
avai
lable
Can
did
uri
a:\
10
3,
12
(21
%);
10
3–10
4,
16
(29
%);
[10
4,
28
(50
%)
Mea
nC
I=
0.4
7ca
ndid
uri
a,
0.8
IC(n
=6)
CI[
0.5
in67
%ca
ndid
uri
a
[10
4
‘‘C
andid
uri
asu
per
ior
or
equal
to10
4U
FC
/ml
asso
ciat
ed
wit
hri
skfa
ctors
may
pre
dic
t
invas
ive
candid
iasi
sin
crit
ical
lyil
lpat
ients
’’
Char
les
etal
.
[38
]
Med
ical
ICU
,
stay
[7
day
s;
pro
spec
tive
92
9W
eekly
1,6
96
(mea
n,
18.4
/
pat
ient)
:C
Ias
soci
ated
wit
h:
day
son
bro
ad-s
pec
trum
anti
bio
tics
[b=
0.0
1(0
.01–0.0
2)]
,
hem
atolo
gic
alm
alig
nan
cy
[b=
0.4
1(0
.09–0.7
3)]
,
candid
uri
a[b
=0.2
(0.0
9–0.3
1)]
‘‘S
ince
most
of
the
invas
ive
candid
iasi
sin
the
ICU
sett
ing
are
thought
tobe
subse
quen
tto
colo
niz
atio
n
inhig
h-r
isk
pat
ients
,
reduci
ng
anti
bio
tic
use
could
be
use
ful
inpre
ven
ting
fungal
infe
ctio
ns’
’
1437
Ta
ble
2co
nti
nu
ed
Ref
eren
ceT
ype
of
pat
ient;
study
des
ign
Num
ber
of
pat
ients
Num
ber
of
invas
ive
candid
iasi
s
case
s
CI
monit
ori
ng
Num
ber
of
surv
eill
ance
cult
ure
s
Mai
nC
Ifi
ndin
gs
Auth
ors
’co
ncl
usi
on
Erg
inet
al.
[59
]M
ixed
ICU
,st
ay
[7
day
s;
pro
spec
tive
100
9(5
candid
emia
,
4uri
nar
y);
42
candid
uri
a
On
incl
usi
on,
then
wee
kly
1,6
91
(mea
n,
17/p
atie
nt)
CI[
0.2
,42
(42
%);
CI[
0.2
and
\0.5
,34
(34
%);
CI
C0.5
,8
(8%
)
Acc
ura
cyof
CI
C0.5
in
pre
dic
ting
invas
ive
candid
iasi
s:se
ns,
100
%;
spec
,64
%;
PP
V,
21
%;
NP
V,
100
%
‘‘C
andid
aco
loniz
atio
nan
d
Candid
aco
loniz
atio
nin
dex
may
be
use
das
use
ful
par
amet
ers
topre
dic
t
invas
ive
Candid
a
infe
ctio
ns’
’
Pro
phyl
axi
sst
udie
s(n
=6)
Lav
erdie
re
etal
.[6
0]
Neu
tropen
ia
(leu
kem
iaor
BM
T);
pro
spec
tive,
double
-bli
nded
,
random
ized
:
pro
phyla
xis
(fluco
naz
ole
,
400
mg/d
ay
ora
lly)
vs.
pla
cebo
266
41:
9/1
35
(6.7
%)
fluco
naz
ole
,
32/1
31
(24.4
%)
pla
cebo
(p\
0.0
01)
On
random
izat
ion,
aten
dof
pro
phyla
xis
1,9
04;
458
(21
%)
posi
tive
Colo
niz
atio
n;
(39–36
%)
fluco
naz
ole
,:
(37–73
%)
pla
cebo
(p\
0.0
001)
CI;
(0.1
8–0.1
6)
fluco
naz
ole
,
:(0
.18–0.3
9)
pla
cebo
(p\
0.0
01)
Acc
ura
cyof
CI
B0.2
5in
pre
dic
ting
ICat
bas
elin
e:
sens,
39
%;
spec
,82
%;
PP
V,
28
%;
NP
V,
88
%
At
end
of
pro
phyla
xis
:se
ns,
76
%;
spec
,69
%;
PP
V,
69
%;
NP
V,
94
%
‘‘F
luco
naz
ole
pre
ven
ted
and
reduce
dfu
ngal
colo
niz
atio
n
of
the
alim
enta
rytr
act
and
subse
quen
tin
vas
ive
fungal
infe
ctio
ns…
Inca
nce
r
pat
ients
,a
colo
niz
atio
n
index
B0.2
5at
the
init
iati
on
of
chem
oth
erap
ycl
earl
y
pre
dic
tsa
low
risk
of
invas
ive
fungal
infe
ctio
n’’
Gar
bin
oet
al.
[40]
Mix
edIC
U,
[2
day
s
mec
han
ical
ven
tila
tion
and
expec
ted
conti
nuat
ion
for
C72
h;
pro
spec
tive,
double
-bli
nded
random
ized
:
pro
phyla
xis
(fluco
naz
ole
,
100
mg/d
ayiv
)
vs.
pla
cebo
plu
s
sele
ctiv
e
dig
esti
ve
dec
onta
min
atio
n
(poly
myxin
B,
neo
myci
n,
van
com
yci
n)
204
6/1
03
(5.8
%)
fluco
naz
ole
,
16/1
01
(16
%)
pla
cebo
(p\
0.0
01)
Dai
lyN
ot
avai
lable
Colo
niz
atio
n:
29/5
5(5
3%
)
fluco
naz
ole
,40/5
1(7
8%
)
pla
cebo
(p=
0.0
1)
CI;
(0.2
6–0.1
3)
fluco
naz
ole
,
:(0
.26–0.5
0)
pla
cebo
(p\
0.0
01)
Mea
npre
-infe
ctio
nC
Iin
pat
ients
wit
hca
ndid
emia
(n=
10):
0.8
9
‘‘…
fluco
naz
ole
pro
phyla
xis
in
sele
cted
,hig
h-r
isk
crit
ical
ly
ill
pat
ients
dec
reas
esth
e
inci
den
ceof
Candid
a
infe
ctio
n,
inpar
ticu
lar,
candid
emia
’’
1438
Ta
ble
2co
nti
nu
ed
Ref
eren
ceT
ype
of
pat
ient;
study
des
ign
Num
ber
of
pat
ients
Num
ber
of
invas
ive
candid
iasi
s
case
s
CI
monit
ori
ng
Num
ber
of
surv
eill
ance
cult
ure
s
Mai
nC
Ifi
ndin
gs
Auth
ors
’co
ncl
usi
on
Norm
and
etal
.
[61
]
Mix
ed
ICU
,[2
day
s
mec
han
ical
ven
tila
tion;
pro
spec
tive,
open
-lab
el,
nyst
atin
vs.
pla
cebo
98
0O
nra
ndom
izat
ion,
then
ever
y
3day
s
Not
avai
lable
Colo
niz
atio
n:
0/5
1nyst
atin
,
12/4
7(2
5%
)pla
cebo
(p\
0.0
1)
CI;
(0.1
–0.0
5)
nyst
atin
,:
(0.1
–0.2
5)
pla
cebo
(p\
0.0
5)
‘‘O
ral
nyst
atin
pro
phyla
xis
effi
cien
tly
pre
ven
ted
Candid
asp
p.
colo
niz
atio
nin
ICU
pat
ients
atlo
wri
skof
dev
elopin
gin
vas
ive
candid
iasi
s’’
Sen
net
al.
[46
]S
urg
ical
ICU
,
recu
rren
t
gas
troin
test
inal
per
fora
tion/
anas
tom
oti
c
leak
age
or
acute
nec
roti
zing
pan
crea
titi
s;
pro
spec
tive,
non-
com
par
ativ
e,
casp
ofu
ngin
pro
phyla
xis
19
1(6
–8
expec
ted
inth
ishig
h-
risk
gro
up
wit
hout
pro
phyla
xis
)
On
incl
usi
on,
then
twic
ein
wee
k1,
then
wee
kly
unti
len
dof
foll
ow
-up
Not
avai
lable
(med
ian,
4si
tes/
pat
ient
scre
ened
)
CI;
(0.5
–0.3
)duri
ng
pro
phyla
xis
(p=
0.0
3)
‘‘D
espit
eli
mit
atio
ns
such
asth
e
open
single
-cen
ter
non-
com
par
ativ
edes
ign
and
the
smal
lsa
mple
size
,th
e
obse
rvat
ions
of
this
pro
of-
of-
conce
pt
study
sugges
t
that
casp
ofu
ngin
may
be
effi
caci
ous
and
safe
for
pre
ven
tion
of
intr
a-
abdom
inal
candid
iasi
sin
surg
ical
pat
ients
wit
ha
hig
h-
risk
pro
file
’’
Gig
lio
etal
.
[62
]
Surg
ical
/tra
um
a
ICU
,[2
day
s
mec
han
ical
ven
tila
tion;
pro
spec
tive,
random
ized
,
open
-lab
el:
pro
phyla
xis
(ora
l
nyst
atin
,
39
1m
illi
on
U/d
ay)
vs.
pla
cebo
99
0O
nin
clusi
on,
then
ever
y3
day
s
unti
len
dof
foll
ow
-up
(day
15)
2,5
69;
746
(29
%)
posi
tive
Over
all:
CI;
(0.1
2–0.0
)
nyst
atin
,:
(0.2
–0.4
4)
pla
cebo
(p\
0.0
5)
Colo
niz
atio
nat
entr
y:
CI;
(0.1
–0.0
)nyst
atin
,:
(0.2
–0.4
2)
pla
cebo
(p\
0.0
5)
‘‘T
he
pre
sent
tria
lsh
ow
sth
at
nyst
atin
pre
-em
pti
ve
ther
apy
insu
rgic
al/t
raum
aIC
U
pat
ients
signifi
cantl
yre
duce
s
fungal
colo
niz
atio
n,
even
in
those
colo
niz
edat
adm
issi
on’’
Chen
etal
.[6
3]
Mix
edIC
U,
mec
han
ical
ven
tila
tion;
pro
spec
tive,
random
ized
,
open
-lab
el:
pro
phyla
xis
(ora
l
nyst
atin
,
39
1m
illi
on
U/d
ay)
vs.
pla
cebo
124
8:
3/6
0(0
.5%
)
nyst
atin
,5/6
4
(7.8
%)
pla
cebo
(p[
0.0
5)
On
incl
usi
on,
ever
y
3day
sunti
len
d
of
foll
ow
-up
(day
9)
Not
avai
lable
;874
posi
tive
CC
Iat
day
6:
0.1
9nyst
atin
,
0.3
9pla
cebo
(p\
0.0
5)
At
day
9:
0.0
nyst
atin
,0.4
5
pla
cebo
(p\
0.0
5)
Len
gth
of
ICU
stay
:
9.6
±3.5
day
snyst
atin
,
11.9
±6.3
day
spla
cebo
(p\
0.0
5)
‘‘N
yst
atin
mig
ht
reduce
the
colo
niz
atio
nby
Candid
a
alb
icans
and
was
asso
ciat
ed
wit
hsh
ort
erIC
Ust
ay’’
CR
PC
-rea
ctiv
epro
tein
,B
MT
bone
mar
row
tran
spla
nt,
MIC
Um
edic
alIC
U,
SIC
Usu
rgic
alIC
U,
sens
sensi
tivit
y,
spec
spec
ifici
ty,
PP
Vposi
tive
pre
dic
tive
val
ue,
NP
Vneg
ativ
epre
dic
tive
val
ue
1439
Ta
ble
3U
seo
fth
eco
lon
izat
ion
index
(CI)
tog
uid
eth
ein
itia
tio
no
fem
pir
ical
anti
fun
gal
trea
tmen
t
Ref
eren
ceT
yp
eo
fp
atie
nt;
stud
yd
esig
nN
um
ber
of
pat
ien
tsN
um
ber
of
invas
ive
cand
idia
sis
case
s
CI
mo
nit
ori
ng
Nu
mb
ero
fsu
rvei
llan
cecu
ltu
res
Mai
nC
Ifi
nd
ing
sA
uth
ors
’co
ncl
usi
on
Inte
rven
tional
studie
s(n
=4
)D
ub
auet
al.
[57]
Su
rgic
alIC
U,
stay
or
anti
bio
tics
[7
day
so
r-
po
sto
per
ativ
efi
stu
laan
dC
RP
[1
00
mg
/ml;
pro
spec
tiv
e,o
pen
,n
on
-ra
nd
om
ized
:em
pir
ical
anti
fun
gal
trea
tmen
tfo
rC
I[
0.5
89
1/3
5w
ith
CI[
0.5
rece
ivin
gem
pir
ical
anti
fun
gal
trea
tmen
t
On
incl
usi
on
,th
enw
eek
ly2
,238
CI[
0.5
:w
eek
1,
6%
;w
eek
2,
25
%;
wee
k3
,4
0%
;w
eek
4,
55
%;
wee
k[
4,
59
%C
I;
rapid
lyaf
ter
star
to
fem
pir
ical
anti
fun
gal
trea
tmen
tin
34
/35
pat
ien
ts
‘‘A
trea
tmen
tw
asst
arte
dw
hen
ever
aco
lon
izat
ion
ind
ex[
0.5
was
asso
ciat
edw
ith
sev
ere
clin
ical
or
bio
log
ical
sig
ns.
Th
isin
vo
lved
anin
crea
seo
fth
eex
pen
seo
fan
tifu
ngal
dru
gs.
Th
ep
ote
nti
alb
enefi
tsco
uld
no
tb
eev
aluat
edfr
om
our
study’’
Pia
rro
ux
etal
.[6
4]
Su
rgic
alIC
U,
stay
C5
day
s;p
rosp
ecti
ve,
op
en,
no
n-
ran
do
miz
ed:
empir
ical
anti
fun
gal
trea
tmen
tfo
rC
CI[
0.4
47
8O
ver
all
IC:
18
[3.8
%;
2/4
55
(7.0
%)
inh
isto
rica
lco
ntr
ols
]IC
U-a
cqu
ired
:0
[10
/45
5(2
.2%
)in
his
tori
cal
con
tro
ls;
p\
0.0
01
)
On
incl
usi
on
,th
enw
eek
ly6
,682
Wit
hem
pir
ical
trea
tmen
t:C
CI[
0.4
,9
6/1
17;
CC
I\
0.4
and
CI
C0
.5,
11
/66
;C
CI\
0.4
and
CI\
0.5
,5
/23
0*
16
0co
mp
lete
my
colo
gic
alsc
reen
ing
san
d1
0p
reem
pti
ve
trea
tmen
tsn
eed
edto
pre
ven
tat
leas
to
ne
pro
ven
SIC
U-a
cqu
ired
can
did
iasi
s
‘‘P
reem
pti
ve
trea
tmen
to
fh
igh
lyco
lon
ized
pat
ien
tsm
ayef
fici
entl
yp
rev
ent
SIC
U-a
cquir
edpro
ven
candid
iasi
s.O
ur
resu
lts
dem
on
stra
teth
efe
asib
ilit
yan
db
enefi
tso
fim
ple
men
tin
ga
larg
esy
stem
atic
myco
logic
alsc
reen
ing
of
SIC
Up
atie
nts
’’E
ren
etal
.[6
5]
Mix
edIC
U,
incl
usi
on
crit
eria
no
tsp
ecifi
ed;
pro
spec
tiv
eo
bse
rvat
ion
al,
inte
rven
tio
nn
ot
spec
ified
(anti
fun
gal
trea
tmen
tfo
rC
I[
0.5
?)
37
0N
ot
spec
ified
19
12
6(7
0%
)w
ith
C.
alb
ica
ns
colo
niz
atio
nC
IC
0.5
in7
(5w
ith
IgM
,Ig
Gp
osi
tiv
ity
)C
I\
0.5
in1
9(3
/12
test
edw
ith
IgM
,Ig
Gp
osi
tiv
ity
)Ig
M,
IgG
fou
nd
in0
/7p
atie
nts
test
edw
ith
out
colo
niz
atio
n,
ou
to
f1
1
‘‘…
foll
ow
-up
of
the
ICU
pat
ien
tsin
term
so
fC
.a
lbic
ans
CI
and
IgM
wo
uld
be
effe
ctiv
efo
rth
ep
rev
enti
on
of
seri
ou
sC
an
did
ain
fect
ion
s’’
Wan
get
al.
[66]
5m
ixed
ICU
s,A
PA
CH
Esc
ore
[1
0;
pro
spec
tiv
era
nd
om
ized
:em
pir
ical
anti
fun
gal
trea
tmen
tfo
rC
CI
C0
.4(i
nte
rven
tio
ng
rou
p)
or
(co
ntr
ol
gro
up
)
110
No
tsp
ecifi
edN
ot
spec
ified
Not
avai
lable
;5
75
po
siti
ve
An
tifu
ng
altr
eatm
ent
for
sep
sis
star
ted
at0
.9±
0.7
day
sin
CC
IC
0.4
,3
.8±
3.6
day
sin
con
tro
l(p
\0
.05)
‘‘A
pp
lica
tio
no
fC
CI
may
enh
ance
the
accu
racy
of
tim
ely
pre
emp
tiv
etr
eatm
ent
for
invas
ive
candid
iasi
san
dfa
cili
tate
the
coll
ecti
on
of
epid
emio
log
ical
dat
ao
fC
an
did
ain
crit
ical
lyil
lp
atie
nts
’’
Sen
sse
nsi
tiv
ity
,sp
ecsp
ecifi
city
,P
PV
po
siti
ve
pre
dic
tive
val
ue,
NP
Vn
egat
ive
pre
dic
tiv
ev
alu
e
1440
currently used widely among non-immunocompromisedcritically ill patients.
In a prospective study of 37 critically ill patients, Erenet al. [65] found a correlation between the CI and thepresence of C. albicans antibodies. Among 26 patientswith C. albicans colonization, immunoglobulin M (IgM)and IgG were found in 5/7 (71 %) patients with CIsgreater than 0.5, compared with only 3/12 (25 %) patientswith CIs less than 0.5 in whom serum could be tested(total, n = 19) and 0/7 non-colonized patients in whomserum could be tested (total, n = 11) [65]. The authorsstated, ‘‘in the follow-up of the patients, no candidemiadeveloped and this was thought to be due to the pre-ventive measures which were taken especially in ICUpatients with CI [ 0.5,’’ suggesting that empirical anti-fungal treatment was given to patients with CIs greaterthan 0.5.
Comparison of the colonization index with otherpredictive tools
The accuracy of the CI has been compared with that ofother surrogate markers of invasive candidiasis, such asthe Candida score, predictive rules, and, more recently,beta-glucan (Table 4) [44, 67–72]. In a prospective studyof 89 high-risk surgical ICU patients (61 recurrent gas-trointestinal tract perforation; 25 severe acutepancreatitis) in whom 29 developed a non-candidemicinvasive candidiasis, Tissot el al. showed that two con-secutive beta-glucan serum levels above 80 pg/mlpredicted early the development of infection with a sen-sitivity, specificity, positive predictive value, and anegative predictive value of 65, 78, 68 and 77 %,respectively [44]. Similarly, a previous study by Posteraroet al. [70] in a cohort of medical and surgical critically illpatients staying more than 5 days and developing a severesepsis, in whom beta-glucan was higher than 80 pg/ml,showed a sensitivity, specificity, positive predictive value,and a negative predictive value of 93, 94, 72, and 99 %,respectively for the early detection of invasive candidiasis(16 episodes, including 13 candidemia). Overall, theaccuracy of CI is lower than other methods for the earlydetection of patients at higher risk of infection. Never-theless, the usefulness of beta-glucan to guide empiricalantifungal treatment remains to be determined.
Three industry-sponsored studies failed to demonstratethe clinical usefulness of predictive rules combining var-ious risk factors. A predictive rule was used in amulticenter, randomized, double-blind, placebo-con-trolled trial comparing caspofungin vs. placebo asantifungal prophylaxis or preemptive treatment in 222critically ill patients [73]. The rule combined ICU stay forat least 3 days, mechanical ventilation, exposure to
antibiotics, the presence of a central line, together with atleast one additional risk factor among the following: par-enteral nutrition, dialysis, surgery, pancreatitis, systemicsteroids, or other immune suppressive agents. The inci-dence of invasive candidiasis was 16.7 % (14/84) vs.9.8 % (10/102) in patients receiving placebo and caspo-fungin, respectively (p = 0.14). When baseline infectionswere included, it was 30.4 % (31/102) and 18.8 % (22/117), for patients receiving preemptive placebo and ca-spofungin, respectively (p = 0.04). Safety, length of ICUstay, antifungal use, and mortality did not differ. Theauthors concluded that caspofungin prophylaxis was safewith a non-significant trend to reduce invasive candidiasis,and that the preemptive therapy approach deserves furtherstudy. Two recent and currently unpublished studiesattempt to use clinical predictive rules to guide empiricalantifungal treatment. The first, entitled ‘‘Pilot FeasibilityStudy With Patients Who at High Risk For DevelopingInvasive Candidiasis in a Critical Care Setting (MK-0991-067 AM1’’ (ClinicalTrials.gov identifier: NCT01045798)and conducted by the Mycoses Study Group, was termi-nated because of a low recruitment rate after the inclusionof only 15 patients and generation of uninterpretable data.Preliminary results of the second study, entitled ‘‘A Studyto Evaluate Pre-emptive Treatment for Invasive Candidi-asis in High Risk Surgical Subjects (INTENSE)’’(ClinicalTrials.gov identifier: NCT01122368), showed ahigh proportion of invasive candidiasis at study entry;independent experts considered, however, that the major-ity of salvage antifungal treatments were unjustified.Globally, the proportion of invasive candidiasis develop-ing in the group receiving preemptive antifungal treatmentwas not different from that of the group receiving placebo,but the number of patients excluded from the analysisresulted in an underpowered analysis.
These outcomes illustrate a well-known paradigm ofaccuracy. The use of the most sensitive test potentiallyincreases the number of patients who are uselessly treatedempirically. In contrast, the use of the most specific testpotentially increases the number of patients who do notreceive empirical antifungal treatment and in whom latetreatment could be associated with a worse outcome.
An ongoing prospective multicenter, double blind,randomized-controlled French trial (EMPIRICUS) iscurrently recruiting critically ill patients at risk of inva-sive candidiasis. Patients mechanically ventilated formore than 4 days with sepsis of unknown origin and withat least one extra-digestive fungal colonization site andmultiple organ failure are eligible for randomization toreceive empirical micafungin or a placebo. This studyincludes prospective determination of the CI among itssecondary endpoints and may help to develop guidelinesfor treating non-immunocompromised patients with fun-gal colonization multiple organ failure and sepsis ofunknown origin [74].
1441
Ta
ble
4C
om
par
iso
no
fth
eco
lon
izat
ion
ind
ex(C
I)w
ith
oth
erm
ark
ers
for
the
pre
dic
tio
no
fin
vas
ive
can
did
iasi
s
Ref
eren
ces
Ty
pe
of
pat
ien
t;st
ud
yd
esig
nN
um
ber
of
pat
ien
tsN
um
ber
of
inv
asiv
eca
nd
idia
sis
CI
mo
nit
ori
ng
Per
form
ance
of
CI
Per
form
ance
of
com
par
ato
rsA
uth
ors
’co
ncl
usi
on
Leo
net
al.
[67]
36
mix
edIC
Us,
stay
C7
day
s;p
rosp
ecti
ve,
ob
serv
atio
nal
coh
ort
1,1
07
58
;1
8/2
40
inw
ho
mB
Gco
uld
be
per
form
ed
At
incl
usi
on
,th
enw
eek
ly(4
,19
8cu
ltu
res)
89
2w
ith
Ca
nd
ida
spp
.co
lon
izat
ion
or
inv
asiv
eca
nd
idia
sis
Dev
elo
pm
ent
of
inv
asiv
eca
nd
idia
sis:
CI\
0.5
,1
6/4
11
(3.9
%);
CI
C0
.5,
42
/48
1(8
.7%
)S
ens,
72
%;
spec
,4
7%
;P
PV
,9
%;
NP
V,
96
%;
AU
C,
0.6
33
20
.8n
eed
edto
trea
tto
avo
ido
ne
epis
od
eo
fin
vas
ive
can
did
iasi
s
89
2w
ith
Ca
nd
ida
spp
.co
lon
izat
ion
or
inv
asiv
eca
nd
idia
sis
Dev
elo
pm
ent
of
inv
asiv
eca
nd
idia
sis:
Ca
nd
ida
sco
re\
3,
13
/56
5(2
.3%
);C
an
did
asc
ore
C3
,4
5/3
27
(13
.8%
)S
ens,
78
%;
spec
,6
6%
;P
PV
,1
4%
;N
PV
,9
8%
;A
UC
,0
.77
48
.7n
eed
edto
trea
tto
avo
ido
ne
epis
od
eo
fin
vas
ive
can
did
iasi
s
‘‘In
this
coh
ort
of
colo
niz
edp
atie
nts
stay
ing
[7
day
s,w
ith
aC
S\
3an
dn
ot
rece
ivin
gan
tifu
ng
altr
eatm
ent,
the
rate
of
ICw
as\
5%
.T
her
efo
re,
ICis
hig
hly
imp
rob
able
ifa
Ca
nd
ida-c
olo
niz
edn
on
-neu
tro
pen
iccr
itic
ally
ill
pat
ien
th
asa
Ca
nd
ida
sco
re\
0.3
’’
Ell
iset
al.
[ 68]
Hem
ato
log
ical
pat
ien
tsw
ith
feb
rile
neu
tro
pen
iaan
d[
3d
ays
bro
ad-
spec
tru
man
tib
ioti
cs;
pro
spec
tiv
e,u
sefu
lnes
so
fse
rial
anal
ysi
sfo
rm
ann
ans/
anti
-m
ann
ans,
b-g
luca
ns
86
12
(14
%)
Tw
ice
aw
eek
CI
C0
.5in
52
/86
(60
%),
8/9
(89
%)
ICA
ccu
racy
no
td
eter
min
ed
Man
nan
s/an
ti-m
ann
ans
(33
pat
ien
ts(2
con
secu
tiv
ep
osi
tiv
ete
sts)
:se
ns,
73
%;
spec
,8
0%
;P
PV
,0
.36
;N
PV
,0
.95
Po
siti
ve
corr
elat
ion
of
inv
asiv
eca
nd
idia
sis
wit
hb-
glu
can
s(r
=0
.28
,p
=0
.01
)
‘‘S
eria
las
say
sfo
rm
ann
ans
and
anti
-m
ann
ans
inp
atie
nts
atri
skfo
rin
vas
ive
can
did
iasi
sm
ayu
sefu
lly
con
trib
ute
toth
em
anag
emen
to
fsu
chp
atie
nts
’’
Cag
gia
no
etal
.[6
9]
Neu
rolo
gic
alIC
U,
stay
C7
day
s;p
rosp
ecti
ve,
lin
kb
etw
een
colo
niz
atio
nan
din
vas
ive
can
did
iasi
s
51
3ca
nd
idem
iaA
tad
mis
sio
n,
then
ever
y3
day
s(t
od
ay1
5)
Co
lon
izat
ion
:7
6%
atst
ud
yen
try
,9
2%
on
day
15
CI
C0
.5:
16
.6%
atst
ud
yen
try
,7
5%
on
day
15
(all
IC)
CC
I:0
.34
atst
ud
yen
try
,0
.60
on
day
15
No
man
nan
/an
ti-m
ann
and
etec
ted
inp
atie
nts
wit
ho
ut
can
did
emia
Man
nan
s:se
ns,
66
.6%
;sp
ec,
10
0%
An
ti-m
ann
ans:
sen
s,1
00
%;
spec
,1
00
%
‘‘T
hu
s,o
ur
exp
erie
nce
sug
ges
tsth
atm
on
ito
rin
gC
Ico
uld
be
hel
pfu
lin
iden
tify
ing
pat
ien
tsat
risk
of
inv
asiv
efu
ng
alin
fect
ion
.In
add
itio
n,
com
ple
men
tin
gth
isw
ith
anti
-Ca
nd
ida
anti
bo
die
sd
etec
tio
nin
imm
un
oco
mp
eten
tp
atie
nts
wit
hC
IC
0.5
incr
ease
sth
ep
osi
tiv
ep
red
icti
vit
yfo
rin
fect
ion
allo
win
gan
earl
yd
iag
no
sis
of
can
did
emia
’’
1442
Ta
ble
4co
nti
nu
ed
Ref
eren
ces
Ty
pe
of
pat
ien
t;st
ud
yd
esig
nN
um
ber
of
pat
ien
tsN
um
ber
of
inv
asiv
eca
nd
idia
sis
CI
mo
nit
ori
ng
Per
form
ance
of
CI
Per
form
ance
of
com
par
ato
rsA
uth
ors
’co
ncl
usi
on
Po
ster
aro
etal
.[7
0]
Mix
edIC
U,
stay
C5
day
s;p
rosp
ecti
ve,
sin
gle
-ce
nte
r,o
bse
rvat
ion
al:
dia
gn
ost
icv
alu
eo
fB
Gas
say
,C
an
did
asc
ore
,C
I
95
16
(14
inv
asiv
eca
nd
idia
sis)
At
entr
y,
day
3,
and
then
wee
kly
CI
C0
.5:
ov
eral
l,3
5%
;IC
,5
6%
;n
oIC
,3
0%
(p=
0.0
4)
Sen
s,6
4%
;sp
ec,
70
%;
PP
V,
27
%;
NP
V,
92
%;
AU
C,
0.6
3(0
.57
–0
.79
)
Ca
nd
ida
sco
reC
3:
ov
eral
l,2
2%
;IC
,7
5%
;n
oIC
,1
1%
(p\
0.0
01
).S
ens,
86
%;
spec
,8
7%
;P
PV
,5
7%
;N
PV
,9
7%
;A
UC
,0
.80
(0.6
9–
0.9
2)
BG
po
siti
vit
y:
ov
eral
l,2
1%
;IC
,9
4%
;n
oIC
,6
%(p
\0
.00
1).
Sen
s,9
4%
;sp
ec,
94
%;
PP
V,
75
%;
NP
V,
99
%;
AU
C,
0.9
8(0
.92
–1
.00
)C
an
did
asc
ore
C3
and
BG
po
siti
vit
y:
sen
s,1
00
%;
spec
,8
4%
;P
PV
,5
2%
;N
PV
,1
00
%
‘‘A
sin
gle
-po
int
BG
assa
yb
ased
on
ab
loo
dsa
mp
led
raw
nat
the
sep
sis
on
set,
alo
ne
or
inco
mb
inat
ion
wit
hC
an
did
asc
ore
,m
ayg
uid
eth
ed
ecis
ion
tost
art
anti
fun
gal
ther
apy
earl
yin
pat
ien
tsat
risk
for
Ca
nd
ida
infe
ctio
n’’
Pem
anet
al.
[71]
6m
ixed
ICU
,h
igh
risk
of
inv
asiv
eca
nd
idia
sisa
;p
rosp
ecti
ve,
use
fuln
ess
of
CA
GT
A
53
0N
ot
spec
ified
CC
IC
0.4
:2
3/5
3(4
3%
)at
stu
dy
entr
y,
41
/53
(77
%)
aten
do
fst
ud
y
CA
GT
Ap
osi
tiv
ity
:2
2/5
3(4
2%
)at
stu
dy
entr
y,
47
/53
(90
%)
aten
do
fst
ud
y
‘‘T
his
stu
dy
iden
tifi
edp
rev
iou
ssu
rger
yas
the
pri
nci
pal
clin
ical
fact
or
asso
ciat
edw
ith
CA
GT
A-p
osi
tiv
ere
sult
s(s
erolo
gic
ally
pro
ven
can
did
iasi
s)an
dem
ph
asis
esth
eu
tili
tyo
fth
isp
rom
isin
gte
chn
iqu
e,w
hic
hw
asn
ot
infl
uen
ced
by
hig
hC
an
did
aco
lon
izat
ion
or
anti
fun
gal
trea
tmen
t’’
Hal
let
al.
[72]
Mix
edIC
U,
sev
ere
acu
tep
ancr
eati
tis;
retr
osp
ecti
ve,
dia
gn
ost
icv
alu
eo
fB
Gas
say
,C
an
did
asc
ore
,C
I
10
11
8(5
die
d)
No
tsp
ecifi
edC
olo
niz
atio
n:
16
/18
(89
%)
IC,
37
/83
(45
%)
no
IC(p
=0
.00
06
).C
olo
niz
atio
nfo
rIC
:O
R,
4.3
3(1
.07
–1
7.5
);p
=0
.04
CI
C0
.5:
sen
s,6
7%
;sp
ec,
79
%;
PP
V,
43
%;
NP
V,
91
%;
AU
C,
0.7
9(0
.69
–0
.87
)
Ca
nd
ida
sco
re:
sen
s,2
3%
;sp
ec,8
5%
;P
PV
,3
9%
;N
PV
,7
2%
;A
UC
,0
.62
(0.5
2–
0.7
1)
Pre
dic
tiv
eru
leb:
sen
s,6
1%
;sp
ec,
49
%;
PP
V,
21
%;
NP
V,
85
%;
AU
C,
0.5
9(0
.49
–0
.69
)
‘‘In
this
stu
dy
the
Ca
nd
ida
colo
nis
atio
nin
dex
sco
rew
asth
em
ost
accu
rate
and
dis
crim
inat
ive
test
atid
enti
fyin
gw
hic
hp
atie
nts
wit
hse
ver
eac
ute
pan
crea
titi
sar
eat
risk
of
dev
elo
pin
gca
nd
idal
infe
ctio
n.
Ho
wev
erit
slo
wse
nsi
tiv
ity
may
lim
itit
scl
inic
aluse
fuln
ess’
’
1443
The CI and other predictive tools (the Candida scoreand predictive rules) have been specifically developed byusing their positive predictive value for the early identi-fication of high-risk patients who will develop invasivecandidiasis. Nevertheless, it is relevant to note that thenegative predictive values of the CI and all other surro-gate markers of invasive candidiasis are much higher thantheir positive predictive values [21]. Among them, onlythe negative predictive value of the Candida score hasbeen validated in a multicenter prospective clinical trial[75].
Advantages and pitfalls of the use of colonizationindex
To summarize the data reviewed on the use of the CI, itsmain advantages are that it has been successfully used tocharacterize colonization dynamics, assess the signifi-cance of candiduria, and to evaluate the impact ofantifungal prophylaxis. A low CI has a high negativepredictive value for the further development of invasivecandidiasis. Among the pitfalls, it should be emphasizedthat it is work-intensive with a limited bedside practica-bility, that only limited data are available for nonsurgicalpatients, and its cost-effectiveness and usefulness for themanagement of critically ill patients remain to be provenin large prospective clinical trials.
Further perspectives
These observations may pave the way for a new paradigmin the research agenda of the early diagnosis of invasivecandidiasis and take advantage of the disease patho-physiology characterized by a 7–10-day delay betweenexposure to risk factors and infection. We propose astepwise approach to optimize the accuracy of the CI andother clinical tools (Fig. 3).
Provided that preliminary results can be confirmed inlarger cohorts of patients [76], preliminary stratificationaccording to specific genetic polymorphisms could beused as a preliminary step to increase the accuracy ofsubsequent steps. The second step is to take advantage ofthe high negative predictive values of the CI, Candidascore, and/or predictive rules. This strategy will reducethe currently large proportion of useless antifungal treat-ments, frequently started for nonobjective emotionalreasons. The enhanced intrinsic risk of invasive candidi-asis in patients with increased risk identified after thissecond step may further improve the accuracy of bio-markers, such as mannan/anti-mannan antibodies and/orb-glucan, applied in the third step. This approach shouldenhance the ability to identify patients who will trulyT
ab
le4
con
tin
ued
Ref
eren
ces
Ty
pe
of
pat
ien
t;st
ud
yd
esig
nN
um
ber
of
pat
ien
tsN
um
ber
of
inv
asiv
eca
nd
idia
sis
CI
mo
nit
ori
ng
Per
form
ance
of
CI
Per
form
ance
of
com
par
ato
rsA
uth
ors
’co
ncl
usi
on
Tis
sot
etal
.[4
4]
Su
rgic
alIC
U,
recu
rren
tg
astr
oin
test
inal
per
fora
tio
n/
anas
tom
oti
cle
akag
eo
rac
ute
nec
roti
zin
gp
ancr
eati
tis;
pro
spec
tiv
e,ac
cura
cyo
fB
Gan
tig
enem
iafo
rd
iag
no
sis
of
intr
a-ab
do
min
alca
nd
idia
sis
89
2O
nad
mis
sio
n,
then
twic
ea
wee
k
CI
C0
.5:
sen
s,2
6%
;sp
ec,
76
%;
PP
V,
35
%;
NP
V,
67
%;
AU
C,
0.6
7C
CI
C0
.4:
sen
s,1
4%
;sp
ec,
77
%;
PP
V,
23
%;
NP
V,
65
%;
AU
C,
0.4
3
BG
[8
0p
g/m
l(2
9):
sen
s,6
6%
;sp
ec,
83
%;
PP
V,
73
%;
NP
V,
78
%;
AU
C,
0.7
9C
an
did
asc
ore
C3
:se
ns,
86
%;
spec
,5
0%
;P
PV
,5
4%
;N
PV
,8
4%
;A
UC
,0
.61
‘‘B
Gan
tig
enem
iais
sup
erio
rto
Ca
nd
ida
sco
rean
dco
lon
izat
ion
ind
exes
and
anti
cip
ates
dia
gn
osi
so
fb
loo
dcu
ltu
re-n
egat
ive
intr
a-ab
do
min
alca
nd
idia
sis’
’
Sen
sse
nsi
tiv
ity
,S
pec
spec
ifici
ty,
PP
Vp
osi
tiv
ep
red
icti
ve
val
ue,
NP
Vn
egat
ive
pre
dic
tiv
ev
alu
e,A
UC
area
un
der
the
rece
iver
op
erat
ing
char
acte
rist
iccu
rve,
CA
GT
AC
an
did
aa
lbic
an
sg
erm
tub
ean
tib
od
ya
(1)
Acu
tep
ancr
eati
tis
wit
h[
7d
ays
evo
luti
on
,(2
)p
rolo
ng
ed([
14
day
s)IC
Ust
ayan
dC
3ri
skfa
cto
rs(d
iab
etes
mel
litu
s,ex
trar
enal
dep
ura
tio
n,
par
ente
ral
nu
trit
ion
,[7
day
sb
road
-sp
ectr
um
anti
bio
tic
ther
apy
,m
ajo
rab
do
min
alsu
rger
y),
(3)
liv
ertr
ansp
lan
tati
on
,(4
)n
eutr
op
enia
or
bo
ne
mar
row
tran
spla
nta
tio
n,
or
(5)
hig
hC
AG
TA
b[7
3]
1444
benefit from early antifungal treatment and reduceunnecessary overexposure to antifungal agents in patientswith a documented limited risk of infection.
Moreover, the high negative predictive value of thepredictive tools (CI, Candida score, and predictive rules)for the further development of an invasive candidiasismay allow one to stop empirical antifungal treatmentpossibly prescribed before evaluating the patient accord-ing to the proposed stepwise approach.
Conclusion
The development of the CI has enabled better appreciationof the dynamics of Candida colonization in patients athigh risk of invasive candidiasis. It can still be used, as it isin many institutions, for the early detection of patients athigh risk of invasive candidiasis and to guide empirical
antifungal treatment. Continued development and execu-tion of clinical trials in this difficult field are important.Until further progress can be achieved with new clinicalstudies of specific biomarkers involving larger patientcohorts, 20 years after its description, the CI remains oneof the best methods of characterizing the dynamics ofCandida colonization and identifying patients at very lowor increasing risk of invasive candidiasis.
Acknowledgments PE received research grants and/or educationalgrants and/or speaker’s honoraria and/or consultant’s honorariafrom (in alphabetic order): Astellas, Merck, Sharp & Dohme-Chi-bret, and Pfizer.
Conflicts of interest DP declares no disclosure of potentialconflicts of interest regarding their contribution to this study.
Open Access This article is distributed under the terms of the
Creative Commons Attribution Noncommercial License which permits
any noncommercial use, distribution, and reproduction in any medium,
provided the original author(s) and the source are credited.
Fig. 3 Proposed study designs to select high-risk patients whowould truly benefit from early empirical antifungal treatment. First,patients should be stratified according to pathophysiologicalcharacteristics specific to invasive candidiasis (i.e., immunocom-promised individuals, after digestive surgery, other critically illpatients). The three-step approach: In a first optional step, increasedintrinsic risk of invasive candidiasis is identified by specific geneticpolymorphisms related to innate immunity. In the second step,
patients at low objective risk of invasive candidiasis are ruled outby using the high negative predictive values of current risk-assessment strategies (CI, Candida score, peritonitis score, predic-tive rule). In the third step, empirical antifungal treatment is startedearly in high-risk patients identified by increased biomarkersvalues, such as beta-glucan performed only in patients retained bywho met criteria outlined in previous steps. A simplified approachrestricted to steps two and three may be also of potential interest
1445
References
1. Eggimann P, Garbino J, Pittet D (2003)Epidemiology of Candida speciesinfections in critically ill non-immunosuppressed patients. LancetInfect Dis 3:685–702
2. Vincent JL, Rello J, Marshall J, Silva E,Anzueto A, Martin CD, Moreno R,Lipman J, Gomersall C, Sakr Y,Reinhart K, EPIC II Group ofInvestigators (2009) International studyof the prevalence and outcomes ofinfection in intensive care units. JAMA302:2323–2329
3. Pfaller M, Neofytos D, Diekema D,Azie N, Meier-Kriesche HU, Quan SP,Horn D (2012) Epidemiology andoutcomes of candidemia in 3648patients: data from the ProspectiveAntifungal Therapy (PATH Alliance�)registry, 2004-2008. Diagn MicrobiolInfect Dis 74:323–331
4. Guery BP, Arendrup MC, Auzinger G,Azoulay E, Borges Sa M, Johnson EM,Muller E, Putensen C, Rotstein C,Sganga G, Venditti M, Zaragoza CrespoR, Kullberg BJ (2009) Management ofinvasive candidiasis and candidemia inadult non-neutropenic intensive careunit patients: part I. Epidemiology anddiagnosis. Intensive Care Med35:55–62
5. Kett DH, Azoulay E, Echeverria PM,Vincent JL, Extended Prevalence ofInfection in ICUSGoI (2011) Candidabloodstream infections in intensive careunits: analysis of the extendedprevalence of infection in intensive careunit study. Crit Care Med 39:665–670
6. Marchetti O, Bille J, Fluckiger U,Eggimann P, Ruef C, Garbino J,Calandra T, Glauser MP, Tauber MG,Pittet D, Fungal Infection Network of S(2004) Epidemiology of candidemia inSwiss tertiary care hospitals: seculartrends, 1991–2000. Clin Infect Dis38:311–320
7. Eggimann P, Bille J, Marchetti O(2011) Diagnosis of invasivecandidiasis in the ICU. Ann IntensiveCare 1:37
8. Bassetti M, Marchetti M, ChakrabartiA, Colizza S, Garnacho-Montero J, KettDH, Munoz P, Cristini F, AndoniadouA, Viale P, Rocca GD, Roilides E,Sganga G, Walsh TJ, Tascini C,Tumbarello M, Menichetti F, Righi E,Eckmann C, Viscoli C, Shorr AF, LeroyO, Petrikos G, De Rosa FG (2013) Aresearch agenda on the management ofintra-abdominal candidiasis: resultsfrom a consensus of multinationalexperts. Intensive Care Med39:2092–2106
9. Gudlaugsson O, Gillespie S, Lee K,Berg JV, Hu J, Messer S, Herwaldt L,Pfaller M, Diekema D (2003)Attributable mortality of nosocomialcandidemia, revisited. Clin Infect Dis37:1172–1177
10. Zaoutis TE, Argon J, Chu J, Berlin JA,Walsh TJ, Feudtner C (2005) Theepidemiology and attributable outcomesof candidemia in adults and childrenhospitalized in the United States: apropensity analysis. Clin Infect Dis41:1232–1239
11. Prowle JR, Echeverri JE, Ligabo EV,Sherry N, Taori GC, Crozier TM, HartGK, Korman TM, Mayall BC, JohnsonPD, Bellomo R (2011) Acquiredbloodstream infection in the intensivecare unit: incidence and attributablemortality. Crit Care 15:R100
12. Morrell M, Fraser VJ, Kollef MH(2005) Delaying the empiric treatmentof candida bloodstream infection untilpositive blood culture results areobtained: a potential risk factor forhospital mortality. Antimicrob AgentsChemother 49:3640–3645
13. Garey KW, Rege M, Pai MP, MingoDE, Suda KJ, Turpin RS, Bearden DT(2006) Time to initiation of fluconazoletherapy impacts mortality in patientswith candidemia: a multi-institutionalstudy. Clin Infect Dis 43:25–31
14. Azoulay E, Dupont H, Tabah A,Lortholary O, Stahl JP, Francais A,Martin C, Guidet B, Timsit JF (2012)Systemic antifungal therapy in criticallyill patients without invasive fungalinfection. Crit Care Med 40:813–822
15. Sipsas NV, Lewis RE, Tarrand J,Hachem R, Rolston KV, Raad II,Kontoyiannis DP (2009) Candidemia inpatients with hematologic malignanciesin the era of new antifungal agents(2001–2007): stable incidence butchanging epidemiology of a stillfrequently lethal infection. Cancer115:4745–4752
16. Lortholary O, Desnos-Ollivier M,Sitbon K, Fontanet A, Bretagne S,Dromer F, French Mycosis Study G(2011) Recent exposure to caspofunginor fluconazole influences theepidemiology of candidemia: aprospective multicenter study involving2,441 patients. Antimicrob AgentsChemother 55:532–538
17. Pappas PG, Kauffman CA, Andes D,Benjamin DK Jr, Calandra TF, EdwardsJE Jr, Filler SG, Fisher JF, Kullberg BJ,Ostrosky-Zeichner L, Reboli AC, RexJH, Walsh TJ, Sobel JD, InfectiousDiseases Society of A (2009) Clinicalpractice guidelines for the managementof candidiasis: 2009 update by theInfectious Diseases Society of America.Clin Infect Dis 48:503–535
18. Cornely OA, Bassetti M, Calandra T,Garbino J, Kullberg BJ, Lortholary O,Meersseman W, Akova M, ArendrupMC, Arikan-Akdagli S, Bille J,Castagnola E, Cuenca-Estrella M,Donnelly JP, Groll AH, Herbrecht R,Hope WW, Jensen HE, Lass-Florl C,Petrikkos G, Richardson MD, RoilidesE, Verweij PE, Viscoli C, Ullmann AJ,Group EFIS (2012) ESCMID guidelinefor the diagnosis and management ofCandida diseases 2012: non-neutropenic adult patients. ClinMicrobiol Infect 18(Suppl 7):19–37
19. Eggimann P, Garbino J, Pittet D (2003)Management of Candida speciesinfections in critically ill patients.Lancet Infect Dis 3:772–785
20. Playford EG, Eggimann P, Calandra T(2008) Antifungals in the ICU. CurrOpin Infect Dis 21:610–619
21. Eggimann P, Ostrosky-Zeichner L(2010) Early antifungal interventionstrategies in ICU patients. Curr OpinCrit Care 16:465–469
22. Montravers P, Dupont H, Eggimann P(2013) Intra-abdominal candidiasis: theguidelines-forgotten non-candidemicinvasive candidiasis. Intensive CareMed 39:2226–2230
23. Pittet D, Monod M, Filthuth I, Frenk E,Suter PM, Auckenthaler R (1991)Contour-clamped homogeneous electricfield gel electrophoresis as a powerfulepidemiologic tool in yeast infections.Am J Med 91:256S–263S
24. Pittet D, Monod M, Suter PM, Frenk E,Auckenthaler R (1994) Candidacolonization and subsequent infectionsin critically ill surgical patients. AnnSurg 220:751–758
25. van de Veerdonk FL, Kullberg BJ,Netea MG (2010) Pathogenesis ofinvasive candidiasis. Curr Opin CritCare 16:453–459
26. Magill SS, Swoboda SM, Shields CE,Colantuoni EA, Fothergill AW, MerzWG, Lipsett PA, Hendrix CW (2009)The epidemiology of Candidacolonization and invasive candidiasis ina surgical intensive care unit wherefluconazole prophylaxis is utilized:follow-up to a randomized clinical trial.Ann Surg 249:657–665
1446
27. Khan ZU, Chandy R, Metwali KE(2003) Candida albicans strain carriagein patients and nursing staff of anintensive care unit: a study ofmorphotypes and resistotypes. Mycoses46:479–486
28. Marco F, Lockhart SR, Pfaller MA,Pujol C, Rangel-Frausto MS, Wiblin T,Blumberg HM, Edwards JE, Jarvis W,Saiman L, Patterson JE, Rinaldi MG,Wenzel RP, Soll DR (1999) Elucidatingthe origins of nosocomial infectionswith Candida albicans by DNAfingerprinting with the complex probeCa3. J Clin Microbiol 37:2817–2828
29. Nucci M, Anaissie E (2001) Revisitingthe source of candidemia: skin or gut?Clin Infect Dis 33:1959–1967
30. Filler SG (2012) Insights from humanstudies into the host defense againstcandidiasis. Cytokine 58:129–132
31. Charlier C, Hart E, Lefort A, Ribaud P,Dromer F, Denning DW, Lortholary O(2006) Fluconazole for the managementof invasive candidiasis: where do westand after 15 years? J AntimicrobChemother 57:384–410
32. Maertens J, Marchetti O, Herbrecht R,Cornely OA, Fluckiger U, Frere P,Gachot B, Heinz WJ, Lass-Florl C,Ribaud P, Thiebaut A, Cordonnier C,Third European Conference onInfections in Leukemia (2011)European guidelines for antifungalmanagement in leukemia andhematopoietic stem cell transplantrecipients: summary of the ECIL3–2009 update. Bone MarrowTransplant 46:709–718
33. Dannaoui E, Desnos-Ollivier M,Garcia-Hermoso D, Grenouillet F,Cassaing S, Baixench MT, Bretagne S,Dromer F, Lortholary O, FrenchMycoses Study G (2012) Candida spp.with acquired echinocandin resistance,France, 2004–2010. Emerg Infect Dis18:86–90
34. Pacholczyk M, Lagiewska B, Lisik W,Wasiak D, Chmura A (2011) Invasivefungal infections following livertransplantation—risk factors, incidenceand outcome. Ann Transplant 16:14–16
35. Schaenman JM (2013) Is universalantifungal prophylaxis mandatory inlung transplant patients? Curr OpinInfect Dis 26:317–325
36. Toth A, Csonka K, Jacobs C, VagvolgyiC, Nosanchuk JD, Netea MG, Gacser A(2013) Candida albicans and Candidaparapsilosis induce different T-cellresponses in human peripheral bloodmononuclear cells. J Infect Dis208:690–698
37. Charles PE, Doise JM, Quenot JP, AubeH, Dalle F, Chavanet P, Milesi N, AhoLS, Portier H, Blettery B (2003)Candidemia in critically ill patients:difference of outcome between medicaland surgical patients. Intensive CareMed 29:2162–2169
38. Charles PE, Dalle F, Aube H, Doise JM,Quenot JP, Aho LS, Chavanet P,Blettery B (2005) Candida spp.colonization significance in critically illmedical patients: a prospective study.Intensive Care Med 31:393–400
39. Eggimann P, Calandra T, Fluckiger U,Bille J, Garbino J, Glauser MP,Marchetti O, Ruef C, Tauber M, PittetD, Fungal Infection Network ofSwitzerland (2005) Invasivecandidiasis: comparison of managementchoices by infectious disease andcritical care specialists. Intensive CareMed 31:1514–1521
40. Garbino J, Lew DP, Romand JA,Hugonnet S, Auckenthaler R, Pittet D(2002) Prevention of severe Candidainfections in nonneutropenic, high-risk,critically ill patients: a randomized,double-blind, placebo-controlled trial inpatients treated by selective digestivedecontamination. Intensive Care Med28:1708–1717
41. Vincent JL, Anaissie E, Bruining H,Demajo W, el-Ebiary M, Haber J,Hiramatsu Y, Nitenberg G, NystromPO, Pittet D, Rogers T, Sandven P,Sganga G, Schaller MD, Solomkin J(1998) Epidemiology, diagnosis andtreatment of systemic Candida infectionin surgical patients under intensive care.Intensive Care Med 24:206–216
42. Solomkin JS, Flohr AB, Quie PG,Simmons RL (1980) The role ofCandida in intraperitoneal infections.Surgery 88:524–530
43. Calandra T, Bille J, Schneider R,Mosimann F, Francioli P (1989)Clinical significance of Candidaisolated from peritoneum in surgicalpatients. Lancet 2:1437–1440
44. Tissot F, Lamoth F, Hauser PM, OraschC, Fluckiger U, Siegemund M,Zimmerli S, Calandra T, Bille J,Eggimann P, Marchetti O, FungalInfection Network of Switzerland(FUNGINOS) (2013) Beta-glucanantigenemia anticipates diagnosis ofblood culture-negative intraabdominalcandidiasis. Am J Respir Crit Care Med188:1100–1109
45. Eggimann P, Francioli P, Bille J,Schneider R, Wu MM, Chapuis G,Chiolero R, Pannatier A, Schilling J,Geroulanos S, Glauser MP, Calandra T(1999) Fluconazole prophylaxisprevents intra-abdominal candidiasis inhigh-risk surgical patients. Crit CareMed 27:1066–1072
46. Senn L, Eggimann P, Ksontini R,Pascual A, Demartines N, Bille J,Calandra T, Marchetti O (2009)Caspofungin for prevention of intra-abdominal candidiasis in high-risksurgical patients. Intensive Care Med35:903–908
47. Goldstein E, Hoeprich PD (1972)Problems in the diagnosis and treatmentof systemic candidiasis. J Infect Dis125:190–193
48. Klein JJ, Watanakunakorn C (1979)Hospital-acquired fungemia. Its naturalcourse and clinical significance. Am JMed 67:51–58
49. Solomkin JS, Flohr AM, Simmons RL(1982) Indications for therapy forfungemia in postoperative patients.Arch Surg 117:1272–1275
50. Vazquez JA, Sanchez V, DmuchowskiC, Dembry LM, Sobel JD, Zervos MJ(1993) Nosocomial acquisition ofCandida albicans: an epidemiologicstudy. J Infect Dis 168:195–201
51. Tran LT, Auger P, Marchand R, CarrierM, Pelletier C (1997) Epidemiologicalstudy of Candida spp. colonization incardiovascular surgical patients.Mycoses 40:169–173
52. Yazdanparast K, Auger P, Marchand R,Carrier M, Cartier R (2001) Predictivevalue of Candida colonization index in131 patients undergoing two differentcardiovascular surgical procedures.J Cardiovasc Surg 42:339–343
53. Eloy O, Marque S, Mourvilliers B, PinaP, Allouch PY, Pangon B, Bedos JP,Ghnassia JC (2006) Contribution of thePittet’s index, antigen assay, IgM, andtotal antibodies in the diagnosis ofinvasive candidiasis in intensive careunit. J Mycol Med 16:113–118
54. Agvald-Ohman C, Klingspor L,Hjelmqvist H, Edlund C (2008)Invasive candidiasis in long-termpatients at a multidisciplinary intensivecare unit: candida colonization index,risk factors, treatment and outcome.Scand J Infect Dis 40:145–153
55. Massou S, Ahid S, Azendour H,Bensghir M, Mounir K, Iken M,Lmimouni BE, Balkhi H, Drissi KamiliN, Haimeur C (2013) Systemiccandidiasis in medical intensive careunit: analysis of risk factors and thecontribution of colonization index.Pathol Biol (Paris) 61:108–112
56. Chabasse D (2001) Yeast count inurine. Review of the literature andpreliminary results of a multicenterprospective study carried out in 15hospital centers. Ann Fr Anesth Reanim20:400–406
1447
57. Dubau B, Triboulet C, Winnock S(2001) Use of the colonization index.Ann Fr Anesth Reanim 20:418–420
58. Sellami A, Sellami H, Makni F, BahloulM, Cheikh-Rouhou F, Bouaziz M,Ayadi A (2006) Candiduria in intensivecare unit: significance and value ofyeast numeration in urine. Ann FrAnesth Reanim 25:584–588
59. Ergin F, Eren Tulek N, Yetkin MA,Bulut C, Oral B, Tuncer Ertem G(2013) Evaluation of Candidacolonization in intensive care unitpatients and the use of Candidacolonization index. Mikrobiyol Bul47:305–317
60. Laverdiere M, Rotstein C, Bow EJ,Roberts RS, Ioannou S, Carr D,Moghaddam N (2000) Impact offluconazole prophylaxis on fungalcolonization and infection rates inneutropenic patients. The CanadianFluconazole Study. J AntimicrobChemother 46:1001–1008
61. Normand S, Francois B, Darde ML,Bouteille B, Bonnivard M, Preux PM,Gastinne H, Vignon P (2005) Oralnystatin prophylaxis of Candida spp.colonization in ventilated critically illpatients. Intensive Care Med31:1508–1513
62. Giglio M, Caggiano G, Dalfino L,Brienza N, Alicino I, Sgobio A, FavaleA, Coretti C, Montagna MT, Bruno F,Puntillo F (2012) Oral nystatinprophylaxis in surgical/trauma ICUpatients: a randomised clinical trial. CritCare 16:R57
63. Chen Z, Yang CL, He HW, Zeng J(2013) The clinical research of nystatinin prevention of invasive fungalinfections in patients on mechanicalventilation in intensive care unit.Zhonghua Wei Zhong Bing Ji Jiu YiXue 25:475–478
64. Piarroux R, Grenouillet F, Balvay P,Tran V, Blasco G, Millon L, Boillot A(2004) Assessment of preemptivetreatment to prevent severe candidiasisin critically ill surgical patients. CritCare Med 32:2443–2449
65. Eren A, Aydogan S, Kalkanci A,Kustimur S (2007) The relationshipbetween Candida albicans colonizationindices and the presence of specificantibodies in non-neutropenic intensivecare unit patients. Mikrobiyol Bul41:253–259
66. Wang DH, Gao XJ, Wei LQ, Xia R,Sun L, Li Q, Peng M, Qin YZ (2009)The preemptive treatment of invasiveCandida infection with reference ofcorrected colonization index incritically ill patients: a multicenter,prospective, randomized controlledclinical study. Zhongguo Wei ZhongBing Ji Jiu Yi Xue 21:525–528
67. Leon C, Ruiz-Santana S, Saavedra P,Galvan B, Blanco A, Castro C, BalasiniC, Utande-Vazquez A, Gonzalez deMolina FJ, Blasco-Navalproto MA,Lopez MJ, Charles PE, Martin E,Hernandez-Viera MA, Cava Study G(2009) Usefulness of the ‘‘Candidascore’’ for discriminating betweenCandida colonization and invasivecandidiasis in non-neutropenic criticallyill patients: a prospective multicenterstudy. Critical Care Med 37:1624–1633
68. Ellis M, Al-Ramadi B, Bernsen R,Kristensen J, Alizadeh H, Hedstrom U(2009) Prospective evaluation ofmannan and anti-mannan antibodies fordiagnosis of invasive Candidainfections in patients with neutropenicfever. J Med Microbiol 58:606–615
69. Caggiano G, Puntillo F, Coretti C,Giglio M, Alicino I, Manca F, Bruno F,Montagna MT (2011) Candidacolonization index in patients admittedto an ICU. Int J Mol Sci 12:7038–7047
70. Posteraro B, De Pascale G, TumbarelloM, Torelli R, Pennisi MA, Bello G,Maviglia R, Fadda G, Sanguinetti M,Antonelli M (2011) Early diagnosis ofcandidemia in intensive care unitpatients with sepsis: a prospectivecomparison of (1?3)-beta-D-glucanassay, Candida score, and colonizationindex. Crit Care 15:R249
71. Peman J, Zaragoza R, Quindos G,Alkorta M, Cuetara MS, Camarena JJ,Ramirez P, Gimenez MJ, Martin-Mazuelos E, Linares-Sicilia MJ, PontonJ, Study group Candida albicans GermTube Antibody Detection in CriticallyIll Patients (2011) Clinical factorsassociated with a Candida albicansgerm tube antibody positive test inintensive care unit patients. BMC InfectDis 11:60
72. Hall AM, Poole LA, Renton B,Wozniak A, Fisher M, Neal T, HalloranCM, Cox T, Hampshire PA (2013)Prediction of invasive candidalinfection in critically ill patients withsevere acute pancreatitis. Crit Care17:R49
73. Ostrosky-Zeichner L, Shoham S,Vazquez J, Reboli A, Betts R, BarronMA, Schuster M, Judson MA, RevankarSG, Caeiro JP, Mangino JE, Mushatt D,Bedimo R, Freifeld A, Nguyen MH,Kauffman CA, Dismukes WE, WestfallAO, Deerman JB, Wood C, Sobel JD,Pappas PG (2014) MSG-01: arandomized, double-blind, placebo-controlled trial of caspofunginprophylaxis followed by preemptivetherapy for invasive candidiasis in high-risk adults in the critical care setting.Clin Infect Dis 58:1219–1226
74. Timsit JF, Azoulay E, Cornet M,Gangneux JP, Jullien V, Vesin A, SchirE, Wolff M (2013) EMPIRICUSmicafungin versus placebo duringnosocomial sepsis in Candida multi-colonized ICU patients with multipleorgan failures: study protocol for arandomized controlled trial. Trials14:399
75. Charles PE, Castro C, Ruiz-Santana S,Leon C, Saavedra P, Martin E (2009)Serum procalcitonin levels in criticallyill patients colonized with Candida spp:new clues for the early recognition ofinvasive candidiasis? Intensive CareMed 35:2146–2150
76. Wojtowicz A, Tissot F, Lamoth F,Orasch C, Eggimann P, Siegemund M,Zimmerli S, Flueckiger UM, Bille J,Calandra T, Marchetti M, Bochud PY(2014) Polymorphisms in TNF-aincrease susceptibility to intra-abdominal candida infection in highrisk surgical ICU patients. Critical CareMed 42:e304
1448