Candida Colonization Index and Subsequent Infection In

Intensive Care Med (2014) 40:1429–1448 DOI 10.1007/s00134-014-3355-z MY PAPER 20 YEARS LATER

Philippe Eggimann Candida colonization index and subsequent Didier Pittet infection in critically ill surgical patients: 20 years later

Received: 20 March 2014 Abstract Introduction: For dec- Results: The recognition of specific Accepted: 23 May 2014 ades, clinicians dealing with characteristics among underlying Published online: 17 June 2014 immunocompromised and critically ill conditions, such as neutropenia, solid The Author(s) 2014. This article is patients have perceived a link organ transplantation, and surgical published with open access at and nonsurgical critical illness, has Springerlink.com between Candida colonization and subsequent infection. However, the enabled the description of distinct pathophysiological progression from epidemiological patterns in the colonization to infection was clearly development of invasive candidiasis. established only through the formal Conclusions: Despite its limited description of the colonization index bedside practicality and before con- (CI) in critically ill patients. Unfortu- firmation of potentially more accurate nately, the literature reflects intense predictors, such as specific biomark- confusion about the pathophysiology ers, the CI remains an important way P. Eggimann of invasive candidiasis and specific to characterize the dynamics of colo- Adult Critical Care Medicine and Burn associated risk factors. Meth- nization, which increases early in Unit, Centre Hospitalier Universitaire patients who develop invasive Vaudois (CHUV), Lausanne, Switzerland ods: We review the contribution of e-mail: [email protected] the CI in the field of Candida infection candidiasis. and its development in the 20 years D. Pittet ()) following its original description in Keywords Candida albicans Faculty of Medicine, Infection Control 1994. The development of the CI Candidemia Invasive candidiasis Program and WHO Collaborating Center on enabled an improved understanding of Colonization Colonization index Patient Safety, University of Geneva the pathogenesis of invasive candidi- Empirical treatment Hospitals (HUG), Rue Gabrielle-Perret- Nosocomial infections Gentil 4, 1205 Geneva, Switzerland asis and the use of targeted empirical e-mail: [email protected] antifungal therapy in subgroups of Tel.: ?41 22 372 9844 patients at increased risk for infection.

Introduction Early empirical treatment of severe candidiasis has improved survival, but is responsible for the overuse of Candida spp. colonization occurs in up to 80 % of critically antifungals [12–14]. This overuse not only contributes to ill patients after 1 week in intensive care [1–3]. This very high a huge financial burden, but has also promoted a shift to proportion contrasts strongly with the low rate of invasive Candida species with reduced susceptibility to antifungal candidiasis in less than 10 % of them [4, 5]. Despite recent agents [15, 16]. Unfortunately, recent guidelines resulting advances in microbiological techniques, early diagnosis of from expert consensus failed to provide high-level rec- invasive candidiasis remains problematic and microbiologi- ommendations about empirical antifungal treatment or to cal documentation often occurs late in the course of infection clarify the nature of such treatment strategies [8, 17, 18]. [6–8]. This explains its high crude and attributable mortality, Despite limited levels of evidence, empirical treatment i.e., in the range reported for septic shock [9–11]. currently relies on the positive predictive value of risk 1430 assessment strategies, such as the colonization index (CI), Solid organ transplant recipients Candida score, and predictive rules based on combina- tions of risk factors [19–21]. Anti-rejection therapy increases the risk for Aspergillus Improved knowledge of the pathophysiological spec- and other filamentous fungal infections in solid organ ificities of invasive candidiasis should promote better use transplant recipients, except among those receiving small of clinical tools in the evaluation of patients who could bowel and liver transplants, in whom additional specific truly benefit from early antifungal therapy [22]. In con- surgical factors further modify the epidemiology of dis- trast to most bacterial infections, invasive candidiasis is ease [34, 35]. Prolonged and intense impairment of characterized by a 7–10-day delay between exposure to cellular T cell-mediated immune response may contribute risk factors and infection development [23–25]. This time to the high proportion of cutaneous and mucosal candi- window provides a unique opportunity for the imple- diasis [36]. This factor has justified the use of systematic mentation of a structured approach to the rigorous antifungal prophylaxis in these patients in whom, as selection of patients likely to benefit from early empirical among neutropenic patients and bone marrow transplant antifungal treatment [19, 26]. recipients, breakthrough Candida spp. infection develops The CI is the most widely studied clinical tool for the with strains resistant to the agent used [18]. early risk assessment of invasive candidiasis among at- risk patients. The primary objective of this review is to highlight its role in this setting. The use of colonization Nonsurgical critically ill patients dynamics, as assessed by the CI, for the early detection of patients likely to benefit from early antifungal treatment is In nonsurgical critically ill patients, the continuous and also explored in comparison with other tools currently prolonged support of failing organs and the selective available in clinical practice. Finally, we also propose a pressure of broad-spectrum antibiotics constitute key risk research agenda for the next decade. factors for invasive candidiasis. Support of organ failure requires the use of numerous devices, such as intravascular catheters, endotracheal tubes, naso- and oro-gastric Pathophysiology of invasive candidiasis tubes, and Foley catheters, which are frequently colonized by Candida spp. as a result of the high affinity of their biofilm [1]. These specificities may explain progressive Nosocomial exogenous transmission of Candida has been colonization in a high proportion of patients after pro- well described [27]. Of note, both endogenous and longed stay in the intensive care unit (ICU) [6, 37, 38]. exogenous colonization can co-exist in the clinical setting They may also explain a higher proportion of catheter- [1]. However, carefully designed studies using genotyping related infections in the absence of severe immune of Candida strains confirmed that endogenous coloniza- impairment [39]. In this context, the role of exposure to tion is responsible for the large majority of severe steroids largely used to improve the speed of recovery candidiasis [28, 29]. from circulatory failure remains to be determined. The epidemiology of invasive candidiasis has some Limited data are available on the usefulness of pro- important differences in immunocompromised and criti- phylaxis and recent guidelines did not include cally ill patients (Fig. 1). recommendation for antifungal prophylaxis in nonsurgical critically ill patients [17, 18, 40]. Neutropenic and bone marrow transplant recipients

Neutrophils are essential in the defense against invasive Critically ill patients recovering from abdominal candidiasis [30]. Their prolonged absence in neutropenic surgery patients or those with functional impairment after bone marrow transplantation, combined with the selective Additional factors should be considered in patients after pressure of frequent and repetitive exposure to antibacte- abdominal surgery [41]. Any perforation or opening of the rial agents, plays a major role in the development of digestive tract results in contamination of the peritoneum invasive candidiasis in these patients. A high density of by bowel flora. In most cases, surgical cleaning of the Candida spp. on mucosal surfaces injured by chemother- abdominal cavity, eventually combined with antibiotics, is apy develops progressively and is responsible for fungal sufficient to allow full recovery [42]. Except for patients translocation with a high risk of candidemia. The use of presenting with nosocomial peritonitis and some of those systematic antifungal prophylaxis in the past three decades presenting with septic shock and multiple organ failure, explains the epidemiology of Candida infection, typi- the identification of Candida spp. has no clinical signifi- cally characterized by breakthrough invasive candidiasis cance under these conditions [22]. Recurrent peritonitis [31, 32]. Infection is almost always caused by a Candida following anastomotic leakage or persistent inflammation strain resistant to the antifungal agent used [15, 33]. may be required for the progression of Candida spp. 1431

Candidaspp. Gram positive Gram negative Red cell Macrophage

Neutropenic patients Solid organ transplant recipients

Candida spp. Gram positive Gram negative Red cell Macrophage Lymphocyte Neutrophil

Critically ill surgical patients Critically ill non-surgical patients

Fig. 1 Pathophysiology of invasive candidiasis in different patient populations. Detailed mechanisms are summarized in the corresponding sections of the text 1432 colonization to invasive candidiasis [43]. However, the 1970s of Candida infection in surgical patients, mani- development of invasive candidiasis requires concomitant festing as ‘‘disseminated candidiasis’’ and characterized exposure to additional factors, such as an increasing by persistent fungemia and multiple disseminated mic- amount of Candida spp. in a non-functioning bowel, roabscesses on autopsy, several authors reported that it prolonged antibiotic treatment, and/or requirement for might be preceded by fungal colonization. Klein and organ support. In contrast to nonsurgical patients, these Watanakunakorn [48] reported the presence of visceral particularities explain why a high proportion of invasive microabscesses in up to 30 % of patients with Candida candidiasis cases do not manifest with candidemia and spp. colonization and recommended treatment of all such develop only late in the course of the disease [44]. patients with critical illness or persistent candidemia. Non-candidemic invasive candidiasis includes intra- Solomkin et al. confirmed this concept in a detailed abdominal abscess and peritonitis defined by one of the review of 63 surgical patients with postoperative candi- following culture results from specimens obtained at sur- demia. Among 51 patients who developed candidemia as a gery: (1) monomicrobial growth of Candida spp.; (2) any late complication of intra-abdominal infection, adequate amount of Candida spp. growth within a mixed-flora antifungal treatment ([3 mg/kg amphotericin B) resulted abscess; or (3) moderate or heavy growth of Candida spp. in in survival of 10/15 patients compared with 6/36 untreated mixed-flora peritonitis treated with appropriate antibacte- patients. Autopsies revealed visceral Candida spp. mic- rial therapy according to susceptibility testing [43–46]. roabscesses in 7/20 patients [42, 49]. Moreover, a digestive Two small prospective studies, including one placebo- fungal source was identified in nearly all patients, and controlled, suggested that antifungal prophylaxis in candidemia was considered to originate from intestinal patients presenting with anastomotic leakage after Candida spp. growth in the presence of mucosal loss. abdominal surgery may prevent the development of Calandra et al. provided further evidence of the clinical invasive candidiasis [18, 45, 46]. significance of Candida colonization of the peritoneum in In patients with severe acute necrotizing pancreatitis, surgical patients [39]. In a series of 49 patients with invasive candidiasis develops with similar pathophysio- spontaneous perforation (n = 28) or surgical opening of logical characteristics. Progressive colonization of the the digestive tract (n = 21), 19 (39 %) patients developed bowel within the first 2 weeks of the disease results in invasive candidiasis, including 7 cases of intra-abdominal translocation into necrotic tissues, and fungal infections abscesses and 12 cases of peritonitis [43]. Invasive can- have been documented in up to 10 % of patients not didiasis developed significantly more frequently in patients exposed to antibiotics. who had undergone surgery for acute pancreatitis than in those with gastrointestinal perforations (90 vs. 32 %, p = 0.005) or other disorders, such as diseases of the bil- Critically ill patients with Candida isolated iary tract or colon (90 vs. 17 %, p = 0.003). Invasive from the respiratory tract candidiasis was associated significantly with an initially high or increasing amount of Candida spp., as assessed by Candida spp. colonization of the airway is frequently semiquantitative cultures. Compared with uninfected reported in mechanically ventilated critically ill patients, patients, Candida spp. showed light, moderate, and heavy and its clinical significance is difficult to evaluate. Can- growth from the first positive specimen in 26 (87 %) vs. 9 dida has a low affinity for alveolar pneumocytes and (47 %), 4 (13 %) vs. 6 (32 %), and 0 vs. 4 (21 %) infected histologically documented pneumonia has been rarely patients, respectively (p = 0.005). Moreover, the amount reported. Hematogenous dissemination in the context of of Candida spp. cultured from subsequent specimens candidemia [1] may be responsible for multiple pulmon- increased in 15 (79 %) infected patients and only 2 (7 %) ary abscesses and should be viewed as a distinct entity. uninfected patients (p \ 0.001) [39]. Hence the existence of true candidal pneumonia is doubtful and recovery of Candida spp. from the respiratory tract should generally be considered as colonization and does not justify antifungal therapy [17–19]. Candida colonization index: the missing link

In the early 1990s, the origin of Candida spp. infection was highly debated [50]. Whether the infection was Candida colonization exogenous or endogenous in origin, and whether patients could become colonized with strains from the surrounding The distinction between patients with Candida spp. col- environment or with their ‘‘own’’ strains of Candida spp. onization who did not require treatment and those likely was uncertain. We tested the hypothesis that yeast colo- to develop invasive infection was identified as a difficult nization would precede infection in critically ill patients clinical challenge in the field of infectious diseases more [23]. We determined the genotypic characteristics of 322 than 40 years ago [47]. After the description in the early Candida spp. collected prospectively from 29 critically ill 1433 patients who developed significant colonization under the total number of body sites cultured [24]. With the routine culture surveillance two to three times weekly exception of blood cultures, samples collected from body over a 6-month period. These patients belonged to a sites other than those routinely screened are also considered cohort of 650 surgical ICU patients. Significant coloni- in the CI (Table 1). Only strains of Candida spp. with the zation was defined as the presence of Candida spp. in same genetic identity are considered in calculating the CI. three or more samples taken from one or more sites on at All 29 included patients were heavily colonized with least two consecutive screening days. Candida spp. Candida spp., but with no significant difference in the strains isolated from an individual patient had an identical reason for ICU admission and comorbidities between genetic pattern, even when isolated from different body infected and uninfected patients. The Acute Physiology sites, and this pattern remained the same over a prolonged and Chronic Health Evaluation II (APACHE II) score and period of up to 140 days. No horizontal transmission duration of antibiotic exposure before colonization were could be demonstrated during the study period. Invasive higher among the 11 patients who ultimately developed candidiasis, including 8 cases of candidemia, occurred in invasive candidiasis. In these patients, colonization pre- 11/29 (38 %) patients. All patients who developed ceded infection by an average of 25 days (range 6–70). infection had previous colonization by strains with iden- The 18 patients who did not develop infection were col- tical genetic patterns. Our intensive surveillance failed to onized for a mean of 29 days (range 5–140). A total of document cross-transmission of Candida spp. between 153 [5.3 ± 1 (range 3–8) per patient] distinct body sites colonized and non-colonized patients. were tested, and this number did not differ significantly between the two groups of patients. The CI was calculated daily, and the highest values Colonization index obtained before invasive candidiasis development were compared with those recorded for patients who did not The development of the CI has been viewed as a major develop candidiasis. The average CI differed significantly conceptual advance in the characterization of supporting between colonized and infected patients (0.47 vs. 0.70, the progression from colonization to infection in surgical p \ 0.01). All patients who ultimately developed infec- patients. tion reached the threshold value of 0.5 before infection The Candida CI is defined as the ratio of the number of (Fig. 2). Importantly, infected patients reached the distinct non-blood body sites colonized by Candida spp. to threshold CI value an average of 6 days (range 2–21)

Table 1 Description of the colonization index

Index Definition Formula Number of sites colonized Colonization index (CI) Ratio of the number of distinct non-blood body sites CI ¼ Number of sites cultured colonized by Candida spp. to the total number of body sites cultured. With the exception of blood cultures, samples collected from body sites other than those routinely screened are also considered. Only strains of Candida spp. with identical electrophoretic karyotypes are considered when Number of sites colonized CCI ¼ calculating the CI Number of sites cultured Corrected colonization Product of the CI and the ratio of the number of Number of sites with heavy growth index (CCI) distinct body sites showing heavy growth (??? from semi-quantitative culture or B105 in urine or Total positive sites gastric juice) to the total number of body sites with Candida spp. heavy growth Use of the CIa In patients perceived to be at risk of developing an invasive candidiasis: twice weekly surveillance culture 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 only b Except for specific invasive candidiasis discussed in the section been performed in the original study. However, different species of on pathophysiology of non-candidemic invasive candidiasis, Can- Candida may contribute to the burden of fungi which will eventually dida spp. isolated from these sites should be considered as result in symptomatic infection. Accordingly, we suggest to con- colonization sider all species of Candida identified for the calculation of the CI 1434

Colonization intex or a methodology derived from its original description to assess the dynamics of Candida colonization in different subgroups of critically ill patients at risk of invasive 0.8 candidiasis. Unfortunately, these data have not been validated in large multicenter trials. 0.6 Several studies have indirectly suggested the validity and potential usefulness of the CI, but almost exclusively 0.4 in surgical patients (Table 2). This index has been used to characterize colonization dynamics [37, 38, 51–55], 0.2 assess the significance of candiduria in critically ill patients [38, 56–59], and evaluate the impact of antifungal prophylaxis in subgroups of patients at risk of Days candidiasis [40, 46, 60–63]. At least four studies have reported the use of the CI to guide empirical antifungal Fig. 2 Colonization index. This index is defined as the ratio of the treatment. Although major methodological flaws may number of distinct non-blood body sites with Candida spp. preclude the validation of their findings, this application colonization to the total number of distinct body sites tested. It was recorded for each patient from the first day of colonization reduced the incidence of ICU-acquired invasive candidi- until discharge from the intensive care unit (uninfected patients) or asis in all four studies (Table 3)[57, 64–66]. severe candidiasis development (infected patients). Black circles Dubau et al. [57] prospectively monitored CIs weekly represent patients who developed severe candidiasis, white circles in 89/669 consecutive patients staying in the ICU for represent patients who remained colonized. Reproduced with more than 7 days after digestive surgery who had pro- permission from Pittet et al. [24] tein C levels greater than 100 mg/l or received antibiotic treatment for more than 7 days. All patients developing before candidiasis. All but one patient reached this a CI above 0.5 were empirically treated with antifungals threshold at least 3 days before the time of infection. The until it decreased below 0.5. The proportion of patients sensitivity, specificity, and positive and negative predic- with CIs greater than 0.5 increased from 6 to 25, 40, 55, tive values of the CI were determined to be 100, 69, 66, and 59 % after 1, 2, 3, 4, and more than 4 weeks, and 100 %, respectively. respectively. Only one of the 35 patients with CIs greater than 0.5 who were empirically treated with Corrected colonization index antifungal medications developed candidiasis; the degree of colonization decreased rapidly in the other 34 In a post hoc analysis, we developed the corrected colo- patients. nization index (CCI) by taking into account the amount of In a before/after trial, Piarroux et al. [64] prospectively Candida spp. recovered by semiquantitative cultures [24]. used weekly CIs and CCIs to assess the intensity of The CCI is the product of the CI and the ratio of the Candida spp. colonization in 478 surgical patients staying number of distinct body sites showing heavy growth in the ICU. Patients with CCIs greater than 0.4 received (??? or at most 105 in urine or gastric juice) to the total empirical antifungal treatment. Compared with an his- of distinct body sites with Candida spp. growth (Table 1). torical cohort of 455 control subjects, the incidence of The mean CCI differed significantly between colo- invasive candidiasis was lower among these patients (7.0 nized (0.16) and infected (0.56) patients (p \ 0.01). CCIs vs. 3.8 %; p = 0.03). Moreover, this strategy completely were less than 0.35 in all colonized patients and at least prevented the development of ICU-acquired invasive 0.4 in all infected patients (p B 0.001). The sensitivity, candidiasis. The proportions of patients treated empiri- specificity, and positive and negative predictive values cally were 87 % of those with CCIs greater than 0.4 were all 100 %. In a multiple logistic regression analysis, (some of them died before the CCI could be calculated), only the APACHE II score [odds ratio (OR), 1.03; 95 % 18 % of those with CCIs less than 0.4 and CIs greater confidence interval, 1.01–1.05 per point; p = 0.007] and than 0.5, and 2 % of those with CCIs less than 0.4. A total CCI (OR, 4.01; 95 % confidence interval, 2.16–7.45; of 25 % of patients received empirical antifungal treat- p \ 0.001) independently predicted the development of ment. The authors estimated that 160 complete invasive candidiasis. mycological screenings and 10 preemptive antifungal treatments were needed to prevent one proven ICU- acquired invasive candidiasis. Despite the absence of randomization, this study is the most valuable ever per- Overview of the usefulness of the colonization index formed using the CI to guide empirical antifungal treatment. However, the intense laboratory work required Since the publication of the paper describing the CI in to perform CI in all ICU patients and the high proportion 1994 in Annals of Surgery, many centers have used the CI of patient treated may explain why this practice is not Table 2 Reported experience with the use of the colonization index (CI)

Reference Type of Number of Number of CI monitoring Number of Main CI ﬁndings Authors’ conclusion patient; patients invasive surveillance study candidiasis cultures design cases

General studies (n = 7) Tran et al. [51] Cardiac surgery; 81 7 Preoperatively, 773 (mean, 8.8/ CI : in 57 % developing IC, ‘‘This study showed that three prospective day 6 or at patient); 116 23 % uninfected; epidemiological factors— discharge (16.3 %) positive CI associated with female sex duration of the intravascular (45 vs. 21 %), catheter catheterization and medical duration (13 vs. 11 days), devices, length of stay in the length of ICU stay (5.7 vs. SICU and the sex of the 4.9 days) patient—were associated in CI did not change in 43 % elective cardiovascular developing IC, 72 % surgical patients with an uninfected increase of the Candida spp. colonization index, which helps to identify high-risk patients for Candida infections’’ Yazdanparast Cardiac surgery; 131: coronary artery 0 Preoperatively, 147 In patients requiring ‘‘Epidemiological data et al. [52] prospective bypass grafting postoperatively extracorporeal circulation, obtained from this study with (n = 72) and (day not : CI associated with more show that coronary artery without (n = 59) specified) antibiotic use (1.50 ± 0.83 bypass grafting with extracorporeal vs. 1.08 ± 0.40; different extracorporeal circulation circulation agents; p = 0.0055) procedure is associated with an increase in the use of antibiotics and subsequently a higher risk of Candida colonization-infection’’ Charles et al. Mixed ICU, 51: 20 MICU, 32 51 Daily for 7 days Not specified (mean, Mean CI in 46 assessed: ‘‘Candidemia occurrence is [37] patients with SICU (p \ 0.05) before 3.2/patient before 0.56 ± 0.39 associated with a high candidemia; candidemia candidemia) CI C 0.50 in 21 (45.6 %) mortality rate among retrospective MICU 0.74 ± 0.31 vs. SICU critically ill patients. 0.45 ± 0.40 (p = 0.01) Differences in underlying Mean CI in survival 0.46 vs. conditions could account for death 0.63 (p = 0.04) the poorer outcome of the Better ICU outcome associated medical patients. Screening with prior surgery [HR, 0.25 for fungal colonization could (0.09–0.67)], antifungal allow identification of such treatment [HR, 0.11 high-risk patients and, in (0.04–0.30)], absence of turn, improve outcome’’ neutropenia [HR, 0.10 (0.02–0.45)] 1435 1436 Table 2 continued

Reference Type of Number of Number of CI monitoring Number of Main CI ﬁndings Authors’ conclusion patient; patients invasive surveillance study candidiasis cultures design cases

Charles et al. Medical ICU, stay 92 1 (septic shock Weekly (mean, 3.2/ 1,696 (mean, CI C 0.50 in 36 (39.1 %), ‘‘Candida spp. multiple-site [38] [7 days; with typical patient) 18.4/patient) almost all with detectable colonization is frequently prospective maculopapular fungal colonization on ICU reached among the critically rash, positive admission ill medical patients. Broad skin biopsy) Predictors of CI C 0.5: spectrum antibiotic therapy Candida colonization at was found to promote fungal admission [OR 18.80 growth in patients with prior (5.21–67.79)], [2 days colonization. Since most of bladder catheterization [OR, the invasive candidiasis in 10.44 (1.61–67.85)] the ICU setting are thought CI : associated with days on to be subsequent to broad-spectrum antibiotics colonization in high-risk [b = 0.01 (0.01–0.02)], patients, reducing antibiotic hematological malignancy use could be useful in [b = 0.41 (0.09–0.73)], preventing fungal candiduria [b = 0.2 infections’’ (0.09–0.31)] Empirical antifungal treatment prescribed in 14/36 (39 %) with CI C 0.5, 7/56 (13 %) with CI \ 0.5 (p = 0.007) Eloy et al. [53] Mixed ICU, risk 75: 4 MICU (2 At admission, 901; 469 (52 %) Accuracy of CI [ 0.5 in ‘‘Serological tests failed to factors 46 MICU, 29 SICU respiratory weekly positive predicting IC in MICU: sens, differentiate infected from (antibiotics or tract, 2 75 %; spec, 35 %; PPV, non-infected patients. The hospital stay urinary); 5 10 %; NPV, 94 % Pittet’s CI identified infected [8 days, SICU (4 In SICU: sens, 100 %; spec, surgical patients (Fisher neutropenia); peritonitis, 1 50 %; PPV, 29 %; NPV, exact test, 0.052), which are prospective urinary) 100 % in the population with CI [ 0.5’’ Agvald-Ohman Mixed ICU, stay 59 10 At admission, then 401; 149 (37 %) 32 (54 %) received antifungals: ‘‘High colonization index and et al. [54] [7 days; weekly positive 10 with IC [mean CI, 0.7; 8 recent extensive gastro- prospective had digestive surgery, 7 had abdominal surgery were CIs [ 0.5 (\0.8 in 6)]; 22 significantly correlated with without IC (mean CI, 0.26; 7 IC. The results indicate that had digestive surgery, 9 had ICU patients exposed to CIs [ 0.5) extensive gastro-abdominal 17/25 (68 %) with CIs C 0.5 surgery would benefit from on day 7 received early antifungal antifungals prophylaxis’’ Predictors of IC: CI [ 0.5 (OR, 19.1 [2.4–435]), digestive surgery [OR, 60 (2.4– infinity)] Table 2 continued

Reference Type of Number of Number of CI monitoring Number of Main CI ﬁndings Authors’ conclusion patient; patients invasive surveillance study candidiasis cultures design cases

Massou et al. Medical ICU, stay 100 15 At admission, then 816 (including 143 CI C 0.5 in 53 (53 %), ‘‘CI has the advantage to [55] [2 days and at weekly blood cultures) predicted only by provide quantified data of least one risk corticosteroids [OR, 5.1 the patient’s situation in factor; (1.02–25.2)] relation to the colonization. prospective Accuracy of CI [ 0.5 in But, it isn’t helpful with predicting IC: sens, 93 %; patients having an invasive spec, 48 %; PPV, 26 %; candidiasis in medical NPV, 98 % intensive care unit’’ Only neutropenia predicted invasive candidiasis [OR, 18.3 (2.9–114)] Candiduria studies (n = 5) Chabasse [56] 15 mixed ICUs, 135 0 At time of Not available Candiduria: \103, 56 (42 %); ‘‘Quantification of candiduria one major or two candiduria and/ 103–104, 21 (16 %); [104, could be useful to select minor risk or candidemia 56 (42 %) patients at high risk for factors and CI [ 0.5 in 36/76 tested (65 % disseminated candidiasis’’ candiduria; with candiduria [104,31% prospective with candiduria \104; p = 0.003) Dubau et al. Surgical ICU, stay 89 1/35 empirically On inclusion, then 2,238 Absence of candiduria: ‘‘The presence of a candiduria [57] or antibiotics treated with weekly CI = 0.3–0.47 (p = 0.008) was significantly associated [7 days or CIs C 0.5; 0 Presence of candiduria: with an increased invasive postoperative with CI = 0.57–0.87 candidiasis’’ fistula and CIs \ 0.5; 22 (p = 0.0001) CRP [ 100 mg/ candiduria ml; prospective Sellami et al. Mixed ICU, stay 162 6; 56 candiduria On inclusion, then Not available Candiduria: \103, 12 (21 %); ‘‘Candiduria superior or equal [58] [3 days; weekly 103–104, 16 (29 %); [104, to 104 UFC/ml associated prospective 28 (50 %) with risk factors may predict Mean CI = 0.47 candiduria, invasive candidiasis in 0.8 IC (n = 6) critically ill patients’’ CI [ 0.5 in 67 % candiduria [104 Charles et al. Medical ICU, 92 9 Weekly 1,696 (mean, 18.4/ : CI associated with: days on ‘‘Since most of the invasive [38] stay [ 7 days; patient) broad-spectrum antibiotics candidiasis in the ICU prospective [b = 0.01 (0.01–0.02)], setting are thought to be hematological malignancy subsequent to colonization [b = 0.41 (0.09–0.73)], in high-risk patients, candiduria [b = 0.2 reducing antibiotic use could (0.09–0.31)] be useful in preventing fungal infections’’ 1437 1438 Table 2 continued

Reference Type of Number of Number of CI monitoring Number of Main CI ﬁndings Authors’ conclusion patient; patients invasive surveillance study candidiasis cultures design cases

Ergin et al. [59] Mixed ICU, stay 100 9 (5 candidemia, On inclusion, then 1,691 (mean, CI [ 0.2, 42 (42 %); ‘‘Candida colonization and [7 days; 4 urinary); 42 weekly 17/patient) CI [ 0.2 and \0.5, 34 (34 %); Candida colonization index prospective candiduria CI C 0.5, 8 (8 %) may be used as useful Accuracy of CI C 0.5 in parameters to predict predicting invasive invasive Candida candidiasis: sens, 100 %; infections’’ spec, 64 %; PPV, 21 %; NPV, 100 % Prophylaxis studies (n = 6) Laverdie`re Neutropenia 266 41: 9/135 (6.7 %) On randomization, 1,904; 458 (21 %) Colonization ; (39–36 %) ‘‘Fluconazole prevented and et al. [60] (leukemia or fluconazole, at end of positive fluconazole, : (37–73 %) reduced fungal colonization BMT); 32/131 prophylaxis placebo (p \ 0.0001) of the alimentary tract and prospective, (24.4 %) CI ; (0.18–0.16) fluconazole, subsequent invasive fungal double-blinded, placebo : (0.18–0.39) placebo infections… In cancer randomized: (p \ 0.001) (p \ 0.001) patients, a colonization prophylaxis Accuracy of CI B 0.25 in index B0.25 at the initiation (fluconazole, predicting IC at baseline: of chemotherapy clearly 400 mg/day sens, 39 %; spec, 82 %; predicts a low risk of orally) vs. PPV, 28 %; NPV, 88 % invasive fungal infection’’ placebo At end of prophylaxis: sens, 76 %; spec, 69 %; PPV, 69 %; NPV, 94 % Garbino et al. Mixed ICU, 204 6/103 (5.8 %) Daily Not available Colonization: 29/55 (53 %) ‘‘…fluconazole prophylaxis in [40] [2 days fluconazole, fluconazole, 40/51 (78 %) selected, high-risk critically mechanical 16/101 (16 %) placebo (p = 0.01) ill patients decreases the ventilation and placebo CI ; (0.26–0.13) fluconazole, incidence of Candida expected (p \ 0.001) : (0.26–0.50) placebo infection, in particular, continuation for (p \ 0.001) candidemia’’ C72 h; Mean pre-infection CI in prospective, patients with candidemia double-blinded (n = 10): 0.89 randomized: prophylaxis (fluconazole, 100 mg/day iv) vs. placebo plus selective digestive decontamination (polymyxin B, neomycin, vancomycin) Table 2 continued

Reference Type of Number of Number of CI monitoring Number of Main CI ﬁndings Authors’ conclusion patient; patients invasive surveillance study candidiasis cultures design cases

Normand et al. Mixed 98 0 On randomization, Not available Colonization: 0/51 nystatin, ‘‘Oral nystatin prophylaxis [61] ICU, [2 days then every 12/47 (25 %) placebo efficiently prevented mechanical 3 days (p \ 0.01) Candida spp. colonization in ventilation; CI ; (0.1–0.05) nystatin, : ICU patients at low risk of prospective, (0.1–0.25) placebo (p \ 0.05) developing invasive open-label, candidiasis’’ nystatin vs. placebo Senn et al. [46] Surgical ICU, 19 1 (6–8 expected On inclusion, then Not available CI ; (0.5–0.3) during ‘‘Despite limitations such as the recurrent in this high- twice in week 1, (median, 4 sites/ prophylaxis (p = 0.03) open single-center non- gastrointestinal risk group then weekly patient screened) comparative design and the perforation/ without until end of small sample size, the anastomotic prophylaxis) follow-up observations of this proof- leakage or acute of-concept study suggest necrotizing that caspofungin may be pancreatitis; efficacious and safe for prospective, prevention of intra- non- abdominal candidiasis in comparative, surgical patients with a high- caspofungin risk profile’’ prophylaxis Giglio et al. Surgical/trauma 99 0 On inclusion, then 2,569; 746 (29 %) Overall: CI ; (0.12–0.0) ‘‘The present trial shows that [62] ICU, [2 days every 3 days positive nystatin, : (0.2–0.44) nystatin pre-emptive therapy mechanical until end of placebo (p \ 0.05) in surgical/trauma ICU ventilation; follow-up (day Colonization at entry: CI ; patients significantly reduces prospective, 15) (0.1–0.0) nystatin, : fungal colonization, even in randomized, (0.2–0.42) placebo those colonized at open-label: (p \ 0.05) admission’’ prophylaxis (oral nystatin, 3 9 1 million U/day) vs. placebo Chen et al. [63] Mixed ICU, 124 8: 3/60 (0.5 %) On inclusion, every Not available; 874 CCI at day 6: 0.19 nystatin, ‘‘Nystatin might reduce the mechanical nystatin, 5/64 3 days until end positive 0.39 placebo (p \ 0.05) colonization by Candida ventilation; (7.8 %) of follow-up At day 9: 0.0 nystatin, 0.45 albicans and was associated prospective, placebo (day 9) placebo (p \ 0.05) with shorter ICU stay’’ randomized, (p [ 0.05) Length of ICU stay: open-label: 9.6 ± 3.5 days nystatin, prophylaxis (oral 11.9 ± 6.3 days placebo nystatin, (p \ 0.05) 3 9 1 million U/day) vs. placebo

CRP C-reactive protein, BMT bone marrow transplant, MICU medical ICU, SICU surgical ICU, sens sensitivity, spec speciﬁcity, PPV positive predictive value, NPV negative predictive value 1439 1440 Table 3 Use of the colonization index (CI) to guide the initiation of empirical antifungal treatment

Reference Type of patient; Number of patients Number of CI monitoring Number of Main CI ﬁndings Authors’ conclusion study design invasive surveillance candidiasis cultures cases

Interventional studies (n = 4) Dubau et al. [57] Surgical ICU, stay or 89 1/35 with On inclusion, 2,238 CI [ 0.5: week 1, 6 %; week ‘‘A treatment was started antibiotics [7 days or - CI [ 0.5 then weekly 2, 25 %; week 3, 40 %; whenever a colonization postoperative fistula and receiving week 4, 55 %; week [4, index [0.5 was associated CRP [100 mg/ml; empirical 59 % with severe clinical or prospective, open, non- antifungal CI ; rapidly after start of biological signs. This randomized: empirical treatment empirical antifungal involved an increase of the antifungal treatment for CI treatment in 34/35 patients expense of antifungal [0.5 drugs. The potential benefits could not be evaluated from our study’’ Piarroux et al. [64] Surgical ICU, stay C5 days; 478 Overall IC: 18 On inclusion, 6,682 With empirical treatment: ‘‘Preemptive treatment of prospective, open, non- [3.8 %; 2/455 then weekly CCI [ 0.4, 96/117; highly colonized patients randomized: empirical (7.0 %) in CCI \ 0.4 and CI C 0.5, may efficiently prevent antifungal treatment for historical 11/66; CCI \ 0.4 and SICU-acquired proven CCI [ 0.4 controls] CI \ 0.5, 5/230 candidiasis. Our results ICU-acquired: 0 *160 complete mycological demonstrate the feasibility [10/455 (2.2 %) screenings and 10 and benefits of in historical preemptive treatments implementing a large controls; needed to prevent at least systematic mycological p \ 0.001) one proven SICU-acquired screening of SICU candidiasis patients’’ Eren et al. [65] Mixed ICU, inclusion criteria 37 0 Not specified 191 26 (70 %) with C. albicans ‘‘…follow-up of the ICU not specified; prospective colonization patients in terms of C. observational, intervention CI C 0.5 in 7 (5 with IgM, albicans CI and IgM not specified (antifungal IgG positivity) would be effective for the treatment for CI [ 0.5?) CI \ 0.5 in 19 (3/12 tested prevention of serious with IgM, IgG positivity) Candida infections’’ IgM, IgG found in 0/7 patients tested without colonization, out of 11 Wang et al. [66] 5 mixed ICUs, APACHE 110 Not specified Not specified Not available; Antifungal treatment for ‘‘Application of CCI may score [ 10; prospective 575 positive sepsis started at enhance the accuracy of randomized: empirical 0.9 ± 0.7 days in timely preemptive antifungal treatment for CCI C 0.4, 3.8 ± 3.6 days treatment for invasive CCI C 0.4 (intervention in control (p \ 0.05) candidiasis and facilitate group) or (control group) the collection of epidemiological data of Candida in critically ill patients’’

Sens sensitivity, spec specificity, PPV positive predictive value, NPV negative predictive value 1441 currently used widely among non-immunocompromised antibiotics, the presence of a central line, together with at critically ill patients. least one additional risk factor among the following: par- In a prospective study of 37 critically ill patients, Eren enteral nutrition, dialysis, surgery, pancreatitis, systemic et al. [65] found a correlation between the CI and the steroids, or other immune suppressive agents. The inci- presence of C. albicans antibodies. Among 26 patients dence of invasive candidiasis was 16.7 % (14/84) vs. with C. albicans colonization, immunoglobulin M (IgM) 9.8 % (10/102) in patients receiving placebo and caspo- and IgG were found in 5/7 (71 %) patients with CIs fungin, respectively (p = 0.14). When baseline infections greater than 0.5, compared with only 3/12 (25 %) patients were included, it was 30.4 % (31/102) and 18.8 % (22/ with CIs less than 0.5 in whom serum could be tested 117), for patients receiving preemptive placebo and ca- (total, n = 19) and 0/7 non-colonized patients in whom spofungin, respectively (p = 0.04). Safety, length of ICU serum could be tested (total, n = 11) [65]. The authors stay, antifungal use, and mortality did not differ. The stated, ‘‘in the follow-up of the patients, no candidemia authors concluded that caspofungin prophylaxis was safe developed and this was thought to be due to the pre- with a non-significant trend to reduce invasive candidiasis, ventive measures which were taken especially in ICU and that the preemptive therapy approach deserves further patients with CI [ 0.5,’’ suggesting that empirical anti- study. Two recent and currently unpublished studies fungal treatment was given to patients with CIs greater attempt to use clinical predictive rules to guide empirical than 0.5. antifungal treatment. The first, entitled ‘‘Pilot Feasibility Study With Patients Who at High Risk For Developing Invasive Candidiasis in a Critical Care Setting (MK-0991- 067 AM1’’ (ClinicalTrials.gov identifier: NCT01045798) Comparison of the colonization index with other and conducted by the Mycoses Study Group, was termi- predictive tools nated because of a low recruitment rate after the inclusion of only 15 patients and generation of uninterpretable data. The accuracy of the CI has been compared with that of Preliminary results of the second study, entitled ‘‘A Study other surrogate markers of invasive candidiasis, such as to Evaluate Pre-emptive Treatment for Invasive Candidi- the Candida score, predictive rules, and, more recently, asis in High Risk Surgical Subjects (INTENSE)’’ beta-glucan (Table 4)[44, 67–72]. In a prospective study (ClinicalTrials.gov identifier: NCT01122368), showed a of 89 high-risk surgical ICU patients (61 recurrent gas- high proportion of invasive candidiasis at study entry; trointestinal tract perforation; 25 severe acute independent experts considered, however, that the major- pancreatitis) in whom 29 developed a non-candidemic ity of salvage antifungal treatments were unjustified. invasive candidiasis, Tissot el al. showed that two con- Globally, the proportion of invasive candidiasis develop- secutive beta-glucan serum levels above 80 pg/ml ing in the group receiving preemptive antifungal treatment predicted early the development of infection with a sen- was not different from that of the group receiving placebo, sitivity, specificity, positive predictive value, and a but the number of patients excluded from the analysis negative predictive value of 65, 78, 68 and 77 %, resulted in an underpowered analysis. respectively [44]. Similarly, a previous study by Posteraro These outcomes illustrate a well-known paradigm of et al. [70] in a cohort of medical and surgical critically ill accuracy. The use of the most sensitive test potentially patients staying more than 5 days and developing a severe increases the number of patients who are uselessly treated sepsis, in whom beta-glucan was higher than 80 pg/ml, empirically. In contrast, the use of the most specific test showed a sensitivity, specificity, positive predictive value, potentially increases the number of patients who do not and a negative predictive value of 93, 94, 72, and 99 %, receive empirical antifungal treatment and in whom late respectively for the early detection of invasive candidiasis treatment could be associated with a worse outcome. (16 episodes, including 13 candidemia). Overall, the An ongoing prospective multicenter, double blind, accuracy of CI is lower than other methods for the early randomized-controlled French trial (EMPIRICUS) is detection of patients at higher risk of infection. Never- currently recruiting critically ill patients at risk of inva- theless, the usefulness of beta-glucan to guide empirical sive candidiasis. Patients mechanically ventilated for antifungal treatment remains to be determined. more than 4 days with sepsis of unknown origin and with Three industry-sponsored studies failed to demonstrate at least one extra-digestive fungal colonization site and the clinical usefulness of predictive rules combining var- multiple organ failure are eligible for randomization to ious risk factors. A predictive rule was used in a receive empirical micafungin or a placebo. This study multicenter, randomized, double-blind, placebo-con- includes prospective determination of the CI among its trolled trial comparing caspofungin vs. placebo as secondary endpoints and may help to develop guidelines antifungal prophylaxis or preemptive treatment in 222 for treating non-immunocompromised patients with fun- critically ill patients [73]. The rule combined ICU stay for gal colonization multiple organ failure and sepsis of at least 3 days, mechanical ventilation, exposure to unknown origin [74]. Table 4 Comparison of the colonization index (CI) with other markers for the prediction of invasive candidiasis 1442

References Type of patient; Number of Number of CI monitoring Performance of CI Performance of Authors’ conclusion study design patients invasive comparators candidiasis

Leon et al. 36 mixed ICUs, stay 1,107 58; 18/240 in At inclusion, 892 with Candida spp. 892 with Candida spp. ‘‘In this cohort of [67] C7 days; whom BG then weekly colonization or colonization or invasive colonized patients prospective, could be (4,198 invasive candidiasis candidiasis staying [7 days, with a observational cohort performed cultures) Development of Development of invasive CS \3 and not invasive candidiasis: Candida receiving antifungal candidiasis: score \3, 13/565 treatment, the rate of IC CI \ 0.5, 16/411 (2.3 %); Candida score was \5 %. Therefore, (3.9 %); C3, 45/327 (13.8 %) IC is highly improbable CI C 0.5, 42/481 Sens, 78 %; spec, 66 %; if a Candida-colonized (8.7 %) PPV, 14 %; NPV, non-neutropenic Sens, 72 %; spec, 98 %; AUC, 0.774 critically ill patient has 47 %; PPV, 9 %; 8.7 needed to treat to a Candida score \0.3’’ NPV, 96 %; AUC, avoid one episode of 0.633 invasive candidiasis 20.8 needed to treat to avoid one episode of invasive candidiasis Ellis et al. Hematological patients 86 12 (14 %) Twice a week CI C 0.5 in 52/86 Mannans/anti-mannans ‘‘Serial assays for [68] with febrile (60 %), 8/9 (89 %) (33 patients (2 mannans and anti- neutropenia and IC consecutive positive mannans in patients at [3 days broad- Accuracy not tests): sens, 73 %; spec, risk for invasive spectrum antibiotics; determined 80 %; PPV, 0.36; NPV, candidiasis may prospective, 0.95 usefully contribute to usefulness of serial Positive correlation of the management of analysis for invasive candidiasis such patients’’ mannans/anti- with b-glucans mannans, b-glucans (r = 0.28, p = 0.01) Caggiano Neurological ICU, stay 51 3 candidemia At admission, Colonization: 76 % at No mannan/anti-mannan ‘‘Thus, our experience et al. [69] C7 days; then every study entry, 92 % on detected in patients suggests that prospective, link 3 days (to day day 15 without candidemia monitoring CI could be between 15) CI C 0.5: 16.6 % at Mannans: sens, 66.6 %; helpful in identifying colonization and study entry, 75 % spec, 100 % patients at risk of invasive candidiasis on day 15 (all IC) Anti-mannans: sens, invasive fungal CCI: 0.34 at study 100 %; spec, 100 % infection. In addition, entry, 0.60 on complementing this day 15 with anti-Candida antibodies detection in immunocompetent patients with CI C 0.5 increases the positive predictivity for infection allowing an early diagnosis of candidemia’’ Table 4 continued References Type of patient; Number of Number of CI monitoring Performance of CI Performance of Authors’ conclusion study design patients invasive comparators candidiasis

Posteraro Mixed ICU, stay 95 16 (14 invasive At entry, day 3, CI C 0.5: overall, Candida score C3: ‘‘A single-point BG assay et al. [70] C5 days; candidiasis) and then 35 %; IC, 56 %; no overall, 22 %; IC, based on a blood prospective, single- weekly IC, 30 % (p = 0.04) 75 %; no IC, 11 % sample drawn at the center, Sens, 64 %; spec, (p \ 0.001). Sens, sepsis onset, alone or in observational: 70 %; PPV, 27 %; 86 %; spec, 87 %; PPV, combination with diagnostic value of NPV, 92 %; AUC, 57 %; NPV, 97 %; Candida score, may BG assay, Candida 0.63 (0.57–0.79) AUC, 0.80 (0.69–0.92) guide the decision to score, CI BG positivity: overall, start antifungal therapy 21 %; IC, 94 %; no early in patients at risk IC, 6 % (p \ 0.001). for Candida infection’’ Sens, 94 %; spec, 94 %; PPV, 75 %; NPV, 99 %; AUC, 0.98 (0.92–1.00) Candida score C3 and BG positivity: sens, 100 %; spec, 84 %; PPV, 52 %; NPV, 100 % Peman et al. 6 mixed ICU, high risk 53 0 Not specified CCI C 0.4: 23/53 CAGTA positivity: 22/53 ‘‘This study identified [71] of invasive (43 %) at study (42 %) at study entry, previous surgery as the candidiasisa; entry, 41/53 (77 %) 47/53 (90 %) at end of principal clinical factor prospective, at end of study study associated with usefulness of CAGTA-positive CAGTA results (serologically proven candidiasis) and emphasises the utility of this promising technique, which was not influenced by high Candida colonization or antifungal treatment’’ Hall et al. Mixed ICU, severe 101 18 (5 died) Not specified Colonization: 16/18 Candida score: sens, ‘‘In this study the Candida [72] acute pancreatitis; (89 %) IC, 37/83 23 %; spec, 85 %; PPV, colonisation index score retrospective, (45 %) no IC 39 %; NPV, 72 %; was the most accurate diagnostic value of (p = 0.0006). AUC, 0.62 (0.52–0.71) and discriminative test BG assay, Candida Colonization for IC: Predictive ruleb: sens, at identifying which score, CI OR, 4.33 61 %; spec, 49 %; PPV, patients with severe (1.07–17.5); 21 %; NPV, 85 %; acute pancreatitis are at p = 0.04 AUC, 0.59 (0.49–0.69) risk of developing CI C 0.5: sens, 67 %; candidal infection. spec, 79 %; PPV, However its low 43 %; NPV, 91 %; sensitivity may limit its AUC, 0.79 clinical usefulness’’ (0.69–0.87) 1443 1444

The CI and other predictive tools (the Candida score and predictive rules) have been speciﬁcally developed by 7 days Candida [ using their positive predictive value for the early identi- ﬁcation of high-risk patients who will develop invasive Candida candidiasis. Nevertheless, it is relevant to note that the negative predictive values of the CI and all other surrogate markers of invasive candidiasis are much higher than

superior to score and colonization indexes and anticipates diagnosis of blood culture-negative intra- abdominal candidiasis’’ their positive predictive values [21]. Among them, only Authors’ conclusion ‘‘BG antigenemia is the negative predictive value of the Candida score has igh CAGTA been validated in a multicenter prospective clinical trial [75]. ): 9 3: sens, C

score Advantages and pitfalls of the use of colonization

80 pg/ml (2 index [ sens, 66 %; spec,PPV, 83 73 %; %; NPV, 78 %; AUC, 0.79 86 %; spec, 5054 %; %; PPV, NPV, 84AUC, %; 0.61 BG comparators Candida To summarize the data reviewed on the use of the CI, its main advantages are that it has been successfully used to characterize colonization dynamics, assess the signiﬁ- cance of candiduria, and to evaluate the impact of antifungal prophylaxis. A low CI has a high negative predictive value for the further development of invasive area under the receiver operating characteristic curve, CAGTA

0.4: sens, candidiasis. Among the pitfalls, it should be emphasized 0.5: sens, 26 %; C

C AUC that it is work-intensive with a limited bedside practica- spec, 76 %; PPV, 35 %; NPV, 67AUC, %; 0.67 14 %; spec, 77PPV, %; 23 %; NPV, 65 %; AUC, 0.43 bility, that only limited data are available for nonsurgical CI CCI

3 risk factors (diabetes mellitus, extrarenal depuration, parenteral nutrition, patients, and its cost-effectiveness and usefulness for the C management of critically ill patients remain to be proven in large prospective clinical trials. then twice a week CI monitoring Performance of CI Performance of Further perspectives negative predictive value,

14 days) ICU stay and These observations may pave the way for a new paradigm NPV [ in the research agenda of the early diagnosis of invasive candidiasis and take advantage of the disease pathophysiology characterized by a 7–10-day delay between Number of invasive candidiasis exposure to risk factors and infection. We propose a stepwise approach to optimize the accuracy of the CI and other clinical tools (Fig. 3). Provided that preliminary results can be confirmed in Number of patients 89 2 On admission, larger cohorts of patients [76], preliminary stratification positive predictive value, according to specific genetic polymorphisms could be used as a preliminary step to increase the accuracy of PPV

7 days evolution, (2) prolonged ( subsequent steps. The second step is to take advantage of

[ the high negative predictive values of the CI, Candida score, and/or predictive rules. This strategy will reduce

speciﬁcity, the currently large proportion of useless antifungal treat-

recurrent gastrointestinal perforation/ anastomotic leakage or acute necrotizing pancreatitis; prospective, accuracy of BG antigenemia for diagnosis of intra- abdominal candidiasis ments, frequently started for nonobjective emotional Spec study design Surgical ICU, reasons. The enhanced intrinsic risk of invasive candidiasis in patients with increased risk identiﬁed after this continued germ tube antibody second step may further improve the accuracy of bio-

] markers, such as mannan/anti-mannan antibodies and/or ] sensitivity, 44 73 b-glucan, applied in the third step. This approach should [ [ (1) Acute pancreatitis with References Type of patient; Table 4 a b broad-spectrum antibiotic therapy, major abdominal surgery), (3) liver transplantation, (4) neutropenia or bone marrow transplantation, or (5) h Tissot et al. Sens albicans enhance the ability to identify patients who will truly 1445

Fig. 3 Proposed study designs to select high-risk patients who patients at low objective risk of invasive candidiasis are ruled out would truly benefit from early empirical antifungal treatment. First, by using the high negative predictive values of current risk- patients should be stratified according to pathophysiological assessment strategies (CI, Candida score, peritonitis score, predic- characteristics specific to invasive candidiasis (i.e., immunocom- tive rule). In the third step, empirical antifungal treatment is started promised individuals, after digestive surgery, other critically ill early in high-risk patients identified by increased biomarkers patients). The three-step approach: In a first optional step, increased values, such as beta-glucan performed only in patients retained by intrinsic risk of invasive candidiasis is identified by specific genetic who met criteria outlined in previous steps. A simplified approach polymorphisms related to innate immunity. In the second step, restricted to steps two and three may be also of potential interest benefit from early antifungal treatment and reduce antifungal treatment. Continued development and execu- unnecessary overexposure to antifungal agents in patients tion of clinical trials in this difficult field are important. with a documented limited risk of infection. Until further progress can be achieved with new clinical Moreover, the high negative predictive value of the studies of specific biomarkers involving larger patient predictive tools (CI, Candida score, and predictive rules) cohorts, 20 years after its description, the CI remains one for the further development of an invasive candidiasis of the best methods of characterizing the dynamics of may allow one to stop empirical antifungal treatment Candida colonization and identifying patients at very low possibly prescribed before evaluating the patient accord- or increasing risk of invasive candidiasis. ing to the proposed stepwise approach. Acknowledgments PE received research grants and/or educational grants and/or speaker’s honoraria and/or consultant’s honoraria from (in alphabetic order): Astellas, Merck, Sharp & Dohme-Chi- bret, and Pfizer. Conclusion Conflicts of interest DP declares no disclosure of potential conflicts of interest regarding their contribution to this study. The development of the CI has enabled better appreciation of the dynamics of Candida colonization in patients at Open Access This article is distributed under the terms of the high risk of invasive candidiasis. It can still be used, as it is Creative Commons Attribution Noncommercial License which permits in many institutions, for the early detection of patients at any noncommercial use, distribution, and reproduction in any medium, high risk of invasive candidiasis and to guide empirical provided the original author(s) and the source are credited. 1446

References

1. Eggimann P, Garbino J, Pittet D (2003) 9. Gudlaugsson O, Gillespie S, Lee K, 17. Pappas PG, Kauffman CA, Andes D, Epidemiology of Candida species Berg JV, Hu J, Messer S, Herwaldt L, Benjamin DK Jr, Calandra TF, Edwards infections in critically ill non- Pfaller M, Diekema D (2003) JE Jr, Filler SG, Fisher JF, Kullberg BJ, immunosuppressed patients. Lancet Attributable mortality of nosocomial Ostrosky-Zeichner L, Reboli AC, Rex Infect Dis 3:685–702 candidemia, revisited. Clin Infect Dis JH, Walsh TJ, Sobel JD, Infectious 2. Vincent JL, Rello J, Marshall J, Silva E, 37:1172–1177 Diseases Society of A (2009) Clinical Anzueto A, Martin CD, Moreno R, 10. Zaoutis TE, Argon J, Chu J, Berlin JA, practice guidelines for the management Lipman J, Gomersall C, Sakr Y, Walsh TJ, Feudtner C (2005) The of candidiasis: 2009 update by the Reinhart K, EPIC II Group of epidemiology and attributable outcomes Infectious Diseases Society of America. Investigators (2009) International study of candidemia in adults and children Clin Infect Dis 48:503–535 of the prevalence and outcomes of hospitalized in the United States: a 18. Cornely OA, Bassetti M, Calandra T, infection in intensive care units. JAMA propensity analysis. Clin Infect Dis Garbino J, Kullberg BJ, Lortholary O, 302:2323–2329 41:1232–1239 Meersseman W, Akova M, Arendrup 3. Pfaller M, Neofytos D, Diekema D, 11. Prowle JR, Echeverri JE, Ligabo EV, MC, Arikan-Akdagli S, Bille J, Azie N, Meier-Kriesche HU, Quan SP, Sherry N, Taori GC, Crozier TM, Hart Castagnola E, Cuenca-Estrella M, Horn D (2012) Epidemiology and GK, Korman TM, Mayall BC, Johnson Donnelly JP, Groll AH, Herbrecht R, outcomes of candidemia in 3648 PD, Bellomo R (2011) Acquired Hope WW, Jensen HE, Lass-Florl C, patients: data from the Prospective bloodstream infection in the intensive Petrikkos G, Richardson MD, Roilides Antifungal Therapy (PATH Alliance) care unit: incidence and attributable E, Verweij PE, Viscoli C, Ullmann AJ, registry, 2004-2008. Diagn Microbiol mortality. Crit Care 15:R100 Group EFIS (2012) ESCMID guideline Infect Dis 74:323–331 12. Morrell M, Fraser VJ, Kollef MH for the diagnosis and management of 4. Guery BP, Arendrup MC, Auzinger G, (2005) Delaying the empiric treatment Candida diseases 2012: non- Azoulay E, Borges Sa M, Johnson EM, of candida bloodstream infection until neutropenic adult patients. Clin Muller E, Putensen C, Rotstein C, positive blood culture results are Microbiol Infect 18(Suppl 7):19–37 Sganga G, Venditti M, Zaragoza Crespo obtained: a potential risk factor for 19. Eggimann P, Garbino J, Pittet D (2003) R, Kullberg BJ (2009) Management of hospital mortality. Antimicrob Agents Management of Candida species invasive candidiasis and candidemia in Chemother 49:3640–3645 infections in critically ill patients. adult non-neutropenic intensive care 13. Garey KW, Rege M, Pai MP, Mingo Lancet Infect Dis 3:772–785 unit patients: part I. Epidemiology and DE, Suda KJ, Turpin RS, Bearden DT 20. Playford EG, Eggimann P, Calandra T diagnosis. Intensive Care Med (2006) Time to initiation of fluconazole (2008) Antifungals in the ICU. Curr 35:55–62 therapy impacts mortality in patients Opin Infect Dis 21:610–619 5. Kett DH, Azoulay E, Echeverria PM, with candidemia: a multi-institutional 21. Eggimann P, Ostrosky-Zeichner L Vincent JL, Extended Prevalence of study. Clin Infect Dis 43:25–31 (2010) Early antifungal intervention Infection in ICUSGoI (2011) Candida 14. Azoulay E, Dupont H, Tabah A, strategies in ICU patients. Curr Opin bloodstream infections in intensive care Lortholary O, Stahl JP, Francais A, Crit Care 16:465–469 units: analysis of the extended Martin C, Guidet B, Timsit JF (2012) 22. Montravers P, Dupont H, Eggimann P prevalence of infection in intensive care Systemic antifungal therapy in critically (2013) Intra-abdominal candidiasis: the unit study. Crit Care Med 39:665–670 ill patients without invasive fungal guidelines-forgotten non-candidemic 6. Marchetti O, Bille J, Fluckiger U, infection. Crit Care Med 40:813–822 invasive candidiasis. Intensive Care Eggimann P, Ruef C, Garbino J, 15. Sipsas NV, Lewis RE, Tarrand J, Med 39:2226–2230 Calandra T, Glauser MP, Tauber MG, Hachem R, Rolston KV, Raad II, 23. Pittet D, Monod M, Filthuth I, Frenk E, Pittet D, Fungal Infection Network of S Kontoyiannis DP (2009) Candidemia in Suter PM, Auckenthaler R (1991) (2004) Epidemiology of candidemia in patients with hematologic malignancies Contour-clamped homogeneous electric Swiss tertiary care hospitals: secular in the era of new antifungal agents field gel electrophoresis as a powerful trends, 1991–2000. Clin Infect Dis (2001–2007): stable incidence but epidemiologic tool in yeast infections. 38:311–320 changing epidemiology of a still Am J Med 91:256S–263S 7. Eggimann P, Bille J, Marchetti O frequently lethal infection. Cancer 24. Pittet D, Monod M, Suter PM, Frenk E, (2011) Diagnosis of invasive 115:4745–4752 Auckenthaler R (1994) Candida candidiasis in the ICU. Ann Intensive 16. Lortholary O, Desnos-Ollivier M, colonization and subsequent infections Care 1:37 Sitbon K, Fontanet A, Bretagne S, in critically ill surgical patients. Ann 8. Bassetti M, Marchetti M, Chakrabarti Dromer F, French Mycosis Study G Surg 220:751–758 A, Colizza S, Garnacho-Montero J, Kett (2011) Recent exposure to caspofungin 25. van de Veerdonk FL, Kullberg BJ, DH, Munoz P, Cristini F, Andoniadou or fluconazole influences the Netea MG (2010) Pathogenesis of A, Viale P, Rocca GD, Roilides E, epidemiology of candidemia: a invasive candidiasis. Curr Opin Crit Sganga G, Walsh TJ, Tascini C, prospective multicenter study involving Care 16:453–459 Tumbarello M, Menichetti F, Righi E, 2,441 patients. Antimicrob Agents 26. Magill SS, Swoboda SM, Shields CE, Eckmann C, Viscoli C, Shorr AF, Leroy Chemother 55:532–538 Colantuoni EA, Fothergill AW, Merz O, Petrikos G, De Rosa FG (2013) A WG, Lipsett PA, Hendrix CW (2009) research agenda on the management of The epidemiology of Candida intra-abdominal candidiasis: results colonization and invasive candidiasis in from a consensus of multinational a surgical intensive care unit where experts. Intensive Care Med fluconazole prophylaxis is utilized: 39:2092–2106 follow-up to a randomized clinical trial. Ann Surg 249:657–665 1447

27. Khan ZU, Chandy R, Metwali KE 37. Charles PE, Doise JM, Quenot JP, Aube 46. Senn L, Eggimann P, Ksontini R, (2003) Candida albicans strain carriage H, Dalle F, Chavanet P, Milesi N, Aho Pascual A, Demartines N, Bille J, in patients and nursing staff of an LS, Portier H, Blettery B (2003) Calandra T, Marchetti O (2009) intensive care unit: a study of Candidemia in critically ill patients: Caspofungin for prevention of intra- morphotypes and resistotypes. Mycoses difference of outcome between medical abdominal candidiasis in high-risk 46:479–486 and surgical patients. Intensive Care surgical patients. Intensive Care Med 28. Marco F, Lockhart SR, Pfaller MA, Med 29:2162–2169 35:903–908 Pujol C, Rangel-Frausto MS, Wiblin T, 38. Charles PE, Dalle F, Aube H, Doise JM, 47. Goldstein E, Hoeprich PD (1972) Blumberg HM, Edwards JE, Jarvis W, Quenot JP, Aho LS, Chavanet P, Problems in the diagnosis and treatment Saiman L, Patterson JE, Rinaldi MG, Blettery B (2005) Candida spp. of systemic candidiasis. J Infect Dis Wenzel RP, Soll DR (1999) Elucidating colonization significance in critically ill 125:190–193 the origins of nosocomial infections medical patients: a prospective study. 48. Klein JJ, Watanakunakorn C (1979) with Candida albicans by DNA Intensive Care Med 31:393–400 Hospital-acquired fungemia. Its natural fingerprinting with the complex probe 39. Eggimann P, Calandra T, Fluckiger U, course and clinical significance. Am J Ca3. J Clin Microbiol 37:2817–2828 Bille J, Garbino J, Glauser MP, Med 67:51–58 29. Nucci M, Anaissie E (2001) Revisiting Marchetti O, Ruef C, Tauber M, Pittet 49. Solomkin JS, Flohr AM, Simmons RL the source of candidemia: skin or gut? D, Fungal Infection Network of (1982) Indications for therapy for Clin Infect Dis 33:1959–1967 Switzerland (2005) Invasive fungemia in postoperative patients. 30. Filler SG (2012) Insights from human candidiasis: comparison of management Arch Surg 117:1272–1275 studies into the host defense against choices by infectious disease and 50. Vazquez JA, Sanchez V, Dmuchowski candidiasis. Cytokine 58:129–132 critical care specialists. Intensive Care C, Dembry LM, Sobel JD, Zervos MJ 31. Charlier C, Hart E, Lefort A, Ribaud P, Med 31:1514–1521 (1993) Nosocomial acquisition of Dromer F, Denning DW, Lortholary O 40. Garbino J, Lew DP, Romand JA, Candida albicans: an epidemiologic (2006) Fluconazole for the management Hugonnet S, Auckenthaler R, Pittet D study. J Infect Dis 168:195–201 of invasive candidiasis: where do we (2002) Prevention of severe Candida 51. Tran LT, Auger P, Marchand R, Carrier stand after 15 years? J Antimicrob infections in nonneutropenic, high-risk, M, Pelletier C (1997) Epidemiological Chemother 57:384–410 critically ill patients: a randomized, study of Candida spp. colonization in 32. Maertens J, Marchetti O, Herbrecht R, double-blind, placebo-controlled trial in cardiovascular surgical patients. Cornely OA, Fluckiger U, Frere P, patients treated by selective digestive Mycoses 40:169–173 Gachot B, Heinz WJ, Lass-Florl C, decontamination. Intensive Care Med 52. Yazdanparast K, Auger P, Marchand R, Ribaud P, Thiebaut A, Cordonnier C, 28:1708–1717 Carrier M, Cartier R (2001) Predictive Third European Conference on 41. Vincent JL, Anaissie E, Bruining H, value of Candida colonization index in Infections in Leukemia (2011) Demajo W, el-Ebiary M, Haber J, 131 patients undergoing two different European guidelines for antifungal Hiramatsu Y, Nitenberg G, Nystrom cardiovascular surgical procedures. management in leukemia and PO, Pittet D, Rogers T, Sandven P, J Cardiovasc Surg 42:339–343 hematopoietic stem cell transplant Sganga G, Schaller MD, Solomkin J 53. Eloy O, Marque S, Mourvilliers B, Pina recipients: summary of the ECIL (1998) Epidemiology, diagnosis and P, Allouch PY, Pangon B, Bedos JP, 3–2009 update. Bone Marrow treatment of systemic Candida infection Ghnassia JC (2006) Contribution of the Transplant 46:709–718 in surgical patients under intensive care. Pittet’s index, antigen assay, IgM, and 33. Dannaoui E, Desnos-Ollivier M, Intensive Care Med 24:206–216 total antibodies in the diagnosis of Garcia-Hermoso D, Grenouillet F, 42. Solomkin JS, Flohr AB, Quie PG, invasive candidiasis in intensive care Cassaing S, Baixench MT, Bretagne S, Simmons RL (1980) The role of unit. J Mycol Med 16:113–118 Dromer F, Lortholary O, French Candida in intraperitoneal infections. 54. Agvald-Ohman C, Klingspor L, Mycoses Study G (2012) Candida spp. Surgery 88:524–530 Hjelmqvist H, Edlund C (2008) with acquired echinocandin resistance, 43. Calandra T, Bille J, Schneider R, Invasive candidiasis in long-term France, 2004–2010. Emerg Infect Dis Mosimann F, Francioli P (1989) patients at a multidisciplinary intensive 18:86–90 Clinical significance of Candida care unit: candida colonization index, 34. Pacholczyk M, Lagiewska B, Lisik W, isolated from peritoneum in surgical risk factors, treatment and outcome. Wasiak D, Chmura A (2011) Invasive patients. Lancet 2:1437–1440 Scand J Infect Dis 40:145–153 fungal infections following liver 44. Tissot F, Lamoth F, Hauser PM, Orasch 55. Massou S, Ahid S, Azendour H, transplantation—risk factors, incidence C, Fluckiger U, Siegemund M, Bensghir M, Mounir K, Iken M, and outcome. Ann Transplant 16:14–16 Zimmerli S, Calandra T, Bille J, Lmimouni BE, Balkhi H, Drissi Kamili 35. Schaenman JM (2013) Is universal Eggimann P, Marchetti O, Fungal N, Haimeur C (2013) Systemic antifungal prophylaxis mandatory in Infection Network of Switzerland candidiasis in medical intensive care lung transplant patients? Curr Opin (FUNGINOS) (2013) Beta-glucan unit: analysis of risk factors and the Infect Dis 26:317–325 antigenemia anticipates diagnosis of contribution of colonization index. 36. Toth A, Csonka K, Jacobs C, Vagvolgyi blood culture-negative intraabdominal Pathol Biol (Paris) 61:108–112 C, Nosanchuk JD, Netea MG, Gacser A candidiasis. Am J Respir Crit Care Med 56. Chabasse D (2001) Yeast count in (2013) Candida albicans and Candida 188:1100–1109 urine. Review of the literature and parapsilosis induce different T-cell 45. Eggimann P, Francioli P, Bille J, preliminary results of a multicenter responses in human peripheral blood Schneider R, Wu MM, Chapuis G, prospective study carried out in 15 mononuclear cells. J Infect Dis Chiolero R, Pannatier A, Schilling J, hospital centers. Ann Fr Anesth Reanim 208:690–698 Geroulanos S, Glauser MP, Calandra T 20:400–406 (1999) Fluconazole prophylaxis prevents intra-abdominal candidiasis in high-risk surgical patients. Crit Care Med 27:1066–1072 1448

57. Dubau B, Triboulet C, Winnock S 66. Wang DH, Gao XJ, Wei LQ, Xia R, 72. Hall AM, Poole LA, Renton B, (2001) Use of the colonization index. Sun L, Li Q, Peng M, Qin YZ (2009) Wozniak A, Fisher M, Neal T, Halloran Ann Fr Anesth Reanim 20:418–420 The preemptive treatment of invasive CM, Cox T, Hampshire PA (2013) 58. Sellami A, Sellami H, Makni F, Bahloul Candida infection with reference of Prediction of invasive candidal M, Cheikh-Rouhou F, Bouaziz M, corrected colonization index in infection in critically ill patients with Ayadi A (2006) Candiduria in intensive critically ill patients: a multicenter, severe acute pancreatitis. Crit Care care unit: significance and value of prospective, randomized controlled 17:R49 yeast numeration in urine. Ann Fr clinical study. Zhongguo Wei Zhong 73. Ostrosky-Zeichner L, Shoham S, Anesth Reanim 25:584–588 Bing Ji Jiu Yi Xue 21:525–528 Vazquez J, Reboli A, Betts R, Barron 59. Ergin F, Eren Tulek N, Yetkin MA, 67. Leon C, Ruiz-Santana S, Saavedra P, MA, Schuster M, Judson MA, Revankar Bulut C, Oral B, Tuncer Ertem G Galvan B, Blanco A, Castro C, Balasini SG, Caeiro JP, Mangino JE, Mushatt D, (2013) Evaluation of Candida C, Utande-Vazquez A, Gonzalez de Bedimo R, Freifeld A, Nguyen MH, colonization in intensive care unit Molina FJ, Blasco-Navalproto MA, Kauffman CA, Dismukes WE, Westfall patients and the use of Candida Lopez MJ, Charles PE, Martin E, AO, Deerman JB, Wood C, Sobel JD, colonization index. Mikrobiyol Bul Hernandez-Viera MA, Cava Study G Pappas PG (2014) MSG-01: a 47:305–317 (2009) Usefulness of the ‘‘Candida randomized, double-blind, placebo- 60. Laverdiere M, Rotstein C, Bow EJ, score’’ for discriminating between controlled trial of caspofungin Roberts RS, Ioannou S, Carr D, Candida colonization and invasive prophylaxis followed by preemptive Moghaddam N (2000) Impact of candidiasis in non-neutropenic critically therapy for invasive candidiasis in high- fluconazole prophylaxis on fungal ill patients: a prospective multicenter risk adults in the critical care setting. colonization and infection rates in study. Critical Care Med 37:1624–1633 Clin Infect Dis 58:1219–1226 neutropenic patients. The Canadian 68. Ellis M, Al-Ramadi B, Bernsen R, 74. Timsit JF, Azoulay E, Cornet M, Fluconazole Study. J Antimicrob Kristensen J, Alizadeh H, Hedstrom U Gangneux JP, Jullien V, Vesin A, Schir Chemother 46:1001–1008 (2009) Prospective evaluation of E, Wolff M (2013) EMPIRICUS 61. Normand S, Francois B, Darde ML, mannan and anti-mannan antibodies for micafungin versus placebo during Bouteille B, Bonnivard M, Preux PM, diagnosis of invasive Candida nosocomial sepsis in Candida multi- Gastinne H, Vignon P (2005) Oral infections in patients with neutropenic colonized ICU patients with multiple nystatin prophylaxis of Candida spp. fever. J Med Microbiol 58:606–615 organ failures: study protocol for a colonization in ventilated critically ill 69. Caggiano G, Puntillo F, Coretti C, randomized controlled trial. Trials patients. Intensive Care Med Giglio M, Alicino I, Manca F, Bruno F, 14:399 31:1508–1513 Montagna MT (2011) Candida 75. Charles PE, Castro C, Ruiz-Santana S, 62. Giglio M, Caggiano G, Dalfino L, colonization index in patients admitted Leon C, Saavedra P, Martin E (2009) Brienza N, Alicino I, Sgobio A, Favale to an ICU. Int J Mol Sci 12:7038–7047 Serum procalcitonin levels in critically A, Coretti C, Montagna MT, Bruno F, 70. Posteraro B, De Pascale G, Tumbarello ill patients colonized with Candida spp: Puntillo F (2012) Oral nystatin M, Torelli R, Pennisi MA, Bello G, new clues for the early recognition of prophylaxis in surgical/trauma ICU Maviglia R, Fadda G, Sanguinetti M, invasive candidiasis? Intensive Care patients: a randomised clinical trial. Crit Antonelli M (2011) Early diagnosis of Med 35:2146–2150 Care 16:R57 candidemia in intensive care unit 76. Wo´jtowicz A, Tissot F, Lamoth F, 63. Chen Z, Yang CL, He HW, Zeng J patients with sepsis: a prospective Orasch C, Eggimann P, Siegemund M, (2013) The clinical research of nystatin comparison of (1?3)-beta-D-glucan Zimmerli S, Flueckiger UM, Bille J, in prevention of invasive fungal assay, Candida score, and colonization Calandra T, Marchetti M, Bochud PY infections in patients on mechanical index. Crit Care 15:R249 (2014) Polymorphisms in TNF-a ventilation in intensive care unit. 71. Peman J, Zaragoza R, Quindos G, increase susceptibility to intra- Zhonghua Wei Zhong Bing Ji Jiu Yi Alkorta M, Cuetara MS, Camarena JJ, abdominal candida infection in high Xue 25:475–478 Ramirez P, Gimenez MJ, Martin- risk surgical ICU patients. Critical Care 64. Piarroux R, Grenouillet F, Balvay P, Mazuelos E, Linares-Sicilia MJ, Ponton Med 42:e304 Tran V, Blasco G, Millon L, Boillot A J, Study group Candida albicans Germ (2004) Assessment of preemptive Tube Antibody Detection in Critically treatment to prevent severe candidiasis Ill Patients (2011) Clinical factors in critically ill surgical patients. Crit associated with a Candida albicans Care Med 32:2443–2449 germ tube antibody positive test in 65. Eren A, Aydogan S, Kalkanci A, intensive care unit patients. BMC Infect Kustimur S (2007) The relationship Dis 11:60 between Candida albicans colonization indices and the presence of specific antibodies in non-neutropenic intensive care unit patients. Mikrobiyol Bul 41:253–259