Candida tropicalis bloodstream infection: Incidence, risk factors and outcome in a population-based surveillance

Journal of Infection (2015) xx, 1e10

www.elsevierhealth.com/journals/jinf

Candida tropicalis bloodstream infection: Incidence, risk factors and outcome in a population-based surveillance*
ndez-Ruiz a,*, Mireia Puig-Asensio b, Jesu
s Guinea c, Mario Ferna ´n Padilla c, Manuel Almela d, Benito Almirante b, Bele ~ o f,
s Rodrı´guez-Ban Ana Dı´az-Martı´n e, Jesu g ´ Marı´a Aguado a, on behalf of the Manuel Cuenca-Estrella , Jose CANDIPOP Projecth, GEIH-GEMICOMED (SEIMC), REIPI Unit of Infectious Diseases, Hospital Universitario “12 de Octubre”, Instituto de Investigacio
n Hospital “12 de Octubre” (iþ12), Universidad Complutense, Madrid, Spain b Department of Infectious Diseases, Hospital Universitari “Vall d’Hebron”, Department of Medicine, Universitat Autonoma de Barcelona, Barcelona, Spain c Department of Clinical Microbiology and Infectious Diseases, Hospital General Universitario “Gregorio Maran~o
n”, Universidad Complutense, Madrid, Spain d Department of Microbiology, Hospital Clinic, Institut d’Investigacions Biomediques August Pi i Sunyer (IDIBAPS), Barcelona, Spain e Clinical Unit of Critical Care and Emergency Medicine, Hospital Universitario “Virgen del Rocı´o”, Sevilla, Spain f Clinical Unit of Infectious Diseases and Microbiology, Hospital Universitario “Virgen Macarena”, Sevilla, Spain g Department of Mycology, Spanish National Centre for Microbiology, Instituto de Salud Carlos III, Majadahonda, Madrid, Spain a

Accepted 16 May 2015 Available online - - -

KEYWORDS Candida tropicalis bloodstream infection;

Summary Objective: To assess the current clinical features and determinants of outcome of Candida tropicalis bloodstream infection (BSI). Methods: A population-based surveillance on Candida BSI was conducted from May 2010 to April 2011 in 29 Spanish hospitals. Antifungal susceptibility testing (EUCAST methodology)

* This study was partially presented at the 24th European Congress of Clinical Microbiology and Infectious Diseases (e-Poster 356), Barcelona, May 10e13, 2014. * Corresponding author. Unit of Infectious Diseases, Hospital Universitario “12 de Octubre”, Centro de Actividades Ambulatorias,
rdoba, s/n, 28041, Madrid, Spain. Tel.: þ34 913908000x4631; fax: þ34 914695775. 2a planta, bloque D, Avda. de Co
ndez-Ruiz). E-mail address: [email protected] (M. Ferna h The members of the CANDIPOP Project are listed in the Acknowledgements section.

http://dx.doi.org/10.1016/j.jinf.2015.05.009 0163-4453/ª 2015 The British Infection Association. Published by Elsevier Ltd. All rights reserved.
ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009


ndez-Ruiz et al. M. Ferna

2 Epidemiology; Minimum inhibitory concentration; Treatment; Outcome

was centrally performed. The characteristics and outcome of C. tropicalis BSI episodes were compared with those due to other species. Results: Fifty-nine out of 752 episodes (7.8%) were due to C. tropicalis (annual incidence: 0.62 cases per 100,000 population). Resistance to fluconazole and voriconazole was found in 23.2% and 26.8% of isolates. Breakthrough BSI occurred in 10.5% of episodes. Risk factors for C. tropicalis BSI were age (odds ratio [OR]: 1.01; P-value Z 0.05), underlying leukaemia (OR: 4.77; Pvalue Z 0.001) and chronic lung disease (OR: 2.62; P-value Z 0.002). There were no differences in clinical failure (persistent BSI for 72 h after initiation of therapy and/or 30-day all-cause mortality) between C. tropicalis (39.6%) and non-C. tropicalis groups (45.6%). The appropriateness of antifungal therapy or the fluconazole MIC values had no significant impact on outcome, whereas early central venous catheter removal exerted a protective effect. Conclusions: C. tropicalis BSI was associated with advanced age, haematological malignancy and respiratory comorbidity. We found no correlation between the unexpectedly high resistance rate to azoles observed and outcome. ª 2015 The British Infection Association. Published by Elsevier Ltd. All rights reserved.

Introduction Candida tropicalis ranks the third to fourth among the most commonly isolated species in patients with Candida spp. bloodstream infection (BSI). Its relative importance seems to be even greater in series reported from Southern Europe 1,2 and Latin American.3,4 As compared to other species, C. tropicalis depicts various distinct features that increase its intrinsic pathogenicity,5 including its capability for secretion of extracellular hydrolytic enzymes (coagulase, phospholipase or proteinase),6,7 biofilm formation7 and phenotypic switching.8 In addition, clinical studies have suggested a particular susceptibility to C. tropicalis BSI among neutropenic patients with haematological and solid-organ malignancies.9e11 Therefore, the outcome of this form of invasive fungal disease is generally assumed to be poorer in comparison to that of BSI due to other Candida species.12 Most of the previous studies on C. tropicalis BSI were limited by their single-centre design,9e11 often exclusively focused on specific patient populations,9,10 and were performed years ago by encompassing long recruitment periods.10,11,13 These drawbacks preclude the assessment of recent trends in antifungal susceptibility patterns or the impact on outcome of the widespread use of newer agents and current recommendations regarding central venous catheter (CVC) management.14,15 In contrast, a population-based surveillance may provide a comprehensive picture of the contemporary characteristics of this condition, which could acquire particular relevance in view of the emerging reports of episodes of breakthrough BSI involving azole-resistant C. tropicalis strains.16,17 The present study was aimed at gaining insight into the epidemiological and clinical features, therapeutic approaches and determinants of outcome of patients with C. tropicalis BSI through analysis of the data derived from a populationbased surveillance program.

Patients and methods Study setting and design The CANDIPOP Project (ClinicalTrials.gov number: NCT01236261) was a prospective population-based

surveillance program on both Candida and non-Candida yeasts causing BSI (yeast fungaemia) carried out between May 2010 and April 2011 in 29 hospitals located in five cities of Spain (Barcelona, Bilbao, Madrid, Seville, and Valencia). In 2011 these centres served an overall population of 9,498,980 inhabitants (about 20% of the Spanish population). Its methodology and main findings have been described elsewhere.18,19 Briefly, all consecutive episodes of yeast fungaemia diagnosed at the participating institutions were deemed eligible for inclusion. Cases were identified by local laboratories and reported to study coordinators, who collected data on clinical variables and 30-day outcome and periodically performed internal audits to ensure complete case capture. Patients were managed according to the standard of care at each participating centre. The study protocol was approved by the local ethics committees, and written informed consent was obtained at enrolment. In this sub-study we compared the demographics and clinical characteristics, predisposing conditions, clinical presentation, therapy and outcome of episodes of C. tropicalis BSI with those of BSI due to other Candida species (non-C. tropicalis BSI group). We also specifically analysed the predictors of outcome in the C. tropicalis BSI group. The primary outcome was clinical failure, defined as: (1) all-cause mortality within the first 30 days from the first positive blood culture, and/or (2) persistent C. tropicalis BSI for 72 h after the initiation of antifungal therapy.20 As secondary outcomes we separately analysed early (0e7 days) and late (8e30 day) all-cause mortality.18

Definitions and data collection An incident episode was defined as the first blood culture yielding a yeast-like fungus throughout the study period (index culture). Episodes of mixed yeast fungaemia in which a second non-C. tropicalis species was simultaneously isolated were excluded from the analysis. The isolation of the same yeast species in a given patient >30 days after the index culture was considered as a distinct episode, as well as the subsequent isolation at any time of a second different species. In order to preserve the assumption of independence of observations, only the first episode of BSI observed for an individual patient was


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009

Epidemiology and outcome in C. tropicalis candidemia included in the outcome analysis. Proven catheter-related Candida BSI was defined as detailed elsewhere,21 whereas the diagnosis of other secondary sources required microbiological documentation of an alternative site of infection. Episodes without apparent portal of entry were deemed as primary. A breakthrough episode was defined by the occurrence of Candida BSI in a patient receiving systemic antifungal therapy for any reason for at least 72 h before the date of the index culture. Severity of infection and Pitt bacteraemia score22 were recorded in all patients on the day of index culture sampling, as well as the Acute Physiology and Chronic Health Evaluation (APACHE) II score in those admitted to the Intensive Care Unit. Comorbidity was measured according to Charlson index.23 Early CVC removal was defined as removal of the line within the first 48 h from the incident BSI (removal within this timeframe of all CVCs in patients with multiple vascular accesses was required). Initial antifungal therapy was that provided within the first 72 h of administration of systemic antifungals for an incident episode of Candida BSI. Adequate antifungal therapy required the administration of the recommended dose of an antifungal drug within the first 48 h after blood culture collection for a susceptible isolate according to the clinical breakpoints (CBPs) proposed by the European Committee on Antimicrobial Susceptibility Testing (EUCAST),24 as detailed below [http://www. eucast.org/clinical_breakpoints/]. Since the EUCAST has not yet established species-specific CBPs for C. tropicalis and caspofungin or micafungin, we tentatively used the CBP of anidulafungin for categorizing the appropriateness of therapy with both agents.

3 Laboratory Standard Institute (CLSI) M27-A3 method.26 This second procedure was carried out at the Hospital Gen~o
n” (Madrid). The eral Universitario “Gregorio Maran comparative results obtained by both methods have been detailed elsewhere.19 Of note, for the purposes of the present study only the EUCAST procedure was taken into account for analysing the adequacy of antifungal therapy and the correlation of antifungal susceptibility testing with outcome.

Statistical analysis Quantitative data were shown as the mean  standard deviation (SD) or the median with interquartile (Q1eQ3) ranges. Qualitative variables were expressed as absolute and relative frequencies. Categorical variables were compared using the c2 test or Fisher’s exact test, whereas Student’s T test or U ManneWhitney test were applied for continuous variables, as appropriate. We analysed the factors associated with C. tropicalis BSI by using a backward stepwise logistic regression model in which variables found to be significant at the univariate level (P-value <0.05) were introduced. We also attempted to specifically assess the predictive factors for clinical failure within the C. tropicalis BSI group. Associations were given as odds ratios (ORs) with 95% confidence intervals (CIs). All the significance tests were two-tailed. Statistical analysis was performed using SPSS, version 15.0 (Statistical Package for Social Sciences Inc., Chicago, IL).

Results Microbiological studies Yeast isolates were initially processed at participating hospitals using local routine methods. All the strains were then forwarded to the Mycology Reference Laboratory at the Spanish National Centre for Microbiology (Majadahonda, Madrid) for species confirmation and antifungal susceptibility testing. The molecular identification was achieved by sequencing the internal transcribed spacer 1 (ITS1) and ITS2 regions from the ribosomal RNA.25 When identification was discordant, the data from the reference laboratory were used.19 Susceptibility to antifungal drugs was assessed according to the EUCAST microdilution broth reference method, by using the following range of concentrations: amphotericin B, caspofungin, micafungin and anidulafungin, 0.03e16 mg/L; fluconazole, 0.125e64 mg/L; and voriconazole and posaconazole, 0.015e8 mg/L. The optical densities of the plates were measured after 24 h of incubation at 35  C. The minimum inhibitory concentration (MIC) value was defined as the antifungal concentration that produced 90% (amphotericin B) or 50% (remaining antifungal agents) of growth inhibition relative to that of the drug-free growth control. Species-specific CBPs proposed by the EUCAST were applied. In detail, C. tropicalis isolates with MIC values of >4 mg/L for fluconazole, >0.12 mg/L for voriconazole, >0.06 mg/L for posaconazole and anidulafungin, and >1 mg/L for amphotericin B were considered as resistant. In addition, antifungal susceptibility testing was also performed in all isolates according to the Clinical

Incidence rates We identified 773 episodes of yeast fungaemia during the surveillance period. Twenty-one case-patients did not consent to participate in the study. Out of the remaining 752 episodes, 59 (7.8%) were due to C. tropicalis, accounting for crude annual incidence rates of 0.62 cases per 100,000 population, 0.07 per 1000 hospital admissions, and 0.10 per 10,000 patients-days. A second Candida species was simultaneously isolated in the incident blood culture in 2 episodes, which were subsequently excluded. Nineteen episodes of fungaemia due to non-Candida yeasts were also excluded from the non-C. tropicalis BSI group. Therefore, we analysed 57 and 689 episodes in the C. tropicalis BSI and non-C. tropicalis BSI groups, respectively.

Underlying conditions Comparison of host characteristics and predisposing factors is shown in Table 1. Patients in the C. tropicalis BSI group were older, more likely to have an underlying haematological malignancy (mainly leukaemia) and to have been diagnosed with a chronic lung disease and congestive heart failure, and to have undergone a haematopoietic stem cell transplantation. Neutropenia and mucositis were also more commonly observed in this group, although these differences did not reach statistical significance.


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009


ndez-Ruiz et al. M. Ferna

4 Table 1

Comparison between episodes of C. tropicalis BSI and non-C. tropicalis BSI.

Age, years [mean  SD] Neonate (1 year) [n (%)] Gender (male) [n (%)] Hospital-acquired episodes [n (%)] Hospital days until diagnosis [median (Q1eQ3 range)] Recent health-care exposure [n (%)]b Charlson comorbidity index [median (Q1-Q3 range)] Comorbidities [n (%)] Diabetes mellitus Solid-organ cancer Resistant/relapsing or metastatic Haematological malignancy Leukaemia Lymphoma Renal failure Renal replacement therapy Liver disease Chronic lung disease COPD Asthma Chronic heart disease Congestive heart failure Solid-organ transplantation Haematopoietic stem cell transplantation Predisposing factors [n (%)] Vascular catheter in place at diagnosis CVC in placec Previous surgeryd Parenteral nutrition Orotracheal intubation at diagnosis Neutropenia (<0.5  103 cells/mL) Mucositis Immunosuppressive therapye,f Prior antibiotic therapyf,g Prior antifungal exposuref Echinocandin exposure Fluconazole exposure Breakthrough BSI Previous or concurrent colonization by the same Candida species

C. tropicalis BSI (n Z 57)

Non-Candida tropicalis BSIa (n Z 689)

P-value

63.0  22.8 4 (7.0) 30 (52.6) 55 (96.5) 18 (10e35) 27 (47.4) 2 (1e3)

53.9  27.9 88 (12.8) 410 (59.5) 652 (94.6) 20 (11e36) 326 (47.4) 2 (0e3)

0.006 0.204 0.311 0.544 0.691 0.998 0.133

16 17 8 11 8 3 14 2 5 21 10 2 24 9 5 3

(28.1) (29.8) (14.0) (19.3) (14.0) (5.3) (24.6) (3.5) (8.8) (36.8) (17.5) (3.5) (42.1) (15.8) (8.8) (5.3)

144 200 70 46 28 18 179 47 98 146 66 10 221 58 43 8

(20.9) (29.0) (10.2) (6.7) (4.1) (2.6) (26.0) (6.8) (14.2) (21.2) (9.6) (1.5) (32.1) (8.4) (6.2) (1.2)

0.205 0.899 0.358 0.002 0.004 0.212 0.814 0.257 0.252 0.006 0.056 0.232 0.121 0.061 0.301 0.045

54 36 27 20 7 5 6 14 54 9 2 5 6 18

(94.7) (63.2) (47.4) (35.1) (12.3) (8.8) (10.7) (24.6) (94.7) (15.8) (3.5) (8.8) (10.5) (31.6)

651 522 353 344 181 29 35 151 639 140 35 97 85 265

(94.5) (76.0) (51.2) (49.9) (26.3) (4.2) (5.1) (21.9) (93.3) (20.3) (5.1) (14.1) (12.3) (38.6)

0.616 0.032 0.575 0.031 0.019 0.109 0.080 0.644 0.469 0.411 0.451 0.262 0.688 0.296

BSI: bloodstream infection; COPD: chronic obstructive pulmonary disease; CVC: central venous catheter; Q1eQ3 range: interquartile range; SD: standard deviation. a The most common species in the this group were C. albicans (346 episodes [50.2%]), C. parapsilosis (187 [27.1%]), C. glabrata (98 [14.2%]) and C. krusei (14 [2.0%]). b Hospitalization for 24 h within the preceding 3 months. c Included non-tunnelled temporary catheters, peripherally inserted central catheters, totally implantable catheters (Port-a-Cath), and skin-tunnelled catheters. d Within the preceding 3 months. e Includes corticosteroids, cytotoxic agents, and other immunosuppressive drugs. f Within the preceding month. g Data not available in 4 episodes.

Conversely, the presence of a CVC in place and orotracheal intubation at infection onset was more frequent in the non-C. tropicalis BSI group, as well as the prior receipt of parenteral nutrition. The multivariate analysis revealed that the factors independently associated with

C. tropicalis BSI were age (OR [per unitary increment]: 1.01; 95% CI: 1.00e1.02; P-value Z 0.05) and the presence of underlying leukaemia (OR: 4.77: 95% CI: 1.96e11.6; Pvalue Z 0.001) and chronic lung disease (OR: 2.62; 95% CI: 1.44e4.77; P-value Z 0.002).


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009

Epidemiology and outcome in C. tropicalis candidemia

Clinical presentation, therapy and outcome There were no significant differences between episodes of C. tropicalis and non-C. tropicalis BSI in the source of infection, nor in the markers of haemodynamic instability or severity of illness, as detailed in Table 2. Haematogenous dissemination leading to invasive disease was documented in only one episode of C. tropicalis BSI (metastatic renal and pulmonary infection). With regard to the therapeutic approaches, overall CVC removal was less commonly performed among episodes of C. tropicalis BSI group, although the rate of early removal was comparable between both

Table 2

5 groups. Again, we found no significant differences in the type of antifungal regimen initially used. Nine episodes (15.8%) in the C. tropicalis BSI group received no antifungal therapy. Initial therapy in the remaining cases consisted of an azole-based regimen in 26 episodes (54.2%; 24 fluconazole, one voriconazole, and one fluconazole followed by voriconazole), an echinocandinbased regimen in 12 (25.0%; 8 caspofungin, 3 anidulafungin, and one anidulafungin followed by micafungin), an amphotericin B-based regimen in 6 (12.5%), and a combination regimen in 4 (8.3%). After excluding the 6 episodes already receiving antifungal drugs at the onset of infection, median

Clinical features, therapeutic approaches, and outcome in both study groups. C. tropicalis BSI (n Z 57)

Source of BSI [n (%)] Primary Catheter-related Abdominal Urologic Severity at onset [n (%)] Sepsis Severe sepsis Septic shock APACHE II score [mean  SD]a Pitt score [median (Q1eQ3 range)] Concomitant bacteraemia in incident blood culture [n (%)] Requirement for ICU admission [n (%)] Haematogenous dissemination [n (%)] Endocarditis Chorioretinitis Otherb Therapeutic interventions [n (%)] CVC removal Early CVC removal (48 h)c Administration of antifungal therapy Initial antifungal therapy (72 h) Azole-based regimen Echinocandin-based regimen Amphotericin B-based regimen Combination regimen Clinical failure [n (%)]e Persistent BSI for 72 h from the initiation of therapy [n (%)]f All-cause mortality [n (%)]e Early (0e7 days) Late (8e30 days)

33 16 3 5

(57.9) (28.1) (5.3) (8.8)

36 (63.2) 11 (19.3) 10 (17.5) 22.2  11.4 2 (1e3) 10 (17.5) 6 (10.5) 0 (0.0) 0 (0.0) 1 (1.8)

Non-Candida tropicalis BSI (n Z 689) 388 238 22 35

(56.3) (34.5) (3.2) (5.1)

472 (68.5) 114 (16.5) 103 (14.9) 19.7  7.2 1 (1e4) 135 (19.6) 53 (7.7) 12 (1.7) 20 (2.9) 20 (2.9)

P-value

0.817 0.322 0.297 0.183 0.707

0.455 0.490 0.707 0.291 0.382 0.200 0.513

23/36 (63.9) 10/23 (43.5) 48 (84.2)d

411/522 (78.7) 176/405 (43.5) 618 (89.7)

0.038 0.998 0.198

26/48 12/48 6/48 4/48 21/53 5/28

284/618 157/618 95/618 82/618 308/676 129/453

(46.0) (25.4) (15.4) (13.3) (45.6) (28.5)

0.272 0.951 0.593 0.326 0.403 0.224

86/687 (12.5) 126/590 (21.4)

0.252 0.670

(54.2) (25.0) (12.5) (8.3) (39.6) (17.9)

10/56 (17.9) 8/43 (18.6)

APACHE: Acute Physiology and Chronic Health Evaluation; BSI: bloodstream infection; CVC: central venous catheter; ICU: intensive care unit; Q1-Q3 range: interquartile range; SD: standard deviation. a Data available for patients admitted to ICU at BSI onset (182 episodes). b Includes metastatic central nervous system (7 episodes), cutaneous (5 episodes), renal (4 episodes), pulmonary and splenic (3 episodes each), and hepatic (one episode) involvement. c Timing of CVC removal were not available in 6 episodes. d Nine episodes were not treated in the C. tropicalis BSI group due to the following reasons: patient death within the first 48 h from blood culture sampling and, therefore, before the results could be notified to the attending physician (4 episodes); the results of blood cultures were received after cessation of therapeutic measures (2 episodes); the patient had been already discharged when the results of blood cultures were received (2 episodes); and no evident reason (one episode). e Seventeen episodes (2.3%) did not have 30-day follow-up data. f Follow-up blood cultures were obtained beyond the first 48 h of antifungal treatment in 481 episodes (73.0% of those in which antifungal therapy was administered and patient survived for 72 h).


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009


ndez-Ruiz et al. M. Ferna

6 interval until initiation of therapy was 1.0 days (Q1eQ3 range: 0.0e2.0). Antifungal therapy was deemed inappropriate in 34 episodes (59.6%). The reasons for therapy being deemed inappropriate were the delay in its initiation for more 48 h from blood culture collection (58.8% [20/34]), the initial use of a drug to which the isolate was found to be non-susceptible in vitro (fluconazole in 8.8% [3/34] and caspofungin in 23.5% [8/34]), or the lack of antifungal therapy (26.5% [9/34]). Six episodes simultaneously met two of these criteria for inappropriateness. Follow-up blood cultures were obtained beyond the first 48 h from the initiation of antifungal treatment in 28 episodes of C. tropicalis BSI and 453 episodes of non-C. tropicalis BSI (59.6% and 73.9%, respectively, of those episodes that received antifungal therapy and in which the patient survived for 72 h). Overall, 39.6% (21/53) and 45.6% (308/676) of evaluable episodes met the criteria for clinical failure (primary study outcome) in the C. tropicalis BSI and non-C. tropicalis BSI groups, respectively. Persistent C. tropicalis BSI for 72 h from the initiation of therapy occurred in 17.9% (5/28) of evaluable episodes. Early and late all-cause mortality rates were 17.9% and 18.6%. The rate of clinical failure in the 9 episodes of C. tropicalis BSI that did not receive antifungal therapy was 66.7% (6/9), and was mainly driven by a high early all-cause mortality (55.6% [5/9]).

Antifungal susceptibility The results of antifungal susceptibility testing for the C. tropicalis isolates are shown in Table 3. According to species-specific CBPs proposed by the EUCAST, the rates of resistance to fluconazole (23.2%), voriconazole (26.8%) and posaconazole (19.6%) were notably high, whereas resistance to anidulafungin was very low (3.6%). No isolates showed resistance to amphotericin B. Of note, the rate of resistance to fluconazole (1.7%) was markedly lower when the MIC values were obtained by the CLSI method, whereas no isolates resistant to voriconazole or posaconazole were identified (data not shown).

Outcome predictors in C. tropicalis BSI In the univariate analysis for predictors of clinical failure, a number of comorbid conditions (renal and congestive heart failure), predisposing factors (prior corticosteroid therapy),

Table 3

and markers of severity of clinical presentation (severe sepsis or septic shock and Pitt bacteraemia score) were associated with a worse outcome, whereas early CVC removal exerted a protective effect (Table 4). Predictors of early and late mortality (secondary study outcomes) also included CVC removal, as well as certain host characteristics (Tables S1 and S2 in Supplementary Data). Multivariate analyses for these outcomes could not be performed due to the low number of events. We did not identify any effect of the appropriateness of antifungal therapy on the odds of clinical failure (OR: 0.45; 95% CI: 0.14e1.47; P-value Z 0.183) or early (OR: 0.56; 95% CI: 0.13e2.43; P-value Z 0.339) or late mortality (OR: 0.39; 95% CI: 0.07e2.24; P-value Z 0.254). Thus, we further analysed the potential impact of reduced fluconazole susceptibility on the outcome. As depicted in Table 5, there was no clear correlation between the MIC values for fluconazole (determined by the EUCAST method) and early or late mortality or the odds of clinical failure, either in the overall cohort or in episodes in which fluconazole was administered as initial antifungal therapy.

Discussion This present study is one of the largest clinical series specifically focused on C. tropicalis BSI reported to date. Our methodological approach takes advantage of a prospective design with a contemporary case-enrolment period and centralized species confirmation and antifungal susceptibility testing, thus overcoming some limitations of previous studies.9e11,13 C. tropicalis was the fourth most frequently isolated species in our surveillance.19 We found an annual incidence close to that reported for 2008 in a retrospective study carried out in one of the participating centres,11 but considerably higher than that from a populationbased surveillance performed a decade ago in Barcelona (0.44 cases per 100,000 population).1 Some insight into this epidemiological trend may be gained by identifying the specific predisposing factors for the occurrence of C. tropicalis BSI. Increased age and certain underlying conditions were found to be independently associated with this event and din conjunction with a lower prevalence of CVC in place and orotracheal intubation at infection onsetd to delineate a distinctive risk profile compared to other Candida species. The increased risk of invasive

MIC distributions among C. tropicalis isolates according to the EUCAST methodology (n Z 56).

Antifungal agent

Caspofungin Anidulafungin Micafungin Amphotericin B Fluconazole Posaconazole Voriconazole

GM (mg/L)

Mode (mg/L)

MIC90 (mg/L)

0.41 0.034 0.034 0.079 1.83 0.047 0.13

0.5 0.03 0.03 0.06 0.5 0.02 0.03

0.5 0.03 0.03 0.12 >64 8 >8

Number of isolates at each MIC (mg/L) 0.015

0.03

0.06

53 53 9

1 1 22

3 15

2 10

0.12 1

40 10

1 20 1 2 6

0.25

0.5

1.0

20 1

30

4 1

4 12 1 3

1 14

2.0

4.0

8.0

16.0

1

1 2 2

5 8

32.0

1 1

1

11 1

5

12

EUCAST: European Committee on Antimicrobial Susceptibility Testing; GM: geometric mean; MIC: minimum inhibitory concentration.


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009

Epidemiology and outcome in C. tropicalis candidemia

7

Table 4 Prognostic factors for clinical failure (all-cause mortality within the first 30 days and/or persistent BSI for 72 h from the initiation of therapy) in 53 evaluable episodes of C. tropicalis BSI. Failure (n Z 21)

Variable

Host factors Age, years [mean (SD)] Charlson comorbidity index [median (Q1eQ3 range)] Solid organ cancer Haematological malignancy Renal failure Liver disease Chronic lung disease Congestive heart failure Predisposing factors Parenteral nutrition Orotracheal intubation at diagnosis Neutropenia (<0.5  103 cells/mL) Corticosteroid therapy (within the preceding month) Clinical features Severe sepsis or septic shock Pitt score [median (Q1eQ3 range)] Therapeutic interventions CVC removala Early CVC removal (48 h)a Initial antifungal therapyb Azole-based regimen Echinocandin-based regimen Adequate antifungal therapy

Not failure (n Z 32)

Univariate analysis OR

95% CI

P-value

1.02 1.06

0.99e1.04 0.80e1.41

0.290 0.683

(34.4) (15.6) (12.5) (3.1) (28.1) (6.2)

0.59 1.69 4.31 7.29 2.81 6.00

0.17e2.06 0.42e6.74 1.09e16.93 0.75e70.58 0.89e8.89 1.08e33.78

0.412 0.345 0.034 0.074 0.075 0.035

10 (47.6) 4 (19.0)

9 (28.1) 3 (9.4)

2.32 2.28

0.73e7.35 0.45e11.40

0.148 0.270

3 (14.3)

2 (6.2)

2.50

0.38e16.42

0.304

8 (38.1)

4 (12.5)

4.31

1.09e16.93

0.034

9 (28.1) 1 (0e2)

3.41 1.40

1.07e10.85 1.02e1.91

0.035 0.035

6/16 (37.5) 1/16 (6.2)

16/18 (88.9) 9/18 (50.0)

0.08 0.07

0.01e0.45 0.01e0.62

0.002 0.006

8/16 (50.0) 6/16 (37.5)

16/29 (55.2) 6/29 (20.7)

0.81 2.30

0.24e2.76 0.59e8.90

0.739 0.192

6 (28.6)

15 (46.9)

0.45

0.14e1.47

0.183

67.1  20.3 3 (1.5e4) 5 5 8 4 11 6

60.1  24.7 2 (1e3)

(23.8) (23.8) (38.1) (19.0) (52.4) (28.6)

11 5 4 1 9 2

12 (57.1) 3 (1.25e4)

BSI: bloodstream infection; CI: confidence interval; CVC: central venous catheter; Q1eQ3 range: interquartile range; OR: odds ratio; SD: standard deviation. a Percentages calculated on evaluable episodes (i.e., 30-day follow-up data) with a CVC in place at BSI onset. b Percentages calculated on evaluable episodes that received antifungal therapy.

disease due to C. tropicalis faced by patients with cancer has been well established.1,9e11,13 In line with previous studies,9,10 we have been able to further circumscribe this association to patients with underlying haematological malignancies and, specifically, leukaemia. The higher prevalence of mucositis and neutropenia in the presence of species-specific virulence factors would overall suggest an endogenous source of infection within the gastrointestinal tract, in contrast to the often device-associated

Table 5

origin of non-C. tropicalis BSI (i.e., vascular catheters and Candida parapsilosis).5 Although the isolation of C. tropicalis from different body sites has good predictive value for the subsequent development of invasive disease,5 we did not find differences in the rate of previous or concurrent colonization by the same species between study groups. However, it should be noted that the intensity of sampling for colonization was not standardized across participating centres.

Correlation of fluconazole MIC values (determined according to the EUCAST method) with outcome.

Fluconazole MIC (mg/L)

All episodes (n [%])

Episodes initially treated with fluconazole (n [%])

Early mortality

Late mortality

Clinical failure

Early mortality

Late mortality

Clinical failure

0.25 0.5e2.0 4.0

7.7 (1/13) 21.4 (6/28) 15.4 (2/13)

18.2 (2/11) 23.8 (5/21) 10.0 (1/10)

25.0 (3/12) 48.1 (13/27) 33.3 (4/12)

0.0 (0/5) 6.2 (1/16) 0.0 (0/3)

40.0 (2/5) 14.3 (2/14) 0.0 (0/2)

40.0 (2/5) 33.3 (5/15) 0.0 (0/2)

MIC: minimum inhibitory concentration.


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009


ndez-Ruiz et al. M. Ferna

8 Hyphal formation and production of secreted hydrolases and haemolysins are among the virulence mechanisms harboured by C. tropicalis which increase its capacity for tissue invasion.5e8 In keeping with this intrinsic pathogenicity, various authors have reported higher mortality rates for episodes of C. tropicalis BSI as compared to those due to other species, even in the absence of apparent differences in concurrent signs of haemodynamic compromise.1,11 In their quantitative review of randomized trials of treatment of invasive candidiasis, Andes et al. demonstrated that infection with C. tropicalis acted as an independent predictor of 30-day mortality after adjusting for APACHE score and treatment-related factors.12 In contrast, we found no differences in the rate of clinical failure or all-cause mortality between C. tropicalis and other species. Such lack of impact on outcome persisted even when comparison was restricted to episodes of BSI due to Candida albicans (data not shown). It should be noted that both the 7- and 30-day mortality rates in our experience (17.9% and 34.0%, respectively) were lower than in previous reports.1,3,11,13 For instance, the early mortality observed in the aforementioned surveillance carried out in Barcelona between 2002 and 2003 (33%) was almost double than that of the present study.1 It could be hypothesized that continuous improvements over the last years in both the acute management of Candida BSI (i.e., introduction of echinocandins or prompt CVC removal) and supportive care of underlying conditions (i.e., haematological malignancies) might have favourably impacted to a greater extent on the outcome of C. tropicalis BSI as compared to other less virulent species. One of the most surprising findings in our study is the unexpectedly high resistance rate to azoles obtained according to the EUCAST guidelines. Enhanced drug efflux activation and point mutations in the erg11 gene have been proposed as the main resistance mechanisms.27 Although there have been increasing reports of azole resistance for C. tropicalis,16,17,27 the present figures are far higher than those previously observed in our setting and others, which ranged from 0% to 2.3% either by using the CLSI11,13 or EUCAST methods.28,29 However, marked discrepancy appeared when the strains of the present study were tested by the CLSI (M27-A3) method.19 Among other methodological issues, both procedures differ at the end-point reading of the microdilution plates to compare the growth inhibition achieved at the tested drug concentration and that of the drug-free control. Whereas the EUCAST method demands an objective measurement of optical density by means of a microdilution plate reader, the CLSI relies on a subjective assessment as determined by the naked eye.24,26 Although one would expect a more accurate and reproducible assessment of the antifungal susceptibility from the EUCAST procedure, the in vivo correlate with the clinical outcome has been found to be poor in the present study. The most plausible explanation for this apparent contradiction lies in the presence of an altered susceptibility phenotype termed as trailing effect or paradoxical growth,30 a phenomenon that mirrors the longknown Eagle effect observed for b-lactams.31 Indeed, we detected a strong in vitro paradoxical growth at increasing azole concentrations and exposure in all the isolates with

discordant susceptibility testing results between both methods, as growth inhibition was below 50% compared to that for the drug-free control.19 The clinical relevance of this phenomenon remains unclear since it seems not to be related with the selection of resistant subpopulations30 and previous studies have been unable to reproduce in vivo the paradoxical effect seen in vitro with caspofungin against C. albicans.32 However, the confounding impact of the trailing effect should be taken into account when assessing azole susceptibility by means of the EUCAST method in C. tropicalis isolates. Thus, on the sole basis of the MIC values obtained by this procedure and by interpreting them according to its species-specific CBPs, antifungal therapy should be deemed as inadequate in as much as 59.6% of evaluated C. tropicalis BSI episodes. Despite this fact, there was no apparent effect of the appropriateness of therapy on any of the outcomes examined, whereas dand in line with previous studies11,12d early CVC removal was the only measure identified to exert a favourable impact. In an attempt to further determine the potential role of this altered phenotype, we found no correlation between fluconazole MIC values dconsidered as a quantitative variabled and outcome, nor even when the analysis was focused on those episodes initially treated with fluconazole in monotherapy. Notwithstanding the relatively large sample size, the present study is limited by the number of episodes of C. tropicalis BSI included, particularly in subgroup analyses, thus precluding us from assessing risk factors for clinical failure in a multivariate way and making our results prone to a type 2 error due to insufficient statistical power. In addition, clinical management (i.e., extraction of followup blood cultures or timing of CVC removal) was not standardized across participating centres, and some findings might be susceptible to confounding by indication bias. In conclusion, the incidence of C. tropicalis BSI seems to have increased over the last years, mainly in older patients with haematological malignancy and chronic respiratory comorbidity. Despite the increased pathogenicity of this species and previous reports in the literature, we found no significant differences in outcome compared to episodes due to other Candida species, suggesting the protective impact of recent improvements in the therapeutic approach to patients with candidemia. Of note, the unexpectedly high resistance rate to azoles obtained by using the EUCAST methodology had no clear correlate with the clinical outcome, even when the initial therapy consisted exclusively of fluconazole.

Funding sources This work was supported by non-restrictive research grants from Gilead, MSD, Astellas, and Pfizer. This study was co
n SEIMC-GESIDA, by the Spanish Minisfounded by Fundacio try of Economy and Competitiveness, Instituto de Salud Carlos III (co-financed by the European Development Regional Fund [ERDF] “A way to achieve Europe”), and by the Spanish Network for the Research in Infectious Diseases (REIPI RD12/0015). These sources of funding had no involvement in the preparation of the manuscript.


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009

Epidemiology and outcome in C. tropicalis candidemia

Conflict of interest
ndez-Ruiz has received honoraria for talks on Mario Ferna
s Guinea has received grant support behalf of Pfizer. Jesu ~a Founfrom Astellas Pharma, MICOLAB, the Mutua Madrilen dation, and the Spanish Health Research Fund (FIS). He has been paid for talks on behalf of Gilead Sciences, Merck Sharp and Dohme, Pfizer, Astellas Pharma, Hickma Pharmaceutica, and United Medical. Benito Almirante has received grant support from Gilead Sciences, Pfizer, and the Instituto de Salud Carlos III. He has received honoraria for talks on behalf of Gilead Sciences, Merck Sharp & Dohme, Pfizer, Astellas Pharma, and Novartis. Bele ´n Padilla has received honoraria for talks on behalf of Gilead Sciences, Merck Sharp & Dohme, Pfizer, Astellas Pharma and Novartis. Manuel Cuenca-Estrella has received grant support from Astellas Pharma, bioMerieux, Gilead Sciences, Merck Sharp & Dohme, Pfizer, Schering Plough, Soria Melguizo SA, Ferrer International, the European Union, the ALBAN program, the Spanish Agency for International Cooperation, the Spanish Ministry of Culture and Education, the Spanish Health Research Fund, the Instituto de Salud Carlos III (Spanish Ministry of Economy and Competitiveness), the Ra~a Foundamon Areces Foundation, and the Mutua Madrilen tion. He has been an advisor/consultant to the Panamerican Health Organization, Astellas Pharma, Gilead Sciences, Merck Sharp & Dohme, Pfizer, and Schering Plough. He has been paid for talks on behalf of Gilead Sciences, Merck Sharp & Dohme, Pfizer, Astellas Pharma, and Schering Plough. Jose ´ Marı´a Aguado has received grant support from Astellas Pharma, Gilead Sciences, Merck Sharp & Dohme, Pfizer, the Instituto de Salud Carlos III (Spanish Ministry of Economy and Competitiveness), and the Mutua ~a Foundation. He has been an advisor/consultant Madrilen to Astellas Pharma, Gilead Sciences, Merck Sharp & Dohme, and Pfizer. He has received honoraria for talks on behalf of Gilead Sciences, Merck Sharp & Dohme, Pfizer, and Astellas Pharma.

Acknowledgements Members of CANDIPOP Project: ~oz and Jesu
s Guinea (Hospital Bele ´n Padilla, Patricia Mun ~o
n, Madrid); Jose General Universitario Gregorio Maran ´
n Pan ~o Pardo, Julio Garcı´a-Rodrı´guez and Carlos Ramo Garcı´a Cerrada (Hospital Universitario La Paz, Madrid);
mez (Hospital Uni
s Fortu
n, Pilar Martı´n and Elia Go Jesu
n y Cajal, Madrid); Pablo Ryan and Carolina versitario Ramo Campelo (Hospital Infanta Leonor, Madrid); Ignacio de los Santos Gil and Ventura Buendı´a (Hospital Universitario La Princesa, Madrid); Beatriz Perez Gorricho and Mercedes ~o Jesu
s, Madrid); Alonso (Hospital Universitario del Nin Francisca Sanz Sanz and Jose ´ Marı´a Aguado (Hospital Universitario “12 de Octubre”, Madrid); Paloma Merino
lez Romo (Hospital Clı´nico San Carlos, and Fernando Gonza
n Madrid); Miguel Gorgolas and Ignacio Gadea (Fundacio Jime ´nez Dı´az, Madrid); Juan Emilio Losa and Alberto
n, Madrid); Antonio Delgado-Iribarren (Hospital de Alcorco
nchez Romero (HosRamos, Yolanda Romero and Isabel Sa pital Universitario Puerta de Hierro-Majadahonda, Madrid); Oscar Zaragoza and Manuel Cuenca-Estrella (Centro

9 Nacional de Microbiologı´a, Instituto de Salud Carlos III, ~o and Ana
s Rodriguez-Ban Majadahonda, Madrid); Jesu Isabel Suarez (Hospital Universitario Virgen Macarena, Sevilla); Ana Loza, Ana Isabel Aller Garcı´a and Estrella Martı´n-Mazuelos (Hospital Universitario Virgen de Valme,
n and Jose Sevilla); Maite Ruiz Pe ´rez de Pipao ´ Garnacho (Hospital Universitario Virgen del Rocı´o, Sevilla); Carlos
n Sagrado Corazo
n, Sevilla); Mo
nica Ortiz (Hospital Quiro
vez and Fernando L. Maroto (Hospital San Juan de Dios Cha
n de Aljarafe, Sevilla); Miguel Salavert and Javier Pema (Hospital Universitari La Fe, Valencia); Jose ´ Blanquer and David Navarro (Hospital Clı´nico Universitario de Valencia); Juan Jose ´ Camarena and Rafael Zaragoza (Hospital Universitario Dr. Peset, Valencia); Vicente Abril and Con
n Gimeno (Consorcio Hospital General Universitario cepcio
ez and Guillermo Ezpeleta (Hosde Valencia); Silvia Herna pital de Basurto, Bilbao); Elena Bereciartua, Jose ´ L.
ndez Almaraz and Miguel Montejo (Hospital UniversiHerna tario de Cruces, Bilbao); Rosa Ana Rivas and Rafael Ayarza (Hospital de Galdakano, Bilbao); Ana Ma Planes, Isabel Ruiz Camps and Benito Almirante (Hospital Universitari Vall d’Hebron, Barcelona); Jose ´ Mensa and Manel Almela (Hos Gurgui and Ferran pital Clı´nic-IDIBAPS, Barcelona); Merce
nchez-Reus (Hospital Universitari de Sant Pau i Santa Sa Creu, Barcelona); Joaquin Martinez-Montauti and Montserrat Sierra (Hospital de Barcelona, Barcelona);  Go
mez (Hospital Juan Pablo Horcajada, Luisa Sorli and Julia del Mar, Barcelona); Amadeu Gene ´ and Mireia Urrea (Hospital Sant Joan de De ´u, Esplugues de Llobregat, Barcelona). Study collaborators: Maricela Valerio, Ana Dı´az-Martı´n, Francesc Puchades and Alessandra Mularoni.

Appendix A. Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.jinf.2015.05.009.

References 1. Almirante B, Rodrı´guez D, Park BJ, Cuenca-Estrella M, Planes AM, Almela M, et al. Epidemiology and predictors of mortality in cases of Candida bloodstream infection: results from population-based surveillance, Barcelona, Spain, from 2002 to 2003. J Clin Microbiol 2005;43:1829e35.
n J, Canto
n E, Quindo
s G, Eraso E, Alcoba J, Guinea J, 2. Pema et al. Epidemiology, species distribution and in vitro antifungal susceptibility of fungaemia in a Spanish multicentre prospective survey. J Antimicrob Chemother 2012;67:1181e7. 3. Colombo AL, Nucci M, Park BJ, Noue ´r SA, Arthington-Skaggs B, da Matta DA, et al. Epidemiology of candidemia in Brazil: a nationwide sentinel surveillance of candidemia in eleven medical centers. J Clin Microbiol 2006;44:2816e23.
n DE, Alvarado-Matute T, Colombo AL, Cornejo4. Corzo-Leo
rez P, Corte Jua ´s J, Echevarrı´a JI, et al. Surveillance of Candida spp bloodstream infections: epidemiological trends and risk factors of death in two Mexican tertiary care hospitals. PLoS One 2014;9:e97325. 5. Negri M, Silva S, Henriques M, Oliveira R. Insights into Candida tropicalis nosocomial infections and virulence factors. Eur J Clin Microbiol Infect Dis 2012;31:1399e412. 6. Rodrigues AG, Pina-Vaz C, Costa-de-Oliveira S, Tavares C. Expression of plasma coagulase among pathogenic Candida species. J Clin Microbiol 2003;41:5792e3.


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009

10 7. Deorukhkar SC, Saini S, Mathew S. Virulence factors contributing to pathogenicity of Candida tropicalis and its antifungal susceptibility profile. Int J Microbiol 2014;2014: 456878. 8. Moralez AT, Franc ¸a EJ, Furlaneto-Maia L, Quesada RM, Furlaneto MC. Phenotypic switching in Candida tropicalis: association with modification of putative virulence attributes and antifungal drug sensitivity. Med Mycol 2014;52:106e14. 9. Leung AY, Chim CS, Ho PL, Cheng VC, Yuen KY, Lie AK, et al. Candida tropicalis fungaemia in adult patients with haematological malignancies: clinical features and risk factors. J Hosp Infect 2002;50:316e9. 10. Kontoyiannis DP, Vaziri I, Hanna HA, Boktour M, Thornby J, Hachem R, et al. Risk factors for Candida tropicalis fungemia in patients with cancer. Clin Infect Dis 2001;33:1676e81. ~ oz P, Giannella M, Fanciulli C, Guinea J, Valerio M, Rojas L, 11. Mun et al. Candida tropicalis fungaemia: incidence, risk factors and mortality in a general hospital. Clin Microbiol Infect 2011;17: 1538e45. 12. Andes DR, Safdar N, Baddley JW, Playford G, Reboli AC, Rex JH, et al. Impact of treatment strategy on outcomes in patients with candidemia and other forms of invasive candidiasis: a patient-level quantitative review of randomized trials. Clin Infect Dis 2012;54:1110e22. 13. Nucci M, Colombo AL. Candidemia due to Candida tropicalis: clinical, epidemiologic, and microbiologic characteristics of 188 episodes occurring in tertiary care hospitals. Diagn Microbiol Infect Dis 2007;58:77e82. 14. Pappas PG, Kauffman CA, Andes D, Benjamin Jr DK, Calandra TF, Edwards Jr JE, et al. Clinical practice guidelines for the management of candidiasis: 2009 update by the Infectious Diseases Society of America. Clin Infect Dis 2009;48: 503e35. 15. Cornely OA, Bassetti M, Calandra T, Garbino J, Kullberg BJ, Lortholary O, et al. ESCMID* guideline for the diagnosis and management of Candida diseases 2012: non-neutropenic adult patients. Clin Microbiol Infect 2012;18:19e37. 16. Myoken Y, Kyo T, Fujihara M, Sugata T, Mikami Y. Clinical significance of breakthrough fungemia caused by azole-resistant Candida tropicalis in patients with hematologic malignancies. Haematologica 2004;89:378e80.  s T, Renier G, Chabasse D, 17. Vandeputte P, Larcher G, Berge Bouchara JP. Mechanisms of azole resistance in a clinical isolate of Candida tropicalis. Antimicrob Agents Chemother 2005;49:4608e15.
n J, Zaragoza R, Garnacho-Montero J, 18. Puig-Asensio M, Pema Martı´n-Mazuelos E, Cuenca-Estrella M, et al. Epidemiology and predictive factors for early and late mortality in Candida bloodstream infections: a population-based surveillance in Spain. Clin Microbiol Infect 2014;20:O245e54.
Escribano P, Martı´n-Mazuelos E, 19. Guinea J, Zaragoza O,
n J, Sa
nchez-Reus F, et al. Molecular identification and Pema antifungal susceptibility of yeast isolates causing fungemia collected in a population-based study in Spain from 2010 to 2011. Antimicrob Agents Chemother 2014;58:1529e37.


ndez-Ruiz et al. M. Ferna 20. Segal BH, Herbrecht R, Stevens DA, Ostrosky-Zeichner L, Sobel J, Viscoli C, et al. Defining responses to therapy and study outcomes in clinical trials of invasive fungal diseases: Mycoses Study Group and European Organization for Research and Treatment of Cancer consensus criteria. Clin Infect Dis 2008; 47:674e83. 21. Mermel LA, Allon M, Bouza E, Craven DE, Flynn P, O’Grady NP, et al. Clinical practice guidelines for the diagnosis and management of intravascular catheter-related infection: 2009 Update by the Infectious Diseases Society of America. Clin Infect Dis 2009;49:1e45. 22. Paterson DL, Ko WC, Von Gottberg A, Mohapatra S, Casellas JM, Goossens H, et al. Antibiotic therapy for Klebsiella pneumoniae bacteraemia: implications of production of extendedspectrum b-lactamases. Clin Infect Dis 2004;39:31e7. 23. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373e83. 24. EUCAST definitive document EDef 7.1: method for the determination of broth dilution MICs of antifungal agents for fermentative yeasts. Clin Microbiol Infect 2008;14:398e405. 25. White T, Bruns T, Lee S, Taylor J. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis MA, Gefland DH, Sninsky JJ, White TJ, editors. PCR protocols: a guide to methods and applications. San Diego, CA: Academic Press; 1990. p. 315e22. 26. Clinical and Laboratory Standards Institute (CLSI). Reference method for broth dilution antifungal susceptibility testing of yeasts. Approved standard (third edition). CLSI document M27-A3. Wayne, PA: Clinical and Laboratory Standards Institute; 2008. 27. Forastiero A, Mesa-Arango AC, Alastruey-Izquierdo A, AlcazarFuoli L, Bernal-Martinez L, Pelaez T, et al. Candida tropicalis antifungal cross-resistance is related to different azole target (Erg11p) modifications. Antimicrob Agents Chemother 2013; 57:4769e81. 28. Cisterna R, Ezpeleta G, Telleria OSpanish Candidemia Surveillance Group. Nationwide sentinel surveillance of bloodstream Candida infections in 40 tertiary care hospitals in Spain. J Clin Microbiol 2010;48:4200e6. 29. Arendrup MC, Dzajic E, Jensen RH, Johansen HK, Kjaeldgaard P, Knudsen JD, et al. Epidemiological changes with potential implication for antifungal prescription recommendations for fungaemia: data from a nationwide fungaemia surveillance programme. Clin Microbiol Infect 2013;19: E343e53. 30. Stevens DA, Espiritu M, Parmar R. Paradoxical effect of caspofungin: reduced activity against Candida albicans at high drug concentrations. Antimicrob Agents Chemother 2004;48: 3407e11. 31. Eagle H. A paradoxical zone phenomenon in the bactericidal action of penicillin in vitro. Science 1948;107:44e5. 32. Clemons KV, Espiritu M, Parmar R, Stevens DA. Assessment of the paradoxical effect of caspofungin in therapy of candidiasis. Antimicrob Agents Chemother 2006;50:1293e7.


ndez-Ruiz M, et al., Candida tropicalis bloodstream infection: Incidence, risk factors and Please cite this article in press as: Ferna outcome in a population-based surveillance, J Infect (2015), http://dx.doi.org/10.1016/j.jinf.2015.05.009