Clin Microbiol Infect17: 259–263
Candida krusei has been documented as an emerging pathogen causing nosocomial outbreaks. The consecutive isolation of C. krusei strains in three patients admitted to the same hospital department within 2 months lead us to consider the possibility of an outbreak. Additionally, C. krusei isolates were collected from the room surfaces, whereas another isolate had been recovered from the blood of one patient 2 years before. HinfI DNA restriction endonuclease-based analysis of all C. krusei isolates was performed and restriction profiles were compared. Surprisingly, isolates from different patients were unrelated, whereas isolates from biological products of the same patient showed indistinguishable HinfI restriction patterns and were similar to those obtained from the surrounding environment of the respective patients. The study approach revealed the endogenous origin of the C. krusei infectious episodes observed and demonstrated that, subsequent to colonizing a patient, C. krusei can be involved in infectious episodes distant in time. The hypothesis of an outbreak was excluded, although we believe that the methodology employed in the present study represents a valuable tool for diagnostic and epidemiological surveys.
In the last two decades, invasive fungal infections in hospitalized patients have increased significantly worldwide. According to data obtained from the USA and Europe, Candida species represent, respectively, the fourth and sixth most frequent cause of invasive healthcare-related infections [1,2], accounting for 8–15% of all episodes of sepsis acquired in hospital settings . Inherent to these types of infection are the extremely high morbidity and mortality rates, particularly among immunocompromised patients [4,5].
Fluconazole is one of the antifungal agents mostly used in both prophylactic and therapeutic protocols. Fluconazole prophylaxis has been associated with a decrease in the prevalence of Candida species such as Candida tropicalis and Candida albicans, and to an increase in that of Candida krusei and Candida glabrata . C. krusei presents intrinsic resistance to fluconazole and, to some extent, reduced susceptibility to amphotericin B .
Infectious outbreaks in hospitals, especially in intensive care units, represent a serious health problem and are mainly due to Candida lusitaniae , C. albicans , Candida parapsilosis , and C. krusei [10,11]. Many factors may account for their occurrence (e.g. barrier loss, lack of proper infection control measures by health care workers when managing patients, resistance to prescribed antifungal drugs, as well as insufficient drug levels).
Molecular methods represent a powerful tool to clarify transmission pathways in health care facilities (i.e. to investigate the occurrence of possible outbreaks). Techniques such as karyotyping, restriction fragment length polymorphism analysis by pulsed-field gel electrophoresis, southern blot hybridization, PCR fingerprinting and randomly amplified polymorphic DNA fingerprinting have been extensively used for Candida typing [12–16]. Restriction endonuclease analysis (REA) has been described in the last decade as a valuable tool for Candida spp. characterization. REA of the mitochondrial DNA (mtDNA) was first applied in the biotechnology industry in order to characterize yeast strains used for wine fermentation [17,18] and, more recently, to discriminate between Candida clinical strains [19–22]. The data obtained demonstrate the relevance of using molecular genetic methods in many different areas, including taxonomic, ecological and clinical surveys.
Recently, we were challenged by a hypothetical outbreak as a result of C. krusei in the neutropaenia unit of the Haematology Department of Hospital S. João, Porto, Portugal. Within a short period of time, several C. krusei isolates were cultured from biological products of three patients. In addition, C. krusei was found in the surrounding environment of the patients. All isolates were compared using mtDNA REA, which is a convenient and powerful tool that allows valid comparisons between isolates of the same yeast species.
Patients and Methods
Patient A was a 41-year-old male with acute lymphoblastic leukaemia (T/NK) diagnosed in July 2006. In August 2008, the patient was initially treated with amoxicillin and ciprofloxacin because of undetermined fever. When symptoms remained unchanged, a myelogram was performed; a first relapse of the haematological disease was diagnosed. He was admitted to the neutropaenia unit for salvage intensive chemotherapy with fludarabine, idarubicin and ara-C, followed by growth factor (G-CSF), administered through a central venous catheter. Upon the onset of the aplastic period (28 August), he received antimicrobial treatment for 2 weeks [ciprofloxacin, amoxicillin-clavulanic acid, imipenem, acyclovir and fluconazole (200 mg/day)]. At the time of fungaemia detection, he displayed the following haematologic parameters: white blood cells 0.03 × 109 cells/L, haemoglobin level 8.5 g/dL and blood platelets 11 × 109 cells/L. As soon as C. krusei was identified, the patient was started on caspofungin for 12 days without improvement; the treatment was then changed to amphotericin B.
This patient had previously developed a fungaemia episode as a result of C. krusei during chemotherapy in 2006 and a corresponding isolate had been stored at −70°C.
The patient remained in room number 1 from August to October 2008.
Patient B was a 53-year-old male with non-Hodgkin’s lymphoma, who was admitted to room number 2 for 3 weeks in September 2008. He was administered cyclophosphamide, doxorubicin, vincristine, and prednisone, followed by G-CSF. On 12 September 2008, he was started on caspofungin but, as a result of sustained fever on day 5, treatment was changed to voriconazole, completing 14 days of antifungal therapy. He never received fluconazole. At that fungaemia episode, he displayed the following haematologic parameters: white blood cells 2.41 × 109 cells/L, haemoglobin 9.3 g/dL and blood platelets 73 × 109 cells/L.
Patient C was a 60-year-old male with acute myeloid leukaemia secondary to myelodisplastic syndrome; he was admitted to room number 2 for 3 weeks in October 2008, subsequent to patient B. He received Ara-C, etoposide and doxorubicin for 5 days, followed by another 5 days of Ara-C; he was also prescribed amoxicillin, ciprofloxacin and allopurinol. In addition, prophylactic treatment included imipenem, vancomycin, fluconazole and acyclovir from 29 September to 20 October 2008. On that latest date, he displayed the following haematologic parameters: white blood cells 0.6 × 109 cells/L, haemoglobin 9.8 g/dL and blood platelets 24 × 109 cells/L.
Patient D served as a typing control; one single C. krusei isolate was recovered from the patient’s bronchial secretions in September 2008 upon his admission to the Internal Medicine Department.
Eighteen C. krusei isolates were collected from different clinical specimens (blood, urine, stools and bronchial secretions) of the above mentioned patients admitted to the neutropaenia unit within a 2-month period. Additionally, a previous C. krusei isolate (2006) from a blood culture of patient A was included in the study, as well as a C. krusei isolate obtained from the control patient D. All isolates were identified using the automatic system Vitek2 YBC identification cards (BioMérieux, Paris, France), stored at −70°C in Brain Heart infusion (Merck KGaA, Darmstadt, Germany) with 10% glycerol and sub-cultured twice on Sabouraud agar (Merck KGaA) to ensure purity prior to experimental assays.
Several environmental samples were collected from the patients’ rooms (1 and 2) using Sabouraud agar contact plates (Merck KGaA); two C. krusei isolates were cultured: one from the bedside table of patient A (room 1) and another from the bed of patient C (room 2). The isolates were characterized and stored as described above for clinical samples. Air samples were collected by filtration with a MAS-100 Eco instrument (Merck Eurolab, Dietlikon, Switzerland), containing Sabouraud agar plates (Merck KGaA); no C. krusei isolates were obtained.
Antifungal susceptibility testing
Voriconazole (Pfizer, New York, NY, USA), posaconazole (Schering-Plough; Kenilworth, NJ, USA), amphotericin B (Bristol Meyers Squibb, New York, NY, USA), caspofungin (Merck, Rahway, NJ, USA) and anidulafungin (Pfizer) stock solutions were prepared according to CLSI protocols (M27-A3)  and maintained at −70°C until use. Minimal inhibitory concentration (MIC) of each antifungal drug was determined according to CLSI protocol M27-A3 .
Total genomic DNA extraction
Yeast cells were cultured overnight at 30°C in 10 mL of YPD liquid medium, with continuous orbital shaking at 180 r.p.m., and subsequently collection by centrifugation. Total DNA was extracted using phenol:chloroform:isoamyl alcohol 25:24:1, precipitated using 100% ice-cold ethanol and redissolved in 200 μL of TE buffer. The DNA was treated with 20 μg of RNase (Applichem, Darmstadt, Germany), incubated at 37°C for 30 min to 1 h. For final precipitation, 20 μL of 4 M ammonium acetate, pH 4.8 (Sigma-Aldrich, Munich, Germany) and 600 μL of ice-cold 100% ethanol (Applichem) were added and samples were incubated overnight at −20°C. The DNA was re-dissolved in TE buffer 1x, assessed in a biophotometer 6131 (Eppendorf, Hamburg, Germany) and adjusted to a final concentration of 2.0–2.5 μg/μL. To assay DNA integrity, approximately 3–5 μg of DNA was run in agarose gel (1%, w/v) (Sigma-Aldrich) in TBE buffer 1x and stained with ethidium bromide (0.5 mg/mL) (Applichem). DNA samples were stored at −20°C for subsequent use.
REA of mt DNA
For each sample, a reaction mixture was prepared containing 1x HinfI enzyme reaction buffer (Metabion, Planegg, Germany), 1 μg/μL RNase, 0.5 U/μL HinfI restriction enzyme (Metabion), approximately 25–30 μg of total DNA and DNAse-RNAse free water up to 20 μL final volume; reaction tubes were incubated overnight at 37°C. Restriction was ended upon HinfI inactivation by incubating for 20 min at 80°C. The total reaction mixture was run on a 1% agarose gel (20 cm × 24 cm) at 120 mV for 3–5 h, stained with ethidium bromide (0.5 mg/mL) and the DNA visualized under UV light.
Restriction patterns were analyzed using the UVIDOC 12.6 software for Windows (Topac Inc., Cohasset, MA, USA) and the resulting groups of strains were compared.
Results and Discussion
There have been an increasing numbers of reports describing non-albicans Candida hospital outbreaks. Given that C. krusei is not the main pathogen causing nosocomial infections, the detection of simultaneous episodes at the neutropaenia unit of our hospital lead us to consider the possibility of an outbreak.
The C. krusei isolates were all susceptible to all the antifungals assayed (fluconazole was not tested because C. krusei presents intrinsic resistance to this agent). Variations in the antifungal MICs for different isolates were not significant. The MICs of amphotericin B were from 0.06 to 1 mg/L; for caspofungin from 0.125 to 1 mg/L; for anidulafungin 0.06 mg/L; for voriconazole from 0.25 to 2 mg/L; and for posaconazole from 0.03 to 0.5 mg/L.
The routine biochemical identification protocols or antifungal susceptibility profiles are usually not sufficient to either corroborate or exclude an outbreak hypothesis. REA for Candida species was first described by Scherer and Stevens  who considered this method to be an extremely valuable tool for epidemiological studies. Fujita et al. described HinfI restriction patterns as exhibiting a superior discriminatory power among distinct Candida isolates compared to patterns obtained with other enzymes such as EcoRI or MspI. Additionally, Sancak et al. established a correspondence of almost 100% between the results obtained with HinfI restriction endonuclease-based analysis and PCR methodologies. In the present study, a total of 22 C. krusei isolates (20 clinical, two environmental) were analyzed using REA. A high number of C. krusei isolates was obtained from different biological products of patient A in distinct periods of time. All of them showed the same restriction pattern, including the isolate recovered in 2006 from a blood culture (ABC1) (Fig. 1, lanes ABC1, ABC2 and ACVCII), indicating that they are the same strain. Most certainly, this patient harbors a reservoir of C. krusei and was colonized throughout a long period of time, as described similarly for Acinetobacter  and Pseudomonas . This is the first report, to our knowledge, describing a long-lasting colonization by C. krusei. These results have implications in terms of prophylactic measures (i.e. fluconazole is not recommended in a patient with previous isolation of C. krusei). Other antifungals, such voriconazole or amphotericin B, are more likely to be efficient in these cases [27,28].
The C. krusei isolates from each patient yielded distinct HinfI restriction patterns, suggesting that the isolates were different strains (Fig. 1, lanes ABC2, BBS, CS), at the same time discarding the hypothesis of an outbreak in the neutropaenia unit where the patients were admitted. The C. krusei strain isolated from patient D showed a pattern distinct from those isolated from the remaining patients (Fig. 1, lane DBS), as expected.
The two environmental C. krusei isolates collected from the surfaces of the rooms where patient A (AE9) (room 1) and patients B and C (CE) (room 2) had stayed were different, as depicted in Fig. 1. This excludes the possibility of different C. krusei strains being transmitted as a result of patient handling by health care workers. However, both the environmental and the clinical C. krusei isolates associated with the same patient displayed an undistinguishable restriction pattern, as shown in Fig. 1 (lanes ACVC11 and AE9 vs. CS and CE), suggesting a putative environmental reservoir and the possibility of subsequent transmission from it to other patients. Our results emphasize the need to enhance preventive infection control measures, both when handling the patients directly and when cleaning patients′ facilities, particularly those admitting neutropenic patients. Indeed, a study by Berrouane et al.  described the relatedness among C. krusei clinical isolates obtained from different biological products from both patients and health care workers.
As described in the present study, we were able to exclude the hypothesis of an outbreak occurring in the neutropaenia unit throughout the time period considered, supporting the usefulness and suitability of the REA methodology. Moreover, the present study provided very useful information concerning C. krusei reservoirs existing in patients and their surrounding environment.
The authors would like to thank I. Santos for the excellent technical support. Part of the results were presented at the 19th European Congress of Clinical Microbiology and Infectious Diseases held in Helsinki, Finland, 2009.
The authors declare that there is no source of funding and no potential conflicts of interest.