Prevalence of Candida dubliniensis Fungemia at a Large Teaching Hospital

Reports of cases of systemic disease due to Candida dubliniensis are increasing, given the availability of new diagnostic tests that can accurately differentiate this pathogen from the phenotypically similar Candida albicans [1–3]. The peptide nucleic acid fluorescent in situ hybridization (PNA FISH) test uses fluorescent-labeled PNA probes that target the highly specific rRNA sequences of C. albicans and C. dubliniensis. A PCR-based method using C. albicans—specific primers derived from the PHR1 gene sequence was successfully used to differentiate between C. albicans and C. dubliniensis isolates with 100% sensitivity and specificity [4]. Analyses of the C. dubliniensis genome have shown that C. dubliniensis encodes a range of virulence genes similar to those encoded by C. albicans, such as the agglutinin-like sequence and secreted aspartyl proteinase families of genes involved in adherence and secretion of proteases, respectively [2]. In addition, considerable attention has been given to C. dubliniensis, because of its ability to develop resistance to fluconazole (the most commonly used drug for treatment of HIV-infected patients with or without AIDS), owing to a readily inducible fluconazole resistance mechanism [1, 2, 5, 6]. Between 1999 and 2003, a total of 19 cases of systemic and blood infections with C. dubliniensis were reported in patients with a variety of underlying diseases. However, only 1 case of C. dubliniensis fungemia in an HIV-infected patient has been reported [1, 7–10]. Here, we describe cases of C. dubliniensis fungemia in patients at the University of Maryland Medical System (Baltimore) determined on the basis of recovery of C. dubliniensis from blood cultures between March and November 2004. Patients and methods. After institutional review board approval was obtained, clinical data were collected for 6 hospitalized patients with symptoms of septicemia. In addition, data were collected for 81 patients with blood cultures positive for yeast during the same period. Yeast isolates recovered from blood specimens were speciated with Vitek II (bioMérieux). However, 7 germ tube—positive, chlamydospore-producing isolates recovered from 6 patients (1 of whom had a second blood culture with a positive result) were not identified as C. albicans and were therefore presumptively considered to be C. dubliniensis. The 7 isolates were tested on ChromAgar Candida (Becton Dickinson) for the ability to grow at 45°C and for colony color. The PNA FISH test (AdvanDx) was performed on microscope slides on smears directly obtained from Gram stain—positive yeast blood cultures, according to the manufacturer's instructions. Briefly, after fixation, smears of culture specimens were incubated with PNA probe, washed, air dried, and examined with a fluorescent microscope. Positive results were indicated by a fluorescent green color. Amplification was performed on genomic DNA from C. albicans ATCC 90029 and LHG1095 strains and from C. dubliniensis type strain CD36 with C. albicans—specific PHR1 primers using Promega PCR Master Mix (Promega). Products were separated on an agarose gel (0.8% wt/vol) and stained with ethidium bromide. Fluconazole susceptibility tests were performed and interpreted according to the guidelines outlined by the NCCLS broth microdilution protocol [11]. Results. Patient demographic and clinical characteristics and information about the antifungal therapy they received are summarized in table 1. The 2 HIV-infected patients were receiving fluconazole before admission to the hospital for oral candidiasis; one had methicillin-susceptible Staphylococcus aureus endocarditis and did not survive, and the other was successfully treated with fluconazole. A patient with methicillin-resistant S. aureus endocarditis did not respond to fluconazole therapy and died. One patient was successfully treated with fluconazole, and 3 required caspofungin therapy. Characteristics of patients with Candida dubliniensis recovered from the bloodstream. A total of 88 Candida isolates recovered from blood cultures were identified. Of these, 35 were C. albicans, 19 were C. parapsilosis, 19 were C. glabrata, 7 were C. dubliniensis, 5 were C. tropicalis, 2 were C. krusei, and 1 was C. lusitaniae. Of the 88 cases of Candida bloodstream infection, 6 were due to C. dubliniensis, with a recovery rate of 7%. The 7 C. dubliniensis isolates produced germ tubes and abundant chlamydospores, grew as dark-green colonies on ChromAgar Candida, failed to grow at 45°C, did not hybridize with the PNA FISH probe for C. albicans, but did hybridize with a PNA FISH probe for C. dubliniensis (figure 1). PCR amplification of DNA from C. albicans strains resulted in a fragment with the predicted size of 1.6 kb, but no amplimers were detected by PCR amplication of DNA from the 7 C. dubliniensis isolates (figure 2). All 7 C. dubliniensis clinical isolates were susceptible to fluconazole, with MICs of 0.125–0.25 µg/mL. Results of peptide nucleic acid fluorescent in situ hybridization analysis for detection of Candida dubliniensis using a C. dubliniensis—specific probe (A) and a Candida albicans—specific probe (B). Amplification products obtained after PCR using a primer set for the Candida albicans—specific PHR1 gene and a Candida template. Lane 1, molecular weight standard; lanes 2 and 3, C. albicans ATCC 90029 and LGH 1095 strains, respectively; lane 4, Candida dubliniensis type CD36; lanes 5–11, seven C. dubliniensis isolates from 6 patients with C. dubliniensis fungemia. Discussion.C. dubliniensis, initially considered to be an opportunistic oral pathogen in HIV-infected individuals, is emerging as a pathogen than can cause invasive disease in patients who have a variety of clinical conditions. Here, we report 6 cases of C. dubliniensis fungemia in which 7 isolates were recovered; 2 isolates were recovered 1 week apart from a patient receiving fluconazole. All patients were hospitalized at the time blood was obtained for culture, and all had at least 1 symptom of septicemia. Two of the patients died, and 1 remained hospitalized at the time of writing. Isolates were confirmed to be C. dubliniensis after hybridization with the C. dubliniensis—specific PNA FISH probe. All 7 isolates were susceptible to fluconazole. During the same 9-month period in which the C. dubliniensis blood isolates were recovered, 81 yeast fungemias were reported in the clinical laboratory. The reported incidence of fungemia due to C. dubliniensis is not accurately known, because not all clinical laboratories distinguish this species from the morphologically similar C. albicans. It is suspected that many cases of systemic infection by C. dubliniensis go undetected or are erroneously reported as C. albicans infection and, therefore, that accurate species identification using molecular-based methods, such as PNA FISH or PCR, can aid in determining the true incidence of C. dubliniensis infection. C. dubliniensis possesses many of the virulence characteristics associated with C. albicans, including the ability to adhere to surfaces. Interestingly, in a study by Zepelin et al. [12], the inclusion of fluconazole in the assay medium resulted in an increased adherence of C. dubliniensis to Vero cells, which could explain the high recovery rate of this species among HIV-infected patients with or without AIDS who are receiving fluconazole therapy. Two large-scale, multicenter surveillance studies on the incidence of Candida bloodstream infections were conducted in the United States during 1995–2000. Results of the first study showed that 2.6% of the isolates initially identified as C. albicans were actually C. dubliniensis, whereas the second study (which analyzed the distribution of Candida species) revealed a C. dubliniensis recovery rate of 0.9% [13, 14]. In a retrospective study done at our institution during a 2-year period between 1998 and 2000, screening of >1200 isolates identified as C. albicans that were recovered from patients with a variety of immunocompromising conditions, including HIV infection, resulted in a 1.2% recovery rate for C. dubliniensis [5]. In the present study, of 8 cases of Candida bloodstream infection, 6 were due to C. dubliniensis, with a 7% recovery rate that surpassed the recovery rates for C. tropicalis (5.7%), C. krusei (2%), and C. lusitaniae (1%). The current data suggest that, because of a lack of awareness of the distinction between C. dubliniensis and C. albicans, the prevalence of C. dubliniensis has been underreported during the past few years. To our knowledge, 19 cases of fungemia and systemic infection due to C. dubliniensis have been reported in the literature [1, 7–10]. Only 1 case was in an HIV-infected individual, and the majority were in patients with a variety of other immunocompromising conditions, including various malignancies, neuroblastoma, disseminated tuberculosis, hepatosplenomegaly with acute renal failure, systemic lupus erythematous with chronic renal failure, liver disease, solid organ transplant, and multiple myeloma with C. dubliniensis recovered from synovial fluid [1, 7–10]. Recently, a relatively high rate of colonization was reported in a population of patients with cystic fibrosis and in a diverse group of patients without known immunocompromising conditions, suggesting that this species is associated with the same spectrum of disease as that for C. albicans and that HIV-infected individuals are not the only population susceptible to C. dubliniensis infection [15]. Our patient population had a variety of underlying conditions, with 2 patients infected with HIV, for a total of 3 reported cases of C. dubliniensis fungemia in patients with HIV infection to date. Concerns have been expressed about the development of resistance to fluconazole in C. dubliniensis, especially among isolates in HIV-infected individuals receiving fluconazole therapy [2, 5]. Resistance to fluconazole has been shown in clinical isolates and in an in vitro study, in which a C. dubliniensis isolate rapidly developed stable resistance to fluconazole after exposure to the drug. On the other hand, attempts to produce a fluconazole-resistant C. albicans strain were not successful [6]. This potential for rapid development of resistance makes it crucial to identify C. dubliniensis as the causative agent in cases of fungemia and to perform antifungal susceptibility testing on each isolate to confirm azole susceptibility. The demonstration that C. dubliniensis has the potential to cause bloodstream infection provides crucial information about the clinical importance and epidemiologic characteristics of this interesting species. In light of these increasing reports about the pathogenic significance of C. dubliniensis, accurate identification of this species in the clinical laboratory is warranted. This task has become feasible and practical with the recent emergence of simple and highly specific tests, such as the PNA FISH test, which can be utilized in large-scale, population-based surveillance for bloodstream infection, to estimate more accurately the incidence of candidemia due to C. dubliniensis. Financial support. National Institute of Dental and Craniofacial Research, National Institutes of Health (grant DE 14424). Potential conflicts of interest. All authors: no conflicts.