Development of echinocandin resistance in Candida krusei isolates following exposure to micafungin and caspofungin in a BM transplant unit

Hematopoietic SCT (HSCT) is associated with a high risk of invasive fungal infection. Fluconazole prophylaxis significantly reduces the frequency of candidiasis.1 Echinocandins are antifungal agents inhibiting 1,3-β-D-glucan synthesis, an essential component of the cell wall. Micafungin at a dose of 50 mg/day is an option for prophylaxis during the neutropenic phase after HSCT.2 Despite the increasing use of echinocandins, resistance remains uncommon in most Candida species, particularly in C. krusei.3, 4 Mutations in the Fks1 gene (encoding 1,3-β-D-glucan synthase) are correlated with reduced susceptibility.5 We report here C. krusei breakthrough fungemia following exposure to prophylactic micafungin in HSC transplanted patients. In two patients, C. krusei with novel mutations in the Fks1 gene and reduced susceptibility to echinocandins emerged in digestive tract surveillance isolates. While conducting a study to evaluate micafungin prophylaxis in the HSCT unit of the University Hospital of Saint Etienne, France, we compared a historical cohort of 75 fluconazole recipients (400 mg/day) treated before 1 April 2011 to 37 micafungin recipients (50 mg/day) who underwent HSCT after this date. No candidemia was diagnosed in the fluconazole group whereas three C. krusei fungemia were observed among 37 patients exposed to micafungin prophylaxis (P=0.034). Clinical characteristics of subjects are presented in Table 1. Both treatment groups were similar with regard to age, transplant type, stem cell source, underlying disease, microbiological surveillance and antibiotic protocol. The first C. krusei fungemia (patient no. 1) was observed on 23 July 2012 in a 38-year-old man becoming febrile on day 5 after pheno-identical HSCT for CLL, despite micafungin prophylaxis since 14 days. Concurrently, fecal cultures were positive for C. krusei. Antibiotic therapy with imipenem-cilastatine was started. Liposomal amphotericin B (LAmB) was added on 24 July and central venous catheter removed 3 days later. A thoracic computed tomography revealed nodular pneumonia. Culture of bronchoalveolar lavage on 2 August yielded C. krusei. The patient clinically improved concurrently with the engraftment (3 August). Two weeks later, numerous isolates of C. krusei were found in the digestive tract. These isolates were susceptible to echinocandins. The patient was discharged from hospital on 17 August and LAmB was maintained for 60 days because of corticosteroïds introduced for aGVHD. This patient is currently devoid of active infection and in hematological CR. The second case (patient no. 2) occurred in a 56-year-old woman who developed candidemia on day 11 after pheno-identical HSCT for multiple myeloma. Fever began on 23 September 2012 and a broad-spectrum empirical antibiotic therapy was started. Blood cultures became positive for C. krusei on 2 October. After 20 days of micafungin prophylactic treatment, antifungal therapy was switched to caspofungin (70 mg/day) on 4 October. Neutrophil engrafment was achieved on 5 October. The next day, the patient was admitted to the intensive care unit (ICU) for bilateral pneumonia with respiratory distress requiring noninvasive ventilation. The patient was discharged from the ICU on 30 October. Six weeks later, on 12 November, C. krusei isolates with decreased susceptibility to echinocandins were identified in the digestive tract. Caspofungin was changed to LAmB. On 29 November, the patient finally died of respiratory distress escalation with septic presentation. The third fungemia (patient no. 3) occurred in a 48-year-old man 6 days after cord blood transplantation for CLL (18 October 2012) and after 15 days of micafungin prophylaxis. The patient had been receiving empirical antibiotic therapy since 11 October. Two days before the candidemia, C. krusei was isolated from digestive tract cultures. Micafungin was stopped, LAmB (3 mg/kg/day) administered (18 October) and central venous catheter removed. Fever and candidemia persisted for 7 days. LAmB was changed to caspofungin (70 mg/day) on 26 October, because of altered renal function. Two weeks later (12 November), whereas caspofungin was maintained, C. krusei with decreased susceptibility to echinocandins was isolated in cultures from the mouth. Because of immunosuppression, antifungal therapy was maintained and caspofungin changed to LAmB. The patient was discharged from hospital on 27 November and is still alive and in CR. Isolates were routinely tested for susceptibility to antifungal drugs with the Etest method (bioMerieux, Craponne, France). For isolates recovered from patients no. 2 and 3, in vitro susceptibility was confirmed by a microdilution technique following the procedure proposed by the Antifungal Susceptibility Testing Subcommittee of EUCAST6 (AFST-EUCAST) modified by using Antibiotic Medium 3 for caspofungin and micafungin.7 Primers and PCR conditions previously described were used to amplify and sequence HS1 and HS2 regions of the Fks1 gene of C. krusei isolates.7 Genotypes of the isolates were determined according to the multilocus sequence typing (MLST) method.8 Antifungal susceptibility results (minimal inhibitory concentrations, MICs) and mutations for patients no. 2 and 3 are presented in Table 2. For each patient, two later colonizing isolates recovered from digestive tract had two unknown missense mutation (S659P heterozygous and S659F homozygous, respectively) localized in the HS1 region.9 These mutations were associated with increased caspofungin MICs in the corresponding isolates. Isolates of a given patient shared similar MLST profiles suggesting that they were genetically linked. This paper describes two phenomena: (a) C. krusei breakthrough fungemia during micafungin prophylaxis and (b) emergence of colonizing isogenic isolates harboring hot-spot resistance alterations and decreased susceptibility to echinocandins. Such observation raises concerns and contradicts previous reports. In patients undergoing HSCT, van Burik et al.2 showed that micafungin prophylaxis (50 mg/day) was more effective than fluconazole in preventing fungemia (1.6% vs 2.4%). Considering the micafungin dose (50 mg/day), one may hypothesize that exposure to low dose micafungin increases the likelihood to select Candida sp. with decreased susceptibility, as previously described with fluconazole.10 In our report, emergence of high MICs C. krusei isolates is observed in two patients treated exclusively with echinocandins, not in patient no.1 treated with LAmB. Reduced susceptibility to echinocandins (suspected with Etest) was confirmed by the microdilution technique AFST-EUCAST and high MICs obtained in vitro were associated with hot-spot mutations of the Fks1 gene, unknown so far. This is the major finding of this report. This observation suggests a propensity to develop reduced echinocandin susceptibility during prolonged echinocandin therapy and under selection pressure. The authors declare no conflict of interest. We thank Catherine Blanc and Damien Hoinard for technical assistance. We thank the Platform Genotyping of Pathogen and Public Health, Institut Pasteur, for sequencing facility.