The ever-evolving landscape of candidaemia in patients with acute leukaemia: non-susceptibility to caspofungin and multidrug resistance are associated with increased mortality.

OBJECTIVES: The epidemiology and clinical course of candidaemia in patients with acute leukaemia, a population frequently exposed to antifungals, have not been extensively studied. In the present contemporary series of acute leukaemia patients, we describe patient characteristics, Candida species and MIC distributions and investigate the association between antifungal resistance and all-cause mortality. METHODS: We performed a retrospective review of medical records and microbiological data of adult patients with acute leukaemia or high-risk myelodysplastic syndrome with at least one positive blood culture for Candida species at the MD Anderson Cancer Center between January 2008 and October 2012. Susceptibility was defined according to the 2012 epidemiological cut-off values and clinical breakpoints. RESULTS: We identified 67 episodes of candidaemia in 65 patients. Almost all episodes (94%) occurred in patients who were receiving antifungal agents, 71% in patients receiving an echinocandin. Almost all isolates (99%) were of non-albicans Candida species [most frequently Candida parapsilosis (32%), Candida tropicalis (23%) and Candida glabrata (20%)]. Caspofungin non-susceptibility was significantly associated with fluconazole resistance (P < 0.001). Non-susceptibility to caspofungin and multidrug resistance were associated with excess 14 day [adjusted HR (aHR) 3.02 (95% CI 1.28-7.09), P = 0.011 and aHR 3.02 (95% CI 1.27-7.14), P = 0.012, respectively] and 30 day [aHR 2.96 (95% CI 1.38-6.37), P = 0.005 and aHR 2.86 (95% CI 1.31-6.21), P = 0.008, respectively] all-cause mortality. CONCLUSIONS: In patients with acute leukaemia, a shift in candidaemia epidemiology was noted with a 99% predominance of non-albicans species. Non-susceptibility of Candida strains to caspofungin or multidrug resistance were independent markers of poor outcome in this patient population.