Candidaemia associated with decreased in vitro fluconazole susceptibility: is Candida speciation predictive of the susceptibility pattern?

Background: Candidaemia is often treated with fluconazole in the absence of susceptibility testing. We examined factors associated with candidaemia caused by Candida isolates with reduced susceptibility to fluconazole. Methods: We identified consecutive episodes of candidaemia at two hospitals from 2001 to 2007. Species identification followed CLSI methodology and fluconazole susceptibility was determined by Etest or broth microdilution. Susceptibility to fluconazole was defined as: full susceptibility (MIC <= 8 mg/L); and reduced susceptibility (MIC >= 32 mg/L). Complete resistance was defined as an MIC >32 mg/L. Results: Of 243 episodes of candidaemia, 190 (78%) were fully susceptible to fluconazole and 45 (19%) had reduced susceptibility (of which 27 were fully resistant). Of Candida krusei and Candida glabrata isolates, 100% and 51%, respectively, had reduced susceptibility. Despite the small proportion of Candida albicans (8%), Candida tropicalis (4%) and Candida parapsilosis (4%) with reduced fluconazole susceptibility, these species composed 36% of the reduced-susceptibility group and 48% of the fully resistant group. In multivariate analysis, independent factors associated with reduced fluconazole susceptibility included male sex [odds ratio (OR) 3.2, P<0.01], chronic lung disease (OR 2.7, P=0.01), the presence of a central vascular catheter (OR 4.0, P<0.01) and prior exposure to antifungal agents (OR 2.2, P=0.04). Conclusions: A significant proportion of candidaemia with reduced fluconazole susceptibility may be caused by C. albicans, C. tropicalis and C. parapsilosis, species usually considered fully susceptible to fluconazole. Thus, identification of these species may not be predictive of fluconazole susceptibility. Other factors that are associated with reduced fluconazole susceptibility may help clinicians choose adequate empirical anti-Candida therapy.