Support for the EUCAST and revised CLSI fluconazole clinical breakpoints by Sensititre® YeastOne® for Candida albicans: a prospective observational cohort study.

OBJECTIVES

Species-specific clinical breakpoints (CBPs) for Candida spp. were established following consideration of clinical outcomes in patients with oesophageal candidiasis. We sought to further determine the validity of the current CBPs based on data from a prospective candidaemia study.

PATIENTS AND METHODS

All Candida albicans candidaemia episodes in patients enrolled in the Australian Candidaemia Study and who were treated with fluconazole monotherapy were included. Fluconazole MICs were established using Sensititre(®) YeastOne(®).

RESULTS

Two hundred and seventeen evaluable episodes were identified, 93.5% of which occurred in adult patients. Fluconazole was commenced within 72 h of blood culture positivity in 96.3% (209/217) of episodes. Fluconazole doses were appropriate in 89.9% (195/217) of episodes and the median duration of therapy was 14 days (IQR 8-21 days) for the whole cohort. The all-cause 30 day mortality was 19.8% (43/217), with 37.2% (16/43) of deaths attributed to candidaemia. Classification and regression tree (CART) analysis identified a fluconazole MIC target of ≥2 mg/L for infection-related mortality and ≥4 mg/L for overall 30 day mortality. Overall mortality was no different in episodes with isolates above or below the identified MIC target, although there was a trend towards significance (P = 0.051). On univariate analysis, infection-related mortality was significantly increased in C. albicans episodes with an MIC ≥2 mg/L compared with those below this MIC target (20.6% versus 4.9%; P = 0.001). This target remained an independent predictor of infection-related mortality (OR 8.2; 95% CI 2.3-29.7; P = 0.001).

CONCLUSIONS

We observed a direct relationship between infection-related mortality and rising fluconazole MIC for C. albicans candidaemia; overall, the data support the EUCAST and revised CLSI fluconazole clinical breakpoints.