Methods: We retrospectively identified all patients with candidemia (based on blood culture) during a five year period (Jan 2012-Dec 2016). Subjects were either diagnosed in the ICU or required ICU admission within 48 hours of developing candidemia. Co-primary endpoints included crude hospital mortality and distribution of species causing candidemia. The minimum inhibitory concentration 90 (MIC90) of isolates to micafungin (MICA) was examined to assess for changes over time. We further explored the impact of delayed and/or inappropriate therapy (DIT) on survival. We defined delayed therapy as initiation of an antifungal > 24 hrs after collecting the eventually positive blood culture and inappropriate therapy as the use of fluconazole (FLU) when isolate in vitro resistant.
Results: ±15.6 years, 47.4% male, 4 with ≥2 candida species). Crude hospital mortality equaled 27.9%. C. albicans accounted for 31.4% of isolates while non-albicans species represented 68.6% of cases. C. glabrata was the most frequent organism (42.3%). Mortality rates varied in a statistically significant way by species: C. parasalopsis (4.2%), C. tropicalis (7.1%), C. albicans (28.8%), C. glabrata (39.7%), (p=0.002). All isolates were in vitro susceptible to MICA with a median MIC90 of 0.032 ng/dl. MICA MIC90s did not change over time or correlate with mortality. Among C. albicans 18.6% of patients received DIT compared to 42.3% of C. glabrata subjects (p=0.032).
Conclusion: Candidemia in the ICU continues to be associated with substantial short-term mortality. C. glabrata now represents the most common cause of candidemia in this setting. Our observations support current guidelines recommending use of echincandins for suspected candidemia in the ICU given the risk for DIT with empiric FLU therapy and apparently stable in vitro activity of MICA despite its broader use.
N. Biru, None
C. Woods, None
A. Shorr, Astellas: Consultant , Consulting fee . Cidara: Consultant , Consulting fee . Pfizer: Consultant , Consulting fee .