Do clinical features allow for accurate prediction of fungal pathogenesis in bloodstream infections? Potential implications of the increasing prevalence of non-albicans candidemia

OBJECTIVE: To describe the evolving epidemiology of fungal bloodstream infections in critically ill and noncritically ill patients and to identify predictors of infection with non-albicans yeast species. DESIGN: Retrospective case series. SETTING: Two academic, tertiary care centers. PARTICIPANTS: All persons during a 4-yr period who developed fungemia. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We initially compared subjects with Candida albicans vs. alternative yeast. In a sensitivity analysis, we compared persons with potentially fluconazole-resistant organisms (Candida glabrata and Candida krusei) to those with other fungi. We also repeated these analyses in the subgroup of persons in the intensive care unit when they developed fungemia. The study cohort included 245 patients (60% in the intensive care unit), and C. albicans accounted for 52% of infections, whereas C. glabrata represented 20% of cases. The distribution of isolates was similar in both intensive care unit patients and those on the wards. In the entire population, no variable, including both previous fluconazole exposure and severity of illness, correlated with the fungemia due to a non-albicans species. In our sensitivity analysis, no factor was independently associated with a potentially fluconazole-resistant yeast. For the subgroup of subjects whose fungemia was diagnosed while they were in the intensive care unit, no variable differentiated C. albicans from non-albicans isolates. CONCLUSIONS: Non-albicans yeast are common both in the intensive care unit and on the wards. Simple clinical factors do not allow the clinician to effectively identify patients likely infected with non-albicans pathogens or with possible fluconazole-resistant fungi.