The changing epidemiology of healthcare-associated candidemia over three decades

We describe the epidemiology of healthcare-associated candidemia (HAC) in our tertiary care hospital, in comparison with both the pre-fluconazole (pre-FLU) and pre-echinocandin (pre-EC) eras. We identified all patients with HAC using microbiology records from 1/2004 to 12/2007, reviewed medical records, and pulled isolates for testing. We compared mortality, underlying illness, . Candida species distribution, and antifungal susceptibility with 2 prior University of Iowa cohorts (88 patients from 1983 to 1986 [pre-FLU], and 108 from 1997 to 2001 [pre-EC]). Of 108 patients with HAC from 2004 to 2007, species distribution was 47% . C. albicans, 29% . C. glabrata, 12% . C. parapsilosis, 6% . C. tropicalis, and no . C. krusei. Compared with pre-FLU and pre-EC eras, there was a reduction in % . C. albicans (from 61% and 60%, respectively), an increase in % . C. glabrata (from 0% and 16%), and no change in % . C. parapsilosis over time (12% and 12%). In-hospital mortality was lower in 2004-2007 than both pre-FLU and pre-EC (31% versus 57-61%), and 30-day mortality was also lower (33% versus 48% in pre-EC). Mean Charlson index was lower for the 2004-2007 cohort than pre-EC (3.0 versus 3.4)-fewer patients had leukemia or lymphoma (8% versus 16%) or other malignancies (18% versus 24%), while more were surgical patients (58% versus 48%). Using the new Clinical and Laboratory Standards Institute breakpoints for FLU and caspofungin, we found no caspofungin resistance, and FLU resistance only among . C. glabrata (15% had FLU MICs >32 μg/mL). The epidemiology of HAC is changing at our hospital, with continued emergence of . C. glabrata, fewer cases among oncology patients, and lower in-hospital and 30-day mortality.