Objective: In this study we sought to identify differences in risk factors and outcome of critically ill patients with Candida albicans and non-albicans candidemia.
Methods: Nonimmunosuppressed, nonneutropenic patients with candidemia diagnosed after intensive care unit (ICU) admission were included in a prospective observational study in a medical-surgical ICU at a tertiary academic hospital in Athens, Greece.
Results: During the study period (January 2001 to December 2005), 56 candidemia episodes in 1037 ICU admissions were included (5.4%). Of these patients, 36/56 (64.3%) had candidemia due to C. albicans and 20/56 (35.7%) due to non-albicans species (8/56 [14.3%] C. glabrata, 6/56 [10.7%] C. tropicalis, 3/56 [5.4%] C. parapsilosis, 1/56 [1.8%] C. lusitaniae, 1/56 [1.8%] C. krusei and 1/56 [1.8%] C. dubliniensis). Administration of glucocorticosteroids, central venous catheter placement, and preexisting candiduria were independently associated with candidemia due to C. non-albicans species (Odds ratio [OR]: 45.1, 95% confidence interval [CI]: 3.0-669.9; OR: 26.2, 95% CI: 2.1-334.8; and OR: 16.5, 95% CI: 1.6-173.9, respectively). The treatment response rate differed significantly between patients with C. albicans and patients with C. non-albicans bloodstream infections (29/36 [80.6%] vs 9/20 [45%], P = 0.006). Overall mortality was higher in patients with non-albicans species than C. albicans bloodstream infections (18/20 [90%] vs 19/36 [52.8%], P = 0.005). Multivariable logistic regression analysis revealed that candidemia due to non-albicans species was independently associated with death (OR: 6.7, 95% CI: 1.2-37.7).
Conclusions: In the subset of critically ill nonimmunosuppressed patients, candidemia caused by non-albicans species occurred more frequently in those with medical devices or receiving steroids. Candidemia due to non-albicans species was also associated with higher mortality.