Treatment-related risk factors for hospital mortality in Candida bloodstream infections

Objective: To examine the relationship between treatment-related variables for Candida bloodstream infection and hospital mortality. Design: Retrospective cohort analysis. Setting. Thousand two hundred-bed academic medical center. Patients. A total of 245 consecutive patients with Candida bloodstream infections who received antifungal therapy. Interventions: Identification of treatment-related risk factors: central vein catheter retention, inadequate initial fluconazole dosing, and delayed administration of antifungal therapy. Measurements and Main Results. A total of 245 patients with Candida bloodstream infections who received antifungal therapy were identified. One hundred eleven (45.3%) patients were managed in an intensive care unit and analyzed as a separate subgroup. In the hospital cohort, 72 (29.4%) patients died during hospitalization and 40 (36.0%) patients died in the intensive care unit cohort. In the hospital cohort, logistic regression analysis identified Acute Physiology and Chronic Health Evaluation II scores (1-point increments) (adjusted odds ratio [AOR], 1.18; 95% confidence interval [CI], 1.11-1.25; p = 0.003), corticosteroid use at the time a positive blood culture was drawn (AOR, 3.41; 95% Cl, 1.96-5.93; p = 0.027), inadequate initial fluconazole dosing (AOR, 3.31; 95% Cl, 1.83-6.00; p = 0.044), and retention of a central vein catheter (AOR, 4.85; 95% Cl, 2.54-9.29; p = 0.015) as independent determinants of hospital mortality. In the intensive care unit cohort, logistic regression analysis identified Acute Physiology and Chronic Health Evaluation 11 scores (1-point increments) (AOR, 1.21; 95% Cl, 1.14-1.29; p = 0.001), inadequate initial fluconazole dosing (AOR, 9.22; 95% Cl, 2.15-19.79; p = 0.004), and retention of a central vein catheter (AOR, 6.21; 95% Cl, 3.02-12.77; p = 0.011), as independent determinants of hospital mortality. For both cohorts the incremental presence of treatment-related risk factors was statistically associated with greater hospital mortality. Conclusions. Treatment-related factors, including retention of central vein catheters and inadequate initial fluconazole dosing, were associated with increased hospital mortality in patients with Candida bloodstream infections. These data suggest that optimization of initial antifungal therapy and removal of central vein catheters may improve the outcomes of patients with Candida bloodstream infections. (Crit Care Med 2008; 36:2967-2972)