Background and purpose: Candidemia has been associated with a very high mortality. This study evaluated the predictors of candidemia-related mortality at a teaching hospital in northern Taiwan.
Methods: We conducted a retrospective analysis of adult patients with candidemia between September 2003 and May 2005. A stepwise logistic regression analysis was performed to determine the predictors of candidemia-associated mortality. All Candida isolates were identified to species by use of the ATB ID 32C kit and their susceptibilities to antifungal agents were tested by ATB Fungus 2 system.
Results: 179 episodes in 174 adult patients with candidemia were identified retrospectively. The predictors of mortality included duration of prior antibiotics >or=28 days, Acute Physiology and Chronic Health Evaluation (APACHE) II score >or=23 and retention of central venous catheters (CVCs). There was no statistically significant association between the time to the start of antifungal therapy and mortality from nosocomial candidemia. In addition, there was no significant association between the time to CVC removal and mortality after stratification by APACHE II score.
Conclusions: Despite effective antifungal therapy after the onset of candidemia in more than half of the patients studied, mortality remained very high, especially in the groups with longer duration of prior antibiotic treatment, higher APACHE II score and CVC retention. Timing of CVC removal after onset of candidemia was not correlated with mortality.