Bloodstream infections due to Candida species in the intensive care unit: identifying especially high-risk patients to determine prevention strategies.

Nosocomial bloodstream infections due to Candida species are associated with a 40% crude (total) mortality rate in the United States. The attributable mortality rate defines the number of deaths due directly to the infection and represents the proportion of the total number of deaths influenced by the use of anti-infectives. We used 4 defined risk factors to create a conditional logistic regression model of individual patients' risk of candidemia and then examined the possibility of using a calculated risk threshold to begin treatment with anti-Candida antibiotics. We subsequently examined the number needed to treat (NNT) to prevent a single death related to bloodstream infection due to Candida species. By use of our model and assuming a 33% threshold for treating with an antibiotic, which is assumed to prevent 65% of candidal bloodstream infections, we show that the NNT would be 7 patients. Use of defined risk factors to select patients who are at high risk for candidemia and treatment with effective antibiotics would be an efficient way to reduce infection-related deaths.