Methods: In a case-control design, 100 consecutive, hospitalized patients with candidemia were enrolled at the University Hospital of Cologne. These cases had at least one blood culture positive for Candida spp. >48 hours post admission. We enrolled patients from Jan 2017 backwards until Feb 2014. Controls were patients without candidemia matched for age, sex, calendar year, duration of hospitalization, main admission diagnosis, and Patient Clinical Complexity Level. Risk factors for candidemia captured were malignancy, diabetes, infection other than candidemia, liver cirrhosis, hemodialysis, congestive heart failure, coronary artery disease, chronic lung disease, intensive care, mechanical ventilation, and presence of central lines. For each control patient we considered the day of candidemia of its matched case to compare post diagnosis length of stay. We estimated attributable mortality until day 30 post candidemia diagnosis. We performed Χ²-test for categorical and Students t-test for continuous variables, and defined a two-tailed p-value <0.05 statistically significant.
Results: Cases and controls were 68% males. Median age was 62 and 63 years, and 25th and 75th percentile 55 and 74 years in both groups. Candidemia occurred a median 18 days post admission. For cases and controls, median length of stay post diagnosis was 17 and 15.5 days (p=0.13), for those controls who died 12 and 19 days (p=0.21), and for survivors 24 and 13 days (p=0.006). Day 30 mortality rates were 38% and 11% for cases and controls (p=0.03), thus attributable mortality was 27% (95% CI, 16%-28%).
Conclusion: Attributable mortality of nosocomial candidemia is still substantial, but was lower in our study as compared to literature from before introduction of echinocandins.
P. Koehler, None
F. Koehler, None
S. Mellinghoff, None