Attributable mortality of candidemia after introduction of echinocandins.

OBJECTIVES Candidemia is among the most frequent nosocomial bloodstream infections. Landmark case-control studies on amphotericin B and fluconazole estimated attributable mortality rates of 38% and 49%, respectively. After introduction of echinocandins, these may have decreased. METHODS In a case-control design, 100 consecutive, hospitalised patients with candidemia were enrolled at the University Hospital of Cologne, Germany between 2014 and 2017. Controls were patients without candidemia matched for age, sex, year and duration of hospitalisation, main admission diagnosis and Patient Clinical Complexity Level (PCCL). Main data captured were risk factors for candidemia, attributable mortality rates and diagnostic and therapeutic adherence according to the EQUAL Candida score. RESULTS Overall mortality rates for cases and controls were 43% and 17% (P < .001), respectively; day 30 mortality rates were 38% and 11% (P = .03), accounting for an attributable mortality of 26% and 27%. Guideline adherence was higher in surviving vs non-surviving patients: while survivors reached a median of 17 (IQR: 16-19) points, non-surviving cases reached a median 16 (IQR: 14-18) points out of 22 maximum achievable points (P = .028). Risk factors for candidemia were more frequent in cases compared to control patients, especially chronic pulmonary disease (25% vs 16%; P = n.s.), chronic liver disease (21% vs 6%; P = .002), stay on intensive care unit (70% vs 64%; P = n.s.), respiratory failure (56% vs 50%; P = n.s.) and central venous catheter (97% vs 35%; P < .001). CONCLUSIONS Attributable mortality of nosocomial candidemia is still substantial but has decreased compared to previous studies with similar design.