Epidemiology and mortality of candidemia both related and unrelated to the central venous catheter: a retrospective cohort study.

Our objective was to compare mortality, epidemiology, and morbidity in hospitalized patients with candidemia which was both related and unrelated to the central venous catheter (CVC). This was a monocentric, retrospective cohort study of candidemia. The sample consisted of 103 patients with laboratory-confirmed nosocomial candidemia hospitalized between 2006 and 2013 in a tertiary care public hospital. We included 65 (63.1 %) patients (24 in the CVC-positive group, 41 in the CVC-negative group). Demographic data and risk factors were recorded using a structured case report form. In the group of candidemia associated to the CVC, survival at day 50 was 58.6 ± 11.9 %, compared to 26.5 ± 8.9 % for the CVC-negative group (p-value = 0.012); the hazard ratio of death was 0.38 (95 % confidence interval 0.17-0.85, p-value = 0.019). Compared with the CVC-positive patients, CVC-negative patients were often colonized with yeast (41.5 % vs. 16.7 %, p-value = 0.041), had a shorter previous in-hospital stay (20 days vs. 34 days, p-value = 0.023), and were more severely ill (severe sepsis 85.4 % vs. 58.3 %, p-value = 0.016). In this study, when the origin of candidemia was not the CVC, patients were more seriously ill, had a higher mortality rate, and the removal of the catheter seemed to lead to disappointing results. It would be useful to explore the impact of retention of the CVC on survival in the CVC-negative patients, where the CVCs are essential to treating these patients.