30 Patient age and incubation time to positivity were the only independent predictors of mortality in a multivariate analysis controlling for several other known risk factors (e.g. APACHE II score, neutropenia, catheter removal). Mortality increased by 2% per hour of
incubation time elapsed. Median time to initiation of antifungal therapy after notification of culture positivity was 7 h, which indicates another delaying factor with room for improvement. The cohort analysis performed by Morrell et al.  previously identified delays of the start of therapy after retrieval of blood culture sample as a risk factor for hospital mortality. Delaying the initiation of therapy for more than 12 h after retrieval of the blood sample that later yields positive results was associated with almost this website threefold increase in hospital mortality (from 11% to 30%) and was identified as an independent risk factor for mortality in multiple logistic regression analysis, as Selleck STA-9090 were APACHE II score and prior antibiotic therapy. Another cohort analysis by Garey et al.  used different time categories and found that delays in the initiation of antifungal therapy beyond 24 h
after blood sampling significantly increased the mortality from 15% to 24% (therapy started on day 2) and 37% (day 3). The influence of timely initiation of antifungal treatment was confirmed in a neutropenic animal model. Increasing the delay in drug administration gradually reduced the therapeutic efficacy to a point at which the drug effect was completely abolished.39 Taken together, these data clearly underscore the need for early and – in septic shock – immediate initiation of therapy. The European sepsis guidelines issued in 2008 advocate the immediate Interleukin-3 receptor use of antifungals in septic shock patients at high risk of candidaemia, albeit somewhat indirectly: they argue that calculated antimicrobial therapy should be started within 1 h after recognition of septic shock or severe sepsis without shock and that clinicians should consider
whether Candida is a likely pathogen when choosing the initial regimen.40 In the light of a 33-h median time to blood culture positivity (see above), we either need innovative diagnostic tools for much earlier identification of Candida in the bloodstream or we have to enhance our ability to identify patients being at high risk of having candidaemia when developing signs and symptoms of systemic infection. The difficulties of identifying patient groups at high risk are illustrated by a recent prospective randomised trial. Schuster et al.  compared empirical fluconazole with placebo in ICU patients deemed at risk for IC. Inclusion criteria were: ICU stay of ≥96 h, APACHE II score of ≥16, 4 days of fever, broad-spectrum antibiotics for ≥4 days and presence of a central venous catheter.