Community-onset candidemia at a university hospital, 1995-2005

Background and Purpose: Although not all candidemias are hospital-acquired, data on clinical epidemiology for the community-onset candidemia are limited. This retrospective study was conducted to describe predisposing factors and outcomes of community-onset candidemias. Methods: Medical records of patients who were admitted to the National Taiwan University Hospital between January 1, 1995 and May 31, 2005 and had Candida isolated from their blood in the outpatient setting and/or within 48 h of hospitalization (community-onset) were reviewed. Results: A total of 56 episodes of candidemia were reviewed, which included 8 episodes (14.3%) of true community-acquired candidemia occurring in patients with no record of hospitalization within the previous 30 days and without histories of invasive procedures either just before or at the time of admission, and 48 episodes (85.7%) that were health care-associated. The latter included 24 episodes (42.9%) in patients recently discharged from hospitals (within 2-30 days of current admission), 23 episodes (41.1 %) associated with invasive procedures and/or central intravascular lines placed for outpatient therapy, and 1 episode (1.8%) in patients admitted from nursing homes. Gastrointestinal bleeding (46.4%), immunosuppressive therapy (42.9%) and previous antibiotics use (37.5%) were the most common predisposing factors. Diabetes was the single most important predisposing factor in true community-acquired candidemia (62.5%) and had a significantly higher prevalence among these patients than in those with health care-associated candidemias (p=0.035). Candida albicans was the most common isolate (39.7%), followed by Candida tropicalis (22.4%) and Candida glabrata (17.2%). The overall case fatality rate was 55.4% (31/56), and 58.1% (18/31) of this was attributable to candidemia. Multivariate analysis identified higher severity score and lack of antifungal therapy as having an independent and adverse influence on outcome. Conclusions: Up to 85.7% of community-onset candidemias are health care-associated. There is a conceptual and practical need for a new classification for the spectrum of acquisition of infection, wherein the new category of health care-associated infection will have implications for the selection of empirical therapy.