Despite a recent focus on prevention of health care-associated infections, rates of Candida blood stream infections (BSI) in adults have remained largely unchanged. We examined trends in Candida BSI and compared them to those of Staphylococcus aureus BSI at a single site to assess changes over time.
We identified inpatients at the Atlanta Veterans Administration Medical Center from 1999-2013 retrospectively for total and hospital-onset (HO) Candida and S. aureus BSI. We ascertained the implementation dates of infection control practices through a survey to the infection preventionist and hospital epidemiologist. We report intravenous (IV) antibiotic and antifungal drug use from 2006-2012. χ2 test was used to compare rates over time.
We identified 311 Candida BSI. Rates (cases/1,000 unique patients) of Candida BSI declined from 0.39 in 2011 to 0.06 in 2013 (p <0.0001); without changes in Candida species distribution. The HO-Candida BSI rate (cases/1,000 patient days) declined in parallel from 0.42 in 2011 to 0.07 in 2013 (p<0.0001). This change mirrors the recent decline in HO-S. aureus BSI from 0.28 in 2011 to 0.11 in 2013 (p <0.0001), although those rates have been decreasing for several years. Rates (defined daily doses/1,000 patient days) of both predominantly gram-positive (141 to 213, p=0.002) and gram-negative (108 to 155, p=0.04) IV antibiotics increased significantly from 2006 to 2012. However, oral and IV fluconazole use decreased significantly over the same time period (72 to 30, p <0.0001). An ongoing hand hygiene campaign started in 1999, use of antiseptic-coated central venous catheters (CVCs) started in 2002, whereas the Institute for Healthcare Improvement's central line insertion bundle was implemented starting in 2002. Chlorhexidine bathing began in 2010. For CVCs, chlorhexidine dressing use started between 2011-2013, and alcohol cap use began in 2013 (Figure 1).
Candida BSI rates decreased at our institution from 1999-2013. Those declines cannot be explained by decreasing broad spectrum antibiotic use or increasing antifungal prophylaxis, and are temporally associated with several infection control interventions that relate to CVC maintenance rather than insertion.
R. Gaynes, None
D. Rimland, None