Background: Intravascular catheters (IVCs) are a major cause of hospital onset (HO) BSIs but the contributions of IVCs to community onset (CO) BSI is less well characterized.
Objective: To evaluate the secular trends in BSI at a single 500-bed university teaching hospital over 3 decades, comparing HO to CO infections, and to assess impact of IVC on these trends.
Methods: Data collected as part of routine active infection control surveillance between 1980 and 2009 at Emory University Hospital Midtown were reviewed. BSIs were classified as CO (including healthcare-associated infections) or HO until 2005, after which time only HO-BSI data were available. Each BSI was attributed to IVC or other sources using contemporary CDC definitions.
Results: During the study period, there were 12,030 BSIs including 6915 (57%) CO-BSI. CO-BSI increased from 143 in 1980 to 472 in 2005 (figure 1). Much of the increase in CO-BSIs was due to IVC-related BSIs which increased from 0% (0/143) of BSI in 1980 to 24% (115/472) in 2005. After more than doubling from 98 to 210 (1980-1989), HO-BSI remained relatively constant (range: 173-241 BSI) for the next 15 years, followed by a decrease during the last 5 years (range: 144-205). The proportion of HO-BSI attributed to IVC increased dramatically during the 1980s, from 7% (7/105) in 1980 to 46% (95/207) in 1990. IVC have remained a significant cause of HO-BSI over the past 2 decades. Among the CO-BSI, 491/833 (59%) of IVC-associated BSI were related to hemodialysis catheters, 21% with ports and 9% with peripherally inserted central catheters (PICCs).
Conclusions: Although IVC remain a common cause of HO-BSI, the overall rate of HO-BSI has decreased in the past 5 years. The shift of intravenous therapy to the outpatient setting, including use of hemodialysis catheters, ports and PICCs has contributed to increases in CO-BSI.