Infection and Natural History of Emergency Department–Placed Central Venous Catheters

Abstract

Central line-associated bloodstream infection (CLABSI, hereafter referred to in this paper as "bloodstream infection") is a leading cause of hospital-acquired infection. To our knowledge, there are no previously published studies designed to determine the rate of bloodstream infection among central venous catheters placed in the emergency department (ED). We design a retrospective chart review methodology to determine bloodstream infection and duration of catheterization for central venous catheters placed in the ED. Using hospital infection control, administrative, and ED billing databases, we identified patients with central venous catheters placed in the ED between January 1, 2007, and December 31, 2008, at one academic, urban ED with an annual census of 57,000. We performed a structured, explicit chart review to determine duration of catheterization and confirm bloodstream infection. We screened 4,251 charts and identified 656 patients with central venous catheters inserted in the ED, 3,622 catheter-days, and 7 bloodstream infections. The rate of bloodstream infection associated with central venous catheters placed in the ED was 1.93 per 1,000 catheter-days (95% confidence interval 0.50 to 3.36). The mean duration of catheterization was 5.5 days (median 4; range 1 to 29 days). Among infected central venous catheters, the mean duration of catheterization was 8.6 days (median 7; range 2 to 19 days). A total of 667 central venous catheters were placed in the internal jugular (392; 59%), subclavian (145; 22%), and femoral (130; 19%) veins. The sensitivity of using ED procedural billing code for identifying ED-placed central venous catheters among patients subsequently admitted to any ICU was 74.9% (95% confidence interval 71.4% to 78.3%). The rate of ED bloodstream infection at our institution is similar to current rates in ICUs. Central venous catheters placed in the ED remain in admitted patients for a substantial period.