Bacteremia associated with central venous catheter infection is not an independent predictor of outcomes

Background: Infection is the leading complication of central venous catheters. In the setting of suspected line infection, the CDC recognizes only catheter-related bloodstream infection but not catheter infection without bacteremia, which is designated colonization. To evaluate the hypothesis that catheter-related bloodstream infection has worse outcomes than catheter infection without bacteremia, we compared demographics, clinical data, and outcomes. Study Design: Analysis of catheter infections was performed on data collected prospectively for all episodes of infection occurring from December 1996 to September 1999 on the surgical services at a university hospital. Catheter tips were cultured only when infection was suspected. Catheter infection without bacteremia was defined as systemic evidence of infection, the presence of at least 15 colony-forming units on the catheter tip by a semiquantitative technique, and absence of bloodstream infection with the same organism as the catheter. Catheter-related bloodstream infection required the presence of bacteremia with the same organism as the catheter tip. Results: The 59 patients with catheter-related bloodstream infection had more coexistent infections than the 91 patients with catheter infection without bacteremia (2.9 +/- 0.1 versus 1.7 +/- 0.1; p = 0.0001), most commonly pneumonia (37.3% versus 16.5%, p = 0.004) and urinary tract infections (28.8% versus 8.8%, p = 0.001). Catheter-related bloodstream infection was associated with an increased proportion of gram-negative organisms compared with catheter infections without bacteremia (29.5% versus 16.9%, p = 0.04) and a trend toward fewer gram-positive organisms (61.5% versus 73.7%, p = 0.07). There were no differences in APACHE II score, WBC, length of hospital stay, time from admission to fever, time from fever to treatment, normalization of WBC, days of antibiotics, defervescence, gender, presence of comorbidities, occurrence of colonization while in an ICU, or mortality rate (18.6% with bacteremia, 24.2% without; p = 0.42). Conclusions: The presence of bloodstream infection in addition to catheter infection does not appear to alter outcomes. The definition of catheter infection perhaps should be extended to include catheter infections without bloodstream infection in the presence of systemic illness without another source. (J Am Coll Surg 2000;190:671-681. (C) 2000 by the American, College of Surgeons).