Reducing Central Venous Catheter–Associated Primary Bloodstream Infections in Intensive Care Units Is Possible: Data From The German Nosocomial Infection Surveillance System

Abstract Background And Objective: The German Nosocomial Infection Surveillance System (KISS) began in 1997 as a nationwide surveillance project for voluntary registration of nosocomial infections in intensive care units (ICUs). This study investigates trends in the rates of central venous catheter (CVC)–associated primary bloodstream infections (BSIs) in ICUs since participation in KISS. Methods: Eighty-four ICUs that had participated in KISS for at least 24 months were considered for more detailed analysis. Monthly rates of primary BSI for the 84 ICUs were pooled for the 24 months. The best model for describing the curve of reduction was sought. Additionally, incidence densities were compared using the z test. Results: For the 212 ICUs participating, a relative 25.7% decrease (from 2.1 to 1.6 primary BSIs per 1,000 CVC-days) was observed from January 1997 to June 2001. The 84 ICUs that participated in KISS for a minimum of 24 months accumulated 552,359 patient-days and 404,897 CVC-days during their 24 months. A linear regression model was selected to explain the curve of primary BSI reduction in the 84 ICUs. It showed a decrease from 2.1 to 1.5 primary BSIs per 1,000 CVC-days, meaning an overall relative reduction of 28.6% during the 2-year observation period. These results were significant (Student's t test for the monthly reduction coefficient; P = .04). The reduction of primary BSIs was shown for both clinical sepsis and laboratory-confirmed, CVC-associated primary BSIs. Conclusion: Performing surveillance with KISS was associated with a reduction of the rates of CVC-associated primary BSIs in ICU patients (Infect Control Hosp Epidemiol 2003;24:501-505).