Objectivesa) To reduce the rate of catheter-related infection, using improved skin preparation and catheters impregnated with silver sulfadiazine and chlorhexidine; b) to decrease the number of unnecessary guidewire exchanges of existing catheters by substituting suspected catheter-related sepsis for fever alone as an indication to change an indwelling catheter; and c) to decrease the hospital costs associated with guidewire exchanges and new catheter insertions.DesignSequential, prospective, descriptive studies using a continuous quality management approach.SettingA 20-bed trauma intensive care unit at a university teaching hospital.PatientsPatients (n = 147) admitted from July 1 to December 31, 1992 (phase 1); 34 patients admitted in June and September 1993 (phase 2); and 156 patients admitted between January 1 and June 30, 1994 (phase 3).InterventionsPhase 1: Proportions of catheter-related infections and catheter-related bacteremia were compared with our prior reported results. Indications for guidewire exchange were analyzed, and the rate of catheter-related infection for each indication was derived. Phase 2: The rate of catheter-related infection was determined for a trial group of triple-lumen catheters impregnated with silver sulfadiazine and chlorhexidine. Phase 3: Four components were altered. Impregnated triple-lumen catheters were used instead of unprotected catheters. Chlorhexidine skin cleanser was substituted for povidone-iodine solution during initial aseptic preparation during catheter insertion and subsequent guidewire exchange. Suspected catheter-related sepsis was substituted for fever as an indication for guidewire exchange. The "safe" period (the time before considering changing a catheter because catheter-related sepsis was suspected) was extended from 2 to 4 days.Measurements and Main Results10 days compared with lesser durations. Fever was the indication for 42% of all guidewire exchanges. In phase 2, the catheter-related infection rate was 2% (one positive of 48 catheters), significantly (p = .0231) lower than the phase 1 rate. In phase 3, the overall rate of catheter-related infection was 8.6%, significantly (p = .0067) lower than the 15% rate in phase 1. The proportion of protected triple-lumen catheter catheter-related infections decreased significantly (p = .0024) from 15% to 6%. The rate of catheter-related infection for introducers was the same in both phases (p = .33). The days of catheterization for all catheters increased from 4.5 +/- 2.6 to 5.4 +/- 3.6 (p < .0001). The days for triplelumen catheters increased from 4.7 +/- 2.7 to 7.0 +/- 3.9 (p < .0001). For introducers, there was no difference in the days of catheterization. The proportion of catheters changed for suspected catheter-related sepsis decreased significantly (p < .0001) to 23% from the 42% changed for fever in phase 1. The proportion of catheter-related infections for catheters changed for fever was 18% in phase 1. The proportion of catheter-related infections for catheters changed for suspected catheter-related sepsis was 13% in phase 3 (p = .43). The total number of catheters used per patient in phase 3 was 1.9 +/- 1.4, significantly lower than the 2.6 +/- 2.7 catheters used in phase 1 (p = .0018). The number of triple-lumen catheters decreased from 1.8 +/- 1.2 to 1.0 +/- 1.2 in phase 3 (p = .0001).ConclusionsCatheters impregnated with silver sulfadiazine and chlorhexidine had a smaller proportion of catheter-related infection compared with unprotected catheters. Fever alone as an indication for guidewire exchange resulted in an increased number of unnecessary procedures. Using protected catheters and suspected catheter-related sepsis, together with an increase in the safe period before guidewire exchange, decreased the rate of catheter-related infection and increased the duration of catheterization. Together, these factors significantly decreased the number of catheters used for guidewire exchange and new catheter insertions per patient. The cost savings to the hospital were approximate $4,750 per month.(Crit Care Med 1996; 24:1660-1665)