Effect of frequent guidewire changes on triple-lumen catheter sepsis.

Frequent guidewire changes of single-lumen (SLC) and triple-lumen (TLC) catheters have been proposed to decrease catheter sepsis. We placed TLC in 126 patients needing total parenteral nutrition (TPN) and multiple venous access, prospectively randomizing them to two groups: group I received a guidewire change every 3 days, and group II received guidewire changes for mechanical or septic complications only. Tips were cultured at each line change and tips and blood for each septic episode. Catheter sepsis was defined by the criteria of the Association for Practitioners in Infection Control (APIC). There were 67 positive cultures in 52 patients, but most produced very few colonies or grew the same organisms in other infection sites. Forty-seven% of all cultures grew Staphylococci, and 23% grew Candida. APIC-defined catheter sepsis was detected in 12.7% of group I and 15.9% of group II. Although we observed no statistically significant difference in the two techniques, if we assume that a 20% difference in the incidence of catheter-induced sepsis would be important to detect, the probability of failing to detect such a difference is 0.24 with an 0.05 level of significance (two-sided). Prophylactic guidewire changes did not alter the incidence of catheter sepsis in patients with TLC who required TPN. The high rate of sepsis and Candida infection may be due to the critical illness of the immunocompromised population studied.