One hundred fifty-six triple-lumen central venous catheters placed into 65 different sites in 52 surgical or trauma patients were prospectively evaluated to determine the frequency of catheter-related infection and the efficacy of multiple guidewire exchanges. Thirty-four anatomic sites in 33 patients without clinical signs of sepsis were compared to 31 anatomic sites in 19 patients with a clinical diagnosis of sepsis based on clinical, microbiologic, and radiologic evaluation. Anatomic sites were used indefinitely unless a positive semiquantitative catheter culture (greater than or equal to 15 cfu) or catheter-associated bacteremia developed. Seven catheter-related infections developed in five septic patients (26.3%) while no infections developed in the nonseptic group (p = 0.004). Three of 31 anatomic sites (9.6%) were associated with bacteremia. The number of hospital days to initial catheter insertion and the length of catheter site use were considerably higher in the septic patients compared to the nonseptic group. We conclude that triple-lumen catheter-related infections are uncommon in patients with no evidence of infection and can be safely guidewire-exchanged to prolong site use. Catheter-related infections in septic patients are much more frequent and unpredictable. Although guidewire exchange has a role in septic patients, multiple exchanges cannot routinely be recommended.