Sepsis from triple- vs single-lumen catheters during total parenteral nutrition in surgical or critically ill patients.

We prospectively studied the infection rates for 59 triple-lumen (TLC) and 68 single-lumen (SLC) subclavian catheters during the administration of total parenteral nutrition (TPN) to surgical or critically ill patients. A standard protocol was used for catheter insertion and maintenance. The infection control committee determined independently whether patients had catheter-related sepsis, an infected insertion site only, or no catheter infection. The TLCs had an increased incidence of catheter sepsis (19%) compared with the SLCs (3%). Low rates (5% for TLCs and 3% for SLCs) of infected catheter sites only indicated that the catheter care was comparable for both groups. The patients in the two groups were similar but not identical; those with TLCs appeared to be sicker and, therefore, at greater risk to develop catheter sepsis than patients with SLC. However, since TLCs were involved in six times more catheter sepsis than were SLCs, limiting the use of a subclavian catheter to giving TPN only and strict adherence to a TPN protocol are necessary to minimize the risk of catheter sepsis.