In a three-year controlled trial of subcutaneous catheter tunnelling as a method of reducing total parenteral nutrition (TPN) catheter sepsis 99 silicone catheters (52 tunnelled, 47 untunnelled) were inserted into the subclavian (94%) or jugular (6%) veins under aseptic conditions. The influence of a nutrition nurse, who joined the nutrition team after 18 months, on catheter sepsis rate was also documented. Catheter sepsis was confirmed in 13 of 47 (28%) untunnelled catheters and only 6 of 52 (11·5%) tunnelled catheters (p