A 2 year experience of a nutritional support service: prospective study of 229 non-intensive care patients receiving parenteral nutrition.

During a 2 year period, 229 non-intensive care patients received total parenteral nutrition (TPN) for a mean of 23.2 days. Nutritional parameters of weight, triceps skin-fold thickness (TSF), mid-arm muscle circumference (MAMC), total lymphocyte count (TLC), serum albumin and serum transferrin were measured prior to commencement of TPN and then at weekly intervals. All parameters showed an upward trend during the period of TPN. This was small and not statistically significant for weight, TSF, MAMC and albumin; the trends were greater and statistically significant for transferrin (P = 0.001) and TLC (P = 0.002). In contrast, in the 14.4% of patients who died, albumin, transferrin and TLC all fell. Patients who died had a significantly lower initial albumin (P = 0.05), transferrin (P = 0.04) and TLC (P = 0.04). The last values obtained in patients who died were very significantly lower for albumin (P less than 0.001), transferrin (P less than 0.001) and TLC (P = 0.003). Single-lumen tunnelled subclavian catheters had a significantly lower incidence of catheter sepsis (4.8%) compared with double-lumen (17.4%) and triple-lumen (13.7%) catheters (P = 0.01). There was also a greater incidence of mechanical and thrombotic complications with multiple-lumen catheters compared with single-lumen catheters (P = 0.02). This study shows that the nutritional indices albumin, transferrin and TLC have prognostic significance. Single-lumen rather than multiple-lumen catheters should be used for administration of TPN whenever possible.