Long-term parenteral nutrition. II. Catheter-related complications.

Catheter-related complications were investigated in 70 patients receiving long-term parenteral nutrition (LTPN) for 1 to 63 months (median, 4.5 months) with a total observation period of 816 patient-months. Two hundred and three central venous catheters were used: 52 Broviac silicone rubber catheters and 151 Intracaths or Argyle baby feeding tubes made from polyvinyl chloride with plasticizers. The median duration was longer for Broviac catheters (5.0 months) than for the other catheters (1.1 months). Pneumothorax and/or subcutaneous emphysema occurred in 14 of 113 subclavian vein punctures (12%). Forty-eight episodes of catheter sepsis occurred in 27 patients. The incidence of sepsis was lower for Broviac catheters (0.3 per catheter-year) than for the other catheter types (0.9 per catheter-year). Catheter-induced thrombosis of a central vein was shown by phlebography 35 times among 25 (52%) of te 48 patients investigated, corresponding to an overall incidence of 1 in 22 patient-months. Total occlusion of a central vein occurred in 14 cases. The incidence of thrombosis was lower for Broviac catheters (0.5 per catheter-year) than for the other catheters (1.6 per catheter-year). Partial venous occlusions usually resolved after heparin therapy and/or catheter exchange, whereas total occlusions usually persisted for years. No correlation was found between incidence of complications and catheter lifetime. Seventeen patients died while receiving LTPN. In three patients death was related to LTPN: catheter sepsis, subdural haematoma possibly due to anticoagulant therapy, and respiratory failure caused by pulmonary infection after iatrogenic pneumothorax. For LTPN we recommend Broviac catheters, which showed longer duration, lower complication rate, and higher patient compliance. Catheter sepsis should be treated with both catheter exchange and antibiotics. Because of the high incidence of thrombosis we recommend that all patients on LTPN receive anticoagulant therapy. However, the value of anticoagulant therapy is not proved in a prospective, controlled study.