12 July 2003

Dear Editor,

Central venous catheter insertion is technically an easy procedure for a neonatal registrar, but to achieve optimal position for the catheter tip can be difficult. The recommended correct positions are either at the junction between the superior vena cava and the right atrium or immediately above the diaphragm if inserted into the inferior vena cava1. Extravasations from catheters in the inferior vena cava have been described to cause ascites and acute abdomen in neonates2. We report a case of malpositioning of a peripherally inserted central catheter (PICC), delayed presentation of an abdominal complication related to it and complete recovery on removal of the long line.

A 29-week-old gestation male infant, weighing 1.06 kg, was admitted and electively ventilated with administration of two doses of surfactant. Umbilical arterial and venous catheters were inserted on day 1. Non-nutritive orogastric feeding was started on day 5. On day 8, abdominal distension was noted and orogastric aspirates were bilious. Feeds were then stopped. An abdominal X-ray showed a thickened bowel wall. Blood counts were normal. He was treated for early necrotizing enterocolitis (NEC) after removal of umbilical lines.

A PICC 27-G (Vygon, Nutriline, Germany) was inserted on day 10, entering the saphenous vein just medial to the left knee for a length of 15 cm. Plain radiography of the abdomen reported the tip of long line in the common iliac vein. Total parenteral nutrition (TPN) was started through the long line. Abdominal distension improved and there was no more bilious aspirate. On day 14, a redness of the epigastric and hypochondrial areas was noted. The periumbilical area was normal. A plain X-ray of the abdomen showed normal gaseous pattern. Cloxacillin was added and NEC measures continued as before. Surgical opinion sought at that time was that of a progressing NEC with associated abdominal wall erythema. The redness and induration of the anterior abdominal wall continued to increase. By day 18, a localized area of cellulitis was noted superior and to the left of the umbilicus. This became necrotic within 24 h and abdominal distension increased. Under sterile precautions, a thin eschar was removed and whitish fluid drained out. Microscopic examination of the fluid did not show any pus cells or bacteria and culture was sterile. Dressings applied over the drained area became continuously wet with a whitish discharge. A repeat microscopic examination of the fluid the next day showed flat globules. A deliberate diagnostic purge of intralipid being given through the long line resulted in visible oozing at the site of the ulcer on the abdomen. This confirmed a communication of the long line with the ulcer.

It was noted that there had been no obvious migration of the line as indicated by the PICC markings externally, and there was no resistance while the long line was removed. All abdominal X-rays had shown no change in the site of the catheter tip. There had been no syringe pump alarms of high pressure at any time. Throughout this period, blood counts and biochemical parameters were normal and the culture of blood and the catheter tip were negative.

The abdominal wall erythema completely resolved in 2 days. Feeds were then restarted. The ulcer healed quickly leaving only a small scar. The infant tolerated feeds well and started gaining weight. He was discharged home on day 51 with a weight of 2020 g.

A review of the abdominal X-ray of this case, taken on the day of insertion of the venous catheter, showed coiling of the  catheter at the level of the head of femur. This would explain the length of 15 cm long line inserted. We think that the catheter had taken an abnormal route through a tributary of the external iliac vein. We considered possible factors contributing to the late appearance of signs and recognition of the problem. Upon insertion, we were able to aspirate blood back freely into the line before the line was secured. At some point in time the line must have perforated the vein. The high concentration of TPN in a small vein probably increases the risk of vessel erosion. The TPN solution used in this present case was a standard strength comprising of 10% dextrose with 40 g aminoacids/1000 mL and 20% intralipid with electrolytes and trace elements of unremarkable concentrations. The parenteral nutrients had extravasated, but abdominal signs only appeared 4 days after its insertion. It is likely that absorption of TPN by the peritoneal cavity allowed maintenance of normoglycaemia and contributed to the slow appearance of abdominal distension. This contrasts with the case reported by Baker et al. where hypoglycemia was a problem3.

In conclusion, we should confirm the tip of long lines ideally using a contrast medium. If the appropriate amount of contrast is injected, the catheter would be outlined and a flare of contrast can be seen leaving the tip. If there is coiling of the long line, as in our case, the line should be pulled back until the lengths of coils are straightened. This will reduce the risk of migration. A long line has too much slack within these coils and could be pushed further along if the coils spontaneously straighten. This may occur with active movement of the limb. If a repositioning of the line is considered too difficult or the position of the tip is in doubt, the long line should be removed completely.


The authors thank Dr HT Wickramasinghe and Mr Chua Hock Beng of RIPAS Hospital for their advice in this case.