Introduction
Intestinal fistulae produce fluid and electrolyte deficits, septic complications, and metabolic and nutritional disorders. They are classified as internal when they communicate with another viscus and as external when they drain onto the skin of the abdominal wall. Internal fistulae are usually asymptomatic unless they communicate between high and low segments of bowel. Symptomatic internal fistulae often present with malnutrition and will not close spontaneously. Therefore, all symptomatic internal fistulae require operative repair. The management of external fistulae, on the other hand, is not as clear-cut and external fistulae occur in clinical situations requiring difficult decision-making with good judgement.
Over the past five months Surgical Firms A and B have had 7 cases of intestinal fistulae. Two were internal, one associated with an underlying colonic malignancy and the other with ileocolic Crohn’s Disease. Both internal fistulae were treated by resection and anastomosis and both did well.
Postoperative external fistulae are catastrophic, probably moreso for surgeon than patient. Five external fistulae were postoperative complications and fortuitously none had the primary operative procedure performed here at the University Hospital. They were essentially small bowel fistulae and will be discussed further.
Cases
Case I
This 41 year-old male patient attended a private hospital with a bleeding duodenal ulcer confirmed on endoscopy. He had truncal vagotomy and pyloroplasty but ten days later rebled. He then had partial gastrectomy and Billroth II reconstruction. Nine days after gastric resection he developed a low output duodenal fistula and was transferred to the University Hospital where he received 7 days of parenteral nutrition with, on average, 2,000 calories per day. After 9 days his fistula closed spontaneously.
Case II
This 41 year-old lady had two previous Caesarian sections followed by hysterectomy and incidental appendicectomy. She attended a rural hospital with lower abdominal pains and an ultrasonographically detected right ovarian cyst. She had laparotomy with lysis of adhesions and right ovarian cystectomy. Inadvertently injured small bowel was resected, anastomosed and repaired. Postoperatively she developed intestinal obstruction and on day 8 had relaparotomy with lysis of adhesions and repair of new small bowel injuries. Postoperatively she was treated for a pneumonia. Six days after relaparotomy she was noted to have a wound infection and a distal ileal fistula. She was then transferred to the University Hospital where she tolerated an adequate oral diet. Her fistula closed spontaneously within 11 days.
Case III
Forty-five year-old L.S. had laparotomy at a rural hospital where he was found to have a perforated duodenal ulcer with severe intra-abdominal infection. He had drainage of abscesses and patch closure of the perforated ulcer. Postoperatively he went into respiratory failure with right lung consolidation and collapse. He was transferred to the University Hospital for ventilatory support which he had after tracheostomy. Five days after laparotomy he had a burst abdomen and a duodenal fistula. The duodenal fistula could not easily be exposed. A sump drain was therefore sited in the right upper quadrant and the abdominal incision was re-sutured. Fluid and electrolyte balance was maintained and parenteral nutrition was given for 39 days with on average 1,920 calories per day. Subsequently the patient developed bilateral subclavian vein thrombosis and parenteral nutrition was discontinued. The high fistulous drainage did not decrease and at laparotomy, six weeks after admission, the patient was noted to have a duodenal fistula, two jejunal fistulae and a transverse colon fistula all draining into a common tract. The duodenal fistula was intubated. The two fistulated loops of jejunum were mobilized, the lateral fistulae closed and the loops laid away from the fistulous tract. The transverse colon fistula was exteriorized in the incision. An inadvertent laceration to the jejunum was repaired. A distal jejunostomy was sited with a feeding catheter brought out to one side of the incision.
After 6 days of enteral nutrition intestinal fistulae recurred. Jejunostomy feeds refluxed through the fistulae and enteral nutrition was discontinued. The incision disrupted and the deep fistulae were treated thereafter by the “open” laparostomy method. Nine days later a severely malnourished patient faded away.
Case IV
One year before presentation this fifty-six year-old gentleman sufferred large bowel obstruction for which he had sigmoid colostomy and subsequent closure. On his recent admission he had recurrence of presumptive rectosigmoid obstruction. A sigmoid colostomy was refashioned and injured small bowel was repaired. Eleven days later he had relaparotomy for intestinal obstruction. The small bowel was again injured and repaired, and an obstructed small bowel loop was bypassed. Four days later he developed a low-output ileal fistula with wound disruption. He was then transferred to the University Hospital where he received parenteral nutrition supplemented by oral feeding. After 28 days the ileal fistula closed spontaneously only to be followed three days later by an exposed jejunal fistula in the granulating wound. Although 52 days of parenteral nutrition with 2,800 calories per day, and 36 days of jejunostomy and oral supplementation have been given, the jejunal fistula has failed to close after 11 weeks. In the interim there has been one documented episode of central venous catheter sepsis. Presently the patient is on the ward receiving jejunostomy and oral feeds.
Case V
This 48 year-old male patient attended hospital with generalised peritonitis. At operation his appendix was removed but ten days later he went into septic shock with generalised suppurative peritonitis and a blown appendix stump. At relaparotomy a tube caecostomy was sited. His intra-abdominal sepsis persisted and he developed a caecal fistula. At relaparotomy six days later he had a right hemicolectomy with ileotransverse anastomosis. Inadvertently injured small bowel was repaired. Right upper quadrant and pelvic drains were sited. One week later the patient had a partial wound disruption, a wound infection and a high output intestinal fistula. Two weeks after diagnosing the fistula the patient was transferred to the University Hospital for further management. Parenteral nutrition lasted 29 days with an average daily delivery of 2,000 calories. This was supplemented with a high-protein, high-calorie oral diet. One month after admission the patient began bleeding via the fistula. He also had gastrointestinal bleeding. He was transfused and remained stable but his high fistulous drainage persisted. Two days later he was found pulseless. A postmortem examination was not performed.
Discussion
Possible misjudgements and technical errors
All cases of inadvertent small bowel injury may be considered technical errors. The more often relaparotomy is performed after obliterative peritonitis the more likely are bowel injuries to occur. These injuries will often lead to fistula formation.
In Case III the duodenal fistula should have been intubated at the second laparotomy. At this time a gastrostomy and jejunostomy should also have been sited. The multiple intestinal fistulae of this case were most likely produced by the suction of sump drainage in the depts of the right upper quadrant. The end of the suction catheter should be placed in the fistulous tract as superficially as possible. At the third laparotomy the fistulated jejunum and colon should have been resected and the ends exteriorized or defunctioned away from the incision. After failure of subclavian and internal jugular catheterisation due to vein thrombosis the femoral or saphenous veins could have been catheterised utilising the same subcutaneous tunnelling technique.
Case IV has no definitive diagnosis accounting for recurrent large bowel obstruction. Does this patient have Ogilvie’s syndrome? Whatever is the underlying cause of his large bowel obstruction sigmoid colostomy at best provides only temporary decompression. Gastrointestinal contrast studies and lower gastrointestinal endoscopy are indicated. The anatomical characteristics of the exposed jejunal fistula in this patient make spontaneous closure unlikely. The fistulous orifice is larger than one cm and has everted mucosa. Operative intervention will be required at an opportune time. This may be as early as 4 weeks after resolution of all intra-abdominal infection. Delaying surgical repair after 4 to 6 weeks can only be justified if the patient’s nutritional status continues to improve on enteral nutrition. Operative procedures in this case may be staged or definitive. In a preliminary laparotomy the fistula may be defunctionalized or exteriorized away from the incision.
Case V appears to have had an inappropriate primary operative procedure for what may have been a perforated transverse colon carcinoma. At the third operation a right hemicolectomy with primary ileotransverse anastomosis in the presence of intra-abdominal sepsis was a gross error of judgement. The bowel ends should have been exteriorized. Then after four weeks of parenteral nutrition the fistulous drainage showed no sign of diminution. The late haemorrhage into the fistula was most likely due to vessel erosion by the effluent. These are two indications for operative intervention. If haemorrhage was this patients fatal complication it was a preventable cause of death.
Conclusion
The management of intestinal fistulae at the University Hospital can improve. With mastery of the local care and control of difficult fistulae, with aggressive surgical intervention when indicated and with enteral nutrition when possible and parenteral nutrition supplementation delivered via properly cared and subcutaneously tunnelled central venous catheters, more patients will heal their fistulae and hopefully more will survive.
Acknowledgement: Dr. Hugh Barned helped with the review of the patients above.
SurgicalInSite 