Alfred E. Stillman M.D. Robert Painter M.D. and
David W. Hollister M.D.
Day Kimball Hospital, Putnam, Connecticut

Ehlers-Danlos syndrome type IV is a heritable disease of type II collagen metabolism. This diagnosis is suspected in a patient with a combination of clinical manifestations and family history, but it is confirmed only by culture of the patient's skin fibroblasts and demonstration of a defect in type III collagen metabolism. The disease may rarely present with spontaneous colonic perforation, a complication traditionally treated by primary closure of the perforated segment and creation of an end colostomy. Attempts at bowel reanastomosis have often resulted in repeated colon perforations. We present the first patient with Ehlers-Danlos type IV syndrome to develop a colon perforation proximal to an end colostomy, and discuss the surgical strategy to prevent recurrences of this and other post-operative complications associated with the syndrome.

Introduction
Ehlers-Danlos syndrome (EDS) is a heterogeneous group of inherited diseases of which only three (EDS types IV, VI, and VII) have been shown to be due to defects of collagen metabolism defects, but there are no conclusive data to support this inference. Patients with EDS type IV (ED IV), a disease of collagen type III metabolism, have the most severe manifestations, the most dramatic of which are rupture of major arteries and hollow internal organs at a relatively early age (2). Spontaneous bowel perforation is a rare but particularly troublesome complication of ED IV because of its tendency to recur despite aggressive surgical management.

We present a unique patient with documented ED IV and multiple episodes of colon perforation, the last of which occurred proximal to a protective end colostomy. This is the first reported instance of perforation proximal to such a colostomy in ED IV patients. This patient's experience forms the basis for a discussion of the recognition and diagnosis of ED IV and new recommendations for the surgical management of spontaneous colon perforation in these patients.

Case Report
A 32 year old man was hospitalized with sudden onset of left lower quadrant pain which spread rapidly over his entire abdomen. On physical examination, there was marked guarding and rebound tenderness; bowel sounds were absent. His white blood count was 26,000 with a marked left shift. An abdominal film demonstrated free intraperitoneal air. Laparotomy revealed a sigmoid perforation for which sigmoid colectomy and end colostomy were performed, and a Hartmann pouch was created. Examination of the colon at surgery and by the pathologist revealed that the perforation had occurred in previously normal bowel. The colostomy was reanastomosed 2 months later. Surgery was complicated by excessive bleeding at the site of colostomy closure. The same symptoms and physical signs recurred 2 years later. Repeat laparotomy revealed another sigmoid perforation, for which, again, a sigmoid colectomy and end colostomy were performed and a Hartmann pouch created. Again, perforation had occurred in apparently normal bowel. This time there was no plan to close the colostomy.

Because of the patient's additional history of clubfoot at birth (previously corrected) and two spontaneous pneumothoraces at ages 17 and 30 years, the possibility of EDS was considered. The diagnosis was confirmed when the patient's cultured skin fibroblasts were shown to have intracellular accumulation and modestly decreased secretion of type III collagen. Type I collagen metabolism was normal.

He was seen again 1 year later with sudden onset of severe abdominal pain, rigidity, and rebound tenderness; his colostomy would not admit the tip of the examiner's fifth finger. At surgery, a perforation 13 cm from the end colostomy was found, and a left colectomy and colostomy revision were performed. However, the patient was re-explored several hours later when he became hypotensive. Massive intra-abdominal bleeding due to multiple small arterial bleeding points was found and repaired. Three months later, severe abdominal pain and vomiting led to re-exploration, with discovery and lysis of massive adhesions producing small bowel obstruction.

None of the patient's bower perforations had been traumatically produced. He denied a history of gay or bisexual behavior and had not inserted foreign objects into his rectum or colostomy. There was no family history suggestive of EDS. Interim physical examination had revealed uniformly thin skin with unusually prominent veins but normal skin extensibility. The small joints of his hands showed minimal hyperextensibility but other joints were normal.

Discussion
The diagnosis of ED IV depends on a combination of clinical manifestations, family history, and demonstration of quantitative and/or qualitative defects in type III collagen metabolism. In common with other EDS types, ED IV patients have skin and joint manifestations; however, they are not generally as severe as those seen in the majority of other EDS subjects. Unlike other EDS types, patients with ED IV do not have hyperextensible skin, and only the small joints of their hands are hyperextensible. However, the skin of ED IV patients is particularly thin, permitting visualization of the venous pattern over the anterior trunk. Easy skin bruisability is also characteristic of ED IV. The association of both club foot and spontaneous pneumothorax with ED IV has also been noted.

The diagnosis of ED IV is often difficult and may be made especially so by the appearance of acute abdominal pain due to major complications of spontaneous rupture of major arteries, spontaneous bowel perforation, and rarely, rupture of the uterus near term in pregnancy. Arterial rupture is rare before the third decade, whereas uterine rupture generally doesn't occur before the fourth decade. Bowel perforations usually occur between the late teens and early forties, but rare cases have occurred in childhood. The colon has been involved almost exclusively; most cases have occurred in the sigmoid. Ominously, perforations often recur after either simple closure or after reanastomosis of a colostomy performed to defunctionalize the perforated bowel segment. This tendency has resulted in recommendations to maintain a permanent colostomy or eventually to perform a subtotal colectomy with construction of an ileoproctostomy or cecoproctostomy (to preserve the ileocecal valve) because spontaneous rectal perforations in ED IV have been rare. Stool softeners have also been considered an important therapeutic adjunct.

Our patient is the first reported instance of spontaneous colon perforation in ED IV that has occurred with a colostomy in place. This perforation apparently occurred due to pressure proximal to a stenotic colostomy stoma and was probably unrelated to ED IV per se. However, an end colostomy can no longer be considered a guarantee against colon perforation in ED IV patients and must always be watched for developing stomal stenosis. Perhaps the most simple and safe life-saving surgical procedure would be a one-stage subtotal colectomy with construction of a Hartmann pouch or mucus fistula and permanent ileostomy. Should diversion colitis appear in the excluded rectosigmoid segment, a mucus fistula would offer a convenient stoma for therapy with short chain fatty acids or other agents. More extensive surgery, such as initial primary closure of the perforated segment and end colostomy construction followed by subtotal colectomy and ileoproctostomy, might increase in these patients the already formidable incidence of excessive bleeding disruption of sutures, and wound dehiscence without increasing their benefit. The risk of dehiscence can be decreased by use of stay sutures placed at a distance from the incision and by decreasing intra-abdominal pressure (i.e., prophylaxis against ileus, cough, bladder outlet obstruction).

ED IV is genetically heterogeneous. Most subjects have an autosomal dominant inheritance; recessive inheritance is rare. However, in common with several other lethal genetically transmitted conditions manifest at a young age, spontaneous mutations may occur too. Ultimate proof depends on culture of the patient's skin fibroblasts and demonstration of defects in the steps of type III collagen synthesis and secretion.

return to top