The Mysteriously Bleeding Ileostomy: A Lesson in Too Much Endoscopy and Not Enough Looking?

Wai-Kit Lo, MD; David L. Carr-Locke, MD, FRCP


January 20, 2010

Introduction and Case Presentation


Defining the origin of obscure gastrointestinal bleeding can be a frustrating task. Previous gastrointestinal surgery introduces potential sources of bleeding that can be overlooked. In this case report, a patient presented with hematochezia and melena via an ileostomy stoma and was thoroughly evaluated for a source of bleeding using multiple endoscopic techniques, but the diagnosis was eventually made by eye without the use of an endoscope.

Case Presentation

A 71-year-old woman with a history of ulcerative colitis, colon cancer treated by total colectomy with ileostomy in 1968, cryptogenic cirrhosis, and coronary artery disease treated with bypass (and currently taking aspirin and clopidogrel) presented with iron-deficiency anemia in 2001. From 2001 to 2006, she underwent upper endoscopy, enteroscopy, video capsule endoscopy, and colonoscopy for evaluation of anemia and intermittent hematochezia and melena via her ileostomy. Endoscopic evaluation revealed no source of bleeding.

In 2008, the patient returned with complaints of bright red bleeding from her ileostomy that occurred intermittently (every 48 to 72 hours). Over the course of a month, 3 colonoscopies, an upper endoscopy, and a video capsule endoscopy were performed in response to multiple episodes of bleeding that required several blood transfusions. The major findings were nonbleeding grade I esophageal varices, portal hypertensive gastropathy, and a single small jejunal ulcer, all of which were unlikely to be contributing to the patient's symptoms.

Prolapse of the ileostomy site was noted, as well as peristomal blood, but the ileostomy mucosa appeared healthy. Further workup, including abdominal CT scan, did not reveal a source of the bleeding. However, targeted abdominal transcutaneous ultrasound at the stoma did reveal small periportal varices and a large peristomal vein. No varices were evident on endoscopic ultrasound (EUS) performed at the stoma following an episode of bleeding. However, repeat peristomal transcutaneous ultrasound the next day showed large peristomal veins up to 5 mm in diameter (Figures 1a and 1b).

Figures 1a and 1b (inset). Ileostomy. The inset of a transcutaneous ultrasound shows small peristomal veins (colored areas).

The negative EUS was attributed to variceal decompression following a recent active bleeding episode. Careful evaluation of the skin surrounding the stoma site revealed dilated subcutaneous vessels, but, more importantly, there were also many prominent veins on the mucosa of the ileostomy itself. These were considered to be the source of bleeding because they would explain the pattern and presentation of her bleeding episodes (Figures 2, 3a, and 3b).

Figure 2. Prolapsing ileostomy showing blood on the superior mucosal surface.

Figure 3a. Prominent veins on ileostomy.

Figure 3b. Closer view of ileostomy in Figure 3a.

A decision was made to inject Dermabond® (N-butyl-cyanoacrylate; Ethicon, Inc., New Brunswick, New Jersey) under direct vision with a 1-mL syringe and a 25-gauge needle until the vein became firm to digital palpation (Figure 4). The cycle of frequent recurrent bleeding immediately stopped.

Figure 4. Manual injection of cyanoacrylate into the ileostomy mucosal veins.

Since 2008, over an 18-month period, the patient has undergone 13 sessions of cyanoacrylate injection into the ileostomy mucosa. No serious complications have occurred, and the patient has tolerated the procedures with minimal discomfort. Treatments were supplemented occasionally with argon plasma coagulation to achieve hemostasis at the stoma-cutaneous junction when glue injections caused bleeding ulceration (Figures 5 and 6). The patient continues to be retreated as needed for new mucosal veins.


Figure 5. Argon plasma coagulation applied to bleeding ulceration caused by glue injection at stoma-cutaneous junction.

Figure 6. Healing site after argon plasma coagulation as shown in Figure 3.