Endoscopic Management of Complications

Endoscopic treatment of complications after Gastrointestinal Surgery

Gastrointestinal operations generally result in a suture or staple line. The primary acute complications of these repairs are leak, bleeding, or stricture. Emerging endoscopic techniques now allow treatment of these complications to avoid reoperation or to simplify management.The greatest impact of endoscopic techniques has been made in anastomotic or resection line leaks following upper gastrointestinal surgery. The technology that has allowed this is the development of self-expanding covered stents. These stents have a metal or plastic mesh material that can be compressed to a small diameter and then expanded to a predetermined length and diameter. This allows accurate placement under fluoroscopic control. The walls of the stents are covered with silicone allowing complete control of the anastomotic leak. The stent can be removed once healing is thought to be complete. Currently available covered stents in the United States are made of polyester (Polyflex, Boston Scientific, Inc.) or Nitenol (Alveolus, Inc.).
Anastomotic leaks following bariatric surgery often require reoperative management and specialized nutritional support [1]. Endoscopic stents allow oral nutrition to be taken while healing is taking place. Our group has been using these stents for anastomotic leaks following Roux-en-Y gastric bypass (RYGB) since January 2006. Between January 2006 and June 2007 we treated 11 patients for acute anastomotic leaks with covered stents [2]. All leaks eventually closed although one of the leaks took 40 days to heal. The duration of stent placement was 25 days. In addition, two chronic anastomotic enterocutaneous fistulas were treated with stents resulting in healing of one of the fistulas at 45 days.

The mean procedure time for stent placement at our institution is 47 minutes. A standard sized upper endoscope is used to identify the area of leak or stricture. The site of the leak and the Z-line are identified and marked with external metallic markers. A flexible wire is then placed through the scope well into the Roux limb. The scope is then removed and the stent delivery system is placed over the wire and positioned under fluoroscopy. As our series progressed we placed stents that were larger in diameter and usually placed two stents to decrease the rate of migration. Under fluoroscopy the first stent is placed so its distal end is 3-4 cm beyond the leak. A second stent is placed more proximal but overlapping the first stent and extending 3-4 cm above the Z-line. The primary complication seen with stents is migration. Most migration is over short distances and the stents can be retrieved endoscopically or passed through the GI tract. We have had 3 of 34 stents migrate into the small bowel and require laparoscopic retrieval because of failure to advance.

Endoscopic stents greatly simplify the management of anastomotic leaks after bariatric procedures. Similar results have been found for the treatment of esophageal anastomotic leaks and perforations [3]. Early bleeding after gastrojejunal anastomosis can also occur following RYGB. This is generally self limiting and can be treated with observation. Occasional patients develop hemorrhagic shock or require multiple blood transfusions and more aggressive treatment is warranted. Endoscopic treatment of the gastrojejunal bleed is usually successful [4]. While data is sparse it is my
opinion that endoscopic clips are the treatment of choice for discrete anastomotic bleeding. The use of cautery or epinephrine would in theory increase the risk of anastomotic ischemia and leak. The final anastomotic complication seen after upper GI repair is stricture. This usually does not occur immediately but 4-8 weeks following surgery. It is relatively common, occurring in 5-10% of patients after RYGB. These patients usually present with nausea and vomiting and EGD is the diagnostic test of choice. If a stricture is found, dilation up to 2x the diameter of the stricture is
generally safe. A final diameter of 15-18mm is usually sufficient to eliminate symptoms. If an acute ulcer is present, aggressive treatment of the ulcer should be done before dilation of the stricture. Stricture dilation is generally effective but a 2-4% risk of perforation is present [5].

Occasionally strictures do not respond to repeated dilation. In the past this could only be corrected with revisional surgery. We have used stent treatment of seven refractory gastrojejunal strictures [2]. The stents are often poorly tolerated due to pain and the mean duration of stent placement was only seven days. However, our early results were encouraging with 6 of 7 having improved food toleration at a mean of 2.3 months. Longer follow up is needed for definitive recommendations. The use of stents for refractory benign esophageal strictures has been reported [6]. Stents in this situation can be helpful. However, the complication rate is high (severe chest pain 11%, bleeding 8%, perforation 6%). Careful selection of patients is required.
Colonoscopic treatment of anastomotic complications is much less developed than in upper GI surgery. We have conducted a randomized trial in pigs showing better results with a stent in an anastomotic leak model [7]. There is at least one case report of a successful treatment of an anastomotic leak with a covered stent [8].The literature on stent placement for benign colonic strictures is also quite limited. A recent review of the literature documented 63 reports using both uncovered and covered stents for colonic stricture [9]. Stent placement is generally possible and can often be helpful. Uncovered metal stents appear to have a higher complication rate and are not recommended. Migration continues to be the primary complication for covered stents. Techniques to avoid stent migration need to be developed.