Esophageal leaks: extending our toolbox?

Traumatic or iatrogenic esophageal leaks are always associated with high morbidity and mortality. Besides intensive medical support and administration of broad-spectrum antibiotics, until the 1990 s surgery was almost the only solution for treatment of large leakages. One could divide the treatment principles into two categories, the first goal being to ensure closure of the defect. This could be done either directly through primary repair, if necessary reinforced by tissue interposition as muscle or gastric fundoplication, or through a delayed reconstruction after diversion (cervical esophagostomy and gastro/jejunostomy). The second goal was to ensure cleaning of the mediastinal or pleural space through surgical debridement and drainage.

However, surgery is associated with high morbidity and mortality, particularly in patients with a delayed diagnosis and mediastinal or pleural contamination [1]. Over the last two decades, endoscopy has taken a place in the armamentarium for managing such situations. It offers several advantages compared with surgery: reduced need for prolonged anesthesia, direct visualization of the leakage and, potentially, organ preservation with less morbidity. Many case reports or case series are now available to demonstrate the feasibility of several endoscopic approaches for leakage treatment.

Most of these endoscopic treatments were designed to fulfil the first rule of surgical treatment, namely closure of the defect. Insertion of a self-expandable esophageal stent across the leakage region was the first and is currently the most popular technique that avoided surgery. It enables diversion of the esophageal content from the wound cavity with the secondary intention of promoting closure of the leakage. The first cases reported the use of self-expandable metallic stents (SEMSs), and were limited to patients unfit for surgery [2] [3]. The main reason for this was concern about the removability of the stents, since they are difficult or impossible to remove when their ends are deeply embedded in the esophageal wall. This problem was solved with the advent of self-expandable plastic stents (SEPS) that are designed to be easily removed. Although several case series demonstrated their efficacy for leakage closure [4] [5], they tend to migrate and are not very effective when used alone with large leakages. However as our group has reported previously [6], they may also be used to facilitate extraction of metallic stents by inducing pressure necrosis of hyperplasia developed in SEMS. Indeed, in trickier situations such as post-bariatric surgery leakages, partially covered SEMS are still a better option, offering greater watertightness thanks to development of hyperplasia. Besides covering the wall defect, stents may also help to treat a distal stenosis when present and responsible for persistence of the fistula.

Clipping has also been tried for direct closure of esophageal leakage in the case of small defects (less than 2 cm). As in surgical procedures, clipping should preferably be attempted on fresh perforations such as iatrogenic leaks [7]. Placement of clips is still technically challenging and often has to be repeated. The limited area offered by the esophageal cylinder makes their tangential application more difficult that in the other intestinal cavities. Use of larger clips such as the new over-the-scope clip (OTSC) [8] could reduce the number of clips that must be placed, and the dedicated forceps presented with this device may help to overcome the difficulties in apposing the two edges of the leakage .

Successful use of sealants such as fibrin glue or Surgisis in order to plug the leakage has also been reported for closure of esophageal fistula. Endoscopic sealants are generally reserved for nonacute situations with no active infection and without distal esophageal obstruction. Cyanoacrylate, in our opinion, should be avoided in such cases, since it may induce a foreign-body reaction and promote further infection.

Few endoscopic series have addressed the second goal of treatment, regarding mediastinal drainage. In some situations, as with a fresh iatrogenic defect, this aspect of treatment is unnecessary and a broad-spectrum antibiotic will be prescribed to reduce the risk of mediastinal infection. In other cases, as in a post-surgical situation, drainage is ensured by the surgical drain. It has to be kept in mind that some fistulas are maintained by the traumatic effect of transthoracic drain extremities, which should always be pulled back or removed after endoscopic closure. That being said, the external drainage may be inefficient or even nonexistent. In these situations it is pointless to close the defect without ensuring the external or internal drainage of infected cavities. As the placement of transthoracic drains in the mediastinal space can be challenging, other than at the time of thoracic surgery, an endoscopic solution that dealt efficiently with this part of the treatment would certainly be most welcome.

The concept of transluminal drainage of the mediastinal or pleural cavity had already been described during the 1990 s by two surgical teams [9] [10] who both reported a good success rate for abscess healing in a small series of postsurgical esophageal leaks. The treatment was limited to the insertion of a nasogastric tube through the esophageal defect into the cavity. In the past five years, beginning with post pancreatitis necrosectomy, the endotherapeutic approach to periluminal abscesses has acquired another dimension with the insertion of the endoscope into cavities [11] [12]. In this issue of Endoscopy, Ahrens et al. report their experience of treating esophageal leakages by endoscopically inserting polyurethane sponges into the thoracic wound cavity after endoscopic exploration and cleaning. They were inspired by rectal abscess treatment and two esophageal case reports, and describe outcomes in five patients presenting with major anastomotic or iatrogenic leakage. After insertion into the cavity, sponges were connected via a nasal drainage tube to a device providing continuous suction. In this way the infected cavities were continuously drained by means of the sponges, presumably avoiding the trauma risk of a rigid plastic drain and also stimulating granulation. With a median of 9 sponge exchanges, with the sponges cut to the cavity size each time, and a median of 28 days of treatment, the authors demonstrated a progressive reduction in size of the wound cavity along with an improvement in inflammation markers and with clinical recovery.

Despite the large volume of some cavities, and leaks extending to up to one third of the anastomosis circumference, all five patients were treated successfully without the need for surgical re-intervention. Two patients developed anastomotic stricture, and at the time of dilation one of them died of major bleeding from an aortoesophageal fistula at the site of the esophagogastric anastomosis.

This paper highlights the role of infection control and pus evacuation. In these patients, wound closure, is not the priority, and because of possible difficulties in providing adequate surgical drainage, designing a transluminal drainage is justifiable. Although it is too early to draw conclusions about the role of the polyurethane sponge, it is conceptually interesting that possible injury in this area, that is loaded with large vessels and major organs, might be reduced by the interposition of a soft harmless sponge. This device might also stimulate the surrounding tissue for granulation and secondary healing. Nevertheless this remains a critical situation as shown by the fatal bleeding that occurred after stricture dilation, and these treatments should always be discussed in multidisciplinary teams.

Endoscopic treatment of intestinal wall disruption and infected paraintestinal cavities is at present only an experimental solution. Several case series demonstrate its feasibility. Nevertheless, a Boerhaave syndrome with a late diagnosis and associated with empyema in an elderly patient is clearly different from a post-esophagectomy anastomotic leak and even more different from an iatrogenic perforation after mucosal resection or endoscopic submucosal dissection (ESD). Factors influencing the treatment option are the delay between diagnosis and treatment, size of the breach, presence of a stenosis, and presence of infected cavities [13]. Each situation will have different requirements and need a different approach. This heterogeneity justifies a broad choice of tools. The different devices are clearly complementary rather than competitive. While stenting may often be an appropriate solution, in some situations direct closure could be required or, on the other hand, a more progressive closure with transluminal drainage. The precise role of the new device demonstrated by Arhens et al. will need further clinical evaluation. But yes, the attempt to extend our toolbox is still justified. Nevertheless, a multidisciplinary approach and specific expertise will remain the pillars of best clinical management for these high-risk situations.

Competing interests: None

References

• 1 Brinster C J, Singhal S, Lee L. et al . Evolving options in the management of esophageal perforation.  Ann Thorac Surg. 2004;  77 1475-1483
• 2 Dumonceau J M, Deviere J, Cappello M. et al . Endoscopic treatment of Boerhaave’s syndrome.  Gastrointest Endosc. 1996;  44 477-479
• 3 Siersema P D, Homs M Y, Haringsma J. et al . Use of large-diameter metallic stents to seal traumatic nonmalignant perforations of the esophagus.  Gastrointest Endosc. 2003;  58 356-361
• 4 Gelbmann C M, Ratiu N L, Rath H C. et al . Use of self-expandable plastic stents for the treatment of esophageal perforations and symptomatic anastomotic leaks.  Endoscopy. 2004;  36 695-699
• 5 Schubert D, Scheidbach H, Kuhn R. et al . Endoscopic treatment of thoracic esophageal anastomotic leaks by using silicone-covered, self-expanding polyester stents.  Gastrointest Endosc. 2005;  61 891-896
• 6 Eisendrath P, Cremer M, Himpens J. et al . Endotherapy including temporary stenting of fistulas of the upper gastrointestinal tract after laparoscopic bariatric surgery.  Endoscopy. 2007;  39 625-630
• 7 Qadeer M A, Dumot J A, Vargo J J. et al . Endoscopic clips for closing esophageal perforations: case report and pooled analysis.  Gastrointest Endosc. 2007;  66 605-611
• 8 Voermans R P, Deprez P, Le Moine O. et al . Endoscopic closure of iatrogenic perforations of the gastrointestinal tract using the over-the-scope-clip: a prospective multicenter human trial.  Gastrointest Endosc. 2010;  71 AB132-AB133
• 9 Infante M, Valente M, Andreani S. et al . Conservative management of esophageal leaks by transluminal endoscopic drainage of the mediastinum or pleural space.  Surgery. 1996;  119 46-50
• 10 Jorgensen J O, Hunt D R. Endoscopic drainage of esophageal suture line leaks.  Am J Surg. 1993;  165 362-364
• 11 Charnley R M, Lochan R, Gray H. et al . Endoscopic necrosectomy as primary therapy in the management of infected pancreatic necrosis.  Endoscopy. 2006;  38 925-928
• 12 Seewald S, Groth S, Omar S. et al . Aggressive endoscopic therapy for pancreatic necrosis and pancreatic abscess: a new safe and effective treatment algorithm (videos).  Gastrointest Endosc. 2005;  62 92-100
• 13 Swinnen J, Eisendrath P, Rigaux J. et al . Self-expandable metallic stents (SEMS) for the treatment of benign upper GI leaks and perforations.  Gastrointest Endosc. 2010;  71 AB133