Grade 1: Any deviation from the normal postoperative course without the need for pharmacological treatment or surgical, endoscopic, and radiological interventions
Grade 2: Requiring pharmacological treatment with drugs other than those allowed for grade I complications. Blood transfusions and total parenteral nutrition are also included
Grade 3: Requiring surgical, endoscopic, or radiological intervention
Grade 4: Life-threatening complication (including CNS complications) requiring IC/ICU management
Grade 5: Death of a patient
Anastomotic leaks can be classified by the following criteria:
Localization [upper versus lower GI (gastrointestinal) tract]
Type of previous operation and anastomotic technique (e.g., esophageal resection with gastric conduit, gastrectomy, rectal resection)
Time point of postoperative diagnosis: acute versus chronic leak
Size of leak (given in percent of circumference)
Presence or absence of a leak cavity
Severity of the complication (Clavien–Dindo classification, ◘ Table 6.1)
Therapeutic Algorithms: Operation, Endoscopy, or Conservative Management?
The severity of the complication and the condition of the patient determine if endoscopic therapy is an option for the management of a postoperative leak. With regard to the Clavien–Dindo classification, the domain of endoscopic therapy are grade 3 complications. These are anastomotic leaks which cannot be managed by parenteral nutrition, antibiotic therapy, and placing of a gastric tube alone. The patient is in a septic condition, but does not fulfill criteria for a grade 4 complication (organ dysfunction). A partial dehiscence of an esophagogastric anastomosis following esophagectomy with viable gastric conduit is a typical example. This is a situation in which endoscopic therapy has replaced surgical management in most cases and has become the gold standard. In this example, endoscopic stent placement is a well-defined standard procedure; recently, endoscopic vacuum therapy is increasingly used in such cases. A typical example of a grade 3 complication in the lower gastrointestinal tract is the rectal anastomotic leak with a leak cavity in the small pelvis with the presence of a diverting ileostomy. In this case, endoscopic vacuum therapy is the accepted standard therapy.
The success of endoscopic therapy combined with low morbidity and mortality of these procedures has even shifted the indications toward grade 4 complications. By now, critical patients with organ dysfunction and ICU therapy are managed by endoscopic means in selected cases. However, in these cases it has to be critically evaluated if endoscopic therapies are sufficient to manage the life-threatening sepsis. If the septic condition cannot be controlled by endoscopic means or if the local condition of the anastomosis is not suitable for endoscopic therapy (e.g., necrosis of the gastric conduit after esophagectomy), surgical management is still mandatory.
In case an endoscopic therapy is indicated, the choice of method depends on the abovementioned criteria, especially localization of the leak (stent or endoscopic vacuum therapy in the upper gastrointestinal tract, endoscopic vacuum therapy in the lower gastrointestinal tract), the local condition of the anastomosis, and the presence of an infected leak cavity. These aspects are discussed in the context of the respective endoscopic techniques.
6.2 Stent Therapy
M. Colombo-Benkmann
Indication, Evidence, and Significance of Endoluminal Stenting
Implantation of self-expanding endoprotheses, i.e., stents, is indicated in the treatment of leakages of esophagogastrostomies as well as esophagojejunostomies, perforations of the esophagus (van Boeckel et al. 2011), and leakages after bariatric operations (Puli et al. 2012). So far, stenting is not established in the therapy of leaks or perforations of the duodenum, jejunum, or ileum.
At present, stents are made of wires of alloys, e.g., nitinol with a memory effect, which are woven as a cylindrical mesh. Stents can be covered either partially or over their total length by a silicone sheet to prevent ingrowth of the mucosa into the mesh. Each end of the stent should contain a circular thread to allow stretching. This will result in simultaneous reduction of the diameter of the stent, enabling adjustments of its intraluminal position or its removal. The advantage of nitinol stents is that they can be implanted easily without the need for any preparation, in contrast to former endoprostheses made exclusively from plastic.
Further advantages comprise their ability to cover multiple leaks of suture lines such as after sleeve gastrectomy, as well as easy endoscopic removal and amendment in case of a misplacement.
The most common indications of stent insertion are anastomotic leaks after esophageal resections (51%), followed by iatrogenic perforations due to diagnostic or interventional endoscopy (25%) occurring during gastroscopic balloon dilatation or bougienage of stenoses after endoscopic mucosal resection or endoscopic retrograde cholangiopancreaticography, Boerhaave syndrome (17%), and benign fistula, e.g., to the trachea and bronchi (4%) (van Boeckel et al. 2011). Even anastomotic dehiscences comprising up to 100% of the luminal circumference can be treated successfully. This holds true also for leaks after bariatric surgery such as gastric bypass, gastric sleeve, and biliopancreatic diversions (Puli et al. 2012).
Covered stents represent a physical barrier between leakage and lumen, preventing contact of endoluminal secretions with the leak. This represents a crucial prerequisite of leak closure. In addition, patients can receive enteral nutrition 24–48 h after implantation, initially by a simultaneously implanted jejunal tube followed by natural ingestion of food. This prevents the necessity of parenteral nutrition and its associated complications (Puli et al. 2012).
Despite low levels of evidence due to the lack of prospective not to mention randomized studies and due to small patient cohorts, endoluminal stents are the gold standard in the treatment of postoperative leaks and fistula.
Requirements of Manpower, Instrumentation, and Organization
Implantation of intraluminal stents requires at least two, ideally three, persons with expertise in endoluminal stenting: the implanting physician and two assistants who are knowledgeable in the technique of implantation.
Instruments include a gastroscope and a stiff guide-wire with a flexible spiral tip (e.g., Eder–Puestow) which yields when coming into contact with the tissue. Due to the soft spiral tip, the risk of incidental perforation of the hollow organ reduces. The guide-wire should have a length of 200 cm.
Sterile warm water should be injected into the core of the delivery system, to ensure fast expansion of the stent once it is released. In our practice, stents are delivered under fluoroscopic guidance. Thus epicutaneous radiopaque pins, e.g., made from lead, are needed to mark the position of the leak, the eophageal introitus in case of leaks in the vicinity of the upper esophageal sphincter, the esophagogastric junction, or the pylorus depending on the location of the leak. If adjustment of the stent’s position after delivery is required, an alligator forceps should be used.
Ideally, stents should be implanted using a fluoroscopy system. This allows monitoring and documentation of the respective phases of stent implantation from the positioning of the leak until expansion of the stent in its final position.
Implantation is carried out in a supine position of the patient in analgosedation. Intubation should be used to avoid aspiration if there is significant reflux, in case of respiratory failure or fistulas. Constant monitoring of oxygen saturation is mandatory; electrocardiography should be used additionally in patients with cardiac failure or significant cardiac risk factors.
In general, patients who breathe spontaneously during stent implantation should be given oxygen continuously during the procedure by a nasal applicator. This can prevent decrease of oxygen saturation during the procedure. Since many patients already suffer from pre-existent cardiopulmonary morbidity, sedation may result in respiratory failure during implantation. As a consequence, the equipment for manual resuscitation such as a respiratory mask, a resuscitation bag with oxygen supply, and an emergency case with the possibility of endotracheal intubation are to be provided at the site of the procedure.
In case of extraction of the stent, disconnection of the grasping forceps from the stent on the level of the pharynx can result in acute respiratory obstruction, resulting in asphyxia. The attempt of immediate endoscopic extraction by an endoscopic forceps is unlikely to be successful. Instead, we recommend instant insertion of a laryngoscope as used for endotracheal intubation, for visualization of the stent, and a strong needle holder to enable immediate stent removal. Thus, we recommend having these instruments ready to hand.
With regard to selection of specific type of stent, it should be taken into account that in case of the treatment of leaks, it is recommended to remove stents 6 weeks after implantation. Fully covered stents have the advantage that they can be extracted generally without damaging the mucosa. In partially covered stents, there is a significant risk of ingrowth of the mucosa into the mesh. This impedes not only stent extraction, but can result in considerable trauma to the mucosa. Thus, single cases of leak caused by extraction of partially covered stents have been described.
On the other hand, fully covered stents have a significant risk of dislocation, due to their smooth surface allowing them to slide on the mucosa. Choosing an adequate diameter of the shaft and the ends of the stent may impede dislocation. In general, we use fully covered stents with diameters of the shaft of 25 mm and of both ends of at least 30 mm.
Procedure
At first a diagnostic endoscopy is carried out in the sedated or intubated patient. This should comprise not only the esophagus, but all parts of the digestive tract which can be reached by the gastroscope. Independent from the endoscopic verification of a leak, water-soluble contrast dye during fluoroscopy should be applied, since this enables reliably the detection of fistulas in the respiratory tract. The contrast dye is applied by a catheter, which is inserted into the gastroscope. After aspiration of the contrast dye, the leak is marked under fluoroscopy by epicutaneous markers (◘ Fig. 6.1). These have to be attached onto the patient’s skin by an adhesive tape. Markers should not be attached onto the patient’s clothing. If the leak is close to the upper esophageal sphincter, the latter should be marked in the same way as well, to avoid misplacement of the upper end of the stent, e.g., into the pharynx. When using stents in the treatment of a leak after sleeve gastrectomy, the pylorus should be marked as well, to ensure that the aboral end of the stents is positioned reliably beyond the pylorus.
Fig. 6.1
Procedure of stent implantation. a Anastomotic leak following esophagectomy with gastric conduit, located at 7 o’clock. b The level of anastomosis is indicated by an epicutaneous marker. c The stent is placed with the leak being located in the middle portion of the stent. d Endoscopic view on the upper opening of the stent. e Immediately after stent removal which was performed 6 weeks later: multiple erosions can be seen, and the leak is completely closed. f Endoscopic view 3 weeks later: all erosions have resolved, and the anastomosis is healed
In the meantime, warm sterile water is instilled into the delivery systems through a respective opening. Subsequently, a guide-wire (Eder–Puestow) is introduced through the gastroscope. The end of the wire is placed aborally to the intended position of the aboral end of the stent. The wire is secured by the assisting staff during retraction of the endoscope, to prevent an incidental dislocation of the wire. Very importantly, the risk of facial and eye injury by the extracorporeal end of the wire is to be considered.
The well-lubricated delivery system is introduced over the wire, and the stent is delivered in a way that leak or fistulas are covered with a sealing effect. After the stent has disconnected from the delivery system during its expansion, the latter is pulled out together with the guide-wire. Care has to be taken that the stent is not dislocated by this maneuver. If dislocation occurs or if the position of the stent has to be corrected, an endoscopic grasper should be used.
The final endoscopic exam should be limited to documenting the distance of the upper end of the stent from the front teeth in centimeters, to verify dislocation if it should occur. It is not necessary to intubate a stent not fully expanded, since this carries a high risk of dislocation. Exceptions are the necessary adjustments of the stent’s position.
If there is suspicion of a newly occurred stent dislocation, it can be verified by endoscopy or fluoroscopy. Insufficient sealing of the leak can be confirmed by fluoroscopy and water-soluble contrast dye. Occasionally if the diameter of the chosen stent is too small, the failure of sealing can be recognized by a gap between the stent and the hollow organ.
There is no need for a special follow-up if the patient is asymptomatic and receiving his habitual nutrition.
Stent removal is carried out by an endoscopic grasper pulling at the upper thread. If the stent is adherent to the inner layer of the stent due to mucosal overgrowth of the ends of the stent, these adhesions can be eliminated either mechanically by graspers or thermically.
Technical success of stent implantation with complete sealing of leaks and fistulas in non-bariatric patient is between 98% and 100% (van Boeckel et al. 2011).
In non-bariatric patients, the time the stent is left in place is 6 weeks on average, published times are between 3 and 17 weeks on average (van Boeckel et al. 2011) and between 6 and 8 weeks in bariatric patients (Puli et al. 2012).
If the interval chosen to leave the stent in place is too short, this will result in incomplete closure of the leak, while choosing an excessively long interval may result in stent migration or mucosal overgrowth by epithelial cell. This significantly impedes removal of the stent and contains a considerable risk of injury to the hollow organ. In addition to this, if too long intervals are chosen, this can lead to dysphagia.
Removal of fully covered stents in non-bariatric patients is almost completely without any complication. After removal of a partially covered stent, 8% of patients will experience complications (van Boeckel et al. 2011). In bariatric patients, successful stent extraction occurs in 92% of patients (Puli et al. 2012).
The objective of complete closure of leaks and fistulas solely by stents can be achieved in 85% of non-bariatric patients and in 88% of bariatric patients. In this context, successful treatment is defined by complete closure of leaks and fistulas as shown by fluoroscopy with contrast dye, if after stent removal no extraluminal contrast dye can be seen.
If leaks are persistent, re-stenting can be carried out without any problems in general.
Possible Complications and Treatment
Complications associated with implantation of stents such as intraluminal bleeding or perforation are rare and occur in only 3% of patients (van Boeckel et al. 2011).
Dislocation of the stent is one of the most common complications. In fully covered stents, this occurs in 26% and in partially covered stents in 13% of non-bariatric patients. In bariatric patients, the dislocation rate is 16% and 9% (Puli et al. 2012). In bariatric patients, stents can migrate into the jejunum. In such cases, surgery is required to remove the stent; occasionally, stents have been egested by defecation. If repositioning is not successful, a stent with a larger diameter should be chosen or another method is to be applied.
In contrast, partially covered stents are more often overgrown by the epithelium (12%) than fully covered stents (7%) (van Boeckel et al. 2011). As a consequence, the stent cannot be removed (Puli et al. 2012).
Endoscopic reinterventions are necessary in 26% of patients with fully covered stents and in 13% of patients with partially covered stents (van Boeckel et al. 2011). Occasionally, stents are obstructed by food (Puli et al. 2012). If possible, the bolus should be dislocated aborally, to be digested. If this is not possible nor indicated, extraction should be strived.
Surgical therapy is necessary in 13% of non-bariatric patients, since leaks do not close and due to complications associated with the procedure or the stent (van Boeckel et al. 2011). Mortality after stent implantation is due to septicemia associated with the leakage and not due to the endoprostheses. Its incidence in non-bariatric patients is 18% (van Boeckel et al. 2011).
Endoluminal stenting has been the standard treatment of the abovenamed complications for more than 10 years. However, new therapies are becoming available due to the technological progress in medicine. Thus, it can be expected that in the next few years, indications for specific treatment options will be specified, especially if they become more available.
6.3 Endoscopic Vacuum Therapy (EVT)
M.G. Laukoetter
Indication and Evidence
Vacuum therapy (endoscopic vacuum therapy (EVT), VacuSeal, vacuum-assisted closure [VAC] therapy, negative pressure wound therapy [NPWT]) for wound healing simply consists of a sponge-based drainage system connected to negative pressure, leading to decrease of bacterial contamination, secretion, local edema, and promotion of granulation tissue (Holle et al. 2007). Since introduction of this treatment technique in the early 1990s, endoscopic vacuum therapy, as an alternative treatment option for even desolate wounds in almost every localization, has been established in nearly all surgical disciplines (Argenta and Morykwas 1997). After initially being considered and established as a treatment modality for infected superficial skin defects of different sizes and extent (Argenta and Morykwas 1997; Vikatmaa et al. 2008), the first intracorporeal endoscopic vacuum therapy was established successfully for anastomotic leaks after rectal resection (Weidenhagen et al. 2008; Willy et al. 2006). The close proximity of the sphincter and of the anastomotic region in such cases leads to permanent congestion of infected secretion and intestinal gas, leading to potential severe local peritonitis in the pelvic region. In such cases where there is local lower abdominal peritonitis with an endoscopically accessible cavity, the Endo-SPONGE treatment can be applied (◘ Fig. 6.2). An overtube is placed into the cavity, and, after the endoscope has been withdrawn from the overtube, the sponge is brought down by a pusher. The cavity is drained subsequently by the endoscopically introduced Endo-SPONGE® system. The open pores of the sponge allow the suction to be transferred over all tissues in contact with the sponge surface.
Fig. 6.2
Endo-SPONGE® system (By courtesy of Braun Melsungen AG)
The insertion of a polyurethane sponge into the defect zone, connected transanally to an external vacuum system, does, in contrast to the treatment of superficial skin defects, not require the presence of an airtight sealing, since the pelvic wound cavity seals itself after start-up of the drainage system. Closure rates of >90% avoid reoperations in those patients characterized by a complicated postoperative course (Glitsch et al. 2008; Weidenhagen et al. 2008). Perforations and fistulas of the upper gastrointestinal (GI) tract occur as postoperative complications (anastomotic dehiscence or fistula), during diagnostic or interventional endoscopy, iatrogenic as a consequence of other therapeutic measures (e.g., gastric tube placement, percutaneous endoscopic gastrostomy, transesophageal echocardiography), or spontaneously (ulcers, tumors, Boerhaave syndrome, and others). These perforations often lead to severe septic conditions which are difficult to treat and give rise to a high morbidity and mortality, especially if leading to mediastinitis or peritonitis (Junemann-Ramirez et al. 2005). In particular, reported leak rates after esophagectomy vary widely from 1% to 30% (Ahrens et al. 2010; Whooley et al. 2001). Anastomotic leakage accounts for approximately 40% (Miller et al. 1997; Pross et al. 2000) of all postoperative fatalities and is highly challenging to treat: control of the septic focus is essential; thus, the already critically ill patient often requires intensive additional measures that themselves are associated with high morbidity, adding to the clinical burden (Junemann-Ramirez et al. 2005).
A number of competing treatment modalities ranging from conservative to surgical approaches are available for the management of this situation. The surgical treatment options include revision of the anastomosis, closure of the defect and perifocal drainage, or complete surgical deviation and creation of a cervical stoma. These procedures are usually difficult and carry a high risk for severe complications associated with high morbidity and mortality rates. Therefore reoperation is not always a reasonable option.
In this context, numerous minimally invasive treatment options have more recently become available to treat a variety of secondary surgical complications. Conservative management may be advantageous if reliable endoscopic methods are available. Endoscopic clips (Mennigen et al. 2013; Rodella et al. 1998), fibrin glue injection, absorbable plugs, and endoscopic suturing (EndoCinch) (Adler et al. 2001; Fritscher-Ravens et al. 2010) have been used to close smaller defects. At present, the placement of completely covered metal or plastic stents (Doniec et al. 2003; Hunerbein et al. 2004) is still the favored conservative treatment option for esophageal leakage. The implantation of these stents has been thoroughly studied and has been proven to be effective (Tuebergen et al. 2008; van Boeckel et al. 2011). However, stent implantation does not always lead to a sufficient sealing of the leakage (van Boeckel et al. 2011), and dislocation rates of up to 40% (Kauer et al. 2008) have been reported. Another important complication is failure of stent extraction due to ingrowth of granulation tissue and/or secondary strictures due to scarring (Doniec et al. 2003; Loske and Muller 2009; Schubert et al. 2005). While stents bridge the defect intraluminally and prevent further leakage, continuous local drainage is necessary to prevent inflammatory fluids from remaining in the perianastomotic tissues and maintaining inflammation. The well-established stent therapy is now being challenged increasingly by endoscopic vacuum therapy (EVT). While it can already be considered as standard therapy for leakages of lower colorectal anastomoses, its use in the upper GI tract only evolved several years later. Yet soon after first reports of the technical feasibility of endoscopic vacuum therapy in the upper GI tract, several case series with good success rates in the management of esophageal leaks were published. However, most series include heterogeneous types of leaks and are not focused on anastomotic leaks. All publications report excellent success rates (healing of leaks and perforations in 84–100%) and virtually no procedure-related complications in these patient cohorts. The technique appears to have potential as a first-line therapy for postoperative upper GI leaks (◘ Table 6.2).
Table 6.2
Endoscopic vacuum therapy (EVT) for leaks of different etiology
Literature | Patients (n) | Indication for EVT | Success rate (closure of leak by EVT) |
|---|---|---|---|
Weidenhagen et al. | 6 | 6× a. l. | 6/6 (100%) |
Wallstabe et al. | 1 | 1× a. l. | 1/1 (100%) |
Brangewitz et al. | 32 | 30× a. l. 1× perf. 1× b. s. | 27/32 (84%) |
Schniewind et al. | 17 | 17× a. l. | 15/17 (88%) |
Bludau et al. | 14 | 8× a. l. 6× perf. | 12/14 (87%) |
Smallwood et al. | 6 | 1× a. l. 5× perf. | 6/6 (100%) |
Schorsch et al. | 35 | 21× a. l. 7× perf. 1× b.s. 6× o.o. | 32/35 (91%) |
Kuehn et al. | 21 | 11× a. l. 8× perf. 2× b.s. | 19/21 (91%) |
Seyfried et al. | 1 | 1× b. surg. | 1/1 (100%) |
Total | 133 | 95× a. l. 27× perf. 4× b. s. 1× b. surg. 6× o. o. | 119/133 89.5% |
Since its first description by Wedemeyer et al. and Loske et al. the abovementioned principle is used by all authors, with only small variations in the procedure. Recently, a commercially available and certified drainage system using the overtube principle has been distributed (Eso-SPONGE®, Braun Melsungen AG).
Resources and Organizational Requirements
In case of anastomotic leakage or perforation in the upper gastrointestinal tract, interventional endoscopy has evolved as an effective alternative treatment modality (Maish et al. 2005). Endoscopic vacuum therapy requires a competent, experienced endoscopic team and a well-equipped endoscopic unit permitting additional peri-interventional radioscopy as well as an examiner who is well trained in the field of EVT. EVT can be done under conscious sedation or general anesthesia, depending on the general condition of the patient.
The following tools and equipment have to be provided (◘ Fig. 6.3a):
Fig. 6.3
Endoscopic vacuum therapy (EVT) in the upper gastrointestinal tract. a Arrangement of the necessary materials. b Open-pore polyurethane sponge—sponge preparation. c Sponge mounted on a gastric tube for endoscopic vacuum therapy. d Mounted sponge—L loop for easy positioning. e Principle of sponge drainage insertion into the esophagus using a forceps in a «backpack method»
Materials
1× open-pore polyurethane sponge (e.g., VivanoMed® Foam, Paul Hartmann AG, Heidenheim, Germany; V.A.C. GranuFoam, KCI-Kinetic Concepts, Inc., TX, USA)
1× electronic vacuum pump system(e.g., VivanoTec®, Paul Hartmann AG, Heidenheim, Germany)
1× polyvinyl chloride (PVC) gastroduodenal tube (e.g., Covidien™ Salem Sump™, 14 Fr/Ch (4.7 mm) × 114 cm, Covidien™, MA, USA)
2× suture material (e.g., Ethibond Excel, Ethicon, Johnson & Johnson MEDICAL GmbH)
1× scissors, 1× clamp, 1× needle holder, 1× Magill forceps, 1× laryngoscope, 1× metal pin for the Redon drainage, 1× tube for nasal diversion, 1× endoscopic forceps, and lubricant
Endoscopic Vacuum Therapy (EVT): Procedure
EVT is performed under conscious sedation or general anesthesia, depending on the general condition of the patient. After endoscopic assessment of the geometry of the leakage and the cavity, a polyurethane foam sponge is cut into the corresponding shape (◘ Fig. 6.3b). The sponge is fixed to the tip of a polyvinyl chloride (PVC) gastroduodenal tube with a suture at the proximal and distal ends of the sponge (◘ Fig. 6.3c) allowing communication between the side ports of the tube with the sponge. An additional suture loop (L loop) is placed at the tip of the sponge (◘ Fig. 6.3d). Thus, the additional loop at the tip of the sponge serves as a purchase for the endoscopic forceps and facilitates manipulation of the sponge into difficult-to-access cavities and hollow spaces. After final shaping of the sponge (◘ Fig. 6.3d), the loop is grasped with a forceps (◘ Fig. 6.3e) and pulled close to the endoscope, and the sponge is placed in the leakage cavity under direct endoscopic vision. If the defect is initially not wide enough to accommodate the endoscope (<10 mm) and an abscess cavity is suspected, the opening can be dilated by endoscopic balloon dilatation (Esophageal Balloon Dilatation Catheter, 10–12 mm, Boston Scientific, Ratingen, Germany) to allow extraluminal inspection by the standard endoscope and examination of the extraluminal septic focus. After sponge placement, the vacuum drainage tube is diverted through the nose. Continuous suction of 100–125 mmHg generated by an electronic vacuum pump system (e.g., VivanoTec®, Paul Hartmann Ag, Heidenheim, Germany) is connected to the drainage tube, allowing the sponge to stay in position due to continuous suction. Optionally, with the sponge drainage system in place, parenteral feeding, a transnasal enteral feeding tube, a percutaneous endoscopic gastrostomy (PEG,) or a jejunostomy feeding tube ensure enteral nutrition (◘ Fig. 6.4a). A scheduled change of sponges should take place every 3rd to 5th day; and at each session, the size of the defect has to be assessed and be treated with an individually prepared sponge, cut to fit the lesion’s dimensions. After each discontinuation of suction, the tube has to be diverted through the mouth and removed simply by pulling. It is advisable to flush the tube with 0.9% saline solution to dissolve the granulation tissue from the pores of the sponge prior to removal. In some cases, remnants of the sponge have to be removed by endoscopic forceps. Over the course of the treatment and with diminishing defect size not allowing an access with the scope, sponge placement can be changed from its initial intracavitary position to intraluminal position onto the defect at any time. Secretion is then drained endoluminally, and the continuous suction force results in temporary complete occlusion of the intestinal passage. Especially in the absence of an extraluminal wound cavity (e.g., with early diagnosis of a transmural defect in case of a Boerhaave syndrome), it is advisable to use an intraluminal sponge drainage covering the whole defect zone within the lumen of the upper GI tract. EVT can be stopped when the defect size becomes too small for further sponge placements and the defect is finally lined with surface epithelium (◘ Fig. 6.4b–d). Complete healing of the anastomosis should be assessed by endoscopy and additional X-ray contrast study showing no clinical signs of persistent leakage. Usually the defect completely closes within 1–2 weeks.
