Successful treatment of large cavity esophageal disruptions with transluminal washout and endoscopic vacuum therapy: a report of two cases

Introduction

Esophageal perforation and postoperative leak after esophagogastrostomy are associated with high morbidity and mortality due to the development of mediastinitis and thoracic contamination. The development of this life-threatening complication has no uniform treatment guidelines, but typically relies on percutaneous drainage, cessation of oral intake, and repair of the luminal defect.^1,2 Endoscopic vacuum therapy (EVT) is based on well-established vacuum-assisted surgical closure techniques and is a novel, minimally invasive approach to drain mediastinal or thoracic abscesses and fluid collections endoscopically. ³ Recent literature has shown that this approach provides effective transnasal treatment for small collections without the need for percutaneous hardware and yields encouraging results.^4,5 The extent of just how much contamination is potentially treatable by this means however is unknown.

Here, we report two cases demonstrating the successful treatment of severe and widespread mediastinal and thoracic contamination by serial debridement and washout and EVT with encouraging results. We demonstrate that even large fluid collections and abscesses could be managed by EVT when the approach is coupled with serial endoscopic washout and debridement.

Methods

Patients

In this retrospective analysis, two patients with large esophageal defects were treated using EVT and serial debridement and washout between February and August 2018. Patient 1 developed esophageal perforation from Boerhaave syndrome with a 2-cm disruption and 10.7-cm cavity, and Patient 2 developed postoperative anastomotic leak after Ivor-Lewis esophagectomy with two separate disruptions: a 2-cm proximal disruption with a 4-cm mediastinal cavity and 14-cm gastric pull up disruption with a cavity encompassing the entire central and lower right thorax (Figure 1(a)). Both patients received broad-spectrum antibiotics, nasogastric (NG) tube decompression, and chest tubes. Patient 1 received a self-expanding metal stent (SEMS) to optimize EVT. A WallFlex™ stent was used and renewed twice over the 6-week course of endoscopic therapy. The mean age of the patients was 69.5 (range: 63–76) years. The primary outcome of this study was esophageal defect closure.

OPEN IN VIEWER

Figure 1.

(a) Coronal CT of Patient 2 demonstrating a large collection of fluid and solid debris dominating the central and lower right thoracic cavity (arrowheads) with chest tube (blue arrow) and collapsed residual lung in the upper right thorax (yellow arrow). (b) Endoscopic view of the mediastinal cavity with purulent debris prior to washout and debridement (c) now washed and debrided. (d) Right thoracic cavity after washout and debridement with percutaneous drainage tube seen in the center. (e) EVT polyurethane sponge and nasogastric tube assembly. (f) Endoscopic insertion of the EVT assembly down the esophagus using rat tooth alligator jaw forceps for intracavitary placement. (g) Follow-up swallow study 2 months after serial washouts and EVT demonstrating relatively normal progression of contrast through the remnant esophagogastric pull up to the small bowel with minimal residual extraluminal fluid collection.

Endoscopic vacuum therapy

Once the esophageal defects were located, the cavities were explored and lavaged with normal saline and any necrotic and purulent debris were removed with a suction cap (Figure 1(b)). Once the cavities were cleaned of solid debris (Figure 1(c) and (d)), an NG tube was inserted through the nare and retrieved through the mouth to attach an open pore polyurethane sponge obtained from a vacuum-assisted closure (VAC) medium dressing kit (V.A.C.^® GranuFoam™; KCI, San Antonio, TX, USA). The sponge was trimmed in size so that it could fit across the upper esophageal sphincter and into the desired cavity and then was sutured to the tip of the NG tube (Figure 1(e)). The NG tube with polyurethane sponge was inserted into the esophagus and directed toward the cavity using rat tooth and alligator jaw grasping forceps (FG-44NR-1; Olympus, Co., Tokyo, Japan) (Figure 1(f)). Intracavitary insertion was preferred due to the large size of the thoracic and mediastinal cavities (Figure 2). After ensuring proper sponge placement, the NG tube was cut and connected to a VAC suction device and placed to 125 mmHg of continuous suction. Serial washout and EVT sponge exchanges were performed every 5–7 days. At the conclusion of EVT, the cavities were well granulated and collapsed allowing removal of the EVT sponge assembly. A gastrografin swallow study was subsequently performed to ensure leak closure (Figure 1(g)).

OPEN IN VIEWER

Figure 2.

Illustration of EVT in Patient 2 with bilateral transnasal wound NG tube-sponge assemblies placed endoscopically into both the anterior mediastinal and right thoracic cavities for decompression and drainage thus allowing healing by secondary intention. The esophagogastric anastomosis (blue star) with the superior EVT sponge is seen in the anterior mediastinal cavity. Also noted is the large lateral gastric pull-up disruption (yellow star) with the inferior EVT sponge assembly leading to the large right thoracic collection.

Results

A total of two patients with large cavities greater than 8 cm underwent EVT over the study time period (Table 1). The median size of the collections via longest cross-sectional diameter was 10.7 cm as determined by a combination of endoscopic assessment and cross-sectional imaging. Complete leak closure was confirmed in both cases by gastrografin swallow study. Patient 1 completed a 3-month follow-up after hospital discharge with no complications. Patient 2 died 6 weeks after completion of EVT due to aspiration pneumonia from oropharyngeal dysphagia.

OPEN IN VIEWER

Table 1.

Patient Demographics and Procedure Details.

Patient	Sex, age	Type of leak	Etiology	Size of disruption (cm)	Size of collection(s) (cm)	Location	Additional therapies	Time to intervention (days)	Number of EVT procedures	Time to resolution (days)	Outcome	Adverse events	3-month sustained resolution
1	F, 63	Spontaneous perforation	Boerhaave syndrome	2	10.7 × 3.3	Mediastinum	Abx, chest tube, NG decompression, SEMS	15	3	42	Complete closure	None	Yes
2	M, 76	Anastomotic leak	Ivor-Lewis esophagectomy	2 (M); 14 (T)	8.4 × 4.7 × 11.5 (M); 5.6 × 5.2 × 1.9 (T)	Mediastinum; thorax	Abx, chest tube, NG decompression	27	10	55	Complete closure	None	n/a; patient expired

Abx, antibiotics; EVT, endoscopic vacuum therapy; M, mediastinum; NG, nasogastric; SEMS, self-expanding metal stent; T, thorax.

Discussion

Esophageal anastomotic leaks and other types of esophageal disruptions with associated mediastinal and thoracic contamination carry high rates of morbidity and mortality, and the development of effective, minimally invasive, endoscopic techniques to control and treat these fluid collections and abscesses stands to significantly improve the management of these complications and the mortality rates of this patient population. The development of EVT is an attractive approach to decompressing and draining these contaminated collections entirely internally and transnasally without the need for percutaneous hardware in many circumstances in appropriately selected patients and has demonstrated potential to significantly reduce morbidity and mortality in this population.^4,5 Investigators have reported encouraging results using this approach, allowing closure of the esophageal defect in the intestinal lumen, and adequate drainage, rapid cleaning, and granulation in the wound cavity until the defect closes by secondary intent. ⁶

According to Mennigen and colleagues, ⁷ EVT offers a safe and effective approach for the management upper gastrointestinal (GI) perforations and anastomotic dehiscence. The technique involves introducing a polyurethane sponge attached to an NG tube into the leakage endoscopically and application of continuous suction via the tube, thus resulting in effective drainage of the contaminated cavity and induction of wound healing. This approach has proven to be effective in other anatomical sites as well, such as in the treatment for rectal anastomotic insufficiency, with the proposed mechanism of the beneficial effects credited to fluid removal, edema reduction, stimulation of granulation tissue formation, and decrease of bacterial contamination. ⁸

Current studies lack evidence that EVT is successful in treating upper intestinal leaks and perforations with large cavities and severe contamination. One earlier report claims that treatment of larger disruptions is less effective due to size restraints of a swallow-to-mouth sized sponge and in openings which have copious secretions. ⁹

It is demonstrated in our study that a combination of serial transluminal washout and debridement in conjunction with EVT provided successful treatment of large, widespread fluid collections and abscesses that measured up to 11.5 cm in cross-sectional diameter. Our 6- to 8-week time to completion is longer than what is reported in the literature, which describes a mean treatment of 18 days. ¹⁰ While technically successful, the clinical outcome demonstrates the seriousness of this situation

The case of esophageal perforation due to Boerhaave syndrome in Patient 1 illustrates that severe thoracic contamination can be treated with EVT with a relatively short duration of treatment (6 weeks) and time to complete recovery. It should be noted that we placed an SEMS over the NG tube in Patient 1 to effectively seal off the cavity both to ensure adequate intracavitary negative pressure for EVT and also for preventing recontamination of the cavity with gastric and esophageal fluids.

In the case of postoperative anastomotic leak in Patient 2, we describe the use of bilateral NG tube-sponge assemblies for decompression and drainage to treat two separate collections, one of which involved a large thoracic collection and the other a mediastinal collection. The etiology of the massive disruption was likely a combination of gastric ischemia and axial/radial pressure necrosis from the stent originally placed, leading to a massive 14-cm gastric pull-up sidewall disruption with associated massive contamination of the entire right lower thorax. The dual approach to EVT with successful management of two separate disruptions demonstrates that multiple collections can be treated simultaneously, even large cavities with copious secretions when combined with endoscopic washout and debridement.

Our technique involving EVT coupled with serial washout and debridement has the potential to change the current approach to the management of anastomotic dehiscence and other types of esophageal disruptions associated with severe mediastinal or thoracic contamination. If washout and debridement of necrotic debris is planned, the collection should be sufficiently walled off, otherwise negative-pressure should first be applied, after which any required washout and debridement could then be safely performed. This method is an important means of preventing spread of infection. These cases demonstrate the feasibility of treating even large mediastinal or thoracic fluid collections with EVT. At our institution, we perform serial endoscopy for EVT every 5–7 days to replace the sponge. This provides opportunity to provide repeated lavage of the contaminated cavity and removal of debris, potentially contributing to an improved outcome.

In conclusion, although Patient 2 with anastomotic leak did not survive long term, these cases demonstrate that the use of endoscopic washout and debridement followed by EVT can be effective for large, even multiple, thoracic and mediastinal contaminations following esophageal perforation and gastroesopagheal anastamotic dehiscence and leaks in appropriately selected patients. We recommend that endoscopists consider measuring the cavity prior to initiating EVT as this information can be used to assess treatment progress. If the treatment is not progressing or an intraprocedural adverse event is suspected, the endoscopist should have a low threshold for repeating enhanced cross-sectional imaging for assessment and to ensure that the drain sponge is in the correct position. Computed tomographic (CT) imaging is also important in helping to identify complete resolution and treatment cessation. The use of this technique can be laborious and time consuming however, and so an improvement in the efficiency of the procedure is needed to facilitate more ideal outcomes in this patient population in need of improved therapeutic options.