Esophageal Exclusion and Retrosternal Bypass in Management of Post-Pneumonectomy Esophagopleural Fistula: A Case Report

Introduction

Esophagopleural fistula (EPF) is a rare complication following pneumonectomy, occurring in up to 1% of patients [1] and with few cases reported in the literature. It is a highly morbid condition characterized by a pathological connection between the esophageal lumen and the pleural cavity leading to potentially life-threatening sepsis with a mortality rate of 49% to 63% [1–5].

Treatment of post-pneumonectomy EPFs can be conservative, endoscopic, or surgical, depending on the onset and characteristics of the fistula, as well as the clinical status of the patient [1].

We describe a case of surgical treatment of a large esophagopleural fistula after pneumonectomy for lung cancer, after failure of endoscopic treatment.

Case Report

A 65-year-old woman with no prior medical history was diagnosed with lung adenocarcinoma of the left lower lobe, infiltrating the fissure and with positive N2 lymph nodes, for which she received neoadjuvant chemotherapy and immunotherapy. Toxicity was observed during the 4th cycle of chemotherapy with Paclitaxel, after which surgical treatment was proposed. During the surgical procedure, the tumor was seen to originate from the apical segment of the left lower lobe and extended into the fissure and upper lobe with infiltration of the bronchial and vascular branches for the upper lobe. Furthermore, the tumor infiltrated the adventitia of the lateral part of the aortic arch, requiring an en-bloc resection of the left lung, parietal pleura and portion of aortic adventitia, as well as standard lymphadenectomy.

On the 8th postoperative day (POD), the patient showed onset of fever, increased inflammatory markers and evidence of purulent fluid and food particles in the chest tube. After oral administration of methylene blue dye and its subsequent drainage from the pleural space, the patient underwent esophagogastroduodenoscopy (EGD), confirming the diagnosis of an EPF at 30 cm from the incisors. An initial endoscopic treatment was attempted with positioning of a self-expandable metallic stent anchored with Stentfix; the patient was made nil by mouth and fed by total parenteral nutrition. Culture of the pleural fluid resulted positive for Enterobacter cloacae, Enterococcus faecium, and Streptococcus oralis; therefore, antibiotic therapy was optimized, and daily lavages of the pleural cavity were performed with saline solution and antibiotics.

After an initial period of clinical improvement after stent positioning, a subsequent decline in the patient’s general status warranted another EGD, in which the stent had to be repositioned. The endoscopic treatment proved unsuccessful, and 26 days after the initial diagnosis of EPF an open thoracostomy was performed with debridement of the pleural cavity, where the fistula was seen to extend beyond the repositioned esophageal stent. The stent was therefore removed and the patient transferred to our unit.

Once admitted to our unit, another EGD was performed to reevaluate the extent of the EPF and to place a nasojejunal feeding tube. At this stage, the patient’s nutritional status was greatly debilitated, with a body mass index of 17.5, and required optimization before undergoing the planned surgical procedure of esophageal exclusion and restoration of the continuity of the gastrointestinal tract using the stomach. This was achieved with continuous infusion of hypercaloric enteral nutrition through the nasojejunal feeding tube 24 hours per day, while simultaneously addressing the infectious issues. The antibiotic therapy was optimized, and daily lavages of the pleural cavity with disinfectant solution continued to be performed. An updated CT scan was also done (Figure 1).

Once the clinical condition had improved, the inflammatory markers were reduced and the nutritional status was satisfactory, the patient underwent the planned surgical procedure, which involved 3 different surgical approaches.

The initial approach was through laparoscopy, during which the stomach was prepared for gastric conduit creation by sectioning the left gastric and left gastroepiploic vessels. Preservation of the right gastroepiploic vessels was crucial at this stage to ensure adequate vascularization to the gastric tube, which was to be fashioned out of the greater curvature. This step was done with the aid of intravenous indocyanine green dye to help identify the vessels and to visualize the perfusion of the stomach once completely mobilized. The stomach was then sectioned from the esophagus using a mechanical stapler (EndoGIA® 60 mm with purple reload).

The second approach was through left cervicotomy with mobilization of the cervical portion of the esophagus and its sectioning in the most caudal portion using a linear mechanical stapler (GIA® 60 mm with green reload). The cranial stump was then prepared for subsequent anastomosis. With this technique the thoracic esophagus was left in place but lacking continuity with the rest of the gastrointestinal tract.

The third and final approach was with minilaparotomy, performed by extending the subxiphoid laparoscopic incision. The stomach was extracted through this incision and transformed into a tube by positioning a 36 Fr calibration tube along the greater curvature and then using a linear mechanical stapler (GIA® 75 mm with green reloads) to divide the stomach from the fundus to the antrum along the border of the calibration tube (Figure 2). Adequate vascularization of the gastric conduit was confirmed once again using indocyanine green angiography.

The newly fashioned gastric tube was subsequently transposed through the thorax in the retrosternal space until its proximal stump reached the stump of the cervical esophagus. The anastomosis was then performed with a mixed technique using a mechanical stapler (Echelon® 45mm with green reload) for the posterior wall and completed with a running suture in Prolene 4-0. Figure 3 illustrates the completed procedure. A 15 Fr Jackson-Pratt drain was positioned adjacent to the anastomosis and the cervicotomy incision was closed. Lastly, a jejunostomy was fashioned for adequate nutrition postoperatively and all the abdominal incisions were closed after positioning a subdiaphragmatic 15 Fr drain on the left side.

At the end of the surgical procedure, the patient stayed in the intensive care unit until the 3rd POD, when she was transferred back to our ward. On the 5th POD the abdominal drain was removed and on the 7th POD the nasogastric tube placed to protect the anastomosis was removed. Eight days after the esophageal exclusion, the patient underwent a final surgical procedure in which the open thoracotomy was closed.

On the 10th POD after the esophageal surgery, the patient resumed oral intake of liquids without any complications and with subsequent reintroduction of progressively more solid foods. On the 16th postoperative day, the cervical drain was removed.

An increase in cholestasis markers was observed during reintroduction of soft foods and an MRCP was performed that excluded obstruction of the biliary and pancreatic ducts.

The patient was discharged 28 days postoperatively with inflammatory markers within the normal range, without further antibiotic treatment, instructed to eat small and frequent meals throughout the day and to continue nighttime enteral nutrition via jejunostomy for a subsequent month.

The patient returned to our attention for surgical follow-up 3 months postoperatively. The esophagogram performed (Figure 4) showed a good caliber of the esophagogastric anastomosis, no leakage of contrast from the esophageal lumen, and she had a significant improvement in nutritional status, with a BMI within the normal range.

Discussion

EPF is a rare but important complication of pneumonectomy. The literature is limited on this subject; therefore, treatment approaches are heterogenous. The main challenges to face in the management of this pathology are acute and/or chronic sepsis and malnutrition [6].

The median time reported from pneumonectomy to onset of EPF is of 12.5 months (range 0–600 months) [1], which contrasts with the 8 days observed in the reported case. This most likely suggests an iatrogenic cause for the EPF resulting from surgical injury during the pneumonectomy procedure. EPFs occurring soon after pulmonary surgery can arise because of direct injury to the esophagus itself or its blood supply during extensive dissection [7,8]. Although no apparent injury to the esophageal wall was observed in the initial surgery, damage resulting in its devascularization may have occurred during the laborious dissection of hilar structures and/or during lymphadenectomy of nodal stations 7, 8, and 9.

Conservative, endoscopic, and surgical strategies have been described in the treatment of EPFs, with conservative treatment generally being reserved for patients who are asymptomatic or systemically unfit to undergo any interventions [3]. Regardless of the approach selected, most cases have chest drain insertion and/or irrigation of the pleural cavity as an initial management step [1]. This is a particularly important step in treating EPFs after pulmonary resection to avoid further complications like bronchial stump failure and bronchopleural fistula due to infection [8].

Endoscopic treatment is recommended by the American Gastroenterological Association (AGA) 2021 guidelines on the management of perforations of the gastrointestinal tract. In the case of esophageal perforations under 2 cm, the use of through-the-scope clips is recommended, while endoscopic suturing should be used for defects larger than 2 cm. Self-expandable metallic stents (SEMS) are recommended in cases where primary closure is not possible [9]. Studies have shown varying success rates of 48% to 100% in the use of stents for treating benign esophageal ruptures and anastomotic leaks, with no statistically significant differences between partially-covered SEMS, fully-covered SEMS, and self-expanding plastic stents (SEPS) [10].

One important advantage of endoscopic treatment is that it is generally available in a wider range of centers compared to specialized thoracic and/or gastrointestinal surgery. However, stent placement is frequently complicated by stent migration and/or tissue ingrowth or overgrowth, as well as having varying rates of success [10]. Furthermore, Debourdeau et al showed a correlation between the size of non-malignant esophageal fistulae treated endoscopically and mortality, with no mortality in the group with punctiform fistulae versus 71.4% mortality in the group of large fistula orifices, suggesting that surgical treatment should be preferred for large-orifice fistulae [11].

Phelan et al recently published a systematic review on the treatment of post-pneumonectomy EPFs which summarized the type of surgical procedures described in the literature to treat this condition. These included direct closure of the fistula with or without muscle/fat/omentum flap, thoracoplasty/thoracomyoplasty, and esophageal resection/reconstruction using either the stomach or colon [1]. In our patient, direct closure of the EPF was not feasible due to the large extent of the fistulous tract, lack of viable tissue for reinforcement of the suture, and high risk of dehiscence.

While esophagectomy can be a valid approach, it was deemed technically hazardous and unsafe in this patient due to the presence of thick mediastinal inflammatory tissue resulting from recent surgery and infection. Esophageal exclusion, firstly described by Postlethwait for unresectable advanced esophageal cancer [12], has been reported as a viable alternative to esophagectomy in the treatment of benign strictures, perforations, and fistulae [13–17], and in most cases the retained esophagus scars down without further complications [15]. However, even if rare, the onset of esophageal mucocele [13] and malignant transformation [18] could be possible long-term complications and should be evaluated in the postoperative follow-up. It is a technique that has been used in the past by the authors in the surgical treatment of Boerhaave syndrome without long-term complications.

In our experience, a narrow gastric tube is the best choice for reconstruction in terms of postoperative management and complications. However, colonic transposition with esophago-colonic anastomosis can be performed if the stomach is involved or unusable [19].

Furthermore, the chosen surgical procedure with a cervical esophagogastric anastomosis allows for earlier detection of potential anastomotic leakage and for safe radiation therapy to the thorax if needed as part of further oncological treatment.

The postoperative course and subsequent follow-up show that this is also a viable procedure for the treatment of EPFs, with good outcomes. Nevertheless, it is an invasive and extensive surgical procedure that must be performed by specialized, expert surgeons in patients whose clinical condition is not prohibitive. Possible postoperative complications include anastomotic stenosis, anastomotic leakage, dumping syndrome, and gastric conduit necrosis.

Conclusions

Esophagopleural fistulas are a rare complication after pneumonectomy that require prompt treatment. While there is no standardized approach for this type of pathology, surgical treatment is often used either as first-line therapy or after endoscopic treatment failure. Esophageal exclusion and reconstruction of the digestive tract using the stomach may be the best course of action when the extent of the fistulous tract presents a problem for direct repair. This approach presents a single-step solution for excluding the EPF without entering the chest cavity and thus potentially causing further damage to an already fragile area, while offering a reconstruction solution adequate for long-term nutrition.