|
|
||||||||
a Division of Thoracic and Vascular Surgery, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland
b Division of Visceral Surgery, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland
c Division of Gastroenterology, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland
Accepted for publication May 6, 2009.
* Address correspondence to Dr Ris, Service de Chirurgie Thoracique et Vasculaire, Centre Hospitalier Universitaire Vaudois, 1011 Lausanne, Switzerland (Email: hans-beat.ris{at}chuv.ch).
| Abstract |
|---|
|
|
|---|
Methods: Intrathoracic esophageal leaks with mediastinitis and systemic sepsis occurred after delayed spontaneous perforations (n = 7) or surgical and endoscopic interventions (n = 12). Six patients presented with fulminant anastomotic leaks. Seven patients had previous attempts to close the leak by surgery (n = 4) or stenting (2) or both (n = 1). The debrided defects measured up to 2 x 12 cm or involved three quarters of the anastomotic circumference and were closed either by a full thickness diaphragmatic flap (n = 13) or a pediculated intrathoracically transposed extrathoracic muscle flap (n = 6). All patients had postoperative contrast esophagography between days 7 and 10 and an endoscopic evaluation 4 to 6 months after surgery.
Results: There was no 30-day mortality. During follow-up (4 to 42 months), 16 patients (84%) revealed functional and morphological restoration of the esophagointestinal integrity without further interventions. One patient required serial dilatations for a stricture, and 1 underwent temporary stenting for a persistent fistula; both patients had normal control endoscopy during follow-up. A third patient requiring permanent stenting for stenosis died from gastrointestinal bleeding due to stent erosion during follow-up.
Conclusions: Intrathoracic esophageal leaks may be closed efficiently by a muscle flap onlay approach in the presence of mediastinitis and where a primary repair seems risky. The same holds true for fulminant intrathoracic anastomotic leaks after esophagectomy or other surgical interventions at the gastroesophageal junction.
| Introduction |
|---|
|
|
|---|
Richardson and coworkers [6] have described the successful closure of esophageal perforations in the context of mediastinal sepsis by use of muscle flaps without attempting a primary repair in situations where direct closure of the defect was judged risky because of tissue friability. We have adopted this technique for delayed (more than 72 hours) esophageal perforation and for postoperative fulminant intrathoracic esophageal leaks with mediastinal and systemic sepsis to restore gastrointestinal continuity while avoiding esophageal exclusion. In our institution, intrathoracically transposed muscle flaps have been successfully used as airway substitutes even in the presence of empyema or infected residual pleural spaces [7–9].
This report summarizes our experience with the pediculated muscle flap onlay approach for the closure of intrathoracic esophageal leaks in the presence of mediastinal and systemic sepsis in a consecutive series of 19 patients.
| Patients and Methods |
|---|
|
|
|---|
The study was approved by the Local Ethics Committee, individual consent was waived.
Patients were separated into two groups: patients with spontaneous esophageal perforations and patients with esophageal leakage after surgical interventions or endoscopic procedures.
Patients With Spontaneous Esophageal Perforations
Spontaneous perforations occurred in 7 patients and were all localized in the lower third of the esophagus (Table 1). Six patients presented with a Boerhaave's syndrome, 4 with a left-sided and 2 with a right-sided perforation. One of the patients with a right-sided perforation had a concomitant right lower lobe abscess by continuity. The seventh patient presented with a 12-cm right-sided spontaneous perforation due to the rupture of an infected pancreatic pseudocyst into the mediastinum. The delay between the onset of symptoms of perforation and closure varied between 72 hours and 12 days, and the size of the esophageal defect after debridement varied from 3 x 4 cm to 3 x 12 cm. Two patients with Boerhaave's syndrome underwent unsuccessful attempts to close the leak either by a primary surgical repair or by endoscopic stenting.
|
|
Middle third esophageal leaks
Three patients presented with right-sided intrathoracic anastomotic disruption after esophagectomy and esophagogastric reconstruction at the level of the azygous vein. All 3 patients had preoperative chemotherapy for esophageal cancer and presented with fulminant leaks several days after the initial operation with some degree of conduit necrosis. The size of the debrided defects at the level of the anastomosis was one half to three quarters of the anastomotic circumference.
Lower third esophageal leaks
Three patients had a left-sided intrathoracic anastomotic disruption at the gastroesophageal junction, 2 after gastrectomy for malignant disease and 1 after multiple operations including gastrectomy for failed bypass surgery. The size of the debrided defects at the level of the anastomosis involved as much as three quarters of the anastomotic circumference. One patient presented with a persistent left-sided supradiaphragmatic esophageal leak after resection of an esophageal diverticulum that had failed to be controlled by primary repair. One patient had a persistent left-sided supradiaphragmatic esophageal leak after multiple antireflux procedures and failed closure by endoscopic stenting (Fig 1). One patient presented with a right-sided esophageal leak after right lower bilobectomy. Two patients had lower-third esophageal perforations after endoscopic pneumatic dilatation for achalasia and a caustic structure, respectively. The esophageal defects ranged from 3 x 4 cm to 3 x 10 cm after debridement. Five patients had previous unsuccessful attempts to close the leak either by surgery (n = 3) or stenting (n = 1) or both (n = 1).
|
|
Closure of right-sided middle third leaks by latissimus dorsi or serratus anterior onlay approach
Middle third esophageal leaks were accessed by right-sided thoracotomy through the sixth intercostal space and were closed preferentially by an intrathoracically transposed latissimus dorsi flap, or a serratus anterior muscle flap if the latissimus dorsi muscle had been divided during a previous thoracotomy. The pedicled muscle flaps were harvested during the thoracotomy and were transposed intrathoracically through the bed of a resected segment of the second rib, as previously described [7–9]. After decortication and debridement of the mediastinum and the esophageal defect, the muscle flap was sutured into the esophageal defect and wrapped around the debrided mediastinal structures (Fig 2A). Three patients with anastomotic leakage after esophagectomy underwent debridement of the esophageal wall and the gastric conduit into healthy tissues, which required resection of as much as three quarters of the anastomotic circumference. A serratus anterior muscle flap was then sutured into the debrided anastomotic defect without attempting a primary repair and was wrapped around the anastomosis and the gastric conduit (Fig 2B). Two additional patients underwent closure of an esophageal leak by extrathoracic muscle flaps through right-sided thoracotomy, 1 with a concomitant abscess of the right lower lobe who in addition required a right lower lobectomy, and 1 with a 12-cm-long esophageal perforation after rupture of a pancreatic pseudocyst into the mediastinum. Patients with right-sided thoracotomy underwent primary closure of the chest and chest tube drainage.
Closure of lower third leaks by diaphragmatic muscle onlay approach
A thoracoabdominal incision through the seventh intercostal space with division of the costal arch was performed, and the ipsilateral diaphragm was detached on its hemicircumference at its insertion along the anterior and lateral chest wall. A full-thickness diaphragmatic flap was created while the posterior and mediastinal insertion was kept intact. The flap was rotated into the mediastinum and sutured into the debrided esophageal defect or anastomosis after mediastinal debridement and decortication had been performed (Fig 2C). The diaphragmatic flap was then wrapped around the reconstructed esophagus or anastomosis and covered the lower part of the mediastinum. A VAC device was cut according to the shape and size of the diaphragm and was applied on the initial site of the diaphragm. On the right side, it was applied directly on the liver surface; on the left side, the colon and the spleen were covered by omentum before application of the VAC device to avoid direct suction on these structures. No attempt was made to close the thoracoabdominal incision primarily. The skin, however, was adapted over the VAC device, leaving a small pleurostomy of several cm for vacuum application. The VAC device was repeatedly changed once or twice a week until granulation tissue covered the mediastinum and the reconstruction site. No diaphragmatic reconstruction by mesh was necessary as the VAC device also induced the formation of granulation tissue at the initial site of the diaphragm.
All patients with esophageal leaks at the gastroesophageal junction underwent closure by diaphragmatic flaps, including 3 who had anastomotic disruptions. One patient with esophageal rupture after pneumatic dilation for caustic stenosis underwent closure of the perforation site with a diaphragmatic flap that was used at the same time to enlarge the caustic stricture by the muscle patch.
Follow-Up
All patients underwent clinical, radiologic, and endoscopic follow-up. Morbidity and 90-day mortality were recorded. Postoperative contrast esophagography and flexible endoscopy with biopsies harvested at the level of the muscle patch were routinely performed between day 7 and 10, and 4 to 6 months after the operation, respectively.
| Results |
|---|
|
|
|---|
Postoperative contrast esophagography performed on postoperative day 7 to 10 revealed a restored integrity of the esophageal wall without leakage in 18 of 19 patients (94%). One patient had minor leakage, which was successfully treated by temporary stenting with a covered stent for 2 months without leakage or stenosis during follow-up.
Control endoscopy during follow-up revealed a patent esophagus or anastomotic site without stenosis and a well-integrated muscle flap in 17 of 19 patients (89%). Histologic assessment of biopsies harvested at the site of the muscle flaps during control endoscopy 4 to 6 months after the operation demonstrated a malpighian reepithelialization of the muscle patch associated with hyperplasia (Fig 3). Two patients had stenosis at the site of reconstruction during follow-up; 1 patient was successfully treated by four dilations without further stenosis, and 1 patient underwent stenting but subsequently died of severe gastrointestinal bleeding due to stent erosion 12 months after the operation. All surviving patients had the perception of normal swallowing without dysphagia and no regurgitation or reflux disease.
|
| Comment |
|---|
|
|
|---|
For most institutions, reinforced primary repair nowadays remains the surgical treatment of choice for thoracic oesophageal perforations [1, 3–6, 14–16] Various types of vascularized tissue have been used to reinforce and buttress the primary repair, including pleural flaps, strips of diaphragm, rhomboid and latissimus dorsi muscle flaps, intercostal muscle flaps, or pericardial fat pads [3–6, 10, 14–16]. A primary repair was also shown to be successful in situations of time delays longer than 24 hours between perforation and treatment [5]. However, the leakage rate after primary repair with or without reinforcement varied between 25% and 83% in these patients [1, 6]. A recent review dealing with the management of esophageal perforations revealed that the mortality rate after primary repair ranged from 0% to 31% and averaged 12%, with and without reinforcement. The mortality was 4% for a time delay between perforation and treatment of less than 24 hours and increased to 14% for a delay more than 24 hours [1]. A primary repair may thus be hazardous in the presence of friable tissue, which is typically observed after delayed diagnosis, after previous unsuccessful surgical or endoscopic attempts to close the leak and in the presence of mediastinitis. Moreover, adequate debridement may result in defects of more than one third of the esophageal circumference that may then result in narrowing of the esophageal lumen after oversewing [23]. In these circumstances, surgical options consist in esophageal resection with immediate or delayed reconstruction, exclusion and diversion procedures, or drainage alone in critically ill patients. Esophageal resection is a valid therapeutic option in the presence of concomitant esophageal obstruction or when previous surgical interventions have failed to control mediastinal sepsis. Although esophageal resection with immediate reconstruction may be associated with low mortality in specialized centers [23], it is generally associated with a postoperative mortality of 17% [1]. Esophageal exclusion and diversion procedures are advocated for patients with severe mediastinal contamination but are associated with a 24% postoperative mortality. Drainage alone has been recommended for critically ill patients but bears a mortality rate as high as 37% [1].
The therapeutic dilemma of intrathoracic esophageal leakage in the presence of mediastinal and systemic sepsis may be resolved by the muscle flap onlay technique, in which pedicled muscle flaps are sutured into the debrided esophageal defect without attempting a primary repair. That allows a safe closure of the debrided esophageal defect without risk of esophageal narrowing or leakage and avoids esophageal resection or exclusion in critically ill patients. In addition, the muscle can be wrapped around the reconstructed structures and the debrided mediastinum. However, this technique requires relatively large flaps with mechanical stability because they are used as esophageal wall substitutes. Intrathoracic tissues such as pericardium, pleura, intercostal muscle flaps, or omentum that may be used for primary repair reinforcement are less likely to fulfil these criteria.
Richardson and coworkers [26] have successfully treated delayed (more than 48 hours) esophageal perforations by the technique of a muscle flap onlay approach. Experimental studies on a feline model have shown that esophageal defects as much as 60% of the circumference can be closed with muscle flaps without stricture formation and that after a few weeks, the muscle flaps were covered by a malpighian epithelium with preservation of oesophageal function [27].
Our results endorse these previous findings. Nineteen patients underwent successful closure of intrathoracic esophageal leaks as long as 12 cm by pedicled muscle flaps in the presence of mediastinal and systemic sepsis with no postoperative mortality. Restoration of esophageal integrity without further interventions was obtained in 84% of the patients, and endoscopic and histological evaluation during follow-up revealed epithelialization of the muscle patch. However, 1 patient required deployment of a permanent stent for postoperative stenosis and succumbed from gastrointestinal bleeding due to stent erosion 12 months after the operation. In this patient, treated at the beginning of our experience, only a strip of diaphragm was created for esophageal closure and might have led to traction on the esophageal reconstruction with subsequent stricture formation. In all other patients undergoing repair by use of a diaphragmatic flap, a hemicircumferential desinsertion of diaphragm at the level of the anterior and lateral chest wall was performed through a thoracoabdominal incision, and no subsequent stricture development was observed.
Our series included also 12 patients with intrathoracic esophageal leaks after previous surgery such as esophagectomy, lung resections, or other surgical interventions at the level of the gastroesophageal junction, with and without gastric resection. Five of these patients had previous unsuccessful surgical or endoscopic attempts to control intrathoracic leakage before undergoing muscle flap onlay repair. In all these patients, restoration of the esophageal and intestinal integrity was achieved. Of particular interest were the 6 patients with fulminant intrathoracic anastomotic leak after esophagogastric or esophagointestinal reconstructions after Lewis esophagectomy or gastrectomy. Because some degree of ischemia was observed at the site of the anastomosis, the defects encompassed as much as three quarters of the anastomotic circumference after debridement in healthy tissues. Reconstruction of the defects by muscle flap onlay technique without attempting a primary reanastomosis resulted in permanent restoration of intestinal continuity without further leakage or stenosis.
Contained intrathoracic anastomotic leaks have also been successfully treated by nonoperative management [28] or by self-expanding covered stents [2]; however, fulminant leaks and poorly or noncontained leaks require prompt operative intervention. Local resection and primary repair of the anastomosis with additional buttressing may be considered in focal necrosis but a takedown of the anastomosis with cervical esophagostomy is recommended in the presence of more extensive anastomotic disruption [2]. The muscle flap onlay technique may be an attractive alternative for patients with fulminant intrathoracic anastomotic leaks and gross mediastinal sepsis after esophagectomy or other surgical interventions in which takedown of the anastomosis and restoration of gastrointestinal continuity at a later date may otherwise be required.
In conclusion, intrathoracic esophageal leaks may be efficiently closed by use of pedicled muscle flaps in the presence of mediastinal and systemic sepsis and in situations where a primary repair seems risky owing to the friability of tissues. In addition, this technique may be an attractive alternative in patients with fulminant intrathoracic anastomotic leaks and gross mediastinal sepsis after esophagectomy or other surgical interventions. However, because the muscle flap onlay approach may require additional dissection and operation time in already critically ill patients and the sacrifice of a functioning diaphragm, it should be considered as a salvage procedure in situations for which less invasive procedures are not applicable or have failed.
| References |
|---|
|
|
|---|
Related Article
This article has been cited by other articles:
![]() |
J. Liu, X. Zhang, D. Xie, A. Peng, X. Yang, F. Yu, and D. Liu Acute Mediastinitis Associated with Foreign Body Erosion from the Hypopharynx and Esophagus Otolaryngology -- Head and Neck Surgery, January 1, 2012; 146(1): 58 - 62. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Diana, M. Hubner, H. Vuilleumier, P. Bize, A. Denys, N. Demartines, and M. Schafer Redistribution of Gastric Blood Flow by Embolization of Gastric Arteries Before Esophagectomy Ann. Thorac. Surg., May 1, 2011; 91(5): 1546 - 1551. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Saadi, J. Y. Perentes, M. Gonzalez, A. C. Tempia, Y. Wang, N. Demartines, H.-B. Ris, and T. Krueger Vacuum-Assisted Closure Device: A Useful Tool in the Management of Severe Intrathoracic Infections Ann. Thorac. Surg., May 1, 2011; 91(5): 1582 - 1589. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. Hofstetter Invited Commentary Ann. Thorac. Surg., September 1, 2009; 88(3): 973 - 973. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |