Ann Thorac Surg 2008;85:322-325. doi:10.1016/j.athoracsur.2007.07.069
© 2008 The Society of Thoracic Surgeons
Case Reports
Benign Esophagobronchial Fistula With and Without Esophageal Obstruction: Two Ends of the Surgical Spectrum
Timothy L. Van Natta, MDa,*,
Kalpaj R. Parekh, MDc,
Caitlin G. Reed, MDb,
Saad A. Shebrain, MDa,
Bassam O. Omari, MDa
a Division of Cardiothoracic Surgery, Department of Surgery, Harbor-UCLA Medical Center, Torrance, California
b Division of Infectious Disease, Department of Internal Medicine, Harbor-UCLA Medical Center, Torrance, California
c University of Iowa Hospitals and Clinics, Iowa City, Iowa
Accepted for publication July 24, 2007.
* Address correspondence to Dr Van Natta, Harbor-UCLA Medical Center, Department of Surgery, Box 42, Division of Cardiothoracic Surgery, 1000 W Carson St, Torrance, CA 90509 (Email: timothy.vannatta{at}gmail.com).
 |
Abstract
|
|---|
Acquired esophagobronchial fistula (EBF) is uncommon and its surgical remediation is challenging. Management depends on the cause, degree of pulmonary involvement, and existence of esophageal obstruction. We report management of two EBF cases representing extremes of the surgical spectrum. One patient with EBF secondary to mediastinal fungal infection underwent pulmonary resection and esophageal repair. Another, who was positive for human immunodeficiency virus, required esophageal resection and fistula closure, but no pulmonary resection. Successful outcome was achieved in both patients.
|
Introduction
|
|---|
Benign esophagobronchial fistula (EBF) is an uncommon disorder with a variety of causes. Surgical repair is often required for cure, and complexity of an EBF may pose considerable challenges for the thoracic surgeon. Operative decisions must take into account the cause of the fistula and location, severity of pulmonary involvement, and presence or absence of esophageal obstruction. We present two cases that illustrate these issues.
|
Case Reports
|
|---|
Patient 1
A 25-year-old previously healthy man, employed in poultry processing, was evaluated for chest pain and dyspnea. For months he had relapsing fevers and a productive cough consistently and immediately provoked by eating or drinking. Antibiotics only transiently helped. He was acutely ill on presentation and had diminished left breath sounds. Results of serology for human immunodeficiency virus (HIV) were negative. Chest roentgenogram showed complete left lower lobe consolidation. Chest computed tomography (CT) demonstrated a 7 x 4-cm mid-mediastinal mass extending into the left lower pulmonary lobe, abutting the left atrium. Although the left upper lobe was spared, the lower lobe was essentially destroyed (Fig 1). Endoscopy showed the left lower bronchi were inflamed, more so than the upper ones, and they contained abundant purulent secretions.
View larger version (117K):
[in this window]
[in a new window]
|
Fig 1. A chest computed tomography scan of patient 1 shows left lower lobe destruction and upper lobe preservation.
|
|
A sputum culture grew Streptococcus milleri, and the patient was treated with ampicillin. Results for acid-fast bacillus (AFB) and fungal cultures were negative. A subcarinal biopsy specimen showed only inflammation.
An esophagram revealed a 7-cm tract extending from the midesophagus to the left lower lobe bronchi and no esophageal stenosis (Fig 2). Esophagoscopy verified a fistula to the left at 27 cm from the incisors. Endoscopic ultrasound (EUS) with biopsy confirmed a benign subcarinal mass. Fistula closure and left lower lobe resection were advised. Left thoracotomy would be required for left lower lobectomy, and it was considered likely that an additional right thoracotomy would be needed to address the midesophagus.
View larger version (122K):
[in this window]
[in a new window]
|
Fig 2. An esophagram from patient 1 demonstrates a tract from the esophagus to the left lower bronchial tree.
|
|
At left posterolateral thoracotomy, the upper lobe was unremarkable, but the lower lobe was an inflammatory mass. No malignancy was present. With some difficulty, the esophagus was dissected away from lung. A lower lobectomy was performed, aided by intrapericardial vein isolation. An immediate right posterolateral thoracotomy then addressed the subcarinal esophagus. A fistula tract was identified, extending inferolaterally to the left. Esophageal myotomy allowed assessment of the perforation, which was débrided and repaired (stapled) over a 44F Maloney dilator. The myotomy was closed with sutures and the pleura was interposed between esophagus and carina.
Pathology revealed a necrotizing granulomatous process teeming with Histoplasma capsulatum, consistent with occupational exposure. Itraconazole was given. A postoperative S milleri empyema required rib resection and drainage, after which the patient progressed well.
Patient 2
A 33-year-old man presented with dysphagia and cough exacerbated by liquid ingestion. History was notable for HIV (CD4 count of 256, 25%; never on antiretroviral therapy), and a S mitis brain abscess requiring craniotomy 8 months prior, without residual deficit. Lower extremity venous thrombosis after the craniotomy had prompted Coumadin therapy (Bristol-Myers Squibb, Princeton, NJ).
A chest roentgenogram showed a prominent left lower infiltrate. CT revealed a paraesophageal mass from carina to gastroesophageal junction, subcarinal adenopathy, left lower lobe consolidation, esophageal dilation, and left atrial compression (Fig 3). The patient presented with upper gastrointestinal bleeding shortly thereafter and required transfusion of 2 U of packed red blood cells. Coumadin was stopped, and an inferior vena cava filter was placed.
View larger version (147K):
[in this window]
[in a new window]
|
Fig 3. A computed tomography scan from patient 2 shows left lower lobe consolidation with preservation of much of the parenchyma.
|
|
Esophagoscopy showed inflamed, friable midesophageal mucosa and distal esophageal narrowing that was difficult to traverse. No esophageal perforation was identified. There were no gastroduodenal mucosal abnormalities, and biopsy specimens of the esophageal mucosa showed no malignancy. Bleeding resolved with acid suppression therapy and Coumadin cessation.
A repeat esophagoscopy showed decreased inflammation but marked distal esophageal compression. EUS-guided paraesophageal biopsy yielded purulent fluid and reactive lymph nodes, but no malignancy. Cultures were acid-fast bacilli-negative but grew multiple streptococcal species, Gram-negative rods, and Candida albicans. Antimicrobial drugs were initiated.
An esophagram showed high-grade distal esophageal stenosis with an EBF to the left lower lobe bronchial tree (Fig 4). Surgical repair was advised but initially declined. Esophageal stent placement was offered. The esophagus was pneumatically dilated to 12 to 13.5 mm, and a 16 x 90-mm Polyflex stent (Boston Scientific, Natick, MA) was placed. Contrast studies showed no esophageal leak. Swallowing improved, but the patient presented 2 weeks later with dysphagia, cough, and prerenal azotemia.
View larger version (104K):
[in this window]
[in a new window]
|
Fig 4. An esophagram from patient 2 reveals a fistula from the esophagus to the left lower bronchial tree.
|
|
A chest roentgenogram showed marked proximal stent migration, requiring device removal. Noted on subsequent esophagoscopy were marked inflammation and ischemia of the esophageal mucosa. Two perforations were present, through which contrast again passed into the left bronchial tree.
Esophagectomy through a right posterolateral thoracotomy was necessary. The proximal esophagus was dilated but mobile, but the lower esophagus was encased in dense scar. Neither a discrete mediastinal mass nor acute infection were present. The lower esophagus was resected submucosally because malignancy was absent. By leaving the muscularis propria, the atrium was avoided. The fistula orifices were closed with fibrin glue (Evicel, Johnson & Johnson, Somerville, NJ). Hiatal dissection was performed external to the muscularis, through fibrotic tissue.
After temporary chest closure, the patient was placed supine for gastric mobilization, pyloromyotomy, and jejunostomy. Returning to thoracotomy, a supraazygous esophagogastric anastomosis was made with a No. 31 EEA stapler (US Surgical Corp, Norwalk, CT). Given lack of overt infection in the right pleural space, cervical gastroesophageal anastomosis was unnecessary. The result of a postoperative barium swallow was satisfactory. An oral diet was tolerated well by 1 week, obviating tube feedings. Pathology showed no malignancy, tuberculosis, or fungal disease. The left pneumonitis gradually resolved after many weeks of antibiotic therapy, and he remains well 8 months later.
|
Comment
|
|---|
EBF, congenital or acquired, is uncommon. Its manifestations may be overt or protean. Of signal importance is cough provoked by swallowing liquids [1]. An epithelium-lined tract surrounded by muscularis mucosa and minimal inflammation characterizes congenital EBF [2]. Of the acquired variety, 80% result from chest malignancy [1]. Given generally poor prognosis of malignant EBF, palliative strategies have emerged. Esophageal and airway stents are commonly used to manage symptoms [3].
Benign EBF causes include esophageal and mediastinal inflammatory conditions, trauma, radiation, foreign bodies, and surgical complications. Tuberculosis and fungal disease comprise a large percentage. Benign EBF is amenable to surgical cure [2]. Although stents have been deployed successfully, long-term complications can result [4]. Resection is often preferable. Diagnosis is suggested by history and confirmed by imaging and endoscopic studies. Pulmonary symptoms predominate; however, esophageal obstruction may occur, as in patient 2. A chest roentgenogram indicates pulmonary involvement, and a chest CT defines this and evaluates mediastinal disease. Barium esophagram localizes the fistula and demonstrates esophageal caliber. Endoscopy, with or without ultrasound imaging, further characterizes the fistula and facilitates biopsies.
Griffo and colleagues [2] summarized management keys as sepsis control, nutritional support, respiratory therapy, and surgical resection. Further goals are to establish the cause, determine the need for pulmonary or esophageal resection, or both, and assess for cardiac encroachment. Fistula obliteration is necessary, and this can often be achieved by separating the esophagus from the airway, suturing closed the respective openings, and interposing a pedicled tissue flap [1]. However, pulmonary tissue destruction and esophageal obstruction require more complex approaches. Mediastinal inflammation makes dissection tedious and treacherous. Pulmonary and esophageal resections are correspondingly challenging. For benign disease, esophageal muscularis preservation allows fistula identification and closure while limiting hazards.
The cases presented here represent EBF extremes. Patient 1 had such severe pulmonary inflammatory changes that the lower lobe was essentially destroyed. As in surgical therapy for tuberculosis, a destroyed lobe or lung must be resected at thoracotomy for bronchopleural or other fistulas. Left lower lobe resection required an ipsilateral thoracotomy. It was not possible to approach the subcarinal esophagus from the left, and unfortunately, an additional right thoracotomy was necessary. Intense mediastinal reaction to histoplasmosis likely produced an EBF because of traction forces. Lack of obstruction allowed primary esophageal repair. We favor a stapled repair of mucosa/submucosa for esophageal perforations, with suture closure of the overlying myotomy.
Patient 2 had marked compromise of the distal esophageal lumen and prompt filling of the left bronchial tree on esophagography. However, he had considerably less pulmonary involvement than did the first patient. It was anticipated that with removal of the esophageal communication and appropriate antimicrobial therapy, the lung would mend. The healing mediastinal abscess apparently drew the esophagus into an inflammatory matrix that produced a stenosis. The cause of the mediastinal abscess is unclear, and no primary esophageal disease was demonstrated pathologically. Dilatation and stent placement only exacerbated the disease process, necessitating esophageal resection.
Patient 2 was HIV-positive. Benign EBF has been reported in this patient group, often secondary to tuberculosis [5]. Although successful tuberculosis treatment can eliminate an EBF [5, 6], resection may be necessary. Esophageal resection in HIV patients is rarely reported. Successful esophagectomy for necrotizing esophagitis with EBF in HIV disease has been accomplished [7]. With this article, we report esophagectomy for EBF associated with HIV and an EBF after primary mediastinal infection in this patient population. Efficacy of antiretroviral therapy has changed the course of HIV and acquired immunodeficiency syndrome. With HIV infection becoming a chronic disease, surgeons will increasingly resect the esophagus in these patients [8].
|
References
|
|---|
- Kaul DR, Orringer MB, Saint S, Jones SR. The drenched doctor—a 55-year-old male physician was seen in August because of a 1-week history of fever and night sweats N Engl J Med 2007;356:1871-1876.[Free Full Text]
- Griffo S, Stassano P, Iannelli G, et al. Benign bronchoesophageal fistula: report of four cases J Thorac Cardiovasc Surg 2007;133:1378-1379.[Free Full Text]
- van den Bongard HJ, Boot H, Baas P, Taal BG. The role of parallel stent insertion in patients with esophagorespiratory fistulas Gastrointest Endosc 2002;55:110-115.[Medline]
- Gaissert HA, Grillo HC, Wright CD, Donahue DM, Wain JC, Mathisen DJ. Complications of benign tracheobronchial strictures by self-expanding metal stents J Thorac Cardiovasc Surg 2003;126:744-747.[Abstract/Free Full Text]
- Porter JC, Friedland JS, Freedman AR. Tuberculous bronchoesophageal fistulae in patients infected with the human immunodeficiency virus: three case reports and review Clin Infect Dis 1994;19:954-957.[Medline]
- Alkhuja S, Miller A. Tuberculous bronchoesophageal fistulae in patients infected with the human immunodeficiency virus: a case report and review Heart Lung 1998;27:143-145.[Medline]
- Gaissert HA, Roper CL, Patterson GA, Grillo HC. Infectious necrotizing esophagitis: outcome after medical and surgical intervention Ann Thorac Surg 2003;75:342-347.[Abstract/Free Full Text]
- Issa RA, Podbielski FJ, Fontaine JP, Connolly AE, Walsh WV, Fraire AE. Esophagectomy in a patient with AIDS Dis Esophagus 2004;17:270-272.[Medline]
Top
Abstract
Introduction
