Ann Thorac Surg 2000;69:967-970
© 2000 The Society of Thoracic Surgeons
Current Review
Surgical management of primary aortoesophageal fistula secondary to thoracic aneurysm
Michael J. Reardon, MDa,
Robert J. Brewer, MDa,
Scott A. LeMaire, MDa,
John C. Baldwin, MDa,
Hazim J. Safi, MDa
a Department of Surgery, Baylor College of Medicine, The Methodist Hospital, Houston, Texas, USA
Address reprint requests to Dr Reardon, Department of Surgery, Baylor College of Medicine, 6560 Fannin, Suite 1619, Houston, TX 77030
e-mail: reardonm{at}bcm.tmc.edu
 |
Abstract
|
|---|
Aortoesophageal fistula, secondary to thoracic aortic aneurysm, is an uncommon cause of gastrointestinal bleeding that is uniformly fatal without surgical intervention. These may be primary fistulas, in cases of thoracic aortic aneurysm without previous repair, or secondary fistulas occurring after surgical repair of thoracic aortic aneurysm. Surgical treatment has been successful in a small number of cases of primary aortoesophageal fistula, secondary to thoracic aortic aneurysm, but techniques used have varied. We report a successful repair of primary aortoesophageal fistula, secondary to descending thoracic aortic aneurysm, and review the evolution of management since the three previously reported successful repairs at our institution.
|
Introduction
|
|---|
Aortoesophageal fistula is a rare cause of upper gastrointestinal bleeding with catastrophic consequences without treatment. Described causes of this entity include thoracic aortic aneurysm, foreign body, malignancy, postoperative trauma, and others [1]. Thoracic aortic aneurysm is the cause of aortoesophageal fistula in over half of the reported cases [1]. These include primary fistulas in cases of unoperated thoracic aortic aneurysms, and secondary fistulas occurring after surgical repair of thoracic aortic aneurysms. The entity was first described in 1818, in a soldier who swallowed a piece of rib, which perforated his descending thoracic aorta, causing death [2]. The Chiari triad of midthoracic pain, herald arterial bleed, and final exsanguination after a symptom-free period, represents the typical clinical presentation [3]. The symptom-free interval, ranging from hours to months, allows the opportunity for diagnosis and definitive surgical intervention. The first two successful repairs of primary aortoesophageal fistula, secondary to thoracic aortic aneurysm, were reported from our institution in 1983 [4], and a third in 1990 [5]. We report a fourth successful repair and review the evolution and current status of our management of this entity.
|
Case report
|
|---|
A 48-year-old female presented with midthoracic pain and hematemesis of bright red blood. She required a transfusion of five units of packed red blood cells. She had a history of a type B aortic dissection 6 months prior to admission treated medically. Upper endoscopy, performed prior to transfer, demonstrated an extrinsic mass compressing the midesophagus with overlying clot (Fig 1). The patient was transferred to our institution with a diagnosis of aortoesophageal fistula.
On admission to our institution, the patient was not actively bleeding and was hemodynamically stable. She was admitted to the intensive care unit where appropriate venous access and monitoring catheters were inserted. No prior imaging studies accompanied the patient. She was scheduled for total aortogram, but had an episode of massive hematemesis accompanied by hypotension. She was transferred immediately to the operating room where she underwent a left posterior lateral thoracotomy. A 7-cm descending thoracic aortic aneurysm was noted with a neck just distal to the left subclavian artery and a relatively normal caliber aorta at the diaphragm. The aorta was clamped just distal to the left subclavian and at the diaphragm, and the aneurysm was opened longitudinally. A 3-cm aortic fistula to the esophagus was apparent, 10 cm distal to the left subclavian artery. The descending thoracic aneurysm was replaced with a 28-mm Dacron graft (Hemashield graft; Boston Scientific, Boston, MA) from just distal to the left subclavian to the tenth intercostal space. The aorta cross-clamp time was 21 minutes. The esophageal opening was carefully identified and debrided, and a nasogastric tube was passed by the anesthesia service beyond the opening, into the stomach under direct vision. The esophageal opening was then repaired with a running 3-0 polypropylene suture and the repair bolstered with aneurysm wall. The chest and mediastinum were copiously irrigated and drained with two 36F chest tubes. The left thoracotomy was closed in standard fashion. The patient was placed in the supine position, and a lateral cervical esophagostomy was performed in the left neck for proximal diversion with the nasogastric tube remaining in place for distal decompression (Fig 2).
View larger version (22K):
[in this window]
[in a new window]
|
Fig 2. Artist’s rendition of the repair technique, illustrating the primary repair of the esophagus and buttressing with aortic aneurysm wall, proximal diversion with lateral cervical esophagostomy, distal drainage through nasogastric tube, and in situ repair of thoracic aneurysm.
|
|
Total parenteral nutrition was begun in the immediate postoperative period, and she was maintained on perioperative antibiotic coverage of vancomycin, ceftazidine, and clindamycin. Seven days postoperatively, a gastrograffin esophagram was performed through the lateral esophagostomy, which showed no esophageal leak. To allow for enteral nutrition, a gastrostomy was placed on postoperative day 8. The patient was discharged on gastric tube feeds and intravenous clindamycin for a 6 week course of intravenous antibiotics.
Six weeks postoperatively, a follow-up esophagram demonstrated healing of the fistula without leak or stricture. Her cervical esophagostomy was closed at this time, and she tolerated a regular diet without swallowing difficulties. She is currently on long-term treatment with oral clindamycin without evidence of graft infection, normal swallowing, and is doing well at 1 year follow-up.
|
Comment
|
|---|
Repair of aortoesophageal fistula, secondary to thoracic aortic aneurysm, with long-term survival, continues to be an unusual occurrence with most patients dying prior to diagnosis. Early surgical survival is plagued with late septic complications and death secondary to the esophageal fistula. In addition to our current case, we have identified 9 cases from the English literature of surgical repair of primary aortoesophageal fistula secondary to thoracic aortic aneurysm with long-term survival (Table 1) [4–11]. Survival depends on establishing an early and accurate diagnosis, successful surgical approach to both the aortic aneurysm and the esophageal fistula, as well as appropriate antibiotic coverage. The evolution of our strategy, over the 4 successful cases from our institution, is discussed.
An early diagnosis is important in allowing surgical repair, since the period between the herald bleed and the final exsanguinating hemorrhage is unpredictable, and historically most patients have died prior to diagnosis [12]. Endoscopy is the procedure of choice in the initial evaluation of significant upper gastrointestinal bleeding. Esophagogastroduodenoscopy is the most sensitive and specific test for aortoesophageal fistula. Endoscopy in aortoesophageal fistula secondary to thoracic aortic aneurysm will usually demonstrate esophageal compression at the 25 to 30-cm level from the incisors. A pulsating mass covered with clot is often seen, as in the current case. A high index of suspicion for aortoesophageal fistula is imperative, as biopsy of this mass will likely lead to fatal hemorrhage. For patients with massive hemorrhage, endoscopy within an operating room suite is recommended. Radiographic exams are also useful in the evaluation of aortoesophageal fistula, particularly in patients without significant active bleeding. Barium swallow esophagram, total aortography, computed tomographic and magnetic resonance scanning of the chest do not routinely demonstrate the fistula, but often delineate and define pathology known to be associated with aortoesophageal fistula, and help in planning repair. We recommend endoscopy as the primary initial measure in establishing the diagnosis of aortoesophageal fistula and computed tomographic scanning of the chest to delineate associated pathology, in this case the descending aortic aneuysm.
After the diagnosis of aortoesophageal fistula is confirmed, surgical correction should be undertaken. Repair consists of two phases: 1) management of the thoracic aortic aneurysm and control of bleeding through the fistula as the initial concern; and 2) management of the esophageal tear. Repair of the thoracic aortic aneurysm is complicated by the fact that aortoesophageal fistula represents an infected surgical field. Early in our experience with infected grafts in the descending thoracic aortic position, we excised the grafts and established an extra anatomic bypass from the ascending aorta to the abdominal aorta [13]. Due to the complexity of this approach, and the success of treating aortoduodenal fistula with in situ graft replacement, we recommend an in situ replacement of the thoracic aortic aneurysm as has been done in all of the reported cases. Although we often use adjuncts, such as atrial femoral bypass and cerebral spinal fluid drainage, in certain cases of extensive descending thoracic and thoracoabdominal aortic aneurysm, the emergent and usually unstable nature of these patients makes this difficult. This usually requires a clamp and sew approach to stop the bleeding from the fistula as expeditiously as possible, and remove the immediate risk of death. If the patient is not actively bleeding at the time of diagnosis, and has a Crawford type I or II aneurysm, we would recommend consideration of atrial femoral bypass and cerebral spinal fluid (CSF) drainage [14].
Treatment of the esophageal tear, to allow swallowing and avoid the infectious complications, has been more problematic. The two basic approaches are primary repair and esophageal resection. Extensive esophageal wall necrosis may preclude repair and necessitate esophageal resection. This adds significantly to the surgical procedure in these critically ill patients, and necessitates reconstruction at the time of operation or later. Resection was done in the first 2 cases reported from our institution, 1 after a failed primary repair [4]. When resection is necessary and reconstruction contemplated, we would recommend cervical esophagogastrostomy, when possible, to decrease the risk of intrathoracic sepsis if an anastomotic leak should occur. Primary repair will be feasible in some patients presenting with less esophageal wall necrosis. When successful, this will simplify the procedure and allow better swallowing than an esophageal reconstruction. However, unsuccessful repair risks sepsis and death [6]. Primary repair was done in the last case reported from our institution [5] and the current one. We recommend this approach when technically feasible. When primary repair is done, we recommend measures to attempt to minimize the risk of leak and the potentially catastrophic consequences thereof. Cervical lateral esophagostomy is a simple and rapid procedure for proximal diversion, and should be considered in these cases. Subsequent reversal is also simple, safe, and rapid. Tube gastrostomy for distal drainage and omental transposition into the chest to cover the graft, are both useful and reasonable maneuvers if the patient is stable enough to tolerate these extensions of the surgical procedure. Because of instability in the operating room, we choose distal drainage through a nasogastric tube in this case, and avoided the extra dissection needed for omental mobilization. A later gastrostomy was added for enteral feeding during healing and prior to reversal of the cervical esophagostomy. Management of the esophageal pathology should be individualized based on the extent of esophageal destruction.
Although early death is usually the result of bleeding or operative complications, long-term survival depends on avoiding intrathoracic sepsis and graft infection. Antibiotic coverage should be started, beginning with the diagnosis of aortoesophageal fistula, to provide high-dose broad spectrum coverage of oral floral including anaerobes. Subsequent antibiotic choice is based on operative cultures. We use a minimum of 6 weeks of IV antibiotics followed by life-long oral antibiotic suppression due to the devastating consequences of graft infection when it occurs [15].
Aortoesophageal fistula, secondary to thoracic aortic aneurysm, is unusual and always fatal without surgical intervention. Prompt diagnosis followed by expeditious operation is imperative for survival. Endoscopy is our procedure of choice for the diagnosis of aortoesophageal fistula and chest computed tomographic scanning used to delineate associated pathology. We recommend in situ replacement of the thoracic aortic aneurysm, despite the potentially infected field. Esophageal repair should be individualized based on the extent of esophageal wall necrosis. When possible, primary repair simplifies the procedure and results in better long-term swallowing function. When primary repair is performed, adjuncts such as cervical lateral esophagostomy for esophageal diversion should be considered. The evolution of our surgical approach through 4 survivors of this entity is described.
|
References
|
|---|
-
Hollander J.E., Quick G. Aortoesophageal fistula. Am J Med 1991;91:279-287.[Medline]
-
Dubrueil Observations sur la perforation de l’oesophage et de l’aorte thoracique par une potion d’os avale. J Univ Sci Med 1818;9:357-363.
-
Chiari H. Ueber Fremdkorperverletzung des Oesophagus mit Aortenperforation. Berlin Klin Wschr 1914;51:7-9.
-
Snyder D.M., Crawford E.S. Successful treatment of primary aorta-esophageal fistula resulting from aortic aneurysm. J Thorac Cardiovasc Surg 1983;85:457-463.[Abstract]
-
Coselli J.S., Crawford E.S. Primary aortoesophageal fistula from aortic aneurysm. J Vasc Surg 1990;12:269-277.[Medline]
-
Von Oppell U.O., de Groot M., Thierfelder C., Zilla P., Odell J.A. Successful management of aortoesophageal fistula due to thoracic aortic aneurysm. Ann Thorac Surg 1991;52:1168-1170.[Abstract]
-
Bogey W.W., Jr, Thomas J.H., Hermreck A.S. Aortoesophageal fistula. J Vasc Surg 1992;16:90-95.[Medline]
-
Peck J.J., Eidemiller L.R. Aortoenteric fistulas. Arch Surg 1992;127:1191-1194.[Abstract]
-
Tkebuchava T., von Segesser L.K., Turina M.I. Successful repair of primary concomitant aortobronchial and aortoesophageal fistulas. Ann Thorac Surg 1997;63:1779-1781.
-
Goto H., Utoh J., Hongoh H., Hirata T., Kondoh K., Sun L.B., Hara M. Successful treatment of aortoesophageal fistula resulting from aneurysm of the aortic arch. J Cardiovasc Surg (Torino) 1998;39:425-427.[Medline]
-
Amin S., Luketich J., Wald A. Aortoesophageal fistula. Dig Dis Sci 1998;43:1665-1671.[Medline]
-
Myers H.S., Silber W. Oesophageal bleeding from aortoesophageal fistula due to aortic aneurysm. S Afr Med J 1983;63:124-127.[Medline]
-
Crawford E.S., Reardon M.J., Williams T.W. Surgical considerations of infection following operations involving the descending thoracic aorta. World J Surg 1980;4:669-677.[Medline]
-
Safi H.J., Miller C.C., Reardon M.J., et al. Operation for acute and chronic aortic dissection. Ann Thorac Surg 1998;66:402-411.[Abstract/Free Full Text]
-
Yamamura M.S., Gentile A.T., Westerband A., Mills J.L. Surgical management of aortic graft infection. Surgical Rounds 1998;21:274-278.
This article has been cited by other articles:
|
|
|
|
 
T. Inoue, T. Nishino, Y.-F. Peng, and T. Saga
Successful one-stage operation of aortoesophageal fistula from thoracic aneurysm using a rifampicin-soaked synthetic graft
Interactive CardioVascular and Thoracic Surgery,
April 1, 2008;
7(2):
322 - 324.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Kant
Omental Wrap for Aortic Suture Lines
Asian Cardiovasc Thorac Ann,
April 1, 2006;
14(2):
174 - 174.
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Chocron, L. Stoica, S. Koch, J.-F. Bonneville, D. Kaili, B. Heyd, and J.-P. Etievent
Is endovascular stent grafting for descending thoracic aortic disease recommendable in patients with dysphagia?
J. Thorac. Cardiovasc. Surg.,
December 1, 2002;
124(6):
1239 - 1241.
[Full Text]
|
|
|
|
|
|
|
L. Kotsis
Types of esophageal repair after primary aortoesophageal fistula resection
Ann. Thorac. Surg.,
February 1, 2001;
71(2):
759 - 759.
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. J. Reardon
Types of esophageal repair after primary aortoesophageal fistula resection: reply
Ann. Thorac. Surg.,
February 1, 2001;
71(2):
759 - 759.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
