Aortobronchial fistula: a rare etiology for hemoptysis. (Original Article).

Abstract

Aortobronchial fistula is an extremely rare cause of hemoptysis. Aortobronchial fistula occurs in patients who have a history of thoracic vascular surgery. Because its symptoms are nonspecific, a high index of suspicion is critical if the physician is to detect it. The results of imaging studies (e.g., plain films, computed tomography, and angiography) and bronchoscopy are sometimes, but not always, diagnostic--another reason the diagnosis is difficult. Left untreated, mortality in patients with aortobronchial fistula is 100%. Patients can be salvaged by a variety of techniques, including the placement of an endovascular stent. We describe the case of a 52-year-old man who came to us with hoarseness and hemoptysis, which proved to be underlying symptoms of aortobronchial fistula. He was treated successfully.

Introduction

Hemoptysis, defined as "the spitting of blood derived from the lungs or from the bronchial tubes," can be life threatening. (1) The common causes of hemoptysis are bronchogenic carcinoma and other bronchial tumors, bronchiectasis, and pulmonary tuberculosis. (2) One condition that can mimic hemoptysis is epistaxis that drains posteriorly and is then coughed up. A rare cause of hemoptysis is an aortobronchial fistula.

An aortobronchial fistula is an anomalous communication between the aorta and a segment of the tracheobronchial tree. This is a rare condition, as only 71 cases were reported prior to 1995. (3) The primary symptom of aortobronchial fistula is hemoptysis. Some patients with aortobronchial fistula experience intermittent hemoptysis that eventually progresses to a massive and often fatal episode. Timely diagnosis and surgical intervention can be life saving, but the correct diagnosis of aortobronchial fistula is established premortem in only 44 to 56% of cases. (4) In this article, we describe a case of aortobronchial fistula that was treated successfully, and we review the literature on this subject.

Case report

A 52-year-old military veteran came to the Carl T. Hayden Medical Center ENT Clinic in 1999 for evaluation of hoarseness and hemoptysis. He had brought with him ajar full of clots that he had collected following an episode of hemoptysis that had occurred that morning. While in the examination room, the patient experienced a dramatic episode of hemoptysis, but he remained hemodynamically stable. Upon further questioning, he revealed that he had been experiencing intermittent hemoptysis for the previous 7 weeks--generally on a daily basis. He also experienced chest pain that radiated into his back. He had recently undergone esophagoduodenoscopy (results were negative), and his primary care physician subsequently started him on an [H.sub.2] blocker. His surgical history included a graft repair of a traumatic thoracic aneurysm in 1984. He later developed a pseudoaneurysm, which was repaired in 1988.

Our examination revealed that the patient's voice was hoarse and that he had a left vocal fold paralysis. Pertinent lab data--including his platelet count and International Normalized Ratio---were normal. We suspected an aortobronchial fistula and obtained a computed tomography (CT) scan. CT demonstrated an abnormal lucency that was suspicious for a fistulous tract.

The patient was transferred to the intensive care unit at the Arizona Heart Hospital for management of the suspected aorta bronchial fistula. Plain chest films were negative, as was bedside transesophageal echocardiography (TEE). Bronchoscopy detected no active bleeding, but it did reveal an abnormal area in the superior segment of the left lower lobe bronchus. Angiography was subsequently performed, and an aortic wall defect was noted just distal to the distal anastomosis.

The patient was taken to surgery, where he underwent bi-iliac angiography. The diameter of the right iliac artery was adequate for the introduction of an Excluder endoprosthesis. Intraoperative intravenous sonography along the length of the aorta revealed the presence of a defect in the aortic wall just past the distal anastomosis. An anterior jet was noted on TEE. A 28 x 7-cm endoluminal Excluder endoprosthesis was placed with the proximal end at the level of the take-off of the left subclavian artery. The graft extended 3 cm past the distal end of the old graft. Balloon angioplasty was performed, and the anterior jet previously seen on TEE disappeared. No defect or leak was seen on repeat intraoperative angiography. Postprocedural CT showed no evidence of leak and confirmed that the graft was situated in the proper position. The patient had an uneventful postoperative course, and he was discharged home in stable condition.

Discussion

The otolaryngologist is often consulted to help determine the etiology of a particular case of hemoptysis. If the source of the hemorrhage is unclear, the upper aero-digestive tract must be thoroughly examined. Our patient was initially thought to have a laryngeal tumor, but after a precise examination, the source was believed to be in the tracheobronchial tree.

Girardet first described an aortobronchial fistula in 1914. (5) In 1934, Keefer and Mallory reported six cases of aortobronchial fistula in a series of autopsies. (6) In 1947, Jones performed the first successful repair of an aortobronchial fistula in an 11-year-old girl who had undergone ligation of a patent ductus arteriosus. (6) 1959, DeProphetis repaired an aortobronchial fistula that had been caused by a ruptured thoracic aortic aneurysm, but his patient later died from a recurrent rupture. (6) Davey successfully repaired an aortobronchial fistula in 1962. (7)

Most aortobronchial fistulas involve the descending thoracic aorta and the left tracheobronchial tree. Some cases of involvement of the right pulmonary lobe and left mainstem bronchus have been reported. (8) Aortobronchial fistulas arise as a complication of an aneurysm or pseudoaneurysm, which can be attributed to any of several etiologic processes. Prior to 1960, the most common infectious causes were tuberculosis, syphilis, and fungal infections. (4) Since then, most cases have occurred in the context of atherosclerosis. Other causes of aneurysm and pseudoaneurysm include reactions to sutures and graft materials and suture dehiscence. A pseudoaneurysm can develop as a result of a failed anastomosis (e.g., following patch aortoplasty or graft replacement), prosthetic infections, or erosion of a graft into the lung parenchyma. (9) An aortobronchial fistula that is caused by a traumatic aneurysm, such as the one that occurred in our patient, is rare.

The most common predisposing factor in the development of an aortobronchial fistula is surgical repair of a thoracic aneurysm, particularly one that involves patch aortoplasty. (10,11) Aneurysms often occur at the level of the ligamentum arteriosum just distal to the take-off of the left subclavian artery. Other surgical factors associated with the later development of aortobronchial fistula are coarctation of the aorta, (11) correction of congenital abnormalities (e.g., in a patient with a patent ductus arteriosus), (12) and stent placement for pediatric bronchomalacia. (13) An aortobronchial fistula can occur anytime between 4 months and 17 years following such a procedure.

As an aneurysm expands, it causes a pulsatile compression of the adjoining structures; this occurred in the tracheobronchial tree in our patient. Inflammation ensues, which leads to adherence and pressure necrosis and ultimately results in fistula formation. When the fistula is new, the small amount of bleeding associated with it generally causes hemoptysis of small amounts of blood. Once clot formation occurs, the hemoptysis ceases. However, if the clot becomes lysed or dislodged, hemoptysis will recur and the process will repeat itself; this is called intermittent hemoptysis. Meanwhile, the fistula continues to grow and eventually becomes large enough that a hemorrhage through it becomes life threatening. (6) Such an event is sometimes associated with cardiac arrest. (6) The average length of time between an initial episode of hemoptysis and a fatal episode is 5 months. (3)

The early symptoms of aortobronchial fistula are nonspecific, which contributes to the difficulty in diagnosing it and which can lead to a delay in treatment. Intermittent hemoptysis is the most common symptom, occurring in 95% of patients. (14) Intermittent hemoptysis is usually followed by massive hemoptysis, which is characterized by the expectoration of 300 to 400 ml of blood during a 24-hour period; massive hemoptysis occurs in nearly 50% of patients. (6) Unfortunately, a massive episode is sometimes the first and only manifestation of the disease.

Other symptoms include chest pain (45% of patients), cough (44%), and dyspnea (38%). (8) The differential diagnosis includes such entities as neoplasm, bronchiectasis, pulmonary abscess, tuberculosis, bronchitis, pulmonary embolism, and fungal infection. A careful history and examination will often serve to rule out these conditions.

A number of imaging studies can aid in the diagnosis, including plain chest films, CT, and angiography. Bronchoscopy is another option to consider.

Chest films demonstrate an aneurysm in only about 50% of patients. X-ray findings are usually nonspecific (e.g., infiltrates, masses, consolidation, and the presence of pleural fluid). (5)

CT, particularly helical CT, is frequently used to identify an aortobronchial fistula, and some authors recommend it as the procedure of choice. (15) On CT, an aortic aneurysm and the ground-glass appearance of the adjacent lung suggests the presence of an aortobronchial fistula, although the fistula itself is usually not seen. But here again, detection rates approach only 50%. (3) CT can also detect soft-tissue densities that represent clots.

Angiography is more sensitive in detecting the aneurysm itself. It can visualize vascular leaks and the presence of contrast media in the bronchial tree. However, the clot often prevents the flow of contrast from the aorta into the bronchial tree, and thus angiography yields falsenegative results in many instances. Angiography also carries the risk of dislodging a clot during catheter manipulation. Still, some authors believe angiography is the most definitive of these studies, while others reserve it for situations where CT is negative or equivocal. (3)

Bronchoscopy can identify the source of the bleeding, but the fistulous tract might not be readily visible. This invasive procedure carries the risks of perforation and clot dislodgment, which could lead to an exsanguinating hemorrhage. (6) Therefore, it is imperative that the means are in place to take the patient to operation emergently if necessary. Yet despite its risks, some authors advocate bronchoscopy as the diagnostic procedure of choice. (4)

Prior to the use of stenting techniques, surgeons performed open procedures via a thoracotomy or sternotomy. The usual approach was through a left thoracotomy where the aortic end of the fistula can be corrected by patch closure, direct repair, grafting, or subclavian artery flap repair. Current endovascular techniques, as described in our case, offer a safer, more rapid procedure for repairing a life-threatening aortobronchial fistula. The successful use of endovascular aortic stent grafts, which has been described elsewhere, represents an optional approach to patients whose operative risk during an open procedure would be prohibitively high. However, the long-term efficacy of this approach has yet to be defined.

When aortobronchial fistulas are diagnosed and repaired in a timely manner, survival rates of 73 to 80% have been reported (8) There have been reports of recurrence 1 to 2 years following the reparative procedure, but they are extremely rare. In affected patients whose condition is undetected, mortality is 100%. This certainty underscores the importance of having a high index of suspicion when faced with a patient who might have an aortobronchial fistula and the importance of minimizing any delay in diagnosis.

References

(1.) Stedman's Medical Dictionary. Boston: Houghton Mifflin, 1995:364.

(2.) English GM, ed. Otolaryngology. Philadelphia: Harper and Row, 1987.

(3.) Foster CL., Kalbhen CL, Demos TC, Lonchyna VA. Aorobronchial fistula occurring after coarctation repair: Findings on aortography, helical CT, and CT angiography. AJR Am J Roentgenol 1998;171:401-2.

(4.) Szolar DH, Riepl T, Stiskal M, Preidler KW. Aortobronchial fistula as a late complication of posttraumatic chronic aortic aneurysm. AJR Am J Roentgenol 1995; 164:1511-3.

(5.) Kazerooni EA, Williams DM, Abrams GD, et al. Aortobronchial fistula 13 years following repair of aortic transection. Chest 1994;106:1590-4.

(6.) MacIntosh EL, Parrott JC, Unruh HW. Fistulas between the aorta and the tracheobronchial tree. Ann Thorac Surg 1991;51:515-9.

(7.) Ishizaki Y, Tada Y, Takagi A, et al. Aortobronchial fistula after an aortic operation. Ann Thorac Surg 1990;50:975-7.

(8.) Demeter SL, Cordasco EM. Aortobronchial fistula: Keys to successful management. Angiology 1980;31:431-5.

(9.) Milano A, De Carlo M, Mussi A, et al. Aortobronchial fistual after coarctation repair and blunt chest trauma. Ann Thorac Surg 1999;67:539-41.

(10.) Riancho JA, Gutierrez JA, Echevarria S, et al. Aortobronchial fistula: An uncommon cause of haemoptysis. Postgrad Med J 1989;65:171-3.

(11.) Garniek A, Morag B, Schmahmann S, Rubinstein ZJ. Aortobronchial fistula as a complication of surgery for correction of congenital aortic anomalies, Radiology 1990;175:347-8.

(12.) Wood AE, Stevenson HM, Cleland J. Aortobronchial fistula: A late complication of division of the persistent ductus arteriosus. Ann Thorac Surg 1984;38:402-5.

(13.) Cook CH, Bhattacharyya N, King DR. Aortnbronchial fistula after expandable metal stent insertion for pediatric bronchomalacia, J Pediatr Surg 1998;33:1306-8.

(14.) Miyazaki M, Hiraga S, Kitamura M, et al. Aortobronchial fistula complicated with an aortic aneurysm in hemodialysis patient. Nephron 1990;56: 101-2.

(15.) Urschel JD. The diagnostic importance of computed tomography in aortobronchial fistula--a case report. Angiology 1993;44:817-9.

From the Department of Surgery, Carl T. Hayden VA Medical Center, Phoenix.

Reprint requests: Randy Oppenheimer, MD, Department of Surgery, Carl T. Hayden VA Medical Center, 650 E. Indian School Rd., Phoenix, AZ 85012. Phone: [phone]; fax: [phone]: c-mail: [e-mail]