Endovascular repair of bleeding aortoenteric fistulas: A 5-year experience James A. Burks, Jr, MD, Peter L. Faries, MD, Edwin C. Gravereaux, MD, Larry H. Hollier, MD, and Michael L. Marin, MD, New York, NY Purpose: Aortoenteric fistula (AEF) is an uncommon but catastrophic complication that can occur either primarily or after aortic reconstruction. Untreated, it is uniformly fatal. Conventional surgical management is associated with a perioperative mortality rate of 25% to 90% and frequent major complications. We reviewed our experience with the endovascular treatment of both primary and secondary AEFs in an effort to determine whether endovascular repair is a less morbid alternative to traditional surgical treatment in select patients. Methods: In a 5-year period, seven high-risk patients who had bleeding and an AEF documented by means of radiology or endoscopy (2 primary, 5 secondary) were treated with coil embolization (1) or placement of an endovascular aortic stent graft (3 aortouniiliac, 2 tube, 1 bifurcated). One patient underwent computed tomography (CT)-guided percutaneous catheter drainage of an infected perigraft collection. The average follow-up period was 27 months (range, 1166 months), and follow-up consisted of physical examination, complete blood count, and contrast-enhanced helical CT scanning at 3, 6, and 12 months and yearly thereafter. All patients were treated with intravenous antibiotics perioperatively and were prescribed life-long oral antibiotics on discharge. Results: There was one perioperative death (14%) caused by fungal sepsis. Persistent sepsis after stent-graft placement necessitated laparotomy and bowel resection in one patient. One patient had three bouts of recurrent sepsis that were successfully treated with a change of antibiotic. There were three late deaths (43%) unrelated to the procedure or AEF. Three patients (43%) were alive and well an average of 36 months (range, 23-67 months) after the procedure, with no clinical or radiologic evidence of recurrent bleeding or infection. Conclusion: Endovascular management of AEFs is technically feasible and may be the preferred treatment in select patients with bleeding and no signs of sepsis. In the setting of gross infection, it may also be considered in high-risk patients as a bridge to more definitive treatment after hemodynamic stabilization and optimization. (J Vasc Surg 2001;34:1055-9.)
Aortoenteric fistula (AEF) remains a catastrophic complication with a high morbidity and mortality, despite numerous strategies for surgical treatment. Primary AEFs are uncommon and occur when a large, previously untreated aneurysm erodes into the adjacent bowel. The more common secondary AEF occurs in 0.3% to 1.6%1-8 of patients after aortic prosthetic reconstruction. Conventional treatment of an AEF consists of extra-anatomic bypass grafting, aortic ligation, and, for a secondary AEF, subsequent graft removal. This technique is associated with a 25% to 90% operative mortality rate,5-8,9-19 an amputation rate of 5% to 25%,9-15 and an aortic stump rupture rate of 10% to 50%.10 In-situ aortic reconstruction has been proposed as a lessmorbid alternative to ligation and bypass grafting. However, the average perioperative mortality rate for insitu reconstruction is reported to be 27% to 30%.9,10,20 From the Division of Vascular Surgery, Department of Surgery, Mount Sinai School of Medicine. Competition of interest: nil. Presented at the Twenty-fifth Annual Meeting of the Southern Association for Vascular Surgery, Rio Grande, Puerto Rico, Jan 24–27, 2001. Reprint requests: Michael L. Marin, MD, Henry Kaufman Professor of Vascular Surgery, Department of Surgery, Mount Sinai School of Medicine, 5 East 98th Street, Box 1273, New York, NY 10029 (e-mail:
[email protected]). Copyright © 2001 by The Society for Vascular Surgery and The American Association for Vascular Surgery. 0741-5214/2001/$35.00 + 0 24/1/119752 doi:10.1067/mva.2001.119752
Several recent case reports describe the endovascular management of AEFs in high-risk patients and suggest lower perioperative morbidity and mortality rates than traditional surgery.21-25 This study documents the use of endovascular techniques as a means of treatment for seven patients with primary and secondary AEFs in a 5-year period at a single institution. PATIENTS AND METHODS Demographics. In a 5-year period, 467 patients underwent endovascular repair of an abdominal aortic aneurysm at the Mount Sinai Medical Center and were entered prospectively into a vascular registry. All patients were considered for endografting in accordance with an investigator-sponsored Investigational Device Exemption from the US Food and Drug Administration, with the approval of the Mount Sinai Medical Center Institutional Review Board. Informed consent was obtained from all patients. Of these, seven patients were treated for an AEF (Table I). There were six men and one woman, whose ages ranged from 48 to 92 years (mean, 76 years). All patients were considered to be at high-risk for surgical repair because of significant comorbid disease (Table II). Two patients had a primary AEF, and five patients had a secondary AEF. All five patients with a secondary AEF had their initial operation for an abdominal aortic aneurysm or thoracoabdominal aneurysm, four electively and one emergently because of aneurysm rupture. The mean inter1055
JOURNAL OF VASCULAR SURGERY December 2001
1056 Burks et al
A
B
Fig 1. A, CT scans showing the loss of the fat plane between bowel and aorta (grey arrow) and the extravasation of contrast at the distal anastomosis in this same patient (black arrow), 8 years after tube graft repair of an AAA. B, Diagram (left) illustrating the angiogram (right) that was performed via a catheter placed from the right femoral artery. The aortouniiliac stent-graft has been deployed from the left femoral artery. Before placement of an excluder in the right common iliac artery, this angiogram, which shows free extravasation of contrast from the excluded aortic sac to the bowel lumen, was performed.
val between initial repair and occurrence of an AEF in these five patients was 8 years (range, 5-12 years). Five patients had fistulization to the duodenum or small intestine, one to the thoracic esophagus after thoracoabdominal aneurysm repair and one to the sigmoid colon. All patients had active gastrointestinal bleeding; four of these patients were in frank hemorrhagic shock, defined as a systolic blood pressure less than 100 mm Hg and a heart rate of more than 120 beats per minute. Three patients had clinical evidence of infection, defined as a temperature of more than 101°F, white blood cell count of more than 10,000, or perigraft fluid or air found by means of a computed tomography(CT) scan. All patients underwent upper and lower endoscopy before intervention. All AEFs were documented either by means of endoscopic or radiology (Table III). Endoscopic confirmation consisted of visualization of the graft and was diagnostic in three of the seven patients. Radiologic confirmation consisted of either extravasation of contrast on aortic angiog-
A
B
Fig 2. A, Mesenteric arteriography identifies a large celiac artery aneurysm in a patient with multiple recurrent episodes of gastrointestinal bleeding. B, Selective catheterization of the aneurysm identifies a communication with the duodenum. Thrombogenic coils are placed within the aneurysm.
raphy, which was seen in two patients, or by the loss of the fat plane between bowel and aorta on abdominal CT scan (Fig 1), which was diagnostic in six of the seven patients. Procedures and devices. The devices used included Gianturco coils in one patient, custom-made Parodi/ Palmaz aortouniiliac devices in three patients, Talent LPS (Medtronic, Minneapolis, Minn) tube grafts in two patients, and the bifurcated Gore Excluder (W.L. Gore, Flagstaff, Ariz) in one patient. One patient was treated with thrombogenic coils placed percutaneously (Fig 2). The remaining six patients were treated with an aortic stentgraft placed with fluoroscopic guidance via bilateral femoral artery cutdowns. Device selection was made on the basis of patient anatomy. One patient underwent CTguided percutaneous drainage of the aneurysm sac immediately after stent-graft placement because of the presence of air and fluid within the sac noted on a preoperative CT scan (Fig 3). An indwelling catheter was left for 6 weeks and was irrigated daily with an antibiotic solution. All patients received life-long suppressive oral antimicrobial therapy. The antibiotics were selected based on blood culture and sensitivity data. Recurrent sepsis required a change of antibiotic agent because of new microbial culture and sensitivity data. Patients were given intravenous
JOURNAL OF VASCULAR SURGERY Volume 34, Number 6
Burks et al
1057
Table I. Presentation, procedures, and outcomes Patient
Type
Location
1
1°
2
2°
3
1°
Celiac aneurysm to duodenum Distal anastomosis to sigmoid colon AAA to duodenum
4
2°
5 6 7
Proximal anastomosis to duodenum 2° Distal anastomosis to small intestine 2° Proximal thoracic anastomosis to esophagus 2° Proximal anastomosis to duodenum
Interval* (years) Sepsis Unstable Device§ –
Adjunctive procedures
Major complications
Follow-up
–
+
4
–
–
67 months
+
+
1
–
–
–
1
Diverting ileostomy –
Persistent sepsis –
12
–
–
3
–
–
Died at 18 months (CVA) Died at 13 months (MI) 34 months
5
–
–
1
–
–
23 months
7‡
+
+
2
–
8
+
+
2
9†
Persistent sepsis, Died at 26 death days (MOSF) CT-guided Persistent Recurrent sepsis × 3; drainage of AAA sac sepsis, ATN died at 11 months (MI)
*For 2° AEF, time interval between initial aortic surgery and occurrence of AEF. †Initial surgery for ruptured AAA. ‡Initial surgery for thoracoabdominal aortic aneurysm §Device: 1, Aortouniiliac; 2, Talent Dacron tube endograft; 3, Gore Excluder bifurcated endograft; 4, Gianturco coils. AAA, Abdominal aortic aneurysm; CVA, cerebrovascular accident; MI, myocardial infarction; MOSF, multiple organ system failure; CT, computed tomography; ATN, acute tubular necrosis.
antibiotics and then changed to life-long oral antibiotics when signs of sepsis resolved. All patients were prepared and draped in a sterile manner for possible conversion to open repair. Anesthesia consisted of locally injected lidocaine in three patients, epidural anesthesia in three patients, and general endotracheal anesthesia in one patient. Followup studies included a physical examination, hematocrit level, white blood cell count, and contrast-enhanced spiral CT scanning at 3, 6, and 12 months and yearly thereafter. RESULTS In a 5-year period, endovascular techniques were used as a means of treating seven patients with an acutely bleeding AEF (Table I). Follow-up included all seven patients and ranged from 11 to 67 months (mean, 27 months). All patients had immediate cessation of bleeding after stentgraft placement. The three patients who had obvious signs of graft infection either died or required adjunctive measures to control persistent sepsis. One patient died in the perioperative period of overwhelming fungal sepsis. Persistent sepsis necessitated a proximal diverting ileostomy on postoperative day 2 in another patient. Signs of sepsis quickly resolved, and he was discharged home 3 weeks later. A third patient with a large perigraft collection containing pockets of air (Fig 3) underwent percutaneous CT-guided catheter drainage immediately after stent-graft placement. The catheter was irrigated daily with an antibiotic solution. Signs of sepsis resolved in 2 weeks, and the patient was discharged home. The percutaneous drain was removed 6 weeks later. Three subsequent bouts of recurrent sepsis were successfully treated with a change of antibiotic. There were three late deaths unrelated to the procedure, bleeding, or graft infection caused by myocardial infarction, stroke, and pneumonia. Three patients were alive and well at 63, 30, and 19 months postoperatively without clinical evidence of bleeding or graft infection and with normal hematocrit levels and white blood cell counts and no evidence of
Table II. Contraindications to conventional surgical repair Patient Age/sex 1 2 3 4 5 6 7
Contraindications
48/male Morbid obesity, total gastrectomy, subtotal colectomy, ileostomy, granulating open abdomen 73/male Earlier aortic arch reconstruction and aortic valve replacement, CAD 82/male CAD, HTN, > 10 cm AAA 92/male 99% left main stenosis 73/female Unreconstructible 3-vessel CAD 76/male CAD, COPD, CRI, earlier TAA repair 88/male CAD, COPD, CRI, thoracic and suprarenal aortic aneurysm
CAD, Coronary artery disease; HTN, hypertension; AAA, abdominal aortic aneurysm; COPD, chronic obstructive pulmonary disease; CRI, chronic renal insufficiency; TAA, thoracoabdominal aneurysm.
aortic pathology found by means of CT scanning. A comparison with open repair was not performed. There were no early or late amputations, embolic events, or graft thromboses. There were no recurrent AEFs or aortic ruptures. DISCUSSION There is little consensus about the optimal management of AEFs. Historically, reported morbidity and mortality rates have been disturbingly high. When patients are hemodynamically unstable and actively bleeding, results are even more dismal. Although endovascular management of AEFs is clearly not definitive, it does provide rapid control of bleeding with minimal physiologic insult to the patient. Most of these patients are elderly and physiologically compromised, with multiple comorbid conditions and a limited life expectancy. In our study, three of six patients (50%) who survived the procedure were dead within 2 years because of cardiopulmonary disease unrelated to their AEF. The other three patients were alive at a
JOURNAL OF VASCULAR SURGERY December 2001
1058 Burks et al
A
Table III. Positive preoperative diagnostic studies Patient 1 2 3 4 5 6 7
Helical CT – + + + + + +
Angiography
Endoscopy
+ – – – + – –
– + – – + + –
CT, Computed tomography.
B
C
Fig 3. A, Air surrounding a previously placed aortic graft (arrow). B, Predeployment (left) and postdeployment (right) intraoperative angiography. A Talent tube graft was placed just below the renal arteries, excluding the earlier proximal anastomosis that was the site of the aortoenteric fistula. C, Percutaneous drainage catheter (arrow) placed with CT guidance immediately after graft deployment.
mean of 37 months postoperatively and had no evidence of infection. Without removal of all aortic prosthetic and debridement of all infected tissue, eradication of infection is not likely to occur. However, with broad spectrum antibiotics and percutaneous drainage, a long-lasting suppression of the infection is possible.
However, all the major complications in this series were caused by sepsis and occurred in the three patients with obvious signs of graft infection. The only perioperative death occurred in this group, in a 76-year-old man with chronic obstructive pulmonary disease and coronary artery disease who was transferred to our institution with upper gastrointestinal bleeding, fever, and an elevated white blood cell count 7 years after elective repair of a thoracoabdominal aortic aneurysm. The diagnosis of a thoracic AEF was confirmed by means of upper endoscopy. He was taken to the operating room, and a Talent tube graft was placed, with immediate cessation of bleeding and hemodynamic stabilization. Postoperatively, progressive sepsis developed, and 26 days later he died of overwhelming Aspergillus fungemia and multiple organ system failure. Only five patients have been reported to have survived this uncommon and frequently lethal condition.26 With mediastinal esophageal contamination, endovascular repair is an unlikely means of providing adequate treatment unless combined with wide drainage and esophageal repair or diversion. In retrospect, uncontrolled sepsis in this patient was likely caused by mediastinitis similar to that seen with primary esophageal perforation. In instances of mediastinal sepsis, definitive open management should follow initial endovascular control of hemorrhage. The other two septic complications were successfully managed with adjunctive procedures after successful stentgraft deployment, cessation of bleeding, and hemodynamic optimization. One patient was treated with CT-guided percutaneous drainage of the infected aneurysm sac. The other patient had an aortosigmoid fistula and required a diverting ileostomy performed in the distal ileum to minimize further enteric soilage of the graft. Neither patient had any further bleeding, and both died at home of causes unrelated to graft infection or gastrointestinal bleeding (pneumonia, stroke). A review of the literature discloses five previously reported cases of endovascular repair of AEFs. Two of these patients were reportedly alive and well, without evidence of infection, after 6 and 18 months.24,25 One patient died after 10 months of myocardial infarction, without evidence of recurrent infection or bleeding.23 One patient underwent successful axillobifemoral bypass grafting followed by device removal and aortic ligation because of recurrent infection with fever and chills 8
JOURNAL OF VASCULAR SURGERY Volume 34, Number 6
months after stent-graft placement.21 One patient underwent successful device removal and aortic homograft placement 4 weeks after stent-graft placement because of the discovery of air within the aneurysm sac by means of CT scanning.22 None of these patients died or sustained significant morbidity as a result of aortic stent-graft placement. Three of these five patients may need no further intervention. The two patients in whom endovascular treatment “failed” had elective surgery under optimal physiologic conditions rather than emergent surgery in the setting of hemorrhagic shock, acute renal insufficiency, and myocardial ischemia. The results of this study suggest that endovascular aortic stent-graft placement provides a safe and effective method of rapidly controlling an acutely hemorrhaging AEF. Without debridement of all contaminated prosthetic and retroperitoneal tissue, it is not definitive therapy. The newly placed prosthetic endograft is contaminated by gut flora. Complete eradication of infection is probably impossible. However, with life-long antibiotic treatment, longterm suppression of sepsis may be feasible. As seen in this series, once bleeding is controlled, many patients die of coexistent cardiopulmonary disease before infection or aortic degeneration becomes a significant problem. Adjunctive treatments, such as percutaneous drainage and proximal bowel diversion, may be necessary to contain sepsis in the setting of gross aortic graft infection. Patients with fistulas solely to segments of the small bowel were successfully treated without diversion. However, in the patient with a colonic fistula, diversion ultimately became necessary. The patient with an esophageal fistula died of mediastinitis, but might have benefitted from esophageal diversion had it been performed. Those patients in whom these measures fail can still be converted to traditional surgical repair at a later date, on an elective basis in an optimized physiologic state. This may translate into an improved outcome for this unfortunate group. However, no subsequent conversions have been required in our experience. CONCLUSION The treatment of AEFs has improved in recent years. However, despite the multiple surgical techniques reported, many of these patients do not survive or are left debilitated after treatment. Endovascular repair allows for hemodynamic stabilization, and when combined with broad spectrum antibiotics and drainage or diversion when necessary, sepsis may be controlled for prolonged periods. Although not curative, endovascular repair may provide a bridge to more definitive repair at a later time or provide long-term palliation in patients who are too sick to undergo major surgery. REFERENCES 1. Elliott JP, Smith RF, Szilagyi DE. Aortoenteric and paraprostheticenteric fistulas. Arch Surg 1974;108:479-90. 2. Thompson WM, Jackson DC, Johnsrude IS. Aortoenteric and paraprosthetic-enteric fistulas: radiologic findings. AJR Am J Roentgenol 1976;127:235-42.
Burks et al
1059
3. Plate G, Hollier LA, O’Brien P, Pairolero PC, Cherry KJ, Kazmier FJ. Recurrent aneurysms and late vascular complications following repair of abdominal aortic aneurysms. Arch Surg 1985;120:590-4. 4. Bergqvist D, Alm A, Claes G, et al. Secondary aortoenteric fistulas— an analysis of 42 cases. Eur J Vasc Surg 1987;1:11-8. 5. Bergeron P, Espinoza H, Rudondy P, et al. Secondary aortoenteric fistulas: value of initial axillofemoral bypass. Ann Vasc Surg 1991;5:4-7. 6. Champion MC, Sullivan SN, Coles JC, Goldbach M, Warson WC. Aortoenteric fistula. Incidence, presentation, recognition and management. Ann Surg 1982;195:314-7. 7. Higgins SD, Steed DL, Julian TB, Makaroun MS, Peitzman AB, Webster MW. Management of aortoenteric paraprosthetic fistulae. J Cardiovasc Surg 1990;31:81-6. 8. Peck JJ, Eidemiler LR. Aortoenteric fistulas. Arch Surg 1992;127: 1191-4. 9. Pipinos II, Carr JA, Haithcock BE, Anagnostopoulos PV, Dossa CD, Reddy DJ. Secondary aortoenteric fistula. Ann Vasc Surg 2000;14: 688-696. 10. Kuestner LM, Reilly LM, Jicha DL, Ehrenfeld WK, Goldstone J, Stoney RJ. Secondary aortoenteric fistula: contemporary outcome using extraanatomic bypass and infected graft excision. J Vasc Surg 1995;21:184-96. 11. McCann RL, Schwartz LB, Georgiade GS. Management of abdominal aortic graft complications. Ann Surg 1993;217:729-34. 12. Perdue GD, Smith RB, Ansley JD, Costantino MJ. Impending aortoenteric hemorrhage: the effect of early recognition on improved outcome. Ann Surg 1980;192:237-43. 13. Trout HH III, Kozloff L, Giordano JM. Priority of revascularization in patients with graft enteric fistulas, infected arteries, or infected arterial prostheses. Ann Surg 1984;199:669-83. 14. Umpleby HC, Britton DC, Turnbull AR. Secondary arterio-enteric fistulae: a surgical challenge. Br J Surg 1987;74:256-9. 15. Yeager RA, McConnell DB, Sasaki TM, Vetto RM. Aortic and peripheral prosthetic graft infection: differential management and causes of mortality. Am J Surg 1985;150:36-43. 16. Jacobs MJHM, Reul GJ, Gregoric I, Cooley DA. In-situ replacement and extra-anatomic bypass for the treatment of infected abdominal aortic grafts. Eur J Vasc Surg 1991;5:83-6. 17. Lundbom J, Andresande E, Myhre HO. Aortoduodenal fistula reconstructed with autogenous tissue: case report. Acta Chir Scand 1989;155:545-7. 18. O’Mara CS, Williams GM, Ernst CB. Secondary aortoenteric fistula: a 20-year experience. Am J Surg 1981;142:203-9. 19. Ricotta JJ, Faggioli GL, Stella A, et al. Total excision and extraanatomic bypass for aortic graft infection. Am J Surg 1991;162: 145-9. 20. Gordon LL, Hagino RT, Jackson MR, Modrall JG, Valentine RJ, Clagett GP. Complex aortofemoral prosthetic infections: the role of autogenous superficial femoropopliteal vein reconstruction. Arch Surg 1999;134(6):615-20. 21. Chuter TAM, Lukaszewicz GC, Reilly LM, Kerlan RK, Faruqi R, Sawhney R, et al. Endovascular repair of a presumed aortoenteric fistula: late failure due to recurrent infection. J Endovasc Ther 2000;7:240-244. 22. Curti T, Freyrie A, Mirelli M, Rossi C, Paragona O, Resta F, et al. Endovascular treatment of an ilio-enteric fistula: a “bridge” to aortic homograft. Eur J Vasc Endovasc Surg 2000;20:204-6. 23. Kinney EV, Kaebnick HW, Mitchell RA, Jung MT. Repair of mycotic paravisceral aneurysm with a fenestrated stent graft. J Endovasc Ther 2000;7:192-7. 24. Grabs AJ, Irvine CD, Lusby RJ. Stent-graft treatment for bleeding from a presumed aortoenteric fistula. J Endovasc Ther 2000;7:236-9. 25. Deshpande A, Lovelock M, Mossop P, Denton M, Vidovich J, Gurry J. Endovascular repair of an aortoenteric fistula in a high-risk patient. J Endovasc Surg 1999;6:379-84. 26. Bogey WM, Thomas JH, Hermreck AS. Aortoesophageal fistula: report of a successfully managed case and review of the literature. J Vasc Surg 1992;16:90-5. Submitted May 23, 2001; accepted Aug 2, 2001.
