Secondary Logo

Journal Logo

Primary Aortoenteric Fistula: Should Enhanced Computed Tomography be Considered in the Diagnostic Work-Up?

Wood, Alice, BSc, MBBS*; Bendjelid, Si-Mhamed, MD; Bendjelid, Karim, MD, MS*

doi: 10.1213/01.ANE.0000175769.11560.FE
Critical Care and Trauma: Case Report

We present the case of a 66-yr-old woman who was admitted to a surgical intensive care unit with life-threatening rectal bleeding. Despite the use of angiography and repeated computed tomography scans, the diagnosis of a primary aortoduodenal fistula secondary to an abdominal aortic aneurysm was not made, leading to a delay in diagnosis until the time of surgery. The reliability of radiological investigations and the importance of being alert to the possibility of this extremely rare condition are discussed.

IMPLICATIONS: We describe the delayed diagnosis of an aortoduodenal fistula related to an abdominal aortic aneurysm. We emphasize that computed tomography scan results may be unreliable for detecting this pathology in patients who are hemodynamically unstable.

*Division of Surgical Intensive Care, Department APSIC, University Hospitals of Geneva, Switzerland; †Department of Radiology, Hospital of Moulins, France

Accepted for publication March 30, 2005.

Address correspondence and reprint requests to Karim Bendjelid, MD, MS, Surgical Intensive Care Division, University Hospital of Geneva, CH-1211, Geneva 14, Switzerland. Address e-mail to

In the aortoenteric fistula (AEF) the bowel wall is eroded in the vicinity of the aorta, with the formation of a direct communication between the aorta and the intestinal lumen. Abdominal aortic aneurysms (AAA) account for more than half of all cases of AEF. The diagnosis should be considered in anyone older than 60 yr who presents with major rectal bleeding, especially if the patient has a history of aortic aneurysm. Without surgical intervention, AEF has a mortality rate of nearly 100% (1).

Diagnosis of this entity may be difficult because of the subtleties of its clinical presentation and evolution. Therefore, surgical treatment is often delayed, representing a challenge even for the experienced vascular surgeon. Thus paraclinical examinations for formal identification of AEF in the acute phase are needed. The usual imaging modalities available include angiography and contrast-enhanced computed tomography (CT) scan. Of these, CT appears to be the investigation of choice. Here, we report the case of a patient in whom the diagnosis was not made until surgery, although repeated contrast-enhanced CT scans had been performed.

Back to Top | Article Outline

Case Report

A 66-yr-old woman in good general health presented to her doctor with a 2-mo history of abdominal pain. She underwent an abdominal ultrasound, which was reported as normal. A few days later she passed a small amount of blood per rectum and was admitted to her local hospital for investigation. Esophagogastroduodenoscopy (OGD) was performed, showing esophagitis with no evidence of recent bleeding. The patient was discharged home only to be readmitted 2 days later with profuse rectal bleeding and hypovolemic shock. Her hemoglobin level was low, and she underwent massive transfusion, requiring an epinephrine infusion to maintain mean arterial blood pressure. A contrast-enhanced CT showed an aortic aneurysm but no fistula. OGD showed fresh blood in the duodenum but no active bleeding. The source of the blood was not clear, and the patient was therefore transferred to the surgical intensive care unit for monitoring.

The patient remained relatively stable for 2 days. A further CT scan confirmed the presence of a 6-cm AAA, in contact with the duodenum, with a possible primary aortoduodenal fistula (Fig. 1). Angiography did not show any evidence of contrast extravasation (Fig. 2), which would have indicated active bleeding. Surgery was finally performed 3 days after the episode of rectal bleeding, and a sub-renal AAA, with an aortoduodenal fistula, was found. The aneurysm was repaired and the fistula was closed (endovascular aortic graft placement and closure of an aortoduodenal fistula). The patient had a good recovery.

Figure 1.

Figure 1.

Figure 2.

Figure 2.

Back to Top | Article Outline


This report describes a 66-yr-old woman with no medical history of vascular or gastrointestinal disease who presented with massive gastrointestinal bleeding secondary to a primary AEF (PAEF). Surprisingly, this diagnosis was not made with repeated enhanced CT scan but was only confirmed at the time of operation.

This case highlights two important issues. First, although PAEF are not the most common cause of gastrointestinal bleeds, their existence must not be forgotten. Indeed, PAEF is a life-threatening cause of upper gastrointestinal bleeding that may affect patients with no history of vascular disease. Second, diagnosis of PAEF may be difficult (2) and radiological techniques are not totally reliable.

Secondary AEF, which account for the majority of AEF, are complications of reconstructive abdominal aortic surgery. Commonly, the fistula is the result of infection near the proximal anastomosis between the aorta and the prosthesis. PAEF, which are extremely rare, are usually the result of erosion of a preexisting AAA and rupture into the gastrointestinal tract. This is often associated with an inflammatory or infective aortitis (3), which was not found in our patient; however, PAEF in the absence of inflammation has also been reported (1). In our patient, the assumption is that an undiagnosed atherosclerotic AAA had eroded into the adjacent duodenal wall and eventually led to a gastrointestinal bleed. Indeed, atherosclerosis remains the most common etiology, accounting for more than 3/4 of the cases reported (4). Moreover, para-aortic radiation therapy can, in uncommon cases, lead to PAEF with histological signs of radiation damage (4).

It is common for AEF to present initially with a minor, so-called “herald” bleed, which is later followed by a catastrophic, life-threatening bleed (1). This was the sequence of events in our patient’s case. It is also possible that, in her case, the two episodes of bleeding may have been unrelated. The first, minor, episode may have been a result of the documented esophagitis, whereas the second may have been a complication of the OGD. Visceral perforation is a known complication of OGD (although esophageal perforation is far more common than duodenal) (5); it is less common in diagnostic than in therapeutic OGD. Logically, if visceral perforation can occur, an AEF can also be formed, although this complication has not been reported (5). Unfortunately, in our patient’s case, it was impossible to tell whether the AEF was genuinely primary or secondary to the initial OGD.

Both PAEF and secondary AEF are difficult to diagnose, and the same diagnostic tests are used (although suspicion should be increased in patients with a previous AAA repair). More information is available regarding the use of diagnostic tests in secondary AEF. Therefore, in the discussion of AEF and diagnostic tests these two conditions will be considered together.

In the past, OGD was considered to be the most useful diagnostic test (6,7), although it is not, in fact, reliable in establishing the diagnosis (8). Aortoangiography has also been used but is not always useful (9), as noted in our patient. Nevertheless, this technique may be useful to define arterial anatomy and to allow the planning of aortic reconstruction. At present, enhanced CT is thought to be the diagnostic modality of choice (10). However, in our patient even this examination did not definitively confirm the communication between duodenum and aorta (Fig. 1). Our finding is in line with a few reports in which CT scan did not help and in which the diagnosis was made at the time of emergency laparotomy (11,12).

The sensitivity and specificity of CT scan in detecting AEF have not been prospectively studied (13–15). Two studies have compared CT scan findings with operative results and concluded that this examination has a sensitivity of 94% and a specificity of 85% in detecting AEF (14,16). One report demonstrated that a three-phase CT (delayed scan) is able to visualize the extravasating contrast media from the aorta to the duodenum and the late opacification of the bowel by blood coming from the aorta (17). Indeed, the presence of dye in the duodenum when IV contrast has been given has already been used to affirm the diagnosis of PAEF (18). In our patient, it was later noted that on each occasion that the CT scan was performed, the patient’s mean arterial blood pressure was <60 mm Hg. This may have accounted for the lack of sensitivity of this examination in this particular patient—the decrease in the gradient between systemic arterial blood pressure and abdominal pressure may have affected the flow of the contrast between the aneurysm and the duodenum.

In conclusion, even if CT scan is at present considered to be a valuable tool for diagnosing PAEF, suspicion is the most important diagnostic aid to prevent the overlooking of this rare and often fatal disease. In the context of hemodynamic instability, emergency surgery should be performed in all cases of gastrointestinal bleeding associated with an abdominal aneurysm.

Back to Top | Article Outline


1. Lemos DW, Raffetto JD, Moore TC, Menzoian JO. Primary aortoduodenal fistula: a case report and review of the literature. J Vasc Surg 2003;37:686–9.
2. Saillen P, Mosimann F, Friedlender J. Aorto-enteric fistulas: report of 12 cases. J Chir (Paris) 1991;128:290–3.
3. Kelliher C, Kavanagh DO, Belton F, et al. Primary aorto-duodenal fistula following staphylococcal septicaemia. Eur J Vasc Endovasc Surg 2004;27:679–81.
4. Drognitz O, Pfeiffenberger J, Schareck W, et al. Primary aortoduodenal fistula as a late complication of para-aortic radiation therapy: a case report. Chirurg 2002;73:633–7.
5. Hart R, Classen M. Complications of diagnostic gastrointestinal endoscopy. Endoscopy 1990;22:229–33.
6. Marquez Galan JL, Herrera JM, Vega P, et al. Aorto-duodenal fistula: endoscopic diagnosis. Rev Esp Enferm Dig 1990;78:96–8.
7. Vollmar JF, Kogel H. Aorto-enteric fistulas as postoperative complication. J Cardiovasc Surg (Torino) 1987;28:479–84.
8. Dossa CD, Pipinos II, Shepard AD, Ernst CB. Primary aortoenteric fistula: Part I. Ann Vasc Surg 1994;8:113–20.
9. Peck JJ, Eidemiller LR Aortoenteric fistulas. Arch Surg 1992;127:1191–3.
10. Mark AS, Moss AA, McCarthy S, McCowin M. CT of aortoenteric fistulas. Invest Radiol 1985;20:272–5.
11. Ikeda K, Abe T, Itou M, et al. Successful surgical treatment of primary aorto-duodenal fistula associated with inflammatory abdominal aortic aneurysm: a case report. Ann Thorac Cardiovasc Surg 1999;5:194–7.
12. Pabst TS 3rd, Bernhard VM, McIntyre KE Jr, Malone JM. Gastrointestinal bleeding after aortic surgery: the role of laparotomy to rule out aortoenteric fistula. J Vasc Surg 1988;8:280–5.
13. Kukora JS, Rushton FW, Cranston PE. New computed tomographic signs of aortoenteric fistula. Arch Surg 1984;119:1073–5.
14. Mark A, Moss AA, Lusby R, Kaiser JA. CT evaluation of complications of abdominal aortic surgery. Radiology 1982;145:409–14.
15. Raptopoulos V, Cummings T, Smith EH. Computed tomography of life-threatening complications of abdominal aortic aneurysm: the disrupted aortic wall. Invest Radiol 1987;22:372–6.
16. Low RN, Wall SD, Jeffrey RB Jr, et al. Aortoenteric fistula and perigraft infection: evaluation with CT. Radiology 1990;175:157–62.
17. Sklair-Levy M, Spira RM, Simanovsky N. Primary aorto-duodenal fistula: evaluation with computed tomography. Eur Radiol 1999;9:1005.
18. Lichtenstein S, Singh S, Dhruva VN. Primary aortoenteric fistula in the emergency department. J Emerg Med 2002;23:209–10.
© 2005 International Anesthesia Research Society