ÌÇÐÄvlog

The patient underwent an esophagogastroduodenoscopy to identify the source of the gastrointestinal bleeding, and graft material was seen at the base of a duodenal diverticulum with no active bleeding, indicating an aortoduodenal fistula (Figure 1). Understanding the risks, the patient refused any surgery. She was treated with a nasogastric tube and intravenous vancomycin hydrochloride and Zosyn (piperacillin sodium and tazobactam sodium; Wyeth, Madison, New Jersey). Her peritonitis resolved within 4 days. She then agreed to a palliative repair of the endoleak (Figure 2A and C). She was treated with an Excluder bifurcated endograft (W. L. Gore & Associates, Inc, Flagstaff, Arizona). A second Excluder bifurcated endograft was inserted into the first endograft, with the orientation such that the second device was rotated 180° compared with the first device, ensuring complete overlap of the 2 main bodies and converting the bifurcated device into an aorto-uni-iliac endograft. A completion angiogram showed complete resolution of the type IIIb endoleak (Figure 2B). The patient tolerated the procedure well without any complications and with complete resolution of the melanotic stool and endoleak confirmed by a computed tomographic angiographic scan with 3-dimensional reconstruction (M2S Inc, West Lebanon, New Hampshire). The patient was discharged from the hospital with oral antibiotics. At 1-year follow-up, the patient was doing well, without any complaints of abdominal pain or gastrointestinal bleeding, and tolerating a normal diet. There is no evidence of an endoleak or an expansion of the aneurysm on the 1-year follow-up computed tomographic scan with 3-dimensional reconstruction (Figure 2D).

Complications of stent-graft implantation include endoleak, endograft migration, aneurysmal rupture, distal embolization, contrast allergy, renal insufficiency, wound complications such as hemorrhage, infection, and, in rare circumstances, the development of a secondary aortoenteric fistula (AEF), which is a devastating complication with a high mortality rate.¹ Bergqvist et al² identified 14 cases of AEF. The etiology of the AEF has been explained as a direct injury to the bowel due to the defective stent graft (in 5 cases), the erosion of the bowel (in 2 cases), or an expanding aneurysm with endoleak or rupture (in 6 case); in 1 case, however, the etiology was unclear.² In our case, the AEF could be the result of continued expansion of the aneurysm because of a persistent type IIIb endoleak through a hole in the body of the endograft and because of the aneurysm's close proximity to the duodenum with local inflammation. The standard treatment would include the removal of the impaired stent graft and either in situ aortic replacement with a new graft or aortic closure in combination with an axillobifemoral bypass procedure.² The mortality rates for patients who undergo either procedure are as high as 59%. Our 90-year-old patient who underwent hemodialysis was clearly not an acceptable candidate for open repair. Endovascular repair of an AEF may offer rapid control of bleeding and may help to avoid severe morbidity and high mortality associated with the complicated open procedures.³ Antoniou et al⁴ demonstrated the association of endovascular repair with a high incidence of infection postoperatively. They identified 33 reports with 41 cases of AEF. There was a 44% risk of hemorrhage and/or infection after a mean follow-up period of 13 months. This risk of infection was almost 3-fold higher in secondary AEF than in primary AEF. Although our patient continues to do well without evidence of bleeding or infection at 1-year follow-up, she is still at significant risk of recurrent infection and/or bleeding.

In summary, AEF is one of the most difficult vascular complications to treat, with a very poor outcome. Although open repair continues to be the gold standard, endovascular repair may offer rapid control of bleeding, may help to stabilize the patient, and may offer a bridge to a better-planned complicated open repair. In rare circumstances, endovascular repair might be the only endurable intervention for severely debilitated and/or elderly patients.

Return to .

The Editor welcomes contributions to the Image of the Month. Manuscripts should be submitted via our online manuscript submission and review system (). Articles and photographs accepted will bear the contributor's name. Manuscript criteria and information are per the Instructions for Authors for Archives of Surgery (). No abstract is needed, and the manuscript should be no more than 3 typewritten pages. There should be a brief introduction, 1 multiple-choice question with 4 possible answers, and the main text. No more than 2 photographs should be submitted. There is no charge for reproduction and printing of color illustrations.

Correspondence: Mahmoud B. Malas, MD, Department of General Surgery, Division of Vascular Surgery, Johns Hopkins University, 4940 Eastern Ave, Bldg A/5, Baltimore, MD 21224 (bmalas1@jhmi.edu).

Accepted for Publication: May 22, 2010.

Author Contributions:Study concept and design: Malas, Glebova, and Perler. Acquisition of data: Malas, Choo, Qazi, Meguid, and Reifsnyder. Analysis and interpretation of data: Malas and Freischlag. Drafting of the manuscript: Malas, Choo, Qazi, Glebova, Meguid, and Perler. Critical revision of the manuscript for important intellectual content: Malas, Reifsnyder, Perler, and Freischlag. Administrative, technical, and material support: Malas, Choo, Qazi, Meguid, and Reifsnyder. Study supervision: Malas, Perler, and Freischlag.