Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Aortoenteric fistula (AEF) is defined as a communication between the aorta and gastrointestinal (GI) tract. AEF is classified as primary or secondary based on the underlying cause leading to the fistula development. Primary AEF is a communication between the native aorta and GI tract; secondary AEF is a communication between a reconstructed aorta (for either aneurysmal or occlusive disease) and the GI tract.
Sir Astley Cooper was the first to describe a primary AEF in a publication in 1829. He referred to it as a “sometimes but serious complication of an aneurysmal aorta.” Brock published the first report of a secondary AEF in 1953, wherein he described a fistula between the proximal anastomosis of an aortic homograft and the duodenum. Heberer performed the first repair of a primary AEF in 1954 by primary closure, and MacKenzie et al. performed the first successful repair of a secondary AEF in 1958.
AEF presents a particularly complicated challenge for the vascular surgeon. When it is left untreated, the outcome is almost universally fatal. However, surgical repair is fraught with complications, and despite continued advances in medicine and critical care, morbidity and mortality rates remain high.
In addition to the difficulties encountered with surgical treatment, AEF is notoriously difficult to diagnose, and as such, a high index of suspicion is required in approaching any patient with GI hemorrhage and history of aortic disease. This chapter reviews the pathogenesis, etiology, diagnostic methods, and treatment options for AEF.
The incidence of primary AEF has been reported at 0.04% to 0.07% in large autopsy series. The body of literature available on the subject of primary AEF is small, with approximately 251 cases available in the published literature. Despite the rarity of primary AEF in the population overall, the incidence in patients with aneurysms of the abdominal aorta is 0.69% to 2.36%. In the majority of cases (83%), an aneurysmal aorta is associated with primary AEF; foreign bodies, tumors, radiotherapy, infection (historically due to tuberculosis and syphilis but now most commonly caused by Klebsiella and Salmonella ), and GI tract disease (peptic ulcer disease and perforating biliary stones) account for the remainder of AEF. The mean diameter of the aorta with primary AEF is 6.2 cm; the mean age of patients is 64 years, with a male-to-female ratio of 3:1.
The most commonly described GI tract location for primary AEF is the third and fourth portion of the duodenum (54%). It is presumed that this is due to the tethering effect of the ligament of Treitz, leaving this portion of the duodenum exposed to the direct pulsatile pressure of the aorta. Primary AEF has also been described in the following locations: esophagus (28%), small and large bowel (15%), and stomach (2%).
Overall, the pathogenesis of primary AEF is uncertain. The proposed mechanisms are mechanical, infectious, and inflammatory ( Fig. 51.1 ). In the majority of cases, the mechanical component is caused by the pulsatile pressure of an expanding aorta against the wall of the GI tract. This leads to local compression and ischemia, with weakening of the wall and eventual erosion with fistula formation.
Secondary AEF is more common than primary AEF and is related to prior open vascular surgery and endovascular repair, with an incidence of 0.36% to 1.6% reported after open abdominal aortic graft reconstruction. The interval from aortic reconstruction to onset of symptoms is on average 2 to 6 years after graft placement. Secondary AEF involves fistulization between the GI tract and a prior vascular reconstruction for either aneurysmal or occlusive disease. Although it is typically described as occurring in the setting of a synthetic graft, secondary AEF has also been reported with aortic homograft reconstruction and allografts. Similar to primary AEF, the most common location described for secondary AEF is the distal duodenum and proximal jejunum. However, secondary AEF has been described at multiple GI sites depending on the location of the prosthetic graft.
Secondary AEF has been further classified based on the location of the fistula regarding the suture line of the graft ( Fig. 51.2 ). A fistula that has a direct communication between the arterial circulation and the GI tract at the level of the suture line is classified as a graft enteric fistula. Communication between the GI tract and the graft interstices (but not at the suture line) is referred to as a graft enteric erosion. Both entities are highly morbid complications that require urgent management; however, the etiology and presentation may differ. The graft enteric fistula’s involvement of the suture line will disrupt the arterial anastomosis and causes dramatic hemorrhage, whereas graft enteric erosion may be manifested first with infectious symptoms as a result of direct contact of the graft material with the GI tract, allowing bacterial translocation to occur between the interstices of the graft from the GI tract.
The etiology of secondary AEF includes both patient and surgical factors that occur during the initial placement of the graft. A history of multiple vascular procedures, wound complications, infection, emergency operation, and technical error have been found to predispose to secondary AEF formation.
The proposed mechanisms of secondary AEF pathogenesis are infection, pulsatile pressure, and technical error.
Infection and downstream inflammation appear to play a significant role in the development of AEF. Studies lack uniformity in description of methods particularly in regard to bacterial culture which may differ between blood stream, graft, and aortic wall within these patients. Infection may result at the time of initial surgery, or through secondary infection and bacteremia. Further, varying immune and inflammatory features of patients may contribute to the development of graft infection and subsequent AEF. Best evidence suggests that secondary AEFs after open aortic reconstructions are polymicrobial in nature in upwards of two-thirds of patients. Importantly, improved culture methods have demonstrated that yeast, particularly Candida , appear frequently in the setting of secondary AEF. Candida is also present in two-thirds of patients, with similar rates in patients who have undergone endovascular as compared to open intervention. , Graft infection leads to local inflammation and results in breakdown of the suture line, pseudoaneurysm (PSA) formation, and eventual rupture into surrounding structures. It is important to acknowledge that characteristics of bacterial species, such as ability to produce collagenase and other locally destructive proteases, may contribute to the development of fistula complications. More research is required with improved bacterial methods to confirm this statement.
Pulsatile pressure also remains a factor in the proposed mechanism of AEF formation. As described above, pressure from a noncompliant prosthesis against the bowel wall may lead to ischemia to the surrounding tissue and eventual erosion. Another mechanism described is suture line disruption, leading to the formation of an expanding PSA, compressing surrounding structures and eventually eroding into the bowel. Despite the intuitive nature of this mechanism, the majority of studies continue to point to the issue of infection as the root cause of AEF.
The final factor in the formation of AEF is technical error. This includes inoculation of the prosthesis at time of implantation as well as duodenal injury, serosal thinning, and ischemia during the operation. Changes in standards in patient preparation and appropriate antibiotic administration intraoperatively have decreased the overall rate of aortic graft infection. Operative trauma to the bowel, however, remains an issue. The bowel can be injured during the initial dissection and exposure either directly with sharp injury and heat transfer from cautery or indirectly by over-retraction and tension. Retraction injury may occur from the retractor blades directly; however, it may also occur by over-compression of the bowel, causing ischemia, or by allowing the bowel to become dehydrated outside of the body. Methods to reduce bowel trauma include wrapping the bowel in warm moist laparotomy sponges during the procedure (being mindful to re-soak the sponges intermittently), keeping the bowel intraperitoneal by packing it above the liver (which can be difficult in larger patients), and using an isolation bag to keep the bowel moist and protected (but still visible) while it is outside the abdomen. Another method to reduce bowel trauma is by using a retroperitoneal approach and interposing tissue between the aorta and the graft at the end. Proponents of the retroperitoneal approach believe that avoiding division of the parietal peritoneum to expose the aorta limits the devitalization of tissue and development of adhesions between the graft material and bowel. Regarding interposition of tissue between the aorta and graft, aortic sac closure followed by a second layer of retroperitoneal tissue and parietal peritoneum is the preferred approach. If that is unavailable, a flap of greater omentum or a layer of bovine pericardium patch material may be used. Omental coverage may be necessary more commonly in end-to-side graft anastomoses to the aorta, rather than in end-to-end anastomoses, and in thin patients.
There are an increasing number of reports documenting AEF after endovascular aneurysm repair (EVAR) and thoracic endovascular aneurysm repair (TEVAR). Proposed mechanisms in the case of endovascular repair include persistent endoleak with growth of the residual sac, multiple coiling efforts to repair an endoleak, erosion of the stent-graft through the aorta, endotension, and infection at the time of graft placement. , The Multicenter Study on Aortoenteric Fistulization After Stent Grafting of the Abdominal Aorta (MAEFISTO) evaluated 3932 patients who underwent EVAR between 1997 and 2013 at eight Italian centers with EVAR programs. Of the patients who underwent EVAR at the participating centers, 22 presented with AEF on follow up, 15 of which underwent EVAR for atherosclerotic aneurysmal disease and 7 for post-surgical PSA. On subgroup analysis, they found that the incidence of AEF development of EVAR was 0.46% in patients undergoing EVAR for atherosclerotic aneurysmal disease and 3.9% when performed for PSA. Anastomotic PSA as indication for EVAR, and urgent/emergency EVAR were significantly associated with AEF development. There were no significant associations between choice of endograft and fistula formation. The increased risk of AEF after EVAR for PSA does suggest that a preexisting infectious process or even subclinical AEF formation may have already been present at the time of EVAR placement in these cases. In urgent/emergency surgery, the presence of a hematoma compressing surrounding structures or the overall increase in local inflammation in emergency repair may contribute to the formation of AEF. Most recently, systemic review of known AEFs after EVAR revealed that approximately 37% were associated with an endoleak and/or persistent sac expansion. Interestingly analysis of these devices identified that 30% of AEFs after EVAR were associated with a defect in the aortic stent graft that includes fracture, erosion, and angulation. As our endovascular experience continues to grow, it appears technical success at the initial operation may be an important factor in the prevention of AEF formation.
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here