Aortoenteric Fistula and Visceral Artery Aneurysms

Incidence, Demographics, and Causes of Primary Aortoenteric Fistulas

An autopsy study of the general population showed that the incidence of primary AEF is 0.04% to 0.07%. Those with an abdominal aortic aneurysm have a higher incidence, from 0.69% to 2.36%. The majority of primary AEFs occur in the older population, with a mean age of 64 years, and predominantly in men, with a ratio of 3 : 1. Fistulas are most often located in the third and fourth portions of the duodenum (54%), followed by the esophagus (28%), small and large bowel (5%), and the stomach (2%). The higher likelihood of a fistula forming between the aorta and duodenum is attributed to their proximity to one another, as the duodenum directly overlies the aorta within the retroperitoneum.

Prior to eradication of infections with modern medical practice and antibiotics, diseases such as tuberculosis, salmonella, syphilis, and mycoses were the most common causes of primary AEFs. Currently, the most common causes of primary AEFs are atherosclerotic aortic aneurysms. Less common etiologies include radiation, tumors, trauma, and ingestion of foreign material.

Incidence, Demographics, and Subtypes of Secondary Aortoenteric Fistulas

The incidence of secondary AEFs is higher than primary AEFs, with a reported range of 0.36% to 1.6%. Although most secondary AEFs form after an open aortic surgical intervention, a small and increasing number of AEFs have been reported after endovascular aortic repairs. A fistula can form if there is an endoleak with subsequent recurrence of the aortic aneurysm, primary endograft infection, kinking or migration of the graft, or physical breakdown of the graft material. There are two main categories of secondary AEFs. Type 1, or graft-enteric fistula, occurs as the result of an erosion of the proximal aortic suture, with or without the presence of a pseudoaneurysm, into the adjacent bowel. There is direct erosion into the aortic lumen with the potential for massive GI hemorrhage if not repaired emergently. Type 2 secondary AEFs, also known as aortoenteric erosions (AEEs), graft enteric erosions, or paraprosthetic erosions, can occur where the aortic graft mechanically erodes into the overlying bowel. In this type of secondary AEF, no fistula is formed with the lumen of an arterial structure. The bleeding ensues from the edges of the eroded bowel and often presents as a chronic GI bleed.

Pathogenesis

There are several theories as to the pathogenesis of AEFs and AEEs. In primary AEFs, pressure necrosis may occur where the aorta and bowel come into contact with one another, because of the repetitive mechanical trauma from aortic pulsations, which leads to degradation of the nearby tissue.

Secondary AEFs may form due to suture line breakdown, usually from suture technique that fails to adequately incorporate healthy aortic tissue. Failure to close the aneurysm sac and retroperitoneum will also leave the suture line in direct contact with the duodenum, increasing the odds of erosion of the suture line into the bowel. Another hypothesis involves the presence of chronic infection of the reconstructed aorta, which initiates abscess formation with eventual erosion of the graft into the adjacent bowel ( Fig. 88.1 ). Chronic infection may also originate from the GI tract. Foreign-body ingestion, carcinoma, or infections such as diverticulitis can lead to adhesions that bring the bowel in closer proximity to the aorta. The mechanical pulsations cause thinning of the bowel wall with eventual translocation of enteric organisms into the suture line. If a pseudoaneurysm or aneurysm forms adjacent to the suture line, the weakened blood vessel wall leaves it susceptible to breakdown.

Diagnosis

Imaging

After an AEF or AEE is suspected, a diagnostic work-up should be obtained expediently if the patient is stable. A hemodynamically unstable patient should undergo resuscitation while being prepped for an emergency exploratory laparotomy. Indeed, the exploratory laparotomy is the gold standard, with 100% sensitivity and specificity. If the patient is stable, further diagnostic tests should be obtained without delay. Currently, computed tomography (CT) and endoscopy are used as first-line diagnostic modalities when an AEF is in the differential.

CT has been endorsed as the preferred diagnostic tool for the evaluation of AEF. CT is a useful tool for evaluating perigraft infections, and its utility in determining the presence of AEFs is also promising. Its advantages are widespread accessibility and rapidity of image acquisition. Intravenous contrast should be routinely used when assessing for an AEF with both arterial and venous or delayed phases; the utility of oral contrast is in debate. Although oral contrast may be useful in distinguishing bowel wall thickening, it may obscure visualization of any arterial extravasation into the GI tract. The overall sensitivity and specificity of CT for AEF are 94% to 100% and 50% to 85%, respectively. Among the most specific signs for an AEF are extravasation of contrast from the aorta into the bowel lumen and enteric contrast found in the periaortic space ( Fig. 88.2 ). Other signs that may be present on CT scan include periaortic soft tissue edema, periaortic fluid, periaortic stranding, focal bowel wall thickening, pseudoaneurysm formation, and disruption of the aortic wall or aneurysmal wrap. Some of the difficulty in evaluating for AEFs is a result of the overlapping CT findings in AEFs and periaortic graft infections. Periaortic gas, fluid, and soft tissue edema also may be observed with graft infections and are normal findings immediately after the operation. Periaortic gas is abnormal if present 3 to 4 weeks after the operation and may signify a graft infection with or without fistulization or erosion into the GI tract. A perigraft fluid collection that persists beyond 3 months also indicates a possible graft infection and warrants further investigation. The key to determining the relevance of these findings is to correlate the radiologic findings with other clinical signs, such as concurrent GI bleeding.

Endoscopy has also been used as a first-line diagnostic tool in evaluating for AEFs and AEE in the hemodynamically stable patient. A patient exhibiting transient GI bleeding may benefit from an extensive work-up that includes attempts at complete visualization of the GI tract. If there is a high suspicion for an AEF, the endoscopy should be performed in the operating room, in the event of a catastrophic hemorrhage. Particular attention should be paid to the third and fourth portions of the duodenum because AEFs and AEEs are most commonly located in these areas. Endoscopy may reveal a portion of the graft protruding into the bowel, active bleeding, ulcerations, petechiae, blood clot, or extrinsic pulsating mass. Lack of endoscopic findings does not preclude the possibility of an AEF or AEE, and further investigation may be warranted. In addition, alternative bleeding sites can be uncovered; however, care must be taken not to discount the possibility of an AEF or AEE in conjunction with these findings.

Numerous other diagnostic tools have been used with limited success and are often used in the initial work-up of abdominal complaints or as an adjunct to CT or endoscopy. Plain abdominal radiographs may show pneumoperitoneum from the perforated bowel and can be one of the first signs that alert the clinician to the severity of the situation. However, this finding is not common in the presence of AEF, and there are numerous other causes of free air.

Abdominal ultrasounds are not particularly useful for diagnosing AEFs, and it is currently unclear as to the utility of magnetic resonance imaging (MRI) for determining the presence of an AEF. Increased signal intensity in T1- and T2-weighted images indicates localized inflammation, and perigraft fluid seen after the initial postoperative period may signal a graft infection. Although perigraft air also can be detected, it can easily be obscured by motion artifact and can be difficult to distinguish from calcified plaque. In addition, MRI is an expensive imaging technique, not as widely available, and requires more technical proficiency than CT scan.

Indium 111–labeled white blood cell scans or technetium 99m-hexametazime white blood cell scans are useful adjuncts to CT and endoscopy in detecting low-grade graft infections. In patients who do not have overt signs of graft infection, radiolabeled white blood cell scans show promising results, with 100% sensitivity and 94% specificity. Tagged red blood cell scans are also beneficial in localizing fistulas in patients with active GI bleeding. Upper GI studies with barium may demonstrate an AEF with active extravasation of contrast; however, the contrast may be detrimental because it may obscure other diagnostic tests that are more sensitive.

Although rarely used, aortography may show aortic contrast extravasation into the bowel lumen, which is pathognomonic for an AEF. Aortography is not often used in diagnosing AEFs because it is difficult to localize the fistula, and its usefulness lies mostly in preoperative planning with visualization of the anatomy. This modality is occasionally therapeutic because it affords the ability to stem temporarily high-volume GI bleeding by placing a stent over the fistula or embolizing a small artery.

Treatment

The mainstay of treatment for an AEF is surgery, with almost certain death if intervention is delayed or deferred. A clear distinction should be made between the hemodynamically stable patient and one who is actively bleeding and requires an expedient laparotomy. If the patient is hemodynamically unstable, an arterial line and central line should be placed and rapid resuscitation initiated while preparing the patient for an exploratory laparotomy. The patient should be typed and crossmatched and receive empiric intravenous broad-spectrum antibiotics that cover both gram-positive and enteric organisms. In stable patients, a diagnostic work-up may be initiated with a comprehensive operative plan set in place. Because the treatment for an AEF or AEE is a serious undertaking, if the patient is competent and stable, the surgeon should initiate a conversation regarding the potential complications of the operation and the patient's goals of care. A review of the patient's comorbidities will aid in operative planning because there are several different approaches that are available.

Surgical Technique

The goals of AEF repair are rapid and effective control of hemorrhage, preservation of perfusion to the legs, and treatment and containment of infection. Because good visualization of the surgical field is essential, the initial incision should extend from the xiphoid process to the pubic symphysis. A retroperitoneal approach through a left flank incision can be considered in patients who have had previous abdominal operations or currently have hostile abdomens. Proximal and distal control of the aorta is imperative prior to manipulation of the fistula site to prevent uncontrollable hemorrhage. Systemic heparinization is performed prior to aortic cross-clamping to prevent thromboembolic events. Control is obtained with either aortic clamps or occlusion of the aorta with a Fogarty balloon. The aorta is cautiously exposed and proximal control undertaken either at the subdiaphragmatic level or preferably at the infrarenal level to maintain perfusion to the visceral vessels and renal arteries. The aorta is dissected out and distal control obtained either at the aortic or iliac level.

Repair of the Enteric Fistula or Erosion

Because a fistula or erosion may be found anywhere along the alimentary tract, the entire GI tract is inspected to determine its location. Again, careful attention should be paid to the third and fourth portions of the duodenum because the majority of AEFs and AEEs occur in this region. Sharp dissection of the bowel is performed away from the aorta with vigilance sustained to minimize the amount of enteric spillage. Contamination of the field provides a nidus of infection and may lead to disastrous postoperative outcomes, such as aortic stump rupture or recurrence of infection at the reconstruction site. After the bowel has been separated from the aorta, inspect the bowel for areas of necrosis or nonviability. Even though most of the enteric defects are small enough to close primarily with a two-layered transverse closure, a resection may be necessary with anastomosis and proximal diversion as the situation dictates. Meticulous débridement of all infected graft material and tissue within the perigraft area is undertaken. Cultures should routinely be obtained from the aorta, periaortic tissue, and any gross purulence in the field to aid in narrowing antibiotic coverage after the reconstruction.

Prior to closure, a thorough washout decreases the likelihood of infection by decreasing the bacterial load. A nasogastric tube should be placed and left for decompression until bowel function has returned. The retroperitoneal region may require placement of a drain if an abscess was present.

Depending on the location of the AEF and other patient factors, several different approaches to aortic reconstruction have been used and are described next.

Extraanatomic Bypass Grafting With Aortic Ligation and Graft Excision

The traditional and most commonly used surgical approach to AEFs and AEE is the extraanatomic bypass graft. Blaisdell et al. first described the use of an axillobifemoral bypass in 1965 as a treatment for aortic aneurysms. This technique uses a two-team approach to establish an extraanatomic bypass, most often as an axillofemoral bypass, followed by the removal of the infected aortic graft with oversewing of the aorta to create an aortic stump. The aortic stump is closed with two layers of monofilament sutures. Healthy aortic tissue should be incorporated into the stump closure to lessen the risk of aortic rupture. The aortic stump should be protected with omentum, preferably with 360-degree coverage. If omentum is unavailable, a sartorius or rectus femoris muscle transposition may be used.

A staged procedure involves a short delay between the construction of the extraanatomic bypass and resection of the infected graft. This operative plan may be beneficial in those who have chronic aortic graft infections and should be used only if the patient is hemodynamically stable. After the extraanatomic bypass is established, resection of infected graft is performed, preferably within the next few days, depending on the stability of the patient. One study had a median interoperative interval of 5 days, with a range of 2 to 31 days. Mortality rates for a staged extraanatomic bypass range from 11% to 27% and are highly dependent on the patient's preoperative condition, infection load, and hemodynamic status.

One of the most dreaded complications of the extraanatomic bypass in this setting is aortic stump blowout or rupture. It occurs in 5% to 25% of patients and is almost certainly fatal. In addition, pelvic ischemia can occur as a consequence of a lack of perfusion to the internal iliac arteries, and inadequate perfusion to the lower extremities can lead to high amputation rates. The graft itself is subject to both infection and thrombosis, with axillofemoral graft infection rates reported to be between 22% and 40%. The primary patency rate of extraanatomic bypass grafts is approximately 43% at 3 years, with the secondary patency rate improved to 65%. With the staged procedure these numbers have improved, with the 5-year primary patency between 64% and 73%, and secondary graft patency rate of 92% to 100%. Although the extraanatomic bypass grafting technique is the most widely used and has the longest history, it has been criticized for significant mortality and complications. Hence there has been a search for alternative operative methods.

In Situ Graft Replacement

In situ graft replacement is a surgical option that is particularly attractive for patients who have a number of comorbidities, wherein a long operative time may be detrimental. The infected graft is removed and replaced with an in situ graft. Common conduits include femoral veins, aortic homografts, and prosthetic grafts, which have been soaked in antibiotics, typically rifampin. The advantages to these alternatives are decreased risk of limb loss and pelvic ischemia, lower amputation rates, avoidance of stump blowouts, and better long-term patency. The greatest drawbacks to these methods are the possibility of recurrent graft infections, proximal anastomotic rupture, deep vein thrombosis, and chronic limb edema.

Autogenous reconstruction is a technique that has been used with construction of a neoaortoiliac system. The most commonly used conduits are the femoropopliteal veins because of their high patency and large caliber. Superficial veins have fallen out of favor as a consequence of high stenosis rates and intimal hyperplasia that occlude these conduits. The veins are used reversed or nonreversed with lysed valves and anastomosed if a bifurcated graft is necessary for the aortoiliac junction. The decreased infection risk, circumvention of stump blowout, and lower amputation rates as compared with extraanatomic bypass grafting make this an attractive alternative. The technical complexity of harvesting and lack of appropriate conduits pose obstacles in some cases, and long operative times make this approach not ideal for patients with significant comborbidities. However, a large series from Europe routinely using the neoaortoiliac system reconstruction and tensor fascia lata flap coverage of the proximal anastomosis in 55 patients with AEF over 13 years resulted in a 90% 30-day survival with a minimal number of reinterventions or limb thrombosis.

Cryopreserved arterial homografts are an alternative material for in situ arterial reconstruction. They have been used in the treatment of aortic valve endocarditis by cardiothoracic surgeons, with good success. The cryopreserved arterial homograft is maintained at low temperatures prior to implantation, and may be impregnated with neomycin to heighten its bacterial resistance. Thirty-day mortality rates are approximately 9% to 12%, whereas 30-day survival rates range from 67% to 81%. Its major benefit arises from the extremely low infection rates as a result of the antibacterial properties of the allograft. Complications can include aneurysmal degeneration of the homograft and allograft rejection leading to dilation and rupture. Other disadvantages to the cryopreserved arterial homograft include its lack of availability in all centers and relative expense.

The use of antibiotic-soaked prosthetic grafts is also increasing in popularity, partly because of higher flow rates and the higher patency rates seen in these grafts compared with extraanatomic bypass grafts. Grafts may be silver coated, rifampin bonded, or amikacin loaded for greater antibacterial resistance. Rifampin covers a broad spectrum of both gram-positive and gram-negative bacteria, including S. aureus, which is why it remains a popular choice. Antibiotic-coated prostheses have been used with moderate success; however, the concern of infection is higher in patients growing methicillin-resistant S. aureus . Rifampin has been used in the treatment of methicillin-resistant S. aureus but never as a monotherapy because of the rapid emergence of resistance. One approach is to place the graft in a 45- to 60-mg/mL solution of rifampin for 15 minutes at room temperature prior to implantation. The perioperative mortality rate for in situ prosthetic graft replacement ranges from 13% to 21%, with a 30-day mortality rate ranging from 8% to 26%.

Endovascular Repair

Endovascular repair has become increasingly popular as an alternative technique in selected situations. In the presence of severe hemorrhage, it can be used to temporize prior to definitive management. This percutaneous technique is less invasive and may be beneficial to those who have significant comorbidities and cannot tolerate significant shifts in their circulatory status, have a limited life expectancy, or have a hostile abdomen from inflammation or previous operations. Placement of a covered aortic stent graft within the surgical graft excludes the AEF from the circulation, leaving the infected graft in place. Endovascular repair is less likely to be helpful in cases of graft erosion because the bleeding originates from the bowel edges rather than an abnormal aortic fistula formation.

Although most of the literature describes the use of endovascular repair as a temporizing measure, there have been cases in which AEFs were managed solely with endovascular repair in high-risk patients. Significant concerns remain because of the high recurrence rate of bleeding and infection. The infected graft remains in place with this repair, and the fistula is not removed, thus providing a continued source of infection. The inability to débride the area of infected tissue leads to a greater risk of long-term infectious complications. Recurrent or new infections or recurrent hemorrhage have been reported to occur in 44% to 60% of patients with an endovascular approach. Those with known preoperative signs of infection are more likely to have a poor outcome, and patients with secondary AEFs (compared with those with primary AEFs) tend to have a higher rate of infection, presumably from the retained prosthetic graft.

Numerous adjunctive strategies have been used with endovascular repairs to help to minimize risk of progressive infection. Although there is no clear evidence that lifelong antibiotics prevent sepsis, most vascular surgeons will place their patients on lifelong suppressive antibiotic therapy and maintain close follow-up to monitor for signs of recurrent or progressive infection. Patients have also been treated with CT-guided percutaneous drainage of the aneurysm sac with or without placement of drains within the sac for irrigation with saline or antibiotics. A diverting ileostomy may be attempted to prevent further contamination of the graft site. One report of endoscopic injection of fibrin sealant into the fistula tract alleged good success.

Infection of Endografts

In an era when the majority of aortic aneurysms are repaired with endovascular technology, an increasing proportion of AEFs are associated with previously placed endoluminal stent grafts ( Fig. 88.3 ). In a multicenter American review examining more than 200 endovascular aortic graft infections, more than a quarter presented as enteric fistulas. This included aortoesophageal fistulas occurring in a high proportion of infected thoracic endografts ( Fig. 88.4A and B ). In this population, aortic fistula in univariate and multivariate analysis was significantly associated with all-cause mortality. In a multicenter Italian study, the incidence of AEF after EVAR in approximately 4000 patients over 16 years was 0.8%. Pseudoaneurysm, as the initial indication for EVAR, and urgent or emergent timing of the initial procedure were significantly associated with later development of AEF.

Management of these patients is parallel to those with prior open aortic reconstruction, with the majority undergoing in-line aortic reconstruction, primarily with antibiotic-soaked prosthetic grafts, although a significant proportion do undergo reconstruction with cryopreserved allograft or autogenous tissue and creation of a neoaortoiliac system.

Infected fenestrated and branched grafts have been reported, and as more centers place these grafts for complex aortic aneurysmal disease, enteric fistulization will mandate creative management solutions.