|Home | About | Journals | Submit | Contact Us | Français|
Primary aortoenteric fistula is a rare clinical situation with a high mortality rate. One should suspect that condition when an abdominal aortic aneurysm is known to be present. We describe the case of a 60 year old man who presented with upper gastrointestinal bleeding as the first and sole manifestation of an abdominal aortic aneurysm, due to the rupture of the aneurysm in the jejunum.
The patient was admitted with hematemesis and melena. He reported no abdominal pain. Upper gastrointestinal endoscopy disclosed no bleeding or lesions of the stomach and duodenum. Bleeding stopped the following day, only to recur 4 days later. The patient was then subjected to abdominal CT scan, which revealed the presence of a subrenal aortic aneurysm, with fistulization to the small intestine. At laparotomy the aortic aneurysm was adherent to the first centimeters of jejunum. The diseased aorta was excised and replaced with a Dacron Y graft in situ.
Primary aortojejunal fistulas are only rarely encountered. They usually are the result of a nonspecific aneurysm of the abdominal aorta. They usually manifest with premonitory bleeding, followed by catastrophic hemorrhage few days later. Even with surgery the mortality rate is high.
Diagnosis of primary aortoenteric fistula requires a high index of suspicion in cases of upper gastrointestinal bleeding, especially when endoscopy is negative and there is no knowledge of the existence of an abdominal aortic aneurysm. CT scan and prompt surgery are crucial to the survival of these patients.
Primary aortoenteric fistulas are a rare cause of massive bleeding from the upper gastrointestinal tract, with about 350 cases reported in the literature until 2006.1–3 Most of them (about 70%) are the result of fistulization between the abdominal aorta and the duodenum, due to a non-specific abdominal aortic aneurysm of atherosclerotic etiology.1 Less frequently they may be formed between the abdominal aorta and other portions of the alimentary tract (jejunum, colon or even esophagus). Less than half of the reported cases have been treated by emergency laparotomy and the operative mortality approaches 50%.4 Without treatment mortality is 100%. This shows that early diagnosis is frequently missed, with catastrophic results for the patients. A large autopsy series reported an incidence of 0.04–0.07% of primary aortoenteric fistulas, occurring in 0.69–2.36% of patients with an aneurysm of the abdominal aorta.5 Secondary aortoenteric fistulas, that is after surgical repair of an abdominal aortic aneurysm, have been reported to occur in 0.4–4.0% of patients.6 We describe the case of a 60 year old man who presented with upper gastrointestinal bleeding as the first and sole manifestation of an abdominal aortic aneurysm, due to the rupture of the aneurysm in the jejunum, just a few centimeters from the ligament of Treitz.
The patient, a 60 year old man, was admitted to the hospital, after an episode of upper gastrointestinal bleeding and melena. He reported periumbilical abdominal pain. His past history was unremarkable, except hypertension controlled with medication. He reported no known abdominal aneurysm. He reported no haematemesis or melena or abdominal pain in the past and he had no abdominal surgery. His blood pressure was 120/80, his heart rate was 110 bpm and his hemoglobulin level was 10.1 mg/dl. He was transfused with 2 units of blood and scheduled for upper gastrointestinal endoscopy.
Emergency esophagogastroduodenoscopy was performed. It revealed the presence of large clots in the stomach, but it did not disclose any active bleeding from the stomach or duodenum or any other pathology. The stomach and duodenum were irrigated for cleansing the blood and thrombi.
The patient continued having hematemesis and melena for another 2 days, at which time bleeding ceased, and the patient's condition seemed to improve. 4 days after admission bleeding from the upper gastrointestinal tract recurred. At that time the patient was subjected to abdominal CT scan with iv administration of contrast, which revealed the presence of an infrarenal aortic aneurysm, with fistulization to the small intestine (Fig. 1). The patient collapsed during performance of CT scan with loss of consciousness. He was intubated and transferred immediately to the operating room. Blood gas hemoglobulin was 7.2 mg/dl at that time.
At the operating table the patient was set at the supine position. An initial left anterior thoracotomy incision was chosen, because the patient was extremely hemodynamicaly unsteady, with systolic blood pressure 50 mm Hg and slowing heart rate with obvious change of the ECG to asystoly. The pleura around the descending thoracic aorta, close to the diaphragm, was incised and the descending aorta was clamped by the means of a long vascular Satinsky clamp. Thereafter the pericardium was opened anterior to the phrenic nerve and massage of the failing heart was performed. The patient's hemodynamic state improved dramatically. At that point we proceeded to the midline laparotomy incision. No blood was evident in the peritoneal cavity. An infrarenal abdominal aortic aneurysm was present, adherent to the 4th portion of the duodenum and the first few centimeters of jejunum. The neck of the aneurysm just below the renal arteries, as well as the common iliac arteries were dissected and controlled with vessel loops. The clamp at the descending aorta was transferred at the abdominal aorta just below the renal arteries and the common iliac arteries were clamped as well.
At that point the aneurysm was incised longitudinally with a #11 blade and consequently with Metzelbaum scissors throughout its length. After evacuation of blood and clots, a communication between the aorta and the jejunum was evident at the right lateral wall of the aorta. As much as of the diseased wall of aorta was excised and the aorta was replaced by a Dacron Y graft (end to end anastomosis centrally with running 4-0 prolene stitch and bilateral end to end anastomoses to the iliac arteries with running 5-0 prolene stitch). Thereafter the hole in the jejunum was closed in two layers with absorbable material, after meticulous debridement of the its edges. After final hemostasis, the peritoneum was closed over the vascular graft. A chest tube was inserted in the left hemithorax and was laid posteriorly. Thereafter the incisions were closed as per routine. The patient was transferred to the ICU supported with large doses of inotropes. Unfortunately, the patient succumbed several hours later at the ICU, due to multi-organ failure.
Primary aortoenteric fistula is a rare clinical situation. Most of the times the cause is a, frequently unknown, atherosclerotic abdominal aortic aneurysm.7 In as much as of 50% of cases, the diagnosis is not evident until the time of emergency surgery for uncontrollable hemorrhage.4 The retroperitoneal third portion of the duodenum is the bowel segment most frequently affected, due to its proximity with the abdominal aorta. The mechanism of fistula formation between the aorta and the intestine most probably involves ischemia and subsequent necrosis of the intestinal wall, as a result of bowel wall injury due to the pulsating movements of the adjacent aortic aneurysm. Subsequently, rupture of the expanding aneurysm results in the formation of a fistulous communication of the aorta with the bowel and rapid exsanguination, if untreated.8 Other much rarer causes of aortoenteric fistulas may be of infectious etiology (bacterial, syphilitic, and tuberculous forms of aortitis), as well as cancer, radiation injury, peptic ulcer, gallstones, diverticula, and foreign bodies.8
The classic triad of abdominal pain, gastrointestinal hemorrhage and a pulsatile abdominal mass can be found in 10–23% of the patients.7,9 The most important thing to remember is that the clinical picture of primary aortoenteric fistulas is characterized by a “herald” or “premonitory” hemorrhage, that is followed a few days later by catastrophic hemorrhage, as was the case in the patient we describe. The herald bleeding is the result of a small fistula that is temporarily tamponaded by thrombus. The natural history of this condition is one of expansion of the fistula and dislodgement of the thrombus due to the high blood pressure within the aorta, with resultant catastrophic exsanguination.8 Survival without treatment is 70% at 6 h after the initial bleeding episode and 50% at 24 h, so intensive diagnostic workup and prompt treatment are indicated.10
Upper gastrointestinal endoscopy should be the first diagnostic examination performed (overall sensitivity may be up to 50%), followed by CT scan with iv contrast.11 Detection of contrast within the intestine is the definite pathognomonic sign, but other more subtle abnormalities detected on CT have been described as well.1,11
Mortality is 100% without surgical intervention. Emergency surgery should be undertaken, as soon the diagnosis is evident or there is strong suspicion. Mortality rates after surgery range,1–11 but the postoperative mortality rate may be high. At surgery we recommend for culture and debridement of the aneurysmal aorta, repair with an in situ prosthetic graft, and a primary repair of the gastrointestinal tract in situations where there is no gross contamination of the peritoneal cavity. Postoperatively, the patient should be treated with intravenous broad spectrum antibiotics for up to 6 weeks according to the culture results.4 In grossly contaminated cases, one should consider the use of extraanatomic vascular bypass.1 Endovascular treatment has been described as well.12
Diagnosis of primary aortoenteric fistula requires a high index of suspicion in cases of upper gastrointestinal bleeding, especially when endoscopy is negative and there is no knowledge of existence of an abdominal aortic aneurysm. We suggest that abdominal CT scan should be readily performed in such cases, because failure to establish the diagnosis results in high mortality rate. Emergent surgical treatment is the best chance of survival of these sick patients.
No conflicts of interest.
No sources of funding.
Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request.
G. Philippakis and M. Moustardas have contributed equally at the study design, data collection, data analysis. G. Philippakis wrote the case report. He was also the operating surgeon.
The case was operated when the author was a consultant cardiothoracic and vascular surgeon at the Department of Thoracic and Vascular surgery, General Hospital of Chania, Chania, Greece.