Pseudoaneurysm of the gastroduodenal artery secondary to diverticulitis of the first part of duodenum: a rare presentation of upper gastrointestinal bleed

Abstract

Acute upper gastrointestinal (UGI) bleeding is one of the most frequent presentations to a surgical emergency. Most of them respond to initial resuscitation, and a definite diagnosis is established as soon as possible, thereby helping the clinician in management. We present the diagnostic challenges that we faced with a 70-year-old man who presented with UGI bleed. He initially responded to resuscitation, but later deteriorated and became haemodynamically unstable. The source of the UGI bleed on evaluation was found to be pseudoaneurysm of the gastroduodenal artery (PsGDA) and treated successfully by coil embolisation. The cause of the PsGDA was diverticulum arising from the first part of duodenum with changes of diverticulitis. Diverticulum originating from the first part of the duodenum is seldom reported. Moreover, diverticulitis involving this part and causing PsGDA has not been reported so far.

Background

The aetiology of an upper gastrointestinal (UGI) bleed is often established promptly, which helps clinician decide on the management. Rupture of the splanchnic pseudoaneurysm leading to UGI bleed is rare. Majority of the aneurysms are secondary to pancreatitis and are false aneurysms. Any visceral artery can get affected and develop pseudoaneurysm. Splenic, renal and hepatic arteries are involved most of the time. Pseudoaneurysm of the gastroduodenal artery (PsGDA) is rare, and if it occurs, then the underlying aetiology is usually pancreatitis. Duodenal diverticulitis induced PsGDA is an infrequent cause of UGI bleed, and not yet reported in the literature. CT of the abdomen is diagnostic. Surgery is the traditional treatment and is replaced by endovascular methods as the first line of treatment. We report a case of duodenal diverticulitis with PsGDA presented with acute UGI bleeding managed with angiographic coil embolisation of the GDA.

Case presentation

A 70-year-old man with no known comorbidities had presented to the emergency with a history of haematemesis. It was a single bout of around 800 mL vomitus comprising fresh blood. He had no other complaints. He did not have any similar history in the past. He was a chronic alcoholic and a smoker. His last binge of alcohol intake was around 2 months before the day of presentation. He did not have any significant family history.

On general examination, he had severe pallor. Abdominal examination revealed mild guarding in the upper abdomen. Per rectal examination revealed melena. The rest of his systemic examination was normal.

On arrival, he was resuscitated initially with fluids, and later with 1 U of packed cells. His systolic BP initially picked up to 100 mmHg. However, it did not maintain, and he was started on inotropes.

A nasogastric (NG) tube was inserted, and around 1 L of dark altered blood drained.

Investigations

Chest and abdomen X-rays done on arrival were normal.

Ultrasonography abdomen showed features of bilateral medical renal disease without evidence of decompensated liver disease.

His blood investigations at the time of admission revealed haemoglobin of 5.3 g/dL, urea of 197 mg/dL, creatinine of 2.72 mg/dL, prothrombin time of 17.3 s and an international normalised ratio of 1.33.

Upper gastrointestinal endoscopy (UGIE) is the standard of care for patients with UGI bleeding, even though it is severe. The patient was on inotropic support, and his general condition was not good enough to warrant a UGIE. Also, the patient was not co-operative. The patient’s co-operation under such circumstances would not only have been dismal but would have been under very high risk. Hence UGIE was not attempted.

Because of ongoing bleed in the form of persistent melena, dropping haemoglobin and worsening haemodynamic parameters, and nil NG output, small bowel pathology was suspected, and a triple-phase contrast-enhanced CT (CECT) abdomen was done. It showed a focal outpouching from the first part of the duodenum (D1) and 0.6×0.5 cm PsGDA adjacent to the diverticulum with extravasation of blood into the diverticulum (figures 1 and 2). Inflammatory changes were seen in the diverticular wall, with the possibility of an impending rupture.

Figure 1.

Arterial phase of the contrast-enhanced CT showing the gastroduodenal artery pseudoaneurysm (yellow arrow mark) abutting the diverticulum present in the first part of the duodenum (red arrow mark).

Figure 2.

Coronal section of the arterial phase of the contrast-enhanced CT showing gastroduodenal artery pseudoaneurysm supplying the first part of the D1 diverticulitis.

Digital subtraction angiography (DSA) was done, confirming the diagnosis of an aneurysm arising from the GDA.

Treatment

On confirming the presence of PsGDA by DSA, the coil was deployed at the aneurysm site (figures 3 and 4). Post-procedure angiography showed no distal flow in the GDA (video 1). Diverticulitis was addressed with intravenous ceftriaxone.

Figure 3.

Digital subtraction angiography showing pseudoaneurysm of gastroduodenal artery.

Figure 4.

Angiographic coil deployed at the aneurysm site.

Video 1.

Download video file ^{(7MB, mp4)}

DOI: 10.1136/bcr-2020-238232.video01

Outcome and Follow-up

The patient responded immediately with improving haemodynamic parameters, which finally became normal. His haemoglobin improved as well. As his general condition improved, he was discharged 1 week post-procedure. The patient was asymptomatic at the end of 6 months’ follow-up.

Discussion

Aneurysms of the splanchnic vessels are uncommon, with a reported incidence of 0.01% to 0.2% in autopsies.¹ They occur more in middle-aged men in their 50s, with a male:female ratio of 4.5:1.² The most commonly found aneurysms are in the splenic artery (46%), renal artery (22%) and hepatic artery (16.2%).³ Aneurysm of the GDA constitutes an approximate of only 1.5%.⁴ The majority of them are secondary to pancreatitis (47%) and are pseudoaneurysms, which involve only one or more layers of the vessel wall.⁵ The peripancreatic inflammation can lead to the release of pancreatic proteolytic enzymes, which implicates in weakening the vessel wall.⁶

Duodenal diverticula are outpouchings of the duodenal wall and can be found as much as 2% to 5% in patients undergoing UGIE, and 3% to 22% in autopsies.⁷ However, their exact incidence is not known with certainty. Their incidence increases with age and is comparable almost equally among males and females. Their most common site of occurrence is in the second part of the duodenum (67%) in the juxta-papillary area.⁸ Ninety per cent of them are asymptomatic. A duodenal diverticulum can complicate by causing diverticulitis, obstruction, bleeding and very rarely perforation. One case of diverticulitis present in the second part of the duodenum, causing pseudoaneurysm of the pancreaticoduodenal artery has been reported.⁹ Bleeding from this aneurysm can be missed if not promptly looked for, as in our case. Barium meal follow-through (BMFT) has the highest sensitivity in diagnosing a duodenal diverticulum. CECT, although not so sensitive as BMFT in diagnosing diverticulum, can establish the presence of inflammation in diverticulitis and rule out other pathologies in nearby structures.

Only 7.5% of the GDA aneurysms are asymptomatic.¹⁰ When symptomatic, the most common symptom is UGI bleed (52%) from the rupture of the GDA. Mortality from this bleed can be as high as 40% and depends on the severity of the bleed. The prognosis is notably worse when there is a bleed into the duodenum with a reported mortality of 21%. Abdominal pain is the second most common reported symptom (46%). It can also present with gastric outlet obstruction and with vague symptoms like nausea and vomiting when there is intraperitoneal or retroperitoneal bleeding. Presentation with hemosuccus pancreaticus and a pulsatile abdominal mass is rare but reported.¹⁰

The gold standard modality for diagnosing a pseudoaneurysm is a visceral angiography, with a sensitivity of 100%.¹¹ It could also serve as a therapeutic platform since interventions would be possible. Magnetic resonance angiography reported having equal sensitivity.¹² Triple-phase CT has a sensitivity of 67% in diagnosing a PsGDA. However, it has the added advantage of diagnosing the antecedent cause of the aneurysm, for example, pancreatitis and diverticulitis, as in the case presented.¹³

The treatment of PsGDA needs to be tailored to the situation in which the patient presents. It can be treated either with endovascular techniques or an open approach. Endovascular techniques have the advantages of being less invasive, thereby avoiding usual postoperative complications.¹⁴ Open procedures include ligation and excision of the aneurysm and bypass procedure. Attempts and re-attempts at embolisation leading to failure can be addressed by open surgery. Open surgical approaches, despite having its disadvantages, have better outcomes, with a reported 88.2% of patients being healthy at 5 months of follow-up.¹⁵

The inciting factor, diverticulitis, should be addressed as well. While the surgical intervention advocated earlier, intravenous antibiotics and bowel rest seem to address asymptomatic diverticulum and uncomplicated diverticulitis.¹⁶ When there is complicated diverticulitis and not responding to conservative management, surgical treatment may be warranted. Surgical management in the form of excision of the diverticulum and primary closure of the site can be done.¹⁷ In case of complication or presence of significant duodenal and retroperitoneal inflammation, diversion procedure may be needed in the form of duodenojejunostomy, gastrojejunostomy, tube duodenostomy, pyloric exclusion or sometimes even pylorus-preserving pancreaticoduodenectomy. In the present case, acute UGI bleed was treated by emergency angiography and coil embolisation of the PsGDA. Diverticulitis was treated conservatively without further complications.

Patient’s perspective.

My family members and I got scared and little shocked when I vomited out blood. We immediately rushed to the hospital. Doctors at the hospital did some tests. I thought I would get better, but I didn’t get any better in the first 2 to 3 days of admission. They told me they were trying to find out where I was bleeding and that it was not clear. I was apprehensive. However, one late night all of a sudden they got my sign in consent form saying they’ve found a potential source of the bleed and that I might be needing emergency coiling. I gave consent and underwent the procedure and immediately felt better. I was observed for a week and then discharged. My relatives supported me all the while, especially by donating blood since I had to undergo a lot of transfusions, especially before I underwent coiling. This was to match the blood I was losing and to keep me alive, the doctors said. At six months now, I feel normal with no similar complaints again. I’m grateful the doctors could finally find out the source of bleed that night which indeed saved my life.

Learning points.

• Pseudoaneurysm of the gastroduodenal artery (PsGDA) is a rare entity, with duodenal diverticulitis as its aetiology not being reported before.

• Rupture of the pseudoaneurysm leading to upper gastrointestinal bleed is its most frequent complication.

• Although contrast-enhanced computed tomography has a lower sensitivity for the diagnosis of the aneurysm, it has added clinical advantage of diagnosing the underlying cause of the aneurysm.

• Angiography plays a significant role in the confirmation of the PsGDA due to its highest sensitivity and has added therapeutic advantage.

• Angiographic coil embolisation of the aneurysm should be considered as the first line of treatment, and surgery should be reserved for failed cases.