Open resection of gastroduodenal artery aneurysm with fistulization into the duodenum 17 years after coil embolization in a patient with Caroli disease

Abstract

Gastroduodenal artery aneurysm (GDAA) is a very rare vascular condition with a high risk of fatal rupture. Coil embolization is the recommended treatment for GDAA. This is a unique case of a 57-year-old woman with autosomal recessive polycystic kidney disease requiring renal transplantation, and Caroli disease who presented with hemorrhage related to enteric fistulization of a GDAA 17 years after coil embolization. Unsuccessful endoscopic and endovascular management necessitated open surgical resection and omental patching. The patient had an uneventful recovery and is still alive without recurrence 3 years later.

Visceral artery aneurysms (VAAs) are rare, with an incidence estimated at 0.01% to 0.2% in the general population, and mostly discovered incidentally.1, 2, 3 Gastroduodenal artery aneurysms (GDAAs) comprise only 1.5% of all VAAs and thus are extremely uncommon.⁴ Coil embolization is the treatment of choice for all GDAAs, regardless of size, due to the high risk of rupture and fatal bleeding.⁵ However, the radiopaque embolizing agents create significant radioartifact on follow-up imaging, which poses challenges for surveillance and accurately determining reperfusion of the aneurysmal sac.⁵ The patient provided consent for publication.

Case report

A 57-year-old female presented with a recurrent episode of hematemesis, hematochezia, and signs of hemorrhagic shock. She had a complex past medical history of autosomal recessive polycystic kidney disease (ARPKD), polycystic hepatic disease (PKHD1 genetic variant), portal hypertension, and Caroli disease. Caroli disease is a rare congenital disorder that causes cystic dilation of the intrahepatic bile ducts and leads to liver cirrhosis and portal hypertension. The patient’s past medical history is also notable for end-stage renal disease, requiring kidney transplantation twice at 24 and 10 years prior to presentation. She also had annual surveillance for multiple small VAAs involving the celiac trunk (2 cm), splenic artery (2 cm), and underwent coil embolization of a 4.8-cm GDAA with Guglielmi Detachable Coils (Boston Scientific) 17 years prior to presentation.

The patient experienced the first episode of hematemesis and rectal bleeding 6 months prior to the current presentation, during which an esophagogastroduodenoscopy (EGD) failed to identify the source of bleeding. She was managed conservatively at that time, given her history of portal hypertension.

After stabilizing the patient, a computed tomography angiography (CTA) of the abdomen and pelvis showed a poorly visualized, previously coiled GDAA that had enlarged to 5.5 cm (Fig 1). An urgent EGD revealed a significant amount of blood and clots in the stomach, as well as a 5.5-cm subepithelial mass in the second portion of the duodenum with ulceration and an actively oozing vessel (Fig 2). Hemostasis with endoscopic clipping and banding was unsuccessful.

Fig 1.

Preoperative computed tomography angiography (CTA) of the abdomen and pelvis. Yellow arrows point to a 5.5-cm previously coiled gastroduodenal artery aneurysm (GDAA) with poor visualization due to radiopaque artifact.

Fig 2.

Esophagogastroduodenoscopy (EGD) of the second part of the duodenum. (A) Subepithelial mass; (B) Ulcer with actively oozing vessel.

A mesenteric angiography revealed persistent high flow in the previously embolized GDAA, with inflow from a branch of the superior mesenteric artery (SMA) and outflow through a branch of the GDA (Fig 3). Repeat embolization through the SMA inflow branch using Penumbra Packing and Interlock Fibered coils achieved only partial occlusion, as the aneurysm’s large size and turbulent flow caused preferential coil deposition on one side of the sac. Outflow coiling through the celiac trunk could not be achieved, so the procedure was stopped.

Fig 3.

Preoperative mesenteric angiography. The yellow arrow indicates a high-flow previously coiled gastroduodenal artery aneurysm (GDAA). Retrograde inflow is noted from a branch of the superior mesenteric artery (SMA), likely the inferior pancreaticoduodenal artery or other gastroduodenal artery collaterals (white arrow). Outflow occurs through a branch of the gastroduodenal artery (blue arrow). The red arrow marks the previously coiled inflow, which shows no contrast flow.

Given the patient’s history of multiple VAAs, complete aneurysmal degeneration of the GDA, and the imminent risk of bleeding, the decision was made to proceed with open resection of the GDAA and repair of the arterio-duodenal fistulization. A subcostal Chevron incision was made, and the upper abdomen was explored. The liver appeared granular and hard as a result of polycystic disease, and the aneurysm was palpated posterior to the hepatic flexure. The bowel was mobilized using a Cattell-Braasch maneuver. The aneurysm was heavily calcified, inflamed, and adherent to the duodenum (Fig 4, A). Thereafter, the feeding arteries to the aneurysm were identified and controlled with vessel loops. The aneurysm was dissected from the duodenum except at the site of the arterio-enteric fistulization (AEF), where it was completely adherent. Then, the aneurysm was opened circumferentially, and several coils from prior embolization were removed (Fig 4, B and C). The aneurysm was then completely resected, except for a small portion of the aneurysm wall at the site of fistulization. To avoid resection of the wall of the duodenum, which could cause a stenosis or a stricture, the area was closed primarily, incorporating part of the aneurysm sac that fistulized to buttress the repair. An additional omental patch was used also for coverage.

Fig 4.

Intraoperative images of the abdomen. (A)Black arrow points to the closed 5.5-cm gastroduodenal artery aneurysm (GDAA) sac. (B)Blue arrow points to the opened aneurysmal sac with several coils. White arrow points to the duodenum. (C)Inferior blue vessel loop marks the feeding artery.

The postoperative period was uneventful. Upon clinical improvement on day 6, an upper gastrointestinal series showed no extravasation of contrast. The nasogastric tube was removed, and the patient was started on a clear liquid diet, which was advanced to a low-residue diet on day 8. The patient was discharged after 10 days of hospitalization.

The first follow-up imaging study was at 6 months, followed by annual surveillance. At 3-year follow-up, CTA showed no evidence of GDAA recurrence (Fig 5). The patient remains alive and functional and walks 3 miles daily, with a working renal transplant.

Fig 5.

Three-year follow-up computed tomography angiography (CTA) demonstrated no evidence of gastroduodenal artery aneurysm (GDAA) recurrence.

Discussion

This report presents a unique case of a patient with a complex history of ARPKD, Caroli disease, and renal transplant presenting with hemorrhage from fistulization of a large GDAA into the duodenum 17 years after coil embolization. Open resection was life-saving and definitive, durable treatment.

Both ARPKD and Caroli disease result from mutations in the same gene (PKHD1), and very rarely present with VAAs compared with the autosomal dominant variant. ElChediak et al reported on two relatively young patients (a 36-year-old male and a 17-year-old female) with ARPKD, Caroli disease, and VAAs.⁶ One of the patients failed coil embolization of the gastric artery, necessitating an open laparotomy and oversewing the vessel, but eventually succumbed to the disease.⁶ Mandal et al reported a 20-year-old male with Caroli disease and portal hypertension who underwent open splenectomy and excision of a splenic artery aneurysm.⁷

It remains uncertain whether the development of VAAs (including GDAAs) is directly caused by the gene mutation or is a consequence of chronic kidney disease and portal hypertension in patients with ARPKD and/or Caroli disease. However, the occurrence of VAAs at such a young age is likely associated with genetic predisposition and compounded by hemodynamic factors.

Although guidelines recommend coil embolization for all GDAAs,⁵ coiling a ruptured GDAA remains challenging—both technically, given the high flow inside the aneurysm, and clinically, given the ongoing life-threatening bleeding and risk of hemorrhagic shock. A meta-analysis of 224 cases of GDAA, showed that cases with complicated GDAA had lower rates of coil embolization (62.8% vs 77.9%; P < .001), compared with patients with uncomplicated GDAAs.⁸

Open resection has been reported as a definitive and often lifesaving treatment for ruptured GDAAs. Failure of coil embolization has also been documented. Harris et al reported a case of ruptured GDAA that failed coil embolization, necessitating open laparotomy and ligation of the aneurysm,⁹ whereas Savage et al described a spontaneous rupture of GDAA, which was successfully managed with open resection.¹⁰

The presence of an AEF arising from a previously coiled GDAA increased the complexity and rarity of our case. Almadwahi et al reported a spontaneous GDAA rupture with arterio-duodenal fistulization managed primarily with open resection of the aneurysm, with repair of the duodenum.¹¹ Even though the open repair was similar to the repair performed in our case report, the GDAA was only 3.5 cm in size and had initial presentation of rupture and fistulization without prior endovascular therapy.

This case highlights the importance of considering long-term outcomes following coil embolization for GDAA, especially in patients with complex medical histories. The requirement for open surgical repair nearly 20 years after the initial procedure, for the development of an AEF, highlights the risk of late failures of endovascular therapy. These findings emphasize the need for vigilant postembolization surveillance and careful evaluation to determine the appropriate timing for reintervention or conversion to open repair.

Conclusions

VAAs, including GDAAs, are rarely associated with ARPKD and Caroli disease. AEF is a fatal complication that warrants early diagnosis and management. Although endovascular techniques have improved and are widely used, their long-term durability in young patients remains poorly defined. Open surgical repair is the definitive and durable treatment of VAAs and can be performed safely even in emergent settings.

Funding

None.

Disclosures

C.I.O.C. reports consultant for the Society for Vascular Surgery-Patient Safety Organization and EnVVeno Medical; has intellectual property of patent U.S.S.N. 10,524,89; and has received research support from the Yale Department of Surgery, Society for Vascular Surgery, American Venous Forum, CT Innovation, Vascular Study Group of New England, National Institutes of Health, Boston Scientific, Medtronic, EnVVeno Medical, and Inari Medical. The remaining authors report no conflicts.

From the Society for Clinical Vascular Surgery