Selective embolization for bleeding visceral artery pseudoaneurysms in patients with pancreatitis

Harsheet Sethi, Praveen Peddu, Andreas Prachalias, Pauline Kane, John Karani, Mohamed Rela and Nigel Heaton

London, UK

Original Article / Pancreas

Selective embolization for bleeding visceral artery pseudoaneurysms in patients with pancreatitis

Harsheet Sethi, Praveen Peddu, Andreas Prachalias, Pauline Kane, John Karani, Mohamed Rela and Nigel Heaton

London, UK

BACKGROUND:Pancreatitis is associated with arterial complications in 4%-10% of patients, with untreated mortality approaching 90%. Timely intervention at a specialist center can reduce the mortality to 15%. We present a single institution experience of selective embolization as first line management of bleeding pseudoaneurysms in pancreatitis.

METHODS:Sixteen patients with pancreatitis and visceral artery pseudoaneurysms were identified from searches of the records of interventional angiography from January 2000 to June 2007. True visceral artery aneurysms and pseudoaneurysms arising as a result of post-operative pancreatic or biliary leak were excluded from the study.

RESULTS:In 50% of the patients, bleeding complicated the initial presentation of pancreatitis. Alcohol was the offending agent in 10 patients, gallstones in 3, trauma, drug-induced and idiopathic pancreatitis in one each. All 16 patients had a contrast CT scan and 15 underwent coeliac axis angiography. The pseudoaneurysms ranging from 0.9 to 9.0 cm affected the splenic artery in 7 patients: hepatic in 3, gastroduodenal and right gastric in 2 each, and left gastric and pancreaticoduodenal in 1 each. One patient developed spontaneous thrombosis of the pseudoaneurysm. Fourteen patients had effective coil embolization of the pseudoaneurysm. One patient needed surgical exclusion of the pseudoaneurysm following difficulty in accessing the coeliac axis radiologically. There were no episodes of re-bleeding and no in-hospital mortality.

CONCLUSIONS:Pseudoaneurysms are unrelated to the severity of pancreatitis and major hemorrhage can occur irrespective of their size. Co-existent portal hypertension and sepsis increasethe risk of surgery. Angiography and selective coil embolization is a safe and effective way to arrest the hemorrhage.

(Hepatobiliary Pancreat Dis Int 2010; 9: 634-638)

pseudoaneurysm; embolization; acute pancreatitis; chronic pancreatitis; vascular complications

Introduction

Major hemorrhage is a rare but lethal complication of pancreatitis, usually as a result of the development of visceral artery pseudoaneurysms. These pseudoaneurysms result from infective erosion of an adjacent artery. A fibrous capsule progressively enlarges owing to arterial pressure. Rupture may present with fatal bleeding either into a pseudocyst, the peritoneal cavity, and the retroperitoneum, or erosion into the stomach, pancreatic duct (hemosuccus pancreatitis) or bile ducts (hemobilia).

Arterial complications occur in 4%-10% of patients with pancreatitis.[1-5]Survival of the patients depends on an early diagnosis as untreated mortality reaches 90%, in contrast to 15%-50% when treated effectively.[5-9]Prompt diagnosis and the sequence of diagnostic investigations remain a challenge and the available evidence reflects single-institution experiences.

We present a retrospective study of 16 patients admitted to our institution over 7 years with visceral artery pseudoaneurysms related to acute and chronic pancreatitis.

Methods

Study population

Sixteen patients undergoing selective mesenteric angiography for suspected vascular complications of pancreatitisduring the period of January 2000 to June 2007 at our institution were identified from searches of records of interventional angiography and cross checked with clinical databases and patient records. True visceral artery aneurysms and pseudoaneurysms arising as a result of post-operative pancreatic or biliary leak were excluded from the study.

Characteristics of the acute pancreatitis group

Of the 16 patients, 8 (5 males and 3 females) developed vascular complications during the course of an attack of severe acute pancreatitis diagnosed on CT scan. Their mean age at presentation was 55 years (range 21-71 years). Of these patients, 3 had acute necrotising pancreatitis, and 1 underwent necrosectomy prior to the detection of pseudoaneurysm. One patient developed a pseudoaneurysm 6 months post-pancreatitis and underwent cystogastrostomy and peritoneo-venous shunt for pancreatic ascites. The etiology included alcohol in 2 patients, gallstones in 3, trauma in 1, drug-induced (paracetamol overdose) in 1 and idiopathic pancreatitis in 1.

Characteristics of the chronic pancreatitis group

Of the 16 patients, 8 (6 males and 2 females) were known to have alcohol-induced chronic pancreatitis, 7 presented with pseudocysts, and 1 with acute necrotising pancreatitis. The mean age of the patients at presentation was 48 years (range 24-66 years). The diagnosis of chronic pancreatitis was established by the presence of calcification or characteristic ductal lesions on abdominal CT.

Clinical presentation

The most common presenting symptoms were abdominal pain (acute 7; chronic 7), vomiting (acute 5; chronic 8), shock (acute 5; chronic 5), upper gastrointestinal bleeding (acute 4; chronic 4), jaundice (acute 2; chronic 2), and portal hypertension (acute 1; chronic 3).

Two patients developed abdominal distension and raised intra-abdominal pressure owing to intraperitoneal rupture of the pseudoaneurysm. All patients were anemic at presentation with a median hemoglobin level of 7.9 g/dl (range 2.6-11.5). Blood transfusion ranged from 0 to 23 units, with a median of 4 units. Three patients(acute) had adult respiratory distress syndrome and renal failure and were ventilated prior to transfer.

Management protocol

All patients transferred to our institution with septic shock were admitted to the ICU.

All patients were subjected to a contrast CT scan of the abdomen for identifying the cause of bleeding (Fig. 1).

Patients suspected of bleeding pseudoaneurysm on CT underwent a mesenteric angiography with a view to selective embolization. The superior mesenteric artery, coeliac axis and portal vein were imaged on angiography performed through a femoral puncture by experienced interventional radiologists (Fig. 2).

A microcatheter was used to selectively catheterize the pseudoaneurysm, and super-selective embolization was performed with stainless steel coils (0.18) alone or a combination of coils and polyvinyl alcohol (PVA-contour, 150-355 microns). Both the sac and the feeding vessel were occluded with coils. Embolization was regarded as successful when cessation of blood flow in the target vessel was demonstrated radiologically (Fig. 3).

CT scans were performed at 24 hours and after 1 week to look for effective occlusion of the pseudoaneurysm, extravasation of contrast and evidence of ischemic changes post-embolization. Persistent filling of the pseudoaneurysm sac was managed by repeat embolization (Fig. 4).

Fig. 1. CT demonstrating necrotising interstitial pancreatitis with a pseudonaneurysm (short arrow) arising from the left gastric artery (long arrow).

Fig. 2. Coeliac axis angiography demonstrating the left gastric artery pseudoaneurysm (arrow).

Fig. 3. Coil occlusion of the left gastric artery.

Fig. 4. CT post embolisation demonstrates coils within the left gastric artery (arrow). No filling or enhancement of the aneurysm is evident.

Surgical exclusion was resorted to when embolization was not possible.

Patients were followed up for at least 6 months after treatment. Follow-up data were obtained by direct contact with the patients, from clinic records, and from information provided by physicians.

Results

The initial diagnosis of a pseudoaneurysm was made on ultrasound scan in 3 patients and CT scan in 12, and it was confirmed by angiography. The size of the pseudoaneurysm ranged from 0.9 to 9.0 cm. The interval ranging from 2 to 48 hours between diagnosis and embolization was dependent upon availability of interventional radiologists and the clinical condition of the patients. Angiographic findings are presented in Table. Angiography was performed in 15 patients, of whom 14 (93%) showed effective embolization. Two patients underwent embolization of more than one vessel. Embolization was not technically possible in one patient owing to occlusion of the coeliac axis.

One patient with chronic pancreatitis had spontaneous thrombosis of a gastroduodenal artery pseudoaneurysmbleeding into a pseudocyst, and this was managed conservatively.

Table. Radiological findings and primary treatment

There were no episodes of re-bleeding after embolization, continued filling of the pseudoaneurysm on CT in 3 (20%) patients was effectively managed by further coil embolization.

No ischemic complications were noted after embolization. One patient developed a pulmonary embolism and needed anticoagulation therapy.

Two patients had emergency surgery for uncontrolled bleeding and abdominal compartment syndrome. One patient underwent a necrosectomy and washout, and the other was thought to have a bleeding gastric ulcer that was over-sewn in addition to splenectomy for segmental portal hypertension. Both patients continued to bleed in the immediate post-operative period and then underwent an emergency angiography. One was successfully embolized and the other had to return to the operating theater for surgical exclusion of the pseudoaneurysm due to coeliac axis occlusion and lack of radiological access to the splenic artery.

The average hospital stay was 18 days (range 5-64 days).

Outcome

Ninety-three percent of the patients with bleeding pseudoaneurysms were effectively treated by coil embolization, with a success rate of 77% after the first attempt. There were no episodes of re-bleeding causing hemodynamic compromise after embolization although CT identified continued filling of the pseudoaneurysm in 20% of the patients. Thirteen percent of the patients underwent primary surgery owing to hemodynamic instability, continued to bleed and required angiography and embolization or further surgery.

There was no procedure-related mortality. One patientwith chronic pancreatitis underwent embolization of the splenic artery and was anticoagulated following a pulmonary embolism post-procedure. This patient was re-admitted to the local hospital 2 weeks later for intraabdominal hemorrhage. At laparotomy, the patient was found to have an ooze from the pancreatic bed that was packed. Inability to achieve complete reversal of anticoagulation led to the death of the patient 35 days after embolization.

Discussion

The presentation of our patients reflects the norm for hemorrhagic pseudoaneurysms (abdominal pain, gastrointestinal bleeding and shock). The distribution of the visceral artery pseudoaneurysms related to pancreatitis has been fairly consistent in the literature, with splenic artery involvement (40%), gastroduodenal (30%), pancreatico-duodenal (20%), gastric (5%), hepatic (2%), and others including superior mesenteric, jejunal, colic, and aortic (1%-3%).[10-14]Our series also demonstrated a predominance of splenic artery pseudoaneurysms (44%), but we found a higher incidence of gastric (19%) and hepatic (19%) involvement and a lower incidence of gastroduodenal (13%) and pancreatico-duodenal (6%) artery involvement than reported elsewhere.[10-14]These differences need to be highlighted as this could have been missed at surgery without prior localization of the bleeding point.

Prompt diagnosis and the sequence of diagnostic investigations remain a challenge. Co-existent portal hypertension with varices and sepsis can delay diagnosis, and upper gastrointestinal endoscopy is usually the first investigation. The possibility of a bleeding pseudoaneurysm should be investigated by a CT scan even after apparent control of hemorrhage as gastric varices may co-exist (27% in our series), and bleeding ulcers may actually represent erosion of a pseudoaneurysm into the stomach.

Failure to arrest hemorrhage endoscopically should be followed up with CT and angiography if indicated to embolize the bleeding point.[5]Hemodynamic instability may require immediate damage control surgery. However, angiography should be performed post-procedure as severe peri-pancreatic inflammation and adhesions from previous operations make identification of a specific bleeding point challenging.

It is well established that rupture of pseudoaneurysms is unrelated to their size or the severity of pancreatitis.[14]This implies that surgeons have to be aware of this complication even in patients with "moderate" pancreatitis and therapeutic strategies have to be modified accordingly.

There have been scattered reports of spontaneous thrombosis of pseudoaneurysms[15]and one patient in our study was managed conservatively as he remained hemodynamically stable with no change in the size of the thrombosed pseudoaneurysm.

Angiographic control of bleeding should be attempted as early as possible. There are many reports of successful transcatheter embolisation in hemodynamically unstable patients. The reported embolization success rates range from 79% to 100% and mortality rates after embolization from 6% to 33%.[5-7,16-18]In our series the embolization rate was 93%, and there was no procedurerelated mortality. Refilling of the pseudoaneurysm after embolization has been effectively controlled by repeated attempts at embolization using a combination of coils and spongistatin. Udd et al[17]have reported an overall success rate of angioembolization of 67% including reembolization for recurrent bleeding in patients with chronic pancreatitis. In our series, 3 patients needed a second attempt at embolization with the size of pseudoaneurysm making no discernible difference. One patient had 2 feeding vessels, of which one (pancreaticoduodenal artery) was missed at initial angiography.

Complications due to embolization in the pancreatic and visceral area are limited. The risk of ischemia in the upper gastrointestinal tract is minimal if the normal collateral pathways are patent and appropriate embolic agents are used. Splenic artery occlusion leads to sectorial infarcts that are of little clinical significance and usually resolve spontaneously. Occlusion of the common hepatic artery has no clinical consequences in the presence of a normally patent portal vein. A self-limiting postembolization syndrome is often encountered in clinical practice. Splenic and intestinal necrosis has been reported by Mauro et al[12]and Boudghène et al.[16]

A negative angiography, a source that cannot be treated by or re-bleeding after embolization is the usual indication for surgery.[18-20]We found that persistent bleeding into a partially embolized pseudoaneurysm can be successfully managed by repeat angiography and embolization, as often there is a second feeding vessel that needs to be targeted.

Operative mortality rates of 10%-50% have been reported in the literature with the initial hemodynamic state, transfusion requirement and category of pancreatitis affecting survival.[4-7,21-23]Tsiotos et al[13]have shown a significant correlation between the frequency of bleeding episodes and the number of debridements for acute necrotising pancreatitis. Bergert et al[5]have suggested that embolization in combination with interventional drainage for management of infected necrosis may leadto a better outcome compared to an aggressive operative approach.

Pseudoaneurysms related to large pseudocysts are probably best treated with surgical excision because of the difficulty of embolizing these cavities, in addition to the need for surgical drainage of the pseudocyst. Our group of patients did not require surgical intervention for associated complications of pancreatitis, which were effectively managed by radiological or endoscopic drainage.

In conclusion, pseudoaneurysms are unrelated to the severity of pancreatitis and major hemorrhage can occur irrespective of their size. At high suspicion of the disease, early specialist referral and radiological intervention is recommended.

Funding:None.

Ethical approval:Not needed.

Contributors:PA proposed the study. SH wrote the first draft. All authors contributed to the design and interpretation of the study and to further drafts. HN is the guarantor.

Competing interest:No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

1 White AF, Baum S, Buranasiri S. Aneurysms secondary to pancreatitis. AJR Am J Roentgenol 1976;127:393-396.

2 El Hamel A, Parc R, Adda G, Bouteloup PY, Huguet C, Malafosse M. Bleeding pseudocysts and pseudoaneurysms in chronic pancreatitis. Br J Surg 1991;78:1059-1063.

3 Stroud WH, Cullom JW, Anderson MC. Hemorrhagic complications of severe pancreatitis. Surgery 1981;90:657-665.

4 Kiviluoto T, Kivisaari L, Kivilaakso E, Lempinen M. Pseudocysts in chronic pancreatitis. Surgical results in 102 consecutive patients. Arch Surg 1989;124:240-243.

5 Bergert H, Hinterseher I, Kersting S, Leonhardt J, Bloomenthal A, Saeger HD. Management and outcome of hemorrhage due to arterial pseudoaneurysms in pancreatitis. Surgery 2005;137: 323-328.

6 Gambiez LP, Ernst OJ, Merlier OA, Porte HL, Chambon JP, Quandalle PA. Arterial embolization for bleeding pseudocysts complicating chronic pancreatitis. Arch Surg 1997;132:1016-1021.

7 Beattie GC, Hardman JG, Redhead D, Siriwardena AK. Evidence for a central role for selective mesenteric angiography in the management of the major vascular complications of pancreatitis. Am J Surg 2003;185:96-102.

8 Reber PU, Uhl W, Büchler MW. Pancreatic pseudocysts in chronic pancreatitis. Differential diagnosis and therapy. Chirurg 1997;68:881-887.

9 Stabile BE, Wilson SE, Debas HT. Reduced mortality from bleeding pseudocysts and pseudoaneurysms caused by pancreatitis. Arch Surg 1983;118:45-51.

10 Mallick IH, Winslet MC. Vascular complications of pancreatitis. JOP 2004;5:328-337.

11 Giles RA, Pevec WC. Aortic pseudoaneurysm secondary to pancreatitis. J Vasc Surg 2000;31:1056-1059.

12 Mauro MA, Jaques P. Transcatheter management of pseudoaneurysms complicating pancreatitis. J Vasc Interv Radiol 1991;2:527-532.

13 Tsiotos GG, Munoz Juarez MM, Sarr MG. Intraabdominal hemorrhage complicating surgical management of necrotizing pancreatitis. Pancreas 1996;12:126-130.

14 Mortelé KJ, Mergo PJ, Taylor HM, Wiesner W, Cantisani V, Ernst MD, et al. Peripancreatic vascular abnormalities complicating acute pancreatitis: contrast-enhanced helical CT findings. Eur J Radiol 2004;52:67-72.

15 Vanlangenhove P, Defreyne L, Kunnen M. Spontaneous thrombosis of a pseudoaneurysm complicating pancreatitis. Abdom Imaging 1999;24:491-493.

16 Boudghène F, L'Herminé C, Bigot JM. Arterial complications of pancreatitis: diagnostic and therapeutic aspects in 104 cases. J Vasc Interv Radiol 1993;4:551-558.

17 Udd M, Leppäniemi AK, Bidel S, Keto P, Roth WD, Haapiainen RK. Treatment of bleeding pseudoaneurysms in patients with chronic pancreatitis. World J Surg 2007;31:504-510.

18 Nicholson AA, Patel J, McPherson S, Shaw DR, Kessel D. Endovascular treatment of visceral aneurysms associated with pancreatitis and a suggested classification with therapeutic implications. J Vasc Interv Radiol 2006;17:1279-1285.

19 de Perrot M, Berney T, Bühler L, Delgadillo X, Mentha G, Morel P. Management of bleeding pseudoaneurysms in patients with pancreatitis. Br J Surg 1999;86:29-32.

20 Woods MS, Traverso LW, Kozarek RA, Brandabur J, Hauptmann E. Successful treatment of bleeding pseudoaneurysms of chronic pancreatitis. Pancreas 1995;10:22-30.

21 Carr JA, Cho JS, Shepard AD, Nypaver TJ, Reddy DJ. Visceral pseudoaneurysms due to pancreatic pseudocysts: rare but lethal complications of pancreatitis. J Vasc Surg 2000;32: 722-730.

22 Deshmukh H, Rathod K, Garg A, Sheth R, Kulkarni S. Transcatheter embolization as primary treatment for visceral pseudoaneurysms in pancreatitis: clinical outcome and imaging follow up. Indian J Gastroenterol 2004;23:56-58.

23 Flati G, Andrén-Sandberg A, La Pinta M, Porowska B, Carboni M. Potentially fatal bleeding in acute pancreatitis: pathophysiology, prevention, and treatment. Pancreas 2003; 26:8-14.

March 4, 2010

Accepted after revision July 31, 2010

Author Affiliations: Department of Hepatobiliary Surgery (Sethi H, Prachalias A, Rela M and Heaton N) and Department of Hepatobiliary and Interventional Radiology (Peddu P, Kane P and Karani J), Institute of Liver Studies, Kings College London School of Medicine at Kings College Hospital, Denmark Hill, London SE5 9RS, UK

Prof. Nigel Heaton, Department of Hepatobiliary Surgery, Institute of Liver Studies, Kings College London School of Medicine at Kings College Hospital, Denmark Hill, London SE5 9RS, UK (Email: Nigel.Heaton@kch.nhs.uk)

© 2010, Hepatobiliary Pancreat Dis Int. All rights reserved.