Original Paper Received: October 4, 2006 Accepted after revision: February 21, 2007 Published online: August 14, 2007
Pancreatology 2007;7:360–369 DOI: 10.1159/000107396
Endovascular Treatment of Arterial Bleeding in Patients with Pancreatitis G. Mansueto a D. Cenzi a M. D’Onofrio a R. Salvia b L. Gottin c A.A. Gumbs d R. Pozzi Mucelli a a
Department of Morphological and Biomedical Sciences, Radiology Institute, b Department of Surgery Sciences, Department of Anethesiology and Surgery, Intensive Care Unit, University of Verona, Verona, Italy; and d Department of Surgery, New York-Presbyterian, The University Hospitals of Columbia and Cornell, New York, N.Y., USA c
Key Words Pancreatitis Arterial disruption Pseudoaneurysm Bleeding pseudocyst Transcatheter embolization
ture, we found that embolization is less invasive and, at least, as successful as surgery. Thus, it should be considered the first choice in pancreatitis arterial complications. Copyright © 2007 S. Karger AG, Basel and IAP
Abstract Purpose: To assess the technical and clinical success of endovascular treatment of arterial bleeding in pancreatitis. Materials and Methods: From 1992 to 2005, 28 patients with pancreatitis underwent endovascular treatment of associated arterial lesions. Fifteen patients were affected by acute pancreatitis and 13 by chronic pancreatitis. The diagnosis was obtained according to medical history and clinical and laboratory evidence of disease. Arterial involvement was diagnosed by non-invasive imaging and angiography. After treatment, all patients underwent CT scanning at a minimum of 15, 30 and 90 days. We evaluated the feasibility of embolization and patients’ survival at 90 days. Results: Transcatheter embolization was feasible in 26/28 patients (93%). In 2 patients with acute pancreatitis, selective catheterization failed so we could not proceed with the angiographic approach. After treatment, there were 3/26 rebleeds (11.5%), all of whom died within the first week. At 90 days’ follow-up, 21/26 patients (81%) were alive. Two of 26 patients (8%) suffered splenic complications. Among the 13 patients with acute pancreatitis, 8 (61.5%) were alive after 90 days. All 13 patients with chronic pancreatitis were alive after 90 days. Conclusions: Comparing our results with the surgical litera-
© 2007 S. Karger AG, Basel and IAP 1424–3903/07/0074–0360$23.50/0 Fax +41 61 306 12 34 E-Mail [email protected]
Accessible online at: www.karger.com/pan
Major arterial hemorrhagic complications in pancreatitis are an infrequent but life-threatening condition [1, 2]. As defined by the 1992 Atlanta International Symposium on Acute Pancreatitis , massive bleeding in severe acute pancreatitis is a consequence of the erosion of peripancreatic arteries (arterial disruption) by the activated proteolytic enzymes . When pancreatitis is associated with a pseudocyst, the progressive enlargement of the pseudocyst induces necrotizing arteritis because of the added pressure on the vessel walls and from the action of proteolytic enzymes. This leads to the erosion of the vessel walls, resulting in hemorrhage into the pseudocyst. In the literature, this condition is referred to as a pseudoaneurysm or a bleeding pseudocyst. If the arterial bleeding fills the entire lumen of the pseudocyst, it behaves more like a pseudoaneurysm with an inflammatory thickened wall. If there is evidence of bleeding inside the pseudocyst, but a large amount of proteolytic enzymes still fills its lumen, this condition is better defined as a bleeding pseudocyst.
Giancarlo Mansueto Department of Morphological and Biomedical Sciences Institute of Radiology, University Hospital ‘GB Rossi’ Piazza L.A. Scuro, 10, I–37134 Verona (Italy) Tel. +39 045 807 4343, Fax +39 045 827 7808, E-Mail [email protected]
The literature reports the presence of pseudoaneurysms in 10–21% of patients with chronic pancreatitis [2, 4]. The incidence of bleeding pseudocysts in chronic pancreatitis has been reported to be 6–10% . Hemorrhagic complications are expected in 6–31% of patients with pancreatic pseudocyst  and in 7–14% of those suffering from chronic pancreatitis . Hemorrhagic shock is a common clinical presentation of ruptured pseudoaneurysm, with a mortality rate of more than 50%; therefore, immediate diagnosis and treatment is paramount . Surgery or interventional treatments have been proposed, but the optimal treatment is still controversial because only observational studies are available for analysis [9, 10]. The aim of this retrospective study was to assess the feasibility of transcatheter embolization for arterial complications in pancreatitis and to evaluate its clinical success.
Materials and Methods This retrospective study was approved by our internal review board. Oral and written consent for the endovascular procedure was obtained for all the conscious and hemodynamically stable patients. In unconscious patients, the informed consent for the treatment was obtained by consulting with the patient’s relatives. At our institution 28 consecutive patients (22 males and 6 females) with pancreatitis and arterial lesions were treated in the angiography suite over a 14-year period, from January 1992 to December 2005. The mean age was 52 years (range 26–72 years). The diagnosis of acute or chronic pancreatitis was obtained according to clinical and laboratory evidence of disease as well as past medical history. Fifteen patients were affected by severe acute pancreatitis and 13 by chronic pancreatitis. All the patients had a clinical history of gastrointestinal hemorrhage, which was reported as melena and/or hematemesis or diagnosed by a decrease in hemoglobin levels. Endoscopic examination to determine the source of bleeding was the first examination in 8/28 patients. Endoscopy was inconclusive in 4 patients; in 2 patients there was evidence of recent gastrointestinal bleeding, but there was no definitive localization of the bleeding site. In 1 patient, the source of bleeding was identified and placement of a clip was attempted endoscopically. Because this patient suffered a rebleed after 3 days, a transcatheter embolization was performed. In another patient, wirsungorrhage was identified so the patient eventually underwent endovascular treatment. All patients were evaluated by ultrasonography, color Doppler, angio-CT or angio-MR, followed by digital subtraction angiography. Non-invasive vascular imaging (Doppler US, angio-CT, angio-MR) could accurately diagnose the site of arterial bleeding in 22/28 patients (79%). In the remaining 6/28 patients (21%), the arterial lesions as well as the site of bleeding were only clearly identified by angiography.
Endovascular Treatment of Arterial Bleeding in Patients with Pancreatitis
Angiography was performed in 17/28 patients (61%) in shock, defined as hypotension (systolic blood pressure !90 mm Hg or mean arterial pressure !65 mm Hg), tachycardia, or signs of tissue hypoperfusion (lactic acidosis, oliguria, impaired mental status). In these patients angiography was done under general anesthesia, with invasive hemodynamic monitoring of the central venous and arterial pressure. Peripheral oxygen saturation and end-tidal CO2 were also monitored. The remaining 11/28 patients were hemodynamically stable. In these cases angiography was performed with monitored anesthesia care and local anesthesia (intravenous midazolam administration, and subcutaneous lidocaine 2% at the site of percutanous access). In all patients trans-femoral percutaneous arterial access is obtained. We always follow the same standard protocol, using a 5F introducer sheath and a 5F J curve catheter for diagnostic angiography. We perform the study of the superior mesenteric artery and the celiac trunk first, with a 30-ml bolus injection of contrast medium at 5 ml/s. Delayed images are taken in the venous phase whenever possible, to study the porto-systemic veins. This technique permits a diagnostic angiography of the complete pancreatic vascular anatomy. If the arterial lesion is identified, we proceed with selective catheterization to achieve embolization of the arterial lesion. Otherwise, we perform superselective injections starting with the most likely site according to clinical or previous noninvasive imaging. The same 5F J Curve catheter and a 0.035 J hydrophilic guide wire (Terumo Corporation, Tokyo, Japan) are used for superselective catheterization of the splenic, common hepatic, gastroduodenal and pancreaticoduodenal arteries, with a 12–30 ml bolus injection of contrast medium at 3– 5 ml/s. When the arterial lesion was found, we proceeded with endovascular management, using the coaxial catheter technique. All afferent vessels to the arterial lesion were selectively catheterized with a 3F microcatheter (Tracker-18 unibody Target Therapeutics, Fremont, Calif., USA) to enable placement of microcoils (Tornado Embolisation Coils, W Cook Europe, Bjaeverskov, Denmark) or injection of acrylic glue (Histoacryl, B. Braun, Melsungen, Germany) supplemented with Lipiodol UF (Guerbet, Aulnay-Sous-Bois, France). A maximum of 15 metal coils and a range of 1–2 ml of Histoacryl were used. Coils were introduced into the coaxial catheter using an insertion kit and were deployed via a wire pusher. Histoacryl was mixed in a 3-way chamber with Lipiodol in a 1: 1 ratio. Injection was performed selectively via a microcatheter until stasis of the arterial blood flow was achieved under fluoroscopic guidance. To avoid polymerization within the microcatheter, the system was flushed with a 20% glucose solution before and immediately after the embolization procedure. After treatment, the microcatheter was removed and a control angiography was performed via the indwelling mother catheter. Embolization was considered successful when the cessation of blood flow in the target area was demonstrated radiologically. After the endovascular procedure, all patients were clinically admitted to an intensive care unit for at least 24 h to provide hemodynamic stability and to monitor clinical and laboratory parameters in order to diagnose any recurrent rebleeding. Central venous blood pressure, arterial pressure, heart rate, and peripheral oxygen saturation were monitored, as well as serial hematocrit, hemoglobin, blood gas analysis and lactate levels. Rebleeding was defined as a recurrent episode of bleeding from the previously identified vessels after a period of hemodynamic stability,
which in our experience occurred after a period of at least 12 h. When there was clinical evidence of recurrent hemorrhage (continuous low arterial pressure, tachycardia, decrease of Hgb rate) urgent CT scan was obtained and consequent emergency laparatomy was performed. As a result, the definitive diagnosis of the site of rebleeding was obtained during surgery. In contrast, patients who recovered without any clinical or laboratory evidence of hemorrhage underwent CT scanning at 15, 30 and 90 days. We evaluated both the technical and clinical success of transcatheter arterial embolization. The technical success was identified with the feasibility of transcatheter embolization, while the clinical success was regarded as its effectiveness, as well as freedom from pancreatitis-related death at 90 days.
Table 1. Arterial involvement: topographic localization
Splenic artery Gastroduodenal artery Middle colic artery Pancreaticoduodenal arcades Celiac trunk Gastroepiploic artery Left gastric artery Total
Severe acute pancreatitis
Chronic Total pancreatitis
5 4 3 1 1 1 –
6 4 – 2 – – 1
11 8 3 3 1 1 1
Transcatheter arterial embolization was attempted in 28 patients. Among 15 patients affected by acute severe pancreatitis, 10 patients developed an arterial disruption and the remaning 5 a bleeding pseudocyst. Among the 13 patients with chronic pancreatitis, 6 developed a bleeding pseudocyst and the remaining 7 a pseudoaneurysm. In total, we observed 10 arterial disruptions, 11 bleeding pseudocysts and 7 pseudoaneurysms. The localization of arterial involvement is reported in table 1. The interventional procedure was completed in 26/28 patients (feasibility rate 93%). In 2 patients, a proximal dissection of the bleeding vessel was induced by the attempts of selective catheterization for embolization and even though temporary cessation of bleeding was achieved (immediate technical failure 7%), the procedures had to be stopped. It is worth noticing that both patients were affected by acute necrotizing pancreatitis and arterial disruption. Furthermore, the dissected vessel was the same, the middle colic artery. After dissection, one patient was treated with intravenous fluids and underwent immediate open surgery to repair the arterial disruption and reduce necrotic collections; he completely recovered in a couple of months. The other patient remained hemodynamically unstable despite intravenous resuscitation and emergency laparatomy, and died secondary to circulatory failure. Among the patients who underwent feasible endovascular management, coils were placed proximally and distally to the bleeding site in 19/26 patients (73%) (fig. 1), and we used acrylic glue for the endovascular treatment of 6 patients (23%) (fig. 2). In order to obtain a cessation of arterial hemorrhage more quickly, coil and acrylic glue were used simultaneously in only 1 case (4%) (table 2). The mean treatment time was 70 min (range 30–90 min). 362
There were 3/26 rebleeds (rebleeding rate 11.5%) after embolization; they all occurred within the first week. Specifically, rebleeding occurred in 2 patients after coil positioning and one after embolization with acrylic glue. One patient died secondary to hemorrhagic shock despite resuscitation within 24 h after endovascular treatment. In the other 2 patients who had clinical signs of hemorrhage, rebleeding was found on CT scan. As there was evidence of multiple sites of hemorrhage, it was decided together with the surgeon to proceed with emergency laparotomy. Unfortunately, both patients subsequently died. In conclusion, the effectiveness of transcatheter embolization was 88.5%. Notably, all these rebleeds occurred in patients with an arterial disruption in the setting of severe acute pancreatitis. Another 2 patients affected by severe acute pancreatitis died secondary to the development of MODS (multiorgan dysfunction syndrome), which was not associated with rebleeding, at 25 and 35 days after endovascular treatment. At 90 days’ follow-up 21/26 patients were still alive (overall survival: 81%). Among 13 of 26 patients treated for arterial damage secondary to severe acute pancreatitis (50% of our experience), only 8 of 13 (61.5%) were still alive after 90 days. In contrast, among patients treated for an arterial complication in the setting of chronic pancreatitis (13/26, 50%), all 13 patients were still alive after 90 days (table 3). Furthermore, during the follow-up, splenic infarctions occurred in 2 of 26 patients who were embolized successfully (complication rate 8%). This was considered a complication of the radiological treatment and both patients underwent splenectomy. In both cases the arterial lesions treated were located next to the splenic hilum. One patient suffered abdominal pain and developed fever one week after endovascular treatment: a CT Mansueto et al.
Fig. 1a–f. Bleeding pseudocyst in chronic pancreatitis: diagnosis and treatment. a MR T2-weighted fast SE images highlight a large
hyperintense bilobated cystic lesion at the pancreatic head (arrows): the walls are thick and the content is quite heterogeneous, with some hypointense inclusions. b, c Post-gadolinium T1weighted image as well as CT scan in arterial phase, highlight a small hypervascular lesion next to the gastroduodenal artery (arrowhead), which is pathognomic for bleeding inside the pseudocyst. d Selective angiography of common hepatic artery confirms the arterial lesion, which is fed by pancreaticoduodenal arcades (arrowheads). e Selective angiography of the common hepatic artery after the treatment demonstrates coils inside the embolized vessels (arrows), reproducing vascular anatomy; the complete ex-
Table 2. Endovascular treatment of
clusion of the arterial lesion is highlighted. f, g At 1 week after the embolization the patient underwent endoscopical transduodenal pseudocyst drainage. After the direct pucture of the pseudocyst (asterisk), the lumen is completely enhanced, without any spreading into the arterial vessels. Metallic coils (arrows) are highlighted next to the antero-inferior wall of the pseudocyst (f). An endoprothesis (arrowhead) is then placed between the pseudocyst and the duodenum, to realize the endoscopical drainage of pseudocyst (g). h At 3 months’ follow-up post-gadolinium MR sequences demonstrate the reduction of the pseudocyst, as well as no evidence of rebleeding. Coils are still in place, as highlighted by the evidence of small metallic artifacts (arrowheads).
Severe acute pancreatitis
Coils Acrylic glue Coils + acrylic glue
6/13 2/13 0/13
3/13 1/13 1/13
9/13 0/13 0/13
1/13 3/13 0/13
arterial lesions: materials and techniques
scan at 15 days revealed a hypodense intrasplenic collection, which was regarded as suspicious for abscess secondary to the splenic infarction. The other patient recovered and remained asymptomatic; however, at 1 month’s follow-up, CT scan demonstrated a massive
splenic infarction. Because of the high risk of septic complications and splenic rupture, splenectomy was undertaken. No other complications related to embolization were observed.
Endovascular Treatment of Arterial Bleeding in Patients with Pancreatitis
Fig. 2. Radiological treatment of a bleeding pseudocyst in chronic pancreatitis, after pseudocyst-jejunum anastomosis, using acrylic glue. a Axial CT scan in the arterial phase highlighted a large pseudocyst which displaces the gastroduodenal artery (white arrow). A slight hyperdensity is seen next in the most declive part of the pseudocyst (white arrowheads). These signs are suggestive of a bleeding pseudocyst, even though no evidence of active hemorrhage inside the pseudocyst is demonstrated. b Selective angiography of celiac trunk highlights a metallic clip in the site of endoscopic tying of gastroduodenal artery (arrowhead). Gastroepiploic artery at its origin is revascularized by small arte-
rial vessels. Active bleeding is demonstrated at this site (arrow). c In a more advanced angiographic phase, contrast medium enhances the pseudocyst and the jejunum (arrowheads), as a consequence of previous pseudocyst-jejunum anastomosis. d, e After injection of acrylic glue with additional lipiodol in gastroduodenal artery, at native (d) and digital subtraction angiography (e), all the small arterial vessels appear occluded upstream and downstream to the bleeding site (arrows). Embolization material is demonstrated in gastroepiploic artery, inside the pseudocyst and in the anastomotic jejunum ansa (arrows).
Table 3. Results of radiological treatment of pancreatitis-associated arterial lesions at 3-months follow-up
Severe acute pancreatitis Chronic pancreatitis Total
arterial disruption bleeding pseudocyst bleeding pseudocyst pseudoaneurysm
90 days’ mortality
0/8 0/5 2/9 0/4
5/8 5/5 9/9 4/4
5/8 0/5 0/9 0/4
Mansueto et al.
Patients with chronic pancreatitis were discharged after a mean length of stay (LOS) of 19 days (range 10–30 days), while patients suffering from acute necrotizing pancreatitis were discharged after a mean LOS of 49 days (range 25–90 days). All 21 patients remaining showed complete resolution of their arterial lesions on repeat CT scans at 90 days’ follow-up. The mean followup in our experience was 30 months (range 4–111 months).
Arterial bleeding is a serious, albeit rare, complication of pancreatitis with a high mortality. Usually, it occurs late in the time course of disease and in the postoperative period [10, 11]. According to our experience, spontaneous bleeding always occurs in the same sites, at the pancreatic tail next to the splenic hilum or at the head of the pancreas. The more commonly involved vessels are the splenic artery, the gastroduodenal artery, the pancreaticoduodenal arcades and, lastly, the left gastric artery. Anatomy explains these clinical presentations. On the left side, bleeding arterial lesions usually involve the last tract of the splenic artery, immediately upstream from the origin of the left gastroepiploic artery and the collateral omental branches. At this site, retroperitoneal splenic branches become intraperitonal, so they are more mobile and may induce traction until they break down over the stable retroperitoneal vessels. The same mechanism can be observed in right-sided bleeds. We look for arterial lesions at the point in which the gastroduodenal artery divides into the right gastroepiploic artery and the pancreaticoduodenal branches because this site corresponds to a locus minori resistentiae, where the artery enters via the retroperitoneum into the omental root and becomes intraperitoneal (fig. 1). Occasionally, the diagnosis can be made during the follow-up period of acute or chronic pancreatitis. It is worth noting that pseudoaneurysms always rupture, therefore, treatment is mandatory. Multiple recurrences of gastrointestinal hemorrhage, especially in association with episodes of pancreatic pain, should lead to a suspicion of a pseudoaneurysm or a bleeding pseudocyst. Endoscopy has become the method of choice for the diagnosis and treatment of gastrointestinal hemorrhage, secondary to etiologies such as peptic ulcer disease or varices. In instances where localization is impossible, it is well known that angiography or surgery should be employed . Negative endoscopic findEndovascular Treatment of Arterial Bleeding in Patients with Pancreatitis
ings of peptic ulcer disease in a patient with pancreatitis should alert the clinician to the possibility of a pseudoaneurysm or bleeding pseudocyst. The bleeding may be due to a communication between a small pseudoaneurysm or bleeding pseudocyst and the main pancreatic duct. The pseudoaneurysm can erode into the pancreatic duct, with hemorrhage coming from the ampulla, this is referred to as hemosuccus pancreaticus or ‘Wirsungorrhage’ [13–15]. This can sometimes be diagnosed by direct endoscopic visualization of the intermittent hemorrhagic flow from the major papilla. In these cases treatment is mandatory. Arterial bleeding in acute pancreatitis that is not associated with a pseudocyst is defined as ‘arterial disruption’ and appears because of the maceration of the arterial wall by necrotic fluid collections (fig. 3). In the course of severe acute pancreatitis, it is characterized by a severe drop in hematocrit. This critical condition corresponds to the so-called ‘hemorrhagic pancreatitis’ and is considered the most severe manifestation in the spectrum of vascular complications . In our experience, embolization was performed only when the site of bleeding was identified. We found that in patients in hemorrhagic shock and suspected gastrointestinal bleeding, angiography can localize the site of bleeding if a standard protocol is followed. Conversely, if a patient develops a mild recurrent bleeding, angiographic localization is much more difficult and can be inconclusive. As a result, we prefer not to proceed with embolization and observe the patient. We repeat angiography as soon as there is a clinical and/or laboratory finding consistent with re-bleeding. When a definitive site of hemorrhage cannot be found, embolization should be avoided because the ‘blind’ endovascular tying of suspected arteries can result in useless or unsafe ischemia. Coils appear to be extremely efficient in the treatment of pseudoaneurysms that have a fibrotic pseudowall. Embolization with coils should be preferred because permanent arterial occlusion can be attained . The occlusion of all the efferent vessels and afferent vessels to the lesion with coils is technically similar to surgical tying, but is clearly less invasive (fig. 1). We prefer not to place coils inside the pseudoaneurysm unless it is necessary to occlude the efferent vessels. As a general rule, coils can be placed only when selective catheterization of all efferent and afferent vessels to the arterial lesion is possible . Acrylic glue is the best agent for embolizing the afferent vessel of the arterial lesion whenever it is technically impossible to catheterize the efferent vessel, or in Pancreatology 2007;7:360–369
Fig. 3. Radiological treatment of arterial disruption in severe acute pancreatitis. a Contrast-enhanced CT scan highlights a fluid collection at the site of the pancreatic head, almost completely destroyed by necrosis. A tiny hyperdensity inside the collection induces the suspicion of arterial disruption, but the bleeding source cannot be clearly identified. b Selective angiography demonstrates an active bleeding from the gastroduodenal artery (arrow). Angiography is pathognomic for an arterial disruption in acute pancreatitis, with contrast medium spreading directly into the peripancreatic fluid collection (arrowheads). c, d Emergency radiological treatment is carried out positioning microcoils downstream and upstream to the arterial lesion (arrows), at the origin of the gastroepiploic artery, superior pancreaticoduodenal artery and into gastroduodenal artery, until its origin from common hepatic artery (c). Selective angiography of the superior mesenteric artery is performed to definitely exclude the arterial lesion placing coils along the inferior pancreatoduodenal arcades (arrows) (d). e, f At the end of the procedure, selective angiographies of superior mesenteric artery (e) and celiac trunk (f) confirm complete embolization of the arterial lesion.
cases where multiple collaterals of small diameter revascularize the damaged arterial tract immediately downstream from the selective catheterization . The glue is liquid and has progressive polymerization, so that when it is injected upstream to the arterial lesion, it is 366
able to reach and occlude the arterial tract downstream. As a result, it is possible to occlude the downstream vessel of the arterial lesion with controlled polymerization as well as fill the lesion and then occlude the proximal vessels (fig. 2) . Re-absorbable embolic materials, Mansueto et al.
Table 4. Mortality rate and complications reported in the literature: treatment in pancreatitis
Stabile et al. 
Gambiez et al. 
embolization surgery embolization
2 8 14
0% (0/2) 13% (1/8) 7% (1/14)
De Perrot et al. 
Beattie et al. 
Deshmukh et al. 
surgery embolization surgery embolization surgery embolization
8 3 7 11 8 30
13% (1/8) 33% (1/3) 14% (1/7) 33% (3/11) 25% (2/8) 3% (1/30)
Bergert et al. 
embolization surgery embolization
16 19 26
19% (3/16) 21% (4/19) 19% (5/26)
1 technical failure 1 failure, 1 sepsis/death 3 failed embolization, 1 duodenal necrosis, 1 aortic thrombosis/death 1 hepatic failure/death 1 rebleeding 2 subphrenic abscesses 1dissection, 3 rebleedings 2 rebleedings 2 rebleedings, 1 technical failure, 2 partial splenic infarcts 2 rebleedings 7 rebleedings 3 rebleedings, 2 technical failures, 2 splenic infarcts
such as gelatin sponge, should not be utilized because proteolytic enzymes rapidly degrade them. When it is not technically possible to catheterize the efferent vessels, direct ultrasonic-guided percutaneous embolization of pseudoaneurysms with thrombin may be attempted [19–22]. In our experience, embolization was successful in all patients with a pseudoaneurysm and/or a bleeding pseudocyst, specifically in 13 patients affected by chronic pancreatitis and 5 patients affected by acute pancreatitis. Only 4 of 10 patients (40%) affected by severe acute pancreatitis and subsequent arterial disruption were still alive at 3 months of follow-up. This can be explained by the fact that acute bleeding associated with arterial disruption is due to an excessive amount of vessel maceration and as a result it becomes very difficult to identify sites that have undergone no enzymatic lytic action. Furthermore, the fluid collections often erode more arteries, making embolization extremely difficult (fig. 3) [2, 23]. Our data suggest that morbidity and mortality are affected by the type of arterial lesion rather than by its development in acute or chronic pancreatitis. Coils and/or acrylic glue provide clinical results similar to the best surgical reports, but are much less invasive. Emergency surgery for arterial disruption has a high failure rate, with a greater than 80% risk of recurrent bleeding [23–27] and a mortality rate greater than 50% . In our experience, the overall survival rate with radiologic treatment is 81%. If we consider the endovascular treat-
ment of arterial lesions in chronic pancreatitis, the 3month survival is 100%, while in acute pancreatitis it is 61.5%. Transcatheter arterial embolization appears to be a safe treatment modality in patients with arterial bleeding secondary to pancreatitis. In our study only 2 patients experienced complications related to the endovascular approach; notably, both arterial lesions were next to the splenic hilum. This complication is comparable to that reported in the embolization of distal splenic artery aneurysms . Positioning coils distally to the lesion and adjacent to the hilum may reduce the feeding collateral circulation at the end of the organ. Distal embolic events related to endovascular procedures may also play a role, such as splenic infarct and infection, as occurred in our study . In such cases the best treatment is pseudocyst-jejunal anastomosis and splenectomy after surgical tying or embolization of the splenic artery for hemodynamic stabilization. Conversely, the endovascular approach can be an alternative to surgery in patients who develop arterial lesions in vessels whose embolization could lead to ischemia of other organs, such as the superior mesenteric artery or hepatic artery in the setting of portal thrombosis. In such cases, surgery is traditionally considered as the treatment of choice. According to our previous experience in patients with hemorrhage after main pancreatic surgery , we think that placing a covered stent at the site of the arterial lesion should be considered as a potential alternative to surgery.
Endovascular Treatment of Arterial Bleeding in Patients with Pancreatitis
In the interpretation of the findings in this retrospective study some bias must be considered. Firstly, only patients with arterial complications in pancreatitis and with angiographic evidence of hemorrhage were treated and enrolled in our study. No patient who was treated endoscopically or surgically was enrolled. Therefore, we did not directly compare different treatment modalities. However, this is the same limitation reported by the majority of studies in the published literature. It is extremely difficult to collect a sufficient number of cases in the same center and conduct a prospective comparative clinical trial. In order to reduce the impact of this bias, we compared our results with other reports in the literature, both surgical and endovascular [5, 10, 25, 29–35] (table 4).
According to our experience, transcatheter embolization is a feasible and effective treatment for patients with hemorrhagic complications of pancreatitis. High clinical success rates at the 3-month follow-up as well as low morbidity and mortality rates confirm the results reported in the literature (table 4) [8, 31]. Comparing our results with surgical ones in the literature, endovascular treatment appears to be less invasive and at least as successful as surgery. Therefore, we think it should be considered as the treatment of choice in arterial complications of pancreatitis, especially for pseudoaneurysms and bleeding pseudocysts.
References 1 Frey CF, Stanley JC, Eckhauser F: Haemorrhage; in Bradley EL (ed): Complications of Pancreatitis. Philadelphia, Saunders, 2002, pp 96–123. 2 Procacci C, Mansueto G, D’Onofrio M, Gasparini A, Ferrara RM, Falconi M: Non-traumatic abdominal emergencies: imaging and intervention in acute pancreatic conditions. Eur Radiol 2002;12:2407–2434. 3 Bradley EL III: A clinically based classification system for acute pancreatitis. Arch Surg 1993;128:586–590. 4 Burke JW, Erickson SJ, Kellum CD, Tegtmeyer CJ, Williamson BR, Hansen MF: Pseudoaneurysms complicating pancreatitis: detection by CT. Radiology 1986; 161: 447–450. 5 Stabile BE, Wilson SE, Debas HT: Reduced mortality from bleeding pseudocysts and pseudoaneurysms caused from pancreatitis. Arch Surg 1983; 118:45–51. 6 Sankaran S, Walt AJ: The natural and unnatural history of pancreatic pseudocyst. Br J Surg 1975;62:37–44. 7 Bresler L, Boissel P, Grosdidier J: Major haemorrhage from pseudocysts and pseudoaneurysms caused by chronic pancreatitis: surgical therapy. World J Surg 1991; 15:649– 653. 8 Lendrum R: Chronic pancreatitis; in Misiewicz JJ, Pounder RE, Venables CW (eds): Disease of Gut and Pancreas. London, Blackwell Scientific, 1994, pp 441–454. 9 Bergert H, Dobrowolski F, Caffier S, Bloomenthal A, Hinterseher I, Saeger HD: Prevalence and treatment of bleeding complications in chronic pancreatitis. Langenbecks Arch Surg 2004;389:504–510.
10 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. 11 Mortelè KJ, Mergo PJ, Taylor HM, Wiesner W, Cantisani V, Kalantari BN, Ros PR: Peripancreatc vascular abnormalities complicating acute pancreatitis: contrast-enhanced helical CT finding. Eur J Radiol 2004;52:67– 72. 12 Kramer SC, Gorich J, Rilinger N, Siech M, Aschoff AJ, Vogel J, Brambs HJ: Embolization for gastrointestinal hemorrhages. Eur Radiol 2000;10:802–805. 13 Nabi Yattoo G, Sultan Khuroo M, Ahmad Wani N, Alam Wani K, Ahmad Bhat F: Haemosuccus pancreaticus: a clinical challenge. J Gastroenterol Hepatol 1999;14:172–175. 14 Dasgupta R, Davies NJ, Williamson RC, Jackson JE: Haemosuccus pancreaticus: treatment by arterial embolization. Clin Radiol 2002;57:1021–1027. 15 Feng DH, Mauro MA: SIR 2003 film panel case 6: hemosuccus pancreaticus secondary to chronic pancreatitis. J Vasc Interv Radiol 2003;14:803–805. 16 Yamaguchi K, Futagawa S, Ochi M, Sakamoto I, Hayashi K: Pancreatic pseudoaneurysm converted from pseudocyst: transcatheter embolization and serial CT assessment. Radiat Med 2000; 18:147–150. 17 Yamakado K, Nakatsuka A, Tanaka N, Takano K, Matsumura K, Takeda K: Transcatheter arterial embolization of ruptured pseudoaneurysms with coils and n-butyl cyanoacrylate. J Vasc Interv Radiol 2000; 11: 66–72.
18 Schoder M, Cejna M, Langle F, Hittmaier K, Lammer J: Glue embolization of a ruptured celiac trunk pseudoaneurysm via the gastroduodenal artery. Eur Radiol 2000; 10: 1335– 1337. 19 Sparrow P, Asquith J, Chalmers N: Ultrasonic-guided percutaneous injection of pancreatic pseudoaneurysm with thrombin. Cardiovasc Intervent Radiol 2003; 26:312–315. 20 Puri S, Nicholson AA, Breen DJ: Percutaneous thrombin injection for the treatment of a post-pancreatitis pseudoaneurysm. Eur Radiol 2003;13(suppl 4):L79–L82. 21 Geoghegan T, Tuite D, McAuley G, O’Keeffe S, Torreggiani WC: Percutaneous thrombin injection for the treatment of a post-pancreatitis pseudoaneurysm of the gastroduodenal artery. Eur Radiol 2004; 14:2144–2145. 22 Krueger K, Zaehringer M, Lackner K: Percutaneous treatment of a splenic artery pseudoaneurysm by thrombin injection. J Vasc Interv Radiol 2005; 16:1023–1025. 23 Testart J, Boyet L, Perrier G, Clavier E, Peillon C: Arterial erosions in acute pancreatitis. Acta Chir Belg 2001;101:232–239. 24 El Hemel A, Parc R, Adda G, Bouteloup PY, Huguet C, Malafosse M: Bleeding pseudocysts and pseudoaneurysms in chronic pancreatitis. Br J Surgery 1991; 78:1059–1063. 25 De Perrot M, Berney T, Buhler L, Delgadillo X, Mentha G, Morel P: Management of bleeding pseudoaneurysms in patients with pancreatitis. Br J Surg 1999; 86:29–32. 26 Stanley JC, Frey CF, Miller TA, Lindenauer SM, Child CG 3rd: Major arterial haemorrhage: a complication of pancreatic pseudocysts and chronic pancreatitis. Arch Surg 1976;111:435–440.
Mansueto et al.
27 Mansueto G, Rozzanigo U, Iacono C, D’Onofrio M, Serio G, Procacci C: Gastroduodenal artery stump hemorrhage following pylorus-sparing Whipple procedure treatment with covered stents. Dig Surg 2002;19:237–240. 28 Saltzberg SS, Maldonado TS, et al: Is endovascular therapy the preferred treatment for all visceral artery aneurysms? Ann Vasc Surg 2005;19:507–515. 29 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.
Endovascular Treatment of Arterial Bleeding in Patients with Pancreatitis
30 Stroud WH, Cullom JW, Anderson MC: Hemorrhagic complications of severe pancreatitis. Surgery 1981;90:657–665. 31 Marshall GT, Howell DA, Hansen BL, Amberson SM, Abourjaily GS, Bredenberg CE: Multidisciplinary approach to pseudoaneurysms complicating pancreatic pseudocysts: impact of pretreatment diagnosis. Arch Surg 1996;131:278–283. 32 Gambiez LP, Ernst OJ, Merlier OA, Porte HL, Chambron JP, Quandalle PA: Arterial embolization for bleeding pseudocysts complicating pancreatitis. Arch Surg 1997; 132: 1016– 1021.
33 Sand JA, Seppanen SK, Nordback IH: Intracystic hemorrhage in pancreatic pseudocysts: initial experiences of a treatment protocol. Pancreas 1997; 14:187–191. 34 Elton E, Howell DA, Amberson SM, Dykes TA: Combined angiographic and endoscopic management of bleeding pancreatic pseudoaneurysms. Gastrointest Endosc 1997; 46: 544–549. 35 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.