Gastrointestinal bleeding complications in chronic pancreatitis are frequently attributed to peptic ulcer, erosive gastritis, or esophageal and gastric fundus varices. A bleeding pseudoaneurysm is a serious and rare complication of chronic pancreatitis. Consequently, randomized trials for assessing the relative benefits of operative vs. radiologic intervention are difficult. Clinicians therefore must rely on data available from observational studies – such that provided in this study.
The following three mechanisms account for pseudoaneurysms related to pancreatitis: 1) severe inflammation and enzymatic autodigestion of a pancreatic or peripancreatic artery producing arterial disruption with pseudoaneurysm formation; 2) an established pseudocyst eroding into a visceral artery, thereby converting the pseudocyst into a large pseudoaneurysm; and 3) a pseudocyst eroding the bowel wall with bleeding from the mucosal surface [5
]. The splenic artery is most commonly involved, followed by the gastroduodenal, pancreaticoduodenal, and hepatic arteries [1
Pseudoaneurysms can rupture into the gastrointestinal tract, peritoneal cavity, retroperitoneum, biliopancreatic ducts or pseudocysts. (bleeding into the biliopancreatic ducts is known as hemobilia or hemosuccus pancreaticus.) Clinically, a bleeding pseudoaneurysm typically manifest as silent anemia with melena or as intermittent massive bleeding into the gastrointestinal tract or abdominal cavity, both of which require emergency laparotomy.
Early localization of a pseudoaneurysm via imaging studies is crucial to further treatment. Ultrasound is of little diagnostic value for a bleeding pseudoaneurysm. A dynamic bolus computed tomographic scan is a useful noninvasive approach for detecting pseudoaneurysms [6
] and associated pseudocysts [11
]. However, small pseudoaneurysms can escape detection. In the past 2 decades, angiography has improved its diagnostic accuracy for pseudoaneurysms [1
]; in this series angiography reach a sensitivity rate of 100% (7/7). Two patients (22%) in this series underwent laparotomy without pre-operative diagnosis of a bleeding pseudoaneurysm. One patient (case 2) received emergency laparotomy without prior to surgery. The remaining patient (case 8) presented with a huge pseudocyst with abdominal pain; diagnosis of a bleeding pseudoaneurysm was made during surgery. Heightened clinical suspicion in patients with chronic pancreatitis with acute abdominal pain and signs of acute bleeding in combination with appropriate diagnostic modalities is mandatory for early and precise diagnosis.
The reported success of embolization is 79–100% and the reported mortality rate after embolization is 12–33% in patients with acute or chronic pancreatitis [2
]. The etiology, pathophysiology, clinical presentation, and treatment and prognosis of hemorrhagic complications for acute necrotizing or chronic pancreatitis differ. Reporting the bleeding complications associated with the 2 categories of pancreatitis together can result in confusion in treating bleeding pseudoaneurysms. In this series, 2 patients initially treated by embolization and 3 patients with rebleeding undergoing subsequent embolization had an overall treatment success rate of 20% (1/5) and a mortality rate of 20% (1/5). The low success rate of embolization in this series may be in part explained due to the definition for successful embolization in this study. Angiographic hemostasis was defined as successful when bleeding stopped with no rebleeding during follow-up; this definition differs from the definition in other series – no rebleeding within 48 hours after embolization [1
]. The selection bias and the small number of cases in this study may likely also influence outcomes for embolization.
Some authors consider surgery as the treatment of choice for bleeding pseudoaneurysms in patients with chronic pancreatitis [5
]. This opinion is supported by this series in that of the 9 surgical interventions to treat bleeding pseudoaneurysms, only 1 patient (1/8, 12.5%) rebled and no patient died because of surgery. By contrast, angiographic embolization as the treatment modality for bleeding pseudoaneurysms resulted in a rebleeding rate of 66.7% (2/3). These observational results can be explained partly by the relatively young age in patients receiving surgical treatment in this series (median age, 35 years). Nevertheless, the small number of patients in this study precludes a comparison of results between surgical intervention and arteriographic embolization.
Debate still exits about the best surgical procedure for treating bleeding pancreatic pseudoaneurysms. Some researchers suggest that proximal and transcystic ligation of a bleeding vessel with internal or external drainage of the cyst is superior to pancreatic resection [14
]. Conversely, others have suggested that pancreatic resection should be employed as it is the only certain way of preventing the very common problem of rebleeding [2
]. In this study, a pancreatectomy was performed in 4 patients, drainage of pseudocyst in 4, arterial ligation in 2, and resection of pseudoaneurysm followed by arteriorrhaphy in 2 (Table ). Only 1 patient rebled because of pseudocyst formation after distal pancreatectomy. We suggest that surgical treatment for bleeding pseudoaneurysms should be individualized.
Location of a pseudoaneurysm is a major issue when selecting a treatment course (arterial ligation or resection of the diseased pancreas) and is related to patient outcome. Distal pancreatectomy and splenectomy should be employed to treat bleeding lesions located in the pancreatic tail as these procedures have low morbidity and mortality rates [5
]. Pancreaticoduodenectomy should be limited to select situations and to patients for which less invasive procedures are not technically feasible [5
]. Patient outcome is better for patients with lesions in the pancreatic body and tail (mortality, 16%) than for those with lesions in the pancreatic head (mortality, 43%) [17
]. In this series, bleeding pseudoaneurysms were located at the pancreatic head in 3 patients. In these patients who were treated by a total of 2 surgical procedures and 4 embolizations, the mortality rate was 33.3% (1/3) and the success rate for controlling bleeding was 33.3% (1/3). In contrast, the remaining 6 patients, whose lesions were located in the pancreatic body and tail and were surgically treated, had an 83.3% (5/6) rate of bleeding control and no mortality.
Several factors can increase risk of acute bleeding in chronic pancreatitis: duration of disease; proximity of a vessel to a pseudocyst; communication with the bile or pancreatic duct; and, splenic vein occlusion from thrombosis [18
]. This observation underscores the importance of treating unresolved or progressively enlarged pseudocysts associated with chronic pancreatitis prophylactically to prevent life-threatening bleeding [19
]. Because chronic pancreatitis is an ongoing inflammatory process, operative intervention should be performed as soon as possible [6
]. In this study, pseudocysts occurred either prior to treatment (n = 3) or following intervention (n = 2) in 5 patients; rebleeding pseudoaneurysms developed in 2 of these cases. Drainage procedures were employed in 4 patients, none of whom rebled during the follow-up period (range, 4 – 87 months).