To assess the efficacy and safety of arterial embolotherapy in patients with massive duodenal hemorrhage.

Between January 1999 and June 2002, 25 patients (age: 34-81, mean 58, male: 19, female: 6) underwent arterial embolization for duodenal hemorrhage after failed endoscopic therapy. The hemorrhage originated from duodenal ulcer in sixteen patients, from cancer with duodenal invasion in five patients, from endoscopic sphincterectomy in two patients, and from pseudoaneurysm complicating acute pancreatitis in two patients. Hemorrhage was detected at endoscopy and an attempt was made to treat it endoscopically in all patients, but failed in each case. At angiography, direct bleeding signs such as contrast extravasation or pseudoaneurysm were demonstrated in nineteen patients. In the six patients without angiographic evidence of bleeding, blind embolization of the gastroduodenal artery was performed based on the endoscopic examination. Microcoil and gelfoam particles were used as embolic agents.

Hemostasis was achieved immediately after embolotherapy in 21 patients (84%). Bleeding recurred in 4 patients (16%), and of these cases, one was successfully treated purely by endoscopic means, a second was reembolized three times due to bleeding from the collateral vessels of the tumor and the two others were treated by surgery. After the procedure, six patients died (24%). The causes of death were disseminated intravascular coagulopathy, multiorgan failure, sepsis and acute renal failure. The underlying diseases of the deceased patients were cancers with duodenal invasion (n=4) and abdominal aortic aneurysm with ischemic colitis (n=1).

Transarterial embolotherapy in the case of massive duodenal hemorrhage is a safe and effective procedure. Even in the absence of angiographic evidence of bleeding, blind embolization of the gastroduodenal artery is effective for patients in the surgically high risk group.