Transcatheter control of intractible gastrointestinal bleeding
Chang, Y.C.; Tsang, Y.M.; Kung, K.L.; Choi, W.M.; Huang, K.M.; Hsu, C.Y.; Wei, T.C.; Wang, T.H.; Chou, Y.H.
Journal of the Formosan Medical Association 90(1): 81-87
1991
ISSN/ISBN: 0929-6646 PMID: 1679114 Document Number: 382255
A series of 23 patients with intractible gastrointestinal (GI) bleeding were managed by the transcatheter method. The series included 5 with hemobilia, 8 with upper GI (UGI) bleeding, 5 with lower GI (LGI) bleeding and 5 with variceal bleeding. The etiology of the hemobilia was surgery, or percutaneous transhepatic cholangiography and drainage (PTCD) complicated by various degrees of biliary tract infection. The causes of UGI bleeding included erosive gastritis, gastric and duodenal ulcers, and traumatic duodenal laceration. All 5 LGI bleedings were due to ischemic colitis and all 5 variceal bleedings were due to hyperdynamic portal hypertension from arterio-portal (A-P) shunting for hepatocellular carcinomas (HCC). Intra-arterial vasopressin infusion was performed on 17 (4, hemobilia; 8, UGI; and 5 LGI bleeding) of these 23 cases as initial management. The success rate for vasopressin in hemobilia, UGI and LGI bleeding was 75% (3/4), 38% (3/8), and 60% (3/5), respectively. The overall initial success rate of vasopressin was 52% (9/17). The relatively poor success rate of vasopressin infusion for UGI bleeding was due to ulcers and laceration. The incidence of rebleeding for vasopressin infusion was 22% (2/9) including one case each of UGI and LGI bleeding. Three patients (1 hemobilia and 2 UGI bleeding) among these 17 cases underwent transarterial embolization (TAE) after failure of intra-arterial vasopressin infusion. One of these 23 cases with hemobilia underwent TAE for initial transcatheter control of the GI bleeding. Five cases of active esophageal variceal bleeding, caused by A-P shunting in HCC, were all successfully controlled by TAE. The initial success rate TAE was 89% (8/9) including 2 cases of hemobilia (2/2), 1 case of UGI bleeding (1/2) and 5 cases of variceal bleeding (5/5). The rebleeding rate for TAE was 12.5% (1/8), which occurred in one case of hemobilia. Of the 5 patients with hemobilia, 2 patients received TAE and 3 patients received vasopressin infusion. The bleeding was controlled successfully except for 2 patients, one had underlying intrahepatic stones and the other had a common bile duct tumor, who received TAE but succumbed due to hepatic failure. After angiographic demonstration of the cause of the bleeding, transcatheter control of acute GI bleeding would be a rational choice, especially when conservative methods have failed or surgical intervention is too risky for the patient.