Arterial embolotherapy for upper gastrointestinal hemorrhage:: Outcome assessment

PURPOSE: To identify predictors of clinical outcome after arterial embolotherapy for upper gastrointestinal (UGI) hemorrhage. MATERIALS AND METHODS: Seventy-five consecutive patients (mean age, 62.5 y) underwent arterial embolization for acute UGI hemorrhage. Bleeding was detected at endoscopy and angiography in 22 patients, at endoscopy alone in 29 patients, and at angiography alone in 24 patients. As such, embolization was directed by angiography in 46 patients (61.3%) and by endoscopy (referred to as "blind" embolization) in 29 patients (38.7%). The embolic agents used were metallic coils, polyvinyl alcohol particles (size range, 355-710 mum), gelatin sponge, and tissue adhesive. Predictors of bleeding recurrence and mortality were analyzed with logistic regression and Cox models, respectively. RESULTS: The technical success rate of embolization was 98.7%. Primary clinical success was achieved in 57 patients (76%). Secondary clinical success occurred in five additional patients (82.5%) after repeat embolization. There were four (5.3%) complications: two cases of self-resolving duodenal ischemia, one hepatic infarct, and one inguinal hematoma. The periprocedural mortality rate was 34.6% (26 of 75), mostly related to underlying illness. Early recurrence of bleeding (within 30 days of embolization) was associated with coagulation disorders (international normalized ratio >1.5, partial thromboplastin time >45 seconds, or platelet count <80,000/<mu>L; odds ratio, 19.46; P = .001) and with the use of coils as the only embolic agent (odds ratio, 7.73; P = .01). Cirrhosis and cancer shortened the overall survival of patients after embolic therapy. The mean patient follow-up time was 34.5 months. CONCLUSION: Arterial embolotherapy for UGI hemorrhage is safe, effective, and durable. Coagulopathy and the use of coils as the only embolic agent were associated with a higher risk of early bleeding recurrence.