Angiographic embolization of nonvariceal upper gastrointestinal bleeding: Predictors of clinical failure

Background: Angiographic embolization (AE) has emerged as an important therapy for patients with nonvariceal upper gastrointestinal bleeding (UGIB). We hypothesized that discrete factors predictive of AE failure could be identified. Methods: A retrospective review was performed for patients with nonvariceal UGIB who underwent AE from 1999 to 2009 at Penn State Milton S. Hershey Medical Center. AE clinical failure was defined as requirement for another intervention (surgery, endoscopic therapy, or another AE) for nonvariceal UGIB and/or death from bleeding after AE. Statistical analysis was performed using Fisher's exact test and Student's t test to explore the risk of AE failure. Results: Of 48 total AE cases, 17 patients (35.4%) had clinically failed AE. Mortality rate was significantly higher in patients with AE clinical failure than in patients with AE clinical success (64.7% vs. 12.9%, p = 0.001). Factors associated with AE clinical failure include anticoagulant use before admission (p = 0.001), use of corticosteroids before admission (p = 0.045), pre-AE vasopressor use (p = 0.038), and embolization using either coils alone (p = 0.05) or using coils with or without additional embolic materials (p = 0.018). AE clinical failure portends poor prognosis. Caution should be exercised when considering AE, particularly AE using coils, in patients with a history of anticoagulant, corticosteroid, or vasopressor use.