Arterial Embolotherapy for Endoscopically Unmanageable Acute Gastroduodenal Hemorrhage: Predictors of Early Rebleeding

Background & Aims

Severe bleeding from gastrointestinal ulcers is a life-threatening event that is difficult to manage when endoscopic treatment fails. Transcatheter embolization has been proposed but factors that influence the angiographic outcome are not well documented. We aimed to identify predictors of recurrent bleeding within 30 days after transcatheter embolization for refractory hemorrhage from gastroduodenal ulcers.

Methods

This retrospective single-center study of 60 consecutive emergency embolization procedures included hemodynamically unstable patients (41 men, 19 women; mean age, 69.4 ± 15 y), referred from 1999 to 2008 for selective angiography after failed endoscopic treatment. Predictors of early rebleeding were tested with univariate analysis and a multivariate logistic regression model.

Results

The procedural success rate was 95%, the primary clinical success rate was 71.9% (41 of 57), and secondary clinical success was achieved in 3 patients (77.2%) after repeat embolization. No major catheterization-related complications occurred. Periprocedural mortality was 26.7% (16 of 60). Early bleeding recurrence was associated with coagulation disorders (P = .007), longer time to angiography (P = .0005), greater preprocedural blood transfusion volume (P = .0009), 2 or more comorbidities (P = .005), and use of only coils (P = .003). Two factors were independent predictors of embolization failure: coagulation disorders (odds ratio, 6.18; P = .027) and the use of coils as the only embolic agent (odds ratio, 6.24; P = .022). The median follow-up time was 7 months (range, 1 day to 103 months).

Conclusions

Angiographic embolization should be performed early in the course of bleeding, and not with coils alone, in critically ill patients. It is important to correct coagulation disorders throughout the embolization procedure.

Abbreviations used in this paper: GDA, gastroduodenal artery, OR, odds ratio