Transcatheter Arterial Embolization for Upper Gastrointestinal Nonvariceal Hemorrhage: Is Empiric Embolization Warranted?

Abstract

To determine whether transcatheter arterial embolization performed in the setting of active gastric or duodenal nonvariceal hemorrhage is efficacious when the bleeding source cannot be identified angiographically. Records of 115 adult patients who underwent visceral angiography for endoscopically documented gastric (50 patients) or duodenal (65 patients) nonvariceal hemorrhage were retrospectively reviewed. Patients were subdivided into three groups according to whether angiographic evidence of arterial hemorrhage was present and whether embolization was performed (group 1 = no abnormality, no embolization; group 2 = no abnormality, embolization performed [empiric embolization]; and group 3 = abnormality present, embolization performed). Thirty-day rates and duration of primary hemostasis and survival were compared. For patients with gastric sources of hemorrhage, the rate of primary hemostasis at 30 days after embolization was greater when embolization was performed in the setting of a documented angiographic abnormality than when empiric embolization was performed (67% vs. 42%). The rate of primary hemostasis at 30 days after angiography was greater for patients with duodenal bleeding who either underwent empiric embolization (60%) or embolization in the setting of angiographically documented arterial hemorrhage (58%) compared with patients who only underwent diagnostic angiogram (33%). Patients with duodenal hemorrhage who underwent embolization were less likely to require additional invasive procedures to control rebleeding (p = 0.006). Empiric arterial embolization may be advantageous in patients with a duodenal source of hemorrhage but not in patients with gastric hemorrhage.