Role of Transcatheter Arterial Embolization in Acute Refractory Non-variceal Upper Gastrointestinal Bleeding Not Controlled by Endoscopy: A Single-Center Experience and a Literature Review

Abstract

Introduction Acute upper gastrointestinal bleeding (UGIB) is a medical emergency and a common cause of hospital admissions worldwide. It has traditionally been treated with resuscitation and endoscopic intervention as the first-line therapy. In this study, we assessed the adjunctive role of transcatheter arterial embolization (TAE) in patients with uncontrolled UGIB after an endoscopic intervention. Material and methods A retrospective chart review of patients requiring TAE of UGIB which was not controlled by endoscopic intervention in BronxCare Health System from 2018 to 2021 was done. Patients who were more than 18 years of age and required TAE during the time period of the study were included in the study. Patients' charts were reviewed for patients' demographics, comorbidities, hospital course, imaging findings, esophagogastroduodenoscopy findings and intervention, and interventional radiology intervention and clinical outcome. Results A total of 10 patients were included in the study. A majority of the patients were male. Transcatheter atrial embolization was successful in all the 10 patients. Coils were used in seven patients while particulate polyvinyl alcohol 500 micron particle was used in two patients and vascular plug was used in two patients. Out of the 10 patients, four expired during the hospital course. None of the patients died secondary to UGIB. Three of the patients expired due to severe sepsis with septic shock secondary to pneumonia while one patient died because of respiratory failure due to lung collapse secondary to endobronchial lesion. Conclusion Refractory acute UGIB is associated with significant morbidity and mortality. TAE is a minimally invasive measure that should be considered early in the treatment of UGIB which is refractory to conventional endoscopic management. Our case highlights the importance of TAE in a patient with refractory UGIB after endoscopic intervention.

Keywords: anemia; endovascular embolization; interventional radiology guided embolization; morbidity; mortality; refractory upper gastrointestinal bleeding; upper endoscopy.

Figure 1. Esophagogastroduodenoscopy showing actively oozing duodenal ulcer in the duodenal bulb.

Figure 2. A: CT angiogram axial image of the abdomen demonstrated irregularly dilated and exposed mucosal duodenal arterioles corresponding to actively oozing crated duodenal ulcer in the duodenal bulb seen on prior endoscopy. B: Sequential selective vascular catheterization of celiac and GDA angiograms demonstrated duodenal bleeding territory with many irregularly dilated arteriolar feeders from branches of the GDA with focal blush as seen on delayed venous phase (image C). D: Post-embolization celiac angiogram demonstrated no evidence of active bleeding signs, and complete resolution of the above findings after 13 coils' deployment in “sandwiching” appropriate positions, beyond the takeoff of main GDA, with adequate hepatic arterial inflow at the end of the embolization.

CT: computed tomography; GDA: gastroduodenal artery.

Figure 3. A: Esophagogastroduodenoscopy showing spurting duodenal ulcer. B: Repeat esophagogastroduodenoscopy showing oozing duodenal ulcer.

Figure 4. A: CT angiogram of the abdomen demonstrated hyperdense materials/blood products within the second portion of the duodenum adjacent to multiple endoscopic clips. B: Super-selective GDA angiograms using microcatheter systems demonstrated duodenal culprit bleeding with many irregular tortuous arteriolar feeders from superior anterior duodenal arterial branches of the GDA, uncontrolled by adjacent clips placed earlier by endoscopy. C: Post-embolization selective common hepatic angiograms demonstrated no evidence of active bleeding signs, and complete resolution of the above findings after five coils' deployment in “sandwiching” appropriate positions, beyond the takeoff of main GDA and through the proximal aspect of right gastroepiploic artery, with adequate hepatic arterial inflow.

CT: computed tomography; GDA: gastroduodenal artery.

Figure 5. Esophagogastroduodenoscopy showing large amount of bleeding in the second part of the duodenum.

Figure 6. A: CT coronal image demonstrated duodenal fungating mass. B: Super-selective GDA angiograms using microcatheter System demonstrated many pseudoaneurysmal arteriolar tumoral feeders from the GDA branches. C: Post-embolization vascular catheterization for celiac angiogram demonstrated no evidence of active bleeding signs, and complete resolution of the above findings after coils' deployment in “sandwiching” appropriate positions, beyond the takeoff of main GDA, with adequate hepatic arterial inflow at the end of the embolization.

CT: computed tomography; GDA: gastroduodenal artery.

Figure 7. Small bowel enteroscopy showing diffuse mucosal oozing in the second portion of the duodenum.

Figure 8. A: CT angiogram abdomen with contrast showed hyperdense material seen on the portal venous phase within the second portion of the duodenum, not present on the noncontrast phase of the CT as seen in image B. C: Super-selective GDA angiograms using microcatheter system demonstrated duodenal culprit bleeding from lateral superior branch of the GDA. D: Post-embolization selective and super-selective angiograms demonstrated no evidence of active bleeding signs, and complete resolution of the above findings with four deployed coils in appropriate positions, just beyond the takeoff of an accessory right hepatic artery coming off as anatomical variant from the GDA.

CT: computed tomography; GDA: gastroduodenal artery.

Figure 9. A and B: Esophagogastroduodenoscopy showing nonbleeding cratered gastric ulcer with a visible artery in gastric body and antrum. C: Showing large amount of blood in the gastric fundus and antrum.

Figure 10. A: CT demonstrated duodenal outpouching filled with oral contrast correlating with duodenal ulceration seen on endoscopy. B: Super-selective GDA arterial and venous angiograms using microcatheter System demonstrated many pseudoaneurysmal duodenal culprit bleeding territories from IPDA coming off the GDA. C: Post-embolization vascular catheterization for celiac angiogram demonstrated no evidence of active bleeding signs, and complete resolution of the above findings after three coils' deployment along the IPDA in appropriate positions, beyond the takeoff of gastroepiploic artery, with adequate hepatic arterial inflow at the end of the embolization.

CT: computed tomography; GDA: gastroduodenal artery; IPDA: inferior pancreaticoduodenal artery.

Figure 11. Esophagogastroduodenoscopy showing oozing duodenal ulcer.

Figure 12. A: CT angiogram of the abdomen showed hyperdense material seen within the second portion of the duodenum compatible with oozing blood seen on endoscopy. B: Selective celiac angiograms demonstrated irregularly tortuous duodenal arterioles as the culprit of bleeding from branch of the GDA with hypertrophic IPDA shunting toward the SMA, therefore no coils were used to avoid non-target migration and bowel ischemia. C: Super-selective GDA angiograms using microcatheter system for embolization using particulate polyvinyl alcohol (PVA) 500 micron particle with full stasis of flow. D: Post-embolization selective SMA angiogram demonstrated patent jejunal, IPDA and ileocecal and colic arteries with no signs of backflow duodenal bleeding.

CT: computed tomography; IPDA: inferior pancreaticoduodenal artery; SMA: superior mesenteric artery; GDA: gastroduodenal artery.

Figure 13. Esophagogastroduodenoscopy showing oozing duodenal ulcer in the duodenal bulb.

Figure 14. A: CT of the abdomen demonstrated dense-fluid-filled duodenal wall thickening and ulceration crater as seen on endoscopy. B: Super-selective GDA angiograms demonstrated many focal bleeding blush adjacent to endoscopic clip. C: Post-embolization celiac angiogram demonstrated no evidence of active bleeding signs, and complete resolution of the above findings after deployment of one coil and two vascular plugs in appropriate positions along the GDA.

CT: computed tomography; GDA: gastroduodenal artery.

Figure 15. Esophagogastroduodenoscopy showing actively bleeding ulcer in the second part of the duodenum.

Figure 16. A: CT scan with contrast of the abdomen showed hypertrophic submucosal arterioles along the second portion of the duodenum compatible with oozing seen on endoscopy. B: Selective celiac angiograms demonstrated irregularly tortuous duodenal arterioles as the culprit of bleeding from branch of the GDA with shunting toward the SMA, therefore no coils were used at this time to avoid non-target migration and bowel ischemia. C: Super-selective GDA angiograms with embolization using particulate polyvinyl alcohol (PVA) 500 was done. Post-embolization selective celiac and SMA angiograms demonstrated patent jejunal and ileocecal and colic arteries with replaced right hepatic artery with no signs of backflow duodenal blushes or bleeding.

CT: computed tomography; GDA: gastroduodenal artery; SMA: superior mesenteric artery.

Figure 17. Esophagogastroduodenoscopy showing large friable necrotic malignant-appearing infiltrative mass in the second part of the duodenum.

Figure 18. A: CT axial arterial phase demonstrated infiltrative mass with active bleeding in the second part of the duodenum with layering of contrast in the dependent aspect of the duodenum on venous phase as seen on image B. C: Super-selective GDA angiograms using microcatheter system demonstrated active duodenal bleeding from IPDA coming off the GDA with contrast flowing freely and actively from the GDA branches to the duodenum, as seen on CT scan. D: Post-embolization celiac angiogram demonstrated no evidence of active bleeding signs, and complete resolution of the above findings after coils deployment along IPDA in “sandwiching” appropriate positions, beyond the takeoff of gastroepiploic artery, with adequate hepatic arterial flow.

CT: computed tomography; IPDA: inferior pancreaticoduodenal artery; GDA: gastroduodenal artery.

Figure 19. Esophagogastroduodenoscopy showing one bleeding cratered duodenal ulcer with a visible vessel in the duodenal bulb.

Figure 20. A: CT demonstrated hypertrophic submucosal arterioles along the duodenum compatible with oozing ulceration as seen on endoscopy. B: SMA angiogram demonstrating large active and delayed portal venous phase (image C). D: Post-embolization celiac angiogram demonstrated no evidence of active bleeding signs, and complete resolution of the above findings after deployment of two coils and one vascular plug in appropriate positions along the GDA.

CT: computed tomography; SMA: superior mesenteric artery; GDA: gastroduodenal artery.