Introduction Prolonged overt obscure gastrointestinal bleeding (OGIB) after an initial normal oesophagogastroduodenoscopy and colonoscopy can be difficult to manage. “Real-world” studies with all of the endoscopic (capsule endoscopy, device-assisted enteroscopy), radiological (radionuclide red cell scan, CT angiography and angiographic embolisation) and surgical interventions or therapies are lacking.

Methods We studied the investigation and treatment of such patients, requiring transfusion with ≥1 inpatient stay of 7 days between 2004 and 2012 at St. Vincent’s Hospital and Epworth Eastern Hospital, Melbourne, Australia.

Results Twenty-eight patients presented at a median age of 67.5 years. The median blood transfusion requirement per patient from symptom presentation to diagnosis or census was 26 units. Anti-platelet and anticoagulation therapy was taken by 50% patients. Twenty-four had diagnoses made (21 small and 3 large intestinal). These included angioectasias in 8 patients (6 small and 2 large bowel) who were >65 years and six of whom were taking anti-platelet therapy for cardiac disease; portal hypertensive enteropathy/ small bowel varices in four patients who were <60 years; and small intestinal tumours in 5 patients (2 gastrointestinal stromal tumours and 3 carcinoid tumours), the latter of which needed surgery for diagnosis and treatment in all cases. Repeat gastroscopy allowed histoacryl glue injection of peri-anastomotic varices in one case and repeat colonoscopy permitted treatment of angioectasias in two elderly patients. Radionuclide red cell scans had the highest radiological diagnostic yield (51%) but were beneficial only in conjunction with other tests. CT angiography (diagnostic yield 30%) resulted in successful angiographic embolisation in 3/9 cases (a small intestinal angioectasia and bleeding associated with colonic diverticula and a pancreaticoduodenal artery pseudoaneurysm). Capsule endoscopy had the highest endoscopic diagnostic yield (53%). In two patients repeat examination was diagnostic (an angioectasia and a gastrointestinal stromal tumour). Antegrade double balloon enteroscopy had the best enteroscopic yield (31%). In 2 cases it allowed argon plasma coagulation of small intestinal angioectasias, which were missed by prior enteroscopy. Surgery had a diagnostic and therapeutic yield of 60%.

Conclusion Prolonged overt OGIB is difficult to manage. There may be clues to the underlying diagnosis from the history and clinical features. Capsule endoscopy is a good first-line test, which can guide enteroscopy. Similarly CT angiography can guide angiographic embolisation. Surgery is best as a last resort but is not always productive. Management should be individualised with consideration given to repeating investigations.

Disclosure of Interest None Declared.