Computed Tomography Enterography (CTE) Versus Capsule Endoscopy for Overt, Obscure Gastrointestinal (GI) Bleeding

Up to 5% of patients with recurrent gastrointestinal (GI) bleeding remain undiagnosed by upper endoscopy and colonoscopy, the presumed source of bleeding in these patients being the small intestine. These patients fall under the category of "obscure gastrointestinal bleeding," and frequently require an extensive diagnostic work-up. Obscure gastrointestinal bleeding (OGIB) refers to bleeding undiagnosed by upper endoscopy and colonoscopy. In 40-70% of cases of OGIB, a bleeding lesion is localizable to the small bowel. In OGIB, capsule endoscopy (CE) has a diagnostic yield of 40-80%, and has demonstrated diagnostic superiority to push enteroscopy, barium studies, angiography, CT angiography, and routine abdominal CT scan. When CE is non-diagnostic, however, the subsequent diagnostic algorithm is not well-defined. There is currently no established role for cross-sectional imaging for this indication. CT enterography (CTE) combines the spatial and temporal resolution of CT with an orally administered neutral enteric contrast material that permits detailed visualization of the small bowel. Unlike other imaging modalities such as nuclear medicine techniques and catheter angiography, CT is less labor-intensive, more readily available, and provides precise anatomic localization. A novel OGIB-protocol available at Brigham and Women's Hospital for CTE utilizes a dual-phase, dual energy technique that obtains images at two time points to better identify active bleeding in the mesentery. We, the investigators, plan to prospectively study an algorithm that employs CTE and compare to capsule endoscopy to investigate the effectiveness of both modalities and to evaluate the potential role of CTE in OGIB. The goal of our study is to determine observationally the contribution of both CE and the new protocol for CTE to the evaluation and management of overt obscure GI bleeding and accordingly revise the clinical algorithm. We hypothesize that CTE will be as or more effective than C