Diagnostic Yield, Clinical Predictors and Therapeutic Impact of Capsule Endoscopy in a Large Cohort of Patients with Obscure GI Bleeding

Abstract

Background: Identifying a source of bleeding in obscure GI blood loss remains a challenge. Video capsule endoscopy is frequently used as an adjunct diagnostic modality. However, there is limited published data on the clinical benefit of this technology in such patients. Aims: 1) Estimate the diagnostic yield of capsule endoscopy. 2) Identify clinical predictors for a positive finding. 3) Estimate frequency of change in treatment following a positive finding in patients with obscure GI bleeding. Methods: Clinical, demographic and procedural data were abstracted retrospectively from patients with obscure GI bleeding who underwent capsule endoscopy. Logistic regression was used to identify clinical predictors for a positive test. Results: Sixty-five (32 females and 33 males) patients with a mean age of 67 years were included in the study. Median duration of bleeding was 24 months. 50 patients presented with overt bleeding, predominantly melena, while the remaining 15 had occult blood loss. 55% of patients had significant co-morbidities, 25% had NSAID use and 40% aspirin intake. 43% of patients had been hospitalized for the bleeding episode and 91% had blood transfusions. 43%, 65% and 20% of patients had more than one EGD, colonoscopy and extended EGD respectively. Prior to capsule endoscopy 55% had a small bowel x-ray, 20% angiography, 31% technetium labeled red cell scan and 9% intra-operative endoscopy. Transit of capsule into the colon was reported as complete in 41 patients (63%), indeterminate in 7 (11%) and incomplete in 17 (26%). The median transit time through the small bowel was 250 minutes (range 328-80). 34 (52%) patients had a positive finding, the more common findings were vascular ectasia (14), and erosive or ulcerative disease (12). Blood was seen in 13 (20%) patients of which 9 (14%) were associated with the lesion. Following the capsule study, only 13 (20%) of all patients had therapy based on a positive finding (endoscopic therapy (7), surgery (3), change in medication (1), referred for surgery (2)). Using logistic regression age, gender, co-morbidities, NSAID and aspirin use, duration of bleeding, overt blood loss and hospitalization were not significant predictors of a positive capsule study (p>0.05). Conclusions: 1. Video capsule endoscopy in obscure GI bleeding appears to have an acceptable presumed diagnostic yield. 2. No clinical factors were predictive of a positive exam. 3. Its impact on subsequent therapy remains low.