Diagnostic considerations in evaluation of patients presenting with melena and nondiagnostic esophagogastroduodenoscopy.

Proper evaluation of patients with melena and nondiagnostic esophagogastroduodenoscopy is comparatively undefined. We sought to determine the percentage of patients with melena and nondiagnostic upper endoscopy and assess the yield of further evaluation. Of 209 patients presenting with melena, 180 underwent esophagogastroduodenoscopy as the initial study, which was nondiagnostic in 43 cases (24%). Further evaluation was pursued in 30. A presumed source of melena was found in 11 patients (37%), identified by colonoscopy in seven, bleeding scan in three, and barium enema plus flexible sigmoidoscopy in one. Nearly all such defined cases originated from the right colon. Small bowel contrast studies, flexible sigmoidoscopy of barium enema alone, and angiography failed to reveal a source. Our findings suggest that many (24%) patients presenting with melena will have nondiagnostic upper endoscopy; further evaluation reveals a potential source in 37% of this group, with the right colon being the most likely location of pathology; and colonoscopy is the test of choice in this cohort.