BACKGROUND: Location of bleeding can present a diagnostic challenge in patients without hematemesis more so than those with hematemesis.

AIM: To describe endoscopic diagnostic yields in both hematemesis and non-hematemesis gastrointestinal bleeding patient populations.

METHODS: A retrospective analysis on a cohort of 343 consecutively identified gastrointestinal bleeding patients admitted to a tertiary care center emergency department with hematemesis and non-hematemesis over a 12-month period. Data obtained included presenting symptoms, diagnostic lesions, procedure types with diagnostic yields, and hours to diagnosis.

RESULTS: The hematemesis group (n = 105) took on average 15.6 h to reach a diagnosis versus 30.0 h in the non-hematemesis group (n = 231), (p = 0.005). In the non-hematemesis group, the first procedure was diagnostic only 53% of the time versus 71% in the hematemesis group (p = 0.02). 25% of patients in the non-hematemesis group required multiple procedures versus 10% in the hematemesis group (p = 0.004). Diagnostic yield for a primary esophagogastroduodenoscopy was 71% for the hematemesis group versus 50% for the non-hematemesis group (p = 0.01). Primary colonoscopies were diagnostic in 54% of patients and 12.5% as a secondary procedure in the non-hematemesis group. A primary video capsule endoscopy yielded a diagnosis in 79% of non-hematemesis patients (n = 14) and had a 70% overall diagnostic rate (n = 33).

CONCLUSION: Non-hematemesis gastrointestinal bleeding patients undergo multiple non-diagnostic tests and have longer times to diagnosis and then compared those with hematemesis. The high yield of video capsule endoscopy in the non-hematemesis group suggests a role for this device in this context and warrants further investigation.