Objective: Previous studies show conflicting results in the diagnostic yield of oesophagogastroduodenoscopy (OGD) and colonoscopy (bi-directional) in identifying potential bleeding sources (PBS) in patients investigated for occult gastrointestinal bleeding (OGIB). The aims of this study were to evaluate the diagnostic yield of bi-directional endoscopy in patients presenting with OGIB and to assess the factors predictive of a positive yield.
Material and methods: Patients with OGIB referred to the gastroenterology unit were prospectively included in the study. Colonoscopy was immediately followed by OGD. Predetermined criteria for the diagnosis of a PBS were used. Potential clinical factors predictive of positive yield were assessed.
Results: Of the 219 patients (mean age 65 years, 34% M), 110 (50%) had at least one PBS. Colonoscopy revealed 87 PBS in 73 patients (33%), the most common being adenoma. OGD detected 49 PBS in 48 patients (22%), gastric ulcer being the most common. Of the patients with PBS at OGD, 23% also had a PBS at colonoscopy, including 3 patients with colonic cancers. Patients presenting with either a positive faecal occult blood test (FOBT) or iron-deficiency anaemia (IDA) only had a significantly lower yield for PBS than patients with combined positive FOBT and IDA. The percentages of patients with a PBS increased substantially with age.
Conclusions: A PBS was found in only 50% of the patients. Colonoscopy had a slightly higher diagnostic yield, and as expected, resulted in a significantly higher cancer detection rate than OGD. In older patients, colonoscopy should be done irrespective of the findings at OGD. Gastrointestinal-specific symptoms and the use of ASA/NSAIDs were not predictive in finding or localizing PBS.