OBJECTIVES: To prospectively compare the global and specific diagnostic yields of push and wireless videocapsule enteroscopy for small bowel lesions in patients with obscure digestive bleeding after esogastroduodenoscopy and colonoscopy. METHODS: The patients studied had unexplained chronic iron-deficient anemia or digestive blood loss after routine investigations. Small bowel investigation was performed first with the wireless-capsule (M2A, Given Imaging) and then with the push-enteroscope (Olympus SIF100). RESULTS: Twenty-one patients were included in the protocol (14 females and 7 males), whose mean age was 60 years (range: 18 to 81). All patients had iron-deficient anemia with occult bleeding (n = 16) or overt bleeding (n = 5). A digestive lesion was observed in 14 of 21 cases (66%). Lesions were: esophageal varices (n = 2), reflux esophagitis (n = 1), upper gastrointestinal tract ulcerations (n = 9), intestinal angioectasia (n = 4), ileal varices (n = 1), cecal angioectasia (n = 1) and tumor-like angioma in the jejunum (n = 1). These 19 lesions were discovered by both methods in 10 cases (52%), by push-enteroscopy only in 6 (31%) and by wireless-capsule endoscopy only in 3 (17%). The global diagnostic yield was therefore slightly but not significantly higher for push wireless-capsule enteroscopy (61 vs 52%; NS) and the specific diagnostic yield was similar (20%). Interobserver agreement on the wireless-capsule recordings reached 85% for detection of findings. CONCLUSIONS: In patients with obscure digestive bleeding, no significant difference in diagnostic yield was evidenced between push and wireless-capsule endoscopy. The main advantage of the latter method versus the former was the detection of distal lesions in the small bowel. Wireless-capsule enteroscopy is mandatory for patients with active unexplained bleeding and negative push-enteroscopy, or for defining the extension of a disease involving, for instance, the presence of angioectasia.