OBJECTIVES: A radiological examination of the small bowel is often performed in case of gastrointestinal bleeding of obscure origin. More recently, push-type enteroscopy has been reported as a valuable tool in this indication. The purpose of this study was to compare the diagnosis efficiency of these two procedures. METHODS: From February 1994 to February 1996, 40 patients (mean age: 52 years) with obscure gastrointestinal bleeding (iron-deficiency anemia without obvious cause of blood loss or malabsorption: n = 17; macroscopic gastrointestinal bleeding: n = 23) were examined by small bowel follow-through and push-type enteroscopy (jejunoscopy n = 19; double way examination n = 21). Each patient had negative upper and lower gastrointestinal tract endoscopies prior to small bowel examinations. RESULTS: Small bowel follow-through revealed only one lesion potentially responsible for blood loss (2.5%), corresponding to a jejunal leiomyoma. Push-type enteroscopy detected small bowel lesions potentially responsible for blood loss in 6 patients (15%). The lesions were located in the jejunum in 5 cases (arteriovenous malformations: n = 3; metastasis: n = 1; leiomyoma: n = 1), in the ileum in 1 case (leiomyoma). The efficiency of push-type enteroscopy for the detection of a small bowel lesion was of 22% in case of macroscopic bleeding and of 6% in case of iron-deficiency anemia. Push-type enteroscopy also revealed lesions previously undetected by gastroscopy or colonoscopy in 8 patients (20%). CONCLUSION: Push-type enteroscopy was more effective than small bowel follow-through to detect the origin of obscure gastrointestinal bleeding. Push-type enteroscopy revealed a cause of bleeding in 35% of patients, located in the small bowel in only 15% of the patients.