Background: In chronic gastrointestinal bleeding, success rates in the range of 48% to 76% have been reported for diagnosing clear bleeding sources using capsule endoscopy. The influence of patient selection on the numbers of positive findings yielded by capsule endoscopy is as yet unclear. Methods: From April 2001 to June 2003, capsule endoscopy was carried out in 74 of a total of 127 patients (58% ) who presented for capsule endoscopy with a high suspicion of gastrointestinal bleeding in the small- bowel region. Seventy of the 74 patients were included in the analysis. This group of patients was divided into a study group (32 patients) and a post- study group (38 patients), and the two groups were compared. Results: Stricter patient selection was carried out in the study group than in the post- study group: 49% of those in the study group underwent capsule endoscopy, compared with 65% of those in the post- study group. The overall success rate for detecting relevant bleeding sources was 54% ; in the study group, the rate was 66% ; whereas in the post- study group, it was only 45% . Significant differences observed between the two groups of patients included the lowest hemoglobin value (5.9 ± 1.4 g/dL in the study group, compared with 7.7 ± 2.1 g/dL in the post- study group) and transfusion requirements (a median of 10 units in the study group, vs. 2 units in the poststudy group). None of the patients who had a minimum hemoglobin value a 10 g/dL had a positive capsule result. Conclusions: Patient selection has a considerable influence on the success rate of capsule endoscopy. In patients with chronic gastrointestinal bleeding, the minimum hemoglobin value and transfusion requirements appear to be the decisive parameters. Background: In chronic gastrointestinal bleeding, success rates in the range of 48% to 76% have been reported for diagnosing clear bleeding sources using capsule endoscopy. The influence of patient selection on the numbers of positive findings yielded by capsule endoscopy is as yet unclear. Methods: From April 2001 to June 2003, capsule endoscopy was carried out in 74 of a total of 127 patients (58%) who presented for capsule endoscopy with a high suspicion of gastrointestinal bleeding in the small-bowel region. Seventy of the 74 patients This group of patients was divided into a study group (32 patients) and a post-study group (38 patients), and the two groups were compared. Results: Stricter patient selection was carried out in the study group than in the post-study group: 49% of those in the study group underwent capsule endoscopy, compared with 65% of those in the post- study group. The overall success rate for detecting relevant bleeding sources was in the study group, the rate was 66%; while in the post-study group, it was only 45%. in the study group, compared with 7.7 ± 2.1 g / dL in the post-study group) and transfusion requirements (a median of 10 units in the study group, vs. 2 units in the poststudy group). None of the patients who had a minimum hemoglobin value a 10 g / dL had a positive capsule result. Conclusions: Patient selection has a significant influence on the success rate of capsule endoscopy. In patients with chronic gastrointestinal bleeding, the minimum hemoglobin value and transfusion requirements appear to be the decisive parameters.