To assess the diagnostic efficiency of capsule endoscopy in a large group of patients with different indications, to weigh the reliability of the procedure for excluding small bowel lesions, and to identify factors associated with the likelihood of obtaining a definitive diagnosis.

Three hundred four consecutive patients (141 female, mean age 55 years, range 12-91 years) underwent capsule endoscopy in two different Gastroenterology Units, for a total of 314 procedures, and were followed-up for a median period of 15 months. Referrals were obscure occult/overt gastrointestinal bleeding (203 patients), suspected small bowel disease (74), gastrointestinal polyposis (18), suspected/previous intestinal or endocrine malignancies (13), previously diagnosed intestinal lymphangectasia (3), and vascular abnormalities (3).

Adequate visualization of the small bowel was obtained in 96% of patients, although the capsule did not visualize cecum in 20% of cases. Non-natural excretion of the capsule was observed in 4 patients, all of whom underwent laparotomy for intestinal stenosis. Diagnostic yields were 58% for obscure gastrointestinal bleeding and 31% for patients with suspected small bowel disease. Capsule endoscopy was able to rule out small bowel disease in 14% of patients, and a definitive diagnosis was achieved in 65% of patients. The only parameter associated with the likelihood of reaching a conclusive diagnosis was the indication to the procedure (overall chi-square 13.5, P = .004).

Capsule endoscopy represents a reliable tool for verifying the state of the small bowel. Accurate selection of indications and critical evaluation of the results are essential to fully exploit this procedure.