Background and study aims: In patients with obscure digestive bleeding, the bleeding source is frequently located in the small bowel. Capsule endoscopy (CE) is an effective method of investigating the whole small bowel in such patients. In the present study, a diagnostic approach was tested in patients with obscure digestive bleeding in which CE was included as the initial examination of the small bowel when the esophagogastroduodenoscopy (EGD) and colonoscopy findings were normal.
Patients and methods: Patients admitted between October 2000 and February 2002 for obscure digestive bleeding underwent CE as the initial intestinal investigation, and the further management was decided on the basis of the results. After 12 months, follow-up data were obtained from all patients and referring physicians. The positive predictive value was calculated as the percentage of patients in whom CE detected a relevant lesion, and the negative predictive value as the percentage of patients with normal CE in whom no intestinal lesion was detected during the follow-up period.
Results: Forty-four patients (21 men, 23 women, aged 63 +/- 17 y) were included in the study. Twenty-two had overt bleeding and 22 had occult bleeding. CE detected an intestinal lesion in 18 patients (41.9 %)-- nine with angiomas, five with ulcers, one with a tumor, two with portal hypertension, and one with ischemic ileitis. The findings were normal in 17 patients (39.5 %). CE detected upper gastrointestinal lesions missed at EGD in four patients and blood in the stomach in two patients or in the proximal colon in three, leading to new endoscopies. Intestinal lesions detected at CE were treated as follows: push enteroscopy with treatment in eight patients, surgery in four, and medical treatment in six. In eight patients who underwent push enteroscopy, the procedure did not reveal lesions missed by CE. After 1 year, 15 of the 18 patients treated for intestinal lesions had no further bleeding and no anemia; one died after surgery (for resection of an ischemic intestinal loop), one relapsed, and one was diagnosed with a different intestinal condition. In all patients with extraintestinal lesions or blood detected at CE, further endoscopies led to diagnosis and therapy, with a favorable outcome. In patients with normal CE, no intestinal lesion was detected, but an extraintestinal source of bleeding was diagnosed and treated in nine patients (in the upper gastrointestinal tract in five cases and in the colon in four). Three patients had anemia of hematological origin and four had inadequate iron intake. The positive predictive value of CE was 94.4 % in patients with intestinal lesions, and the negative predictive value was 100 % in patients with normal CE findings.
Conclusions: In patients with obscure digestive bleeding, CE positively predicted the intestinal diagnosis or normal status in 95.5 % of cases. A diagnostic approach to obscure digestive bleeding that includes CE after the initial endoscopic work-up thus appears to be a valid strategy for small-bowel examinations.