Small bowel: preliminary comparison of capsule endoscopy with barium study and CT

Abstract

Purpose: To retrospectively compare capsule endoscopic (CE) findings with findings at barium studies or computed tomography (CT) in patients without evidence of small-bowel stricture at barium examination.

Materials and methods: Fifty-two patients underwent CE, and the majority (43 of 52) did so for obscure gastrointestinal bleeding. Forty-two (81%) of 52 patients underwent 36 small-bowel follow-through examinations; four, enteroclysis; and 19, contrast material-enhanced CT of the abdomen and pelvis within 6 months of CE (either before or after CE). Imaging results were retrospectively reviewed and compared with CE, standard endoscopic, and surgical results. Findings of any examinations between CE and imaging that were discrepant were retrospectively reviewed by a radiologist not blinded to CE results. Proportion of positive findings at CE was compared with proportion of positive findings at barium studies and CT in the same patients. Statistical significance was calculated with McNemar chi2 statistic.

Results: Barium examination findings were positive in one (3%) of 40 patients; CE findings were positive in 22 (55%) (P <.001). CT demonstrated small-bowel findings in four (21%) of 19 patients, but CE demonstrated findings in 12 (63%) of 19 patients (P =.02). The most common CE finding, angioectasia (n = 11), was not detected at any imaging study. More ulcers (n = 8) were detected with CE than with barium study (one of eight) and CT (three of six). At CE, three of five surgically confirmed masses (carcinoid, intussusception, lymphangioma) were identified, but two jejunal tumors were not detected in a patient with poor bowel preparation. At barium study, no masses (zero of five) were detected; at CT, one of four masses was detected.

Conclusion: In patients without a small-bowel stricture at barium study, more small-bowel disease was found at CE when findings were retrospectively compared with barium examination and CT findings.