Comparison of magnetic resonance enteroclysis and capsule endoscopy with balloon-assisted enteroscopy in patients with obscure gastrointestinal bleeding

Background and study aims: New modalities are available for visualization of the small bowel in patients with possible obscure gastrointestinal bleeding (OGIB), but their performance requires further comparison. This study compared the diagnostic yield of magnetic resonance enteroclysis (MRE) and capsule endoscopy in patients with OGIB, using balloon-assisted enteroscopy (BAE) as the reference standard. Patients and methods: Consecutive consenting patients who were referred for evaluation of OGIB were prospectively included. Patients underwent MRE followed by capsule endoscopy and BAE. Patients with high grade stenosis at MRE did not undergo capsule endoscopy. The reference standard was BAE findings in visualized small-bowel segments and expert panel consensus for segments not visualized during BAE. Results: Over a period of 26 months, 38 patients were included (20 female [53 %]; mean age 58 years, range 28-75 years). Four patients (11 %) did not undergo capsule endoscopy due to high grade small-bowel stenosis at MRE (n=3; 8%) or timing issues (n=1; 3%). Capsule endoscopy was non-diagnostic in one patient. The reference standard identified abnormal findings in 20 patients (53 %). MRE had sensitivity, specificity, and positive and negative likelihood ratios of 21%, 100%, infinity, and 0.79, respectively. The corresponding values for capsule endoscopy were 61%, 85%, 4.1, and 0.46.The reference standard and capsule endoscopy did not differ in percent positive findings (P=0.34), but MRE differed significantly from the reference BAE (P<0.001). Capsule endoscopy was superior to MRE for detecting abnormalities (P=0.0015). Conclusion: Capsule endoscopy performed better than MRE in the detection of small-bowel abnormality in patients with OGIB. MREmay be considered as an alternative for the initial examination in patients with clinical suspicion of small-bowel stenosis.

• View at Publisher
• View at Pubmed
• View at Scopus
• View at Web of Science