Long-term rebleeding rate and predictive factors of rebleeding after capsule endoscopy in patients with obscure GI bleeding.

ABSTRACT

BACKGROUND AND AIMS: The incidence of rebleeding in obscure GI bleeding (OGIB) remains unclear. This study used capsule endoscopy (CE) to determine the long-term rebleeding rate and predictive factors for rebleeding in patients with OGIB. METHODS: This single-center, observational study enrolled consecutive patients with OGIB who underwent CE as the first small intestinal examination between March 2004 and December 2015 and were followed up through medical records or letters. RESULTS: Three hundred eighty-nine patients were included in the analysis. Survival curve analysis showed that the overall cumulative rebleeding rate in OGIB during the 5 years was 41.7%. Multivariate analysis using the Cox proportional hazards model revealed that overt OGIB (hazard ratio [HR], 2.017; 95% confidence interval [CI], 1.299-3.131; P = .002), anticoagulants (HR, 1.930; 95% CI, 1.093-3.410; P = .023), positive balloon-assisted enteroscopy findings after CE (HR, 2.927; 95% CI, 1.791-4.783; P < .001), and iron supplements without therapeutic intervention (HR, 2.202; 95% CI, 1.386-3.498; P = .001) were associated with rebleeding, whereas a higher minimum hemoglobin level (HR, .902; 95% CI, .834-.975; P = .009) and therapeutic intervention (HR, .288; 95% CI, .145-.570; P < .001) significantly reduced the risk of rebleeding. Among the Charlson Comorbidity Index components, liver cirrhosis was an independent predictor associated with rebleeding in patients with OGIB (HR, 4.362; 95% CI, 2.622-7.259; P < .001) and in patients with negative CE findings (HR, 8.961; 95% CI, 4.424-18.150; P < .001). CONCLUSIONS: Rebleeding is common during the long-term follow-up of patients with OGIB. Careful follow-up is required for patients with liver cirrhosis or previous massive bleeding.