Re-bleeding and its predictors after capsule endoscopy in patients with obscure gastrointestinal bleeding in long-term follow-up

BACKGROUND: Capsule endoscopy (CE) is the preferred diagnostic test of choice in the investigation of obscure gastrointestinal bleeding (OGIB). Although, a conservative strategy is recommended in the short-term, for cases with a negative result from CE, the impact of CE on long-term re-bleeding still remains unclear. Hence, the aim of this study was to determine the long-term re-bleeding rate along with predictors after CE in patients with OGIB. METHODS: We retrospectively reviewed 216 patients with OGIB, whom had received a CE examination, so as to investigate the cause of obscure GI bleeding; between July 2008 and March 2018. The patient’s characteristics, medication use, CE finding, treatments strategy, re-bleeding episodes and follow-up information were collected from the institutional electronic medical chart and CE database. Re-bleeding free survival was evaluated using Kaplan-Meier curves with log rank test, whilst predictors associated with the re-bleeding episodes were analyzed via the use of Cox proportional hazard model. RESULTS: One hundred and thirty-three patients with OGIB, having received CE were enrolled in the analysis. The pool rate of re-bleeding was 26.3% (35/133) during a follow-up duration of 26 months after CE. Patients with positive CE study, without specific treatment, had higher rates of re-bleeding (47.6%) than those with positive study whom received specific treatment (25.7%), and negative study (20.8%) (p = 0.042). Although, the re-bleeding free survival was not significantly different among the groups (log rank test; P = 0.10). Re-bleeding events occurring within 6, 12, and 24 months after CE were 36, 64 and 92%, respectively. The high-frequency re-bleeding etiologies were the small bowel angiodysplasias and abnormal vascular lesions. Furthermore, independent predictors for re-bleeding after CE were patients with cirrhosis (hazard ratio, HR 4.06), incomplete CE visualization (HR 2.97), and a history of previous GI bleeding (HR 2.80). CONCLUSIONS: The likelihood of re-bleeding after CE was higher in patients with positive CE study than those with negative study. Specific treatments, or therapeutic interventions for patients with detectable lesions reduced the probability of re-bleeding episodes in long-term follow-up. Close follow-up for recurrent bleeding is recommeded for at least 2 years after CE.