Frequency and risk factors for rebleeding events in patients with small bowel angioectasia

Small bowel angioectasia is reported as the most common cause of bleeding in patients with obscure gastrointestinal bleeding. Although the safety and efficacy of endoscopic treatment have been demonstrated, rebleeding rates are relatively high. To establish therapeutic and follow-up guidelines, we investigated the long-term outcomes and clinical predictors of rebleeding in patients with small bowel angioectasia.

A total of 68 patients were retrospectively included in this study. All the patients had undergone CE examination, and subsequent control of bleeding, where needed, was accomplished by endoscopic argon plasma coagulation. Based on the follow-up data, the rebleeding rate was compared between patients who had/had not undergone endoscopic treatment. Multivariate analysis was performed using Cox proportional hazard regression model to identify the predictors of rebleeding. We defined the OGIB as controlled if there was no further overt bleeding within 6 months and the hemoglobin level had not fallen below 10 g/dl by the time of the final examination.

The overall rebleeding rate over a median follow-up duration of 30.5 months (interquartile range 16.5-47.0) was 33.8% (23/68 cases). The cumulative risk of rebleeding tended to be lower in the patients who had undergone endoscopic treatment than in those who had not undergone endoscopic treatment, however, the difference did not reach statistical significance (P = 0.14). In the majority of patients with rebleeding (18/23, 78.3%), the bleeding was controlled by the end of the follow-up period. Multiple regression analysis identified presence of multiple lesions (≥3) (OR 3.82; 95% CI 1.30-11.3, P = 0.02) as the only significant independent predictor of rebleeding.

In most cases, bleeding can be controlled by repeated endoscopic treatment. Careful follow-up is needed for patients with multiple lesions, presence of which is considered as a significant risk factor for rebleeding.