Favorable Long-term Outcomes of Repeat Endotherapy for Small-intestine Vascular Lesions by Double-balloon Endoscopy

Background: GI bleeding secondary to small-intestine vascular lesions (SIVLs) is associated with rebleeding despite initial hemostasis by endotherapy applied at double-balloon endoscopy (DBE).

Objective: To evaluate the long-term outcomes of DBE endotherapy of SIVL (as described by the Yano-Yamamoto classification). The impact of repeat DBE endotherapy for recurrent bleeding also was assessed.

Design: Retrospective cohort study. The mean (± standard deviation [SD]) follow-up period was 4.9 (± 1.7) years (range 2.4-9.1 years).

Setting: Tertiary-care referral center.

Patients: A total of 43 patients, who underwent 69 sessions of DBE endotherapy of SIVLs.

Intervention: DBE endotherapy.

Main Outcome Measurements: Frequency of overt rebleeding after the initial DBE endotherapy.

Results: Overt rebleeding occurred in 16 of 43 patients (37%). Patients with multiple SIVLs showed a significantly higher rate of overt rebleeding than did those with a solitary SIVL (12/23 [52%] vs 4/20 [20%]; P = .017). The trend toward frequency of rebleeding after the first DBE hemostasis appeared to be higher for patients with type 1a SIVLs than for those with type 1b or type 2 lesions: type 1a (8/16, 50%) versus type 1b (5/19, 26%) (P = .12) and type 1a (8/16, 50%) versus type 2 (2/7, 29%) (P = .31), respectively. In 12 of 16 patients (75%) who underwent repeat DBE endotherapy at each episode of overt rebleeding (median 3 times, range 2-6), the frequency of rebleeding decreased significantly after the first year of follow-up, as compared with the remaining 4 patients who did not undergo repeat DBE; mean (± SD) 0.12 (± 0.19) versus 0.52 (± 0.33) times per year per patient (P = .006).

Limitations: Single-center, retrospective study.

Conclusion: Although the presence of multiple SIVLs was associated with rebleeding, repeat DBE endotherapy resulted in an improved long-term outcome in patients with refractory SIVL bleeding.