Delayed Bleeding Rate According to the Forrest Classification in Second-Look Endoscopy After Endoscopic Submucosal Dissection

Forrest classification is a valid tool to predict rebleeding rate in peptic ulcer, not in post-endoscopic resection ulcer. We evaluated the delayed bleeding rate in Forrest classification II and III lesions when they were not treated in second-look endoscopy. Between July 2011 and February 2012, 706 lesions in 656 consecutive patients who underwent second-look endoscopy performed on the second day after endoscopic submucosal dissection (ESD) were prospectively investigated. Endoscopic findings were described according to Forrest classification, and late delayed bleeding was defined as bleeding from second-look endoscopy to 1 month. We evaluated the rate of late delayed bleeding in untreated Forrest classification II and III lesions during second-look endoscopy. Among the 706 gastric tumors analyzed, late delayed bleeding after ESD occurred in 29 lesions (4.1 %). At second-look endoscopy, Forrest I lesions (immediately treated by endoscopic hemostasis) were found in 63 cases [Ia, 8 lesions (1.1 %); Ib, 55 lesions (7.8 %)]; there was no further bleeding after discharge. Forrest II and III lesions (not treated in second-look endoscopy) were found in 643 cases [IIa, 62 lesions (8.8 %); IIb, 119 lesions (16.9 %); IIc, 460 lesions (65.2 %); III, 2 lesions (0.3 %)]; and there was no significant difference in the late delayed bleeding rate between these groups [IIa, 2/62 (3.2 %); IIb, 5/119 (4.2 %); IIc and III, 22/462 (4.8 %); P = 1.000]. The rate of late delayed bleeding of post-ESD ulcers with non-bleeding visible vessels was not significantly different from that of ulcers with non-visible vessels (http://cris.nih.go.kr, identifier KCT0000268).