Background The effectiveness of prophylactic clip closure in reducing the risk of delayed bleeding after colorectal endoscopic submucosal dissection (ESD) is uncertain among patients on anticoagulants. We therefore aimed to assess this effectiveness using data from a large multicenter study. Methods We used the ABCD-J study database to analyze delayed bleeding among 34 455 colorectal ESD cases from 47 Japanese institutions. Delayed bleeding rates among the no/partial and complete closure groups were compared in patients on direct oral anticoagulants (DOACs) or warfarin. Propensity score matching was used for baseline characteristics to reduce the effects of selection bias. Results Overall, data from 1478 patients on anticoagulants who underwent colorectal ESD were examined. After propensity score matching, the complete and no/partial closure groups were compared in 212 patients on DOACs and 82 on warfarin. The complete closure group showed a significantly lower delayed bleeding rate in patients receiving DOACs (10.8 % vs. 5.2 %, absolute risk reduction [ARR] 5.7 %, P = 0.048) and warfarin (17.1 % vs. 6.1 %, ARR 11.0 %, P = 0.049). Additionally, complete closure significantly reduced the risk of delayed bleeding among patients taking DOACs for right-sided lesions (ARR 6.7 %, P = 0.04), whereas no risk reduction was observed for left-sided (P > 0.99) or rectal (P = 0.50) lesions. A similar trend was observed among patients on warfarin. Conclusions Prophylactic complete clip closure after colorectal ESD significantly reduced the delayed bleeding rate in patients receiving DOACs or warfarin. It should be performed after ESD, particularly for right-sided lesions.

Objective: A newly launched endoscopy system (EVIS X1, CV‐1500; Olympus) is equipped with texture and color enhancement imaging (TXI). We aimed to investigate the efficacy of TXI for the visibility and diagnostic accuracy of non‐polypoid colorectal lesions.Methods: We examined 100 non‐polypoid lesions in 42 patients from the same position, angle, and distance of the view in three modes: white light imaging (WLI), narrow‐band imaging (NBI), and TXI. The primary outcome was to compare polyp visibility in the three modes using subjective polyp visibility score and objective color difference values. The secondary outcome was to compare the diagnostic accuracy without magnification.Results: Overall, the visibility score of TXI was significantly higher than that of WLI (3.7 ± 1.1 vs. 3.6 ± 1.1; p = 0.008) and lower than that of NBI (3.7 ± 1.1 vs. 3.8 ± 1.1; p = 0.013). Color difference values of TXI were higher than those of WLI (11.5 ± 6.9 vs. 9.1 ± 5.4; p < 0.001) and lower than those of NBI (11.5 ± 6.9 vs. 13.1 ± 7.7; p = 0.002). No significant differences in TXI and NBI (visibility score: 3.7 ± 1.0 vs. 3.8 ± 1.1; p = 0.833, color difference values: 11.6 ± 7.1 vs. 12.9 ± 8.3; p = 0.099) were observed for neoplastic lesions. Moreover, the diagnostic accuracy of TXI was significantly higher than that of NBI (65.5% vs. 57.6%, p = 0.012) for neoplastic lesions.Conclusions: TXI demonstrated higher visibility than that of WLI and lower than that of NBI. Further investigations are warranted to validate the performance of the TXI mode comprehensively.