S2300 Characteristics of Post-Endoscopic Treatment Bleeding in Resected Colorectal Tumors Stratified by Resection Methods

Introduction: Cold snare polypectomy (CSP), endoscopic mucosal resection (EMR), and endoscopic submucosal dissection (ESD) are widely used clinically. While effective, post-procedural bleeding remains an unresolved issue. Prior studies typically examined long-term outcomes, during which endoscopist skill may vary. This study aimed to investigate factors contributing to bleeding in a short, fixed period where operator skill was presumed constant. Methods: A total of 982 patients with 2793 colorectal lesions (carcinoma, adenoma, SSL, TSA, hyperplastic polyps) resected endoscopically at Nippon Medical School Hospital from January to December 2020 were included. Lesions were classified into CSP, EMR (including hot polypectomy and precutting), and ESD groups. In Study 1, we compared bleeding rates, patient characteristics, and endoscopist level across groups. In Study 2, we compared CSP vs EMR and subtypes within the EMR group. Statistical tests included Fisher’s exact test, Mann–Whitney U, and Kruskal–Wallis. Results: There were 1551 CSP, 1141 EMR, and 101 ESD cases. Median (IQR) tumor sizes were 4 mm (3–5), 8 mm (6–12), and 26 mm (21–40), respectively. Antithrombotic use was 24% (CSP), 24% (EMR), and 23% (ESD). Carcinoma rates were 0% (CSP), 11% (EMR), and 70% (ESD). Post-treatment bleeding occurred in 0.26% (4 lesions) of CSP, 0.88% (10 lesions) of EMR, and 3.0% (3 lesions) of ESD (P < 0.001). Excluding ESD, EMR had a significantly higher bleeding rate than CSP (P = 0.027). Among bleeding cases, 100% (5/5) in CSP, 0% in EMR, and 33% in ESD were on antithrombotics. Regarding operator level, 75% (3/4) of CSP, 89% (8/9) of EMR, and 100% (3/3) of ESD bleeding cases were performed by experts. In Study 2, patients had 1 (35%), 2 (24%), 3 (15%), 4–5 (15%), or ≥6 (11%) lesions resected. Among 11 bleeding cases: 6 had 1 lesion, 4 had 4 lesions, and 1 had 9 lesions. No significant difference in lesion count was found between bleeding and non-bleeding cases (median 1 vs 2 lesions, P = 0.12). Conclusion: Bleeding rates aligned with prior reports, even when most procedures were performed by experts. Post-treatment bleeding occurred regardless of lesion number or anticoagulant use, suggesting that even expert-performed resections require careful management to reduce bleeding risk.