Clip Closure After Resection of Large Colorectal Lesions With Substantial Risk of Bleeding

Abstract

It is not clear whether closure of mucosal defects with clips after colonic endoscopic mucosal resection (EMR) prevents delayed bleeding, although it seems to have no protective effects when risk is low. We performed a randomized trial to evaluate the efficacy of complete clip closure of large (≥2 cm) nonpedunculated colorectal lesions after EMR in patients with an estimated average or high risk of delayed bleeding. We performed a single-blind trial at 11 hospitals in Spain from May 2016 through June 2018, including 235 consecutive patients who underwent EMR for large nonpedunculated colorectal lesions with an average or high risk of delayed bleeding (based on Spanish Endoscopy Society Endoscopic Resection Group score). Participants were randomly assigned to groups that received closure of the scar with 11-mm through-the-scope clips (treated, n= 119) or no clip (control, n= 116). The primary outcome was proportion of patients in each group with delayed bleeding, defined as evident hematochezia that required medical intervention within 15 days after colonoscopy. In the clip group, complete closure was achieved in 68 (57%) cases, with partial closure in 33 (28%) cases and failure to close in 18 (15%) cases. Delayed bleeding occurred in 14 (12.1%) patients in the control group and in 6 (5%) patients in the clip group (absolute risk difference, reduction of 7% in the clip group; 95% confidence interval, -14.7% to 0.3%). After completion of the clip closure, there was only 1 (1.5%) case of delayed bleeding (absolute risk difference, reduction of 10.6%; 95% confidence interval, -4.3% to 17.9%). In a randomized trial of patients with large nonpedunculated colorectal lesions undergoing EMR, we found that clip closure of mucosal defects in patients with a risk of bleeding can be a challenge, but also reduces delayed bleeding. Prevention of delayed bleeding required complete clip closure. ClinicalTrials.gov ID: NCT02765022.