Extensive circumferential colorectal endoscopic submucosal dissection carries a high risk of stricture formation. While prophylactic closure reduces this risk, evidence in high circumferential resections (>75%) remains limited. We evaluated the association between prophylactic closure and stricture formation in high circumferential colorectal ESD.

Retrospective analysis 1,267 consecutive colorectal ESD from a single tertiary center was conducted. High circumferential resection (>75% or complete circumferential) cases were identified. Primary outcome was stricture formation requiring dilatation. Secondary outcomes included perforation, delayed bleeding, and composite adverse events. Multivariable Firth logistic regression adjusted for lesion size, location (rectum vs colon), malignancy, and procedure duration.

138 (10.9%) of lesions required >75% circumferential resection. Prophylactic closure was performed in 40/138 (29.0%) cases using predominantly suture-based methods: OverStitch (n=29, 72.5%), Hemoclips (n=7, 17.5%), V-LOC (n=4, 10.0%). En bloc resection was achieved in 137/138 (99.3%). Mean resection duration was significantly shorter in the closure group (118 vs 179 min, p<0.05). Stricture rates increased with circumferential extent: 0/560 (<50%), 0/569 (50-75%), and 10/138 (7.2%) in >75% resections (p<0.001 for trend). Among high circumferential cases, stricture occurred in 0/40 (0%) with closure vs 10/98 (10.2%) without closure (p=0.036) After adjustment for lesion size (OR 1.019 per mm, 95% CI 1.006-1.032, p=0.004), rectal location (OR 7.38, 95% CI 0.40-135), and malignancy (OR 0.62, 95% CI 0.15-2.48), closure showed a protective trend (adjusted OR 0.061, 95% CI 0.002-2.17, p=0.125). Number needed to treat was 10 patients (95% CI 5-20). Stricture risk increased significantly with lesion size quartiles: Q1 (smallest) 2.6%, Q2 0%, Q3 6.1%, Q4 (largest) 20.6% (p=0.007). The protective effect of closure was most pronounced in the largest quartile (0/7 vs 7/27, p=0.165). Closure was associated with significantly shorter hospital stay (median 2 vs 3 days, mean 2.7±1.42 vs 3.47±1.63 days, p=0.007). No delayed bleeding occurred in either group. Perforation occurred in 1/40 (2.5%) with closure vs 0/98 without (p=0.29). Composite adverse events were numerically lower with closure: 1/40 (2.5%) vs 10/98 (10.2%), p=0.176.

Prophylactic closure achieved zero strictures versus 10.2% without closure and significantly shorter hospital stay in high circumferential colorectal ESD with a number needed to treat of 10. Our findings support routine consideration of prophylactic closure for extensive circumferential colorectal ESD, particularly for large lesions. Prospective multicenter studies with standardized protocols are warranted to validate our findings.