One of the main challenges after colorectal endoscopic submucosal dissection (ESD) is the management of oral anticoagulation. Heparin-bridging therapy (HBT) has been associated with an increased risk of delayed bleeding in gastric ESD, but its impact in the colorectal setting remains controversial.

The aim of the study was to evaluate the impact of HBT on the rate of delayed bleeding after colorectal ESD in patients treated with vitamin K antagonists.

We performed a multicenter retrospective cohort study based on a consecutive prospective registry (11/2012–04/2025). We included patients ≥18 years undergoing colorectal ESD. Delayed bleeding was defined as any bleeding event within 30 days requiring urgent evaluation, hospitalization, or endoscopic reintervention. Procedures aborted or referred to surgery for non-bleeding reasons were excluded.

A total of 855 ESDs were included; 49 patients were on vitamin K antagonists, of whom 38 (77.6%) received HBT. Mean age was 73.8 years (SD 8), and 79.2% were male. The timing of anticoagulation resumption was similar between groups (day 1 in 78.9% with HBT vs 72.7% without HBT; p=0.98).

Among vitamin K antagonist users, no significant differences were observed between HBT and non-HBT groups in baseline characteristics, concomitant antiplatelet therapy, lesion morphology, or lesion location.

The overall delayed bleeding rate was 4.5% (95% CI 3.3–6.1%), and 33.3% (95% CI 21.7–47.5%) among patients on vitamin K antagonists (p<0.001). Median DEBE score was 6 (IQR 4–8) and 17 (IQR 13–19), respectively (p<0.001).

No significant differences were found between HBT and non-HBT groups in delayed bleeding rates (34.3% vs 30%; p=1) or in time to bleeding (median 4 vs 3 days; p=0.57). Two patients (12.5%) required transfusion, and no thromboembolic events occurred.

Reinitiation of anticoagulation using heparin-bridging therapy was not associated with a higher risk of delayed bleeding compared with direct reinitiation. However, the limited sample size affects the robustness of the conclusions and supports the need for prospective studies with a larger sample size.