A clinical tool to predict delay post-polypectomy bleeding after resection of colorectal polyps

Background Delayed post-polypectomy bleeding (DPB) is a common and impactful complication following colonoscopic polypectomy, often leading to unplanned care utilization. Despite its clinical significance, validated predictive tools for DPB remain scarce. This study sought to develop a nomogram that integrates patient, polyp, and procedural factors to estimate DPB risk. Methods Retrospective data on patient demographics, polyp characteristics, and procedural details were collected. The study population was randomly split into a training cohort for model development and a validation cohort for performance testing. Multivariable logistic regression was employed to identify independent predictors of DPB in the training set, which were then used to construct a nomogram. The nomogram's predictive accuracy was validated using the independent test cohort. Results A total of 1738 patients with 3650 polyps were included in the study. The rate of DPB in this study was 1.9% (71/3650). In the training group, the independent predictors were age(HR 0.384), prophylactic clipping(HR 0.384), polyp morphology(HR 2.546), polyp size(HR 2.491), thrombin time (HR 4.263). The derived risk score comprised age < 60 years, prophylactic clipping, polyp morphology of Ip/Isp, polyp size ≥ 10mm, TT ≥ 21 sec. In the validation cohort, the C-index for DPB prediction was 0.802. Subgroup analysis showed that prophylactic clipping is not suitable for small (< 10mm) polyps. Conclusions We derive a readily applicable risk score that identifies patients with high risk of DPB. This score may help improving the management of those undergoing polypectomy and assist in allocating resources more cost-effectively to prevent and manage DPB. In addition, we found prophylactic clipping may increase the risk of DPB. Further well-designed researches are needed to demonstrate it.