Baseline characteristics.

Background and aims
Emergency endoscopic hemostasis for colonic diverticular bleeding is effective in preventing serious consequences. However, the low identification rate of the bleeding source makes the procedure burdensome for both patients and providers. We aimed to establish an efficient and safe emergency endoscopy system.
Methods
We prospectively evaluated the usefulness of a scoring system (Jichi Medical University diverticular hemorrhage score: JD score) based on our experiences with past cases. The JD score was determined using four criteria: CT evidence of contrast agent extravasation, 3 points; oral anticoagulant (any type) use, 2 points; C-reactive protein ≥1 mg/dL, 1 point; and comorbidity index ≥3, 1 point. Based on the JD score, patients with acute diverticular bleeding who underwent emergency or elective endoscopy were grouped into JD ≥3 or JD <3 groups, respectively. The primary and secondary endpoints were the bleeding source identification rate and clinical outcomes.
Results
The JD ≥3 and JD <3 groups included 35 and 47 patients, respectively. The rate of bleeding source identification, followed by the hemostatic procedure, was significantly higher in the JD ≥3 group than in the JD <3 group (77% vs. 23%, p <0.001), with a higher JD score associated with a higher bleeding source identification rate. No significant difference was observed between the groups in terms of clinical outcomes, except for a higher incidence of rebleeding at one-month post-discharge and a higher number of patients requiring interventional radiology in the JD ≥3 group than in the JD <3 group. Subgroup analysis showed that successful identification of the bleeding source and hemostasis contributed to a shorter hospital stay.
Conclusion
We established a safe and efficient endoscopic scoring system for treating colonic diverticular bleeding. The higher the JD score, the higher the bleeding source identification, leading to a successful hemostatic procedure. Elective endoscopy was possible in the JD <3 group when vital signs were stable.

背景与研究目的
针对结肠憩室出血的急诊内镜止血术，可有效预防严重不良结局。然而，出血源识别率偏低，令患者与医护人员均承受较大负担。本研究旨在构建一套高效且安全的急诊内镜诊疗体系。

方法
我们基于既往病例诊疗经验，前瞻性评估了一款评分系统——自治医科大学憩室出血评分（JD评分）的应用价值。JD评分由四项指标构成：CT显示造影剂外渗（3分）、口服任何类型抗凝剂（2分）、C反应蛋白（C-reactive protein）≥1 mg/dL（1分）以及合并症指数≥3（1分）。依据JD评分结果，将接受急诊或择期内镜检查的急性憩室出血患者分为JD评分≥3组与JD评分<3组。本研究的主要与次要终点分别为出血源识别率与临床结局。

结果
JD≥3组与JD<3组分别纳入35例、47例患者。JD≥3组的出血源识别率及后续止血操作实施率均显著高于JD<3组（77% vs. 23%，p<0.001），且JD评分越高，出血源识别率亦越高。两组临床结局无显著差异，但JD≥3组出院后1个月再出血发生率更高，需接受介入放射学治疗的患者数量更多。亚组分析显示，成功识别出血源并完成止血操作可缩短患者住院时长。

结论
本研究构建了一套用于结肠憩室出血诊疗的安全高效内镜评分系统。JD评分越高，出血源识别率越高，止血操作成功率亦随之提升。对于JD<3组患者，若生命体征稳定，则可实施择期内镜检查。