实用肝脏病杂志 ›› 2026, Vol. 29 ›› Issue (1): 85-88.doi: 10.3969/j.issn.1672-5069.2026.01.022

• 肝硬化 • 上一篇    下一篇

EVL治疗食管胃底静脉曲张破裂出血患者再出血风险预测模型的构建*

王新平, 吴辉, 冯辉, 周圆, 解格, 郭玲, 刘建文   

  1. 410004 长沙市 南华大学附属长沙中心医院消化内科(王新平,吴辉,冯辉,周圆,解格);内科(郭玲);湘潭市中医医院消化内科(刘建文)
  • 收稿日期:2025-04-30 出版日期:2026-01-10 发布日期:2026-02-04
  • 通讯作者: 郭玲,E-mail:805561649 @qq.com
  • 作者简介:王新平,男,主任医师。主要从事消化内镜诊断与治疗学研究。E-mail: kynpi32@163.com
  • 基金资助:
    *湖南省中医药科研计划研究项目(编号:2020213)

Risk factors impacting re-bleeding in patients with esophageal and gastric varices bleeding after endoscopic variceal ligation treatment

Wang Xinping, Wu Hui, Feng Hui, et al   

  1. Department of Gastroenterology,Central Hospital Affiliated to University of South China, Changsha 410004, Hunan Province, China
  • Received:2025-04-30 Online:2026-01-10 Published:2026-02-04

摘要: 目的 分析内镜下治疗肝硬化并发食管胃底静脉曲张破裂出血(EGVB)患者后再出血的风险因素。方法 2021年1月~2024年1月南华大学附属长沙中心医院收治的196例EGVB患者,其中146例为建模集,另50例为验证集,均接受内镜下曲张静脉套扎术(EVL),部分患者联合聚桂醇注射硬化治疗,随访14 d。采用Logistic回归分析影响术后再出血的危险因素,应用受试者工作特征(ROC)曲线评估预测效能。结果 治疗后,196例EGVB患者发生再出血55例(28.1%),其中建模集40例(27.4%),验证集15例(30.0%,P>0.05);再出血组静脉曲张(EV)直径为(1.0±0.2)cm,套扎环数为(12.2±3.8)和套扎次数为(2.4±1.2),均显著大于未再出血组【分别为(0.7±0.1)cm、(10.4±3.1)和(2.4±1.2),P<0.05】;多因素Logistic回归分析显示,曲张血管直径、套扎环数和套扎次数是EGVB患者术后再出血的危险因素(P<0.05);ROC曲线分析显示,建模集风险评估模型预测再出血的曲线下面积(AUC)为0.830(95%CI:0.767~0.881),其灵敏度为84.2%,特异度为71.3%,在50例验证集,其AUC为0.823(95%CI:0.761~0.875),灵敏度为87.8%,特异度为74.5%。结论 了解内镜下治疗后EGVB患者发生再出血的风险因素,做好防治工作,可能提高救治成功率。

关键词: 食管胃底静脉曲张破裂出血, 内镜下套扎术, 再出血, 风险因素, 预测

Abstract: Objective The purpose of this study was to investigate risk factors impacting re-bleeding in patients with esophageal and gastric varices bleeding (EGVB) after endoscopic variceal ligation (EVL)treatment. Methods A total of 196 patients with EGVB were admitted to Changsha Central Hospital Affiliated to University of South China between January 2021 and January 2024, and 146 cases were assigned to training set and another 50 cases were assigned to validation set. All patients received EVL, some of them were treated with polidocanol injection in combination, and followed-up for 14 days. Multivariate Logistic regression analysis was applied to find risk factors, and receiver operating characteristic curve (ROC) was drawn to evaluate prediction efficacy. Results Re-bleeding was found in 55 patients (28.1%), of which, 40 cases (27.4%) in training set and 15 cases (30.0%, P>0.05) in validation set, out of our 196 patients with EGVB after treatment; in re-bleeding group, the diameter of esophageal varices (EV) was(1.0±0.2)cm, the number of ligation rings was (12.2±3.8) and the number of ligation times was (2.4±1.2), all significantly greater than [(0.7±0.1)cm, (10.4±3.1) and (2.4±1.2), respectively, P<0.05] in non-re-bleeding group; multivariate Logistic regression analysis showed that diameter of EV, ligation rings and ligation times were all the independent risk factors for re-bleeding(P<0.05); the ROC analysis demonstrated that the area under the curve (AUC) was 0.830(95%CI:0.767-0.881), with sensitivity (Se) of 84.2% and specificity (Sp) of 71.3% when the model we established was used to predict in the training set, and the AUC was 0.823(95%CI:0.761-0.875), with Se of 87.8% and Sp of 74.5% in validation set. Conclusion Clinicians should take risk factors for re-bleeding into consideration in clinical practice, and take an appropriate measures to deal with it for improving efficacy of hemostasis.

Key words: Esophageal and gastric fundus varices, Endoscopic variceal ligation, Re-bleeding, Risk factors, Prediction