实用老年医学 ›› 2026, Vol. 40 ›› Issue (8): 781-786.doi: 10.3969/j.issn.1003-9198.2026.08.006

• 临床研究 • 上一篇    下一篇

老年上消化道出血并发多器官功能障碍的危险因素分析以及风险预测模型构建

杨柳, 梁志芳, 王燕, 魏静, 成晓梅, 李妮, 荊玉洁, 李昊   

  1. 710054 陕西省西安市,空军军医大学西京医院第九八六医院消化内科(杨柳,王燕,魏静,成晓梅,李妮,荊玉洁);肿瘤血液科(李昊);
    710038 陕西省西安市,西安医学院第二附属医院超声科(梁志芳)
  • 收稿日期:2026-01-04 出版日期:2026-08-20 发布日期:2026-08-24
  • 通讯作者: 李昊,Email:13992823409@163.com

Risk factor analysis and prediction model construction for multiple organ dysfunction syndrome in elderly patients with upper gastrointestinal bleeding

YANG Liu, LIANG Zhifang, WANG Yan, WEI Jing, CHENG Xiaomei, LI Ni, JING Yujie, LI Hao   

  1. Department of Gastroenterology (YANG Liu, WANG Yan, WEI Jing, CHENG Xiaomei, LI Ni, JING Yujie); Hematology and Oncology Department (LI Hao), 986th Hospital of Xijing Hospital, Air Force Medical University, Xi’an 710054, China;
    Department of Ultrasound (LIANG Zhifang), the Second Affiliated Hospitial of Xi’an Medical University, Xi’an 710038, China
  • Received:2026-01-04 Online:2026-08-20 Published:2026-08-24
  • Contact: LI Hao, Email: 13992823409@163.com

摘要: 目的 探讨老年上消化道出血患者并发多器官功能障碍综合征(MODS)的危险因素,并构建列线图预测模型,为临床早期识别 MODS 高危患者提供循证依据。 方法 回顾性选取2022年3月至2025年3月空军军医大学西京医院第九八六医院收治的365例老年上消化道出血患者临床资料,根据是否并发MODS将患者分为MODS组(55例)和非MODS组(310例)。采用LASSO回归筛选变量,应用多因素logistic回归分析探讨影响老年上消化道出血患者并发MODS的因素,并基于该结果构建列线图预测模型。 结果 通过LASSO回归筛选出6个变量:失血量、失血性休克、血红蛋白、全身炎症反应综合征(SIRS)、心血管疾病和序贯器官衰竭评估(SOFA)评分,并将其纳入多因素logistic回归模型。结果显示,失血性休克、合并SIRS、合并心血管疾病是老年上消化道出血患者并发MODS的危险因素(P<0.05),高血红蛋白水平是保护因素(P<0.05)。基于上述因素构建的列线图预测模型,经验证,该模型拟合度良好(Hosmer-Lemeshow检验χ2=4.544,P=0.631),校准度高(C-index指数为0.972),预测效能良好(AUC为0.972),临床实用性强。 结论 失血性休克、合并SIRS、合并心血管疾病、低血红蛋白水平是老年上消化道出血患者并发MODS的危险因素,基于此构建的列线图预测模型可有效预测老年上消化道出血患者MODS的发生风险,为高危人群的早期干预提供依据。

关键词: 上消化道出血, 多器官功能障碍综合征, 危险因素, 预测模型, 老年人

Abstract: Objective To investigate the risk factors for multiple organ dysfunction syndrome (MODS) in elderly patients with upper gastrointestinal bleeding (UGIB) and to develop a nomogram prediction model for early clinical identification of the high-risk patients. Methods A retrospective cohort study was conducted involving 365 elderly UGIB patients admitted to 986th Hospital of Xijing Hospital, Air Force Medical University from March 2022 to March 2025. All the patients were divided into a MODS group (n=55) and a non-MODS group (n=310). LASSO regression was used to select variables, and multivariate logistic regression analysis was applied to identify risk factors for MODS in elderly UGIB patients. Based on these results, a nomogram prediction model was constructed. Results Six variables [blood loss, hemorrhagic shock, hemoglobin level, systemic inflammatory response syndrome (SIRS), cardiovascular disease, and Sequential Organ Failure Assessment (SOFA) score] were selected through LASSO regression and incorporated into a multivariate logistic regression analysis, which showed that hemorrhagic shock, SIRS, and cardiovascular disease were independent risk factors for MODS in elderly UGIB patients (P<0.05), while high hemoglobin level was a protective factor (P<0.05). The nomogram prediction model constructed based on these factors showed good fit (Hosmer-Lemeshow test:χ2=4.544,P=0.631), high discriminative ability (C-index=0.972), excellent predictive performance (AUC=0.972), and strong clinical applicability. Conclusions Hemorrhagic shock, SIRS, cardiovascular disease and low hemoglobin level were independent risk factors for MODS in elderly UGIB patients. The nomogram model, developed from these factors, demonstrated good performance in predicting MODS risk among elderly UGIB patients, potentially enabling timely intervention for high-risk individuals.

Key words: upper gastrointestinal bleeding, multiple organ dysfunction syndrome, risk factors, prediction model, aged

中图分类号: