通讯作者: 刘超

DOI：10.12201/bmr.202607.00019

Prediction Study of VTE Risk in Patients with Acute Cerebral Hemorrhage in Neurosurgery Based on Clinical Indicators and Modified Caprini Model

• 摘要：目的 分析神经外科急性期脑出血患者发生静脉血栓栓塞症(VTE)的独立危险因素,并以此为基础构建和验证一个针对脑出血患者的改良Caprini风险评估模型。方法 采用回顾性分析方法,选取2022年1月至2023年12月浙江省某三甲医院神经外科收治的312例脑出血急性期患者作为研究对象,并根据是否发生VTE分为对照组(非血栓组,n=156)和观察组(血栓组,n=156)。收集两组患者的基线资料、临床指标和干预措施,先通过单因素分析筛选有统计学意义的变量,再采用多因素Logistic回归分析确定VTE的独立危险因素。基于筛选出的核心变量构建预测模型,并对模型的区分度(受试者工作特征曲线,ROC)和稳定性(5折交叉验证)进行评估。?结果? 多因素Logistic回归分析结果显示,D-二聚体升高(OR=12.486, 95%CI:5.225-29.838, P=<0.05)、运动缺陷(OR=3.373, 95%CI:2.157-5.274, P<0.05)、缺血性脑出血(OR=2.146, 95%CI:1.223-3.766, P<0.05)及感染(OR=2.380, 95%CI:1.607-3.524, P<0.05)是脑出血患者发生VTE的独立危险因素。基于以上5个核心变量构建的改良预测模型,在测试集中显示出较高的预测效能：准确率81.91%,精确率82.61%,召回率80.85%,F1分数81.72%,曲线下面积(AUC)达0.882。5折交叉验证的平均AUC为0.883(±0.068),证明模型具有良好的区分能力和稳定性。结论 高龄、D-二聚体升高、运动缺陷、缺血性脑出血及感染是神经外科脑出血急性期患者发生VTE的关键危险因素。本研究构建的包含上述特异性临床指标的改良预测模型具有较好的临床应用价值,能更精准地识别VTE高危患者,为实现个体化预防提供量化依据。

  关键词： 静脉血栓栓塞症；脑出血；危险因素；预测模型；Caprini模型

   

  Abstract: Abstract Objective To analyze the independent risk factors of venous thromboembolism (VTE) in patients with acute intracerebral hemorrhage in neurosurgery, and to construct and validate a modified Caprini risk assessment model for patients with intracerebral hemorrhage based on the above factors. ?Methods? A retrospective analysis was performed on 312 patients with acute intracerebral hemorrhage admitted to the neurosurgery department of a tertiary Class A hospital in Zhejiang Province from January 2022 to December 2023. The patients were divided into a control group (non-thrombosis group, n=156) and an observation group (thrombosis group, n=156) according to whether VTE occurred. Baseline data, clinical indicators and intervention measures of the two groups were collected. Statistically significant variables were first screened by univariate analysis, and then independent risk factors of VTE were determined by multivariate Logistic regression analysis. A prediction model was constructed based on the screened core variables, and the discrimination (receiver operating characteristic curve, ROC) and stability (5-fold cross-validation) of the model were evaluated. ?Results? The results of multivariate Logistic regression analysis showed that elevated D-dimer (OR=12.486, 95%CI: 5.225-29.838, P<0.05), motor deficit (OR=3.373, 95%CI: 2.157-5.274, P<0.05), ischemic intracerebral hemorrhage (OR=2.146, 95%CI: 1.223-3.766, P<0.05) and infection (OR=2.380, 95%CI: 1.607-3.524, P<0.05) were independent risk factors for VTE in patients with intracerebral hemorrhage. The modified prediction model constructed based on the above 5 core variables showed high prediction efficiency in the test set: accuracy 81.91%, precision 82.61%, recall 80.85%, F1 score 81.72%, and the area under the curve (AUC) reached 0.882. The average AUC of 5-fold cross-validation was 0.883 (±0.068), which proved that the model had good discrimination and stability. ?Conclusion? Advanced age, elevated D-dimer, motor deficit, ischemic intracerebral hemorrhage and infection are the key risk factors for VTE in patients with acute intracerebral hemorrhage in neurosurgery. The modified prediction model containing the above specific clinical indicators constructed in this study has good clinical application value, can more accurately identify patients at high risk of VTE, and provides a quantitative basis for individualized prevention

  Key words: Venous Thromboembolism；Cerebral Hemorrhage；Risk Factor；Prediction Model；Caprini Model

  提交时间：2026-07-10

  版权声明：作者本人独立拥有该论文的版权，预印本系统仅拥有论文的永久保存权利。任何人未经允许不得重复使用。
• html

• 图表

• 刘娇娇, 郭连梅, 孙晓红, 张喆, 田野. 基于机器学习的出血性卒中患者下肢深静脉血栓预测模型构建. 2025. doi: 10.12201/bmr.202510.00023

  赵卫良, 夏吉勇, 潘新永. 高血压脑出血患者术后肺部感染的危险因素分析. 2025. doi: 10.12201/bmr.202511.00046

  杨燕飞, 李娜, 徐梦琪, 张敏, 王峻. 骨科住院病人低蛋白血症风险预测模型研究进展. 2025. doi: 10.12201/bmr.202510.00017

  蔡欣欣, 吴夏阳. 早产儿入院败血症风险预测模型的构建与验证. 2025. doi: 10.12201/bmr.202505.00013

  胡欣. 剖宫产术中寒战风险预测模型及列线图的构建. 2025. doi: 10.12201/bmr.202501.00053

  叶妙淇, 刘建颖, 许毅娇, 王路敏, 夏霄彤. 免疫检查点抑制剂治疗肿瘤患者中血液指标预测静脉血栓风险的研究. 2026. doi: 10.12201/bmr.202601.00053

  张二利, 何兰兰, 李丹阳, 沈立, 吴钟华, 张军, 叶永强. 基于脑小血管病影像学总负荷的急性脑梗死患者住院时间延长列线图模型构建与验证. 2025. doi: 10.12201/bmr.202501.00008

  施明杰, 叶洪舟, 袁琛. 大环内酯类耐药肺炎支原体肺炎患儿发展为重症肺炎的危险因素分析及预测模型构建. 2024. doi: 10.12201/bmr.202409.00021

  陈冰融, 舒文秀, 罗柳飞, 童嘉琦, 乐静. 骨髓瘤患者发生3-4级骨髓瘤骨病的风险预测模型构建. 2024. doi: 10.12201/bmr.202410.00057

  张淑涵, 杨 鹏, 杨正武. 乙型肝炎肝硬化首次并发消化道出血的危险因素及预测模型分析. 2025. doi: 10.12201/bmr.202504.00075

• 序号	提交日期	编号	操作
  1	2026-06-10	10.12201/bmr.202607.00019V1	下载

• 公开评论  匿名评论  仅发给作者 提交评论