脑出血继发性脑损伤预测模型构建  

Construction of prediction model for secondary brain injury caused by cerebral hemorrhage

作　　者：陆燕[1] 杨帆[1] 金慧 Lu Yan;Yang Fan;Jin Hui(Department of Critical Care Medicine,Affiliated Nantong Hospital of Shanghai University,Nantong 226000,China)

机构地区：[1]上海大学附属南通医院(南通市第六人民医院)重症医学科,江苏南通226000

出　　处：《海军医学杂志》2025年第4期362-367,共6页Journal of Navy Medicine

基　　金：江苏省自然科学基金项目(BK20221418)。

摘　　要：目的分析脑出血患者继发性脑损伤(SBI)的临床影响因素,以此构建列线图预测模型并进行验证,以期为临床减少脑出血患者SBI发生,改善患者预后提供一定的指导。方法回顾性分析上海大学附属南通医院(南通市第六人民医院)在2020年1月至2023年9月收治的脑出血患者151例,依据8∶2定律随机分为训练集121例和验证集30例,根据患者在发病后7 d内是否发生SBI将其分为SBI组和非SBI组,其中训练集SBI组37例、非SBI组84例,验证集SBI组9例、非SBI组21例。收集患者临床资料,分析患者发生SBI的影响因素,并以此构建列线图模型预测SBI发生风险;采用受试者操作特征(ROC)曲线下面积(AUC)分析预测模型对SBI的预测效能。结果单因素分析结果显示,SBI组年龄≥60岁占比、出血量≥20 ml占比、格拉斯哥昏迷评分(GCS)<10分占比、全身免疫炎症指数(SII)水平≥952.31×109/L占比均高于无SBI组(P<0.05)。二元Logistic回归分析结果显示,年龄(OR=4.489,95%CI:2.364~8.521)、出血量(OR=3.804,95%CI:1.693~8.546)、SII水平(OR=5.642,95%CI:1.864~17.075)是患者发生SBI的独立危险因素(P<0.05)。基于上述影响因素构建的列线图预测模型经Bootstrap法内部验证显示,C⁃index指数为0.841(95%CI:0.792~0.931),预测患者发生SBI的校正曲线趋近于理想曲线(P>0.05)。训练集ROC显示,列线图模型预测患者发生SBI的灵敏度为87.50%、特异度为90.50%,AUC为0.882(95%CI:0.799~0.965)。验证集ROC显示,列线图模型预测患者发生SBI的灵敏度为89.20%,特异度为86.90%,AUC为0.874(95%CI:0.788~0.959)。结论年龄、出血量、SII水平是脑出血患者发生SBI的独立危险因素,基于上述因素构建的列线图风险预测模型可较好地评估患者SBI发生风险。Objective To explore the influencing factors of secondary brain injury(SBI)in patients with cerebral hemorrhage,and to build and verify a nomogram prediction model,so as to provide a basis for reducing the occurrence of SBI and improving the prognosis of patients with cerebral hemorrhage.Methods A total of 151 patients with cerebral hemorrhage who were admitted to Affiliated Nantong Hospital of Shanghai University(The Sixth People’s Hospital of Nantong)from January 2020 to September 2023 were retrospectively analyzed and randomly assigned(8∶2)to training set(121 cases)and validation set(30 cases).SBI occurred in 46 patients(SBI group,37 in the training set and 9 in the validation set)within 7 days of the onset of cerebral hemorrhage,and the other 105 patients did not suffer from SBI(non⁃SBI group,84 in the training set and 21 in the validation set).The clinical data of the patients were collected,and the factors affecting the occurrence of SBI were analyzed.A nomogram model was constructed to predict the risk of SBI.The area under the receiver operating characteristic(ROC)curve(AUC)was used to analyze the predictive efficacy of the model for SBI.Results Univariate analysis showed that the proportion of patients aged≥60 years old,bleeding volume≥20 ml,Glasgow coma score<10,and systemic immune⁃inflammation index(SII)≥952.31×109/L in the SBI group were higher than those in the non⁃SBI group(all P<0.05).Binary Logistic regression analysis showed that age(OR=4.489,95%CI:2.364⁃8.521),amount of bleeding(OR=3.804,95%CI:1.693⁃8.546)and SII level(OR=5.642,95%CI:1.864⁃17.075)were independent risk factors for SBI(all P<0.05).Bootstrap internal validation of the nomogram prediction model based on the above influencing factors showed that the C⁃index was 0.841(95%CI:0.792⁃0.931),and the calibration curve for predicting SBI was close to the ideal curve(P>0.05).The ROC curve of the training set showed that the sensitivity,specificity,and AUC of the nomogram model for predicting the occurrence of SBI wa

关 键 词：脑出血 继发性脑损伤 影响因素 列线图 预测模型 

分 类 号：R743.9[医药卫生—神经病学与精神病学]