基于真实世界的自发性脑出血后脑积水形成的危险因素及临床风险预测模型研究  

A clinical risk prediction model for hydrocephalus formation after spontaneous intracerebral hemorrhage based on the real world

作　　者：曾凤 黄俊章 孟德浩 雷斗斗 文涛 莫伟 廖壮槟 黄梓雄 龙霄翱 陈兵 ZENG Feng;HUANG Junzhang;MENG Dehao;LEI Doudou;WEN Tao;MO Wei;LIAO Zhuangbin;HUANG Zixiong;LONG Xiaoao;CHEN Bing(Guangdong Medical University,Zhanjiang 524023,Guangdong,China;Department of Neurosurgery,Affiliated Hospital of Guangdong Medical University,Zhanjiang 524001,Guangdong,China;Lianjiang Traditional Chinese Medicine Hospital,Lianjiang 524499,Guangdong,China)

机构地区：[1]广东医科大学,广东湛江524023 [2]广东医科大学附属医院神经外科中心,广东湛江524001 [3]廉江市中医院,广东廉江524499

出　　处：《中国现代手术学杂志》2024年第5期385-389,共5页Chinese Journal of Modern Operative Surgery

基　　金：2023年度广东医科大学高等教育教学研究课题(2FY23006)。

摘　　要：目的应用基于真实世界的数据分析,构建自发性脑出血后脑积水形成的列线图预测模型,为临床决策提供证据支持。方法收集2018年1月至2022年12月期间广东医科大学附属医院收治的433例自发性脑出血患者临床资料,采用多因素logistic回归分析筛选出与自发性脑出血后脑积水形成相关的影响因素,并根据这些影响因素制定风险预测列线图模型。结果在433例自发性脑出血患者中,发生脑积水31例,未发生脑积水420例。通过多因素logistic回归分析发现格拉斯哥昏迷评分(Glasgow coma scale,GCS)、出血部位及破入脑室是独立危险因素(P<0.05),而性别、年龄、病因、高血压病史等不是独立危险因素(P>0.05)。将GCS、出血部位及破入脑室这三个变量纳入R软件构建预测模型,列线图预测模型的受试者工作特征曲线下面积为0.930(95%CI:0.895~0.930,P<0.001),校准曲线显示列线图校准效果良好(P>0.05),建立的模型具有较好预测价值。结论GCS、出血部位及破入脑室是自发性脑出血后发生脑积水风险的危险因素,风险预测列线图模型有助于改进脑出血患者整体的治疗效果。Objective To construct a clinical risk prediction model for hydrocephalus formation after spontaneous intracerebral hemorrhage based on the real world,so as to provide evidence support for clinical decision-making.Methods Clinical data was studied in 433 patients with spontaneous cerebral hemorrhage admitted to the Affiliated Hospital of Guangdong Medical University from January 2018 to December 2022.Multiple factor logistic regression analysis was used to screen the influencing factors related to the formation of hydrocephalus after spontaneous cerebral hemorrhage,and to construct a risk prediction nomogram model based on these influencing factors.Results Among 433 patients with spontaneous cerebral hemorrhage,31 developed hydrocephalus and 420 did not.Through multiple logistic regression analysis,it was found that Glasgow coma scale(GCS),bleeding site,and ventricular rupture were independent risk factors(P<0.05),while gender,age,etiology,and history of hypertension were not(P>0.05).A risk prediction nomogram model was constructed based on these influencing factors by the R programming language.The area under the receiver operating characteristic curve of the prediction nomogram model was 0.930(95%CI:0.895~0.930,P<0.001),and the calibration curve showed that the calibration effect of nomogram was good(P>0.05).The prediction nomogram model had good predictive value.Conclusions GCS,the bleeding site and ventricles rupture are risk factors for hydrocephalus after spontaneous cerebral hemorrhage,and the risk prediction nomogram model can help improve the overall treatment effect of cerebral hemorrhage patients.

关 键 词：脑出血 脑积水 预测 列线图 

分 类 号：R74[医药卫生—神经病学与精神病学]