中国实用神经疾病杂志2025，Vol.28Issue(12)：1494-1499,6.DOI:10.12083/SYSJ.250315
基于血清指标及临床资料构建高血压脑出血发生HE的预测模型及验证
Construction and validation of prediction model for the occurrence of HE in hypertensive intracerebral hemorrhage based on serum indicators and clinical data
摘要
Abstract
Objective To construct a predictive model for hematoma enlargement(HE)in patients with hypertensive intracerebral hemorrhage(HICH)based on clinical data,the microtubule-associated protein Tau(Tau protein),neuron-specific enolase(NSE),and aquaporin 4(AQP4),and conduct empirical analysis on the predictive ability of the model.Methods A total of 204 HICH patients admitted to the Second People's Hospital of Qinzhou from May 2021 to May 2024 were selected and divided into expanded group and non-expanded group according to whether they developed HE or not.The clinical data and laboratory index data in the two groups were collected,and the risk factors for the occurrence of HE in HICH patients were screened by multifactorial LASSO-Logistic regression,and the column-line graph prediction model was constructed,and the predictive efficacy of the model was empirically evaluated.Results Univariate analysis revealed that the time from onset to admission in the expanded group was longer than that in the non-expanded group,and the proportion of patients with diabetes,systolic blood pressure,volume of hemorrhage,proportion of patients with CT island sign,Tau protein,NSE,and AQP4 were higher than those in the non-expanded group(P<0.05).Multivariate analysis using LASSO-Logistic regression indicated that prolonged time from onset to admission,increased volume of hemorrhage,CT island sign,elevated Tau protein,elevated NSE,and elevated AQP4 were all risk factors for the occurrence of HE in HICH patients(P<0.05).The consistency index(C-index)of the prediction model for the occurrence of HE in HICH patients constructed based on the aforementioned risk factors was 0.821.Receiver operating characteristic(ROC)curve analysis showed that the AUC of the model in the training dataset and external validation dataset were 0.821(95%CI:0.757-0.886)and 0.867(95%CI:0.793-0.942),respectively.The calibration curve demonstrated that the model had strong predictive ability.Conclusion The risk factors for HE in HICH patients include prolonged time from onset to hospital admission,increased blood loss,CT island sign,elevated Tau protein,elevated NSE,and elevated AQP4.The HE prediction model constructed based on these factors has strong predictive ability and can effectively identify individuals at high risk of early HE.关键词
高血压/脑出血/血肿扩大/Logistic预测模型/外部验证/内部验证Key words
Hypertension/Cerebral hemorrhage/Hematoma enlargement/Logistic prediction model/Exter-nal validation/Internal validation分类
医药卫生引用本文复制引用
梁宝毅,周达坤,郭海敏,梁立毅,洪允钦,周传凯,禤彩霞,何东明,廖振南..基于血清指标及临床资料构建高血压脑出血发生HE的预测模型及验证[J].中国实用神经疾病杂志,2025,28(12):1494-1499,6.基金项目
广西卫健委自筹经费科研课题(编号:Z-N20241667) （编号:Z-N20241667）
