河北医科大学学报 ›› 2025, Vol. 46 ›› Issue (6): 662-667.doi: 10.3969/j.issn.1007-3205.2025.06.007

• • 上一篇    下一篇

中老年急性缺血性脑卒中早期神经功能恶化预测模型的建立与验证:单中心回顾性研究

  

  1. 1.哈尔滨医科大学附属第一医院神经内科,黑龙江 哈尔滨 150001;2.哈尔滨医科大学附属第一医院CCU,
    黑龙江 哈尔滨 150001;3.哈尔滨医科大学附属第一医院肾内科,黑龙江 哈尔滨150001

  • 出版日期:2025-06-25 发布日期:2025-07-04
  • 作者简介:王翀(1987-),女,黑龙江哈尔滨人,哈尔滨医科大学附属第一医院主管技师,医学学士,从事神经内科疾病诊治研究。

  • 基金资助:
    黑龙江省自然科学基金(H2021K106)

Establishment and validation of a prediction model for early neurological deterioration in middle-aged and elderly patients withacute ischaemic stroke: a single-centre retrospective study

  1. 1.Department of Neurology, the First Hospital of Harbin Medical University, Heilongjiang Province, 
    Harbin 150001, China; 2.Department of Critical Care Unit, the First Hospital of Harbin Medical 
    University, Heilongjiang Province, Harbin 150001, China; 3.Department of Nephrology, 
    the First Hospital of Harbin Medical University, Heilongjiang 
    Province, Harbin 150001, China

  • Online:2025-06-25 Published:2025-07-04

摘要: 目的基于单中心临床数据,初步建立中老年急性缺血性脑卒中(acute ischemic stroke,AIS)患者发生早期神经功能恶化(early neurological deterioration,END)的预测模型,并对其性能进行验证。
方法回顾性收集2020年1月—2023年12月哈尔滨医科大学附属第一医院治疗的318例AIS患者的临床资料,根据溶栓7 d内END发生情况分为END组80例(25.16%),非END组238例(74.84%)。采用二分类Logistic回归分析筛选END的危险因素,并根据Logistic回归分析结果建立预测模型,最后采用受试者工作特征(receiver operating characteristic,ROC)曲线对模型识别END的性能进行评价,评价指标为曲线下面积(area under the curve,AUC)、准确率、敏感度、特异度。
结果多因素Logistic回归分析显示,年龄(OR=1.247,95%CI:1.030~1.510)、美国国立卫生研究院卒中量表(National Institute of Health Stroke Scale,NIHSS)评分(OR=1.872,95%CI:1.325~2.645)、超敏C反应蛋白(OR=1.430,95%CI:1.091~1.876)、高血糖(OR=1.372,95%CI:1.109~1.697)、纤维蛋白原/白蛋白比值(OR=1.537,95%CI:1.184~1.996)为中老年AIS患者发生END的独立危险因素。基于Logistic回归分析筛选的5项参数建立END预测模型,经ROC曲线验证,模型预测中老年AIS患者发生END的AUC为0.879(95%CI:0.831~0.927),准确率为82.71%,敏感度为78.75%,特异度为84.03%。
结论年龄、NIHSS评分、超敏C反应蛋白、血糖、纤维蛋白原/白蛋白比值为中老年AIS患者发生END的独立危险因素,据此建立的模型对于END具有良好的预测价值。


关键词: 缺血性卒中, 早期神经功能恶化, 中老年

Abstract: Objective To initially establish a prediction model for the occurrence of early neurological deterioration (END) in middle-aged and elderly patients with acute ischemic stroke (AIS) based on single-centre clinical data, and to validate its performance. 
Methods The clinical data of 318 AIS patients treated in the First Hospital of Harbin Medical University from January 2020 to December 2023 were retrospectively collected and divided into the END group (n=80, 25.16%) and the non-END group (n=238,74.84%) according to the occurrence of END within 7 d of thrombolysis. The risk factors of END were screened using binary logistic regression, and a prediction model was established based on the results of logistic regression.Finally, the performance of the model in identifying END was evaluated using the receiver operating characteristic (ROC) curve, and the evaluation indexes were AUC, accuracy, sensitivity, and specificity. 
Results Multivariate logistic regression analysis showed that age (OR=1.247, 95%CI: 1.030-1.510), National Institute of Health stroke scale(NIHSS) score (OR=1.872, 95%CI: 1.325-2.645), hypersensitive C-reactive protein(hs-CRP) (OR=1.430, 95%CI: 1.091-1.876), hyperglycaemia (OR=1.372, 95%CI: 1.109-1.697), and fibrinogen to albumin ratio (FAR) (OR=1.537, 95%CI: 1.184-1.996) were the independent risk factors for the development of END in middle-aged and elderly AIS patients. An END prediction model was established based on the five parameters screened by logistic regression analysis, and the are under the curve (AUC) of the model for predicting the occurrence of END in middle-aged and elderly AIS patients was 0.879 (95%CI: 0.831-0.927), with an accuracy of 82.71%, a sensitivity of 78.75%, and a specificity of 84.03%, as validated by the ROC curve. 
Conclusion Age, NIHSS score, hs-CRP, blood glucose, and FAR are independent risk factors for END in middle-aged and elderly patients with AIS, and the model established accordingly has good predictive value for END. 


Key words: ischemic stroke, early neurological deterioration, middle-aged and elderly