脑卒中患者静脉溶栓后神经功能恶化预测模型的系统评价

A Systematic Review of Prediction Models for Neurological Deterioration after Intravenous Thrombolysis in Stroke Patients

摘要: 目的：本研究旨在评价急性缺血性脑卒中患者静脉溶栓后早期神经功能恶化相关预测模型，为患者精准护理与预后评估提供循证依据。方法：系统检索PubMed、中国知网等中英文数据库，由2名经培训的科研人员独立完成文献筛选、资料提取与方法学质量评价。结果：最终纳入21篇文献，含32个预测模型(核心模型21个)，以Logistic回归构建的列线图模型为主，高频预测指标为基线美国国立卫生研究院卒中量表评分、血管病变类型。多数模型受试者工作特征曲线下面积 ≥ 0.70，区分度良好；仅5个核心模型完成外部验证，12个模型未报告校准度，9篇文献偏倚风险不明确。结论：现有早期神经功能恶化预测模型已形成初步研究体系，列线图模型更适配临床护理场景，但其普适性与外推稳定性仍需进一步验证。

Abstract: Objective: To systematically evaluate the construction, validation, and application effectiveness of early neurological deterioration prediction models after intravenous thrombolysis in patients with acute ischemic stroke, providing evidence-based guidance for precise patient care and prognosis assessment. Methods: A systematic search was conducted in both Chinese and English databases such as PubMed and China National Knowledge Infrastructure (CNKI). Literature screening, data extraction, and methodological quality assessment were independently performed by two trained researchers. Results: A total of 21 studies were included, encompassing 32 prediction models (21 core models), primarily nomogram models constructed using Logistic regression. High-frequency predictive indicators included baseline National Institutes of Health Stroke Scale (NIHSS) scores and types of vascular lesions. Most models had an area under the receiver operating characteristic curve (AUC) ≥ 0.70, indicating good discrimination; only 5 core models underwent external validation, 12 models did not report calibration, and the risk of bias was unclear in 9 studies. Conclusion: Existing early neurological deterioration prediction models have formed a preliminary research framework. Nomogram models are more suitable for clinical care scenarios, but their generalizability and extrapolation stability still require further validation.