Abstract:

Objective: To identify the risk factors associated of ischemic recurrence in patients with chronic symptomatic severe intracranial vertebrobasilar artery stenosis and to establish a corresponding risk prediction model (Nomogram model). Methods: Total 387 patients with symptomatic severe intracranial vertebrobasilar artery stenosis from January 2017 to March 2025 in Tianjin Huanhu Hospital were retrospectively analyzed and divided into recurrence group (n=97) and non-recurrence group (n=290) according to whether they developed ischemic cerebrovascular events within 90 d of follow-up. Univariate and multivariate Logistic regression analyses were adopted to identify the risk factors of recurrent ischemic cerebrovascular events in patients with chronic symptomatic severe intracranial vertebrobasilar artery stenosis. A Nomogram model was established. Using JavaScript and other network programming languages, we developed the first mobile prediction platform and real-life services for intracranial vertebrobasilar artery stenosis patients to predict their risk of recurrent ischemic cerebrovascular events. Results: Logistic regression analysis showed that high National Institutes of Health Stroke Scale (NIHSS) score at adminssion (OR=1.426, 95%CI: 1.242-1.637; P=0.000), poor collateral compensation (OR=5.786, 95%CI: 3.088-10.482; P=0.000), high low-density lipoprotein cholesterol (LDL-C; OR=1.021, 95%CI: 1.010-1.031, P=0.000), and high fasting blood glucose(FBG; OR=1.015, 95%CI: 1.006-1.025, P=0.002) were risk factors for recurrent ischemic cerebrovascular events in patients with chronic symptomatic severe intracranial vertebrobasilar artery stenosis. The Nomogram model achieved a concordance C-index of 0.805 (95%CI: 0.752-0.854). The Hosmer-Lemeshow goodness-of-fit test demonstrated a favorable fit (P=0.138), with the calibration curve closely aligning with the actual results. A mobile prediction platform was created at https://cossoss.github.io/calc, and this platform demonstrated good predictive ability. Conclusions: High NIHSS score at admission, poor collateral compensation, high LDL-C, and high FBG were the risk factors for recurrent ischemic cerebrovascular events in severe intracranial vertebrobasilar artery stenosis patients. The Nomogram model established based on the identified risk factors had good prediction value, and the created mobile prediction platform further enhances the practical value of the prediction model.

Key words: Atherosclerosis, Vertebral artery, Basilar artery, Brain ischemia, Recurrence, Logistic models, Nomograms

摘要：

目的: 探讨慢性症状性颅内椎基底动脉重度狭窄患者缺血性脑血管事件复发的危险因素并构建风险预测列线图（Nomogram）模型。方法: 纳入天津市环湖医院2017年1月至2025年3月收治的387例慢性症状性颅内椎基底动脉重度狭窄患者，根据发病90 d内缺血性脑血管事件复发情况分为复发组（97例）和未复发组（290例），单因素和多因素Logistic回归分析筛查发病90 d内缺血性脑血管事件复发的危险因素，并据此构建Nomogram模型，经内部验证后采用JavaScript等网络编程语言将该模型开发转化，创建国内外首个慢性症状性颅内椎基底动脉重度狭窄患者缺血性脑血管事件复发的移动端预测平台。结果: Logistic回归分析显示，入院时高美国国立卫生研究院卒中量表（NIHSS）评分（OR=1.452，95%CI：1.245~1.693；P=0.000）、侧支循环差（OR=6.468，95%CI：3.311~12.633；P=0.000）、高低密度脂蛋白胆固醇（LDL-C）水平（OR=1.020，95%CI：1.008~1.032；P=0.001）和高空腹血糖（OR=1.016，95%CI：1.005~1.027；P=0.003）是慢性症状性颅内椎基底动脉重度狭窄患者缺血性脑血管事件复发的危险因素。据此构建的Nomogram模型一致性指数为0.805（95%CI：0.752~0.854），Hosmer-Lemeshow拟合优度检验具有较好的拟合度（P=0.138），校准曲线与真实结果高度一致；创建的移动端预测平台（https://cossoss.github.io/calc）具有良好的预测能力。结论: 入院时高NIHSS评分、侧支循环差、高LDL-C和高空腹血糖的慢性症状性颅内椎基底动脉重度狭窄患者更易复发缺血性脑血管事件，据此构建的Nomogram模型具有较好的预测效能，创建的移动端预测平台进一步提高Nomogram模型的效能。

关键词: 动脉粥样硬化, 椎动脉, 基底动脉, 脑缺血, 复发, Logistic模型, 列线图

CLC Number: 

• R743