2026 年第 1 期 第 21 卷

阿替普酶静脉溶栓联合替罗非班预防急性脑梗死进展的效果及进展影响因素的随机森林模型分析

Efficacy of alteplase intravenous thrombolysis combined with tirofiban in preventing acute cerebral infarction progression and random forest model analysis of progression influencing factors

作者：张森林徐锋平林晓丽

英文作者：Zhang Senlin Xu Fengping Lin Xiaoli

单位：福建医科大学附属南平第一医院神经内科，南平353000

英文单位：Department of Neurology Nanping First Hospital Affiliated to Fujian Medical University Nanping 353000 China

关键词：急性脑梗死；阿替普酶；替罗非班；进展性脑梗死；随机森林模型

英文关键词：Acutecerebralinfarction;Alteplase;Tirofiban;Progressivecerebralinfarction;Randomforestmodel

• 摘要：

• 目的 探讨阿替普酶静脉溶栓联合替罗非班预防急性脑梗死(ACI)转变为急性进展性脑梗死(PCI)的效果，并构建预测PCI发生风险的随机森林模型。方法 选取2022年1月至2024年12月福建医科大学附属南平第一医院收治的106例ACI患者，依据治疗方案的不同分为对照组（阿替普酶静脉溶栓治疗，26例）和观察组（阿替普酶静脉溶栓联合替罗非班治疗，80例）。比较2组PCI发生率、美国国立卫生研究院卒中量表(NIHSS)评分、凝血指标及不良反应。进一步将观察组分为进展组与未进展组，比较其临床资料与实验室指标，采用Logistic回归分析筛选PCI危险因素，并建立随机森林模型预测PCI风险。结果 对照组26例患者中治疗后转变为PCI者17例(65.4%)，观察组80例患者中治疗后转变为PCI者34例(42.5%)，组间比较差异有统计学意义(P=0.042)。观察组治疗后NIHSS评分低于对照组(P＜0.05)。治疗后，观察组凝血酶时间、活化部分凝血活酶时间、凝血酶原时间均长于对照组(均P＜0.05)。多因素Logistic回归分析结果显示，糖尿病(比值比=20.530，P=0.004)、入院时NIHSS评分(比值比=1.368，P=0.018)、入院时收缩压(比值比=1.053，P=0.041)、纤维蛋白原(FIB)(比值比=29.931，P＜0.001)、血清淀粉样蛋白A(SAA)(比值比=1.116，P=0.033)是影响PCI发生的因素。随机森林模型中，因素重要性排序为FIB、SAA、收缩压、NIHSS评分、糖尿病。随机森林模型AUC为0.957，敏感度94.12%，特异度86.96%，优于Logistic模型（AUC=0.909，P=0.042）。结论 阿替普酶联合替罗非班可有效降低PCI风险，改善神经功能及凝血状态，且不增加不良反应。糖尿病、高NIHSS评分、高收缩压、高FIB及高SAA是PCI的独立危险因素。随机森林模型具有更高预测效能，有助于早期识别高危患者并指导个体化干预。

• Objective To explore the efficacy of alteplase intravenous thrombolysis combined with tirofiban in preventing acute cerebral infarction (ACI) from progressing to progressive cerebral infarction (PCI), and to construct a random forest model for predicting the risk of PCI. Methods A total of 106 ACI patients admitted to Nanping First Hospital Affiliated to Fujian Medical University from January 2022 to December 2024 were selected and divided into control group (alteplase intravenous thrombolysis alone, 26 cases) and observation group (alteplase intravenous thrombolysis combined with tirofiban, 80 cases) according to different treatment regimens. The incidence of PCI, national institutes of health stroke scale (NIHSS) scores, coagulation indicators, and adverse reactions were compared between the two groups. Furthermore, the observation group was divided into progression group and non-progression group. Clinical data and laboratory indicators were compared between the two subgroups. Logistic regression analysis was used to screen the risk factors for PCI, and a random forest model was established to predict the risk of PCI. Results Among the 26 patients in the control group, 17 patients (65.4%) progressed to PCI after treatment, while 34 of the 80 patients (42.5%) in the observation group developed PCI, with a statistically significant difference between the two groups (P=0.042). The post-treatment NIHSS score in the observation group was lower than that in the control group (P<0.05). After treatment, thrombin time, activated partial thromboplastin time, and prothrombin time in the observation group were all longer than those in the control group (all P<0.05). Multivariate Logistic regression analysis showed that diabetes mellitus (odds ratio=20.530, P=0.004), admission NIHSS score (odds ratio=1.368, P=0.018), admission systolic blood pressure (odds ratio=1.053, P=0.041), fibrinogen (FIB)(odds ratio=29.931, P<0.001), and serum amyloid A (SAA)(odds ratio=1.116, P=0.033) were influencing factors for PCI. In the random forest model, the order of factor importance was FIB, SAA, systolic blood pressure, NIHSS score, and diabetes mellitus. The area under the curve (AUC) of the random forest model was 0.957, with a sensitivity of 94.12% and specificity of 86.96%, which was superior to the Logistic model (AUC=0.909, P=0.042). Conclusion  Alteplase combined with tirofiban can effectively reduce the risk of PCI, improve neurological function and coagulation status without increasing adverse reactions. Diabetes mellitus, high NIHSS score, high systolic blood pressure, elevated FIB, and high SAA are independent risk factors for PCI. The random forest model exhibits higher predictive performance, which is conducive to early identification of high-risk patients and guidance of individualized interventions.