急性胸痛缺血风险评分模型的建立和验证  

Development and prospective validation of an ischemic risk score for acute chest pain patients with normal high-sensitivity troponin I levels and without obvious ST-segment deviation

作　　者：马春朋[1] 刘晓丽[2] 王庆胜[1] 聂绍平[3] 

机构地区：[1]秦皇岛市第一医院心内科,066000 [2]秦皇岛市第一医院内分泌科,066000 [3]首都医科大学附属北京安贞医院急诊危重症中心

出　　处：《中华心血管病杂志》2018年第2期131-136,共6页Chinese Journal of Cardiology

基　　金：北京市科技计划（Z141107002514014）

摘　　要：目的为高敏肌钙蛋白I水平正常且心电图ST段无明显偏移的中国急性胸痛人群，建立一种缺血风险评分模型并进行验证。 方法入选2014年9月至2015年7月在北京安贞医院急诊胸痛中心就诊的胸痛患者，收集符合纳入标准（包括高敏肌钙蛋白I水平正常且心电图ST段无明显偏移等）患者的基线资料。研究终点为就诊后3个月发生的主要不良心血管事件（MACE），即急性心肌梗死、经皮冠状动脉介入治疗、冠状动脉旁路移植术和全因死亡组成的联合终点。在回顾性队列研究中，通过logistic回归分析确定预测因素并建立急性胸痛缺血风险评分模型。之后进行前瞻性队列研究，验证该模型。 结果在建立模型队列中，就诊的胸痛患者共1 735例，最终入选患者1 030例。多因素logistic回归分析显示，急性胸痛缺血风险评分模型有5个预测因子，包括性别（β=0.88）、胸痛病史（中和重度怀疑冠心病者的β值分别为2.70和3.51）、心电图（β=0.84）、年龄（β=0.51）和≥3个危险因素（β=0.85）。经权重后，确定模型的评分范围为0~13分。模型的受试者工作特征（ROC）曲线下面积为0.75（95%CI 0.72~0.78）。MACE发生率随着评分增加而显著增加（P〈0.01）。Bootstrap内部验证得到的预测因子与由原始数据得出的评分模型一致，其ROC曲线下面积为0.75（95%CI 0.72~0.78）。在验证模型队列中，低危患者（模型评分为0~3分）、中危患者（模型评分为4~7分）和高危患者（模型评分为8~13分）的MACE发生率分别为1.3%（1/77）、19.0%（22/116）和42.2%（122/289），差异有统计学意义（P〈0.01）。 结论本研究为高敏肌钙蛋白I水平正常且心电图ST段无明显偏移的中国急性胸痛人群建立了一种缺血风险评分模型。建立的缺血风险评分模型能够指导该人群的分诊和管理。Objective To develop and prospectively validate a risk score for acute chest pain patients with normal high-sensitivity troponin I （hs-TnI） levels and without obvious ST-segment deviation in China. Methods Chest pain patients admitted to the emergency department of Beijing Anzhen Hospital from September 2014 to July 2015 were enrolled. Baseline characteristics of patients met inclusion criteria including normal hs-TnI levels and without obvious ST-segment deviation were included. The endpoint （major adverse cardiovascular events） was a composite of acute myocardial infarction, percutaneous coronary intervention, coronary artery bypass graft, and all-cause death within 3 months after initial presentation. Predictors were screened and used to develop the risk score model by logistic regression analysis in a retrospective cohort. Then, the risk score model was evaluated in a prospective validation cohort. Results The study population of derivation cohort included 1 735 consecutive chest pain patients. Finally, 1 030 eligible patients were enrolled. Multivariate regression analysis defined five independent predictors： male gender （β=0.88）; history of chest pain （β value of moderate and high suspicion of coronary heart artery was 2.70 and 3.51 respectively）; electrocardiogram （β=0.84）;p〉60 years old （β=0.51） and 1〉3 risk factors （β= 0.85）.The range of weighted score was set as 0 - 13. The area under a receiver operating characteristic （ROC） curve was 0.75 （95% CI 0.72 - 0.78） in the final model. Major adverse cardiovascular events ratesincreased in proportion to score increase （P〈0.01）. The internal validity used bootstrap technique showed the same predictor factors as the final model, and its area under a ROC curve was 0.75（95%C1 0.72 - 0.78）. MACE rates in the low risk group （score 0 - 3）, intermediate risk group （score 4 - 7）, and high risk group （score 8 - 13） were 1.3% （ 1/77 ）, 19.0% （22/116 ）,and 42.2% （122/289） in the prospective

关 键 词：胸痛 肌钙蛋白 心电描记术 风险评分 

分 类 号：R459.7[医药卫生—急诊医学]