2021 年第 5 期 第 16 卷

急性心肌梗死患者发生室壁瘤或左心室血栓的风险预测模型构建及验证

Construction and validation of a risk prediction model for ventricular aneurysm or left ventricular thrombosis in patients with acute myocardial infarction

作者：刘永国朱伟李安娜晁鹏

英文作者：Liu Yongguo Zhu Wei Li Anna Chao Peng

单位：新疆维吾尔自治区人民医院心血管内科，乌鲁木齐830001

英文单位：Department of Cardiovascular Medicine People′s Hospital of Xinjiang Uygur Autonomous Region Urumqi 830001 China

关键词：

英文关键词：Acutemyocardialinfarction;Ventricularaneurysm;Leftventricularthrombosis;Predictionmodel

• 摘要：

• 目的 探究急性心肌梗死患者发生室壁瘤或左心室血栓（LVT）的风险预测模型构建及验证。方法 选取2014年1月至2017年12月于新疆维吾尔自治区人民医院诊断为急性心肌梗死的2 982例患者作为研究对象,其中310例不能配合研究及随访、139例失访或资料不全，最终纳入2 533例。根据随访期间是否发生室壁瘤或LVT，将其分为无室壁瘤或LVT组（2 348例）和室壁瘤或LVT组（185例）。记录患者的临床资料。随访1年，记录随访期间患者主要终点事件（室壁瘤或LVT）发生情况。采用单因素和多因素Cox回归模型分析急性心肌梗死患者发生室壁瘤或LVT的危险因素。采用列线图方法构建急性心肌梗死患者发生室壁瘤或LVT的风险预测模型，并验证该模型的准确性。结果 室壁瘤或LVT组女性、体重指数≤25 kg/m2、收缩压降低≥30 mmHg（1 mmHg=0.133 kPa）、左心室射血分数（LVEF）≤35%、前壁心肌梗死、肌酸激酶同工酶（CK-MB）升高≥正常参考值上限20倍比例均高于无室壁瘤或LVT组［33.5%（62/185）比19.5%（459/2 348）、22.2%（41/185）比7.9%（185/2 348）、22.7%（42/185）比12.3%(288/2 348)、26.5%（49/185）比10.7%（251/2 348）、17.8%（33/185）比8.6%（202/2 348）、43.2%（80/185）比22.4%（526/2 348）］，差异均有统计学意义（均P＜0.001）。单因素和多因素Cox回归分析结果显示，女性、脑卒中史、体重指数≤25 kg/m2、收缩压降低≥30 mmHg、LVEF≤35%、前壁心肌梗死、CK-MB升高≥正常参考值上限20倍均为急性心肌梗死患者发生室壁瘤或LVT的危险因素（均P＜0.001）。将上述指标构建列线图，获得各项危险因素的得分：女性，61分；脑卒中史，60分；体重指数≤25 kg/m2，80分；收缩压下降≥30 mmHg，58分；LVEF≤35%，100分；CK-MB升高≥正常参考值上限20倍，61分。受试者工作特征曲线分析预测模型的曲线下面积（AUC）为0.709，C指数为0.723。2 533例患者中随机抽取25%作为验证组（633例，其中发生室壁瘤或LVT者31例），其余患者作为模型组（1 900例，其中发生室壁瘤或LVT者154例），ROC曲线分析结果显示，模型组AUC为0.694，验证组AUC为0.739。模型校准曲线与实际曲线重合良好。结论 风险预测模型显示女性、脑卒中史、体重指数≤25 kg/m2、收缩压降低≥30 mmHg、LVEF≤35%、CK-MB升高≥正常参考值上限20倍是急性心肌梗死患者发生室壁瘤或LVT的危险因素。本研究建立的预测模型准确性较高。

• Objective To explore the construction and validation of a risk prediction model for ventricular aneurysm or left ventricular thrombosis(LVT) in patients with acute myocardial infarction(AMI). Methods From January 2014 to December 2017, 2 982 patients diagnosed as AMI in People′s Hospital of Xinjiang Uygur Autonomous Region were enrolled. Among them, 310 cases were unable to cooperate with the study and follow-up, 139 cases were lost to follow-up or incomplete data, and finally 2 533 cases were included. According to the occurrence of ventricular aneurysm or LVT during follow-up, they were divided into non ventricular aneurysm or LVT group(2 348 cases) and ventricular aneurysm or LVT group(185 cases). The clinical data of patients were recorded. All patients were followed-up for 1 year, and the primary endpoint events (ventricular aneurysm or LVT) were recorded. Univariate and multivariate Cox regression model were used to analyze the risk factors of ventricular aneurysm or LVT in patients with AMI. Nomogram method was used to construct a prediction model for ventricular aneurysm or LVT in patients with AMI, and the accuracy of the model was verified. Results The proportions of female, body mass index(BMI)≤25 kg/m2, systolic blood pressure reduction≥30 mmHg, left ventricular ejection fraction(LVEF)≤35%, anterior myocardial infarction and creatine kinase isoenzyme(CK-MB) increase≥20 times of upper limit of normal reference value in ventricular aneurysm or LVT group were higher than those in non ventricular aneurysm or LVT group［33.5%（62/185） vs 19.5%（459/2 348）, 22.2%（41/185） vs 7.9%（185/2 348）, 22.7%（42/185） vs 12.3%(288/2 348), 26.5%（49/185） vs 10.7%（251/2 348）, 17.8%（33/185） vs 8.6%（202/2 348）, 43.2%（80/185） vs 22.4%（526/2 348）］ (all P<0.001). Univariate and multivariate Cox regression analysis showed that female, stroke history, BMI≤25 kg/m2, systolic blood pressure reduction≥30 mmHg, LVEF≤35%, anterior myocardial infarction, CK-MB increase≥20 times of upper limit of normal reference value were the risk factors of ventricular aneurysm or LVT in patients with AMI(all P<0.001). The above indicators were constructed into a nomogram, and the results showed that the scores of indicators were 61 for female, 60 for stroke history, 80 for BMI≤25 kg/m2, 58 for systolic blood pressure reduction≥30 mmHg, 100 for LVEF≤35%, and 61 for CK-MB increase≥20 times of upper limit of normal reference value. The receiver operating characteristic(ROC) curve analysis showed that the area under the curve(AUC) of the prediction model was 0.709, and C-index was 0.723. A random sample of 25% of 2 533 patients served as the verification group(633 cases, including 31 cases with ventricular aneurysm or LVT), and the remaining patients served as the model group(1 900 cases, including 154 cases with ventricular aneurysm or LVT). The ROC curve analysis showed that the AUC of model group was 0.694 and the AUC of verification group was 0.739. The model calibration curve coincided well with the actual curve. Conclusions Female, stroke history, BMI≤25 kg/m2, systolic blood pressure reduction ≥30 mmHg, LVEF≤35%, CK-MB increase≥20 times of upper limit of normal reference value are the risk factors of ventricular aneurysm or LVT in patients with AMI. The accuracy of the prediction model established in this study is high.