急性心肌梗死合并心房颤动患者1年发生缺血性脑卒中风险预测模型的构建及评价

Construction and evaluation of a 1-year risk prediction model for ischemic stroke in patients with acute myocardial infarction and atrial fibrillation

摘要:

目的： 探讨急性心肌梗死合并房颤患者1年发生缺血性卒中的独立危险因素，并构建风险预测模型。方法： 回顾性纳入陆军军医大学附属新桥医院医院心血管内科383例诊断为急性心肌梗死合并房颤的患者，电话随访1年内是否发生缺血性卒中事件，按照随访的先后事件顺序以7∶3进行分组，建模组268例，验证组115例。根据建模组患者是否发生缺血性卒中事件分为发生卒中组及未发生卒中组，比较2组患者危险因素并建立风险预测模型，验证组115例患者验证该缺血性脑卒中风险预测模型的效能。结果： 高龄（年龄>65岁）、吸烟、饮酒、高血压、D-二聚体、低密度脂蛋白、左房内径、中性粒细胞/淋巴细胞（neutrophils-lymphocytes ratio，NLR）、C反应蛋白/白蛋白（C-reactive protein-albumin ratio，CAR）提高了急性心肌梗死合并房颤患者发生缺血性脑卒中的风险。本研究最终建立的模型为P=e^x/（1+e^x），其中e为自然对数，X=-22.547+0.044×年龄+0.653×吸烟+1.277×饮酒+1.134×饮高血压+0.867×D-二聚体+0.702×低密度脂蛋白+0.297×左房内径+0.421×NLR+0.665×CAR。建模组的ROC曲线面积为0.866（95% CI=0.820~0.911，P<0.001），敏感性为87.50%，特异性为73.11%。该模型的预测能力处于中等水平，验证组的ROC曲线下面积为0.913（95% CI=0.854~0.973，P<0.001），敏感性为85.71%，特异性为87.36%。结论： 该模型预测能力较好，可用于急性心肌梗死合并房颤患者1年内发生缺血性卒中的筛查，可为及时识别及干预缺血性卒中的发生提供参考依据。

Abstract:

Objective: To investigate the independent risk factors of ischemic stroke in patients with acute myocardial infarction and atrial fibrillation in 1 year, and to construct a risk prediction model. Methods: We retrospectively enrolled 383 patients diagnosed with acute myocardial infarction and atrial fibrillation in the Department of Cardiology of Xinqiao Hospital of Army Medical University. Telephone follow-up for ischemic stroke within 1 year. According to the sequence of follow-up events, they were divided into groups by 7∶3. There were 268 cases in the modeling group and 115 cases in the verification group. According to whether patients in the modeling group had an ischemic stroke, they were divided into a stroke group and a non-stroke group. The risk factors of the two groups were compared and a risk prediction model was established. A total of 115 patients in the verification group verified the efficacy of the ischemic stroke risk prediction model. Results: Advanced age (age>65 years), smoking, drinking, hypertension, D-dimer, low-density lipoprotein, and left atrial diameter, neutrophils-lymphocytes ratio (NLR), and C-reactive protein-albumin ratio (CAR) increased the risk of ischemic stroke in patients with acute myocardial infarction and atrial fibrillation. The final model established in this study was P=e^x/ (1+e^x), where e was the natural logarithm, X=-22.547+0.044×Age+0.653×Smoking+1.277×Drinking+1.134×Drinking hypertension+0.867×D-dimer+0.702×LDL+0.297×Left atrium ID+0.421×NLR+0.665×CAR. The ROC curve area of the modeling group was 0.866 (95% CI=0.820-0.911, P<0.001), the sensitivity was 87.50%, and the specificity was 73.11%. The predictive power of this model was at a moderate level. The area under the ROC curve of the verification group was 0.913 (95% CI=0.854-0.973, P<0.001), the sensitivity was 85.71%, and the specificity was 87.36%. Conclusion: This model has good predictive ability and can be used for the screening of ischemic stroke in patients with acute myocardial infarction and atrial fibrillation within 1 year, and can provide a reference for timely identification and intervention of ischemic stroke.