冠心病不稳定型心绞痛PCI术后并发恶性心律失常的风险预测列线图模型研究

Nomogram model for predicting risk of malignant arrhythmia in patients with unstable angina pectoris of coronary heart disease after PCI

目的探讨冠状动脉粥样硬化性心脏病(冠心病)不稳定型心绞痛患者经皮冠状动脉介入治疗(PCI)术后并发恶性心律失常的危险因素,并据此建立风险预测列线图模型。方法回顾性分析2018年1月至2021年12月因冠心病不稳定型心绞痛于郑州市第七人民医院心内三病区接受PCI的486例患者的临床资料,根据术后6个月是否并发恶性心律失常将患者分为并发组(63例)和未并发组(423例)。采用多因素Logistic回归分析冠心病不稳定型心绞痛PCI术后并发恶性心律失常的危险因素,建立风险预测列线图模型。绘制受试者工作特征(ROC)曲线以评估列线图模型的预测效能,并绘制校准曲线,采用Bootstrap自抽样法内部验证预测列线图模型的区分度。结果并发组与未并发组比较,患者年龄、超敏C反应蛋白(hs-CRP)、吸烟史、术前心肌梗死溶栓治疗(TIMI)血流分级0级、冠状动脉(冠脉)多支病变、心功能Killip分级Ⅲ~Ⅳ级、并发糖尿病、低血钾、PCI术后无/慢复流占比均高于未并发组(P<0.05),发病至PCI时间≥6 h占比大于未并发组(P<0.05);两组患者的体质指数(BMI)、性别和病程、并发高血压、并发慢性阻塞性肺疾病占比比较均无统计学意义差异(P>0.05);多因素Logistic回归分析结果显示年龄、吸烟史、发病至PCI时间≥6 h、术前TIMI血流分级0级、冠脉多支病变、心功能Killip分级Ⅲ~Ⅳ级、并发糖尿病、hs-CRP、低血钾、PCI术后无/慢复流均是冠心病不稳定型心绞痛PCI术后并发恶性心律失常的危险因素(P<0.05);将上述危险因素作为预测指标,构建冠心病不稳定型心绞痛PCI术后并发恶性心律失常的风险预测列线图模型,ROC曲线分析结果显示,列线图预测冠心病不稳定型心绞痛并发恶性心律失常的曲线下面积(AUC)为0.794(95%CI:0.755~0.829,P<0.05),灵敏度为73.02%,特异度为86.05%;Bootstrap自抽样法内部检验的一致性指数为0.796,该列线图模型区分度良好,校准曲线与标准曲线贴合较好。结论年龄、吸烟史、发病至PCI时间≥6 h、术前TIMI血流分级0级、冠脉多支病变、Killip分级Ⅲ~Ⅳ级、并发糖尿病、hs-CRP、低血钾、PCI术后无/慢复流均是冠心病不稳定型心绞痛PCI术后并发恶性心律失常的危险因素,据此构建的列线图模型对指导临床筛选高风险患者以预防患者术后恶性心律失常的发生具有重要意义。

Objective To discuss the risk factors of malignant arrhythmia in patients with unstable angina pectoris(UAP)of coronary heart disease(CHD)after percutaneous coronary intervention(PCI),and establish a nomogram model for risk prediction.Methods The clinical materials were retrospectively analyzed from UAP patients undergone PCI(n=486)hospitalized in the Third Department of Cardiology in the Seventh People’s Hospital of Zhengzhou City from Jan.2018 to Dec.2021.All patients were divided,according to whether malignant arrhythmia occurred or not,into occurred group(n=63)and non-occurred group(n=423).The risk factors of UAP of CHD complicated by malignant arrhythmia after PCI were analyzed by using multi-factor Logistic regression analysis,and a nomogram model for risk prediction was establish.The predictive efficacy was reviewed through drawing ROC curve,and a calibration curve was drawn.The discrimination of the predictive nomograph model was internally verified by applying Bootstrap self-sampling.Results Age,high sensitivity C-reactive protein(hs-CRP),and percentages of smoking history,grade 0 of preoperative TIMI blood flow grading,multi-vessel lesions of coronary artery,gradesⅢ-Ⅳof Killip classification,complicated by diabetes,hypokalemia and no/slow reflow after PCI were higher,and percentage of duration≥6 h from onset to PCI was higher in occurred group than those in non-occurred group(P<0.05).The differences in body mass index(BMI),gender,disease course and percentages of complicated by hypertension and chronic obstructive pulmonary disease(COPD)had no statistical significance between 2 groups(P>0.05).The results of multi-factor Logistic regression analysis showed that age,smoking history,duration≥6 h from onset to PCI,grade 0 of preoperative TIMI blood flow grading,multi-vessel lesions of coronary artery,grades III-IV of Killip classification,complicated by diabetes,hs-CRP,hypokalemia and no/slow reflow after PCI were all risk factors of UAP of CHD complicated by malignant arrhythmia(P<0.05).Taken above risk factors as predictive indexes,a nomogram model was established for predicting risk of UAP of CHD complicated by malignant arrhythmia.The results of ROC curve analysis showed that AUC of the nomogram model was 0.794(95%CI:0.755~0.829,P<0.05),sensitivity was 73.02%and specificity was 86.05%in predicting UAP of CHD complicated by malignant arrhythmia.The consistency index of Bootstrap self-sampling internal inspection was 0.796,and the nomogram model had good discrimination and the calibration curve fitted well with the standard curve.Conclusion Age,smoking history,duration≥6 h from onset to PCI,grade 0 of preoperative TIMI blood flow grading,multi-vessel lesions of coronary artery,grades III-IV of Killip classification,complicated by diabetes,hs-CRP,hypokalemia and no/slow reflow after PCI are risk factors of UAP of CHD complicated by malignant arrhythmia.The nomogram model established based on those is of great significance to guide clinical screening of high-risk patients to prevent postoperative malignant arrhythmia.