下肢动脉硬化闭塞症患者术后严重不良心血管事件的风险预测

Constructing predictive modelling for the risk of serious adverse cardiovascular events in postoperative patients of symptomatic arteriosclerosis obliterans

目的构建有症状的下肢动脉硬化闭塞症(ASO)患者术后发生严重不良心血管事件(MACE)的风险预测模型。方法选取2018年1月1日至2021年12月31日南京鼓楼医院收治的ASO患者957例,依照是否发生MACE分为MACE组和非MACE组。多因素COX回归分析构建风险预测模型,采用受试者工作特征(ROC)曲线对模型预测效能进行评估,用校准曲线评价模型拟合效果,Bootstrap法进行内部验证。结果发生MACE的患者143例(14.94%)。COX回归分析后筛选出BMI、肌酐清除率、纤维蛋白原、利伐沙班和既往血运重建手术史以构建模型。ROC曲线评估模型的C统计量为0.690(95%CI:0.644~0.736),灵敏度和特异性分别为49.2%和80.7%,约登指数为0.299,最佳截断值为0.086;校准曲线显示预测值与实际值一致;内部验证的C统计量为0.689(95%CI:0.672~0.700);根据最佳截断值将人群分为高风险和低风险组,生存分析显示两组差异有统计学意义。结论基于临床参数所构建的MACE发生的风险预测模型具有较好的预测性和良好判别能力,可为ASO患者发生MACE的评估及诊疗提供参考。

Objective To construct a predictive model for the risk of major adverse cardiovascular events(MACE)after surgery in patients with symptomatic arteriosclerosis obliterans(ASO).Methods From Jan 2018 to Dec 2021,957 patients with symptomatic ASO admitted to Nanjing Drum Tower Hospital were selected and divided into MACE and non-MACE groups according to whether they had a post-op MACE.A risk prediction model was constructed based on a stepwise regression method with multi-factor COX regression analysis.The model was evaluated using the receiver operating characteristic curve(ROC),the calibration curve to assess the model fit,and the Bootstrap method for internal validation.Results MACE occurred in 143 patients(14.94%).After COX regression analysis,BMI,creatinine clearance,fibrinogen,rivaroxaban and previous history of surgery were enrolled into model constructing.The ROC curve assessed the model with a C-statistic of 0.690(95%CI:0.644-0.736),sensitivity and specificity of 49.2%and 80.7%respectively,a Jorden index of 0.299 and an optimal cut-off value of 0.086.Calibration curves showing agreement between predicted and actual observed values.Internally validated C-statistic of 0.689(95%CI:0.672-0.700).The population was divided into high and low risk groups based on the best cut-off value and analysed for survival.The difference between the two groups was statistically different.Conclusion The risk prediction model for the occurrence of MACE based on clinical parameters is simple and convenient,with good predictability and good discriminatory ability,and can provide reference for the assessment and treatment of MACE in ASO patients.