不同评分模型对维吾尔族患者冠状动脉旁路移植术后死亡风险的预测价值

Validation of different cardiac operative risk evaluation models in coronary artery bypass grafting surgeryof Uyghur patients

目的评价欧洲心脏手术风险评估系统(EuroSCORE)additive、logistic评分、EuroSCOREⅡ模型及心脏麻醉风险评分(CARE)对于预测维吾尔族患者冠状动脉旁路移植(CABG)手术在院死亡率的准确性。方法选择2012年9月至2013年6月行CABG的维吾尔族患者61例,术前分别使用EuroSCORE additive评分、logistic评分、EuroSCOREⅡ模型与CARE评分对患者手术死亡风险进行评估,术后比较各种评分模型的预测死亡率与实际在院死亡率。模型预测的校准度采用拟合优度检验,预测的鉴别度采用受试者工作特征(ROC)曲线下面积检验。结果对于实施CABG的维吾尔族患者,EuroSCORE additive评分、logistic评分及CARE评分预测鉴别度均较好,CARE评分最优(ROC曲线下面积分别为0.737、0.754、0.864),EuroSCOREⅡ模型的鉴别度较差(ROC曲线下面积为0.682)。结论EuroSCORE模型中的additive评分、logistic评分以及CARE评分对维吾尔族CABG术患者均表现出了较好的鉴别力,针对维吾尔族CABG手术患者,可以优先采用CARE评分进行风险预测。最新的EuroSCOREⅡ模型不适合于维吾尔族患者,在临床应用中应谨慎。

Objective To assess the accuracy of the European System for Cardiac Operative Risk Evaluation（EuroSCORE） additive model, logistic model, EuroSCORE model Ⅱand CARE in predicting the in-hospital mortality of Uyghur patients undergoing coronary artery bypass grafting surgery. Methods We collected the clinical data of sixty-one Uyghur patients who underwent coronary artery bypass grafting surgery from September 2012 to June 2013, predicted the inospital mortality with additive EuroSCORE model, logistic EuroSCORE model, EuroSCORE model Ⅱ and CARE respectively preoperative, the actual and predicted mortality of each model＇s were studied and coinpared. Calibration of models were assessed by the test of goodness of fit. Discrimination was tested by calculating the area under the receiver operating characteristic （ROC） curve. Results For Uyghur patients, discrimination of additive EuroSCORE model, logistic EuroSCORE model and CARE were all good（the area under the ROC curve were 0. 737, 0. 737, 0. 864）, compared with EuroSCORE model Ⅱ （0. 682）. Conclusion The additive EuroSCORE model, logistic EuroSCORE model and CARE showed good discrimination for Uyghur patients undergoing coronary artery bypass grafting surgery, among them, CARE is the best. For Uyghur patients undergoing coronary artery bypass grafting surgery, CARE should be prior to use for risk prediction. The new EuroSCORE model Ⅱ is not suitable for Uyghur patients, this differences should be considered in clinical practicing.