Wang Chong, Yao Feng, Han Lin, Zhu Ji, Xu Zhi-yun
Department of Cardiothoracic Surgery, Changhai Hospital, Second Military Medical University, Shanghai, People's Republic of China.
J Heart Valve Dis. 2010 Jan;19(1):21-7.
The study aim was to assess the performance of the European System for Cardiac Operative Risk Evaluation (EuroSCORE) model in Chinese patients undergoing heart valve surgery.
Between January 2003 and December 2007, the data from a total of 1726 consecutive patients who underwent heart valve surgery at the authors' center were collected and scored according to the additive and logistic EuroSCORE models. The patients were allocated to three risk subgroups, and the entire cohort and each risk subgroup analyzed. Calibration of the EuroSCORE model was assessed by the Hosmer-Lemeshow (H-L) test. Discrimination was tested by calculating the area under the receiver operating characteristic (ROC) curve.
Completed data from all 1726 patients were analyzed. There were significant differences in the prevalence of risk factors between the study sample and European cardiac surgery populations. The observed mortality was 4.46% overall, compared to 3.51% (additive) and 2.85% (logistic). The additive EuroSCORE model showed good calibration in predicting in-hospital mortality (H-L; p = 0.204), but the logistic EuroSCORE model underpredicted observed mortality (H-L; p = 0.038) in the entire cohort. Both, the additive and logistic EuroSCORE models showed good calibration in predicting in-hospital mortality in the medium- and high-risk subgroups, but overpredicted observed mortality in the low-risk subgroup. The discriminative power of both models for the entire cohort was poor (areas under the ROC curve of 0.644 and 0.647 for the additive and logistic models, respectively).
The additive and logistic EuroSCORE models gave an imprecise prediction for individual operative risk in heart valve surgery patients at the authors' center; thus, use of the EuroSCORE models for risk evaluation at this center may be unsuitable in the future. It will be necessary to re-examine the performance of the EuroSCORE model for predicting operative mortality in heart valve surgery on a multicenter database in China.
本研究旨在评估欧洲心脏手术风险评估系统（EuroSCORE）模型在中国接受心脏瓣膜手术患者中的表现。
2003年1月至2007年12月期间，收集了作者所在中心连续1726例接受心脏瓣膜手术患者的数据，并根据累加式和逻辑EuroSCORE模型进行评分。将患者分为三个风险亚组，并对整个队列和每个风险亚组进行分析。通过Hosmer-Lemeshow（H-L）检验评估EuroSCORE模型的校准情况。通过计算受试者工作特征（ROC）曲线下面积来测试辨别力。
对全部1726例患者的完整数据进行了分析。研究样本与欧洲心脏手术人群的风险因素患病率存在显著差异。总体观察到的死亡率为4.46%，而累加式模型预测的死亡率为3.51%，逻辑模型预测的死亡率为2.85%。累加式EuroSCORE模型在预测住院死亡率方面显示出良好的校准（H-L；p = 0.204），但逻辑EuroSCORE模型在整个队列中对观察到的死亡率预测偏低（H-L；p = 0.038）。累加式和逻辑EuroSCORE模型在预测中、高风险亚组的住院死亡率方面均显示出良好的校准，但在低风险亚组中对观察到的死亡率预测偏高。两个模型对整个队列的辨别力均较差（累加式和逻辑模型的ROC曲线下面积分别为0.644和0.647）。
累加式和逻辑EuroSCORE模型对作者所在中心心脏瓣膜手术患者的个体手术风险预测不准确；因此，未来在该中心使用EuroSCORE模型进行风险评估可能不合适。有必要在中国的多中心数据库上重新审视EuroSCORE模型在预测心脏瓣膜手术手术死亡率方面的表现。
