National University Heart Centre, Singapore, Singapore.
Am Heart J. 2011 Aug;162(2):291-9. doi: 10.1016/j.ahj.2011.05.016. Epub 2011 Jul 18.
Acute myocardial infarction (AMI) is a leading cause of mortality in Asia. However, quantitative risk scores to predict mortality after AMI were developed without the participation of Asian countries.
We evaluated the performance of the Global Registry of Acute Coronary Events (GRACE) in-hospital mortality risk score, directly and after recalibration, in a large Singaporean cohort representing 3 major Asian ethnicities.
The GRACE cohort included 11,389 patients, predominantly of European descent, hospitalized for AMI or unstable angina from 2002 to 2003. The Singapore cohort included 10,100 Chinese, 3,005 Malay, and 2,046 Indian patients hospitalized for AMI from 2002 to 2005.Using the original GRACE score, predicted in-hospital mortality was 2.4% (Chinese), 2.0% (Malay), and 1.6% (Indian). However, observed in-hospital mortality was much greater at 9.8% (Chinese), 7.6% (Malay), and 6.4% (Indian). The c statistic for Chinese, Malays, and Indians was 0.86, 0.86, and 0.84, respectively, and the Hosmer-Lemeshow statistic was 250, 56, and 41, respectively. Recalibration of the GRACE score, using the mean-centered constants derived from the Singapore cohort, did not change the c statistic but substantially improved the Hosmer-Lemeshow statistic to 90, 24, and 18, respectively. The recalibrated GRACE score predicted in-hospital mortality as follows: 7.7% (Chinese), 6.0% (Malay), and 5.2% (Indian).
In this large cohort of 3 major Asian ethnicities, the original GRACE score, derived from populations outside Asia, underestimated in-hospital mortality after AMI. Recalibration improved risk estimation substantially and may help adapt externally developed risk scores for local practice.
急性心肌梗死（AMI）是亚洲地区死亡的主要原因。然而，用于预测 AMI 后死亡率的定量风险评分是在没有亚洲国家参与的情况下制定的。
我们评估了全球急性冠状动脉事件登记处（GRACE）住院期间死亡率风险评分在一个代表 3 大亚洲族群的大型新加坡队列中的表现，该队列直接进行了评分，也进行了重新校准。
GRACE 队列纳入了 2002 年至 2003 年期间因 AMI 或不稳定型心绞痛住院的 11389 例主要为欧洲血统的患者。新加坡队列纳入了 2002 年至 2005 年期间因 AMI 住院的 10100 例华人、3005 例马来人和 2046 例印度人。使用原始 GRACE 评分，预测住院期间死亡率为 2.4%（华人）、2.0%（马来人）和 1.6%（印度人）。然而，实际住院期间死亡率要高得多，分别为 9.8%（华人）、7.6%（马来人）和 6.4%（印度人）。华人、马来人和印度人的 C 统计量分别为 0.86、0.86 和 0.84，Hosmer-Lemeshow 统计量分别为 250、56 和 41。使用源自新加坡队列的平均值中心常数对 GRACE 评分进行重新校准并没有改变 C 统计量，但显著改善了 Hosmer-Lemeshow 统计量，分别为 90、24 和 18。重新校准的 GRACE 评分预测住院期间死亡率如下：7.7%（华人）、6.0%（马来人）和 5.2%（印度人）。
在这个由 3 大亚洲族群组成的大型队列中，源自亚洲以外人群的原始 GRACE 评分低估了 AMI 后的住院期间死亡率。重新校准大大提高了风险估计的准确性，可能有助于为当地实践调整外部开发的风险评分。
