Should Patients With Acute Coronary Disease Be Stratified for Management According to Their Risk? Derivation, External Validation, and Outcomes Using the Updated GRACE Risk Score

Study Questions:

What is the performance of a simplified Global Registry of Acute Coronary Events (GRACE) risk predictor (GRACE score 2.0), which incorporates substitutions for creatinine and Killip class, and which incorporates testing for nonlinear associations between outcomes and continuous variables?


This was a retrospective analysis of the GRACE registry, an unbiased population of patients with acute coronary syndrome (ACS) undertaken over 10 years in 94 hospitals and 14 countries. The original GRACE risk score was derived based on independent predictors of outcome (ST-segment deviation, age, heart rate, systolic blood pressure, creatinine, Killip class, cardiac arrest at admission, and elevated biomarkers of necrosis) and estimated hospital risk of death or the combination of death or myocardial infarction (MI) and the same outcomes up to 6 months postdischarge. The authors developed a simplified version of the risk score with substitutions for creatinine and Killip class with history of renal dysfunction and diuretic usage, respectively. To test for possible nonlinear associations between age, creatinine, pulse, and systolic blood pressure, the authors used the method of restricted cubic splines. The updated GRACE risk score was validated in the French Registry of Acute ST-Elevation and non-ST-elevation Myocardial Infarction (FAST-MI).


Nonlinear associations for the 1-year mortality model were found in all four continuous measures: systolic blood pressure, pulse, age, and creatinine (p < 0.001 vs. linear). The use of nonlinear algorithms improved model discrimination, which was validated externally. Using the FAST-MI 2005 of 2,959 patients, the c-statistics for death exceeded 0.82 for the overall population at 1 and 3 years. The c-indices in the simplified model (using substitutions for Killip class and serum creatinine) were 0.82 at both 1 and 3 years.


The use of nonlinear functions for continuous variables (systolic blood pressure, pulse, age, and creatinine) improves the performance of the original GRACE risk score. A simplified risk score (in which renal dysfunction is substituted for creatinine and diuretic use is substituted for Killip class) performs nearly as well.


The authors report an important update to the GRACE risk score. By investigating nonlinear associations between outcomes and continuous variables, the authors report improved model discrimination. The updated GRACE risk score offers simplicity and lends itself to bedside application and use with mobile devices by use of renal dysfunction (instead of creatinine, which is frequently not available at the time of hospitalization) and diuretic use (instead of Killip class, which is not universally used).

Clinical Topics: Acute Coronary Syndromes, Arrhythmias and Clinical EP, Implantable Devices, SCD/Ventricular Arrhythmias

Keywords: Myocardial Infarction, Acute Coronary Syndrome, Coronary Disease, Blood Pressure, Heart Arrest, Creatinine, Heart Rate, Hospitalization

< Back to Listings