Predictive Power of In-Hospital and Long-Term Mortality of the GRACE, TIMI, Revised CADILLAC and PAMI Score in NSTEMI Patients with Diabetes - Data from TSOC ACS-DM Registry.

Background Risk score is widely used in non-ST segment elevation myocardial infarction (NSTEMI) patients to predict the in-hospital outcome for immediate coronary angiography decision and care of unit selection. Objectives This study compared the performances of the thrombolysis in myocardial infarction (TIMI), Global Registry of Acute Coronary Events (GRACE), Primary Angioplasty in Myocardial Infarction (PAMI), and Revised Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications (Revised CADILLAC) risk scores in predicting in-hospital and long-term outcomes in diabetic NSTEMI patients. Methods A total of 750 diabetic NSTEMI patients from 27 hospitals were enrolled between January 2013 and December 2015 in the nationwide registry initiated by the Taiwan Society of Cardiology. Four score systems were calculated with receiver operator characteristic analysis used to compare outcome discrimination performance. Results No studied risk scores reached acceptable discrimination per area under curve (AUC) in the prediction of in-hospital outcome except for the revised CADILLAC score which reached acceptable discrimination in new-onset cardiogenic shock (AUC = 0.7191) and acute renal failure (AUC = 0.7283). In long-term outcomes, only the revised CADILLAC score reached acceptable discrimination of mortality prediction at 6, 12 and 24 months (AUC = 0.7261 at 6 months, 0.7319 at 12 months, and 0.7256 at 24 months). Subgroup analysis based on the revised CADILLAC score risk class showed a significant difference in adjusted mortality rate between low-risk group/intermediate-risk group and high-risk group. Conclusions Only the revised CADILLAC score showed acceptable accuracy to predict the long-term mortality outcome among the scores studied.