Background: Risk prediction with the GRACE risk model is guideline-recommended clinical practice in patients with acute coronary syndrome (ACS). However, the performance of other contemporary risk models derived from the National Cardiovascular Data Registry (NCDR) and the Acute Coronary Treatment and Intervention Outcomes Network (ACTON). We aimed to compare these modern risk models with the gold-standard GRACE for prediction of in-hospital mortality. 

Methods and Results: In this retrospective analysis, a total of 1,567 patients presenting with Non-ST-segment elevation myocardial infarction (NSTEMI, 1,002 patients) and ST-segment elevation myocardial infarction (STEMI, 565 patients) undergoing coronary angiography were included. Overall in-hospital mortality was 7.5% (NSTEMI 3.7%, STEMI 14.5%). Parameters for risk model calculation and in-hospital mortality were assessed, NCDR, ACTION and GRACE risk scores were then calculated and risk model performance was compared. The NCDR model performed best in discrimination of risk in ACS overall (c-index 0.89, 95% CI 0.87-0.92; P_ACTIONvsNCDR <0.0001; P_GRACEvsNCDR<0.0001) and in NSTEMI and STEMI subgroups (P_ACTIONvsNCDR and P_GRACEvsNCDR both <0.02). Comparison of ACTION and GRACE risk models showed advantages for ACTION in NSTEMI patients (c-index 0.87 vs. 0.84 (GRACE); P_{GRACEvsACTION}=0.02). ACTION showed the most accurate calibration overall. 

Conclusion: While GRACE remains the sole tool for long-term risk prediction, the NCDR and ACTION risk scores show superior performance for prediction of in-hospital mortality in a contemporary German patient cohort with acute coronary syndrome.