Validation of a Troponin I-Adapted Modified GRACE Score : A Cohort Study

A modified Global Registry of Acute Coronary Events (modGRACE) model incorporating continuous rather than binary high-sensitivity troponin T data supersedes the performance of the original GRACE score. However, whether a similar prognostic gain can be achieved using contemporary-sensitive (cTnI) or high-sensitivity troponin I (hs-TnI) remains uncertain.

We analyzed data from 2 independent ACS cohorts of consecutively recruited patients: the Freeman cohort (Newcastle, United Kingdom; n = 2,991) and the Greek Study of ACS (GREECS; Greece; n = 1,630). Risk stratification using the original GRACE score and the modGRACE score incorporating continuous c/hs-TnI values. Model performance was assessed using Harrell's C-index and continuous net reclassification improvement (NRI).

In the Freeman cohort, modGRACE showed improvements in discrimination for in-hospital (Harrell's C-index=0.871 vs 0.777) and 30-days mortality (0.848 vs 0.833), with significant improvement at 1-year (0.721 vs. 0.690, all p<0.05). Reclassification improved for in-hospital (NRI=0.67), 30-day (NRI=0.69), and 1-year mortality (NRI=0.49) (all p< .001). In the GREECS cohort, modGRACE showed improvements for discrimination in 30days-mortality (Harrell's C-index 0.778 vs 0.752) and cardiovascular-death at 20 years (0.682 vs. 0.653, p<0.05 for both); and improved risk reclassification for 30-day mortality (NRI, 0.56), 20-year cardiovascular mortality (NRI=0.85), and 20-year MACE (NRI=0.79) (all p<0.001).

The updated ModGRACE model, incorporating continuous c/hs-TnI alongside re-estimated baseline predictors, demonstrated consistent improvements in risk reclassification and discrimination over original GRACE. Ultimately, these findings support the application of quantitative troponin data across different troponin assays, to improve prognostic accuracy and standardize risk assessment in clinical practice.