Abstract 6091: A Simple TIMI Discharge Risk Score Post ACS Predicts CV Mortality at 2 Years
Improved risk stratification at hospital discharge may help guide better post-ACS management. We used data from 2 lipid-lowering trials post-ACS (PROVE IT-TIMI 22, A2Z) to derive a discharge score for predicting long-term CV mortality and validated the score in an external test set from MERLIN-TIMI 36 (ranolazine in nSTE-ACS). Univariate analyses of data from the index ACS admission were used to identify predictors of post-hospital CV mortality (avg 2 yr f/u). A multivariable model using backward selection was constructed to develop a simple mortality risk score with 1 point assigned to 11 independent risk predictors. 7290 pts (mean age 59 yrs; range 25–92), including 1681 women (23%), and 1538 pts age 70+ (21%) had complete data for all 11 variables. Stepwise increasing numbers of predictors corresponded to higher long-term CV mortality (c-stat 0.76) (FIG⇓). The 35% of pts with scores 0–2 had a 2-year CV mortality <0.7%, while the 18% with scores ≥5 had a 2-yr mortality of >8%. The score performed well in key subgroups such as women (c-stat 0.80), elderly (c-stat 0.77), and STEMI (c-stat 0.70), and in an older/sicker population (n=4222) from MERLIN-TIMI 36 (c-stat 0.73). Our score also discriminated risk of other endpoints including all-cause mortality (c-stat 0.74), death/CHF (c-stat 0.73). The GRACE postdischarge risk score yielded a c-stat of 0.74 in this dataset, despite its greater complexity and use of continuous variables. We developed a simple, convenient discharge risk score that discriminates high and low risk in post-ACS pts for long-term CV mortality, enabling long-term management priorities to be organized in proportion to risk.