GRACE Score Validation in Predicting Hospital Mortality: Analysis of the Role of Sex

Abstract

Background: The GRACE (Global Registry of Acute Coronary Events) risk score is recommended for risk stratification in acute coronary syndrome (ACS). It does not include sex, a variable strongly associated with ACS prognosis. The aim of this study was to examine if sex adds prognostic information to the GRACE score in a contemporary population. Materials and Methods: Analysis of discrimination and calibration of GRACE score in the validation population, derived from the ARIAM-SEMICYUC registry (2012-2015). Outcome was hospital mortality. The uniformity of fit of the score was tested in predefined subpopulations: with and without ST-segment elevation myocardial infarction (STEMI and NSTEMI). Results: A total of 9781 patients were included: 4598 with NSTEMI (28% women) and 5183 with STEMI (23% women). Discriminative capacity of the GRACE score was significantly lower in women with STEMI compared to men (area under the receiver operating characteristic curve [AUC] 0.82, 95% CI 0.78-0.86 vs. AUC 0.90, 95% CI 0.88-0.92, p = 0.0006). In multivariate analysis, female sex predicted hospital mortality independently of GRACE in STEMI (p = 0.019) but not in NSTEMI (p = 0.356) (interaction p = 0.0308). However, neither the AUC nor the net reclassification index (NRI) improved by including female sex in the STEMI subpopulation (NRI 0.0011, 95% CI -0.023 to 0.025; p = 0.928). Conclusions: Although female sex was an independent predictor of hospital mortality in the STEMI subpopulation, it does not substantially improve the discriminative ability of GRACE score.