Validation of the GRACE score (Global Registry of Acute Coronary Events) as a predictor of in-hospital mortality in patients with acute myocardial infarction in Argentina

Abstract

The Global Registry of Acute Coronary Events (GRACE) score is widely used and recommended by clinical practice guidelines to stratify risk and predict mortality in the broad spectrum of acute coronary syndromes, but it is not validated in Argentina. Given the variability of the clinical, ethnic, socio-demographic, cultural or idiosyncratic characteristics of the patients and the health care patterns of different health systems of each country, before using and applying a predictive model of prediction outside the  environment where it was created, validation is necessary to ensure that it does not yield erroneous results. Objective: To validate the GRACE score as a predictor of in-hospital mortality in a cohort of patients diagnosed with acute myocardial infarction (AMI)
in 45 centers in Argentina. Methods: Validation study of the GRACE prognostic score using a retrospective analysis of a
cohort composed of consecutive patients admitted to 45 centers across the country with a diagnosis of AMI between December 2009 and July 2010 using data from the CONAREC XVII registry. The score was calculated at admission, and the predictive value of the score for in-hospital death was estimated. The validity of the models was evaluated through its two components: discrimination
using the area under the ROC curve (AUC) and calibration using the Hosmer-Lemeshow (HL) test. Results: A total of 1127 patients with a diagnosis of AMI (AMI with ST segment elevation (STEMI) 59% and AMI without ST segment elevation (NSTEMI) 41% were analyzed), with a mean age of 63 ± 12.6 years; 77.5% were male. Overall mortality was 7.4%. The median of the score was different according to mortality, being 187 (Pc 25-75 150-244) in those who died in the hospitalization, and 138 (Pc 25-75 119-162) in those who did not, p <0.001. This risk model showed a good capacity to predict in-hospital mortality, with AUC 0.76, 95% CI 0.70-0.82, p <0.001, with a higher discrimination in STEMI with AUC 0.85 (95% CI 0.79- 0.91), than in the NSTEMI, with an AUC of 0.6 (95% CI 0.48-0.71). The HL test applied to the GRACE score in both types of acute coronary syndrome presented a good calibration. Conclusions: In this multicenter study, the GRACE risk score for the prediction of in-hospital mortality has been adequately validated in patients with AMI in Argentina, with good discriminative capacity and good calibration. This suggests its suitability for clinical use in this population.