To the Editor,

In the comments by Pérez et al regarding our article,1 they are correct with regards to the validity of the GRACE scale in the subgroup of patients excluded due to loss during follow-up (79 [6.3%]). On this subject, the robustness of the GRACE scale in our study did not drop when these patients were included (c-statistic = 0.85 [0.784-0.916]; Hosmer-Lemeshow P=.6). The predictive covariables that make up the model were similar in the subgroups with and without valid data on vital status, as indicated by the GRACE score for both subgroups (median, 121 [96-144] vs 117 [94-142]; P=.51).

Although the objective of our study was quite clear, to validate the GRACE scale for predicting 6-month mortality, the commentary by Pérez et al gives us an opportunity to present here, briefly, information on the validity of the GRACE scale for predicting intra-hospital risk of death.2 We have proven the validity of this scale for the total sample and by type of acute coronary syndrome, as well as by subgroups with and without percutaneous coronary revascularization during hospitalization. The validation indexes were adequate as an overall score and by the subgroups explored (c-statistic ≥0.79, Hosmer-Lemeshow P>.1). Given that the GRACE scale for predicting intra-hospital risk of death has been recently updated,3 we have repeated the previous analysis with the point scores corresponding to the modernized model. The results did not differ, the discrimination of the model was >0.8, and the estimations here were substantially adjusted to real values (observed mortality) (Hosmer-Lemeshow P≥.12). Therefore, we conclude that the GRACE score represents a useful and reliable clinical tool in our population for predicting the risk of death during hospital stay and at 6 months after discharge. The lack of data on reinfarctions is a limitation in our work,1 as it did not allow us to validate the GRACE model that estimates the probability of occurrence of the combined event of death or reinfarction.