To the Editor,
We have read with interest the article titled "Validation of the GRACE risk score for predicting death within 6 months of follow-up in a contemporary cohort of patients with acute coronary syndrome," by Abu-Assi et al.1 We find it relevant to clarify certain methodological aspects, given the potential repercussions in certain populations.
The population under study is different in many aspects than the one used for building and later validating the GRACE scale. One of the most noteworthy parameters, the elevated rate of invasive techniques used in the patients, is currently very common in our field. This increases the value of the results presented. Unfortunately, as the authors recognize, the data on the incidence of reinfarctions and those corresponding to the rate of events during hospitalization are missing. Although mortality was undoubtedly the most relevant parameter, the additional information that could be provided by an analysis of reinfarctions would be valuable. Have the authors analyzed these variables with negative results, or are these data missing?
Likewise, the patients lost to follow-up that the authors describe could be a source of bias in the results.2 Although the number appears to be low (79 patients; 6.3%), it is greater than the number of deaths observed (52 patients; 4.4%). We cannot assume that all of the patients lost during the follow-up period passed away, but the mortality rate would probably be higher in this group.3 Can we obtain some clue as to the validity of the GRACE scores in this subgroup of patients? Were the baseline characteristics of these patients similar to those of the group that was monitored? What was the GRACE score for this subgroup?
As the authors conclude in the editorial that accompanies the article,4 the validation of these risk scales in different contemporary populations strengthens their role as a therapeutic tool. However, for this to occur, we must have all the possible data at our disposal.