Key result
Original GRACE score slightly outperforms a gender- and race-modified model for predicting in-hospital ACS mortality.
Why the study?
The GRACE score is a validated risk stratification tool for ACS that does not include race and gender, prompting assessment of whether adding these variables improves its predictive ability.
Does adding gender and race to the GRACE score improve the prediction of in-hospital mortality in patients with acute coronary syndrome?
Cohort (n=46,764)
Yes
Does adding gender and race to the GRACE score improve the prediction of in-hospital mortality in patients with acute coronary syndrome?
Absolute Event Rate: 0.838% vs 0.839%
p-value: p=0.008
The original GRACE score remains highly accurate for predicting in-hospital mortality in ACS patients, and its predictive ability is not substantially improved by adding gender and race.
Original GRACE score suffices for ACS mortality prediction without gender or race; cohort data challenge demographic modifications but should not yet alter practice.
Background Acute coronary syndrome (ACS) causes significant global morbidity and mortality and requires early risk stratification. The global registry of acute coronary events (GRACE) score is a well-known, validated risk stratification system that does not include race and gender. We aimed to assess whether the addition of gender and race could add to the predictability of the GRACE score model. Methods We performed a retrospective cohort study of 46 764 ACS patients from the files of a national healthcare system. We compared the predictability of the GRACE score in conjunction with gender and race versus the original GRACE score. Different possible associations of predictability were investigated and statistically calculated. The accuracy of the prediction models was assessed using the receiver operating characteristic curve and its respective area under the curve (AUC). We compared the AUC of the 2 models, with the significance set at a P value of less than .05. Results Our comparison favored the original GRACE score over the modified prediction model with gender and race added (AUC = 0.838 and 0.839 respectively, P = .008). Although the P value comparing the AUC shows that the original GRACE was superior, due to our large dataset, the actual numbers are similar and may not be clinically significant. Gender and race were significantly associated with in-hospital mortality (P < .001, P = .002, respectively). However, this relationship disappeared in the multivariate analysis. Gender significantly predicted in-hospital mortality, with females 1.167 times more likely to die (P < .001). Non-white racial groups had lower in-hospital mortality than whites (OR: 0.823, P = .03). Conclusion The GRACE score was valid in its original form and its ability to predict mortality was not substantially improved by including gender and race.
No takes yet. Share an insight, caveat, or question.
Ogbu et al. (2023) conducted a cohort in Acute coronary syndrome (n=46,764). Modified GRACE score (including gender and race) vs. Original GRACE score was evaluated on Predictability of in-hospital mortality (AUC) (p=0.008). The original GRACE score was slightly but significantly superior to a modified model incorporating gender and race for predicting in-hospital mortality in ACS patients (AUC 0.839 vs 0.838, P=0.008).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: