The addition of two biomarkers that have emerged as possible prognosticators for acute coronary syndromes (ACS) to the GRACE score did not improve the risk prediction, according to a study published in the September issue of Heart. The GRACE score alone proved accurate for predicting in-hospital and long-term mortality for patients with ACS.
The GRACE score is a tool widely used to evaluate mortality risk for patients with ACS and is included in the European Society of Cardiology's recommendations for stratifying risk in patients with non-ST elevation ACS. Assay of the biomarker high-sensitivity cardiac troponin (hs-cTn) has been shown to detect acute MI (AMI) and overall ACS, and B-type natriuretic peptides (BNP and NT-proBNP) have been shown as prognosticators of ACS.
Christophe Meune, MD, PhD, of the department of internal medicine at University Hospital Basel in Basel, Switzerland, and colleagues designed their study to compare the accuracy of the GRACE score, which is calculated using conventional cardiac troponin (cTn) assays, with the GRACE score plus hs-cTn and with the GRACE score plus BNP.
"In this study, we examined patients across the entire spectrum of ACS and focused on mortality because the GRACE score was developed and validated mainly for this indication," Meune et al wrote. They chose end points of in-hospital death, one-year mortality and a combined death/AMI at one year to measure the predictive capacity of the GRACE score in the short term and long term. To determine predictive ability of the models, they applied both c-statistics and integrated discrimination improvement methods.
The researchers included 370 patients enrolled in the APACE (Advantageous Predictors of Acute Coronary Syndromes Evaluation) prospective study. Of those, 173 patients had unstable angina and 197 had AMI (50 with ST-elevation MI [STEMI] and 147 with non-STEMI).
They calculated a GRACE score of 200 for patients who died compared with a score of 125 for patients who survived to discharge, and a score of 151 for patients who died at one year compared to 104 for survivors. The study found the addition of hs-cTn or BNP did not add significant value to the model.
"The in-hospital and one-year prognostic values of the GRACE score were excellent in our study," Meune and colleagues wrote, "and we observed even higher AUC [area under the curve] than in the validation sets of the GRACE registries."
The authors listed small sample size, the exclusion of patients with renal failure and the fact that the biomarkers may have been affected in cases involving inpatient coronary intervention as limitations.