Biomarkers for Predicting Serious Cardiac Outcomes at 72 Hours in Patients Presenting Early after Chest Pain Onset with Symptoms of Acute Coronary Syndromes

Abstract

Abstract BACKGROUND Most outcome studies of patients presenting early to the emergency department with potential acute coronary syndromes have focused on either the index diagnosis of myocardial infarction (MI) or a composite end point at a later time frame (30 days or 1 year). We investigated the performance of 9 biomarkers for an early serious outcome. METHODS Patients (n = 186) who presented to the emergency department within 6 h of chest pain onset had their presentation serum sample measured for the following analytes: creatine kinase, creatine kinase isoenzyme MB, enhanced AccuTnI troponin I (Beckman Coulter), high-sensitivity cardiac troponin T (hs-cTnT), ischemia-modified albumin, interleukin-6, investigation use only hs-cTnI (Beckman Coulter), N-terminal pro–B-type natriuretic peptide, and cardiac troponin I (Abbott AxSym). We followed patients until 72 h after presentation and determined whether they experienced the following serious cardiac outcomes: MI, heart failure, serious arrhythmia, refractory ischemic cardiac pain, or death. ROC curves were analyzed to determine the area under the ROC curve (AUC) and optimal cutoffs for the biomarkers. RESULTS The AUCs for the hs-cTnI assay (0.86; 95% CI, 0.76–0.96), the AccuTnI assay (0.86; 95% CI, 0.78–0.95), and the hs-cTnT assay (0.82; 95% CI, 0.71–0.94) assays were significantly higher than those for the other 6 assays (AUC values ≤0.71 for the rest of the biomarkers, P < 0.05). The ROC curve–derived optimal cutoffs were ≥19 ng/L (diagnostic sensitivity, 80%; specificity, 88%), ≥0.018 μg/L (diagnostic sensitivity, 75%; specificity, 86%), and ≥32 ng/L (diagnostic sensitivity, 68%; specificity, 92%) for the hs-cTnI, AccuTnI, and hs-cTnT assays, respectively. CONCLUSIONS The optimal cutoffs for predicting serious cardiac outcomes in this low-risk population are different from the published 99th percentiles. Larger studies are required to verify these findings.