Background: The HEART score uses elements from patient History, Electrocardiogram, Age, Risk Factors, and Troponin to obtain a risk score on a 0- to 10-point scale for predicting acute coronary syndromes (ACS). This investigation seeks to improve on the HEART score by proposing the HEARTS(3) score, which uses likelihood ratio analysis to give appropriate weight to the individual elements of the HEART score as well as incorporating 3 additional "S" variables: Sex, Serial 2-hour electrocardiogram, and Serial 2-hour delta troponin during the initial emergency department valuation.
Methods: This is a retrospective analysis of a prospectively acquired database consisting of 2148 consecutive patients with non-ST-segment elevation chest pain. Interval analysis of likelihood ratios was performed to determine appropriate weighting of the individual elements of the HEART(3) score. Primary outcomes were 30-day ACS and myocardial infarction.
Results: There were 315 patients with 30-day ACS and 1833 patients without ACS. Likelihood ratio analysis revealed significant discrepancies in weight of the 5 individual elements shared by the HEART and HEARTS(3) score. The HEARTS(3) score outperformed the HEART score as determined by comparison of areas under the receiver operating characteristic curve for myocardial infarction (0.958 vs 0.825; 95% confidence interval difference in areas, 0.105-0.161) and for 30-day ACS (0.901 vs 0.813; 95% confidence interval difference in areas, 0.064-0.110).
Conclusion: The HEARTS(3) score reliably risk stratifies patients with chest pain for 30-day ACS. Prospective studies need to be performed to determine if implementation of this score as a decision support tool can guide treatment and disposition decisions in the management of patients with chest pain.
Copyright © 2012 Elsevier Inc. All rights reserved.