Abstract 157: Use of a Novel Risk Score in the Emergency Department Discriminates Acute Coronary Syndrome Among Chest Pain Patients with Known Coronary Artery Disease

Abstract

Background: Patients with known coronary artery disease (CAD) presenting to the Emergency Department (ED) with chest pain thought to be of ischemic origin are often admitted to the hospital, yet less than half are eventually diagnosed with acute coronary syndrome (ACS). We assessed whether the use of a novel risk score in the ED could discriminate which of these high-risk patients actually do or do not have ACS. Methods and Results: Chart review was performed on a prospectively defined cohort of 142 patients with known CAD presenting to the ED with chest pain thought to be of ischemic origin, all of whom were admitted to the hospital from December 2012 to April 2013. Known CAD was defined as history of myocardial infarction, PCI, CABG, angiographic coronary stenosis >50%, or a positive stress test. Troponin I was measured using the Beckman Coulter assay. Variables were assessed with logistic regression for their association with ACS as determined by the inpatient attending physician at hospital discharge. The cohort included 59 women (42%) and 90 African American individuals (63%). One-hundred sixteen patients (82%) had a history of revascularization (104 PCI, 53 CABG, 41 both). ACS was eventually diagnosed in 43 (30%) of the patients. Non-ACS patients had a 2.8 day average length of stay and $9,908 average inpatient (post-ED) hospital charges (not including physician fees), which is $980,926 for the 99 (70%) non-ACS patients. A novel risk score, including (1) elevated troponin I (>0.05 ng/mL) in the ED, (2) dynamic ECG changes in the ED, (3) body mass index (BMI), (4) home aspirin use, (5) age older than 65, (6) history of chronic kidney disease (CKD), and (7) associated illness at presentation to the ED (anemia, arrhythmia, hypertension, infection, COPD exacerbation, diabetic ketoacidosis or hyperosmolar hyperglycemic state), discriminated ACS and non-ACS with an area under ROC curve (AUC) of 0.829. In the multi-variable regression, troponin I elevation was the most predictive of ACS (OR 7.22, p <0.001), followed by home aspirin use (OR 6.07, p 0.036), age older than 65 (OR 4.06, p 0.012), dynamic ECG changes (OR 2.68, p 0.046), and BMI (OR 1.09, p 0.008). The presence of an associated illness was associated with decreased likelihood of ACS (OR 0.24, p 0.013), as was CKD (OR 0.17, p 0.008). Conclusions: A novel risk score including elevated troponin I in the ED, dynamic ECG changes in the ED, body mass index, home aspirin use, age older than 65, history of chronic kidney disease, and associated illness at presentation to the ED, is a valuable tool for discriminating between ACS and non-ACS among patients with known CAD presenting to the ED with chest pain. This preliminary analysis provides a foundation for larger and prospective studies for validation. Application of this risk score, along with other clinical factors, may reduce the number of potentially avoidable admissions and associated costs.