Abstract 19451: Evaluating the Diagnostic Accuracy of Clinical Risk Scores to Detect Acute Coronary Syndrome in Patients Evaluated at the Emergency Department for a Chief Complaint of Chest Pain

    loading  Checking for direct PDF access through Ovid


Introduction: The process of diagnosing acute coronary syndrome (ACS) in patients with chest pain, especially with the absence of ST elevation (STE), is lengthy, costly, and consumes significant resources. For this task, multiple risk scores have been developed to identify high-risk patients with suspected ACS. These risk scores have never been simultaneously compared in patients with undifferentiated chest pain. We sought to compare the diagnostic accuracy of established clinical risk scores to identify ACS in patients with chest pain.

Methods: This study prospectively enrolled consecutive, non-traumatic, chest pain patients transported by Emergency Medical Services to one of three UPMC-affiliated tertiary care hospitals in Pittsburgh between 2013 and 2014. A reviewer blinded to outcome data retrospectively reviewed patient charts to compute five risk scores: HEART, TIMI, FRISC, PURSUIT, and GRACE. The primary outcome was the diagnosis of ACS, defined as the detection of cardiac biomarkers or the presence of ECG, clinical, echocardiographic, or nuclear evidence of acute myocardial ischemia. The classification performance of the five risk scores was evaluated using the area under the ROC curve (AUC).

Results: The final sample included 750 patients (age 59±17; 42% females, and 40% Blacks). Overall, 130 patients (17.3%) met the primary study outcome (53 with STE and 77 with no STE). Patients with ACS were older and were more likely to have a prior history of diabetes mellitus and MI. The AUC for HEART, TIMI, FRISC, PURSUIT, and GRACE to identify ACS were 0.87, 0.86, 0.82, 0.77, and 0.73 respectively. After excluding those with STE, the AUC to identify non-STE ACS were 0.84, 0.85, 0.81, 0.71, and 0.69 respectively. A HEART score ≥4 achieved sensitivity/specificity values of 94% / 51% to detect ACS. A TIMI score ≥3 achieved values of 86% / 75% respectively.

Conclusions: HEART and TIMI scores outperform other clinical risk scores in classifying undifferentiated chest pain patients to identify ACS cases, including those with non-STE. Although HEART score is more sensitive, it has poor specificity. TIMI score seems to better balance sensitivity and specificity.

Related Topics

    loading  Loading Related Articles