Utility of nonspecific resting electrocardiographic features for detection of coronary artery stenosis by computed tomography in acute chest pain patients: from the ROMICAT trial.

ABSTRACT

Twelve-lead surface electrocardiography (ECG) and computed tomography (CT) are used to evaluate for myocardial ischemia and coronary artery disease (CAD), respectively. We aimed to determine features on resting ECG that predict coronary artery stenosis by cardiac CT. In 309 acute chest pain patients, we compared the initial triage resting ECG to contrast-enhanced 64-slice cardiac CT angiography. We assessed for 6 quantitative (QT interval, QTc interval, QTc > 440 ms, gender-specific QTc, QT dispersion and QRS duration) and 4 qualitative ECG parameters (ST depression >0.05 to ≤0.1 mV, T wave inversion ≥0.1 mV, T wave flattening, and any T wave abnormalities) and for the presence of coronary stenosis by CT (>50% luminal narrowing). Specificities of these ECG parameters were excellent (83.6-97.0%) while sensitivities were poor (12.2-29.3%). For coronary stenosis detection, the ECG features with the greatest performance were the presence of ST depression (positive likelihood ratio [LR+] 4.09) and T wave inversion (LR+ 4.58). In multivariable analyses, the risk for coronary stenosis increased by 33-41% for every 20 ms prolongation of the QTc interval after adjusting for age, gender, and cardiac risk factors or adjustment for Framingham risk score. Similarly, there was an increase of fourfold with the presence of ST depression >0.05 to ≤0.1 mV or T wave inversion ≥0.1 mV. In acute chest pain patients, resting ECG features of QTc interval prolongation, mild ST depression, and T wave inversion are independently associated with the presence of CT coronary stenosis and their presence suggests an increase risk of CAD.