Currently, the only function of the resting electrocardiogram (ECG) in patients referred for exercise testing is to determine whether imaging is mandated. It is unknown if subtle ECG findings in those patients with clinically normal resting ECGs have prognostic significance.

We performed a single-center cohort study of 18,964 patients without known cardiovascular disease who had a clinically normal resting ECG and who underwent treadmill exercise testing for evaluation of suspected coronary artery disease. Eleven quantitative ECG measures relating to heart rate, conduction, left ventricular mass, or repolarization were collected digitally. The primary outcome was all-cause mortality. The prognostic importance of a composite ECG score was assessed by measuring its impact on the c-index (analogous to area under receiver operating characteristic curve) and by measures of reclassification.

During a median follow-up of 10.7 years, 1,585 patients died. The 4 most predictive digital ECG variables were higher ventricular rate, more leftward QRS axis, more downward ST-segment deviation, and longer QT interval. The ECG score was independently associated with mortality (75th vs 25th percentile hazard ratio 1.36, 95% confidence interval 1.25-1.49, P < .0001). The ECG score had modest impact on discrimination (change in c-index 0.04) and reclassification of risk (3.0% decrease of relative integrated discrimination improvement, P < .001).

Subtle ECG findings relating to heart rate, conduction, left ventricular mass, or repolarization in patients with clinically normal ECGs referred for exercise testing may provide modest additional prognostic information over and above clinical and exercise measures.