We assessed the utility of the 12-lead electrocardiogram (ECG) in identifying severe coronary artery disease (CAD) in patients with depressed left ventricular (LV) systolic function. In 336 patients referred for cardiac catheterization with LV ejection fractions < 0.50, we compared the 12-lead ECG of those with and without CAD by multivariate analysis. The sensitivities, specificities, and positive and negative predictive values of all dichotomous electrocardiographic variables for identifying the presence of severe CAD were determined. In comparison to subjects with CAD, those without disease were more likely to exhibit left-axis deviation (p = 0.01), left bundle branch block (p < 0.001), or LV hypertrophy (p < 0.001), and less likely to exhibit pathologic inferior Q waves (p < 0.001). The presence of anterior or any Q waves was similar between the groups. The presence of any diagnostic Q wave had a positive predictive value of 92%, sensitivity of 57%, and specificity of 80% for identifying severe CAD. In patients with LV systolic dysfunction, the 12-lead ECG is insensitive and nonspecific for identifying those with concomitant severe CAD.