Abstract
BACKGROUND: ECGs are a primary assessment of cardiac structure and function. Alignment of their findings with cardiac imaging is not well defined. This study investigates the correlation of left atrial enlargement and left ventricular hypertrophy (LVH) found on ECG with transthoracic echocardiogram and cardiac magnetic resonance imaging (CMR) findings. METHODS: This is a single-center retrospective study of 1000 patients. It included patients with positive ECG criteria for left atrial enlargement or left ventricular hypertrophy (LVH) with transthoracic echocardiograms done within the same month. CMR data were included if within 6 months of the ECG. Receiver operator curves and area under the curves (AUC) were calculated based on whether ECG findings were confirmed on either imaging. RESULTS: Left atrial enlargement ECG criteria best correlating with transthoracic echocardiograms were p-wave duration of 120+ ms in lead II (AUC, 0.81 [95% CI, 0.79-0.83]; P value <0.0001) followed by negative p-wave deflection of 1+ mm and 40 ms + in lead V1 (AUC, 0.59 [95% CI, 0.57-0.62]; P value <0.0001). CMR mirrored similar findings. For LVH ECG criteria compared with transthoracic echocardiograms, Romhilt-Estes criteria demonstrated the strongest correlation (AUC, 0.72 [95% CI, 0.69-0.74]; P value <0.0001), followed by Sokolow-Lyon (AUC, 0.67 [95% CI, 0.64-0.69]; P value<0.0001), Modified Cornell (AUC, 0.65 [95% CI, 0.63-0.68]; P<0001), and Cornell (AUC, 0.61 [95% CI 0.58-0.63]; P value<0.0001). CMR data corroborated this pattern. LAE and LVH criteria with the highest AUC were least represented by patients. CONCLUSIONS: ECG criteria for left atrial enlargement and LVH correlated poorly with both transthoracic echocardiograms and CMR. Updating current waveform criteria may improve diagnostic accuracy of ECGs.