Left ventricular hypertrophy: relationship of anatomic, echocardiographic and electrocardiographic findings.

Anatomic, echocardiographic and ECG findings of left ventricular hypertrophy (LVH) were compared in 34 subjects. Echocardiographic LV mass correlated weel with postmortem LV weight (r = 0.96) and accurately diagnosed LVH (sensitivity 93%, specificity 95%). In contrast, Romhilt-Estes (RE) point score and Sokolow-Lyon (SL) voltage criteria for ECG LVH were insensitive (50% and 21%, respectively) but specific (both 95%). RE correlated weakly with LV weight (r = 0.64), but SL did not. Echocardiographic LV mass was then compared with RE and SL in an unselected clinical series of 100 subjects, in 28 subjects with severe aortic stenosis (AS) and in 14 with severe aortic regurgitation (AR). Results in the clinical series were comparable to those in the necropsy series. In the AS and AR groups, with a high prevalence of LVH, the low sensitivity of RE point score and Sl criteria led to poor overall results. Analysis of individual ECG variables showed that most voltage information is contained in leads aVL and V1. Correction of voltage for distance from the left ventricle did not substantially improve results. Individual nonvoltage criteria were each nearly as sensitive as RE point score. We could not devise new ECG criteria that improved diagnostic results. We conclude that the ECG is specific but insensitive in recognition of LVH. Moreover, when true LVH prevalence is less than 10%, more false-positive than true-positive diagnoses will be obtained. M-mode echocardiographic LV mass is superior to ECG criteria for clinical diagnosis of LVH.