A reliable noninvasive index of left ventricular mass would be useful in following patients with valvular heart disease and left ventricular hypertrophy. We reviewed concurrent electrocardiograms and echocardiograms from 54 subjects, 39 patients with aortic or mitral valve disease and 15 normal subjects. Pre- and early postoperative echocardiographic estimates of left ventricular mass in 17 patients who had valve replacements correlated well (r = 0.96, p less than 0.001) and demonstrated little change in mean values despite altered left ventricular dimensions. Echocardiographic estimates of left ventricular mass were, therefore, used as a standard for evaluating other noninvasive indices. Precordial electrocardiographic voltage showed a weak correlation with left ventricular mass in the study group as a whole (r = 0.59, p less than 0.001), but no correlation in patients with volume overload (r = 0.36, p = NS). In 18 patients who had preoperative and three separate postoperative studies at least eight weeks apart, changes in left ventricular cross-sectional area (an index of left ventricular mass which corrects for changes in left ventricular volume) closely followed alterations in left ventricular mass. However, changes in posterior wall and interventricular septal thickness often resulted from altered ventricular volume and did not accurately reflect directional changes in left ventricular mass. Serial changes in electrocardiographic voltage were similarly unreliable. We conclude that left ventricular mass and cross-sectional area by echocardiography allow accurate noninvasive assessment of left ventricular mass, whereas wall thickness and electrocardiographic changes do not.