Left ventricular (LV) mass, as estimated from M-mode echocardiography (echo), has previously been shown to be an independent predictor of incident cardiovascular disease morbidity and mortality. We evaluated the relationship at baseline of echo LV mass to relevant cardiovascular disease risk factors and other potential covariates in the Cardiovascular Health Study, multicenter study sponsored by the National Heart, Lung, and Blood Institute of 5201 men and women aged 65 years or older (mean, 73). Two-dimensionally directed M-mode echo LV mass measurements could be obtained in 1357 men and 2053 women (66% of this elderly cohort). Stepwise linear regression analyses of the relationship of echo LV mass to demographic and risk factor, physical activity, electrocardiographic, and prevalent disease variables resulted in a model that explained 37% of the variance for the entire cohort. In order of decreasing importance, factors positively associated with echo LV mass were body weight, male sex, systolic pressure, presence of congestive heart failure, present smoking, major and minor electrocardiographic abnormalities, treatment for hypertension, valvular heart disease, aortic regurgitation by color Doppler, and mitral regurgitation by color Doppler (in men) whereas diastolic pressure, bioresistance (a measure of adiposity), and high-density lipoprotein cholesterol were inversely related to echo LV mass. Although height and weight were both related to LV mass, height added nothing once weight was entered in multiple linear regression analyses. Furthermore, in the multiple regression models, diastolic pressure was inversely and systolic BP positively related to LV mass, with similar magnitudes for their coefficients. In consonance with these findings, pulse pressure was positively related to LV mass in bivariate analyses. Multiple linear regression analyses explained less of the variance for ventricular septal thickness (R2 = .13) and LV posterior wall thickness (R2 = .14) than for LV mass (R2 = .37) and LV diastolic dimension (R2 = .27). Intriguing findings in the elderly Cardiovascular Health Study cohort included the presence of pulse pressure as a positive correlate, and high-density lipoprotein cholesterol as an inverse correlate, of LV mass. Longitudinal studies in the Cardiovascular Health Study cohort will help to clarify the importance of demographic, risk factor, and other variables, and changes in these variables, in predicting changes in echo LV mass and its components as well as the prognostic significance of LV mass in the elderly.