To determine the prevalence and correlates of echocardiographic left ventricular hypertrophy among subjects in a general population, we studied 621 employed subjects. Patients with uncomplicated essential hypertension in a worksite-based treatment program included 145 with borderline hypertension and 316 with sustained hypertension by World Health Organization criteria. Normotensive subjects were randomly selected from members of the same unions. M-mode echocardiographic left ventricular dimensions were used to calculate left ventricular mass and other indexes of left ventricular anatomy. The specificity of 13 echocardiographic criteria of left ventricular hypertrophy was determined in normotensive individuals, and the prevalence of left ventricular hypertrophy by each criterion was assessed in patients with borderline or sustained essential hypertension. The results suggest that the most suitable reference standard for detection of left ventricular hypertrophy in a heterogeneous urban population utilizes sex-specific cutoff values for left ventricular mass index of 110 g/m2 or greater for women and 134 g/m2 or greater for men. With 97% specificity, the prevalence of left ventricular hypertrophy by these criteria is approximately 12% among patients with borderline hypertension and 20% among patients with relatively mild, uncomplicated sustained essential hypertension. Wall thickness measurements performed slightly less well. At similar levels of blood pressure, black patients were more likely than white patients to exhibit concentric left ventricular hypertrophy, especially among borderline hypertensive patients. Left ventricular hypertrophy occurred in patients with sustained hypertension who also exhibited increased cardiac output, strongly associated with low plasma renin activity.