Objective: To identify predictors of arterial hypertension. Patients: One hundred thirty-two normotensive adults from a large employed population. Methods: Echocardiography, standard blood tests, and 24-hour urine collection, at baseline and after an interval of 3 to 6 years (mean, 4.7 +/- 0.8 years). Results: At follow-up, 15 subjects (11%; 7 men, 8 women) had a systolic blood pressure greater than 140 mm Hg or a diastolic blood pressure greater than 90 mm Hg or both (mean, 143 +/- 7 and 87 +/- 6 mm Hg, respectively). At baseline, subjects who developed hypertension had a greater left ventricular mass index than those who did not (92 +/- 25 compared with 77 +/- 19 g/m2 body surface area; P < 0.005) and higher 24-hour urinary sodium/potassium excretion ratio (3.6 +/- 1.7 compared with 2.6 +/- 1.4; P < 0.04); there were no differences in race, initial age, systolic or diastolic blood pressure, coronary risk factors, or plasma renin activity. The likelihood of developing hypertension rose from 3% in the lowest quartile of sex-adjusted left ventricular mass index to 24% in the highest quartile (P < 0.005); a parallel trend was less regular for quartiles of the sodium/potassium excretion ratio (P < 0.04). In multivariate analyses, follow-up systolic pressures in all subjects and in the 117 who remained normotensive were predicted by initial age, systolic blood pressure, black race, and sex-adjusted left ventricular mass index; final diastolic blood pressure was predicted by its initial value, plasma triglyceride levels, urinary sodium/potassium ratio, low renin activity, black race, and plasma glucose level. Conclusions: Echocardiographic left ventricular mass in normotensive adults is directly related to the risk for developing subsequent hypertension. Left ventricular mass improves prediction of future systolic pressure, whereas diastolic pressure is more related to initial metabolic status. Black race is also an independent determinant of higher subsequent blood pressure.