Soy-based diets reduce blood pressure in spontaneously hypertensive rats, but apparently not in hypertensive humans. In the present study, the antihypertensive potential of soy milk (500 mL twice daily) compared with cow's milk was investigated in a 3-mo double-blind randomized study of 40 men and women with mild-to-moderate hypertension. Before initiation of the study, urinary isoflavonoids (measured by HPLC) were undetectable in most cases (for genistein, they were always <100 micromol/L). After 3 mo of soy milk consumption, systolic blood pressure decreased by 18.4 +/- 10.7 mmHg compared with 1.4 +/- 7.2 mmHg in the cow's milk group (P < 0.0001), diastolic blood pressure decreased by 15.9 +/- 9.8 mmHg vs. 3.7 +/- 5.0 mmHg in the cow's milk group (P < 0.0001) and mean blood pressure decreased by 16.7 +/- 9.0 mmHg compared with 3.0 +/- 4.6 mmHg in the cow's milk group (P < 0.0001). Urinary genistein was strongly (r = -0.588) and significantly (P = 0.002) correlated with the decrease in blood pressure, particularly for diastolic values. In conclusion, chronic soy milk consumption had modest, but significant hypotensive action in essential hypertensive subjects. This hypotensive action was correlated with the urinary excretion of the isoflavonoid genistein.