Both experimental and epidemiological studies support the idea of magnesium supplementation in essential hypertension. We added 15 mmol Mg to a free diet in 71 subjects with mild essential hypertension or a high-normal blood pressure in a double-blind, placebo-controlled study over 6 months. The treatment, which raised urinary magnesium excretion 30%, induced no general effects on the blood pressure. However, when the changes in blood pressure in the actively treated group were related to the pretreatment magnesium status, a correlation was found between pretreatment urinary magnesium excretion and the induced change in supine blood pressure (P less than .05) with a blood pressure reduction in subjects with a low pretreatment urinary excretion of magnesium, and a pressor effect in the subjects with the highest pretreatment levels of urinary magnesium. The induced change in blood pressure was furthermore found to be inversely correlated to the changes in serum magnesium and urinary excretion of sodium (P less than .03) induced by treatment indicating that both a direct calcium antagonist action of magnesium at the cellular level as well as a diuretic effect of the increased magnesium load might be involved in the blood pressure effects of magnesium. Pretreatment serum potassium concentration also appeared to be a predictor of the induced change in standing blood pressure (P less than .03). In conclusion, magnesium supplementation does not seem to be effective in unselected mild hypertensive subjects or in subjects with a high-normal blood pressure and can therefore not be generally recommended.(ABSTRACT TRUNCATED AT 250 WORDS)