In newly-diagnosed untreated (n = 24) and thiazide treated (n = 18) hypertensive patients erythrocyte ion content and plasma ion and metabolite levels were determined. Thiazide treated patients had lower arterial blood pressure, plasma ion levels, erythrocyte magnesium and potassium content but higher fasting plasma insulin (66 +/- 7 v 87 +/- 8 pmol/L P < .02), triglycerides (1.88 +/- 0.24 v 2.34 +/- 0.44 mmol/L P < .05), free fatty acids (0.68 +/- 0.11 v 0.81 +/- 0.18 mmol/L P < .05). Subsequently, in a double-blind fashion and in random order thiazide diuretic treated patients were assigned to two groups: the first (n = 9) taking magnesium (15.8 mmol/day) and the other (n = 9) a placebo. Each treatment period lasted 8 weeks. At the end of each treatment period, each patient underwent blood sampling for determination of erythrocyte ion content and plasma ion and metabolite levels and was submitted to an euglycemic hyperinsulinemic (1 mU/kg/min for 120 min) glucose clamp. In this latter test D-3-H glucose infusion and indirect calorimetry allowed determination of glucose turnover parameters and substrate oxidation respectively. Chronic magnesium administration (CMA) raised fasting plasma (0.79 +/- 0.03 v 0.83 +/- 0.02 mmol/L, P < .05) and erythrocyte (1.98 +/- 0.08 v 2.35 +/- 0.03 mmol/L, P < .01) magnesium content. Along with insulin infusion, CMA improved glucose uptake, glucose metabolic clearance rate, and oxidative glucose metabolism. In the multiple linear regression analysis of the pooled basal data (n = 42), erythrocyte magnesium content displayed an independent correlation with basal plasma insulin levels (t = -2.08, P < .05).(ABSTRACT TRUNCATED AT 250 WORDS)