BACKGROUND: Epidemiological studies demonstrate an inverse association between serum magnesium and incidence of cardiovascular disease. Diuretics commonly cause hypomagneseamia. METHOD: We evaluated effects of magnesium supplementation on blood pressure (BP) and vascular function in thiazide-treated hypertensive women in a randomized, double-blind, clinical trial. Hypertensive women (40-65 years) on hydrochlorothiazide and mean 24-h BP at least 130/80 mmHg were divided into placebo and supplementation (magnesium chelate 600 mg/day) groups. Patients were evaluated for nutritional and biochemical parameters, office and ambulatory blood pressure monitoring, brachial flow-mediated dilatation (FMD), peripheral arterial tonometry, assessment of carotid intima-media thickness, central hemodynamic parameters and pulse wave velocity at inclusion and after 6-month follow-up. RESULTS: The magnesium group had a significant reduction in SBP (144 ± 17 vs. 134 ± 14 mmHg, P = 0.036) and DBP (88 ± 9 vs. 81 ± 8 mmHg, P = 0.005) at 6 months, without effect on plasma glucose, lipids, or arterial stiffness parameters. The placebo group showed a significant increase in carotid intima-media thickness (0.78 ± 0.13 vs. 0.89 ± 0.14 mm, P = 0.033) without change in the magnesium group (0.79 ± 0.16 vs. 0.79 ± 0.19 mm, P = 0.716) after 6 months. The magnesium group demonstrated a significant increase in variation of FMD vs. the placebo group (+3.7 ± 2.1 vs. 2.4 ± 1.2%, P = 0.015). There was a significant correlation between the intracellular magnesium variation and FMD (r = 0.44, P = 0.011). CONCLUSION: Magnesium supplementation was associated with better BP control, improved endothelial function and amelioration of subclinical atherosclerosis in these thiazide-treated hypertensive women.