Twenty patients receiving long-term (1.0 to 10.3 years) thiazide diuretic treatment for essential hypertension (Th group) received magnesium supplementation as MgO (600 mg Mg/day) for 4 weeks and placebo for another 4 weeks. Before and at the end of each period, we measured blood pressure; intraerythrocyte magnesium, sodium, and potassium (R-Mg, R-Na, and R-K); and erythrocyte ouabain (10-4 M) sensitive sodium efflux rate constant (Kos). The same measurements were also performed in 21 untreated age-matched essential hypertensives (EHT group). In the Th group, R-Mg was lower, R-Na was higher, and Kos was lower than in the EHT group before magnesium supplementation. Oral magnesium resulted in a significant increase in R-Mg (p less than 0.005) and a decrease in R-Na (p less than 0.01), together with an increase in Kos (p less than 0.005) in the Th group. During magnesium supplementation, the increased Kos was correlated negatively with the lowered R-Na (r = -0.57, p less than 0.01) and positively with the increased R-Mg (r = 0.61, p less than 0.005). Systolic, diastolic, and mean blood pressures were reduced significantly during magnesium administration by a mean of 7.5, 3.0, and 4.5 mm Hg. The results indicate that long-term diuretic treatment may give rise to an intracellular magnesium deficiency and a suppression in cell membrane active transport for sodium, with a resultant increase in intracellular sodium content.(ABSTRACT TRUNCATED AT 250 WORDS)