We demonstrated similar plasma concentrations and urinary losses but lower erythrocyte magnesium concentrations (2.18 +/- 0.04 vs 1.86 +/- 0.03 mmol/L, P less than 0.01) in twelve aged (77.8 +/- 2.1 y) vs 25 young (36.1 +/- 0.4 y), nonobese subjects. Subsequently, aged subjects were enrolled in a double-blind, randomized, crossover study in which placebo (for 4 wk) and chronic magnesium administration (CMA) (4.5 g/d for 4 wk) were provided. At the end of each treatment period an intravenous glucose tolerance test (0.33 g/kg body wt) and a euglycemic glucose clamp with simultaneous [D-3H]glucose infusion and indirect calorimetry were performed. CMA vs placebo significantly increased erythrocyte magnesium concentration and improved insulin response and action. Net increase in erythrocyte magnesium significantly and positively correlated with the decrease in erythrocyte membrane microviscosity and with the net increase in both insulin secretion and action. In aged patients, correction of a low erythrocyte magnesium concentration may allow an improvement of glucose handling.