Thiazide can cause magnesium depletion, which may exaggerate renal potassium wasting and hypokalemia. The purpose of this double-blind, randomized trial was to compare the metabolic effects of potassium-magnesium-citrate (K-Mg-citrate) and potassium chloride (KCl) during long-term treatment with thiazide. Twenty-two normal volunteers received hydrochlorothiazide 50 mg/d. Ten subjects concurrently took K-Mg-citrate (42 mEq K/d and 21 mEq Mg/d), and 12 subjects were given KCl 42 mEq/d. Serum potassium concentration remained unchanged during K-Mg-citrate supplementation, with a change from baseline of 21.7% over 6 months, compared with 26.4% with KCl supplementation. Serum electrolytes were normal and not significantly different between K-Mg-citrate and KCl. During K-Mg-citrate treatment, serum magnesium increased significantly by about 10%, associated with an adequate increase in urinary magnesium and a nonsignificant increase in monocyte and free muscle magnesium. Serum magnesium was unchanged, and monocyte and free muscle magnesium showed a nonsignificant decline during KCl supplementation. K-Mg-citrate provided an alkali load, increasing urinary pH, and reducing urinary undissociated uric acid. It also increased urinary citrate and tended to lower the saturation of calcium oxalate. KCl supplementation lacked these actions. K-Mg-citrate prevents thiazide-induced hypokalemia without provoking metabolic alkalosis. It seems to prevent magnesium depletion. By providing an alkali load, it retards the propensity for the crystallization of uric acid and probably of calcium oxalate. Though not conclusive, KCl supplementation may be less effective than K-Mg-citrate in maintaining normokalemia because of a subtle magnesium wasting. Moreover, KCl is devoid of protective action toward crystallization of stone-forming salts.