Evidence is accumulating that insulin is a hypertensive factor in humans. The involved mechanism may be its sodium-retaining effect. We examined whether insulin causes sodium retention through a direct action on the kidney, as is generally assumed, or indirectly through hypokalemia. Insulin was infused (euglycemic clamp technique) with and without potassium infusion to prevent hypokalemia in six healthy subjects. Without potassium infusion, insulin caused a marked decrease in plasma potassium (-0.75 mmol/L), and decreased urinary sodium and potassium excretions by, approximately 38% and 65%, respectively. Simultaneous potassium infusion largely prevented the decrease in plasma potassium, as well as the decrease in urinary sodium and potassium excretions. These data suggest that the acute antinatriuretic effect of insulin may be largely mediated in an indirect way, ie, through hypokalemia.