Three groups of patients with acute or chronic renal failure (GFR less than 5 ml/min) and hyperkalaemia (K+ greater than or equal to 6 mEq/l), similar in age, serum creatinine and pretreatment K+. Group A (n = 24) received salbutamol 0.5 mg i.v. in 15 min, group B (n = 10) received glucose 40 g i.v. plus 10 units insulin i.v. in 15 min, and group C (n = 10) received salbutamol 0.5 mg i.v., glucose 40 g i.v. and insulin 10 units i.v. over a 15-min period. Serum potassium was measured at 30, 60, 180 and 360 min after administration of treatment. All treatments reduced serum potassium, maximal at 30 or 60 min, and ranging from -0.5 +/- 0.1 to -1.5 +/- 0.2 mEq/l; patients in group C exhibited a significantly greater decrement in serum potassium, when compared to group B at 60 (-1.5 +/- 0.2 vs -1 +/- 0.1 mEq/l, respectively; P less than 0.01) and 180 min (-1.2 +/- 0.2 vs -0.7 +/- 0.1 mEq/l, respectively; P less than 0.05). There were no significant differences between groups A and C. Patients from group C had moderate tachycardia and more prolonged hyperglycaemia than those from group B, but all treatments were well tolerated.