Methods

A literature search of Medline from 1993 to 1999 was performed linking the keywords “Hyperkalaemia”, “Management”, and “Treatment”. Medline was also searched under the subject heading “Hyperkalaemia-Treatment” from 1966 to 1999. The Cochrane library was searched under similar headings and there were no systematic reviews of the subject. References from articles recovered were searched for relevant studies.

The ideal design of a study of a treatment for hyperkalaemia is one that is randomised, blinded, and controlled against a placebo or standard therapy. Table 1 illustrates a summary of studies of the treatment of hyperkalaemia. Two studies^{9, 15} were randomised (the method of randomisation was not stated). Two studies^{2, 7} were compared with placebo. The plasma K in these studies were relatively low as it would be unethical not to treat severely hyperkalaemic patients as insulin with glucose is an established treatment in an otherwise potentially fatal condition. One study¹was double blinded. Six studies^2–15 had a crossover design to ensure uniformity in their comparisons.

Most studies examined patients with CRF; four studies^{4, 8, 11, 14} included patients with ARF. Except for the studies that included paediatric patients only,^{4, 14} the average age of the patients was more than 50 years. All studies excluded patients taking β blockers and most excluded patients taking digoxin. In seven trials^{2, 6, 7, 9, 10, 12, 15} the K was only mildly increased (K<6.0 mmol/l). In four studies^{2, 10, 12, 15} between 10% and 20% of patients did not complete the treatment, mostly because of going on to dialysis. Two studies^{6, 9} excluded non-responders (defined as patients who had a maximal reduction in K⁺ of <0.5 mmol/l after treatment) from their analysis.

Initial management

Treatment options include calcium (Ca) gluconate, insulin with glucose, salbutamol, sodium bicarbonate (NaHCO₃) and sodium (Na) polystyrene sulphonate.

CALCIUM GLUCONATE

There were no clinical studies looking solely at calcium in the management of hyperkalaemia. Ca gluconate antagonises cardiac membrane excitability and does not affect the plasma K.^{1, 19} It is generally accepted that calcium should be given when there are ECG changes associated with hyperkalaemia.^{7, 20} The sensitivity of emergency physicians diagnosing moderate to severe hyperkalaemia (K >6.5) from the ECG is only 62%.¹⁸ The ECG changes^{1, 21} include tall T waves >5mm (K 6–7), small broad P waves or absent P waves, wide QRS complex (K 7–8), sinusoidal QRST (K 8–9) and atrioventricular dissociation or ventricular tachycardia/fibrillation (K >9). Twenty ml 10% Ca gluconate¹⁹ is given intravenously in adults (0.5 ml/kg in children⁸) over 5–10 minutes and may be repeated as necessary. Onset of action is immediate but its duration is only a few minutes.^{1, 19} Hyperkalaemic patients taking digoxin should be given calcium as a slow infusion over 20 to 30 minutes. This avoids hypercalcaemia that may potentiate the myocardial toxicity of digitalis.

INSULIN WITH GLUCOSE

Insulin binds to specific membrane receptors and via an unknown second messenger, stimulates the sodium-potassium (Na-K) adenosine triphosphatase (ATP) pump resulting in intracellular uptake of K.⁵ This effect is independent of its hypoglycaemic action. Uraemia attenuates the hypoglycaemic response to insulin but does not affect its hypokalaemic action.^{22, 23} Insulin has been the traditional temporising treatment against which newer treatments are compared. It is indicated in every case of hyperkalaemia that needs emergency treatment. Ten units (in adults) soluble insulin is given with 40–60 g glucose intravenously as a bolus.

In children, a glucose load of 0.5 g/kg/h (2.5 ml/kg/h) should be given. This is because many of these patients increase their endogenous insulin production with the administration of a glucose load. If the blood glucose rises above 10 mml/l, insulin should be added at 0.05 u/kg/h.²⁴

Seven studies^{3, 5, 7, 10, 12, 17, 18} used insulin with glucose (table 2).

These studies show that the onset of hypokalaemic action is within 15 minutes and lasts for at least 60 minutes.^{7, 17} The reduction in K observed is 0.65–1.0 mmol/l.^{5, 10} Delayed (30–60 minutes post insulin) hypoglycaemia is common (up to 75% of patients¹⁰) if less than 30 g glucose is given.

SALBUTAMOL

Salbutamol binds to β₂ receptors in liver and muscle cells stimulating adenylate cyclase that converts ATP to 3′5′cyclic adenosine monophosphate. This stimulates the Na-K ATP pump resulting in intracellular K uptake.⁴ The response in patients on β blockers and digoxin is attenuated. One author²² has expressed reservation that there may be a hyperkalaemic response in the first three minutes of its administration that could be potentially deleterious. This fear was based on a baboon study²³ that used much higher doses intravenously (100 μg/kg) than have been used therapeutically in humans. Salbutamol 0.5 mg(4 μg/kg in children) is given intravenously or 10 mg of nebulised (Neb) salbutamol (in children: 2.5 mg if <25 kg or 5 mg if >25kg).

Thirteen studies^{2, 4–11, 13–16} used salbutamol (table 3).

Salbutamol produced a reduction in K of 0.87–1.4 mmol/l after intravenous administration^{5, 15} and 0.53–0.98 mmol/l after nebulisation.^{2, 7} Most studies used 10 mg nebulised salbutamol but 20 mg has been shown to be more effective at 120 minutes than 10 mg.² No difference was found when insulin with glucose was compared with intravenous salbutamol⁵ or nebulised salbutamol.¹⁰ Onset of action was within 30 minutes. There was no difference in maximum effect when intravenous salbutamol was compared with nebulised salbutamol.^{6, 9, 15} However, the maximum effect was in 30 minutes for intravenous administration compared with 90 minutes for nebulisation. Tremor and tachycardia were more pronounced after intravenous treatment. Caution has been advised for use in patients with ischaemic heart disease. Some authors^{6, 9} suggest nebulised treatment only in these patients.

Some 12–40% patients^{8, 10} were unresponsive to salbutamol and it should always be used in conjunction with insulin. The cause of this unresponsiveness is unknown.²²

Two studies^{5, 10} combined salbutamol and insulin with glucose (table 4). The combination of salbutamol and insulin was more effective than insulin alone in these studies. The hypoglycaemia associated with insulin was attenuated.

SODIUM BICARBONATE

NaHCO₃ has been recommended in textbooks for many years. It has no significant action on plasma K in the first 60 minutes after administration.^{3, 7, 12, 16} It may be indicated in severe metabolic acidosis (pH <7.2),^{19, 20} which may be an associated feature in ARF. Potential risks in giving NaHCO₃ include hypernatraemia, volume overload and tetany in patients with CRF and coexistent hypocalcaemia.

Four studies^{3, 7, 12, 16} examined the efficacy of NaHCO₃ (table 5) and all failed to show any reduction in K within 60 minutes.

These trials did not include patients with severe hyperkalaemia or severe metabolic acidosis. Allon and Shanklin⁷ reported that NaHCO₃ had no additive effect on the action of insulin or salbutamol while Kim^{3, 16} reported that NaHCO₃ increased the effect of insulin and salbutamol. The patients in Kim's studies had a higher initial plasma K.

SODIUM POLYSTYRENE SULPHONATE (KAYEXALATE)

There were no clinical studies looking specifically at the efficacy of Na polystyrene sulphonate in the management of hyperkalaemia. This resin binds K in the intestinal lumen, especially large bowel and ileum.¹⁹ It may be indicated if haemodialysis is delayed (>2–3 hours).¹⁹ Fifty grams Na polystyrene sulphonate in 100–200 ml 30% sorbitol or 10% glucose at 37°C is given rectally and left for at least 60 minutes. Sorbitol is added as it increases faecal K excretion. The onset of action is slow, approximately one to two hours. One gram resin exchanges 1mEq Na for 1mEq K.¹ Rarely, ulceration of the bowel mucosa can occur, especially in constipated patients, and is probably attributable to to the action of sorbitol.²⁵ Concurrent use of a laxative helps prevent faecal impaction.

HAEMODIALYSIS

This is the definitive and most effective hypokalaemic measure.^{1, 22} It is indicated in severe hyperkalaemia. Mild to moderate hyperkalaemia in CRF may be managed without haemodialysis as an emergency depending on the cause. It has been reported to be of benefit in patients who have had a hyperkalaemic cardiac arrest when drug measures have failed.²⁶ During haemodialysis, plasma K falls rapidly in the first hour and very little thereafter.¹² If a potassium free dialysate is used, serum potassium may decrease as much as 1.2 to 1.5 mEq/h.²⁷ Potassium concentrations show a rebound after dialysis has finished and this rebound may require several hours to reach a plateau.²⁸ If the potassium is lowered with other agents such as salbutamol, this may reduce potassium removal during subsequent dialysis treatment, even though the achieved potassium concentration is lower.²⁹

Conclusions

The emergency management of hyperkalaemia should be tailored to the individual patient. It involves (1) determining the cause and (2) instituting temporising measures to stabilise the myocardium and lower the plasma K by redistribution to the intracellular compartment while (3) arranging haemodialysis if necessary. Urgency of treatment is dependent on rate of rise of K in addition to the plasma concentration.

Using common sense interpretation of the available studies, moderate to severe hyperkalaemia in the emergency department should be treated (after confirmation) with (1) Ca gluconate if there are any ECG changes followed by (2) insulin with glucose and (3) intravenous salbutamol (nebulised if there is evidence of ischaemic heart disease). NaHCO₃ does not lower the K in the first 60 minutes but is given if there is severe metabolic acidosis (pH <7.20). Sodium polystyrene sulphonate can be given if there is an anticipated delay in haemodialysis, which is the definitive treatment in these patients.