An outline of management is presented: this should be tailored to individual circumstances. Management of DKA has passed through 3 stages in the last 100 years:
Mortality with the low dose insulin regime is down to about 2 to 5% overall. In older patients with DKA precipitated by a major medical illness (eg acute pancreatitis, myocardial infarction, septicaemia), the mortality rate is still high due to the severity of the precipitating problem.
Management can be considered in terms of emergency and routine components.
Cerebral oedema is a dangerous complication that occurs in about 1% of children and adolescents with DKA.
Onset of headache and deteriorating level of consciousness typically occurs between 2 and 24 hours after onset of treatment. Onset of symptoms is often sudden. Mortality is about 70% in this group.
Recommended treatment is immediate IV mannitol in a dose of 0.5 to 2.0 g/kg body weight. Dexamethasone or hyperventilation have no proven benefit. (Lebovitz, 1995)
Immediate aim is to restore intravascular volume to improve tissue perfusion.
Replacement solutions (eg Normal saline or Hartmann’s solution) are appropriate for initial management. Subsequently fluids need to be adjusted to provide ‘free water’ to replenish intracellular fluid and to provide glucose. Maintenance fluids such as dextrose-saline or oral fluid intake are appropriate at this later stage depending on the individual circumstances but such solutions should not be used initially.
Colloids are necessary only in shocked patients. Colloids are expensive and have a low but significant risk of reactions. Albumin solutions are not required.
Serum level is commonly normal or high (due to the acidosis) at presentation despite the presence of a large total body potassium deficit (due to renal losses). The best approach is to commence therapy with fluid and insulin and monitor the serum [K+].
Potassium replacement can be commenced when the [K+] falls below 5 mmols/l. Infuse at 10 to 30 mmol/hr dependent on [K+]. Rates greater than 20 mmols/hr are reserved for severe hypokalaemia and require at least hourly [K+] monitoring. Never commence a potassium infusion without checking the level.
Fluid resuscitation is necessary to deliver insulin to its sites of action in liver, muscle and adipose tissue. Rehydration itself will cause a fall in blood glucose level.
A typical regime would be to give a stat dose initially (say 10-20U IV) and commence the patient on a continuous insulin infusion at 5 to 10 U/hr decreasing to 1-3 U/hr to maintain blood glucose at 5 to 10 mmols/l. A paediatric regime would be: insulin at 0.1U/kg IV loading dose then infusion at 0.1U/kg/hr.
The blood glucose always falls on this regime and control of blood glucose is almost never a problem. Insulin reverses the peripheral mobilisation of FFA and alters hepatic metabolism to switch off ketone body production. These effects are maximal at insulin levels of 100 micromoles/l and this level is achieved with the low dose regime. The average rate of fall of plasma glucose at this insulin level is about 4.5 mmol/l/hr. There is no advantage in giving more insulin once the ceiling level is reached. This absence of additional effectiveness with very high insulin levels has been referred to in the past as insulin resistance
Though a total body deficiency is always present, it has not been possible to show that acute phosphate administration makes any difference to outcome. However the occasional patient develops extremely low phosphate levels and phosphate administration is undoubtably necessary in these patients and must be given. Phosphate level on presentation is typically high so phosphate administration should be delayed.
By twelve hours after commencement of treatment, the majority (90%) of patients are hypophosphataemic. Ampoules of phosphate available in my hospital contain about 15 mmoles of phosphate and 20 mmoles of potassium and one ampoule can be diluted in the IV fluids and infused over an hour.
Sodium bicarbonate in DKA has arguably a minor role is in urgent management of serious arrhythmias due to hyperkalaemia in DKA. However, glucose-insulin is the preferred treatment in this patient group.
None of the studies done in DKA have shown any benefit of bicarbonate treatment. Potential problems are sodium overload, CSF acidosis, intracellular acidosis, exacerbation of hypokalaemia, rebound alkalosis and impaired tissue oxygen delivery (shift of oxyhaemoglobin dissociation curve). After treatment of DKA starts, the slowest biochemical parameter to recover is usually the serum bicarbonate - this is especially so when substantial amounts of ketones have been lost in the urine. New bicarbonate is generated when the condition is reversed and the ketones are metabolised. Bicarbonate administration is not necessary.
Management in an Intensive Care Unit is recommended.
Monitoring should include observations of airway, breathing, circulation and level of consciousness, serial blood gases and electrolytes, urinary ketones and urine output. Serum lactate is occasionally useful. A Biochemistry Flowchart of results is strongly recommended.
The commonest precipitants in young diabetics are inadequate insulin (eg first presentation of diabetes, omission of doses) and infection. Often no specific cause can be found. In older diabetics, DKA may be precipitated by a major medical illness (esp infection). Antibiotics or surgical management are necessary in some cases. Patient education to prevent further episodes is very important.