A Type 1 Diabetic in Diabetic Ketoacidosis and a Questionable First-Hour Order Set
A 19-year-old college student with type 1 diabetes of four years' duration was admitted at midnight drowsy and dehydrated. He had gastroenteritis for two days, could not eat, and stopped his insulin pump because he feared hypoglycaemia. On arrival: pulse 118/min, blood pressure 96/60 mmHg, respiratory rate 32/min and deep, breath smells of acetone, capillary glucose 28.6 mmol/L, arterial pH 7.08, bicarbonate 6 mmol/L, potassium 5.6 mmol/L, sodium 130 mmol/L, chloride 96 mmol/L, urea 12 mmol/L, urine ketones 3+, blood ketones 5.2 mmol/L. The admitting intern writes orders for 20 units of regular insulin intravenously as a bolus, 100 mmol of sodium bicarbonate in the first hour, and 1 L of 5% dextrose. The ward stocks 0.9% saline, Ringer's lactate, 5%, 10% and 25% dextrose, and potassium chloride concentrate; there is one ICU bed and the patient is currently on the general ward.
1. Critique the intern's three orders and state what the first hour should actually contain.
Read the worked answer
All three orders are wrong.
- 20-unit insulin bolus: intravenous boluses of this size cause abrupt falling glucose and osmolality, shift potassium into cells, and are associated with cerebral oedema, especially in young patients. Insulin should be an infusion.
- Bicarbonate at pH 7.08: bicarbonate is not indicated above pH 6.9, and its risks are discussed below.
- 5% dextrose as the initial fluid: the patient is profoundly volume-depleted; dextrose alone does not restore circulating volume and will not close the ketone gap.
First hour:
- Assess airway, breathing, circulation; start oxygen if needed; place on hourly monitoring and move to the ICU bed, because this patient needs hourly biochemistry.
- 0.9% saline, roughly 15-20 mL/kg (about 1-1.5 L in an adult) in the first hour, adjusted for cardiac status.
- Draw potassium, sodium, bicarbonate, glucose, blood gas and ketones before treatment and repeat hourly initially.
- Start a fixed-rate intravenous insulin infusion of about 0.1 unit/kg/hour, after the fluid bolus has begun - but not before potassium is known.
- Once potassium is below about 5.5 mmol/L and urine output is confirmed, add 20-40 mmol/L potassium chloride to the fluids; if potassium is below 3.3 mmol/L, give potassium before starting insulin.
- Add 5% or 10% dextrose to the saline once glucose falls to about 10-15 mmol/L, while continuing insulin to clear ketones.
2. Explain why his serum potassium is 5.6 mmol/L when his body is potassium-depleted, and why insulin is dangerous in that setting.
Read the worked answer
Total body potassium is low even though serum potassium is high. Insulin deficiency and hyperosmolality reduce the activity of membrane Na+/K+-ATPase, so potassium leaks out of cells into the extracellular fluid; acidosis adds to this by favouring export of potassium in exchange for hydrogen ions. In parallel, the osmotic diuresis of hyperglycaemia, plus vomiting and poor intake from gastroenteritis, have removed large amounts of potassium in urine and vomit. The serum value therefore reflects distribution, not stores: total body deficit in ketoacidosis is typically several hundred millimoles.
Why insulin is dangerous. Insulin restores Na+/K+-ATPase activity and drives potassium back into cells within minutes. In a patient already depleted, this can drop serum potassium precipitously, causing arrhythmias, muscle weakness and respiratory compromise - deaths in ketoacidosis are more often from potassium than from glucose. His potassium of 5.6 mmol/L is therefore a comfortable-looking number that will fall fast, which is why potassium must be checked hourly at first and replacement started once it is below about 5.5 mmol/L with adequate urine output. Insulin must not be delayed to 'normalise' potassium, but it must not be given without a plan for potassium. Bicarbonate would worsen matters by driving potassium into cells as well.
3. Justify withholding bicarbonate at pH 7.08, and state the conditions under which it would be considered.
Read the worked answer
Why withhold it. Bicarbonate given systemically is converted to carbon dioxide, which crosses the blood-brain barrier far more readily than bicarbonate ion does. The result is a paradoxical fall in cerebrospinal fluid pH while arterial pH rises - a well-recognised trigger of cerebral oedema, which carries high mortality in adolescents and young adults. There are further harms: alkalinisation shifts the oxyhaemoglobin dissociation curve to the left, reducing oxygen delivery to already hypoperfused tissue; it drives potassium and ionised calcium into cells, risking arrhythmia and tetany; it can produce overshoot metabolic alkalosis; and the shift in the oxygen curve can paradoxically worsen tissue acidosis. Randomised evidence has not shown improved survival or faster resolution of ketosis, and resolution depends on insulin and fluid, not on alkali.
When it is considered. Current guidance reserves bicarbonate for severe acidaemia, generally arterial pH below about 6.9, and even then it is given cautiously - for example 100 mmol in 400 mL of isotonic fluid with 20 mmol potassium chloride over about two hours, with repeat blood gases, and with vigilant monitoring for cerebral oedema and hypokalaemia. At pH 7.08 this patient does not qualify.
What he does need is the underlying treatment: volume repletion, a fixed-rate insulin infusion, potassium replacement, treatment of the gastroenteritis precipitant, and structured education that insulin must never be stopped during illness - the 'sick day rules' conversation that would have prevented the admission.