Cerebral oedema during paediatric diabetic ketoacidosis in a four-year-old

Working diagnosis: cerebral oedema complicating diabetic ketoacidosis. The sequence is classic - headache, irritability then declining consciousness, with bradycardia and hypertension (Cushing's physiology) in a child whose glucose is being corrected. It occurs in under 1% of paediatric DKA but accounts for most of the deaths, and it can develop before treatment or up to 24 hours into it.

First 15 minutes.

1. Give hyperosmolar therapy immediately - do not wait for a CT. 3% saline 2.5–5 mL/kg (40–80 mL) over 10–15 minutes, or mannitol 0.5–1 g/kg (8–16 g) over 15 minutes if saline is unavailable. Clinical response matters more than the scan.
2. Halve the fluid rate rather than stopping insulin completely - reduce the maintenance infusion and keep the insulin running while adding dextrose, because stopping insulin restarts ketogenesis. Senior paediatric advice should be sought in parallel.
3. Position: head midline, head of bed up 30 degrees, ensure the airway is protected and that she is not obstructing; suction ready. Call for anaesthesia and intensive care.
4. Check and correct: bedside glucose (exclude hypoglycaemia), repeat sodium and osmolality, pupils every 15 minutes, GCS every 15 minutes.
5. CT head as soon as she is stable enough, to exclude dural sinus thrombosis, haemorrhage or infarction, which can mimic oedema and change management.
6. Avoid bicarbonate, hypotonic fluids, further rapid fluid boluses without shock, and rapid correction of sodium.

Escalation. If she has a further decrease in consciousness, unequal pupils, or respiratory depression, she needs intubation with normocapnia - but with the recognition that positive pressure and sedation can drop her blood pressure and cerebral perfusion, so the airway decision belongs to the most senior person present.

The case for 3% saline here. She has had a limited fluid resuscitation, delayed capillary refill initially and a sodium of 131 mmol/L - although the corrected sodium (measured plus roughly 0.3 × the glucose excess, about 139 mmol/L) shows she is not truly hyponatraemic, she is certainly not volume-overloaded. Hypertonic saline works as a small-volume, intravascular osmotic agent: it draws water from brain parenchyma without causing an osmotic diuresis, so it does not worsen hypovolaemia. It can be given through a peripheral line in an emergency and is easy to repeat.

The case for mannitol. It is familiar, rapid and effective, and 0.5–1 g/kg is a well-established dose. The drawbacks are relevant to this child: it causes an osmotic diuresis, which can produce hypotension in a child who is already volume-depleted and thereby reduce cerebral perfusion pressure; it can precipitate acute kidney injury; and it crystallises if cold, requiring a filter.

My choice: I would give 3% saline first because of the haemodynamic profile, and reserve mannitol for a second dose or if saline is unavailable.

Monitoring.

• Both: GCS, pupils, respiratory pattern and haemodynamics every 15 minutes; sodium, potassium and osmolality every one to two hours; strict input/output chart; glucose hourly.
• 3% saline: watch for hypernatraemia (a rapid rise above about 10–12 mmol/L per day outside acute herniation), fluid overload and pulmonary oedema, and extravasation injury; do not stop it abruptly.
• Mannitol: watch for hypotension, oliguria, rising creatinine, and a widening osmolal gap; replace the urine output with isotonic fluid.

The key principle is that the osmotic agent buys time while the underlying process - too-rapid fluid or glucose correction - is corrected and the airway and circulation are supported.

Appropriate.

• The 10 mL/kg saline bolus. She had tachycardia, delayed capillary refill and a systolic of 95; a single 10 mL/kg bolus over 30–60 minutes is the recommended approach when shock is present. Guidelines deliberately moved away from the old 20 mL/kg routine boluses because repeated large boluses are associated with cerebral oedema. One bolus was right; a second should have required reassessment.
• Isotonic fluid with potassium added. 0.9% saline avoids the free-water load of hypotonic fluid, and potassium supplementation at 20–40 mmol/L is standard once the potassium is known to be under about 5.5 mmol/L and the child is passing urine.
• Starting insulin at 0.05 units/kg/hour without a bolus, one hour after fluids, is correct. Insulin boluses increase the risk of cerebral oedema and of precipitous hypoglycaemia; the delay allowed the potassium to be checked.

What I would change.

• The 1.5 times maintenance rate should be computed explicitly and capped, with the deficit replaced evenly over 48 hours; rapid or front-loaded replacement is a recognised risk factor.
• Hourly glucose and two-hourly neurological observations should have been prescribed in writing, with the parents shown what to watch for. The headache and bradycardia were detectable before the GCS fell.
• Glucose should have been rechecked sooner and 5% dextrose added when it fell below about 14–15 mmol/L. A fall from 33 to 19 mmol/L in four hours is acceptable in rate, but the next four hours matter more.
• The corrected sodium should have been calculated and trended. A falling corrected sodium during treatment is an early warning of cerebral oedema, and it should trigger a review of the fluid tonicity and rate.
• Missing baseline work: HbA1c, thyroid function, coeliac screen, an ECG for potassium, and a documented plan for diabetes education and family support before discharge - because this is a new diagnosis, not just an acute episode.