New-Onset Diabetes at 6 Years: Challenging Two Orders in Pediatric DKA
Leo, 6 years old and 22 kg, arrives in the emergency department at 22:00 after three days of vomiting and abdominal pain with increasing drowsiness. His mother reports two weeks of heavy drinking and urination and a 2 kg weight loss; he was previously healthy with no known diabetes. He is drowsy but rousable, with deep sighing respirations at 34/min, dry mucous membranes, capillary refill of 3 seconds, HR 138, BP 92/58 mmHg, and temperature 37.9 °C. He weighed 24 kg at his five-year visit. Point-of-care glucose is 512 mg/dL. Venous gas: pH 7.14, pCO2 22 mmHg, HCO3 8 mmol/L. Potassium 5.1 mmol/L, sodium 131 mmol/L, chloride 98 mmol/L, BUN 24 mg/dL, creatinine 0.6 mg/dL, large urine ketones. The provider orders: "0.9% NaCl 20 mL/kg IV bolus over 30 minutes; regular insulin 0.1 units/kg IV push, then 0.05 units/kg/h infusion; maintenance fluids with 40 mEq/L KCl once urine output is documented."
1. Which two of these orders do you question, and why? What do you propose instead?
Read the worked answer
Order 1 - the 20 mL/kg bolus over 30 minutes. A 20 mL/kg rapid bolus is an adult and trauma habit. In pediatric DKA the brain is already sitting on an osmolar knife-edge, and rapid fluid shifts are associated with cerebral edema - the leading cause of death in pediatric DKA. For a child who is not in frank shock, the standard is 10 mL/kg over 30–60 minutes, repeated only if circulation demands it, followed by slow rehydration over 24–48 hours. He has tachycardia and delayed refill, so he needs some volume, but 20 mL/kg is twice the recommended first step. If he were in hypotensive shock, a larger bolus would be justified - he is not; he is 92/58 with a normal creatinine. My recommendation is 10 mL/kg of an isotonic balanced solution, reassess, and repeat once if needed.
Order 2 - the insulin bolus. An IV bolus of 0.1 units/kg before fluid resuscitation is precisely what guidelines advise against in pediatric DKA. Reasons: it drives potassium intracellularly in a child whose potassium is about to fall; it accelerates the osmolar drop and the risk of cerebral edema; and it is unnecessary because an infusion alone achieves the same correction more smoothly. Proposed alternative: begin the insulin infusion at 0.05–0.1 units/kg/h after fluid resuscitation has started, typically about an hour in. No bolus.
One more thing I would flag constructively, not as a third objection: the KCl order is directionally correct - total body potassium is depleted even though serum K is 5.1 - but 40 mEq/L should not be hung while the potassium is 5.1. I would recheck the potassium after the first hour of fluids and add potassium once it is below about 5.5 and urine output is confirmed, and I would never add it to a fluid that is also being used as a bolus.
How I raise it: I would not simply refuse. I would say, "I want to give this safely - pediatric DKA protocols use 10 mL/kg for the first bolus and no insulin bolus. Can we confirm the orders before I start?" and document the conversation and the outcome.
2. What is the complication you are most worried about in the next 12 hours, what are its earliest signs, and what will you do if they appear?
Read the worked answer
Cerebral edema. It occurs in roughly 0.5–1% of pediatric DKA episodes and accounts for most of the deaths. It can appear during treatment - which is why treatment itself is the risk factor - and it is notoriously easy to dismiss as "he's just tired."
Earliest signs, in the order I would expect them:
- Headache - in a 6-year-old, this may be reported as "my head hurts" or as irritability and photophobia.
- Change in level of consciousness - decreasing responsiveness, inappropriate speech, restlessness, or new incontinence that was not there on arrival.
- Vital sign pattern: bradycardia with hypertension (the Cushing response), which is a late sign and means the process is already established.
- Respiratory: decreasing oxygen saturation, irregular respirations, or a sudden drop in respiratory effort.
- Pupils: unequal or sluggish - again late.
Laboratory warning: a falling corrected sodium is one of the most useful early signals. His measured sodium is 131 with a glucose of 512, so the glucose-corrected sodium is roughly 137–138 - not low. If the measured sodium drifts down while the glucose falls, the effective osmolality is dropping too fast and the brain is taking on water. I would compute this corrected value at every glucose check, not just read the sodium off the panel.
Actions if signs appear (this is a rapid response, not a page-and-wait):
- Stop or slow the fluid, elevate the head of the bed, and get the provider and the intensivist to the bedside.
- Give mannitol 0.5–1 g/kg IV over 10–15 minutes, or 3% hypertonic saline 2.5–5 mL/kg, as ordered - the order set should be located before it is needed, at 22:00, not at 03:00.
- Reduce the insulin infusion rate as directed; do not stop it and do not give bicarbonate.
- Prepare for intubation and transfer to a pediatric intensive care unit.
Prevention is the real answer: a slow glucose decline (roughly 50–100 mg/dL/h), no bolus insulin, and hourly neurological checks documented in a way that lets the next nurse see the trend.
3. Show how you would manage the potassium and the glucose trend for the first 12 hours, with numbers.
Read the worked answer
Potassium Serum potassium 5.1 looks reassuring and is misleading: insulin and fluids will drive it into cells and total body potassium in DKA is always depleted (he has been vomiting for three days). The dangerous move is adding potassium while the serum value is 5.1; the also-dangerous move is failing to add it once the value falls.
- Now: no potassium in the bolus. Confirm urine output - he is catheterized or monitored, and 0.5 mL/kg/h is roughly 11 mL/h for 22 kg.
- At 1–2 hours: recheck potassium with the repeat chemistry. When it is below about 5.5 and he is making urine, add 40 mEq/L KCl to the maintenance fluid (the order already says this - the timing is what needs correcting).
- If potassium falls below 3.5: hold the insulin and call, because insulin without potassium drives the level into arrhythmia territory.
- Watch the monitor for peaked T waves early and flattened T waves, U waves, and ectopy later; continuous cardiac monitoring is standard.
Glucose trend
- Target decline: 50–100 mg/dL per hour. From 512, I would expect roughly 450 at hour one, 380 at hour two. A drop of 200 mg/dL in an hour is too fast and raises cerebral edema risk.
- Check point-of-care glucose hourly, and also at any change in mental status.
- When glucose reaches about 250–300 mg/dL, add dextrose (typically 5% dextrose in the maintenance fluid) and continue the insulin - do not stop it, because insulin is needed to clear the ketones and closing the anion gap is the endpoint, not the glucose number.
- Track the gap: bicarbonate 8 and pH 7.14 at baseline; repeat the gas as ordered. Resolution is pH above 7.3, bicarbonate above 15, and a closed anion gap. If the glucose falls nicely but the gap does not close, the insulin is not doing the job and the infusion rate is reassessed.
- Do not give bicarbonate - it worsens intracellular acidosis and cerebral edema risk.
Family teaching started early: the first injection, the first fingerstick, and the words "this is type 1 diabetes and it is not your fault" belong in this admission, and the diabetes educator should be involved before discharge, not after.