NEET-SS study materials
NEET-SS study resources

Combined DKA and hyperosmolar state with acute kidney injury in a patient who stopped an SGLT2 inhibitor two days ago

NEET-SS · Put your knowledge to work in context.

All case studies

Combined DKA and hyperosmolar state with acute kidney injury in a patient who stopped an SGLT2 inhibitor two days ago

A 48-year-old man with type 2 diabetes presents after 4 days of polyuria, vomiting and confusion. He stopped empagliflozin 2 days ago and self-discontinued insulin glargine 30 units nightly a week earlier. Glasgow Coma Scale is 12. Blood pressure 96/60 mmHg, pulse 122/min, temperature 38.2 °C, respiratory rate 30/min with deep respiration, urine output 15 mL/h. Weight 85 kg. Investigations: glucose 1180 mg/dL, measured sodium 129 mmol/L, potassium 5.6 mmol/L, bicarbonate 14 mmol/L, venous pH 7.26, beta-hydroxybutyrate 4.8 mmol/L, effective serum osmolality 338 mOsm/kg, creatinine 2.1 mg/dL (baseline 0.9), amylase mildly raised, lipase normal. ECG shows sinus tachycardia with no ischaemic change. No obvious infection source on examination; chest radiograph is clear.

1. Classify the hyperglycaemic emergency, estimate the water deficit, and lay out the first 4 hours of fluid, insulin and potassium management with the correction limits you would enforce.

Read the worked answer

Classification: combined diabetic ketoacidosis and hyperosmolar hyperglycaemic state. pH 7.26 with bicarbonate 14 and beta-hydroxybutyrate 4.8 meets DKA criteria, while glucose well above 600 mg/dL and effective osmolality 338 mOsm/kg meet HHS criteria. Treating it as one or the other leads to error - HHS needs slower correction, DKA needs insulin and ketone clearance.

Water deficit: roughly 100–120 mL/kg = 8.5–10 L in an 85 kg man.

First hour: 1 L of 0.9% sodium chloride (faster if shocked, with vasopressors if MAP remains <65 mmHg). Then reassess.

Hours 2–4: guided by corrected sodium and haemodynamics - approximately 500 mL/h, switching to 0.45% saline if the corrected sodium is high, and continuing isotonic if it is low. Enforce: fall in osmolality ≤3 mOsm/kg/h, sodium rise ≤10 mmol/L/24 h, glucose fall 50–75 mg/dL/h. Use corrected sodium (add 1.6 mmol/L per 100 mg/dL glucose above 100) rather than the measured value.

Insulin: start a fixed-rate infusion of 0.1 U/kg/h after the first fluid bolus and only once potassium is confirmed ≥3.3 mmol/L. Do not give a bolus in this combined picture with hyperosmolality; the aim is a controlled fall, not a rapid one.

Potassium: K 5.6 with poor urine output now, but insulin will drive it intracellularly - check hourly and add 20–40 mmol/L replacement once K falls below 5.0–5.5 and urine output is established. Bicarbonate only if pH <6.9; phosphate only if <1 mg/dL or with cardiac/respiratory depression.

2. Which routine measures and drugs would you actively avoid in this patient, and why?

Read the worked answer

A deliberate 'do-not-do' list:

  • Bicarbonate at pH 7.26. It worsens hypokalaemia, causes paradoxical intracellular and CSF acidosis, shifts the oxygen dissociation curve, and increases the risk of cerebral oedema; the threshold is pH <6.9.
  • Rapid osmolality correction or hypotonic fluid boluses. The brain has adapted with idiogenic osmoles; rapid lowering draws water into the brain. Cerebral oedema is uncommon in adults but catastrophic.
  • Insulin before fluid resuscitation. Insulin pulls glucose and water into cells before the intravascular volume is restored, and can precipitate vascular collapse and shock.
  • Subcutaneous sliding-scale insulin as the main therapy. Absorption is unreliable in a hypoperfused, dehydrated patient; an infusion is required.
  • Restarting empagliflozin now or assuming its effect has gone. Euglycaemic/normoglycaemic ketoacidosis can persist for days after the last dose; hold it until the patient is eating, ketone-free and stable, and counsel on sick-day rules.
  • Potassium-containing fluids while K >5.5, and calcium-containing fluids mixed with bicarbonate in the same line.
  • Phosphate replacement for a mildly low value, and routine antibiotics without a source - pyrexia in DKA can be caused by the metabolic state itself, though cultures should be sent and antibiotics started promptly if sepsis is suspected.
  • Fluid loading without regard to urine output and renal function - he has acute kidney injury and may need less than the naive protocol volumes.

3. Six hours later glucose is 380 mg/dL, bicarbonate 19 mmol/L and GCS 14, but he develops a headache and bradycardia. What are your differential and immediate steps?

Read the worked answer

The differential is short and urgent.

  1. Cerebral oedema from over-rapid osmolality correction - the most feared explanation, especially if the sodium change exceeded 10 mmol/L/24 h or osmolality fell faster than 3 mOsm/kg/h. Headache, declining consciousness and bradycardia with hypertension (Cushing reflex) suggest raised intracranial pressure.
  2. Hypoglycaemia or over-sedation - check capillary glucose immediately; insulin infusions outrun dextrose easily.
  3. Hypokalaemia-related bradyarrhythmia or conduction disturbance - an ECG is mandatory.
  4. Intracranial event - ischaemic stroke, venous sinus thrombosis (dehydration is prothrombotic), or subarachnoid haemorrhage; he also has acute kidney injury and possible uraemia.
  5. Infection - meningitis or a new septic focus, given the earlier pyrexia.

Immediate steps: airway and breathing assessment, full GCS and pupillary examination, capillary glucose, ECG and continuous monitoring, urgent non-contrast CT head (and consideration of CT venography), and review of the sodium and osmolality trend from the fluid charts and labs.

Management while imaging: stop any hypotonic fluid and slow the correction; elevate the head of the bed; treat hypoglycaemia if present; correct potassium and magnesium. If there are signs of raised intracranial pressure with impending herniation, give mannitol 0.5–1 g/kg or hypertonic saline after senior and neurosurgical discussion, and arrange critical care. Do not give further insulin boluses blindly, and do not attribute the bradycardia to 'being in DKA' - in a patient who was improving, the change in neurological status is the alarm.