diabetic ketoacidosis — pathophysiology
Contents
DKA results from absolute or relative insulin deficiency triggering three simultaneous cascades: hyperglycaemia, lipolysis-driven ketoacidosis, and osmotic diuresis with massive electrolyte depletion. Each derangement maps directly to a treatment decision in the first 24 hours.
the central event — insulin deficiency
Either absolute (new-onset T1DM, insulin omission) or relative (intercurrent illness increases requirements beyond supply). Both trigger three simultaneous cascades:
| cascade | mechanism | consequence |
|---|---|---|
| hyperglycaemia | impaired GLUT4 translocation → ↓ peripheral glucose uptake; loss of hepatic insulin suppression → ↑ gluconeogenesis + glycogenolysis | glucose rises, exceeds renal threshold (~10 mmol/L) → glycosuria |
| ketogenesis | unopposed lipolysis → FFAs flood liver → β-oxidation → acetoacetate, β-hydroxybutyrate (BHB), acetone | organic acid accumulation → high anion gap metabolic acidosis |
| osmotic diuresis | glycosuria drags water + electrolytes through nephron | 6–9 L water loss; massive K⁺, Na⁺, PO₄³⁻, Mg²⁺ depletion |
counter-regulatory hormone excess
Insulin deficiency alone won’t do it — the full syndrome needs unopposed counter-regulatory hormones:
- glucagon (most pivotal): drives hepatic gluconeogenesis, glycogenolysis, and ketogenesis; hyperketonaemia correlates directly with plasma glucagon levels
- catecholamines + cortisol: enhance lipolysis and hepatic glucose output
- growth hormone: worsens insulin resistance
Correcting either side — restoring insulin or suppressing counter-regulatory hormones — can reverse DKA. This explains the glucose drop from fluids alone (volume resuscitation damps counter-regulatory hormone drive).
why the potassium is a trap
Total body potassium is always depleted (3–6 mmol/kg deficit) despite a normal or high serum level at presentation.
Three forces shift K⁺ out of cells at presentation:
- insulin deficiency — loss of Na⁺/K⁺-ATPase stimulation
- hyperosmolality — solvent drag pulls K⁺ extracellularly
- acidaemia — H⁺/K⁺ exchange (less significant than historically taught, primarily with mineral acids)
Once treatment starts, all three reverse simultaneously:
- insulin drives K⁺ intracellularly via Na⁺/K⁺-ATPase
- fluid resuscitation lowers osmolality
- acidosis corrects
Result: serum K⁺ can plummet within 1–2 hours of insulin initiation → ventricular arrhythmia, respiratory muscle weakness, cardiac arrest.
Both insulin and bicarbonate shift K⁺ intracellularly. Giving either before repleting K⁺ can be fatal.
the acidosis — where the anion gap comes from
BHB accounts for >75% of the ketone burden. Ketoacid accumulation overwhelms bicarbonate buffering:
Unmeasured anions (BHB⁻, acetoacetate⁻) replace bicarbonate → AG rises. The treatment target is the AG, not glucose — glucose normalises well before ketosis resolves.
the two-phase acidosis of recovery
Failing to recognise this → unnecessarily prolonged insulin infusions.
- phase 1 — high anion gap acidosis: ketoacid accumulation (the DKA itself)
- phase 2 — non-anion gap (hyperchloraemic) acidosis: emerges during treatment from:
- renal excretion of ketoacid anions with Na⁺ (indirect bicarbonate loss)
- large-volume normal saline administration (chloride load)
This hyperchloraemic NAGMA is self-limited. AG closed but pH/bicarb still low? Check the chloride — stop chasing bicarb with more insulin.
osmotic diuresis — the electrolyte wasteland
Once glucose exceeds the renal threshold (~10 mmol/L), glycosuria drives osmotic diuresis. Typical deficits at presentation (70 kg adult):
| electrolyte | typical deficit | clinical consequence |
|---|---|---|
| water | 6–9 L | hypotension, prerenal AKI, ↑ counter-regulatory hormones |
| K⁺ | 3–6 mmol/kg (200–400 mmol) | arrhythmia, respiratory failure (see above) |
| Na⁺ | 7–10 mmol/kg | contributes to volume depletion |
| PO₄³⁻ | 1–1.5 mmol/kg | usually clinically insignificant; replete if <0.3 mmol/L |
| Mg²⁺ | variable | refractory hypokalaemia if not corrected |
Hypomagnesaemia prevents renal potassium conservation. If K⁺ is not responding to replacement, check and replace Mg²⁺ first.
euglycaemic DKA
SGLT2 inhibitors promote glycosuria (lowering glucose) while simultaneously increasing glucagon and reducing insulin → ketogenesis proceeds with near-normal glucose.
The trap: standard protocols tie insulin dose reductions to glucose thresholds. In euglycaemic DKA, glucose normalises early → insulin gets reduced prematurely → ketosis persists → acidosis correction stalls.
Risk factors: SGLT2i use + insulin dose reduction, fasting, low-carbohydrate diet, illness, surgery, dehydration, excessive alcohol.
ketone measurement — BHB vs nitroprusside
| assay | measures | problem in DKA |
|---|---|---|
| nitroprusside (urine dipstick) | acetoacetate only | misses BHB (the dominant ketone); may paradoxically turn more positive during recovery as BHB converts to acetoacetate |
| serum BHB (point-of-care) | β-hydroxybutyrate directly | preferred; correlates with severity; BHB <0.6 mmol/L = resolution |
why bicarbonate doesn’t help
Exogenous bicarb reacts with H⁺ → CO₂. The patient has to blow off that CO₂ to raise pH — but they’re already at maximal respiratory compensation. Additional risks:
- worsens intracellular acidosis (CO₂ crosses cell membranes freely; HCO₃⁻ does not)
- shifts K⁺ intracellularly → arrhythmia risk
- may delay ketone clearance
- no mortality benefit in trials
Most guidelines reserve bicarbonate for pH <6.9 only, and even this threshold is debated.
what proteolysis adds
Insulin deficiency triggers muscle proteolysis → released amino acids (alanine, glutamine) feed hepatic gluconeogenesis. This perpetuates hyperglycaemia even when the patient hasn’t eaten.