Endocrine‎ > ‎



  • Diabetic Ketoacidosis (DKA) is defined by hyperglycemia (blood glucose >200 mg/dl) and metabolic acidosis (pH <7.3 and/or plasma bicarbonate <15 mEq/L) with ketonemia and ketonuria. (ISPAD 2014 Consensus Guidelines)
  • The severity is categorized via the degree of acidosis: mild (pH 7.2-7.3), moderate (pH 7.1-7.2), or severe (pH <7.1).
  • Hyperketosis and hyperosmolarity are other key features. 
  • It is frequently the initial presentation of of children with new onset T1DM (30-67%), particularly in younger children and children from lower socioeconomic backgrounds. It can also occur with T2DM (for example, up to 40% in obese African American adolescents with type 2 DM initially present in DKA)1.


  • Results from a relative lack of insulin with subsequent lipolysis, ketogenesis, gluconeogenesis, and glycogenolysis
  • Missed insulin common cause, particularly in adolescents
  • Stress, via secretion of catecholamines, glucagon, and cortisol leads to glucose and ketoacid production and can precipitate DKA. Hence, infection is commonly an inciting factor in DKA
Figure 1: Pathophysiology of DKA

  • Medications such as corticosteroids and atypical antipsychotics can also precipitate DKA
  • Hyperglycemia -> Osmotic diuresis as renal threshold of glucose is met (~200) -> Hypovolemia, total body depletion of potassium and phosphorous
  • Ketosis (as fuel for the brain) -> anion gap acidosis (AG= Na - HCO3 - Cl with normal ~6-10)
  • Anion gap acidosis-> Respiratory compensation with rapid deep breathing (Kussmaul respirations) to compensate can lead to very low PaCO2 levels
  • Nausea/vomiting/abdominal pain thought to be partially due to ketosis
  • Potassium, though total body deplete, is often high in the serum due to the relative lack of insulin (normally drive K+ into cells) and the acidosis (H+/K+ antiport shifts K+ into the blood and H+ into cells to help buffer the acidosis)


  • Obtain baseline labs (Renal panel, VBG, blood glucose, urine ketones-possibly). This will allow you to calculate the anion gap and classify the severity of DKA
  • 20 cc/kg bolus of normal saline or lactated ringers for initial fluid hydration. This will drop the serum glucose as it should help increase the patient's GFR. Patients rarely need more than this initial bolus as although they are volume depleted, the intravascular space is relatively protected as a result of hyperosmolarity
  • Insulin infusion (standard 0.1 u/kg/hr), generally without insulin bolus
  • Order two bags for IV fluids: NS + 20 mEq/L KCL and 20 Kphos and D12.5 NS + 20 mEq/L KCL and 20 Kphos. This allows you to titrate the amount of dextrose you are giving the patient while maintaining everything else constant simply by adjusting the rate of the two fluids. (Note, we generally use NS in these patients, at least initially as the major complication and concern is cerebral edema. Once patients are improving without evidence of cerebral edema (headache, altered mental status, neurological deficits, etc), it is reasonable to change to 1/2 normal saline or to substitute components of chloride with acetate to avoid iatrogenic hyperchloremia that can cause a non-gap metabolic acidosis that may subsequently delay their transfer out of the PICU.
  • The two fluid rates should total ~1.5X maintenance in most PICU patients with DKA. While it is a worthy academic exercise to estimate percent dehydration via clinical parameters, clinical parameters usually relied upon to assess dehydration are difficult to assess in patients with DKA. For example, urine output is unreliable due to glucose induced osmotic diuresis, capillary refill is less reliable as the intravascular space is relatively preserved due to hyperosmolarity, mucus membranes are frequently dry due to Kussmaul breathing, tachycardia is often present simply due to abdominal pain and vomiting, etc. Hence, using 1.5XM in most moderate/severe DKA patients should be sufficient. In fact, one study looked at physician's ability to assess dehydration in DKA patients and found 70% of patients were inaccurately assessed.2 
Figure 1: Doctors are not very good at estimating deyhdration in DKA
  • The fluids should be titrated to maintain a glucose of ~150-250 mg/dl. For example, for a 60 kg child, the maintenance rate would be 100 ml/hr and 1.5X maintenance would be 150 ml/hr. If the blood glucose was 500, you would run the NS + 20 mEq/L KCL and 20 Kphos at 150 ml/hr and the D12.5 NS + 20 mEq/L KCL and 20 Kphos bag at 0 ml/hr. As their glucose decreased, you would gradually increase the contribution of the D12.5 bag, keeping the total rate at 150 ml/hr
  • Lab schedule should generally be: POC blood glucose q1hr, VBG q4hr with Renal panel q4hr (alternating so you are getting a lab back q2hr). Urine ketones are sometimes recommended but may be more appropriate for outpatient management as they monitor acetoacetate and acetone levels (rather than their precursor, betahydroxybutarate, the primary ketone in DKA). Hence, despite adequate treatment, your urine ketone levels may actually look worse initially. 
  • The anion gap is the primary lab you want to follow to assess resolution of ketoacidosis. While the pH and HCO3 level are useful, they also can be misleading as iatrogenic hyperchloremia from use of isotonic saline can cause a mild non-gap metabolic acidosis of little consequence. Nonetheless, this may delay the transition to subcutaneous insulin and out of the PICU so consider substituting chloride with acetate when the patient is improving
  • Once their anion gap has closed (6-10) and they are willing to eat, it is safe to transition them to subcutaneous insulin (though institutional criteria for transitioning may vary) with recommendations from pediatric endocrinology. In transitioning, give the patient their dose of subcutaneous insulin with their meal and leave the insulin infusion on for 30 minutes before shutting off, ensuring no period without insulin. You can discontinue the IV fluids as patients should be able to drink and rehydrate themselves although some elect to continue non-dextrose containing IV fluids. They are now ready to transition the general care ward.


  • Cerebral edema is the most significant and dangerous complication, accounting for 57-87% of deaths in DKA. The estimated incidence of clinical cerebral edema is ~1% although studies examining ventricular size in asymptomatic DKA patients has demonstrated a 50% incidence of ventricular narrowing suggesting a much higher incidence of subclinical cerebral edema. If your patient is complaining of headache, has some degree of somnolence or mental status changes, or has impaired GCS, they likely have some degree of cerebral edema.
  • Glaser et al identified decreased PaCO2 and elevated BUN as independent risk factors for cerebral edema.4

  • Treatment of cerebral edema includes standard ICP maneuvers such as elevation of the head of the bed to 30 degrees, administration of 3% hypertonic saline (3-5 cc/kg bolus with titration to effect) and/or mannitol (0.5-1 g/kg). The maintenance fluid rate can also be decreased to a total of 1X maintenance although there is less evidence to support this practice. There is an ongoing RCT to evaluate fluid therapy and cerebral edema in DKA. There is some controversy over which hyperosmolar therapy to use with one retrospective analysis showing increased associated mortality with use of hypertonic saline.
  • Intubation of the DKA patient with cerebral edema should be approached cautiously, even in the setting of impaired GCS. Frequently, patients will still be hyperventilating with PaCO2 levels as low as 6 mmHg. It is nearly impossible to hyperventilate a patient to that degree and hence, there exists a risk of the PaCO2 rising, intracranial vessels dilating, ICP increasing, and potential herniation. Nonetheless, if the patient is in extremis or beginning to exhibit CO2 retention, careful neuroprotective (premedication with lidocaine, avoiding use of ketamine and succinylcholine) intubation may be required.
  • Thrombosis is common in DKA patients with one study demonstrating a 50% incidence of venous thrombosis with femoral central venous access. Hence, CVL's are generally not placed in DKA patients unless absolutely necessary.
  • Hypoglycemia can occur with insulin treatment and requires careful monitoring as well as vigilance to ensure that the insulin/fluids/dextrose ordered is actually being given to the patient.
  • Hyperosmolar hyperglycemic state (HHS) or hyperosmolar nonketotic coma (HONK) can mimic DKA although it is distinguished by marked hyperglycemia (>600 mg/dL), lack of significant ketosis, and significant hyperosmolarity. It is more common in adults with poorly controlled Type 2 diabetes but has also been described in African American adolescents with DM2. These patients have severe dehydration that can manifest with hypotension and require careful fluid and electrolyte management.


  1. Sellers EA, Dean HJ. Diabetic ketoacidosis: a complication of type 2 diabetes in Canadian aboriginal youth. Diabetes Care 2000; 23:1202.
  2. Koves IH, Neutze J, Donath S, et al. The accuracy of clinical assessment of dehydration during diabetic ketoacidosis in childhood. Diabetes Care 2004; 27:2485.
  3. Laffel L. Improving Outcomes with POCT for HbA1c and Blood Ketone Testing. Journal of Diabetes Science and Technology 2007; 1:1.
  4. Glaser N, Barnett P, McCaslin I, et al. Risk factors for cerebral edema in children with diabetic ketoacidosis. The Pediatric Emergency Medicine Collaborative Research Committee of the American Academy of Pediatrics. N Engl J Med 2001; 344:264.
  5. Decourcey DD, Steil GM, Wypij D, Agus MS. Increasing use of hypertonic saline over mannitol in the treatment of symptomatic cerebral edema in pediatric diabetic ketoacidosis: an 11-year retrospective analysis of mortality. Pediatric Critical Care Medicine 2013;14(7)694-700.
  6. Dr. Kevin Kuo's DKA Resident/Fellow Lecture
  7. 2014 ISPAD Consensus Guidelines