Definition

A complication of insulin deficiency, with accumulation of ketones severe enough to cause ketoacidosis.

Etiology

• Occurs in the setting of insulin deficiency, either absolute (e.g. patients with type 1 diabetes given insufficient insulin) or relative (less common; e.g. patients with type 2 diabetes faced with a stressor like infection or ischemia, or with ketosis-prone type 2 diabetes).
• Precipitants include insulin deficiency, infection or inflammation, ischemia or infarction, intra-abdominal process such as pancreatitis, cholecystitis, ischemic bowel, etc., and iatrogenesis with glucocorticoids.

Evaluation

• Symptoms/signs: polyuria and polydipsia with resulting dehydration, nausea/vomiting, abdominal pain, tachypnea (Kussmaul’s respirations), “fruity” odor of acetone, tachycardia, hypotension, mild hypothermia, and lethargy followed by stupor and coma.
• Labs: the keys to diagnosis, as the name suggests, are ketosis and acidosis. pH is low, serum HCO3 is low, urine ketones are present, and serum ketones (β-hydroxybutyrate) are high. Hyperglycemia does not have to be extreme and is often as low as 250-300 mg/dL. Hyperkalemia can be from acidosis and hypovolemia. Leukocytosis may be present.
• Also evaluate for precipitants and initiate infectious (BMP/LFTs, lipase, lactate, CBC, UA and urine culture, blood culture, chest or abdominal imaging) and ischemic work-up.
• Euglycemic DKA has now been seen in the setting of SGLT-2 inhibitors.
• Relevant formulas:
  • Corrected sodium = measured sodium + 2.4 x (plasma glucose – 100)/100.
  • Anion gap = Na – (Cl + HCO3) (use measured Na level).
  • Calculated osmolality = 2 (Na + K) + glucose/18 + BUN/2.8. Coma possible when >330 mg/dL.

Management

• The primary goal is to close the anion gap, not to reduce the hyperglycemia, per se. Usually, the serum glucose will normalize before the gap is closed.
• Send initial labs, if not done already in the emergency department: CBC, BMP, Ca, Mg, PO4, serum ketones (β-hydroxybutyrate), serum osm, UA, ABG or VBG, HbA1C.
• Send labs to evaluate for potential underlying triggers: LFTs, lipase, blood cx, urine cx if appropriate.
• Keep patient NPO initially, but give large volume IVF (see below).
• Check electrolytes q 2 hours x 3, then q 4 hours until anion gap closes. Monitor I/O’s q 2 hours.
• Can be helpful to set up a DKA flow sheet with: pH, HCO3, electrolytes, anion gap, IVF and insulin dose. A suggested flow sheet can be found here: https://www.aafp.org/afp/1999/0801/p455.html.
• Consider ICU level care if pH <7.1 or challenges with access, hypotension, oliguria, mental obtundation or coma.

Fluid & electrolyte management

• Volume repletion: assume about 10% dehydration. Generally, give 1 liter/hour for 4 hours, then 250-500 ml/hour for the next 2-4 hours, then 100-250 ml/hour. Correct fluid deficit over 36-48 hours. However, be judicious with fluids if patient has a history of heart failure or low ejection fraction.
  • Give NS initially, once the patient is euvolemic or near euvolemic switch to ½ NS if the corrected sodium is greater than 150 meq/L.
  • Change to D5NS or D5½NS when BG <200 mg/dL. This allows the continued use of insulin to clear ketones, but without resulting hypoglycemia. It also decreases the risk of cerebral edema from too rapid a decline in serum glucose.
  • Once BG <200 mg/dL, use D5½NS if the corrected sodium is normal to elevated.
• Potassium replacement: initially, K appears elevated due to acidosis and hypovolemia, but can and will drop quickly.
  • If K <3.3, need aggressive potassium repletion prior to initiating insulin therapy. Give 30-40mEq/hour (requires central line) until K is 3.3-5mEq/L.
  • If K is between 3.3 and 5.3 mEq/L, add IV KCl 20-30mEq to each liter of fluid.
  • If K > 5.3 mEq/L, wait to replete potassium until it has fallen below 5.3mEq/L, check q2 hours.
• Phosphate replacement: generally, replacement not recommended despite the anticipated fall during days 1 and 2. If serum PO4 <1 mg/dL, replete with sodium phosphate (3 mmol PO4/ml; 4 mEq Na/ml). Give 0.3-0.6 mmol PO4/kg/day, but do not give if patient has hypercalcemia or renal failure. Monitor Ca, PO4, and Na.
• Magnesium replacement: replete if serum Mg <1.8 mg/dL or if patient has tetany.
• Bicarbonate: generally not recommended. May consider use only if pH <7.0 or if cardiac instability. Give 50 mEq NaHCO3 in ½NS with KCl 20 mEq/L over 1 hour, never by IV push.

Insulin management

• Insulin bolus, if not already given in the emergency department: 0.1 units/kg IV push.
• Insulin infusion: begin continuous infusion of regular insulin of 0.1U/kg/hour (this is 5-7 units/hour for many patients). See your hospital-specific protocol for insulin infusions.
• Check BG q 1 hour, per insulin infusion protocol.
• Aim to correct BG by 50-70 mg/dL per hour. If not decreasing appropriately, check IV access to ensure it is being delivered correctly. If still not decreasing, increase insulin dose.
  • At many hospitals, mild/average/resistant scales are used, so consider moving to next aggressive scale.
• When BG <200 mg/dL and IVF is changed to a dextrose 5% (D5)-containing fluid (see above), insulin will need to cover the D5.
• Patients with euglycemic DKA will require both insulin and glucose administration.
• Switch from insulin infusion to subcutaneous insulin when blood glucose is <200mg/dL, and two of the following three are met:
  • Serum anion gap < 12 mEq/L.
  • Serum bicarbonate ≥ 15 mEq/L.
  • Venous pH > 7.30.
• Converting insulin infusion to subcutaneous insulin: remember that a patient who presents with DKA (especially type 1) will need exogenous basal insulin, even if NPO. Only using correctional insulin after the resolution of DKA will precipitate the return of DKA.
• To determine the total daily dose (TDD) of insulin, make two calculations.
  • First, calculate the patient’s TDD based on the most recent insulin use (e.g. look at the last 6 hours of insulin infusion and extrapolate this to 24 hours). For safety purposes, take 80% of that dose.
  • Second, estimate TDD using either the patient's usual home dose (if well controlled) or weight (e.g., TDD = 0.5-0.6 units per kg per day).
  • We compare the insulin use to the estimated home insulin needs because in the setting of DKA multiple factors raise insulin requirements and these may rapidly decline over the following several hours (unless there are other complicating factors that influence stress levels like infection). Thus, calculating the TDD based on the first day of DKA treatment may result in an overestimation of the TDD, especially as the patient recovers and stress levels are reducing back to normal.
  • If the TDD calculation and estimation are similar, use the use-based calculation and divide as below. If the use-based calculation is much higher than the home or weight-based estimation, use the lesser value or consult endocrine for assistance.
    • Basal insulin: divide the 80% of the TDD by 2 for the nightly basal dose.
    • Nutritional insulin: if the patient is eating, divide the 80% of the TDD by 2, then split over three meals for the pre-meal dose. If the patient is not receiving nutrition, do not order nutritional insulin.
    • Correctional insulin: order based on BMI.
• Give the first basal insulin SQ injection 1-2 hours BEFORE the infusion is discontinued. If the transition is being made in the morning, consider using a one-time AM NPH injection or ½ of daily glargine dose to bridge until bedtime glargine or NPH begins.
  • Non-gap metabolic acidosis: even when the anion gap has closed, there will be a non-gap metabolic acidosis due to renal excretion of ketotic anions (HCO3 equivalents), with chloride retention. This will normalize as the body restores its buffer stores.

Key Points

• Primary goal is to achieve “metabolic control” by closing the anion gap prior to normalizing blood sugar levels.
• Start potassium replacement even when potassium in normal range.
• Consider infectious or ischemic precipitants.

Kitabchi AE, Umpierrez GE, Miles JM, et al. Hyperglycemic crises in adult patients with diabetes. Diabetes Care 2009; 32(7):1335.

Nyenwe EA, Razavi LN, Kitabchi AE, et al. Acidosis: the prime determinant of depressed sensorium in diabetic ketoacidosis. Diabetes Care. 2010 Aug;33(8):1837-9.

Umpierrez GE, Kitabchi AE. Diabetic ketoacidosis: risk factors and management strategies. Treat Endocrinol. 2003;2(2):95-108.

Rosenstock J, Ferrannini E. Euglycemic Diabetic Ketoacidosis: A Predictable, Detectable, and Preventable Safety Concern With SGLT2 Inhibitors. Diabetes Care. 2015;38(9):1638-1642. doi:10.2337/dc15-1380