Nuances in Resuscitation Part III: Diabetic Ketoacidosis

By Justin Bright, MD
Senior Staff Physician
Henry Ford Hospital, Dept. of Emergency Medicine
Detroit, MI

Edited by Alex Koyfman, MD

 
Thus far we have discussed resuscitation in trauma and sepsis.  What distinguishes those two from the resuscitation goals in DKA is timing.  In trauma and sepsis, it’s all about early recognition, aggressive and quick optimization, and understanding all the possible treatment options at your disposal.  In the management of DKA, it’s quite the opposite.  If you remember anything from this discussion, it’s that slow and steady wins the race!  In fact, overaggressive resuscitation is what leads to the most significant morbidity and mortality in DKA patients.  Patients in DKA don’t die from the disease process – they die because we kill them!

DKA is defined as an anion-gap metabolic acidosis, with elevated serum ketones (usually measured as beta-hydroxybutyrate), blood glucose > 250 mg/dL, pH < 7.3, and a serum HCO3 < 18 mEq/L.  It is the reason for over 50% of diabetic admissions, and many DKA patients begin their inpatient hospital course in the ICU.  The leading causes of DKA are medication non-compliance, underlying infection, new-onset diabetes (i.e. DKA is the first presenting illness), or underlying medical/surgical stress.  In general, DKA patients will present to the ED relatively early in their disease process because the ketones produced by the body induce vomiting, prompting the patient to seek treatment.  This is in contrast to hyperosmolar non-ketotic coma patients (HONK) that present much later in their illness because there are no ketones in the blood to induce vomiting and alert the patient or his/her family that something is wrong.

The mainstays of DKA management are fluid replenishment, glycemic control, correction of any other metabolic anomalies, and treatment of any underlying cause for the glycemic derangement.  There are definitely some differences in approach compared to the other types of critical illness.  First and foremost, virtually every other hemodynamic situation recommends an initial fluid bolus of 2L of fluid, or in the pediatric population, a fluid bolus of 20 cc/kg.  In DKA, overaggressive fluid hydration is associated with the development of cerebral edema, particularly in the pediatric population (controversial; need for more rigorous research/data).  This is the single biggest cause of morbidity and mortality in DKA patients.  As a result, the initial fluid bolus in the pediatric DKA patient is recommended to be only 10 cc/kg.

It is also important to understand what the composition of your IV fluid is.  .9NS has 154 mmol/L of sodium and 154 mmol/L of chloride.  Chloride is a negative ion, and a large amount of chloride can contribute to worsening acidosis over time.  Therefore a switch over to ½ NS is recommended after your initial fluid bolus, and you should run a maintenance rate that produces urine output at 1-2 cc/kg/hr.  Some providers like using lactated ringers.  The composition of LR is 130 mmol/L of sodium and only 109 mmol/L of chloride, so there is less iatrogenic contribution to the acidosis.  In addition, the LR promotes bicarbonate formation, which further assists in neutralizing the acidosis.  Personally, I avoid LR because bicarbonate supplementation is not recommended due to increased incidence in cerebral edema in pediatric patients.  I may not be choosing to intravenously push amps of bicarbonate, but I also don’t want to use a solution that will enhance bicarbonate formation in a nebulous and hard to track amount.

The primary method of glycemic control is an insulin drip.  There is variability in treatment here too.  The recommended dose for the insulin drip is 0.1 units/kg/hr.  Some providers use a loading bolus of 10 units of insulin, but I recommend against it.  Again, the goal here is not rapid reduction.  Ideal reduction happens at 50-100 points per hour.  Once the patient’s serum blood glucose gets to 250 or below, fluids should be changed to D5 ½ NS until the anion gap has closed and pH has normalized.  At that point, you can discontinue the insulin drip.

Another source of iatrogenic morbidity and mortality is lack of respect for the potassium.  Patients with DKA have a relative hypokalemia.  The acidosis causes the potassium ions to leak from the cells into the serum.  As a result, you need to be vigilant and not just quickly scan through your lab values on your computer.  Your patient’s K level won’t trigger a warning alert on lab reports because it often is read as within normal limits.  But as soon as you initiate insulin therapy, the acidosis is going to correct, driving the potassium back into the cells and the patient will get hypokalemic very quickly if you’re not ready for it.  Remember 10 units of insulin is a mainstay in the treatment of hyperkalemia.  This is another reason why I don’t like using an insulin bolus prior to starting the insulin drip.  The recommendations for potassium replenishment are as follows:

  • If K > 6, don’t supplement
  • If K is 4.5-6, administer 10mEq potassium in your IV fluids
  • If K is 3-4.5, administer 20mEq of potassium
  • If K is < 3, replenish the potassium before beginning insulin therapy

Most institutions have protocols for fluid resuscitation and electrolyte correction in DKA.  The Milwaukee formula is particularly useful, and is worth referencing if you are faced with challenging decisions to make in your management of a DKA patient.

Another important aspect in the management of DKA patients is ventilator settings (click here for further reading: http://www.emdocs.net/airway-subtleties-critically-ill-patients/).  Most DKA patients are tachypneic.  This is not due to respiratory distress, but rather is a respiratory compensation for the metabolic acidosis.  The body is trying to breathe off carbon dioxide to neutralize the pH.  If a patient’s mental status dictates protecting their airway, you absolutely need to make sure you communicate with your respiratory therapist regarding ventilator settings.  Often times RTs are not involved in a patient’s care until after intubation when they are asked to set up a ventilator.  The initial settings are usually protocol-driven, or in some cases, set in a “one size fits all” manner.  If you sedate and chemically paralyze your patient, you are now taking away a natural compensatory mechanism.  If, for example, the patient was breathing at a rate of 26 times per minute in an effort to blow off CO2, and now the ventilator is set at a rate of 16 times per minute, CO2 levels are going to elevate quickly in the patient.  This can cause increased acidosis, hemodynamic instability, neurologic sequelae, and in some instances, cardiac arrest.

To sum it all up, not all resuscitations are created equal.  You cannot expect to follow a protocol or a cookbook recipe.  Our patients and their pathologies are just too diverse and complicated by many variables.  Our medical knowledge and critical decision-making ability is what defines us as emergency medicine physicians.  It is incredibly important to understand what is going on at the physiologic and biochemical level so we can best help our patients.  After all, they came to us to live, not to die.

8 thoughts on “Nuances in Resuscitation Part III: Diabetic Ketoacidosis”

  1. Thoughts on .14U/Kg insulin infusion rate?

    Kitabchi AE etal. Is a priming dose of insulin necessary in a low-dose insulin protocol for the treatment of diabetic ketoacidosis? Diabetes Care. 2008 Nov;31(11):2081-5.

    Basal insulin rates in a normal person are about 150-200 microU/ml. They found if you gave a bolus, you had a high peak and a low plateau. If you gave an infusion of 0.07 units/kg, you had a low plateau with infusion alone. If you doubled the infusion rate to 0.14units/kg, you had a desirable plateau.

    If you subscribe, EMRAP.org had a great podcast regarding this: May 2013, “DKA Myths”

  2. Thoughts on .14U/Kg insulin infusion rate?

    Kitabchi AE etal. Is a priming dose of insulin necessary in a low-dose insulin protocol for the treatment of diabetic ketoacidosis? Diabetes Care. 2008 Nov;31(11):2081-5.

    Basal insulin rates in a normal person are about 150-200 microU/ml. They found if you gave a bolus, you had a high peak and a low plateau. If you gave an infusion of 0.07 units/kg, you had a low plateau with infusion alone. If you doubled the infusion rate to 0.14units/kg, you had a desirable plateau.

    If you subscribe, EMRAP.org had a great podcast regarding this: May 2013, “DKA Myths”

  3. Justin,
    1/2 NS is still going to give a metabolic acidosis. It lacks the strong ion diff ( refer to Weingart’s talk.) We should be using balanced solutions to fluid resuscitate like LR or plasmalyte. Much love from St Vs though

    1. Agreed! The SID (cations – anions) of 0.45% NS is still = 0. I believe that even though the absolute amount of electrolytes are different the acid-base effect (SID/Liter) is the same and will yield a nearly identical non-gap acidosis. As a hypotonic solution, however, it may have slight less acidifying effect as less of the IVF will remain intravascular. Still, if NAGMA is the reason you’re abandoning NS, 0.45% NS confers little benefit.

  4. Justin,
    1/2 NS is still going to give a metabolic acidosis. It lacks the strong ion diff ( refer to Weingart’s talk.) We should be using balanced solutions to fluid resuscitate like LR or plasmalyte. Much love from St Vs though

Leave a Reply

Your email address will not be published. Required fields are marked *