Myths in DKA Management

Anand Swaminathan, MD, MPH (@EMSwami) is an assistant professor and assistant program director at the NYU/Bellevue Department of Emergency Medicine in New York City.

Review questions are available at the end of this post.


Each year, roughly 10,000 patients present to the Emergency Department in diabetic ketoacidosis (DKA).  Prior to the advent of insulin, the mortality rate of DKA was 100% although in recent years, that rate has dropped to approximately 2-5%.1  Despite clinical advances, the mortality rate has remained constant over the last 10 years. With aggressive resuscitative measures and appropriate continued management this trend may change.

DKA is defined as:

  • Hyperglycemia (glucose > 250 mg/dl)
  • Acidosis (pH < 7.3)
  • Ketosis

In the absence of insulin, serum glucose rises leading to osmotic diuresis. This diuresis leads to loss of electrolytes including sodium, magnesium, calcium and phosphorous.  The resultant volume depletion leads to impaired glomerular filtration rate (GFR) and acute renal failure.

In patients with DKA, fatty acid breakdown produces 2 different ketone bodies, first acetoacetate, which then further converts to beta-hydroxybutyrate, the latter being the ketone body largely produced in DKA patients.

With this background in mind, let’s take a look at four urban legends in the management of DKA and the evidence that dispels these legends.

Here’s our case:

45 yo woman w/ DM I presents w/ AMS
Vitals: 73/42, 147, 93%, 96oF, FS > 451

Although this presentation likely represents DKA, a blood gas is typically obtained to confirm the diagnosis. Often, the question arises as to whether an arterial or venous blood gas is adequate.

Urban Legend #1 – An ABG is necessary for the diagnosis and treatment of DKA

  1. ABG gets you pH, PaO2, PaCO2, HCO3, Lactate, electrolytes and O2Sat
  2. VBG gets all this except for PaO2 (but we have pulse ox so this isn’t really an issue) – Question is, is it accurate?

There’s a lot of literature on the use of VBG instead of ABG.2-6 Here’s the summary:

  • pH ABG – VBG: 0.03 – 0.05
  • HCO3 ABG – VBG: – 1.5 – 2.0
  • CO2 ABG – VBG: – 6 mm Hg

The Joint British Diabetes Society 2011 Guidelines endorse the use of VBG instead of ABG for initial assessment of acid/base status as well as continued monitoring during management.7

Bottom Line: VBG just as good as ABG in the diagnosis and management of DKA.

Let’s return to our patient. The VBG: 6.87, 15, Bicarb < 5, K = 2.8. IV fluids have been started and the nurse asks you if you can start the insulin drip.

Urban Legend #2 – After fluids, insulin is the next most critical step in treatment

As discussed earlier, the pathophysiology in DKA involves osmotic diuresis and extensive electrolyte loss. Total body potassium can be severely depleted but extracellular potassium levels can be falsely reassuring as acidosis leads to shift of potassium out of cells. Regardless, hypokalemia is a critical cause of morbidity and mortality and should be immediately addressed.

Insulin administration can exacerbate serum hypokalemia by shifting potassium intracellularly. Thus, potassium should always be repleted prior to starting insulin. About 5-10% of patients with DKA will present with hypokalemia (K < 3.3 mEq/L).8 Insulin therapy should be held until a serum potassium > 3.5 mEq/L is documented. If the potassium is < 5.5 mEq/L but > 3.5 mEq/L, it’s safe to start both IV and PO (if patient tolerates) potassium repletion along with your insulin.

For a review of repletion of severe hypokalemia in DKA, check this post from Reuben Strayer at EM Updates.

Other Electrolytes

  1. Sodium – Serum concentration diluted as a result of osmotic gradient of glucose pulling more water into extracellular space.
  2. Phosphate – If < 1.0 mEq/L, start repletion.
  3. Magnesium – All patients who are hypokalemic are hypomagnesemic. Replete together as long as kidney function intact.

The Bottom Line: Electrolyte replacement, particularly potassium is more critical early in DKA management than insulin.

In addition to the severe hypokalemia, our patient is markedly acidemic. Guidelines recommend administration of bicarbonate when the pH falls below 7.1 or 7.0.9

Urban Legend #3 – Once the pH falls below 7.1/7.0, bicarbonate infusion is required

A number of studies have retrospectively examined patients with DKA looking for differences in groups that got bicarbonate versus those that did not.10-12 Patients in these studies, some with pH as low as 6.9, had no benefit from bicarbonate therapy. In order for bicarbonate to increase the serum pH, the lungs have to blow off CO2. Patients with DKA are already maximally ventilating CO2 and are not able to increase this rate. Thus, the administration of bicarbonate does not significantly raise the serum pH. Additionally, there is harm to administration of bicarbonate as well.  Bicarbonate administration delays the improvement of ketosis13 and worsens hypokalemia and intracellular acidosis.14

Bottom Line: There is no established role for administration of sodium bicarbonate to patients with DKA regardless of their pH.

Back to our patient. Bicarbonate therapy is held and a repeat potassium comes back at 3.9 mEq/L. The insulin drip is ready to hang and the patient’s nurse asks you how much of a bolus you would like to give.

Urban Legend #4 – A bolus of insulin should always be given along with the infusion

Once again, let’s return to pathophysiology. Insulin is administered to stop ketosis and eventually, will aid in “closing the gap.” What we would like to do is to give insulin in a way that leads to near normal circulating levels of around 150 – 200 microunits/ml. Giving a bolus and infusion, however, leads to a high peak insulin level with a low plateau. With an infusion rate of around 0.14 units/kg, the patient will reach a steady plateau state that is near the normal circulating level.15

As far as outcomes, Goyal et al found no difference in glucose change or anion gap change in the patients who received a bolus of insulin followed by an infusion.16 Additionally, the bolus insulin group had longer lengths of stay and a 6-fold increase in hypoglycemic episodes (6% vs. 1%).

Bottom Line: Bolus insulin does not result in a more rapid improvement of the patient and is associated with more episodes of hypoglycemia and a longer length of stay.


  • VBG as good as ABG for diagnosis and treatment
  • Aggressively replete potassium prior to starting insulin
  • Bicarbonate is unnecessary in DKA treatment and potentially harmful
  • A bolus of insulin is unnecessary in DKA treatment and potentially harmful

Review Questions

References / Further Reading

  1. Lebovitz HE: Diabetic ketoacidosis.  Lancet 1995; 345: 767-772.
  2. Brandenburg, Mark A, Dire, Daniel J: Comparison of Arterial and Venous Blood Gas Values in the Initial Emergency Department Evaluation of Patients with Diabetic Ketoacidosis.  Annals of Emergency Medicine 1998; 31:4: 459-465.
  3. Gokel, Yuksel; Paydas, Saime; Koseoglu, Zikret; Alparslan, Nazan; Seydaoglu, Gulsah: Comparison of Blood Gas and Acid-Base Measurements in Arterial and Venous Blood Samples in Patients with Uremic Acidosis and Diabetic Ketoacidosis in the Emergency Room.  American Journal of Nephrology 2000; 20:319-323.
  4. Malatesha G, Singh N, Bharija A, Rehani B, Goel A.  Comparison of arterial and venous pH, bicarbonate, PCO2 and PO2 in initial emergency department assessment.  Emergency Medicine Journal 2007. 24: 569-571.
  5. Middleton P, Kelly A-M, Brown J.  Agreement between arterial and central venous values for pH, bicarbonate, base excess, and lactate.  Emergency Medicine Journal 2006 23: 622-624.
  6. Ma JO, Rush MD, Godfrey MM, Gaddis G. Arterial blood gas results rarely influence emergency physician management of patients with suspected diabetic ketoacidosis. AEM 2003; 10: 836-41.
  7. Savage MW, Datary KK, Culvert A, Ryman G, Rees JA, Courtney CH, Hilton L, Dyer PH, Hamersley MS; Joint British Diabetes Societies.  Joint British Diabetes Societies guideline for the management of diabetic ketoacidosis. Diabet Med. 2011 May;28(5):508-15.
  8. Aurora S, Cheng D, Wyler B, Menchine M. Prevalence of hypokalemia in ED patients with diabetic ketoacidosis. Am J Emerg Med 2012; 30: 481-4.
  9. Kitabchi AE, Umpierrez GE, Murphy MB et al.  Hyperglycemic crises in adult patients with diabetes: a consensus statement from the American Diabetes Association.  Diabet Care 2006; 29 (2): 2739-2748.
  10. Morris LR, Murphy MB, Kitabchi AE.  Bicarbonate therapy in severe diabetic ketoacidosis. Ann Intern Med. 1986;105(6):836.
  11. Duhon B, Attridge RL, Franco-Martinez AC, Maxwell PR, Hughes DW. Intravenous sodium bicarbonate therapy in severely acidotic diabetic ketoacidosis. Ann Pharmacother 2013; 47: 970-5.
  12. Green SM, Rothrock SG, Ho JD et al.  Failure of adjunctive bicarbonate to improve outcome in severe pediatric diabetic ketoacidosis.  Ann Emergency Medicine 1998; 31: 41-48.
  13. Okuda Y, Drogue HJ, Field JB et al.  Counterproductive effects of sodium bicarbonate in diabetic ketoacidosis.  J Clinical Endocrinology Metabolism 1996; 81: 314-320.
  14. Villon A, Zuni F, Plafond P et al.  Does bicarbonate therapy improve management of severe diabetic ketoacidosis?  Crit Care Med 1999; 27: 2690-2693.
  15. Kitabchi AE, Murphy MB, Spencer J, Matteri R, Karas J, Is a priming dose of insulin necessary in a low-dose insulin protocol for the treatment of diabetic ketoacidosis?  Diabetes Care. 2008;31(11):2081.
  16. Goyal N, Miller J, Sankey S, Mossallam U. Utility of Initial Bolus insulin in the treatment of diabetic ketoacidosis.  Journal of Emergency Medicine, Vol 20:10, p30.
Edited by Alex Koyfman, MD

21 thoughts on “Myths in DKA Management”

  1. Hi Swami

    Great post.

    One question: dosing of insulin at 0.14 U/hr. Is that actual body weight, ideal body weight in the obese patient?


    1. Great question. The studies didn’t specify but I dose for ideal body weight and then titrate up if needed.

  2. Hi Swami,
    Just a quick question. The paper in Journal of Emergency Medicine examining the utility of an insulin bolus…it doesn’t look like there was a statistically significant difference in rates of hypoglycemia or ED/hospital length of stay between the two groups. Do you still think it’s enough of a valid argument against the bolus? Thanks.

  3. Would you still hold off on an insulin bolus if the patient was hyperkalemic without EKG changes? with EKG changes? as we do for run-of-the-mill hyperkalemia?

  4. Great piece. Two quibbles: phosphate repletion. We used to do this years ago, but last time I looked for literature supporting repletion only found studies showing no changes in important outcomes. Other is you really don’t need to know pH. Doesn’t change what you do, and your chemistry panel gives you bicarb and anion gap. VBG’s only advantage is if you’re somewhere you can get one quickly as opposed to waiting for a chem panel from the lab.

  5. Highly fortuitous discussion for me. As the Medical Director of an ICU, I have a large meeting set up next week to discuss a new DKA management protocol 2 ‘old school’ endocrinologists are trying to impose upon the institution. Pretty much, their protocol is built upon the foundation of every single myth in your article. I plan to use your article & references as a template to form my team’s argumentation.

    Thank you

  6. Highly fortuitous discussion for me. As the Medical Director of an ICU, I have a large meeting set up next week to discuss a new DKA management protocol 2 ‘old school’ endocrinologists are trying to impose upon the institution. Pretty much, their protocol is built upon the foundation of every single myth in your article. I plan to use your article & references as a template to form my team’s argumentation.

    Thank you

Leave a Reply

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