Closing the Gap: DKA Management Pearls


Diagnosis is not always straightforward, here’s what you need to consider.

Case: A 27-year-old female with a history of insulin-dependent diabetes and recurrent episodes of diabetic ketoacidosis presents with polyurea, polydipsia, blurry vision, and “high” blood glucose readings. You diagnosed her with diabetic ketoacidosis (DKA) due to a pH of 7.18, positive serum ketones, and low bicarbonate. You remember something about using basal insulin early in management. Should you start basal insulin, and what other components of management should you consider?

The diagnosis and management of DKA seem simple… Metabolic acidosis, low bicarbonate, high glucose for the diagnosis, and ketones… Replenish intravascular volume, replace potassium, and start insulin. Unfortunately, this is not always the case.


Diagnosis is easy, right?

Recent literature has questioned our approach to DKA. Many patients will have obvious DKA with known diabetes, large anion gap and ketones, and a history and exam suggesting DKA.  While DKA is associated with hyperglycemia, acidosis, and low bicarbonate, patients can have normal glucose (euglycemic DKA) and normal pH and normal bicarbonate (ketoacidosis with metabolic alkalosis from fluid depletion and vomiting).[1-6]

Interestingly, the Canadian DKA guidelines state that “there are no definitive criteria for diagnosis of DKA.”[7]  The American Diabetes Association (ADA) defines DKA as glucose > 250 mg/dL, pH < 7.3, serum bicarbonate < 18 mmol/L, and anion gap > [10,8] while the UK guidelines use glucose > 200 mg/dL (or known diabetes), positive ketones, pH < 7.3, and bicarbonate < 15 mmol/L, but no anion gap.[9]


An easier way to diagnose DKA is to use a venous blood gas (chemistry, anion gap), and serum ketones, preferably beta-hydroxybutyrate. Utilizing urine ketones can be challenging, as this test mainly assesses the presence of acetoacetate, which may not be necessarily present.[5,6,10] Beta-hydroxybutyrate is the predominant ketone in DKA, which is converted to acetoacetate. Urine ketone assays measure acetoacetate, and thus patients presenting early in the disease may have negative urine ketone levels.[1,5,6]

Interestingly, you can use end tidal CO2 (ETCO2) to assist. An ETCO2 > 35 mm Hg can rule out DKA with 100% sensitivity, while a level < 21 mm Hg is 100% specific for diagnosis.[11,12] Along with diagnosis, determining the etiology is vital. The most common triggers include infection and medication/insulin noncompliance, but other stressors include pregnancy, alcohol/other substance use, surgery, trauma, pancreatitis, infarction/ischemia, and many others.[1,8,13-17]

Case: You have completed a focused physical exam and cannot find a source of infection, the EKG is normal other than tachycardia, and the patient has no other complaints or signs/symptoms other than the polydipsia and dry mouth. The chest x-ray is negative, and the abdominal exam and lipase are normal. Serum potassium is 4.4 mEq/L.

What should you use for your initial resuscitation?


Our patient appears dehydrated on exam. The more IV peripheral access you can obtain, the better. Most patients with DKA can be adequately treated with peripheral access only. Unfortunately, these patients may be prone to difficult access, and an external jugular IV line, internal jugular peripheral IV placement, or central access using the femoral vein may be required.[6,18]

Patients with DKA are often severely intravascularly depleted due to the osmotic diuresis, vomiting, and reduced oral intake, on average 3-6L.[8,19] Start with 10-20 cc/kg based on actual body weight of lactated ringers (LR) or plasmalyte.[13] We recommend a balanced crystalloid, rather than normal saline, which can potentially worsen non-anion gap metabolic acidosis.[20-24]  While there are little data directly comparing LR and NS in DKA resuscitation, one study found a trend towards faster improvement in serum pH with LR, as well as serum bicarbonate increase with LR.21 NS reduced serum bicarbonate.[21] Plasmalyte possesses more data, with improved serum bicarbonate and less hyperchloremia compared with NS.[22,23] Repeat 500 cc to 1L boluses until the patient’s fluid status has improved.[5] Be careful using urine output as a measure of resuscitation, as patients will have increased urine output due to the osmotic diuresis.

What should you consider with potassium?

Before we start insulin, you should ensure serum potassium is normal or high (up to 5% of patients will have low serum potassium in DKA).[24] If hypokalemia is present (< 3.5 mEq/L), hold insulin and replace potassium. You can administer up to 20 mEq/L bilaterally through separate IVs, or through central access. Infusions up to 40 mEq/L can be utilized as long as careful monitoring of the IV site is possible, based on the current literature.[20,25] Most patients suffer from severe nausea with DKA, and repleting with oral potassium may not be possible.[5,6,20]

If serum potassium is 3.5-5.5 mEq/L, start potassium replacement with insulin.[5,6,8] Keep in mind that resuscitation of the patient with DKA will result in rapid potassium decrease, and aim for the higher level of normal for serum potassium (4.5-5 mEq/L).20 If the patient is hyperkalemic with ECG changes such as PR or QRS widening or hemodynamic instability, you should provide an IV fluid bolus and IV calcium gluconate, but insulin and fluid resuscitation are also vital therapies in this setting.

What are the finer points of insulin administration?

Once the potassium is normal or high, insulin is necessary. Just remember…your goal is to correct the ketoacidosis, not necessarily the serum glucose.[5,20] There are two schools of thought: 1) administering 10 units insulin bolus followed by an infusion of 0.1 units/kilogram/hour (U/kg/hr), or 2) administering an insulin infusion at 0.14 U/kg/hr without the bolus.[6,8,20,26,27]

Insulin infusion without bolus may be associated with reduced rates of hypoglycemia and hypokalemia, as well as improved resolution of hyperglycemia and acidosis, but if the infusion will take some time to obtain,[26,27] consider 10 units IV as a bolus. For patients with severe acidosis, you may need to set the infusion at 0.2 U/kg/hr. Titrate the infusion to reduce glucose by 50-70 mg/dL per hour.[8,28] If you are unable to reduce glucose levels by this every hour, increase the insulin by 1-2 units per hour. Also, ensure you are basing your infusion on true body weight, not ideal body weight.[5,6,8,20]  For example, if the patient’s glucose drops by 200 mg/dL in an hour, remember that we are treating the ketoacidosis with the insulin.[5] If the pH and serum bicarbonate are not improving despite this decrease in glucose, leave the insulin infusion rate alone and instead provide glucose.[5,20]

Now that you have resuscitated the patient and have insulin infusing, what should you follow? Your target is to close the anion gap, normalize pH, and normalize bicarbonate safely while avoiding hypoglycemia and hypokalemia.[5,6,8] Measure glucose every hour, and watch urine output. If the patient is not producing urine, consider renal failure and/or shock.

Long-acting insulin was often avoided in the past in the ED. Instead, we typically administer an intermediate form with a two-hour overlap with the infusion.[6,8] However, long-acting insulins like glargine or detemir can assist if provided correctly. Literature suggests long-acting insulin used early in the patient’s course can reduce rebound hyperglycemia (after discontinuation of insulin infusion), assist with moving from insulin infusion, and decrease length of stay.[5,8,28-32]

The UK DKA guidelines incorporate using a patient’s home-dose basal insulin in DKA management.[9] If using long-acting insulin, use the patients home regimen (usually a single daily dose).[20] If the patient does not use insulin, start 0.25 U/kg of glargine per day.[5,20] Do not reduce the patient’s home dose, and if the patient takes glargine twice daily, do not stop the insulin infusion until after the second inhospital dose of glargine.[20] Importantly, in patients with an underlying etiology for their DKA that results in physiologic stress and catecholamine release, an increased insulin requirement may be present.  You can also consider combining the two doses into a single once daily dose.[7,20]  Glargine has a delayed onset of action compared to other forms such as intermediate insulins (NPH). Thus, if you elect to use glargine, you should start it earlier than the normal two hours.[7,20] If the insulin infusion is discontinued within this two-hour window, ketoacidosis can potentially worsen.[5,8]

What if I am treating a patient with relatively mild DKA who appears otherwise well?

A major issue with initiating an insulin IV infusion is that this often requires an ICU for admission. However, this may not be always needed. Patients with mild DKA typically only need some IV fluids and insulin; they may even be appropriate for discharge directly from the ED with follow up. Mild DKA is defined by a serum bicarbonate just below 18 or pH slightly less than 7.[3] and patients may be appropriate for discharge home.[5,8,20,33,34]  One treatment option is subcutaneous (SC) rapid-acting insulin. Literature suggests that SC rapid-acting insulin is safe and effective for mild or moderate DKA.[35-41] Studies also show that SC insulin does not increase risk of hypoglycemia or recurrent DKA. Patients can be given 0.3 U/kg SC bolus, followed by 0.1 U/kg SC every hour or 0.2 U/kg every two hours.[35-41]

Monitoring of serum glucose every hour is recommended. If the patient is on a long-acting insulin, this can be utilized. Ensure the patient has adequate insulin for home. After several hours of monitoring with insulin and fluids, recheck labs, including bicarbonate and anion gap. If the gap has closed, the bicarbonate has normalized, and glucose is < 250 mg/dL, the patient can be discharged home with follow up.[5,6,8,20]

What about the patient with a pH < 6.9 or bicarbonate < 5? Should we administer sodium bicarbonate?

The ADA recommends bicarbonate for DKA with pH < 6.9,8 but data does not support this approach.[20] This degree of acidosis can be frightening, but patients typically do not require sodium bicarbonate infusion.[20] To increase the pH, the patient must get rid of CO2, and in severe states the patient is maximizing respiratory compensation.[5,6,20] One study found that serum bicarbonate administered to patients with serum pH > 6.85 can worsen hypokalemia, delay resolution of ketosis, and increase risk of cerebral edema.[42]

Bicarbonate administration can instead worsen intracellular acidosis, and it can further drop serum potassium levels.[20,42] The best means of addressing low serum bicarbonate is to increase the insulin infusion, or bolus 10 units IV. This may require additional potassium and glucose to match the increased insulin. Administer bicarbonate in the following: cardiac arrest and DKA, hyperkalemia with DKA and life-threatening dysrhythmia, and hemodynamic compromise despite fluid resuscitation.[5,6,20]

My patient is tiring and looks like he will require intubation…

Avoid intubation if you can, especially for altered mental status alone (which should improve with IV fluid resuscitation and insulin).[5,6] If possible, delay intubation to optimize volume status and correct ketoacidosis. These patients are at severe risk of hemodynamic compromise, worsening acidosis, and vomiting with aspiration during intubation.[43]

Patients who are unable to control secretions or those in respiratory arrest should be intubated.[20] Just be sure to mitigate the previously mentioned risks with IV fluid resuscitation, starting a vasopressor if needed, use hemodynamic dosing of the sedative, avoid regurgitation by placing an nasogastric tube prior to intubation if the patient has a full stomach, consider sodium bicarbonate 2-3 ampules over 10-15 minutes, use rocuronium and a larger endotracheal tube (8.0), and consider mechanically controlled apneic ventilation.[20,43]

If the patient has increased work of breathing, or for severe metabolic acidosis, attempt high-flow nasal cannula (HFNC). This supports the work of breathing and improves compensation for metabolic acidosis. Set the FiO2 to over 90%, and increase the flow rate to 60 L/min, which blows off more CO2 by reducing dead space.[20,43]  While we often rely on noninvasive positive pressure ventilation with a full face mask, this can be dangerous in patients with severe nausea/vomiting, resulting in aspiration.[7]

What do I do with IV fluids once the patient has been resuscitated and the ketoacidosis has improved?

The patient will require a maintenance fluid infusion, dependent on the serum glucose. If glucose is > 250-300 mg/dL, start LR at approximately 200 mL/hr.[8,13] However, once the glucose decreases to < 250-300 mg/dL, a glucose infusion is needed.[8,13] You have a couple options. You can use D5 ½ NS at 200 mL/hr, or you can reduce the LR infusion to 100 mL/hr and add a D10W infusion at 100 mL/hr.20 This essentially creates a D5 ½ LR solution.[20] These two fluids are compatible, but a premade fluid is typically not available in the ED. You can administer these fluids through separate IV lines, which allows you to titrate the D10W infusion to the patient’s glucose requirements.

Other items to consider include phosphate and magnesium. Phosphate may drop in patients with severe DKA, and this will need to be replaced if the phosphate is < 1 mg/dL.[8,13] Also pay attention to the serum magnesium. Magnesium on the higher level of normal may assist in prevention of arrhythmias (> 2.2 mg/dL).[20]

When should you stop the insulin infusion?

Ensure the anion gap is < 10-12 mEq/L (except for patients with end stage renal disease and uremia, in which patients will have an elevated anion gap due to uremia) and/or use a serum beta-hydroxybutyrate level < 0.6.[5-8,20]  Also make sure the patient has a serum bicarbonate > 18 mEq/L, the patient has received the long acting insulin at least two hours earlier (preferably sooner), the glucose has improved to < 250 mg/dL, and the patient is PO tolerant (and hopefully hungry, which is an indirect way of assessing for the absence of ketoacidosis).[5,6,8] If the patient has gastroenteritis as the etiology for their DKA or gastroparesis, use an infusion of D5W at 75 mL/hr.20  Encourage the patient to eat, and use sliding-scale insulin (0.08 U/kg rapid acting per meal).[5,8]  Keep watching for the recurrence of DKA by following glucose levels and repeat electrolytes.

The anion gap is not closing… what could be wrong?

There can be a variety of problems. Ensure the patient has been adequately resuscitated with fluids, the insulin dose is appropriate, and that the underlying cause has been addressed. For example, if the patient presented with severe abdominal pain and you thought ketones were the sole cause of the pain, is the patient better after fluids and insulin? [5,20] If not, consider imaging. An underlying etiology that has not been addressed can result in failure to close the anion gap.[5,20]

What about a non-anion gap metabolic acidosis (NAGMA)?

NAGMA often develops later in the management of DKA due to use of NS in resuscitation and/or excretion of ketoacid in the urine.[20] If the patient’s bicarbonate remains low but the anion gap has closed, consider NAGMA.  This is often found with a low serum bicarbonate despite closed anion gap.  Treatment of NAGMA includes isotonic sodium bicarbonate infusion.[20]

I’ve heard about euglycemic DKA. What does this entail?

Euglycemic DKA is DKA with glucose < 250 mg/dL, which occurs in up to 10% of patients with DKA.[2,3,44] Causes include SGLT2 inhibitors, decreased hepatic glucose production (starvation, pregnancy), and partial treatment with insulin before the patient presents to the ED. If the anion gap is elevated or ketones are present, consider euglycemic DKA.[2,5,6,20] Treatment is similar to DKA with fluid and insulin, but IV glucose will need to be started with insulin.[20]

Key Points:

  • DKA diagnosis is not always straightforward: be wary in the patient with history of diabetes who has positive ketones, low serum bicarbonate or low pH, or anion gap.
  • Begin therapy with fluid resuscitation, preferably balanced crystalloids. Ensure serum potassium is normal or high before initiating insulin.
  • Insulin infusion at 0.14 U/kg/hr IV is efficacious with no bolus. Long acting insulin can assist with transitioning to SC insulin therapy when provided early in management. Subcutaneous insulin can be used in those with mild to moderate DKA.
  • Intubating a patient with DKA is fraught with danger. Attempt high flow nasal cannula first.
  • While bicarbonate is not necessary for all patients with DKA, it may help with resolving NAGMA.
  • Euglycemic DKA exists. Ask about SGLT2 inhibitors.


1. Dhatariya KK, Vellanki P. Treatment of Diabetic Ketoacidosis (DKA)/Hyperglycemic Hyperosmolar State (HHS): Novel Advances in the Management of Hyperglycemic Crises (UK Versus USA). Curr Diab Rep. 2017;17:33.
2. Munro JF, Campbell IW, McCuish AC, Duncan JP. Euglycaemic diabetic ketoacidosis. BMJ. 1973;2:578–80.
3. Peters AL, Buschur EO, Buse JB, et al. Euglycemic diabetic ketoacidosis: a potential complication of treatment with sodium-glucose cotransporter 2 inhibition. Diabetes Care. 2015;38:1687–93.
4. Guo R-X, Yang L-Z, Li L-X, Zhao X-P. Diabetic ketoacidosis in pregnancy tends to occur at lower blood glucose levels: case–control study and a case report of euglycemic diabetic ketoacidosis in pregnancy. J Obstet Gynaecol Res. 2008;34:324–30.
5. Cardoso L, Vicente N, Rodrigues D, et al. Controversies in the management of hyperglycaemic emergencies in adults with diabetes. Metabolism. 2017 Mar;68:43-54.
6. Fayfman M, Pasquel F, Umpierrez G. Management of Hyperglycemic Crises: Diabetic Ketoacidosis and Hyperglycemic Hyperosmolar State. Med Clin North Am. 2017;101(3):587-606.
7. Canadian D, Goguen J, Gilbert J. Hyperglycemic emergencies in adults. Can J Diabetes. 2013;37 Suppl 1:S72-6.
8. Kitabchi AE, Umpierrez GE, Miles JM, Fisher JN. Hyperglycemic crises in adult patients with diabetes. Diabetes Care 2009; 32:1335–43.
9. Health and Social Care Information Centre. National Diabetes Inpatient Audit (NaDIA)-2015. 2016. Accessed 10 August 2019.
10. Stojanovic, V. Sherri Ihle. Role of beta-hydroxybutyric acid in diabetic ketoacidosis: A review. Can Vet J. 2011 Apr; 52(4): 426–430.
11. Nagler J Krauss B. Capnography: A valuable tool for airway management. Emerg Med Clin North Am, 2008 Nov;26(4):881-97.
12. Chebl BR, Madden B, Belsky J, et al. Diagnostic value of end tidal capnography in patients with hyperglycemia in the emergency department. BCM Emerg Med. 2016 Jan 29;16:7.
13. Van Ness-Otunnu R1, Hack JB. Hyperglycemic crisis. J Emerg Med. 2013 Nov;45(5):797-805.
14. National Center for Health Statistics. National hospital discharge and ambulatory surgery data [article online]. Available from
15. Ennis ED, Stahl EJVB, Kreisberg RA. The hyperosmolar hyperglycemic syndrome. Diabetes Rev 1994;2:115–126.
16. Kitabchi AE, Umpierrez GE, Murphy MB, Kreisberg RA. Hyperglycemic crises in adult patients with diabetes. Diabetes Care 2006;29:2739–2748.
17. DeFronzo RA, Matzuda M, Barret E. Diabetic ketoacidosis: a combined metabolicnephrologic approach to therapy. Diabetes Rev 1994;2:209–238.
18. Zitek T, Busby E, Hudson H, McCourt JD, Baydoun J, Slattery DE. Ultrasound-guided Placement of Single-lumen Peripheral Intravenous Catheters in the Internal Jugular Vein. West J Emerg Med. 2018;19(5):808–812.
19. Hillman K. Fluid resuscitation in diabetic emergencies: a reappraisal. Intensive Care Med 1987;13:4–8.
20. Farkas J. Anatomy of a DKA resuscitation. The Internet Book of Critical Care. 2016. Available at Accessed 03 August 2019.
21. Van Zyl DG, Rheeder P, Delport E. Fluid management in diabetic-acidosis—Ringer’s lactate versus normal saline: a randomized controlled trial. Q J Med 2012; 105:337–343.
22. Chua HR, Venkatesh B, Stachowski E, et al. Plasma-Lyte 148 vs 0.9% saline for fluid resuscitation in diabetic ketoacidosis. J Crit Care. 2012 Apr;27(2):138-45.
23. Mahler SA, Conrad SA, Wang H, Arnold TC. Resuscitation with balanced electrolyte solution prevents hyperchloremic metabolic acidosis in patients with diabetic ketoacidosis. Am J Emerg Med. 2011 Jul;29(6):670-4.
24. Arora S, Cheng D, Wyler B, Menchine M. Prevalence of hypokalemia in ED patients with diabetic ketoacidosis. Am J Emerg Med. 2012;30:481–484.
25. Tran T, Pease A, Wood A, et al. Review of Evidence for Adult Diabetic Ketoacidosis Management Protocols. Front Endocrinol (Lausanne). 2017;8:106.
26. Goyal N, Miller JB, Sankey SS, Mossallam U. Utility of initial bolus insulin in the treatment of diabetic ketoacidosis. J Emerg Med. 2010 May;38(4):422-7.
27. 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):2081–2085.
28. Nyenwe EA, Kitabchi AE. Evidence-based management of hyperglycemic emergencies in diabetes mellitus. Diabetes Res Clin Pract 2011;94:340–51.
29. Hsia E, Seggelke S, Gibbs J, et al. Subcutaneous administration of glargine to diabetic patients receiving insulin infusion prevents rebound hyperglycemia. J Clin Endocrinol Metab. 2012;97(9):3132-3137.
30. Doshi P, Potter A, De L, Banuelos R, et al. Prospective randomized trial of insulin glargine in acute management of diabetic ketoacidosis in the emergency department: a pilot study. Acad Emerg Med. 2015;22(6):657-662.
31. Houshyar J, Bahrami A, Aliasgarzadeh A. Effectiveness of Insulin Glargine on Recovery of Patients with Diabetic Ketoacidosis: A Randomized Controlled Trial. J Clin Diagn Res. 2015;9(5):OC01-5.
32. Rappaport S, Endicott J, Gilbert M, et al. A Retrospective Study of Early vs Delayed Home Dose Basal Insulin in the Acute Management of Diabetic Ketoacidosis. J Endocr Soc. 2019;3(5):1079-1086.
33. Kitabchi AE, Ayyagari V, Guerra SM. The efficacy of low-dose versus conventional therapy of insulin for treatment of diabetic ketoacidosis. Ann Intern Med. 1976;84:633–8.
34. Fisher JN, Shahshahani MN, Kitabchi AE. Diabetic ketoacidosis: low-dose insulin therapy by various routes. N Eng J Med. 1977;297:238–41.
35. Umpierrez GE, Latif K, Stoever J, et al. Efficacy of subcutaneous insulin lispro versus continuous intravenous regular insulin for the treatment of patients with diabetic ketoacidosis. Am J Med. 2004;117:291–6.
36. Umpierrez GE, Cuervo R, Karabell A, et al. Treatment of diabetic ketoacidosis with subcutaneous insulin aspart. Diabetes Care. 2004 Aug;27(8):1873-8.
37. Della Manna T, Steinmetz L, Campos PR, et al. Subcutaneous use of a fast-acting insulin analog: an alternative treatment for pediatric patients with diabetic ketoacidosis. Diabetes Care. 2005 Aug;28(8):1856-61.
38. Ersöz HO, Ukinc K, Köse M, et al. Subcutaneous lispro and intravenous regular insulin treatments are equally effective and safe for the treatment of mild and moderate diabetic ketoacidosis in adult patients. Int J Clin Pract. 2006 Apr;60(4):429-33.
39. Karoli R, Fatima J, Salman T, Sandhu S, Shankar R. Managing diabetic ketoacidosis in non-intensive care unit setting: Role of insulin analogs. Indian J Pharmacol. 2011 Jul;43(4):398-401.
40. Haas RM, Hoffman AR. Treatment of diabetic ketoacidosis: should mode of insulin administration dictate use of intensive care facilities? Am J Med. 2004 Sep 1;117(5):357-8.
41. Teevan C. Evaluation of a diabetic ketoacidosis treatment protocol using subcutaneous insulin aspart. Crit Care Med. 2012 Dec; 40(12):1-328.
42. Chua HR, Schneider A, Bellomo R. Bicarbonate in diabetic ketoacidosis – a systematic review. Ann Intensive Care 2011;1:23.
43. Salim Rezaie. Critical Care Updates: Resuscitation Sequence Intubation – pH Kills (Part 3 of 3). REBEL EM blog, October 3, 2016. Available at: Accessed 10 August 2019.
44. Rawla P, Vellipuram AR, Bandaru SS, Pradeep Raj J. Euglycemic diabetic ketoacidosis: a diagnostic and therapeutic dilemma. Endocrinol Diabetes Metab Case Rep. 2017;2017:17-0081.


Brit Long, MD is an EM Attending Physician at San Antonio Uniformed Services Health Education Consortium.

Alex Koyfman, MD is a Clinical Assistant Professor of Emergency Medicine at UT Southwestern Medical Center and an Attending Physician at Parkland Memorial Hospital. He is also Editor-in-Chief for emDocs.


  1. do you have the rest of the references for “Closing the Gap..
    ” I looked online but it stopped at 16 e.g. 21?

Leave A Reply