Decision Making in Emergency Critical Care

SECTION 10 - Endocrine Critical Care

Diabetic Ketoacidosis and Hyperosmolar Hyperglycemic State

Catherine T. Jamin and Jeffrey Manko


Diabetic ketoacidosis (DKA) and the hyperosmolar hyperglycemic state (HHS) are two potentially devastating complications of diabetes. Although the number of patients diagnosed with DKA or HHS has nearly doubled in recent decades, the age-adjusted mortality of these patients has declined by almost half within the same time period.1,2 This improvement in outcomes is due in large part due to the early recognition and therapeutic interventions delivered in the emergency department.

DKA and HHS are characterized by an imbalance between the effective action of insulin and of counterregulatory hormones such as glucagon, cortisol, catecholamines, and growth hormone.3 This imbalance results in increased gluconeogenesis, impaired peripheral glucose utilization, lipolysis, and increased ketoacid production. In DKA, this produces the triad of hyperglycemia, ketonemia, and metabolic acidosis. In HHS, it is thought that there is sufficient effective insulin to limit lipolysis and ketogenesis, but not enough to facilitate glucose uptake in the tissues (Fig. 42.1). In both conditions, patients undergo a significant osmotic diuresis—HHS with total body water (TBW) deficit of 8 to 10 L and DKA with a TBW deficit of 3 to 6 L—resulting in dehydration and electrolyte shifts.

FIGURE 42.1 Pathogenesis of DKA and HHS. Copyright © 2006 American Diabetes Association From Diabetes Care Vol 29, Issue 12, 2006. Information updated from Kitabchi AE, Umpierrez GE, Miles JM, et al. Hyperglycemic crises in adult patients with diabetes.Diabetes Care. 2009;32:1335. From American Diabetes Association. FFA, free fatty acids.


Classically reported findings in a patient with DKA or HHS include polyuria, polydipsia, weakness, and dehydration. The onset of HHS is usually insidious, occurring over days to weeks, while DKA tends to manifest over a period of hours. Patients with DKA may complain of abdominal pain, nausea, or vomiting, while HHS patients often report mental status changes or confusion. The physical exam in both conditions will reveal evidence of hypovolemia, including hypotension, tachycardia, decreased capillary refill, and poor skin turgor. Patients with DKA will commonly demonstrate deep breathing or Kussmaul respirations, a fruity odor to their breath, and abdominal tenderness. Patients with HHS may present with profound neurologic changes including focal deficits, seizures, or coma. The most common insult precipitating both conditions is infection. Other triggers include insufficient insulin, drugs, and other severe physiologic stresses such as myocardial ischemia, stroke, and pancreatitis.3


When DKA or HHS is suspected, the laboratory evaluation should include plasma glucose, basic metabolic panel, serum osmolarity, venous blood gas, serum lactate, and detection of ketones. A complete blood count, urinalysis, blood and urine cultures, chest radiograph, and electrocardiogram may help detect coexisting or triggering illness.

Hyperglycemia is a cardinal feature of both conditions and is typically more profound in patients with HHS (Table 42.1). Patients with DKA may however present with serum glucose <300 mg/dL; therefore, in a patient clinically suspected of having DKA, laboratory evaluation should always include calculation of the anion gap (AG) and serum ketones.3,4

TABLE 42.1 Diagnostic Criteria for DKA and HHS

aNitroprusside reaction method.

bEffective serum osmolality: 2[Measured Na+ (mEq/L)] + Serum Glucose (mg/dL)/18.

cAnion gap: Na+ − [(Cl + HCO3 (mEq/L)]. Na+, sodium; Cl−, chloride; HCO3−, bicarbonate.

Adapted from Adrogué HJ, Lederer ED, Suki WN, Eknoyan G. Determinants of plasma potassium levels in diabetic ketoacidosis. Medicine (Baltimore). 1986;65(3):163.

Copyright © 2006 American Diabetes Association From Diabetes Care Vol 29, Issue 12, 2006. Information updated from Kitabchi AE, Umpierrez GE, Miles JM, et al. Hyperglycemic crises in adult patients with diabetes. Diabetes Care. 2009;32:1335. From American Diabetes Association.


In the patient with DKA, hepatic fatty acid oxidation produces ketone bodies, specifically acetoacetic acid, beta-hydroxybutyric acid, and acetone. The standard laboratory test used to detect serum ketones uses a nitroprusside reagent. While widely available, this test does not detect beta-hydroxybutyric acid and thus may yield a false-negative result. To avoid false-negative results, serum beta-hydroxybutyric acid should be measured directly, when possible.

Anion Gap Metabolic Acidosis

Patients with DKA will have a metabolic acidosis, with an arterial pH, by definition, of <7.3, and an elevated AG.

The AG reflects the difference between measured cations and anions and is elevated in DKA due to the presence of the ketoacids. Normal AG values are 7 to 11, with >12 considered elevated. Patients with hypoalbuminemia will have a factitiously lower AG due to the partial loss of negatively charged albumin particles.5 The AG should be corrected in patients with hypoalbuminemia using the following calculation:

Arterial versus Venous Blood Gas

Recent studies demonstrate that peripheral venous blood gas (VBG) samples can be used to accurately assess the degree of acidosis in patients presenting to the emergency department.68 Compared with an arterial blood gas (ABG), the VBG will be lower by approximately 0.02 to 0.04 pH units. In general, VBGs and ABGs agree, but periodic correlation should be performed if serial VBGs are being used to monitor a patient's acid–base status.


Unlike patients with DKA, patients with HHS will present with significantly elevated serum osmolarity. Hyperosmolarity is primarily due to the marked free water loss associated with glucose-induced osmotic diuresis. Serum osmolarity is calculated as follows:

A serum osmolarity >320 can result in mental status changes, including stupor and coma. In patients with HHS presenting with neurologic impairment but normal serum osmolarity, a rigorous search for alternative explanations of their altered mental status is required.9,10


Despite presenting with elevated serum potassium levels, patients with DKA and HHS will often have a potassium deficit ranging between 3 and 5 mg/kg.11,12 The potassium deficit is multifactorial and can be attributed to decreased intake and increased urinary and gastrointestinal losses.12 Elevated serum potassium is mechanistically related to insulin deficiency, hyperglycemia, and acidosis, which decrease its regular cellular uptake.12 As patients receive treatment for DKA and HHS, potassium uptake resumes and serum levels will rapidly fall, placing patients at risk for cardiac dysrhythmias and respiratory muscle weakness. Potassium levels should be followed closely at every stage of treatment to prevent these treatment complications.11 Protocols for management of DKA (Table 42.2) include components for potassium replacement and to withhold insulin therapy until serum potassium levels are >3.3 mEq/L.3

TABLE 42.2 Management Guidelines

DKA diagnostic criteria: blood glucose 250 mg/dL, arterial pH < 7.3, bicarbonate 15 mEq/L, and moderate ketonuria or ketonemia. HHS diagnostic criteria: serum glucose >600 mg/dL, arterial pH > 7.3, serum bicarbonate > 15 mEq/L, and minimal ketonuria and ketonemia.

a15–20 mL/kg/h.

bSerum Na should be corrected for hyperglycemia (for each 100 mg/dL glucose > 100 mg/dL, add 1.6 mEq to sodium value for corrected serum value).

Bwt, body weight; IV, intravenous; SC, subcutaneous.

Copyright © 2006 American Diabetes Association From Diabetes Care Vol 29, Issue 12, 2006. Information updated from Kitabchi, AE, Umpierrez, GE, Miles, JM, Fisher, JN. Hyperglycemic crises in adult patients with diabetes. Diabetes Care. 2009;32:1335. From American Diabetes Association.


The hyperglycemia present in both DKA and HHS will initially create an osmotic gradient that draws water from the cellular space, effectively lowering the measured serum sodium. This osmotic effect of glucose on serum sodium should be corrected using the following calculation:

The finding of hypernatremia in either DKA or HHS indicates that a significant free water deficit exists.


Serum phosphate may be normal or elevated in patients with DKA or HHS due to extracellular shifts; however, patients are typically phosphate depleted due to urinary loss and decreased intake.13 As with potassium, insulin therapy will unmask this deficit as it drives phosphate back into the cells. Although phosphate replacement has yet to demonstrate clinical benefit in patients with DKA, patients should be administered phosphorus when cardiac dysfunction, anemia, or respiratory depression is present, or when phosphate levels are <1 mg/dL.3,14,15


Other diagnoses to consider when evaluating a patient with an elevated AG acidosis include lactic acidosis, starvation or alcoholic ketoacidosis, uremic acidosis, and toxic ingestion. Patents with DKA may produce lactate, but will have a predominance of ketone bodies and a less significant elevation of lactate when compared to patients with primary lactic acidosis (e.g., the septic patient). Patients with starvation or alcoholic ketoacidosis will have detectable ketones but without hyperglycemia or glycosuria. Patients with uremic acidosis or toxic ingestion may present with an elevated AG acidosis but will not have the accompanying hyperglycemia, ketonemia, or glycosuria.



All patients with DKA or HHS will be volume depleted and require fluid resuscitation. The free water deficit should be replaced within 24 hours and is calculated as follows:

Fluid resuscitation alone has been shown to improve hyperglycemia, as well as decrease peripheral insulin resistance and the availability of counterregulatory hormones.16 Isotonic fluids are the recommended medium to restore intravascular volume and tissue perfusion in DKA and HHS.3 Colloids are more expensive and have not been shown to improve mortality, while hypertonic fluids have been shown to worsen hyperosmolarity, hypernatremia, and hyperchloremia.1719 Normal saline is the initial resuscitative fluid of choice; use of other isotonic fluids such as Plasma-Lyte, lactated Ringer's or Hartmann solution may benefit patients in whom aggressive resuscitation with normal saline has resulted in hyperchloremic metabolic acidosis; however, robust evidence compelling a switch to one of these choices is lacking.2022

Fluid resuscitation in both DKA and HHS should begin at a rate of 15 to 20 mL/kg/h or 1 to 1.5 L given in the first hour, with the goal of correcting free water deficits in the first 24 hours.3 For patients with a normal or elevated corrected sodium >140 mg/dL, 0.45% NaCl is an appropriate initial resuscitative fluid. For patients with a corrected sodium <140 mg/dL, 0.9% NaCl should be used.3 After the first hour, an appropriate infusion rate of saline will range between 250 and 500 mL/h and will be guided by the patient's hemodynamic status, fluid deficit, urinary output, renal and cardiac function, electrolyte status, and osmolarity correction.3 When serum glucose levels decrease to 200 mg/dL in DKA and 300 mg/dL in HHS, 5% dextrose should be added to the replacement fluids to avoid hypoglycemia. The insulin infusion should not be stopped until the acidosis is corrected, unless potassium levels drop below 3 mg/dL (Table 42.2).


Along with intravenous fluids, insulin is the second essential therapy in DKA and HHS. Regular insulin is typically given as a continuous infusion; a loading dose is not necessary if an initial infusion rate is at least 0.14 units/kg/h.23 Alternatively, a priming dose of 0.1 units/kg may be given prior to the initiation of an infusion of 0.1 units/kg/h of regular insulin. There is evidence for the administration of subcutaneous rapid-acting insulin analogs in place of intravenous insulin therapy.2426 For patients with mild to moderate DKA without severe acidosis, shock, or coma, use of a short-acting insulin such as aspart or lispro given every 1 to 2 hours has been shown to be successful in the treatment of DKA.2426 This approach has the advantage of enabling patient management outside of the intensive care unit; however, its use necessitates cautious patient selection, and further research is warranted before it is implemented widely.

A critical aspect in the management of DKA or HHS is the transition from continuous infusion to subcutaneous insulin. In DKA, the hyperglycemia will typically resolve earlier than the metabolic acidosis. Insulin infusion should continue until the resolution of DKA or HHS, with the addition of 5% dextrose to the replacement fluids when the glucose decreases to 200 or 300 mg/dL in DKA and HHS, respectively. According to the American Diabetes Association (ADA), resolution of DKA and HSS is defined when the following goals are achieved3:

  • HHS: glucose 250 to 300 mg/dL

o    Normal osmolality with normal mental status

  • DKA: glucose <200 mg/dL and two of the following:

o    Serum anion gap ≤12 mEq/L

o    Serum bicarbonate ≥15 mEq/L

o    Venous pH >7.30

When these criteria are achieved, the patient should be transitioned to subcutaneous insulin, with overlapping intravenous insulin for 1 to 2 hours. Patients with known diabetes can be given their usual insulin regimen while insulin naïve patients may be started at 0.5 to 0.8 units/kg/d; both regimens must be dosed according to the type of insulin that is used. Types of insulin and their onset, peak effect, duration of action, and dosing time are summarized in Table 42.3.

TABLE 42.3 Types of Insulin


As noted, patients with DKA or HHS may present as normokalemic or hyperkalemic, despite experiencing overall potassium depletion. The true deficit is unmasked during the treatment of DKA and HHS. Because of this, any patient with an initial serum potassium <3.3 mEq/L should receive potassium replacement prior to the initiation of insulin therapy to avoid triggering cardiac arrhythmias and respiratory muscle weakness.3,11,12 The ADA recommends maintaining potassium between the range of 4 and 5 mEq/L and replacing potassium for patients with initial level <5 mEq/L.3


A recent randomized trial failed to show benefit from the administration of bicarbonate to DKA patients with metabolic acidosis and pH levels of 6.9 to 7.14.27 Similarly, a systematic review examining 44 studies found no evidence of improved glycemic control or clinical improvement with the use of bicarbonate therapy in DKA.28 Moreover, a retrospective analysis of bicarbonate use for DKA and HHS revealed evidence of harm, including transient paradoxical worsening of ketosis, increased need for potassium supplementation, and, in pediatric patients, increased risk of cerebral edema and prolonged hospitalization.28 Notably, no prospective randomized trials have studied the use of bicarbonate in DKA patients with pH <6.9. Due to concern about the effects of severe acidosis on vital organ function, the ADA continues to recommend administering 100 mmol of sodium bicarbonate in 400 mL sterile water with 20 mEq KCl at 200 mL/h for 2 hours, or until venous pH is >7.0.3


Common complications in the treatment of DKA and HHS are hypoglycemia and hypokalemia. Patients with these metabolic derangements are best served in the intensive care setting, where clinicians can more easily provide close monitoring of serum electrolytes and glucose. Another common complication is a non-AG hyperchloremic metabolic acidosis, which can follow aggressive resuscitation with normal saline, but is usually self-limited and rarely consequential.29

A serious complication of treatment that occurs more frequently in pediatric patients is cerebral edema. Symptoms including headache, lethargy, and depressed mental status present within 12 to 24 hours of treatment and may rapidly progress to include seizures, incontinence, and brain herniation. The mortality associated with cerebral edema is as high as 20% to 40%.3 The optimal treatment is preventative, with a focus on fluid and sodium replacements in hyperosmolar patients, and the addition of 5% dextrose once glucose levels reach 200 mg/dL (DKA) or 300 mg/dL (HHS).3 A clinical suspicion for cerebral edema should prompt immediate intensive care unit consultation.


All patients with DKA or HHS should be considered for ICU level of care at presentation and aggressively resuscitated. With proper management in the emergency department, patients who approach normalization of AG, have pH > 7.25, and can protect their airway, may be considered for a higher-acuity floor or step-down bed. Early ICU care should be provided in patients who are severely acidemic or hypokalemic, cannot protect their airway, cannot tolerate fluid resuscitation (e.g., renal or cardiac patients), or have another pathophysiologic process present (e.g., sepsis). In the transfer from the ED to the ICU, important communications include the patients' comorbidities and mental status; their free water deficit and status of fluid resuscitation, insulin requirements and trajectory of AG correction; and electrolyte status and the trajectory of correction.



1. (Accessed Feb 20, 2012)

2.Wang J, Williams DE, Narayan KM, et al. Declining death rates from hyperglycemic crisis among adults with diabetes, U.S., 1985–2002. Diabetes Care. 2006;29(9):2018.

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

4.Miles JM, Gerich JE. Glucose and ketone body kinetics in diabetic ketoacidosis. Clin Endocrinol Metab. 1983;12:303–319.

5.Feldman M, Soni N, Dickson B. Influence of hypoalbuminemia or hyperalbuminemia on the serum anion gap. J Lab Clin Med. 2005;146(6):317.

6.Gokel Y, Paydas S, Koseoglu Z, et al. 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. Am J Nephrol. 2000;20(4):319.

7.Brandenburg MA, Dire DJ. Comparison of arterial and venous blood gas values in the initial emergency department evaluation of patients with diabetic ketoacidosis. Ann Emerg Med. 1998;31(4):459.

8.Malatesha G, Singh NK, Bharija A, et al. Comparison of arterial and venous pH, bicarbonate, PCO2 and PO2 in initial emergency department assessment. Emerg Med J. 2007;24(8):569.

9.Umpierrez GE, Kelly JP, Navarrete JE, et al. Hyperglycemic crises in urban blacks. Arch Intern Med. 1997;157:669–675.

10.Kitabchi AE, Fisher JN. Insulin therapy of diabetic ketoacidosis: physiologic versus pharmacologic doses of insulin and their routes of administration. In: Brownlee M, ed. Handbook of Diabetes Mellitus. New York: Garland ATPM Press; 1981:95–149.

11.Abramson E, Arky R. Diabetic acidosis with initial hypokalemia. Therapeutic implications. JAMA. 1966;196(5):401.

12.Adrogué HJ, Lederer ED, Suki WN, et al. Determinants of plasma potassium levels in diabetic ketoacidosis. Medicine (Baltimore). 1986;65(3):163.

13.Kebler R, McDonald FD, Cadnapaphornchai P. Dynamic changes in serum phosphorus levels in diabetic ketoacidosis. Am J Med. 1985;79(5):571.

14.Fisher JN, Kitabchi AE. A randomized study of phosphate therapy in the treatment of diabetic ketoacidosis. J Clin Endocrinol Metab. 1983;57:177–180.

15.Winter RJ, Harris CJ, Phillips LS, et al. Diabetic ketoacidosis: induction of hypocalcemia and hypomagnesemia by phosphate therapy. Am J Med 1979;67:897–900.

16.Waldhausl W, Klenberger G, Korn A, et al. Effects of rehydration on endocrine derangements and blood glucose concentration. Diabetes. 1979;28:577–584.

17.Martin HE, Smith K, Wilson MI. The fluid and electrolyte therapy of severe diabetic acidosis and ketosis; a study of twenty-nine episodes (twenty-six patients). Am J Med. 1958;24:376–389.

18.Perel P, Roberts I. Colloids versus crystalloids for fluid resuscitation in critically ill patients. Cochrane Database Syst Rev. 2007;17:CD000567.

19.Bauer M, Kortgen A, Hartog C, et al. Isotonic and hypertonic crystalloid solutions in the critically ill. Best Pract Res Clin Anaesthesiol. 2009;23:173–181.

20.Chua HR, Balasubramanian V, Stachowski E, et al. Plasma-lyte 148 vs 0.9% saline for fluid resuscitation in diabetic ketoacidosis. J Crit Care. 2012;27:138–145.

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.Dhatariya K. Editorial. Diabetic Ketoacidosis. Brit Med J. 2007;334:1284–1285.

23.Kitabchi AE, Matteri R, Murphy MB, et al. 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.

24.Umpierrez GE, Latif K, Stoever J, et al. Efficacy of subcutaneous insulin lispro versus continuous intravenous regular insulin for the treatment of diabetic ketoacidosis. Am J Med. 2004;117:291–296.

25.Umpierrez GE, Latif KA, Cuervo R, et al. Treatment of diabetic ketoacidosis with subcutaneous aspart. Diabetes Care. 2004;27:1873–1878.

26.Ersoz HO, Ukinc K, Kose 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;60:429–433.

27.Morris LR, Murphy MB, Kitabchi AE. Bicarbonate therapy in severe diabetic ketoacidosis. Ann Intern Med. 1986;105(6):836.

28.Chua HR, Schneider A, Bellomo R. Bicarbonate in diabetic ketoacidosis—a systematic review. Ann Intensive Care. 2011;1:23.

29.Adrogue HJ, Wilson H, Boyd AE III, et al. Plasma acid–base patterns in diabetic ketoacidosis. N Engl J Med. 1982;307:1603–1610.