Endocrinology

Euglycemic diabetic ketoacidosis: Etiologies, evaluation, and management

a b s t r a c t

Introduction: Diabetic ketoacidosis is an Endocrine emergency. A subset of Diabetic patients may present with rel- ative euglycemia with acidosis, known as euglycemic diabetic ketoacidosis (EDKA), which is often misdiagnosed due to a serum glucose <250 mg/dL. Objective: This narrative review evaluates the pathogenesis, diagnosis, and management of EDKA for emergency clinicians.

Discussion: EDKA is comprised of Serum glucose <250 mg/dL with an Anion gap metabolic acidosis and ketosis. It most commonly occurs in patients with a history of low glucose states such as starvation, Chronic liver disease, pregnancy, infection, and alcohol use. Sodium-glucose cotransporter-2 (SGLT2) inhibitors, which result in in- creased urinary glucose excretion, are also associated with EDKA. The underlying pathophysiology involves insu- lin deficiency or resistance with glucagon release, poor glucose availability, ketone body production, and urinary glucose excretion. Patients typically present with nausea, vomiting, malaise, or fatigue. The physician must deter- mine and treat the underlying etiology of EDKA. Laboratory assessment includes venous blood gas for serum pH, bicarbonate, and ketones. Management includes resuscitation with intravenous fluids, insulin, and glucose, with treatment of the underlying etiology.

Conclusions: Clinician knowledge of this condition can improve the evaluation and management of patients with EDKA.

Published by Elsevier Inc.

Introduction
1. Background

Diabetic ketoacidosis is an endocrine emergency which in- volves hyperglycemia, anion gap metabolic acidosis, and ketosis [1-3]. While the majority of cases occur in patients with type 1 diabetes, up to 23% of cases occur in patients with type 2 diabetes [1,4]. There is a subset of patients with DKA who have normal serum glucose, termed euglycemic diabetic ketoacidosis (EDKA), which is defined by relative euglycemia (serum glucose <250 mg/dL) with metabolic acidosis (serum bicarbonate <18 mEq/L and pH < 7.3) and ketosis [5-9]. This condition was first described in 1973 by Munro et al., followed by the

* Corresponding author.

E-mail address: [email protected] (B. Long).

¹ Present Address: 3841 Roger Brooke Dr., Fort Sam Houston, TX 78234, United States of America.

publication of a larger case series in 1993 [8,9]. The recognition and in- cidence of EDKA have increased in recent years, specifically with the use of sodium-glucose cotransporter-2 (SGLT2) inhibitors in insulin- deficient patients with chronic type 2 diabetes, type 1 diabetes, or latent autoimmune diabetes [10-14]. This class of medications which includes canagliflozin, dapagliflozin, and empagliflozin demonstrates similar or improved efficacy in reducing body weight, serum glucose, hemoglobin A1c levels, and blood pressure [10,15,16]. These medications have also demonstrated a reduction in cardiovascular and all-cause mortality [15,16]. However, this class of medications increases the risk of EDKA by a factor of 7 [17,18]. Of DKA admissions, 2.6-7% of patients are euglycemic, and DKA in insulin-dependent diabetics may be due to SGLT2 inhibitors in 5-12% of cases [5,6,17-24]. Currently, the United States Food and Drug Administration and the European Medicine Agency have issued warnings concerning this risk of EDKA with SGLT2 inhibitor use [25,26].

There are currently no reviews of the available literature on EDKA for emergency clinicians. This narrative review provides an evidence-based

https://doi.org/10.1016/j.ajem.2021.02.015 0735-6757/Published by Elsevier Inc.

summary of EDKA for emergency clinicians, including pathophysiology, etiologies, evaluation, and management.

Methods

The authors searched PubMed and Google Scholar for articles using a combination of the keywords “euglycemic diabetic ketoacidosis” or “euglycemic DKA”. The search was conducted from the database’s in- ception to December 1, 2020. PubMed yielded 91 articles. Authors eval- uated case reports and series, retrospective and prospective studies, randomized controlled trials, systematic reviews and meta-analyses, and other narrative reviews. Authors also reviewed guidelines and supporting citations of included articles. The literature search was re- stricted to studies published in English, with a focus on the emergency medicine and Critical care literature. Authors decided which studies to include for the review by consensus. When available, systematic re- views and meta-analyses were preferentially selected. These were followed sequentially by randomized controlled trials, prospective stud- ies, retrospective studies, case reports, and other narrative reviews when alternate data were not available. A total of 67 resources were se- lected for inclusion in this narrative review by author consensus.

Discussion
1. Pathophysiology and etiologies

The underlying pathophysiology of EDKA includes an absolute insu- lin deficiency or relative insulin deficiency with severe insulin resis- tance [1-3,5]. This causes increased glucagon production and release of free fatty acids, leading to ketogenesis and production of ketone bod- ies and acidosis [1-3,19]. The underlying mechanism of EDKA also in- cludes reduced glucose availability and production during a fasting state (typically associated with some stressor) and/or increased urinary glucose excretion associated with excess counter-regulatory hormones [7,12,27-35]. Thus, any underlying condition resulting in decreased glu- cose availability or production, reduced insulin secretion, and increased counter-regulatory hormone production can result in EDKA. Such con- ditions include starvation state, pregnancy, chronic alcohol use, liver disease, infection/sepsis, and SGLT2 inhibitor use (Table 1) [5,7,27-37]. Bariatric surgery patients with Insulin-dependent diabetes can experi- ence ketoacidosis in 20% of postoperative cases [38]. EDKA can occur rarely in the patient with insulin dependent diabetes who recognizes DKA, gives him/herself a dose of insulin, and presents to the emergency department (ED). Due to the dose of insulin prior to arrival, the serum glucose may be normal.

SGLT2 inhibitors are associated with EDKA due to several mecha- nisms, including the previously discussed reduced insulin secretion

Table 1

Conditions associated with EDKA

Anorexia/fasting state (pre-operative)
Gastroparesis
Glycogen storage disease
Infection/sepsis
Insulin pump use
Intoxication/Ingestion (alcohol, cocaine)
IntraAbdominal pathology (gastroenteritis, pancreatitis, etc.)
Ketogenic diet
Liver disease
Pregnancy
Renal disease
Self-treatment with insulin for DKA prior to presentation
SGLT2 inhibitor use
Surgery

Abbreviations: DKA - diabetic ketoacidosis; ED - emergency depart- ment; SGLT2 - sodium-glucose cotransporter-2.

and increased glucagon release, contributing to free fatty acid produc- tion [19,27-31]. This typically occurs in the setting of some stressor such as infection, ischemia, or starvation, but an underlying etiology is not discovered in all cases of EDKA [19,22,27-31]. SGLT2 inhibitors act on the sodium-glucose cotransporter-2 protein, which is located in the proximal renal tubules and responsible for absorption of up to 90% of fil- tered glucose [5,19,22]. This increases urinary excretion and blocks the reabsorption of glucose, resulting in glucosuria, reduced serum avail- ability of carbohydrates, and Volume depletion [5,19,22,27,30,31]. Moreover, SGLT2 inhibitors reduce clearance of ketone bodies [19,27- 29,31]. The ketoacidosis results in an anion gap metabolic acidosis [27-29]. Because of the urinary glucose losses and the reduction in available serum glucose, serum glucose levels are typically lower than expected compared to other cases of DKA.

SGLT2 inhibitors include canagliflozin, empagliflozin, and dapagliflozin. Canagliflozin demonstrates the highest risk with hazard ratio (HR) of 3.58, compared to empagliflozin (HR of 2.52) and dapagliflozin (HR of 1.86) [39]. Patients on SGLT2 inhibitors with de- creased glycogen stores and lower body mass index have greater risk of EDKA [10]. The typical time to onset of EDKA after initiation of SGLT2 inhibitors is less than 2 months [10,25,40,41].

1. Evaluation

Due to the associated ketoacidosis, patients with EDKA present sim- ilarly to those with DKA [1,3,5]. However, due to a relatively lower serum glucose level in patients with EDKA compared to DKA, nearly 50% of patients experience a Delay in diagnosis [5,40]. Symptoms such as nausea, vomiting, malaise, and fatigue, as well as Kussmaul respira- tions are common, though the most frequent presenting symptom is vomiting [3,5,8]. EDKA should be considered in any patient on an SGLT2 inhibitor or with risk factors for EDKA, such as alcohol intoxica- tion, pregnancy, chronic liver disease, decreased oral intake, or symp- toms such as nausea, vomiting, abdominal pain, confusion, fatigue, malaise, or dyspnea [2,5,6,10,19,42,43]. Diabetic patients often delay presenting to the ED, because their glucose level on home testing will not be significantly elevated [5]. Ketosis typically results in a fruity odor to the patient’s breath [1-3,5]. Patients are commonly dehydrated and demonstrate tachycardia, hypotension, delayed capillary refill, dry mucous membranes, and poor skin turgor on examination [1-3,5].

Laboratory assessment in the ED should include electrolytes, glu- cose, renal function, Liver enzymes, venous blood gas, and serum ke- tones if available. Patients will present with pH less than 7.30, bicarbonate less than 18 mEq/L, and elevated anion gap as well as ke- tone bodies (Table 2) [5-8]. The American College of Endocrinology rec- ommends using serum pH and beta-hydroxybutyrate as diagnostic tools in the assessment of EDKA [10]. An arterial blood gas is not neces- sary, as venous pH is comparable to arterial pH [3]. Ketone assessment evaluates for acetoacetate through nitroprusside testing or beta- hydroxybutyrate. Assessing serum beta-hydroxybutyrate, if available, is preferable rather than relying on serum acetoacetate or urine ketones, as the beta-hydroxybutyrate to acetoacetate ratio in DKA approximates 7-10:1 [3,5,10,20]. Serum beta-hydroxybutyrate levels >3 mmol/L in pediatric patients and > 3.8 mmol/L in adults are reliable for diagnosing DKA [3,44,45]. If serum beta-hydroxybutyrate is not available, serum acetoacetate and/or urine ketones can be utilized, though these have

Table 2

Diagnostic criteria for euglycemic DKA.

Relative euglycemia (< 250 mg/dL)
Acidosis (pH < 7.30, bicarbonate <18 mEq/L)
Ketosis (preferably serum beta-hydroxybutyrate >3 mmol/L if available; serum acetoacetate or urine ketones can be utilized)

Abbreviations: mg/dL - milligrams per deciliter; mEq/L - milliequivalents per liter; mmol/ L - millimoles per liter.

lower specificity and sensitivity for EDKA. Urine ketones may be reabsorbed rather than excreted in the urine in the setting of SGTL2 inhib- itor use, leading to a false negative result [3,10]. Potassium levels typically vary, similar to hyperglycemic DKA [1-3,5]. If the etiology is due to starva- tion or malnutrition, hypomagnesemia and hypophosphatemia may be present, so these levels should also be assessed. There are many precipi- tants of EDKA, and the clinician should also investigate the etiology of the patient’s presentation [1,3].

The differential diagnosis of ketoacidosis with a normal glucose must be considered. The correct diagnosis of the underlying condition driving the ketoacidosis is important so that proper management can be initiated. The other common causes include starvation ketoacidosis and alcoholic ketoacidosis [5]. It is important to inquire about a history of a ketogenic diet, intentional or unintentional starvation, and alcohol use. In general, the acidosis caused by a starvation ketoacidosis is not se- vere with a serum bicarbonate that is usually >18 mEq/L [1,3,46]. Alco- holic ketoacidosis can present with a severe metabolic acidosis, nausea, and vomiting with a low or normal glucose (mean of 118 mg/dL in one small study) and a history of chronic alcohol use with poor nutritional intake [1,47,48]. Therefore, it can be difficult to differentiate EDKA and AKA in the patient with chronic alcohol use.

1. Management

Initial management includes crystalloid fluid resuscitation with 1-2 L during the first 1-2h (Table 3) [1,3,5,49,50]. Balanced crystalloids (e.g., PlasmaLyte, Ringer’s Lactate) are recommended, as normal saline can contribute to hyperchloremic Non-anion gap metabolic acidosis [3,51-53]. Balanced fluids can reduce the time to DKA resolution by sev- eral hours when compared to normal saline [54]. For patients with serum potassium greater than 3.5 mEq/L, clinicians can initiate insulin at a rate of 0.05-0.1 units/kg/h for management of ketosis. Additionally, administration of IV dextrose is essential in patients with EDKA to avoid hypoglycemia [3,5,8,19,55]. Dextrose 5% should be administered con- currently with IV insulin. Dextrose 10% is recommended if hypoglyce- mia occurs despite infusion of dextrose 5%. Insulin is utilized to resolve ketoacidosis, no matter the serum glucose level, and insulin in- fusion is recommended even if patients do not use insulin for home serum glucose control [5,10,19]. If ketosis does not resolve with Insulin infusion, the infusion should be increased [1,3,58-60]. Potassium sup- plementation is recommended at 10 mEq/L IV for those with serum po- tassium 3.5-5.5 mEq/L [3,56,57]. If serum potassium is >5.5 mEq/L, potassium supplementation can be held [3]. If serum potassium is

<3.5 mEq/L, potassium should be repleted to >3.5 mEq/L before the

insulin infusion is initiated [3]. Serum electrolytes and glucose should be monitored every hour. SGLT2 inhibitors should be discontinued, but they can be restarted after resolution of EDKA [5,10]. Sodium bi- carbonate infusion is typically unnecessary, even in the setting of pH < 7.0 [1,61-67]. Sodium bicarbonate may assist in the patient with significant hemodynamic instability associated with acidemia or hyperkalemia [1,3].

The majority of patients recover with appropriate recognition and therapy, with resolution defined by normal serum bicarbonate,

Table 3

Emergency medicine considerations.

Clinicians should consider EDKA in patients with nausea, vomiting, malaise, or fatigue in the setting of alcohol use, chronic liver disease, starvation, pregnancy, infection, or on SGLT2 inhibitor.
A glucose <250 mg/dL should not be used to exclude diabetic ketoacidosis.
If EDKA is suspected, serum pH, bicarbonate, and ketones should be obtained.
Treatment includes intravenous fluids, insulin, and glucose, as well as man- agement of the underlying etiology (e.g., antibiotics for infection).

Abbreviations: EDKA - euglycemic diabetic ketoacidosis; mg/dL - milligrams per deciliter; SGLT2 - sodium-glucose cotransporter-2.

pH > 7.3, ability to tolerate food, and anion gap <12 mmol/L [3,5]. How- ever, Delays in diagnosis and therapy can result in further volume deple- tion, electrolyte abnormalities, severe acidosis, thrombosis, infection, seizures, cerebral edema, Hemodynamic compromise, and prolonged hospital length of stays [1,3]. Due to the need for insulin infusion, hourly electrolyte and glucose monitoring, and the potential for complications, intensive care unit admission is recommended [1,3].

1. Special considerations

Emergency physicians who care for athletes should recommend dis- continuation of SGLT-2 inhibitors at least 24 h prior to strenuous athletic events to reduce the risk of EDKA [10]. Similarly, the medication should be discontinued for patients referred for emergency surgery or kept NPO (nothing by mouth) for a procedure [10].

Conclusions

EDKA is defined by relative euglycemia (serum glucose <250 mg/ dL) with an anion gap metabolic acidosis (pH < 7.3 and serum bicarbon- ate <18 mEq/L) and ketosis. It is commonly associated with low glucose states such as starvation, chronic liver disease, pregnancy, infection, and alcohol use. SGLT2 inhibitors, a relatively newer class of diabetic medi- cations, are also associated with EDKA. This condition should be consid- ered in patients with these risk factors who present with nausea, vomiting, malaise, or fatigue. Testing includes serum venous pH, bicar- bonate, and ketones, as well as evaluation of the underlying etiology. Management includes resuscitation with intravenous fluids, insulin, glucose, and treatment of the underlying etiology.

Declaration of Competing Interest

No conflicts for any author.

Acknowledgements

MG, BL, SL, and AK conceived the idea for this manuscript and con- tributed substantially to the writing and editing of the review. This man- uscript did not utilize any grants or funding, and it has not been presented in abstract form. This clinical review has not been published, it is not under consideration for publication elsewhere, its publication is approved by all authors and tacitly or explicitly by the responsible au- thorities where the work was carried out, and that, if accepted, it will not be published elsewhere in the same form, in English or in any other language, including electronically without the written consent of the copyright-holder. This review does not reflect the views or opinions of the U.S. government, Department of Defense, U.S. Army, U.S. Air Force, or SAUSHEC EM Residency Program.

References