Euglycaemic diabetic ketoacidosis in a patient with pancreatitis and type 2 diabetes on empagliflozin
- 1 Department of Medicine, University of Florida College of Medicine, Gainesville, Florida, USA
- 2 Internal Medicine, University of Florida, Gainesville, Florida, USA
- Correspondence to Dr Matthew D Bloom; matthew.bloom@medicine.ufl.edu
Abstract
Sodium glucose cotransporter-2 (SGLT2) inhibitors are glucose-lowering drugs with proven efficacy in treating type 2 diabetes mellitus, and more recently, have been shown to improve heart failure outcomes in patients without diabetes. A rare complication of SGLT2 inhibitor use is the development of euglycaemic diabetic ketoacidosis (EDKA), characterised by euglycaemia (blood glucose level <250 mg/dL), metabolic acidosis (arterial pH <7.3 and serum bicarbonate <18 mEq/L), and ketonaemia. Given patients with EDKA do not present with the typical manifestations of diabetic ketoacidosis, including marked hyperglycaemia and dehydration, the diagnosis of EDKA may be missed and initiation of treatment delayed. We present the case of a man with recent SGLT2 inhibitor use and multiple other risk factors who developed EDKA.
Background
Sodium glucose cotransporter-2 (SGLT2) inhibitors are glucose-lowering drugs with proven efficacy in treating type 2 diabetes mellitus (T2DM), and more recently, have been shown to improve heart failure outcomes in patients without diabetes.1 An increasing number of patients are being prescribed these medications due to their cardiovascular and renal benefits demonstrated in large-scale randomised controlled trials.2 SGLT2 inhibitors work by inhibiting SGLT2 in the kidney and increasing excretion of glucose. A rare complication of SGLT2 inhibitor use is the development of euglycaemic diabetic ketoacidosis (EDKA), characterised by ‘euglycaemia’ (blood glucose level <250 mg/dL), metabolic acidosis (arterial pH <7.3 and serum bicarbonate <18 mEq/L), and ketosis.3 Other causes and risk factors for EDKA include pregnancy, alcohol use, diet restriction/starvation, pancreatitis, cocaine intoxication, glycogen storage disorders and liver disease.4–6
The SGLT2 inhibitors, such as canagliflozin, empagliflozin and dapagliflozin, may cause EDKA by multiple proposed mechanisms. Enhanced glycosuria from SGLT2 inhibitor use lowers plasma glucose levels, which decrease insulin secretion from pancreatic beta cells. Similar to classic diabetic ketoacidosis (DKA), insulin deficiency and glucagon elevation promote downstream production of free fatty acids in the liver.7 This typically occurs in the setting of a stressor, such as infection, ischaemia or starvation. It has also been shown that SGLT2 inhibitors reduce clearance of ketone bodies.8
Given patients with EDKA do not present with the typical manifestations of DKA, such as marked hyperglycaemia and dehydration, the diagnosis of EDKA may be missed and initiation of treatment delayed. Untreated EDKA can lead to serious complications, including severe dehydration and life-threatening metabolic consequences.4 This case describes a patient with recent SGLT2 inhibitor use and multiple other risk factors who developed EDKA.
Case presentation
A man with T2DM, treated with empagliflozin (10 mg daily), presented with 3 days of burning epigastric pain radiating to his back. He reported consuming 1–4 beers daily. Initial labs revealed a sodium (Na) level of 127 mEq/L, chloride (Cl) 89 mEq/L, bicarbonate 20 mEq/L, anion gap (AG) 18 mmol/L, glucose 239 mg/dL, lipase 233 U/L, beta-hydroxybutyrate 1.9 mmol/L, venous pH 7.38 and partial pressure of carbon dioxide (pCO2) 46 mm Hg (table 1). To further evaluate the pain and elevated lipase level, imaging was obtained. Abdominal CT showed mild acute interstitial pancreatitis and hepatic steatosis. The patient was admitted with diagnosis of alcohol-induced pancreatitis, made NPO (nil per os), and given intravenous fluids (IVF) (ie, lactated ringer solution), pantoprazole, and as needed pain and nausea medications. Empagliflozin was held as the patient was NPO and subcutaneous sliding scale insulin was started. His blood glucose readings remained in the 80–180 mg/dL range over the next 48 hours. The patient continued to complain of epigastric pain. On hospital day 3, he had one episode of non-bloody, non-bilious emesis. He was haemodynamically stable, but routine labs revealed worsening hyponatraemia and further increase of his AG (Na 123 mEq/L, Cl 89 mEq/L, bicarbonate 11 mEq/L, AG 23 mmol/L, and glucose 246 mg/dL).
Lab investigations
| Day 1 | Day 2 | Day 3 | Day 4 | Day 5 | Normal range | |
|---|---|---|---|---|---|---|
| Serum sodium (mEq/L) | 127 | 131 | 123 | 126 | 130 | 136–145 |
| Serum chloride (mEq/L) | 89 | 91 | 89 | 96 | 98 | 98–107 |
| Serum bicarbonate (mEq/L) | 20 | 20 | 11 | 13 | 17 | 22–30 |
| Anion gap (mmol/L) | 18 | 20 | 23 | 12 | 15 | 8–16 |
| Serum glucose (mg/dL) | 239 | 165 | 246 | 145 | 137 | 65–99 |
| Venous pH | 7.38 | 7.22 | 7.19 | 7.34 | 7.26 | 7.30–7.40 |
| Venous pCO2 (mm HG) | 46.4 | 33.1 | 32.0 | 32.7 | 43.0 | 42.0–52.0 |
| Lipase (U/L) | 233 | -- | 145 | -- | -- | 0–70 |
| Beta-hydroxybutyrate (mmol/L) | 1.9 | 7.30 | 8.9 | -- | 3.5 | 0.02–0.27 |
| Urinary ketones (mg/dL) | -- | -- | 80 | 80 | -- | Negative |
| Lactic acid (mmol/L) | -- | -- | 2.9 | 2.4 | 1.6 | 0.3–1.5 |
Investigations
Additional labs were obtained and demonstrated beta-hydroxybutyrate level of 8.9 mmol/L, lactic acid 2.9 mmol/L, venous pH 7.19 and pCO2 26 mm Hg (table 1). While renal and hepatic function and complete blood count (CBC) remained normal, urinalysis showed 500 mg/dL glucose and 80 mg/dL ketones. Toxicology screen was only positive for opiate; however, the patient received hydromorphone on presentation for pain management.
Differential diagnosis
The main differential for EDKA includes alcoholic ketoacidosis and starvation ketoacidosis. Diagnosis of alcoholic ketoacidosis is highly dependent on the medical history, as lab features are similar to EDKA. Patients with alcoholic ketoacidosis typically present after an alcoholic binge and seek medical care once alcohol can no longer be consumed due to nausea, vomiting or abdominal pain.9 These patients are more likely to have hypoglycaemia and can be successfully treated with crystalloid and dextrose without the need for insulin.3 Diagnosis of starvation ketosis is also reliant on the clinical picture, and in practice there may be a considerable degree of overlap between starvation ketosis and EDKA. In starvation ketosis, prolonged starvation is the primary factor, while in EDKA, starvation may be a result of intercurrent illness. In addition, serum bicarbonate concentration is typically not lower than 18mEq/L in starvation ketosis.4 Other causes of increased AG metabolic acidosis, such as toxic serum alcohols (eg, methanol, ethylene glycol), paraldehyde ingestion, uraemia and lactic acidosis, must also be considered and ruled out.
Treatment
Treatment of EDKA does not differ from that of typical DKA. After the diagnosis of EDKA was made in this patient on day 3 of hospitalisation, he was treated with IVF (5% dextrose and 0.45% sodium chloride containing 20 mEq/L potassium) and insulin infusion (regular insulin 250 units in 250 mL 0.9% sodium chloride).
Outcome and follow-up
The patient had marked improvement of his symptoms with minimal abdominal pain within 24 hours of treatment. The abnormal lab values also improved, with increase in venous pH to 7.34, pCO2 33 mm Hg and bicarbonate 17 mEq/L, and decrease in AG to 12 mmol/L (table 1).
The patient was started on a regular diet on hospital day 5. He was eventually discharged on subcutaneous insulin, with recommendations for alcohol cessation, empagliflozin discontinuation and follow-up with primary care and endocrinology. The patient is currently on metformin and insulin for the treatment of his diabetes, and while there is outpatient documentation that he continues to consume alcohol, there have been no documented admissions for EDKA.
Discussion
DKA is a potentially life-threatening complication of both type 1 (T1DM) and T2DM. Although there is heterogeneity in defining DKA, commonly acceptable criteria include hyperglycaemia (blood glucose >250 mg/dL), metabolic acidosis (arterial pH <7.3, serum bicarbonate <15 mEq/L) and ketosis (moderate ketonaemia or ketonuria).5 10 Hyperglycaemia is usually the hallmark for the diagnosis of DKA.5
EDKA is characterised by milder degrees of hyperglycaemia (blood glucose <250 mg/dL) in the presence of metabolic acidosis and ketosis.3 4 Due to the absence of significant hyperglycaemia, diagnosis and time to treatment may be delayed in EDKA. In patients with unexplained acidosis and a high clinical suspicion for EDKA, there should be a low threshold for obtaining ketone levels to aid in the diagnosis.7
Euglycaemia is found in approximately 2.6%–3.2% of patients admitted to the hospital with DKA.3 Despite the presence of euglycaemia, ketoacidosis is a medical emergency and must be treated promptly with fluid resuscitation and insulin and glucose administration, as well as treatment of the underlying aetiology.7 8
It is possible that our patient had EDKA on admission, as he had an AG of 18 and mildly elevated levels of beta-hydroxybutyrate and glucose. EDKA was not considered due to his normal venous pH on presentation. However, arterial blood gas was not obtained. His worsening lab abnormalities over the next 48 hours could have been due to the delay in appropriate EDKA treatment with insulin drip. It is unclear if empagliflozin, an SGLT2 inhibitor, was the only aetiology of his ketoacidosis. He reported alcohol consumption, presented with pancreatitis, and was made NPO, all of which have also been recognised as risk factors for EDKA. Although empagliflozin has a reported half-life of only 12 hours and was discontinued on admission, previous case reports suggest the pharmacodynamic effects of SGLT2 inhibitors can last longer.11–13 Other recognised risk factors for EDKA while on SGLT2 inhibitors include acute illness, infection, surgical stress, decreased carbohydrate intake and insulin discontinuation or dose reduction.14–16
The most effective means of preventing EDKA in patients taking SGLT2 inhibitors is to ensure proper prescribing and withholding under specific scenarios. It is recommended that SGLT2 inhibitors should be stopped immediately in patients with excessive alcohol intake and held at the onset of acute illnesses, for those at risk of dehydration (eg, colonoscopy preparation) or for those on low-carbohydrate diets.15 SGLT2 inhibitors should also be held for at least 24–48 hours prior to major surgical procedures and restarted only in stable clinical conditions.16
Learning points
-
Euglycaemic diabetic ketoacidosis (EDKA) is characterised by mild degrees of hyperglycaemia (blood glucose <250 mg/dL) in the presence of metabolic acidosis and ketosis.
-
Causes and risk factors for EDKA include sodium glucose cotransporter-2 (SGLT2) inhibitor use, pregnancy, alcohol use, diet restriction/starvation, pancreatitis, cocaine intoxication, glycogen storage disorders and liver disease.
-
Diagnosis of EDKA can be missed due to only mild degrees of hyperglycaemia. In patients with unexplained acidosis and a high clinical suspicion for EDKA, there should be a low threshold for obtaining ketone levels.
-
SGLT2 inhibitors should be discontinued in patients with suspected EDKA. Appropriate treatment with fluid resuscitation and insulin and dextrose administration should be promptly initiated.
-
To prevent EDKA, SGLT2 inhibitors should be stopped immediately in patients with excessive alcohol intake and held at the onset of acute illnesses, for those at risk of dehydration, or for those on low-carbohydrate diets. SGLT2 inhibitors should also be held at least 24–48 hours prior to major surgical procedures.
Ethics statements
Patient consent for publication
Footnotes
-
Contributors Supervised by MS. Patient was under the care of MS. Report was written by OB and MDB.
-
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
-
Case reports provide a valuable learning resource for the scientific community and can indicate areas of interest for future research. They should not be used in isolation to guide treatment choices or public health policy.
-
Competing interests None declared.
-
Provenance and peer review Not commissioned; externally peer reviewed.
- © BMJ Publishing Group Limited 2022. No commercial re-use. See rights and permissions. Published by BMJ.
References
Use of this content is subject to our disclaimer