Euglycemic Diabetic Ketoacidosis Caused by Alcoholic Pancreatitis and Starvation Ketosis

Abstract

Starvation ketosis and pancreatitis are uncommon and underrecognized etiologies of euglycemic diabetic ketoacidosis (DKA). Euglycemic DKA is associated commonly with pregnancy, use of insulin en route to the hospital, and use of sodium-glucose cotransporter-2 (SGLT-2) inhibitors. A 58-year-old male with past medical history of type II diabetes mellitus and alcoholism presented with chief complaint of nausea, vomiting, and poor oral intake for several weeks. Despite extensive history of diabetes and no recent SGLT-2 inhibitor use, his labs were consistent with euglycemic DKA. His imaging and clinical history also confirmed alcoholic pancreatitis. The patient was admitted for euglycemic DKA secondary to starvation ketosis and alcoholic pancreatitis. His anion gap and beta-hydroxybutyrate rapidly cleared with initiation of the DKA protocol. This case teaches us that clinicians should consider early initiation of the DKA protocol even in the setting of euglycemia, when a patient presents with high-anion-gap metabolic acidosis, a high beta-hydroxybutyrate level, and a clinical picture of pancreatitis and starvation

INTRODUCTION

Euglycemic diabetic ketoacidosis (DKA) has been brought to light due to the advent of sodium-glucose cotransporter-2 (SGLT-2) inhibitors. The mechanism of euglycemic DKA most commonly occurs via decreased gluconeogenesis due to decreased glycogen storage in the setting of fasting or increased excretion of glucose in the urine. In the setting of reduced serum glucose, an increased glucagon-to-insulin ratio also leads to lipolysis, as well as the stimulation of catecholamines and cortisol in the setting of illness, further contributing to insulin resistance and lipolysis. The diagnosis of euglycemic DKA is made when you have serum glucose < 250 mg/dL, high anion gap metabolic acidosis, and elevated serum ketones, as well as history of diabetes.¹ Euglycemic DKA has most commonly been associated with pregnancy, use of insulin en route to the hospital, and use of recent SGLT-2 inhibitors.² However, frequently overlooked causes of euglycemic DKA include pancreatitis and starvation ketosis (not necessarily starvation ketoacidosis) as demonstrated in our patient.

CASE DESCRIPTION

Our patient was a 58-year-old man with known type 2 diabetes mellitus who presented to the emergency department with several hours of dizziness, nausea, vomiting, and dyspnea. He admitted to drinking a pint of whiskey per day. He stated that his last drink was 2 days ago. He was diagnosed with type 2 diabetes mellitus 6 years ago with a hemoglobin A1c of 11.4%; body mass index (BMI) was 29.28 kg/m². He denied a history of retinopathy, nephropathy, or neuropathy. Two years prior to his presentation, at the emergency room, his hemoglobin A1c was noted to be 7.0% and his diabetic regimen at the time included sitagliptin/metformin 10–1000 mg and glimepiride 4 mg daily, although the patient reported that he has not been taking these medications for the last 3 months due to cost concerns. His vital signs on arrival included a blood pressure of 181/105, pulse of 133, temperature of 97.7 °F, respiration rate of 20, O₂ saturation of 98%, and BMI of 22.60 kg/m² which was a significant decrease from his BMI of 29.28 kg/m² roughly 6 years ago. Physical exam was notable for distressed appearance, muscle wasting, tachycardia, and dry mucous membranes.

The patient’s WBC count was 19.8 K/μL (high), and a basic metabolic panel revealed a sodium of 134 mmol/L (normal), potassium of 3.8 mmol/L (normal), blood urea nitrogen of 13 mg/dL (normal), creatinine of 1.29 mg/dL (high, compared to baseline creatinine of 0.8 mg/dL), serum bicarbonate of 6 mmol/L (low), anion gap of 28 (high), and glucose of 214 mg/dL (high). His plasma lactic acid level was 2.3 mmol/L (high), beta-hydroxybutyrate was 12.92 mmol/L (high), and blood alcohol level was negative (ethylene glycol, methanol, and salicylates were not tested). His liver function panel was significant for an ALT of 22 IU/L (normal), AST of 45 IU/L (normal), albumin of 2.2 g/dL (low), total bilirubin of 1.4 mg/dL (high), direct bilirubin of 0.6 mg/dL (high), and alkaline phosphatase of 57 IU/L (normal). His venous blood gas was significant for a pH of 6.99 (low). A CT of the chest, abdomen, and pelvis with IV contrast showed possible multifocal pneumonia. It also showed hepatic steatosis, and mild peripancreatic edema adjacent to the body and tail, which could not exclude mild acute pancreatitis. A lipase obtained was significantly elevated at 968 IU/L. His HbA1c was 6.1% on admission, compared to hemoglobin A1c of 7.0% 2 years ago and hemoglobin A1c of 11.4% 6 years ago. His antibody to glutamic acid decarboxylase was negative, and C-peptide was less than 0.4 ng/mL initially, consistent with depressed pancreatic function (his repeat C-peptide 2 weeks later was 1.8 ng/mL, consistent with his type 2 diabetes and likely recovery of pancreatic function). Labs were otherwise negative for COVID-19 and influenza A.

The patient’s presentation and initial workup suggested the presence of euglycemic DKA secondary to pancreatitis and starvation ketosis. The patient received several boluses of normal saline and was started on an insulin drip at 3 units/h. This was followed by admission to the medical intensive care unit, where DKA protocol was initiated and sodium bicarbonate 50 mEq injection and insulin drip with dextrose 5% and 0.9% NaCl with KCl 20 mEq infusion were given. This was then switched to an infusion of dextrose 5% and 0.45% NaCl with resolution of the acidosis and bicarbonate levels. As serum glucose levels and electrolytes normalized, the patient was transitioned to a regimen of insulin glargine and insulin lispro.

DISCUSSION

Our case is unique because pancreatitis and starvation ketosis–induced euglycemic DKA are uncommon and underrecognized etiologies. It has been demonstrated that euglycemic DKA can occur with SGLT-2 inhibitor use and glucagon-like peptide-1 receptor agonist–induced pancreatitis.³ In addition, previous cases have described euglycemic DKA induced by pancreatitis and starvation ketosis in a patient without a history of diabetes.^4,5 Our patient had euglycemic DKA and did not use confounding medications, such SGLT-2 inhibitors. He also presented with a clear history of uncontrolled type 2 diabetes mellitus, considering his hemoglobin A1c of 11.4% 6 years ago.

Interestingly, our patient’s anion gap acidosis quickly resolved upon initiation of the DKA protocol, which involved starting insulin drip with D5 and normal saline. It is also worth noting that our patient had an initial serum glucose of 212 mg/dL, but, unlike a patient with class euglycemic DKA, he was not on SGLT-2 inhibitors and had not received insulin en route to the hospital. Instead, he presented with nausea and vomiting, an elevated serum lipase, and imaging findings consistent with pancreatitis, which is commonly associated more with hyperglycemic DKA rather than euglycemic DKA. He also presented with a clinical history of poor oral intake and a serum beta-hydroxybutyrate level, which supports starvation-induced euglycemic DKA.

In starvation ketosis-driven euglycemic DKA, patients are not consuming carbohydrates. This forces the body to use its available glucose, which is already limited due to low glycogen storage in the state of fasting. However, an increased glucagon-to-insulin ratio is still present in the state of starvation as well as insulin resistance in the setting of ketosis, which results in increased free ketosis.⁶ Pancreatitis-induced euglycemic DKA also works via a similar mechanism, since pancreatitis patients have decreased food intake due to abdominal pain; however, they also have reduced beta cell function due to destruction, leading to a state of insulin deficiency.⁴ Our case demonstrates that clinicians should have a low threshold for starting a DKA protocol even in a euglycemic patient when they present with high anion gap metabolic acidosis, high serum ketones, and a clinical picture of pancreatitis or starvation.

Differentiating between starvation ketoacidosis from euglycemic DKA is often challenging, as both present with an anion gap metabolic acidosis and elevated serum ketones. However, starvation ketoacidosis will have a serum bicarbonate level that is usually > 18 mmol/L, hypoglycemia, and a prolonged history of fasting, while euglycemic DKA will present with serum bicarbonate < 18 mmol/L, serum glucose < 250 mg/dL but not hypoglycemia, and a history of diabetes.² The reason behind the less prominent acidosis in starvation ketoacidosis stems from the fact that starvation ketoacidosis has a self-limiting effect since the patient has no history of insulin resistance. In a starvation ketoacidosis patient without any history of diabetes, ketone bodies can stimulate insulin release (despite low glucose levels) which then limits further fatty acid oxidation and ketone production, and adipose tissue develops increased sensitivity to insulin’s inhibitory effect on fatty acid release.^7,8 The end effect is a less prominent ketoacidosis compared to a euglycemic DKA patient. Our patient presented with a serum bicarbonate of 9 mmol/L, a serum glucose level of 212 mg/dL, and an A1c of 6.1% (though prior he had an A1c of 7.0%), which supports our diagnosis of euglycemic DKA rather than starvation ketoacidosis.

In general, patients with starvation ketoacidosis are managed with D5 with 0.9% normal saline to simultaneously replace volume while stimulating release of insulin with the administration of D5. Patients with starvation ketoacidosis are prone to hypoglycemia since they can produce exogenous insulin and do not have baseline insulin resistance, in contrast to a patient with type 2 diabetes mellitus. Although it is possible to have concurrent diabetes mellitus with starvation ketoacidosis, treating starvation ketoacidosis patients with exogenous insulin may put them at risk for hypoglycemia.⁹ Thus, differentiating starvation ketoacidosis and euglycemic DKA requires careful analysis of labs combined with a careful history given that patients often have features of both.

In conclusion, this is a unique case of euglycemic DKA induced by pancreatitis and starvation ketosis. This case teaches us that clinicians should consider early initiation of the DKA protocol even in the setting of euglycemia, when a patient presents with high anion gap metabolic acidosis, a high beta-hydroxybutyrate level, and a clinical picture of pancreatitis and starvation.

Declarations

Conflict of Interest

We do not have any conflicts of interest to disclose.