Severe Euglycemic Diabetic Ketoacidosis After Semaglutide-Associated Poor Intake Despite Empagliflozin Discontinuation: A Case Report

Abstract

Euglycemic diabetic ketoacidosis (EDKA) may be overlooked when serum glucose is near normal. A 68-year-old woman with long-standing type 2 diabetes developed approximately three weeks of nausea, intermittent vomiting and diarrhea, and minimal oral intake after semaglutide initiation at 0.25 mg once weekly for four weeks followed by 0.5 mg once weekly. Patient and family medication reconciliation indicated that she had stopped empagliflozin, metformin, and pioglitazone approximately two weeks before presentation but continued semaglutide, with the last dose approximately eight days before admission. She presented with weakness, dyspnea, slowed speech, and disorientation. Initial testing showed glucose 132 mg/dL, venous pH 7.13, bicarbonate 9 mmol/L, anion gap 26 mmol/L, and beta-hydroxybutyrate 99.5 mg/dL. She improved with intravenous crystalloid, dextrose-containing fluids, insulin infusion, and electrolyte replacement. The temporal association is consistent with a clinically plausible contribution from semaglutide-associated gastrointestinal intolerance and carbohydrate depletion to EDKA despite reported prior empagliflozin discontinuation. Normal glucose should not delay ketone and acid-base testing when compatible symptoms and poor intake are present.

Introduction

Euglycemic diabetic ketoacidosis (EDKA) is characterized by ketoacidosis with normal or mildly elevated glucose, which can delay recognition and treatment. SGLT2 inhibitors are the best-established medication-associated trigger, mediated through glucosuria, lower insulin secretion, higher glucagon-to-insulin ratio, volume depletion, and increased ketogenesis.^1–4 EDKA also occurs during fasting or marked carbohydrate restriction, and severe gastrointestinal symptoms can amplify this physiology.^1,5,6 GLP-1 RAs, including semaglutide, slow gastric emptying and commonly reduce appetite; when complicated by persistent nausea, vomiting, diarrhea, or near-starvation, they may create a metabolic environment favoring ketogenesis.^7–11 We report severe EDKA after semaglutide-associated gastrointestinal intolerance in a patient with type 2 diabetes who had reportedly stopped long-term empagliflozin approximately two weeks before admission. Uncomplicated starvation ketosis generally causes a milder acid-base disturbance; marked ketonemia with substantial anion-gap acidosis in a patient with diabetes favors EDKA, although starvation may act as a precipitant.^1,5,6 The pharmacodynamic effects of SGLT2 inhibitors may outlast their plasma half-lives, and persistent glucosuria and ketogenesis have been described for days after drug discontinuation; therefore, prior exposure may remain contributory even after withholding.^2–4

Case Presentation

Patient Information

A 68-year-old woman with a 17-year history of type 2 diabetes presented to the emergency department with acute shortness of breath, generalized weakness, slowed speech, disorientation, nausea, dry mouth, pruritus, and difficulty ambulating. Family members reported that her speech had been slow for several days and had worsened near presentation. She denied focal neurologic symptoms. Her diabetes had previously been treated with metformin 1000 mg twice daily, pioglitazone 45 mg daily, and long-term empagliflozin 25 mg daily. Primary care records documented semaglutide (Ozempic) initiation approximately seven weeks before presentation at 0.25 mg subcutaneously once weekly for four weeks, followed by 0.5 mg once weekly. According to patient and family medication reconciliation performed for this report, she had not taken empagliflozin, metformin, or pioglitazone for approximately two weeks before admission because of low oral intake and lower home glucose values; she had continued semaglutide, with the last dose approximately eight days before presentation. Her last known hemoglobin A1c was 7.6% about one month before admission. She had never used insulin.

Clinical Findings

For approximately three weeks before admission, she had persistent nausea with intermittent vomiting and diarrhea, heartburn, poor appetite, and minimal food intake, primarily drinking water. On arrival, blood pressure was 143/71 mm Hg, heart rate 99 beats/min, respiratory rate 16 breaths/min, temperature 36.0 deg C, and oxygen saturation 100% on room air. She appeared fatigued but was not in respiratory distress. Cardiopulmonary and abdominal examinations were unremarkable. Neurologic examination showed no focal deficit.

Diagnostic Assessment

Initial testing showed severe high-anion-gap metabolic acidosis with euglycemia: glucose 132 mg/dL, venous pH 7.13, venous bicarbonate 9 mmol/L, serum carbon dioxide 6 mmol/L, anion gap 26 mmol/L, and beta-hydroxybutyrate 99.5 mg/dL (approximately 9.6 mmol/L). Lactate was 1.1 mmol/L, creatinine 0.76 mg/dL, sodium 132 mmol/L, and potassium 3.5 mmol/L. Troponin was negative (Table 1). Chest radiography showed right basal interstitial opacification interpreted as atelectasis or pneumonia, but there was no convincing infectious syndrome and antibiotics were not started. Computed tomography of the head showed no acute intracranial abnormality.

Table 1.

Key Laboratory Findings at Presentation and During Early Treatment

Parameter	Presentation	Early Treatment/Next Value	Interpretation
Glucose	132 mg/dL	87–346 mg/dL during treatment	Euglycemic or mildly elevated range at diagnosis.
Venous pH	7.13	7.15 early in ICU course	Severe acidemia.
Venous bicarbonate	9 mmol/L	8 mmol/L early in ICU course	Severe metabolic acidosis.
Serum CO2	6 mmol/L	7–18 mmol/L during treatment	Low bicarbonate reserve with gradual improvement.
Anion gap	26 mmol/L	20 -> 16 -> 15 -> 12 mmol/L	High-anion-gap acidosis that closed with therapy.
Beta-hydroxybutyrate	99.5 mg/dL (~9.6 mmol/L)	Not repeated in available record	Marked ketonemia.
Lactate	1.1 mmol/L	Not repeated	Lactic acidosis unlikely.
Creatinine	0.76 mg/dL	0.42–0.59 mg/dL	No renal failure to explain acidosis.
Potassium	3.5 mmol/L	2.9–3.4 mmol/L	Required close replacement during insulin therapy.
Phosphorus	Not available initially	<1.0-1.7 mg/dL	Severe hypophosphatemia during treatment.

The working diagnosis was EDKA with severe dehydration and starvation physiology. Metformin-associated lactic acidosis was unlikely because she had not recently taken metformin and lactate was normal. Renal failure and toxicologic causes were not supported by available data. Topiramate can contribute to non-anion-gap metabolic acidosis but did not explain severe ketonemia. Although severe starvation ketoacidosis can occur, uncomplicated starvation ketosis generally produces milder acidemia. The marked beta-hydroxybutyrate concentration and bicarbonate depletion, together with diabetes and improvement with insulin-dextrose therapy, favored EDKA with starvation as a precipitant rather than isolated starvation ketosis.

Therapeutic Intervention

She received crystalloid resuscitation and was admitted to the intensive care unit. Treatment included dextrose-containing intravenous fluids, intravenous regular insulin infusion at 0.05 units/kg/hour, lactated Ringer solution, hourly point-of-care glucose monitoring, and close electrolyte surveillance with replacement of potassium, phosphate, and magnesium. Semaglutide, empagliflozin, metformin, and pioglitazone were discontinued. Topiramate was initially held during the acute metabolic evaluation.

Follow-Up and Outcomes

The anion gap closed with DKA protocol treatment, and her mental status returned to baseline without focal neurologic deficits. She was transitioned from insulin infusion to basal insulin glargine 10 units daily with low-dose correctional insulin during observation, although she did not require correctional insulin before discharge. Diabetes education was provided, including insulin pen teaching, home glucose monitoring, hypoglycemia precautions, and diet counseling. She was discharged on basal insulin only, with semaglutide and empagliflozin discontinued and close primary care follow-up arranged. Longer post-discharge follow-up data were not available at the time of manuscript preparation (Table 2).

Table 2.

Clinical Timeline

Time	Event
Longstanding history	Type 2 diabetes treated with metformin, pioglitazone, and long-term empagliflozin; no prior insulin use.
Approximately 7 weeks before presentation	Semaglutide initiated at 0.25 mg subcutaneously weekly for four weeks, then increased to 0.5 mg weekly.
Approximately 3 weeks before presentation	Persistent nausea, intermittent vomiting and diarrhea, heartburn, dry mouth, fatigue, and markedly reduced food intake; primarily drinking water.
Approximately 2 weeks before presentation	According to patient/family medication reconciliation, empagliflozin, metformin, and pioglitazone were stopped; semaglutide was continued.
8 days before presentation	Last semaglutide dose.
Day 0	Shortness of breath, generalized weakness, slowed speech, and disorientation. Labs showed glucose 132 mg/dL, pH 7.13, bicarbonate 9 mmol/L, anion gap 26 mmol/L, and beta-hydroxybutyrate 99.5 mg/dL.
ICU treatment	Dextrose-containing fluids, insulin infusion, crystalloid resuscitation, and electrolyte replacement.
Discharge	Anion gap closed; mental status returned to baseline; semaglutide and empagliflozin discontinued; basal insulin prescribed.

Discussion

This case illustrates severe EDKA temporally associated with semaglutide-associated gastrointestinal intolerance and near-starvation despite the reported discontinuation of long-term empagliflozin approximately two weeks before admission. The biochemical pattern met accepted EDKA criteria: high-anion-gap metabolic acidosis, marked ketonemia, and glucose well below the traditional hyperglycemic range for DKA.^1,5,6 The normal lactate, preserved renal function, absence of a clear infection source, and improvement with insulin-dextrose therapy supported EDKA rather than lactic acidosis or isolated starvation ketosis.

The causal attribution is clinically important. SGLT2 inhibitors remain the most recognized pharmacologic cause of EDKA, and residual glycosuric or ketotic effects have been reported beyond several half-lives, including cases after medication withholding.^2–4 However, a two-week withholding period, if confirmed, exceeds the usual short perioperative withholding window and shifts attention toward the newly introduced GLP-1 RA and its downstream effect on intake. Semaglutide-associated nausea, vomiting, delayed gastric emptying, and appetite suppression can produce carbohydrate depletion. In susceptible patients with type 2 diabetes, this may lower insulin availability relative to counter-regulatory hormones and accelerate lipolysis and hepatic ketogenesis.^7,9

Published literature increasingly supports this mechanism. Semaglutide-associated euglycemic ketoacidosis has been reported in a patient without diabetes.⁷ EDKA has also been reported after GLP-1 RA initiation in type 2 diabetes, including dulaglutide-associated cases and cases involving GLP-1 RA plus SGLT2 inhibitor exposure or carbohydrate restriction.^8–11 A prior Clinical Case Reports publication described EDKA after SGLT2 inhibitor initiation during an extremely low-carbohydrate diet, underscoring the importance of carbohydrate availability in this syndrome.¹² The present case adds an older patient with long-standing type 2 diabetes, severe GLP-1 RA-associated gastrointestinal intolerance, and a medication history suggesting empagliflozin had been withheld for approximately two weeks.

The circumstances in prior reports were heterogeneous but shared reduced carbohydrate availability or gastrointestinal intolerance. Sood et al described a woman without diabetes who developed ketoacidosis after seven months of semaglutide for weight loss in the setting of a low-carbohydrate diet and several days of vomiting.⁷ Alduraibi et al reported EDKA three days after a first 1.5-mg dulaglutide dose in a patient taking empagliflozin who had markedly restricted carbohydrate intake.⁸ Adrejiya et al described recurrent EDKA after dulaglutide dose escalation in a patient with gastroparesis, vomiting, inability to tolerate oral intake, and SGLT2 inhibitor discontinuation one month earlier.⁹ Lyu described EDKA after semaglutide use and recent enavogliflozin initiation in a patient practicing prolonged intermittent fasting and a low-carbohydrate diet.¹⁰ Gandecka-Pempera et al reported severe EDKA after semaglutide and dapagliflozin were started together, with persistent vomiting, carbohydrate deficiency, and reduced insulin administration.¹¹

The practical lesson is not that semaglutide is a common cause of EDKA, but that GLP-1 RA gastrointestinal intolerance can create the same metabolic danger state as fasting or very-low-carbohydrate intake. Clinicians should ask specifically about recent GLP-1 RA initiation or dose escalation, oral intake, vomiting, diarrhea, insulin omission, and recent or past SGLT2 inhibitor exposure. In patients with diabetes who present with dyspnea, nausea, weakness, altered mentation, or unexplained anion-gap acidosis, normal glucose should not delay blood gas analysis and serum beta-hydroxybutyrate testing.

This report has limitations. Medication timing relied partly on patient and family reconciliation, and the clinical record contained variable descriptions of when oral diabetes agents were stopped. Beta-hydroxybutyrate was not repeated after treatment. A comprehensive toxicologic and endocrine evaluation was not documented, and C-peptide and diabetes autoantibodies were unavailable; therefore, all toxicologic causes and latent autoimmune diabetes could not be definitively excluded, although the available clinical data did not support these alternatives. Follow-up beyond discharge was unavailable. Even with these limitations, the temporal relationship to semaglutide initiation, the severity of reduced oral intake, the absence of alternative high-anion-gap acidosis explanations, and clinical resolution after stopping semaglutide and treating EDKA support a clinically plausible GLP-1 RA-associated contribution.

Conclusion

Recent semaglutide initiation accompanied by persistent gastrointestinal intolerance and near-starvation may contribute to severe EDKA in patients with type 2 diabetes, even when an SGLT2 inhibitor has reportedly been withheld. Clinicians should consider ketonemia and acidemia in symptomatic patients taking GLP-1 receptor agonists when poor intake is present, regardless of glucose level, and should provide sick-day instructions addressing poor intake, vomiting, ketone testing, and when to seek urgent care.

Funding Statement

No public or commercial funding was received.

Patient Perspective

Patient perspective was not available at the time of manuscript preparation.

Abbreviations

EDKA, euglycemic diabetic ketoacidosis; GLP-1 RA, glucagon-like peptide-1 receptor agonist; ICU, intensive care unit; SGLT2, sodium-glucose cotransporter 2.

Data Sharing Statement

Data sharing is not applicable to this article because no datasets were generated or analyzed. Relevant deidentified clinical data are included in this article.

Ethics Approval

Institutional review board (IRB) approval was not required for this case report.

Consent for Publication

Written informed consent was obtained from the patient for publication of this case report and the clinical details contained herein. The patient was shown the contents of the manuscript to be published and consented to their publication.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare no conflicts of interest.