Abstract: TH-PO0388
SGLT2 Inhibitor-Associated Euglycemic Ketoacidosis in Patients with and Without Diabetes
Session Information
- Fluid, Electrolyte, and Acid-Base Disorders: Clinical - 1
November 06, 2025 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Fluid, Electrolytes, and Acid-Base Disorders
- 1102 Fluid, Electrolyte, and Acid-Base Disorders: Clinical
Authors
- Chandler, Andrew, Lenox Hill Hospital, New York, New York, United States
- McGee, James, Lenox Hill Hospital, New York, New York, United States
- Husk, Gregg, Lenox Hill Hospital, New York, New York, United States
- Rosenstock, Jordan L., Lenox Hill Hospital, New York, New York, United States
Background
Sodium-glucose cotransporter 2 inhibitors (SGLT2i) are associated with euglycemic ketoacidosis (eKA). As indications for SGLT2i use broaden, little is known about the risk and course of eKA in non-diabetic (NDM) patients. We explored the incidence and clinical course of patients with and without diabetes (DM) on SGLT2i admitted with eKA.
Methods
The Northwell database was retrospectively reviewed from January 30, 2024 to January 30, 2025 for all admissions with eKA. EKA was defined by an initial blood glucose < 200 mg/dl along with serum bicarbonate of <18mmol/L (or a blood pH <7.30) and urine ketones >40mg/dl. Charts meeting criteria were reviewed. The Northwell database for outpatient prescriptions for SGLT2i was also reviewed.
Results
There were 249,517 total admissions and 564 patients (0.2%) admitted with eKA, of which 41 (7.8%) were taking an SGLT2i. Among patients with SGLT2i use, 38 (93%) had DM. Of the 3 NDM patients, one had a diagnosis of prediabetes and two had a diagnosis of heart failure. One was admitted for a left ventricular assist device alarm, one for viral gastroenteritis, and one for COVID19. The diagnosis of eKA was unrecognized by hospital providers in all 3 cases.
The most common chief complaint for DM patients was gastrointestinal (GI) related (45%). Twenty DM patients were recognized as having eKA while the diagnosis was unrecognized in 18 cases (47%). Patients with a GI related chief complaint were more likely to have recognized eKA (60% v. 28%).
NDM patients had a higher baseline blood bicarbonate (19 vs 15, p<0.001), lower anion gap (15 vs 21, p<0.001) and lower length of stay (3 vs 7 days, p=0.004). Blood pH (7.27 and 7.26) did not differ significantly, nor did time to resolution of acidosis in both groups (1.3 and 1.7 days).
Over the same period, outpatient prescriptions in Northwell practices for SGLT2i were 15,462 for DM (80%) and 3,776 for NDM (20%).
Conclusion
Admissions with eKA were uncommon and fewer than 10% were on SGLT2i. Among NDM patients on SGLT2i, admission with eKA appears extremely rare. There are fewer outpatient SGLT2i prescriptions for NDM patients which could contribute. The NDM patients had higher blood bicarbonate, lower anion gap and shorter length of stay than DM cases. Future studies with a larger NDM sample size are needed to better compare the clinical course and outcomes between these groups.