SGLT2 Inhibitor-Associated Diabetic Ketoacidosis in the Intensive Care Unit

Monday, March 23, 2026

10:15 AM - 11:15 AM Central Time

Location: Connections Central - RST 01

Rowan Rosewarne, PharmD, MPH (she/her/hers)

University of New Mexico Hospital, New Mexico

Disclosure(s): No relevant financial relationship(s) to disclose.
SP

Sarah Perez, MD

Critical Care Medicine Fellow
University of Michigan Hospital and Health Systems

Disclosure information not submitted.
MI

Megan Ingebrigston, PharmD, BCCCP

Clinical Pharmacist Specialist – Medical Intensive Care Unit
University of Michigan Health System

Disclosure information not submitted.
IC

Ivan Co, MD

Clinical Associate Professor
University of Michigan

Disclosure information not submitted.
JM

Jakob McSparron, MD

Clinical Associate Professor
University of Michigan

Disclosure information not submitted.
MK

Michael Kenes, PharmD, FCCM, BCPS, BCCCP

Clinical Associate Professor
University of Michigan

Disclosure information not submitted.

Introduction: Sodium-glucose cotransporter 2 inhibitors (SGLT2i) are a newer class of antidiabetic medications to treat type 2 diabetes. They can increase the risk of developing euglycemic diabetic ketoacidosis (DKA), which may be associated with worsened outcomes compared to hyperglycemic DKA. The purpose of this study is to compare time to resolution of hyperglycemic DKA and SGLT2i-associated DKA.

Methods: This was a retrospective cohort study of adult patients admitted to the intensive care unit (ICU) or emergency critical care center (EC3) with DKA and stratified by those with outpatient use of a SGLT2i versus those without. The primary outcome was time to final DKA resolution defined as laboratory resolution of DKA and final discontinuation of an insulin infusion. Additional outcomes included ICU and hospital length of stay, in-hospital mortality, incidence of hypoglycemia, incidence of electrolyte abnormalities, incidence of proper transition off an insulin drip, and compliance with a DKA order set.

Results: Of the 113 patients included in this study, 13 had SGLT2i-associated DKA. There was no difference in median (IQR) time to final DKA resolution between SGLT2i-associated DKA and hyperglycemic DKA (24.8 [19.4-37.2] vs. 22.1 [15.9-33.0] hours; p = 0.270). No difference was found in median (IQR) time to initial DKA resolution between SGLT2i-associated DKA and hyperglycemic DKA (19.5 [13.9-27.9] vs. 16.0 [11.8-24.9] hours; p = 0.158). A Cox regression analysis found no statistical significance between groups for the primary outcome of time to final DKA resolution when adjusting for APACHE II score, with a Hazard Ratio of 0.960 (95% Confidence Interval: 0.371-1.212; p = 0.186).

Conclusions: The development of DKA in patients with SGLT2i use was not associated with worsened outcomes or prolonged time in DKA compared with hyperglycemic DKA. Additional work is needed to assess if certain characteristics in patients with SGLT2i use impact clinical outcomes in DKA.