Favorable CGM metrics do not exclude SGLT2 inhibitor-associated euglycemic diabetic ketoacidosis: an educational case report with narrative review of literature.

Sodium-glucose cotransporter 2 inhibitors (SGLT2i) carry a risk of euglycemic diabetic ketoacidosis (euDKA) during acute illness. We report an educational case of a 45-year-old man with type 1 diabetes (T1D) receiving ipragliflozin who developed severe euDKA precipitated by COVID-19. During this sick-day period, he independently discontinued basal insulin-partly because his continuous glucose monitoring (CGM) readings remained within the target range-while continuing ipragliflozin. He presented to the emergency department with worsening gastrointestinal symptoms. On arrival, although blood glucose was 246 mg/dL, venous blood gas revealed severe metabolic acidosis (pH 7.040, HCO3- 9.6 mmol/L) and 3+ urinary ketones. Remarkably, CGM data for the 48 hours prior to ambulance arrival showed a time in range of 71%, indicating clinically acceptable glycemic control. Quantitative analysis confirmed minimal glycemic variability (coefficient of variation 15.8%; mean amplitude of glycemic excursions 77 mg/dL). To our knowledge, this is one of the first reports to quantitatively characterize CGM data, demonstrating that favorable CGM metrics do not preclude a metabolic crisis. Although reimbursement for self-monitoring of blood ketones (SMBK) for patients with T1D receiving SGLT2i has been introduced in Japan, National Database reports reveal limited utilization and structural barriers, such as restrictions on test strip supplies. While promoting SMBK is important, the concept of euDKA remains difficult for patients to grasp. Therefore, clinicians must prioritize symptom-based sick-day rules over "normal" glucose numbers to prevent false reassurance from clinically acceptable CGM metrics. This case underscores the inherent limitations of glucose-centric monitoring and the necessity of multifaceted patient education.