Sodium-glucose co-transporter 2 inhibitors (SGLT2i) were first approved in Australia in 2014. Its indications have expanded beyond diabetes treatment to include treatment of heart failure and kidney disease. However, euglycaemic diabetic ketoacidosis (eDKA) can be precipitated by reduced oral intake, systemic illness, and infection, all of which are common in the oncological population. The incidence of eDKA in elective and emergency procedures in surgical inpatients were reported at 0.17% and 1.1% respectively. However, the rate remains unknown in the oncological setting.
This retrospective study examines the prevalence of eDKA in cancer patients who received SGLT2i at Peter MacCallum Cancer Centre, a specialist cancer centre. Pharmacy records of inpatient SGLT2i prescriptions were extracted between September 2016 and February 2023. Variables documented included SGLT2i cessation pre-elective procedure, admission category (emergency or elective), and risk factors for eDKA. eDKA was defined as glucose < 16 mmol/L, ketone > 0.6, bicarbonate ≤ 18 mmol/L or pH ≤ 7.3.
274 admissions from 180 patients (age 67 ± 9 years, 71% male, 37% on Dapagliflozin, 63% on Empaglifozin) were recorded. 58% were elective admissions; 67% had documented reduced oral intake. There were 15/274 (5.5%) episodes of eDKA. Of the 15 unique patients with eDKA (ketone = 4 ± 3.5), mortality rate at follow-up was 47% versus 34% (p = 0.048) in those without eDKA. The risk factor for eDKA was reduced oral intake (7.7% vs 1.1%, p = 0.025). Admission category, withholding SGLT2i, chemotherapy, sex, glucocorticoid use or type of SGLT2i did not alter the incidence of eDKA.
The prevalence of eDKA in oncology patients on SGLT2i was higher than reported in other populations and was associated with higher mortality, reflecting the catabolic state of this cohort. The overall mortality (26% at study cessation) might diminish the long-term cardiovascular and metabolic benefits of this drug class.