Outcomes of SGLT2-inhibitor-induced euglycaemic diabetic ketoacidosis in a North Queensland intensive care unit: A retrospective cohort study

Evidence on sodium-glucose co-transporter-2 inhibitors (SGLT2i)-associated euglycaemic diabetic ketoacidosis (EuDKA) in the intensive care unit (ICU) and its comparison with type 1 diabetes mellitus (T1DM) diabetic ketoacidosis (DKA) is limited. This study aims to describe ICU characteristics, management and outcomes of SGLT2i-associated EuDKA, identify predictors of poorer outcomes and compare findings with T1DM-DKA. Single-centre retrospective cohort (2017-2023) of adults admitted to a Northern Queensland ICU. EuDKA was defined by plasma glucose <13.9 mmol/L plus ⩾1 of acidaemia, low bicarbonate, raised anion gap or ketonaemia, in patients with T2DM on SGLT2i. A comparator cohort comprised ICU admissions with T1DM-DKA. Demographics, precipitants, biochemistry, insulin therapy, and outcomes were extracted and analysed with non-parametric tests. Fourteen EuDKA and 31 T1DM-DKA admissions were included. EuDKA patients were older with more frequent comorbidities; First Nations Australians were over-represented. Presenting features were heterogeneous, with overlapping precipitants (commonly caloric restriction/fasting or intercurrent illness). Compared with T1DM-DKA, EuDKA had milder biochemical derangement and lower 24-h insulin dose but required longer insulin therapy and longer ICU length of stay; time to acidosis resolution was similar. In exploratory analyses, end-organ complications in type 2 diabetes were associated with longer ICU stay, while obesity and pre-admission insulin therapy were associated with prolonged acidosis duration. ICU-admitted SGLT2i-associated EuDKA differs clinically and in management from classical T1DM-DKA. Findings highlight diagnostic challenges, potential high-risk subgroups, and the need for prospective, multicentre studies to inform risk-based management in critical care.