Objectives To compare clinical management and key outcomes of critically ill children with diabetic ketoacidosis (DKA) in two cohorts (2015 cohort: managed according to the 2015 British Society of Paediatric Endocrinology and Diabetes (BSPED) guidelines; 2020 cohort: managed according to the 2020 BSPED guidelines).

Design Retrospective cohort study using prospectively collected data.

Setting A critical care advice and transport service based in London, and referring hospitals within the critical care network.

Patients All children 0–17 years referred for advice and/or critical care transport with a clinical diagnosis of DKA over a 30-month period (from September 2018 to March 2021).

Interventions None.

Main outcome measures Admission to intensive care unit (ICU), clinically diagnosed cerebral oedema and death.

Results There were significant differences in fluid and insulin administration practices between the 2015 and 2020 cohorts (fluid bolus >20 mL/kg: 3% vs 30%, p<0.001; median total fluid given in the first 24 hours: 84 mL/kg vs 117 mL/kg, p<0.01; starting insulin infusion rate 0.1 U/kg/hour: 54% vs 31%, p<0.01). However, these differences were consistent with guideline recommendations (initial fluid infusion rate within 5% of guideline-recommended rate: 80% in the 2015 group vs 84% in the 2020 group). There were no significant differences in outcomes (ICU admission: 26% vs 35%, p=0.2; cerebral oedema: 21% vs 23%, p=0.8).

Conclusions Our study findings indicate that changes to fluid and insulin administration occurred after the 2020 BSPED guideline publication, with strong adherence to the guideline, but these changes were not associated with changes in key outcomes.