Ethnicity is associated with variation in diabetes emergency outcomes: Findings from the DEKODE audit and feedback programme.

Abstract

Background: Although ethnic disparities in diabetes prevalence and chronic complications are well recognised, their association with acute metabolic emergencies and outcomes is not well established. In this study, we investigated the association between ethnicity and clinical outcomes.

Methods: This retrospective multicentre study included adults (≥16 years) admitted with diabetic ketoacidosis (DKA), hyperosmolar hyperglycaemic state (HHS) and hypoglycaemia between 2020 and 2023 in participating centres of the Digital Evaluation of Ketosis and Other Diabetes Emergencies (DEKODE) surveillance programme. Ethnicity was grouped as White, Asian or Black. Multivariable linear and logistic regression models assessed associations between ethnicity and length of stay (LOS), episode duration and in-hospital mortality, adjusting for demographic, biochemical and clinical covariates.

Results: A total of 6859 episodes were analysed (4257 DKA; 191 HHS; 2411 hypoglycaemia). Across these, differences were observed in age at presentation, metabolic parameters, and LOS across ethnic groups. In DKA, Black and Asian ethnicities were associated with longer LOS than White ethnicity (p < 0.001). Black ethnicity was also associated with prolonged DKA duration (p = 0.006). In the HHS cohort, Asian ethnicity was associated with longer LOS (p = 0.014) but lower in-hospital mortality (p = 0.002). In the hypoglycaemia cohort, Black ethnicity was associated with lower mortality (p = 0.037).

Conclusions: Ethnic variation was observed in the presentation and outcomes of diabetic emergencies. After adjustment for demographic and clinical factors, ethnicity remained associated with several outcomes. These findings support further investigation into social, healthcare and pre-hospital factors that may contribute to observed differences.