A UK nationwide study of adults admitted to hospital with diabetic ketoacidosis or hyperosmolar hyperglycaemic state and COVID-19.

AIMS: To investigate characteristics of people hospitalised with coronavirus-disease-2019 (COVID-19) and diabetic ketoacidosis (DKA) or hyperosmolar hyperglycaemic state (HHS), and to identify risk factors for mortality and intensive care admission.

MATERIALS AND METHODS: Retrospective cohort study with anonymised data from the Association of British Clinical Diabetologists nationwide audit of hospital admissions with COVID-19 and diabetes, from start of pandemic to November 2021. Primary outcome was inpatient mortality. DKA and HHS were adjudicated against national criteria. Age-adjusted odds ratios (ORs) were calculated using logistic regression.

RESULTS: 85 confirmed DKA cases, and 20 HHS, occurred among 4073 people (211 type 1 diabetes, 3748 type 2 diabetes, 114 unknown type) hospitalised with COVID-19. Mean(SD) age was 60(18.2)y in DKA and 74(11.8)y in HHS (P < 0.001). A higher proportion of patients with HHS than with DKA were of non-White ethnicity (71.4% vs 39.0% P = 0.038). Mortality in DKA was 36.8% (n = 57) and 3.8% (n = 26) in type 2 and type 1 diabetes respectively. Among people with type 2 diabetes and DKA, mortality was lower in insulin users compared to non-users (21.4% vs. 52.2%; age-adjusted OR 0.13 [95%CI 0.03-0.60]). Crude mortality was lower in DKA than HHS (25.9% vs. 65.0%, P = 0.001) and in statin users vs non-users (36.4% vs. 100%; P = 0.035) but these were not statistically significant after age adjustment.

CONCLUSIONS: Hospitalisation with COVID-19 and adjudicated DKA is four times more common than HHS but both associate with substantial mortality. There is a strong association of prior insulin therapy with survival in type 2 diabetes-associated DKA. This article is protected by copyright. All rights reserved.