Management of diabetes in patients with COVID-19.

The recent Personal View by Stefan Bornstein and colleagues highlights possible susceptibility factors in patients with diabetes and obesity that could predispose to severe COVID-19-related illness. The authors incorporated useful general management guidance, but we feel that the descriptions of acute presentations were not sufficiently detailed to aid clinical management. We wish to expand on the acute metabolic presentations related to hyperglycaemia, which have been challenging to manage in the acute setting. These considerations are based on observations from a large regional clinical network that informed a national diabetes COVID-19 response team to from the Association of British Clinical Diabetologists help produce new specific guidance to aid acute management.

Our clinical network has reported severe acute glycaemic presentations associated with COVID-19 infections. An excess case incidence of diabetes-related ketoacidosis, often with concurrent significant hyperosmolality (>320 mOsm/L), was reported. In some cases, high levels of ketosis (>5 mmol/L) had rapid onset and appeared to be disproportionate to blood glucose levels. These acute presentations have been noted in patients both with and without pre-existing diabetes and in patients of different ethnicities. Early measurement of capillary blood ketone concentration is therefore advised in all patients with acute hyperglycaemia to allow prompt diagnosis. Of note, use of SGLT2 inhibitors for diabetes, which can promote ketosis in a minority of individuals, does not appear to be a feature in the majority of cases (unpublished data). The article by Bornstein and colleagues correctly hypothesises a possible direct COVID-19-related effect on pancreatic β cell function predisposing to ketosis, which has been reported with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), through increased expression of ACE2 receptor.

Fluid replacement decisions in individuals presenting with diabetes-associated emergencies have been complex. There have been concerns about the rapidity of the onset of ketosis, concurrent potential lung injury, and pronounced kidney injury, with reports of rhabdomyolysis. Thus, prudent use of intravenous fluid replacement judged on an individual basis is advised in managing cases of diabetes-related ketoacidosis and hyperosmolality to avoid adverse lung and renal complications.

Extreme insulin resistance was acknowledged by Bornstein and colleagues. Compatible with this, our network have described significant insulin resistance with unusually high insulin requirements in ventilated patients, which is acknowledged in national guidance from the Association of British Clinical Diabetologists. We postulate that very high insulin requirements might be driven by the COVID-19-related hyperinflammatory state, with contribution from feeding regimens and glucotoxicity.

The observed uncontrolled hyperglycaemia in patients with diabetes and COVID-19 might be explained by higher levels of inflammation compared with patients without diabetes, as reported in a retrospective case series from China. Indeed, our unpublished subset of patients presenting with hyperglycaemia consistently showed elevated concentrations of the acute phase protein ferritin, with a median concentration of 1049 μg/L and a range of 744–2594 μg/L (normal range 30–400 μg/L). Pro-inflammatory cytokines have been demonstrably elevated in people with type 2 diabetes. In particular, IL-6, which coordinates the cytokine release syndrome, has been shown to increase blood glucose in a dose-dependent manner.

We postulate that acute presentations of marked ketosis, mixed diabetes-related ketoacidosis with hyperosmolality, and severe insulin resistance might be reiterated in similar populations of diverse ethnicity within and beyond the UK. Although prevention of SARS-CoV-2 infection is a key strategy, early recognition and specific clinical management to mitigate acute manifestations of diabetes could lead to improved patient outcomes.