New onset diabetes is associated with higher in-hospital complications and all-cause mortality as compared to those with euglycaemia or pre-existing diabetes.
New onset hyperglycaemia, even in the absence of diabetes, is associated with worse outcomes as compared to euglycaemic individuals.
Potential pathogenic mechanisms between COVID-19 and diabetes includes the effect of inflammatory pathways on glucose homeostasis.
In this commentary, we reflect on our experience in managing patients with coronavirus disease 2019 (COVID-19) at the Charlotte Maxeke Johannesburg Academic Hospital, from the first wave that started in March 2020 to date. Of the documented 690 COVID-19 patients admitted to our hospital, from 6 March to 31 August 2020, 258 (37%) had diabetes mellitus (DM), and 39 (5.7%) had hyperglycaemia (HG). Patients with DM and even those with HG without DM had a worse prognosis than those with euglycaemia. (1)
The exponential increase in COVID-19 is reaching a cataclysmic scale, and on September 3, 2021, the WHO reported a global total of 218 946 836 cases of COVID-19, including 4 539 723 deaths, with the highest global mortality rate in critically ill patients seen in Africa. (2,3) The African COVID-19 Critical Care Outcomes Study (ACCCOS) showed a mortality rate of 48.2% as compared to the global average of 31.5%.(3) In addition to other previously reported predictors of mortality, including older age, chronic liver disease, chronic kidney disease, human immunodeficiency virus and limited resources, DM was found to be independently associated with higher mortality (odds ratio 1.25).(3) Since the earliest weeks of the COVID-19 pandemic, underlying DM and cardiovascular disease have been considered risk factors for severe COVID-19. An unanswered question is whether the virus causing COVID-19, severe acute respiratory syndrome coronavirus 2 (SARS-Co-V2) per se, causes new onset DM.
New onset DM can be defined in a patient with symptoms of HG as a fasting plasma glucose (FPG) ≥ 7 mmol/L or a random blood glucose (RBG) ≥ 11 mmol/L and / or glycated haemoglobin (HbA1c) ≥ 6.5%.(4) However, variations in this definition occur within the context of COVID -19 and HG. A recent meta-analysis that included over 3700 patients showed a 14.4% prevalence of newly diagnosed DM in a pooled population of hospitalised COVID-19 patients. In this meta-analysis, newly diagnosed DM was defined as new onset diabetes in those with no prior history of DM and a FPG ≥ 7 mmol/L or RBG ≥ 11 mmol/L and HbA1c ≤ 6.5%; and if the HbA1c was ≥ 6.5% patients were classified as having previously undiagnosed DM.(5)
What is clear from our experience, and which is supported by local and international data, is that SARS-Co-V2 induces HG even in those without DM. The concept of “new onset” HG can be further classified into four categories, i) “stress-induced HG” defined as a blood glucose (BG ) > 7.8 mmol/l, ii) “new onset DM” in those without diabetes or previously unrecognised pre-diabetes, iii) HG related to the direct effect of SARS-Co-V2 on the pancreas, and iv) drug-induced HG, especially with the use of corticosteroids in the treatment of COVID-19.(4) Of interest is that this entity of “new onset” HG has not been clearly associated with traditional risk factors for DM or corticosteroid administration.(4,6)
A poorer prognosis has been associated with “new onset” HG with or without DM. Numerous studies have reported worse outcomes in those with new onset DM compared with normoglycaemic individuals, with significantly higher in-hospital complications and all-cause mortality.(7) A recent study by Fandini et al. showed a significant increase in intensive care unit (ICU) admissions and death ( RR 3.06; 95% CI, 2.04-4.57) in those with new onset DM.(7,8) Furthermore, a stronger association was found in those with new onset DM for ICU admission or death (RR 3.06; 95% CI, 2.04-4.57) as compared to those with pre-existing DM (RR 1.55; 95% CI 1.06-2.27).(8) Data from our study showed that in patients classified with HG in the absence of DM, despite having a lower mean admission FPG and mean BG than those with DM, the mortality rate was significantly higher than in those with DM.(1) Furthermore, those with HG had worse outcomes, including a 3 fold higher mortality and need for ventilation, with a 3.5 fold greater chance for ICU admission compared to those with euglycaemia.(1) This is supported by international data. For example, the study by Bode et al. showed that those with HG (with or without DM ) had a higher mortality than those with euglycaemia (28.8% vs. 6.2%) respectively, whereas a sub-analysis showed that individuals with new onset HG without DM had a higher mortality than patients with pre-existing DM (41.7% vs. 14.8%).(9) This is possibly because patients with DM can adapt to high ranges of glucose and may be more tolerant to moderately high BG than those without DM but with HG.(10,11) Despite the evidence that new onset DM is associated with a worse prognosis, an important point to emphasise is that HG regardless of DM is the driving factor for adverse outcomes with COVID-19.
An important question that remains is whether COVID-19 induced DM has a different pathogenesis and whether SARS-Co-V2 is precipitating a “new form” of DM. The clinical phenotype of DM with COVID-19 has been largely attributed to that typical of type 2 DM. However, case reports and data from a multicentre paediatric study in the UK reported new onset type 1 DM and diabetic ketoacidosis associated with COVID-19. (12,13) More research is required to identify the pathophysiological phenotype of COVID-19 associated DM, as the underlying mechanisms remain largely elusive, but potential pathways causing HG and DM are described in (Fig. 1). SARS-CoV-2 gains host cell entry predominantly via angiotensin converting enzyme 2 (ACE2) and its co-factors including transmembrane serine protease 2 (TMPRSS2), transmembrane serine protease 4 (TMPRSS4), neuropilin 1(NRP-1), CD209L and scavenger receptor class B type 1(SR-B1).(14–17) However, data is conflicting with regards to angiotensin converting enzyme 2 (ACE2) expression on pancreatic beta (β) cells, but clear ACE2 expression is shown on endothelial cells and pericytes of the microvasculature, as well as interlobular ducts of the islet and acinar regions.(14–16) Thus, the role of SARS-CoV-2 directly causing diabetes is more complex than β cell destruction, but cannot be excluded. SARS-CoV-2 triggers a cascade of pro-inflammatory cytokines including interleukin (IL)-1 β, IL-6, IL-10, TNFα and acute phase reactants, as well as an increase in counterregulatory hormones.(17–19) These pro-inflammatory pathways could impair insulin signaling and reduce insulin stimulated glycogen synthesis as well as impair glucose uptake in the liver and skeletal muscle and inhibit lipogenesis in adipose tissue.(8,17) Downregulation of ACE2 expression results in unopposed angiotensin II (ANGII) action via angiotensin type 1 (AT1) receptor, with a resultant decrease in blood flow to tissues, impaired insulin signaling and increase oxidative stress.(17) This leads to reduced insulin secretion and increased insulin resistance with an increase in lipolysis, promoting ketogenesis, and enhanced gluconeogenesis and glycogenolysis in the liver and skeletal muscle, with worsening hyperglycaemia and a subsequent increase in viral replication.(17,19) β cells could also be indirectly damaged from hypoxia and inflammation secondary to SARS-CoV-2 infection of the islet microvasculature.(14) Patients with diabetes and COVID-19 have dysregulation of glucose homeostasis with further aggravation of inflammation leading to endothelial dysfunction, increased risk of thromboembolism and cardiorespiratory complications and mortality.(8) This highlights the bi-dierectional relationship between SARS-CoV-2 and diabetes. Long term follow-up will shed light as to whether COVID-19 will be associated with “permanent DM”. There is currently insufficient data to demonstrate whether damage to pancreatic β cells is long lasting or if there is reversibility beyond the acute phase into the post-acute and chronic phases of COVID-19.(20)
Potential mechanisms of COVID-19 causing hyperglycaemia and diabetes. Adapted from Satish T et al (17)
In South Africa, where an estimated one in every 11 adults has DM with a prevalence of 12.8%, and approximately 60% of people being undiagnosed, it is critical to ascertain a complete understanding of the impact of COVID-19 on the risk of developing DM.(21) However, uncontrolled HG, and not just new onset DM influences outcomes with COVID-19, and thus management of these two major pandemics not only requires a greater insight into novel biomarkers for earlier identification of severe disease that will direct therapy but simple measures that include glucose monitoring and the judicious use of insulin should become the focus of all medical protocols. Whilst this may not be a new disease, we need to prepare for the potential upcoming wave of diabetics with a new phenotype resulting from the COVID-19 pandemic.