Effects of hyperglycaemia on complications of <scp>COVID</scp> ‐19: A meta‐analysis of observational studies

Abstract

To the Editor We read with great interest the article by Ceriello et al. regarding the worsened prognosis for coronavirus disease 2019 (COVID-19) in patients with hyperglycaemia and the potential reasons for this phenomenon.1 COVID-19 is currently the greatest public health concern, with global cases surpassing 12 million and attributed deaths reaching over half a million.2 There are concerns that patients with hyperglycaemia may face a high risk of severe COVID-19 with reports of association between high blood glucose and poor prognosis of COVID-19,3, 4 as in the case of the severe acute respiratory syndrome (SARS) epidemic in 2003.5 Despite this, the literature specifically assessing the outcomes of COVID-19 in hyperglycaemic patients is scarce. Hence, this meta-analysis aims to determine if hyperglycaemia may be associated with an increased severity of COVID-19. This review was registered with PROSPERO (CRD42020195937). PubMed, Embase and China National Knowledge Infrastructure were searched for articles examining the outcomes of COVID-19 in hyperglycaemic and normoglycaemic patients. Terms including COVID-19, 2019-n-CoV and SARS-CoV-2 were used in the search strategy, which is shown in supplementary material 1. Original articles that examined the severity of COVID-19 in hyperglycaemic patients, determined by variables such as mortality and ICU admissions were included and full-text articles were reviewed for relevant outcomes. Binary data reporting the COVID-19 outcomes among both hyperglycaemic and normoglycaemic patients diagnosed with SARS-CoV-2 were extracted and analysed with the metaprop function and pooled in random effects.6 Dichotomous data were analysed in random effects using the DerSimonian and Laird model.7 Continuous data measuring hospital stay duration were extracted as mean and standard deviation and pooled as the mean difference.8 In total, 189 articles were retrieved using the search strategy, and subsequently 51 full-text articles were reviewed, of which eight studies satisfied our inclusion criteria to be included in the final analysis.9-16 The results of the analysis are summarized in supplementary material 2, 3 and 4. In total, 1447 hyperglycaemic and 1282 normoglycaemic patients with COVID-19 were included in the analysis. Among hyperglycaemic patients with COVID-19, which included patients with diabetes mellitus and those without the condition, 17% died [confidence interval (CI): 0.10-0.25], 11% required ICU admissions (CI: 0.05-0.19) and 23% required mechanical or non-mechanical ventilation (CI: 0.13-0.36). Furthermore, 72% of patients developed severe or critical COVID-19 (CI: 0.44-0.93), 14% had acute cardiac injury (CI: 0.07-0.24), 5% had acute kidney injury (CI: 0.02-0.09) and 15% developed acute respiratory distress syndrome (CI: 0.06-0.28). The duration of hospital stay trended higher among hyperglycaemic patients compared with normoglycaemic patients but did not reach statistical difference (mean difference: 1.834, CI: −1.234 to 4.901, P = .24). Compared with normoglycaemic patients, hyperglycaemic patients had a higher risk of mortality and would probably need admission to ICU, or mechanical or non-mechanical ventilation (supplementary material 2). Hyperglycaemic patients also probably developed severe or critical COVID-19, acute kidney injury and acute respiratory distress syndrome compared with normoglycaemic patients (supplementary material 2). Among diabetic COVID-19 patients with hyperglycaemia (n = 681), 16% died (CI: 0.09-0.25), 15% required ICU admission (CI: 0.10-0.21) and 16% required mechanical or non-mechanical ventilation (CI: 0.13-0.19). In addition, 89% of patients developed severe or critical COVID-19 (CI: 0.82-0.94), 15% had acute cardiac injury (CI: 0.07- 0.25), 6% acute kidney injury (CI: 0.01-0.16) and 15% developed acute respiratory distress syndrome (CI: 0.08-0.23). When compared with diabetic COVID-19 patients with controlled blood glucose, diabetic COVID-19 patients with hyperglycaemia were more significantly associated with death and the need for mechanical or non-mechanical ventilation. Diabetic COVID-19 patients with hyperglycaemia were also at higher risk of developing acute cardiac injury, acute kidney injury or acute respiratory distress syndrome (supplementary material 3). Of the non-diabetic patients with hyperglycaemia (n = 158), 5% died (CI: 0.02- 0.10), 5% required ICU admissions (CI: 0.02-0.10) and 38% needed mechanical or non-mechanical ventilation (CI: 0.21-0.59). Additionally, 75% of patients developed severe or critical COVID-19 (CI: 0.68-0.82), 20% had acute cardiac injury (CI: 0.14-0.28), 3% acute kidney injury (CI: 0.01-0.08) and 2% developed acute respiratory distress syndrome (CI: 0.00-0.06). When compared with normoglycaemic COVID-19 patients, non-diabetic patients with hyperglycaemia had a higher risk of ICU admission, mechanical or non-mechanical ventilation and severe or critical COVID-19 (supplementary material 4). Hyperglycaemia has been reported to be a predisposing factor for death in COVID-19,17 as well as for the progression of non-critical disease to critical disease.3 Our analysis affirms the higher risk of death among COVID-19 patients with hyperglycaemia, including patients with diabetes and those without diabetes. We urge for close surveillance for hyperglycaemia of patients with COVID-19 individuals in non-diabetics as it might alter the prognosis of these patients. In our analysis, hospitalization stay trended higher among hyperglycaemic patients compared with normoglycaemic patients in our analysis, but did not reach statistical significance. Previous observational studies have also demonstrated the association between hyperglycaemia and a longer hospital stay.18 As such, the non-significance of our results was probably attributed to small sample size. The mechanisms through which hyperglycaemia may contribute to COVID-19 progression include dysregulation of the host immune response, with alterations in the production of cytokines such as interleukin-6, as well as changes in the function of immune cells.12, 19 Ceriello et al. also highlighted the interplay between hyperglycaemia and COVID-19 progression, where cytokine production from COVID-19 infection may induce or worsen insulin resistance, or impair insulin secretion,1 contributing to the hyperglycaemic state. Hyperglycaemia additionally promotes glycosylation of the angiotensin-converting enzyme 2 receptor, which facilitates SARS-CoV-2 virus binding to the host, thereby worsening the extent of host cell infection by SARS-CoV-2.20 Furthermore, hyperglycaemia may elevate gene expression of matrix metalloproteinases, which may promote the spread of inflammation.21 Lastly, hyperglycaemia may promote thrombosis by increasing oxidative stress and decreasing heparan sulphate levels.22 This is supported by findings from Li et al. who reported an elevation of D-dimer levels in patients with poorly controlled blood glucose.11 Diabetes mellitus is a well-established risk factor of severe COVID-19 outcomes.23, 24 Our analysis of diabetic COVID-19 patients supports the notion that they are a vulnerable patient population in the COVID-19 pandemic, as they are at increased risk of adverse COVID-19 outcomes, including death. It is possible that COVID-19 will exacerbate hyperglycaemia through inflammation and counter-regulatory hormones in diabetic patients, leading to a more severe disease. The development of stress hyperglycaemia during critical illness also affects non-diabetic patients, and arises from the increased production of glucose by the liver due to the action of cortisol.25 High serum cortisol has been found associated with an increased mortality in patients with COVID-19,26 which reiterates the interplay between stress response, hyperglycaemia and poor COVID-19 prognosis. We showed that among non-diabetic COVID-19 patients with hyperglycaemia, there was an association between hyperglycaemia and several COVID-19 adverse outcomes. Consequently, this implies that the hyperglycaemia milieu, which influences the course of COVID-19, may not be unique to diabetic patients. In fact, Bode et al. and Li et al. have shown that uncontrolled hyperglycaemia carries a worse prognosis with or without known history of diabetes.9, 11 However, it is unclear whether a diagnosis of diabetes influences the severity of COVID-19 outcomes among patients with hyperglycaemia. Bode et al. found that patients with COVID-19 and with uncontrolled hyperglycaemia but without previous history of diabetes had a significantly higher mortality than those with known diabetes.9 Conversely, Li et al. reported a higher relative risk of mortality in patients with pre-existing or newly diagnosed diabetes compared with hyperglycaemic patients without diabetes.10 Interestingly, in multivariate regression models (adjusted for variables such as age, gender, total cholesterol and the use of antihypertensive agents), patients with known diabetes had a lower relative risk of death from COVID-19 than patients with newly diagnosed diabetes were. However, caution must be taken when interpreting these results as the definition of new onset diabetes may vary across regions and centres and, consequently, patients with the condition may go undiagnosed. With these findings, it is possible that blood glucose levels may serve as a biomarker for risk stratification in the clinical management of patients with COVID-19, as proposed by Wu et al.3 Studies have shown that hyperglycaemia in hospitalized patients confer serious medical consequences, and that aggressive glycaemic control improved outcomes, in particular mortality, multi-organ failure and ICU stays.18 This provides support for the recommendation for screening for hyperglycaemia and the monitoring of blood glucose levels for all patients admitted to hospital for COVID-19 regardless of diabetes status, so that treatment can be rendered if needed.17 So far, Sardu et al. showed that insulin treatment in hyperglycaemic patients had a lower risk of severe disease compared with those not treated with insulin.12 It is currently not clear whether the level of glucose control or the means to achieve glucose control or both would improve the outcomes in patients with COVID-19. Our analysis has several limitations. The first relates to the paucity of literature that has currently reported on the outcomes of COVID-19, stratified by diabetes and hyperglycaemic status. Furthermore, the blood glucose ranges used to classify patients as having hyperglycaemia were not consistent across the studies, which may limit the interpretation of our findings. On a similar note, the definition of severe or critical COVID-19 may be different across nations. However, the studies included in our analysis relied on the guidelines issued by or were modified from the Chinese National Health Commission and the World Health Organization,27, 28 and is thus unlikely to influence our analysis. Lastly, because of the limited studies, we were not able to perform tests for publication bias. In conclusion, this meta-analysis affirms the findings that hyperglycaemia worsens the prognosis of patients with COVID-19 infection, with or without known diabetes mellitus. Early detection and treatment of hyperglycaemia might improve the disease outcomes. None declared. Ming Hui Lee and Chloe Wong participated in the extraction and interpretation of data, as well as writing the manuscript. Cheng Han Ng planned and designed the study, and participated in the analysis and interpretation of data as well as reviewing the manuscript. Amanda Yuan Ling Lim and Chin Meng Khoo planned and designed the study, and participated in the interpretation of data and review of the manuscript. The peer review history for this article is available at https://publons.com/publon/10.1111/dom.14184. The data that support the findings of this study are available in PubMed, at Pubmed reference number 32389027, 32469464, 32623030, 32430456, 32632527, 32457924, 32406594 and 32369736. These data were derived from the following resources available in the public domain: Pubmed Appendix S1. Supporting information Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts