RESUMO
We aimed to assess whether blood glucose control can be used as predictors for the severity of 2019 coronavirus disease (COVID-19) and to improve the management of diabetic patients with COVID-19. A two-center cohort with a total of 241 confirmed cases of COVID-19 with definite outcomes was studied. After the diagnosis of COVID-19, the clinical data and laboratory results were collected, the fasting blood glucose levels were followed up at initial, middle stage of admission and discharge, the severity of the COVID-19 was assessed at any time from admission to discharge. Hyperglycemia patients with COVID-19 were divided into three groups: good blood glucose control, fair blood glucose control, and blood glucose deterioration. The relationship of blood glucose levels, blood glucose control status, and severe COVID-19 were analyzed by univariate and multivariable regression analysis. In our cohort, 21.16% were severe cases and 78.84% were nonsevere cases. Admission hyperglycemia (adjusted odds ratio [aOR], 1.938; 95% confidence interval [95% CI], 1.387-2.707), mid-term hyperglycemia (aOR, 1.758; 95% CI, 1.325-2.332), and blood glucose deterioration (aOR, 22.783; 95% CI, 2.661-195.071) were identified as the risk factors of severe COVID-19. Receiver operating characteristic (ROC) curve analysis, reaching an area under ROC curve of 0.806, and a sensitivity and specificity of 80.40% and 68.40%, respectively, revealed that hyperglycemia on admission and blood glucose deterioration of diabetic patients are potential predictive factors for severe COVID-19. Our results indicated that admission hyperglycemia and blood glucose deterioration were positively correlated with the risk factor for severe COVID-19, and deterioration of blood glucose may be more likely to the occurrence of severe illness in COVID-19.