ABSTRACT
OBJECTIVES: We aimed to identify high-risk factors for disease progression and fatality for coronavirus disease 2019 (COVID-19) patients. METHODS: We enrolled 2433 COVID-19 patients and used LASSO regression and multivariable cause-specific Cox proportional hazard models to identify the risk factors for disease progression and fatality. RESULTS: The median time for progression from mild-to-moderate, moderate-to-severe, severe-to-critical, and critical-to-death were 3.0 (interquartile range: 1.8-5.5), 3.0 (1.0-7.0), 3.0 (1.0-8.0), and 6.5 (4.0-16.3) days, respectively. Among 1,758 mild or moderate patients at admission, 474 (27.0%) progressed to a severe or critical stage. Age above 60 years, elevated levels of blood glucose, respiratory rate, fever, chest tightness, c-reaction protein, lactate dehydrogenase, direct bilirubin, and low albumin and lymphocyte count were significant risk factors for progression. Of 675 severe or critical patients at admission, 41 (6.1%) died. Age above 74 years, elevated levels of blood glucose, fibrinogen and creatine kinase-MB, and low plateleta count were significant risk factors for fatality. Patients with elevated blood glucose level were 58% more likely to progress and 3.22 times more likely to die of COVID-19. CONCLUSIONS: Older age, elevated glucose level, and clinical indicators related to systemic inflammatory responses and multiple organ failures, predict both the disease progression and the fatality of COVID-19 patients.
Subject(s)
Blood Glucose/metabolism , COVID-19/blood , COVID-19/mortality , Disease Progression , Hyperglycemia/blood , Adult , Age Factors , Aged , Aged, 80 and over , Bilirubin/blood , C-Reactive Protein/metabolism , China/epidemiology , Critical Illness , Female , Fever/virology , Humans , Hyperglycemia/complications , L-Lactate Dehydrogenase/blood , Lymphocyte Count , Male , Middle Aged , Proportional Hazards Models , Retrospective Studies , SARS-CoV-2 , Serum Albumin/metabolism , Time FactorsABSTRACT
The pandemic of Coronavirus Disease 2019 (COVID-19) is causing enormous loss of life globally. Prompt case identification is critical. The reference method is the real-time reverse transcription PCR (RT-PCR) assay, whose limitations may curb its prompt large-scale application. COVID-19 manifests with chest computed tomography (CT) abnormalities, some even before the onset of symptoms. We tested the hypothesis that the application of deep learning (DL) to 3D CT images could help identify COVID-19 infections. Using data from 920 COVID-19 and 1,073 non-COVID-19 pneumonia patients, we developed a modified DenseNet-264 model, COVIDNet, to classify CT images to either class. When tested on an independent set of 233 COVID-19 and 289 non-COVID-19 pneumonia patients, COVIDNet achieved an accuracy rate of 94.3% and an area under the curve of 0.98. As of March 23, 2020, the COVIDNet system had been used 11,966 times with a sensitivity of 91.12% and a specificity of 88.50% in six hospitals with PCR confirmation. Application of DL to CT images may improve both efficiency and capacity of case detection and long-term surveillance.