Background: The roles of inflammation and hypercoagulation in predicting outcomes of coronavirus disease 2019 (COVID-19) are unclear.
Methods: Adult patients diagnosed with COVID-19 from 28 January 2020 to 4 March 2020 in Tongji Hospital, Wuhan were recruited. Data on related parameters were collected. Univariate analysis and multivariable binary logistic regression were used to explore predictors of critical illness and mortality.
Results: In total, 199 and 44 patients were enrolled in the training and testing sets, respectively. Elevated ferritin, tumor necrosis factor-α and D-dimer and decreased albumin concentration were associated with disease severity. Older age, elevated ferritin and elevated interleukin-6 were associated with 28-day mortality. The FAD-85 score, defined as age + 0.01 * ferritin +D-dimer, was used to predict risk of mortality. The sensitivity, specificity and accuracy of FAD-85 were 86.4%, 81.8% and 86.4%, respectively. A nomogram was established using age, ferritin and D-dimer to predict the risk of 28-day mortality.
Conclusions: Thrombo-inflammatory parameters provide key information on the severity and prognosis of COVID-19 and can be used as references for clinical treatment to correct inflammatory and coagulation abnormalities.
Keywords: COVID-19; coagulation; inflammation; model; mortality; predictor.