Statistical analysis

The results of continuous variables were presented as median and interquartile range (IQR) values, and categorical variables were presented as counts and percentages. Continuous variables were compared among three groups by using one-way ANOVA on ranks (Kruskal–Wallis test), whereas categorical variables were compared using the chi-square test. The Spearman correlation was used to analyze the relationship between ferritin and other cytokines. The variables with statistical significance were considered for multivariable logistic regression analysis. The regression equation and the coefficient were obtained for each influence factor. The receiver operating characteristic (ROC) analysis was conducted to obtain the best threshold values of the ferritin, procalcitonin (PCT) and C-reactive protein (CRP) to distinguish the death and the survivor with COVID-19. Sensitivity, specificity, and area under the curve (AUC) were observed. Statistical analyses were performed using the IBM SPSS Statistic Version 26.0. p  < 0.05 was considered statistically significant.