COVID-19 Inflammatory Syndrome: Lessons from TNFRI and CRP about the Risk of Death in Severe Disease

Abstract

Background/Objectives: Cytokine storm in severe COVID-19 is responsible for irreversible tissue damage and death. Soluble mediators from the TNF superfamily, their correlation with clinical outcome, and the use of TNF receptors as a potent predictor for clinical outcome were evaluated. Methods: Severe COVID-19 patients had the levels of soluble mediators from the TNF superfamily quantified and categorized according to the clinical outcome (death versus survival). Statistical modeling was performed to predict clinical outcomes. Results: COVID-19 patients have elevated serum levels from the TNF superfamily. Regardless of sex and age, the sTNFRI levels were observed to be significantly higher in deceased patients from the first weeks following the onset of symptoms. We analyzed hematological parameters and inflammatory markers, and there was a difference between the groups for the following factors: erythrocytes, hemoglobin, hematocrit, leukocytes, neutrophils, band cells, lymphocytes, monocytes, CRP, IL-8, IFN-γ, IL-10, IL-6, IL-4, IL-2, leptin MIF sCD40L, and sTNFRI (p < 0.05). A post hoc analysis showed an inferential capacity over 70% for some hematological markers, CRP, and inflammatory mediators in deceased patients. sTNFRI was strongly associated with death, and the sTNFRI/sTNFRII ratio differed between outcomes (p < 0.001; power above 90%), highlighting the impact of these proteins on clinical results. The final logistic model, including sTNFRI/sTNFRII and CRP, indicated high sensitivity, specificity, accuracy, and an eight-fold higher odds ratio for an unfavorable outcome. Conclusions: The joint use of the sTNFRI/sTNFRII ratio with CRP proves to be a promising tool to assist in the clinical management of patients hospitalized for COVID-19.

Keywords: C-reactive protein; COVID-19; SARS-CoV-2; TNF receptors; TNF superfamily; clinical outcomes predictor; inflammatory mediators.

Figure 1

TNF family soluble factors, their relationship to COVID-19 active patient and their kinetic in weeks, after symptoms begin, according to clinical outcomes. TNF family soluble factor levels in the serum of healthy individuals and patients with active COVID-19 by the ELISA method. The results are expressed in pg/mL. In (A,D,G,J), soluble TNF-α, TNFRI, TNFRII, and CD40L levels can be seen, respectively, according to the clinical diagnosis (healthy and active COVID-19). In (B,E,H,K), soluble TNF-α, TNFRI, TNFRII, and CD40L levels are observed, respectively, according to the clinical outcomes (death and discharge). Patients with active disease AP/COVID-19^pos group, n = 214), patients with active disease and evolution to death (DP/COVID-19^pos group, n = 90), patients with active disease and evolution to hospital discharge (SP/COVID-19^pos group, n = 124), and healthy donors (HD/COVID-19^neg group, n = 14). The horizontal line represents the median; the bar shows the 25% to 75% quartiles, whereas the vertical line depicts the 10% and 90% percentiles. The statistical differences between groups are expressed in the graphic. In the graphs on the right, the p-values above the boxplot refer to the statistical differences in relation to the control group (healthy donors). Mann–Whitney test, Kruskal–Wallis test, Dunn’s post-test, and ANOVA were used for this study. In (C,F,I,L), soluble TNF-α, TNFRI, TNFRII, and CD40L levels are shown, respectively, following the symptom onset in weeks, according to the clinical course ( survivor and death). Patients with active disease and evolution to death (DP/COVID-19^pos group, n = 168) and patients with active disease and who survived (SP/COVID-19^pos group, n = 154) are also presented. The points represent standard error of the mean (SEM). The statistical differences between groups during the weeks are expressed in the graphic. Multiple t-tests were used, and the results were considered statistically significant at p < 0.05.

Figure 2

Graphical representation for the dis(similarity) matrix between the explanatory variables for the association with the prognosis of symptomatic patients for COVID-19. A dis(similarity) matrix was generated for the explanatory variables using a multivariate grouping method, the Euclidean distance. In (A), the distribution of groups for patients whose outcome was discharge. In (B), matrix data for the groupings of patients whose outcome was death. TNFRI: tumor necrosis factor receptor I; TNFRII: tumor necrosis factor receptor II; CRP: C-reactive protein.