Table 1.
Demographic and clinical characteristics.
Fig 1.
(A) Box plots illustrating the concentration of cytokines according to the BMI levels in COVID-19 patients with BMI ≤ 25, 25 < BMI < 30, and BMI ≥ 30. The colored squares associated with each cytokine illustrate the different orientations of the immune response (n = 42 patients). (B) Radar plots illustrating the fold increase of selected cytokines in patients with BMI ≤ 25 (n = 11), 25 < BMI < 30 (n = 19), and BMI ≥ 30 (n = 12), as compared to healthy donors (n = 18). (C) Characterization of peripheral blood mononuclear cells according to the BMI categories (above and below the median BMI of 26.8, n = 29) using polychromatic flow cytometry and Uniform Manifold Approximation and Projection representation of the 24 clusters identified. Each C stands for one of the 24 « Cluster » automatically identified and defined by the software. Each cluster corresponds to a group of cells with comparable phenotype i.e that express similar levels of the different markers (at their surface or intracellular) (D) Box plots comparing the SARS-CoV-2 RNAaemia measured at the zenith of the cytokines levels, in patients (n = 50) according to the BMI levels (BMI ≤ 25, 25 < BMI < 30, and BMI ≥ 30). Wilcoxon-Mann-Whitney test used for pairwise comparisons in Fig 1D followed by the Benjamini Hochberg test for multiple testing correction in Fig 1A.
Fig 2.
(A) Left part: box plots illustrating BMI in COVID-19 patients (n = 50) according to the level of respiratory distress (the WHO scale). Wilcoxon-Mann-Whitney test used for pairwise comparisons. (B) Left part: High-resolution axial computed tomography (CT) images of two patients show ground glass opacity (GGO; arrows) with a low score #2 and consolidation (arrowheads) with a moderate score #3. Right part: The correlations between the consolidation scale, GGO scale, and BMI. R for the Kendall rank correlation coefficient. (C) The Spearman correlation heatmap depicting the negative (blue), positive (red), and no (white) correlation between consolidation or the GGO extent evaluated based on chest CT images and the blood cytokines levels in COVID-19 patients (n = 19) according BMI below or above 26, i.e. the median BMI of patients with available CT (D) Bar plots showing the Spearman rank correlation coefficient of IFNγ, IL-6, and IL-8 with consolidation (top) or the GGO extent (bottom) according to the BMI levels (n = 19; *p < 0.05; **p < 0.01).
Fig 3.
(A) Axial computed tomography images showing normal (up) and low attenuation (down) values of the liver parenchyma reflecting steatosis (n = 23). (B) Left panel: violin plot representing the biological hepatic steatosis index (HSI) according to BMI (≤ 25 in green, 25 < BMI < 30 in yellow, ≥ 30 in red). Middle panel: Box plots depicting the plasma leptin levels of patients (n = 51) according to BMI (BMI ≤ 25, 25 < BMI < 30, ≥ 30). Wilcoxon-Mann-Whitney test used for pairwise comparisons Left panel: Box plots representing the HSI levels according to severity of pulmonary distress (the WHO scale). (C) The Pearson correlation heatmap depicts the negative (blue), positive (red), and no (white) correlation between markers of liver dysfunction (AST, ALT, gGT, TBIL, and PT) and the blood cytokines levels in in patients (n = 51) according to BMI (BMI ≤ 25, 25 < BMI < 30, BMI ≥ 30). (D) Linear regression curves estimated between ALT, total bilirubin, and plasma concentration of IL-1β, IL-12 and IL-2 in non-overweight patients (BMI ≤ 25, in light blue) and in patients with obesity (BMI ≥ 30, in dark blue). Patients with 25 < BMI < 30 are presented in grey.
Fig 4.
(A) Top panel: Linear regression curves between plasma concentration of IL-6, IFNγ, TNFα, GM-CSF, and the maximum level of creatinine. Bottom panel: bar plots showing the coefficient of the Spearman correlation of IL-6, IFNγ, TNFα, and GM-CSF with the maximum level of creatinine according to BMI (BMI ≤ 25, 25 < BMI < 30, BMI ≥ 30; *p < 0.05; **p < 0.01, ***p < 0.001). (B) The Spearman correlation heatmap depicts the negative (blue), positive (red), and no (white) correlation between markers of kidney dysfunction (i.e. maximum creatinine, creatinine, the albumin/creatinine ratio with protein/creatinine, the Na/K ratio, and the blood cytokines levels according to BMI (BMI ≤ 25, 25 < BMI < 30, BMI ≥ 30).
Fig 5.
(A) Box plots illustrating the plasma levels of endothelial activation markers (IL-8, endoglin, E and P-selectins, PIGF, Ang-2) according to BMI (BMI ≤ 25, 25 < BMI < 30, BMI ≥ 30). (B) The Spearman correlation heatmap depicts the negative (blue), positive (red), and no (white) correlation between plasma markers of endothelial activation and the blood cytokines levels in COVID-19 patients (n = 38) according to BMI (BMI ≤ 25, 25 < BMI < 30, BMI ≥ 30).
Fig 6.
Diagram representing the cytokines correlated with multiorgan damage of endothelial vessels, lung (consolidation and GGO), liver, and kidney in patients with obesity (BMI ≥ 30) as compared to normal-weight patients (BMI ≤ 25).
Inflammatory cytokines are illustrated in white and noninflammatory in black.