Table 3.
Meta-analysis of 18 studies reporting the association between 16 hematological biomarkers and severity of COVID-19 infection
| Biomarkers | N of Study | SMD (95%CI) | I2 | Chi-square (P-value) | Egger test (P-value) |
|---|---|---|---|---|---|
| Lymphocyte (×109/L) | 9 [21,22,23,25,26.27,28,29,31] | 1.48 (1.30; 1.67) | 97.0 | 289.26 (P < 0.001) | 0.69 (P = 0.51) |
| CD3+ T cell (×1012/L) | 8 21,23,25,27,28,29,30,33 | 1.28 (1.09; 1.47) | 94 | 108.54 (P < 0.001) | 1.69 (P = 0.23) |
| Neutrophils (×109/L) | 9 21,22,23,25,26,27,28,29,31 | -2.06 (-2.26; -1.86) | 96 | 178.38 (P < 0.001) | 0.01 (P = 0.99) |
| PLT (×109/L) | 2 23,25,27 | 0.37 (0.05; 0.69) | 0.0 | 1.81 (P=0.40) | 0.09 (P = 0.94) |
| CD4+ T cell (×1012/L) | 8 21,23,25,27,28,29,30,33 | 2.11 (1.91; 2.31) | 82 | 39.37 (P < 0.001) | 2.84 (P = 0.04) |
| CD8+ T cell (×1012/L) | 8 21,23,25,27,28,29,30,33 | 1.00(0.83; 1.18) | 88 | 58.60 (P < 0.001) | 1.82 (P = 0.11) |
| CRP (mg/L) | 9 22,23,24,25,27,28,31,33,35 | -0.83 (-1.07; -0.58) | 96 | 226.59 (P < 0.001) | 1.2 (P = 0.36) |
| DD-dimer (μg/mL) | 4 23,25,26,28 | -1.75 (-1.99; -1.50) | 98 | 122.56 (P < 0.001) | 1.37 (P = 0.3) |
| NK cell (×106/L) | 3 21,25,30 | 21.21 (18.28; 24.15) | 56 | 4.50 (P < 0.001) | 0.81 (P = 0.46) |
| WBC (×109/L) | 8 21,22,27,28,29,31,32,36 | -1.30 (-.1.54; -1.07) | 95 | 141.24 (P < 0.001) | 0.42 (P = 0.05) |
| FDP (mg/L) | 4 27,36,37,38 | -1.19 (-1.57; -0.81) | 94 | 52.53 (P < 0.001) | 2.02 (P = 0.09) |
| TT (s) | 4 27,36,37,38 | 0.30 (-0.03; 0.64) | 78 | 13.61 (P < 0.001) | 1.06 (P = 0.32) |
| APTT (s) | 4 27,36,37,38 | -0.26 (-0.63; 0.10) | 94 | 51.80 (P < 0.001) | 0.37 (P = 0.73) |
| IL-10 (pg/ml) | 8 21,23,25,27,28,29,30,31 | -1.58 (-1.86; -1.31) | 97 | 221.08 (P < 0.001) | 1.66 (P = 0.34) |
| IL-6 (pg/ml) | 15 21,22,23,24,25,26,27,28,29,30,31,32,33,34,35 | -6.08 (-6.41; -5.75) | 95 | 298.64 (P < 0.001) | 0.91 (P = 0.53) |
| PDW (%) | 4 22,31,37,38 | 0.43 (0.10; 0.75) | 94 | 47.26 (P < 0.001) | 0.67 (P = 0.31) |
N: number of studies used. SMD: Standard Mean Difference. I2 was used for quantifying inconsistency: the larger the value, the stronger the heterogeneity.