Abstract
BACKGROUND: COVID-19 has spread widely worldwide, causing millions of deaths. We aim to explore the association of immunological features with COVID-19 severity. METHODS: We conducted a meta-analysis to estimate mean difference (MD) of immune cells and cytokines levels with COVID-19 severity in PubMed, Web of Science, Scopus, the Cochrane Library and the grey literature. RESULTS: A total of 21 studies with 2033 COVID-19 patients were included. Compared with mild cases, severe cases showed significantly lower levels of immune cells including CD3(+) T cell (× 10(6), MD, - 413.87; 95%CI, - 611.39 to - 216.34), CD4(+) T cell (× 10(6), MD, - 203.56; 95%CI, - 277.94 to - 129.18), CD8(+) T cell (× 10(6), MD, - 128.88; 95%CI, - 163.97 to - 93.79), B cell (× 10(6)/L; MD, - 23.87; 95%CI, - 43.97 to - 3.78) and NK cell (× 10(6)/L; MD, - 57.12; 95%CI, - 81.18 to - 33.06), and significantly higher levels of cytokines including TNF-α (pg/ml; MD, 0.34; 95%CI, 0.09 to 0.59), IL-5 (pg/ml; MD, 14.2; 95%CI, 3.99 to 24.4), IL-6 (pg/ml; MD, 13.07; 95%CI, 9.80 to 16.35), and IL-10 (pg/ml; MD, 2.04; 95%CI, 1.32 to 2.75), and significantly higher levels of chemokines as MCP-1 (SMD, 3.41; 95%CI, 2.42 to 4.40), IP-10 (SMD, 2.82; 95%CI, 1.20 to 4.45) and eotaxin (SMD, 1.55; 95%CI, 0.05 to 3.05). However, no significant difference was found in other indicators such as Treg cell (× 10(6), MD, - 0.13; 95%CI, - 1.40 to 1.14), CD4(+)/CD8(+) ratio (MD, 0.26; 95%CI, - 0.02 to 0.55), IFN-γ (pg/ml; MD, 0.26; 95%CI, - 0.05 to 0.56), IL-2 (pg/ml; MD, 0.05; 95%CI, - 0.49 to 0.60), IL-4 (pg/ml; MD, - 0.03; 95%CI, - 0.68 to 0.62), GM-CSF (SMD, 0.44; 95%CI, - 0.46 to 1.35), and RANTES (SMD, 0.94; 95%CI, - 2.88 to 4.75). CONCLUSION: Our meta-analysis revealed significantly lower levels of immune cells (CD3(+) T, CD4(+) T, CD8(+) T, B and NK cells), higher levels of cytokines (TNF-α, IL-5, IL-6 and IL-10) and higher levels of chemokines (MCP-1, IP-10 and eotaxin) in severe cases in comparison to mild cases of COVID-19. Measurement of immunological features could help assess disease severity for effective triage of COVID-19 patients.