ABSTRACT
Background: In this study, we performed a bidirectional mendelian randomization analysis on circulating cytokines and critically ill COVID-19. Methods: Both the exposure and outcome data were obtained from public genome wide association study (GWAS) database. We extracted independent instrumental variables from exposure at genome level significance (P < 5 × 10-8). Wald ratio or inverse variance weighted (IVW) method were used for estimating the causal relationships between circulating cytokines and critically ill COVID-19. Results: Only IL5 (cytokines to critically ill COVID-19 direction) and bNGF, IL8 (critically ill COVID-19 to cytokines direction) showed suggestive causal relations. However, these associations lost significance after FDR correction. Another validation data set of critically ill COVID-19 did not confirm these associations, either. Conclusions: Our Mendelian randomization did not find causal relationships between analyzable circulating cytokines and critically ill COVID-19.