Objective:To explore the causal relationship between thyroid dysfunction and sepsis based on the bidirectional two-sample Mendelian randomization (MR) method.Methods:The genome-wide association study (GWAS) dataset were selected to screen single nucleotide polymorphisms (SNP) associated with thyroid dysfunction as instrumental variable (IV) for genetic variation, using hypothyroidism and hyperthyroidism as exposure factor and sepsis as outcome factor. Potential causal relationship between thyroid dysfunction and sepsis was analyzed using a bidirectional two-sample MR method primary analysis method of inverse-variance weighted (IVW). Potential pleiotropic analysis of SNP was performed using the MR Egger regression intercept test. Sensitivity analysis was performed using the "leave one out" test. Reverse MR method was used to prove the causal relationship.Results:The GWAS data were screened based on the three main assumptions of MR, resulting in 101 SNP strongly associated with hypothyroidism and 10 SNP strongly associated with hyperthyroidism entering the MR analysis. The results of the MR using the IVW method showed that the risk of sepsis in individuals with hypothyroidism was 2.293 times higher than those without hypothyroidism [odds ratio ( OR) = 2.293, 95% confidence interval (95% CI) was 1.199-4.382, P = 0.012]. There was no significant difference in the risk of sepsis between hyperthyroid and non-hyperthyroid populations ( OR = 1.049, 95% CI was 0.999-1.100, P = 0.560). MR Egger regression intercept test showed that the included SNP did not have pleiotropy, and the MR-PRESSO test did not find outliers. Sensitivity analysis suggested that the results of MR were stable. The results of the reverse MR analysis showed that the reverse causal relationship between hyperthyroidism and sepsis was not proved ( OR = 0.996, 95% CI was 0.988-1.004, P = 0.338), which further confirmed the robust MR analysis result. Conclusion:The results of the bidirectional two-sample MR analysis show that hypothyroidism can increase the risk of sepsis onset, while there is no causal relationship between hyperthyroidism and sepsis.