Abstract
Tuberculosis (TB), caused by infectious agent Mycobacterium tuberculosis (Mtb) seriously poses a great threat to health. An array of metabolites generated by metabolic pathways are essential for Mtb pathophysiology. However, a specific causal relationship between TB and human metabolites remains indistinct. This study aimed to investigate the relationship between 1400 metabolites and TB by Mendelian randomization (MR) analysis. In this study, a total of 1400 metabolites were utilized as exposure factors, while TB-related data served as the outcomes. And TwoSampleMR package and R software were adopted to perform this MR analysis. Various regression fitting methods were employed to conduct MR analysis, including inverse variance weighted (IVW), MR-Egger, weighted median, simple mode, and weighted mode. In addition, potential biases arising from linkage disequilibrium and weak instrumental variables were considered. Metabolites that failed to meet the criteria in both the heterogeneity and pleiotropy tests were considered to have no substantial causal influence on the results, ensuring the robustness and reliability of our analysis. IVW analysis showed that six human metabolites exhibited a significant causal influence (P < 0.05) on TB. Among them, dodecanedioate, myristoleate (14:1n5), and 1-(1-enyl-palmitoyl)-2-arachidonoyl-GPE(p-16:0/20:4) demonstrated a strong causally positive effect on TB, indicating that with the increase of these metabolites, TB progressed robustly. Glycerol 3-phosphate, sphingomyelin (d18:1/20:2, d18:2/20:1, and d16:1/22:2), and 2-methylserine were significantly negatively associated with TB, an increase in these metabolites inhibited TB progression. This is the first time to reveal the causal effects of human metabolites on TB through MR, and the metabolites may be potential biomarkers candidate for TB diagnosis, and monitoring these metabolites might have great clinic significance for TB diagnosis and treatment in the future.