Causal relationships between gut microbiota, metabolites, and chronic kidney disease: a two-step, two-sample Mendelian randomization study
Chronic kidney disease (CKD) is a major global health concern with a multifactorial pathogenesis. Emerging evidence suggests that the gut microbiota and its metabolic products contribute to CKD development. However, robust evidence supporting metabolites as causal mediators in this pathway remains limited. We aimed to investigate causal relationships and mediating effects linking gut microbiota, circulating metabolites, and CKD. Two-sample Mendelian randomization (MR) analyses were conducted using summary statistics from large-scale genome-wide association studies (GWAS) on gut microbiota, plasma metabolites, and CKD-related outcomes. Independent single-nucleotide polymorphisms strongly associated with each exposure were selected as instrumental variables. The inverse-variance weighted method served as the primary analysis, complemented by four additional MR approaches. Heterogeneity and pleiotropy were assessed using Cochran’s Q test, the MR-Egger intercept, and MR-PRESSO. Mediation analysis quantified the proportion of microbiota–CKD effects mediated by specific metabolites. MR analyses identified causal associations among multiple gut microbial pathways, circulating metabolites, and CKD-related traits. Mediation analysis revealed that argininate partially mediated the effect of the microbial pathway PWY-6629 on the urinary albumin-to-creatinine ratio (UACR), accounting for 22% of the total effect. A one–standard-deviation increase in genetically predicted PWY-6629 activity was associated with a 1.3% increase in UACR, mediated through a 7.9% reduction in argininate levels. Conversely, higher genetically predicted argininate levels were associated with a 3.7% reduction in UACR. This study provides genetic evidence that gut microbiota influence CKD risk partially through circulating metabolites, highlighting the gut microbiota–metabolite–CKD axis as a potential therapeutic target.