**Background:** A bidirectional kidney–gut axis has been described in chronic kidney disease (CKD), where gut dysbiosis may promote CKD progression while CKD itself alters gut microbiota composition. Gut microbiota produces short-chain fatty acids, neurotransmitters, bile acids, and other metabolites that influence systemic health. In CKD, dysbiosis triggers production of detrimental metabolites (indoxyl sulfate, p-cresyl sulfate) linked to increased mortality and cardiovascular risk, reduces beneficial short-chain fatty acids, and increases gut permeability leading to systemic inflammation. This systematic review aimed to characterize gut microbiota composition across CKD stages, including end-stage kidney disease (ESKD), and evaluate modulation strategies and their impact on clinical outcomes.
**Methods:** A literature search was conducted in MEDLINE (PubMed), Embase, Scopus, and Cochrane databases from inception to 30 June 2022, using combinations of keywords including "microbiota," "gut," "chronic kidney disease," "hemodialysis," "mortality," "probiotic," and "synbiotic." Both randomized clinical trials and observational studies enrolling patients ≥18 years with CKD of all stages were included. Two independent investigators performed screening and data extraction. Quality assessment used the revised Cochrane risk-of-bias tool (RoB 2) for randomized trials and the Newcastle–Ottawa Scale for observational studies. The protocol was registered in PROSPERO (CRD42022369573).
**Key Results:** From 3290 retrieved references, 69 studies met inclusion criteria (57 observational, 12 randomized trials). Twenty-one studies investigated ESKD patients. Microbiota diversity was significantly decreased in CKD patients compared to healthy individuals. At the genus level, Ruminococcus had good discriminatory power for early-stage CKD (AUC = 0.771, 95% CI 0.771–0.852), while Roseburia accurately identified healthy controls (AUC = 0.803, 95% CI 0.804–0.864). Bacteroides eggerthii showed AUC = 0.80 (95% CI 0.67–0.93) for early-stage CKD, outperforming protein/creatinine ratio (AUC = 0.64). A model based on 25 gut microbiota dissimilarities achieved excellent predictive power for diabetic nephropathy (AUC = 0.972). Roseburia abundance was consistently decreased in CKD patients, especially in ESKD (p < 0.001), and decreased further with disease progression. In ESKD, deceased hemodialysis patients showed lower Simpson and Shannon diversity indices (p = 0.007 and p = 0.028, respectively) and increased levels of Oscillospira, Achromobacter, Agrobacterium, Lactobacillus, Alloscardovia, Anoxybacillus, Devosia, and Yersinia compared to survivors. Lower Bacteroides and Phascolarctobacterium levels were associated with cardiovascular mortality in ESKD (p < 0.05). Peritonitis in peritoneal dialysis patients was associated with higher Bacteroidetes and Synergistetes and lower Bacilli and Lactobacillus abundance. Regarding modulation therapies, synbiotics increased Bifidobacterium levels up to 5-fold from baseline (p = 0.003), though one study reported an eGFR decrease of 3.14 mL/min/1.73 m² (p < 0.01) with synbiotic therapy. Probiotics increased Bacteroidaceae, Enterococcaceae, Lactobacillales, and Bifidobacteria levels (p < 0.001) in hemodialysis patients.
**Clinical Implications:** Gut dysbiosis occurs early in CKD and progresses with disease severity, suggesting that microbiota analysis could serve as a non-invasive diagnostic tool. Specific microbial signatures may identify high-risk ESKD patients beyond traditional risk factors. The consistent depletion of butyrate-producing bacteria like Roseburia and enrichment of pro-inflammatory Proteobacteria highlight potential therapeutic targets. While synbiotic and probiotic interventions show promise for restoring beneficial bacteria, the observed eGFR decline with synbiotics in one study warrants caution. The association between gut dysbiosis and peritonitis in peritoneal dialysis patients suggests microbiome monitoring could help prevent infectious complications. Large randomized clinical trials are needed to establish causal relationships and determine whether microbiota modulation improves hard clinical outcomes including mortality and cardiovascular events.