Gut Microbiota in Chronic Kidney Disease: From Composition to Modulation towards Better Outcomes—A Systematic Review

Background: A bidirectional kidney–gut axis was described in patients with chronic kidney disease (CKD). On the one hand, gut dysbiosis could promote CKD progression, but on the other hand, studies reported specific gut microbiota alterations linked to CKD. Therefore, we aimed to systematically review the literature on gut microbiota composition in CKD patients, including those with advanced CKD stages and end-stage kidney disease (ESKD), possibilities to shift gut microbiota, and its impact on clinical outcomes. Materials and methods: We performed a literature search in MEDLINE, Embase, Scopus, and Cochrane databases to find eligible studies using pre-specified keywords. Additionally, key inclusion and exclusion criteria were pre-defined to guide the eligibility assessment. Results: We retrieved 69 eligible studies which met all inclusion criteria and were analyzed in the present systematic review. Microbiota diversity was decreased in CKD patients as compared to healthy individuals. Ruminococcus and Roseburia had good power to discriminate between CKD patients and healthy controls (AUC = 0.771 and AUC = 0.803, respectively). Roseburia abundance was consistently decreased in CKD patients, especially in those with ESKD (p < 0.001). A model based on 25 microbiota dissimilarities had an excellent predictive power for diabetic nephropathy (AUC = 0.972). Several microbiota patterns were observed in deceased ESKD patients as compared to the survivor group (increased Lactobacillus, Yersinia, and decreased Bacteroides and Phascolarctobacterium levels). Additionally, gut dysbiosis was associated with peritonitis and enhanced inflammatory activity. In addition, some studies documented a beneficial effect on gut flora composition attributed to synbiotic and probiotic therapies. Large randomized clinical trials are required to investigate the impact of different microbiota modulation strategies on gut microflora composition and subsequent clinical outcomes. Conclusions: Patients with CKD had an altered gut microbiome profile, even at early disease stages. Different abundance at genera and species levels could be used in clinical models to discriminate between healthy individuals and patients with CKD. ESKD patients with an increased mortality risk could be identified through gut microbiota analysis. Modulation therapy studies are warranted.


Introduction
Gut microbiota represents one of the most diverse microbiota of the human body and encompasses more than 35,000 bacterial species with 10 million genes [1]. For that reason, disease", "end-stage kidney disease", "hemodialysis", "renal decline", "disease severity", "progression", "mortality", "major adverse cardiovascular events", "inflammation", "inflammatory response", "probiotic", "prebiotic", and "symbiotic". Studies were published from the inception of databases to 30 June 2022. We did not apply any language filters or restrictions in the search strategy. In addition, the ClinicalTrials.gov registry of clinical trials, Google Scholar engine, and references from cited manuscripts were screened and checked for additional eligible studies, which is compliant with PRISMA guidelines. The final search strategy for all databases and references obtained is presented in Table S1.

Eligibility Criteria and Outcomes
A multistep approach was used to assess retrieved references for eligibility. In the first step, two independent investigators evaluated the title and abstracts of articles for inclusion and exclusion criteria. In the next step, the full text of studies that met the eligibility criteria based on title and abstract was appraised.
Several inclusion criteria were pre-defined and were applied for eligibility assessment. (1) Both randomized clinical trials and observational studies were considered for inclusion.
(2) Patients ≥ 18 years old with CKD of all stages were enrolled. (3) Healthy subjects or patients with early stages of CKD were included in the control group (when available). (4) Studies reporting original data on the following outcomes: (a) gut microbiota composition in CKD patients (when available, CKD patients versus healthy subjects or early CKD patients versus ESKD patients); (b) association between identified microbiota species in CKD patients and major adverse cardiovascular events (MACE), mortality, CKD severity, and disease progression; (c) the impact of prebiotics, probiotics, and symbiotics on flora composition and outcomes of CKD patients.
Early CKD patients were considered those presenting with G2 and G3 CKD stages (estimated glomerular filtration rate of 30-90 mL/min/1.73 m 2 ). Prebiotics were defined as "non-digestible food ingredient" which enhance the growth or activity of certain beneficial gut bacteria [16]. Probiotics were defined as "live microorganisms" which exhibit beneficial health effects when they are prescribed in appropriate concentration [17]. Likewise, we defined synbiotics as a "mixture of probiotics and prebiotics" which display beneficial health effects [18].
In addition, some key exclusion criteria were established: unpublished data, studies available only in abstract, overlapping populations, case reports, meta-analyses, editorials, missing data, and inability to extract data regarding the population enrolled and outcomes investigated.

Data Collection and Synthesis
After eligibility assessment, two independent investigators extracted the following data from included studies: first author, year of publication, study design, number of patients enrolled and their age, clinical setting, reported outcome of interest, and followup period. We performed a qualitative synthesis of included studies to provide a better understanding of reported outcomes. Additionally, when available, data were reported as numbers, median and mean values, odds ratio (OR), and p-value.

Quality and Risk of Bias Assessment
In the case of randomized controlled clinical trials, the risk of bias was judged using the revised Cochrane risk-of-bias tool for randomized trials (RoB 2) [19]. The Newcastle-Ottawa scale (NOS) was used to guide the quality assessment of observational, nonrandomized studies. NOS is a tool based on designating stars for signaling questions, which were grouped into three domains: population sampling, comparability of groups, and evaluation of outcomes of interest [20].

Results
We searched the specified databases and retrieved 3290 references. Duplicate records were removed (n = 1927), leaving 1363 references for title or abstract screening. Finally, 121 records were assessed for eligibility in full-text, and 69 studies were included in the present systematic review. The flowchart of the selection process is presented in Figure 1.

Quality and Risk of Bias Assessment
In the case of randomized controlled clinical trials, the risk of bias was judged using the revised Cochrane risk-of-bias tool for randomized trials (RoB 2) [19]. The Newcastle-Ottawa scale (NOS) was used to guide the quality assessment of observational, non-randomized studies. NOS is a tool based on designating stars for signaling questions, which were grouped into three domains: population sampling, comparability of groups, and evaluation of outcomes of interest [20].

Early-Stage CKD
Hu et al. observed different gut microbiota compositions even in patients with earlystage CKD as compared to healthy controls [74]. Gut flora diversity was significantly decreased in early-stage CKD patients (p < 0.001) compared to those without CKD. At the genera level, Ruminococcus had a good power to discriminate between early-stage CKD patients and healthy controls (AUC = 0.771, 95% CI, 0.771-0.852), while Roseburia accurately identified healthy controls (AUC = 0.803, 95% CI, 0.804-0.864) [74]. Wu et al. reported similar discriminatory capacity for early-stage CKD patients in the case of Bacteroides eggerthii (AUC = 0.80, 95% CI, 0.67-0.93), which was higher than in the case of protein/creatinine ratio (AUC = 0.64) [56]. Studies investigating gut microbiota diversity are of particular importance, as decreased diversity could be considered a reliable marker of gut dysbiosis [90]. Consequently, even patients with early-stage CKD had gut dysbiosis (low microbiota diversity) when compared to healthy controls [74]. Thus, improving gut microbiota diversity constitutes an important target for future interventional studies involving CKD patients. Available data on Roseburia suggest a key role in the modulation of gut barrier homeostasis and inflammation [91]. Hence, Roseburia is a marker of a normal gut microbiome, and lower abundance was reported in CKD patients from early stages [74]. Moreover, Roseburia abundance decreases along with CKD progression (lower in ESKD patients compared to early-stage CKD) [53].

ESKD
Some studies reported a shift in gut microbiota profile along with CKD progression, especially in ESKD patients and those requiring renal replacement therapy (RRT). Gao et al. reported a progressively increasing abundance of Bifidobacterium in advanced CKD stages, while Lactobacillus levels decreased [24]. He et al. observed a lower abundance of both Bifidobacterium and Lactobacillus in ESKD patients compared to healthy controls.

Diabetic Nephropathy
Patients with diabetic nephropathy stage 3-4 had a different microbiota profile compared to healthy individuals [23]. Du et al. developed a model based on 25 gut microbiota dissimilarities, which had an excellent predictive power for diabetic nephropathy (AUC = 0.972) [23]. Additionally, in patients with diabetes-associated CKD, Ruminococcaceae and Bacteroidaceae abundance was significantly increased, while Prevotellaceae levels were decreased. These microbiome alterations were observed across all CKD stages, highlighting early gut dysbiosis in diabetic patients, which maintains in advanced CKD stages [38]. In addition, hemodialysis patients with diabetes had an increased abundance of Desulfovibrionaceae, Veillonellaceae, and Lactobacillaceae as compared to hemodialysis patients without diabetes [83].
Outcomes of CKD patients (inflammation, renal function, disease progression, mortality, and peritonitis) linked to gut microbiota composition are displayed in Table 2. The Simpson index and the Shannon index were significantly lower in deceased hemodialysis patients, as compared to those who survived (respectively, p = 0.007 and p = 0.028). Moreover, several bacteria were increased in hemodialysis patients from the non-survivor group (Oscillospira, Achromobacter, Agrobacterium, Lactobacillus, Alloscardovia, Anoxybacillus, Devosia, Yersinia) [40]. Additionally, Luo et al. reported a lower abundance of Bacteroides and Phascolarctobacterium in ESKD patients with cardiovascular mortality compared to those who survived (p < 0.05 for both). Table 2. Outcomes in CKD patients related to gut microbiota. Inflammation In CKD patients, Roseburia spp. and F. prausnitzii were negatively correlated with CRP (respectively, (r = −0.493, p = 0.00; r = −0.528, p = 0.000).

Lin et al., 2020 [41] Inflammation
The Shannon diversity index was negatively corelated with IL-6 (r = −0.253, p = 0.015) and TNFα (r = −0.260, p = 0.011). Faecalibacterium prausnitzii was negatively correlated with TNFα (r = −0.204, p = 0.047).  Peritonitis was associated with altered gut microbiota composition in peritoneal dialysis patients, as was reported by some authors. Zhou et al. observed a higher abundance of Bacteroidetes and Synergistetes in the peritonitis group as compared to patients without peritonitis, while Bacilli and Lactobacillus were decreased [70]. Additionally, Dorea and Clostridium abundance was decreased in peritoneal dialysis patients [45].
Moreover, available gut microbiota modulation interventions (including dietary interventions, probiotics, prebiotics, and synbiotics) are presented in Table 3. Concerning intestinal flora modulation, synbiotics increased Bifidobacterium levels up to 5-fold from baseline (p = 0.003) [87], which was concordant in clinical studies [84,85]. Lactobacillus levels were decreased following synbiotic therapy in one study [85], whereas in another study, Lactobacillus abundance was similar before and after the treatment [87]. Nevertheless, synbiotic therapy was linked to eGFR decrease with 3.14 mL/min/1.73 m 2 (p < 0.01), requiring further research [84]. In 56% of patients in the treatment group, gut microbiota recovered as compared to placebo (grade III dysbiosis was absent after therapy). CRP decreased from 6.8 ± 3.1 g/L to 5.3 g/L in the treatment group.

Borges et al., 2017 [79] Probiotics
Gut microbiota profile was similar in the probiotic group (Streptococcus thermophilus, Lactobacillus acidophilus, and Bifidobacteria longum strains) and placebo group after 3 months of therapy (similar number of bands).

Ebrahim et al., 2022 [80]
β-glucan prebiotic Prevotella tended to increase in the intervention group (β-glucan) as compared to the control group, while Bacteroides and Blautia tended to decrease.
Hu et al., 2022 [66] Dietary intervention HD patients from the protein-energy wasting group had lower abundance of Roseburia as compared to HD patients in the non-protein energy wasting group (p = 0.022).
Escherichia abundance was increased in PD patients from the protein-energy wasting group compared to PD patients from the non-protein-energy wasting group (p = 0.022).
Low-protein diet + inulin Low-protein diet associated with inulin therapy increased Bifidobacteriaceae levels.

Inulin
Inulin decreased Enterobacteriaceae family. PD patients who received a one-month advanced glycation end-product restriction had a lower abundance of Prevotella copri compared to those with a normal diet. CKD = chronic kidney disease; CRP = C-reactive protein; eGFR = estimated glomerular filtration rate; HD = hemodialysis; PD = peritoneal dialysis.
The quality of non-randomized studies was fair to good as assessed by NOS adapted for cross-sectional, case-control, and cohort studies (Tables S2-S4). The risk of bias in randomized trials was appraised using the RoB 2 tool and is displayed in Figure 2.

Discussion
This study assessed gut microbiota alteration across a spectrum of kidney disease through a systematic review. The main findings are: (1) a different microbiota composition and a decreased gut diversity were reported from early stages to advanced CKD; (2) patients with CKD shared the depletion of anti-inflammatory butyrate-producing microbes (i.e., Roseburia, Prevotella, Bacteroides) and the enrichment of pro-inflammatory microbes (Proteobacteria, Actinobacteria); (3) there are limited data regarding the impact of dysbiosis on inflammation, mortality, or cardiovascular risk.
This review was designed to comprehensively illustrate the alteration of the gut microbiome in CKD patients. In more than half of the analyzed studies, gut diversity was significantly decreased. Similar data were reported last year by Zhao et al. [93]. Six (6/9) and four studies (4/6) that included patients with CKD and ESKD, respectively, suggested that the α-diversity of gut microbiota was significantly lower in patients than in healthy controls. Additionally, ten (10/11) studies that analyzed patients with CKD and five (5/6) studies that focused on ESKD reported a significantly altered composition of the gut microbiota in patients as compared to healthy controls [93].
At the phylum level, Fusobacteria, Verrucomicrobia, and Proteobacteria abundances were significantly higher in CKD [39,61]. At the genus and species levels, there were some substantial differences; probably, the individual variances determined by genetic and environmental factors and the etiology and the severity of CKD could explain the variation and inconsistency of the studies.
Proteobacteria, including common bacteria such as Escherichia coli, Salmonella, or Desulfovibrio, are increased in CKD [49,83]. An increased level was associated with abnormal gut barrier function, which might result in increased epithelial permeability, allowing microbial fragments and products to enter the sub-epithelial space and lamina propria, lipopolysaccharides translocation, and enhancing inflammatory response [94]. Additionally, people with an increased level of Desulfovibrio spp. had more severe renal dysfunction [94].
Concerning phylum Firmicutes, we noticed an increase in the genus Streptococcus and a lower abundance of the genus Roseburia, Faecalibacterium, and Prevotella. Roseburia produces butyrate, which could promote the proliferation of extrathymic regulatory T cells

Discussion
This study assessed gut microbiota alteration across a spectrum of kidney disease through a systematic review. The main findings are: (1) a different microbiota composition and a decreased gut diversity were reported from early stages to advanced CKD; (2) patients with CKD shared the depletion of anti-inflammatory butyrate-producing microbes (i.e., Roseburia, Prevotella, Bacteroides) and the enrichment of pro-inflammatory microbes (Proteobacteria, Actinobacteria); (3) there are limited data regarding the impact of dysbiosis on inflammation, mortality, or cardiovascular risk.
This review was designed to comprehensively illustrate the alteration of the gut microbiome in CKD patients. In more than half of the analyzed studies, gut diversity was significantly decreased. Similar data were reported last year by Zhao et al. [93]. Six (6/9) and four studies (4/6) that included patients with CKD and ESKD, respectively, suggested that the α-diversity of gut microbiota was significantly lower in patients than in healthy controls. Additionally, ten (10/11) studies that analyzed patients with CKD and five (5/6) studies that focused on ESKD reported a significantly altered composition of the gut microbiota in patients as compared to healthy controls [93].
At the phylum level, Fusobacteria, Verrucomicrobia, and Proteobacteria abundances were significantly higher in CKD [39,61]. At the genus and species levels, there were some substantial differences; probably, the individual variances determined by genetic and environmental factors and the etiology and the severity of CKD could explain the variation and inconsistency of the studies.
Proteobacteria, including common bacteria such as Escherichia coli, Salmonella, or Desulfovibrio, are increased in CKD [49,83]. An increased level was associated with abnormal gut barrier function, which might result in increased epithelial permeability, allowing microbial fragments and products to enter the sub-epithelial space and lamina propria, lipopolysaccharides translocation, and enhancing inflammatory response [94]. Additionally, people with an increased level of Desulfovibrio spp. had more severe renal dysfunction [94].
Concerning phylum Firmicutes, we noticed an increase in the genus Streptococcus and a lower abundance of the genus Roseburia, Faecalibacterium, and Prevotella. Roseburia produces butyrate, which could promote the proliferation of extrathymic regulatory T cells (Tregs). Tregs, as vital anti-inflammatory lymphocytes, produce interleukin-10, transforming growth factor beta, and interferon gamma. Microbial butyrate has been established to contribute to the pro-and anti-inflammatory balance by inducing Tregs differentiation [95]. Roseburia was negatively related to inflammatory status, renal function, and CKD progression [32,33].
From the phylum Verrucomicrobia, the genus Akkermansia plays an essential role in improving gut-barrier function and viscosity of the mucus. It also promotes the growth of bacteria-producing SCFAs, such as butyrate, by offering them carbon, nitrogen, and energy produced as a consequence of mucus degradation [96]. It also has anti-inflammatory properties. Unfortunately, the abundance was decreased in CKD patients [41,83] and was negatively related to inflammation [46].
Regarding the Bacteroidetes phylum, we observed an increase in the genus Bacteroides and a lower abundance of the genus Prevotella. In the general population, it was noticed that the Western diet was associated with Bacteroides and Clostridiales abundance in the gut microbiome, while rural populations with a high-fiber, low-protein diet tended to have Prevotella, which can produce SCFA. The authors from one study evaluated fecal microbiota composition differences between ESKD patients and 60 healthy controls. They found that Prevotella was enriched in the healthy group, whereas Bacteroides was prevalent in the ESKD group. Moreover, Prevotella was negatively related to inflammatory status and renal function [33]. Bacteroides was increased in most studies, whereas only two studies reported decreased levels [48,53]. Further analysis confirmed that Bacteroides was related to cardiovascular mortality in patients with dialysis. Nevertheless, the authors reported only five cardiovascular-related deaths, thus limiting the extrapolation of the results to all CKD patients. Consequently, these data should be confirmed in large clinical trials [45]. In addition, patients with peritoneal dialysis and peritonitis had a higher abundance of Bacteroides compared with the non-peritonitis group [70].
Our data are comparable to those obtained by Zhao et al.; the same abundance of Proteobacteria in CKD and ESKD was identified [93]. However, we identified a greater abundance of Bacteroides in CKD patients as compared to previous data. Zhao et al. reported enriched levels of bacteria from the Bacteroides genus in 3/11 studies involving CKD patients and in 3/9 studies on ESKD patients [93]. Furthermore, it was connected with infections and cardiovascular mortality [93].
Gut microbial dysbiosis has also been reported in diabetes mellitus. Patients with diabetic nephropathy exhibited increased levels of multiple pathogenic genera such as Acidaminococcus, Lactobacillus, Megasphaera, Clostridium, Sutterella, and Desulfovibrionaceae, while healthy controls had a high abundance of butyrate-producing bacteria [23]. The data are similar to those from a recently published systematic review [97]. Nevertheless, additional studies are warranted to investigate which specific microbes are involved in the pathophysiology of CKD linked to diabetes mellitus.
Data concerning phosphate binders on the gut microbiome in CKD patients are limited. Two studies documented a lower gut flora diversity in HD patients receiving phosphate binder [63,65]. In another study, sucroferric oxyhydroxide supplementation increased levels of Ruminococcus torques, which could influence gut barrier permeability [75,98]. Likewise, after oral iron supplementation, α-diversity and Firmicutes levels decreased, while Bacteroides increased [76]. These findings warrant further safety analysis of phosphate binders and iron supplementation in terms of gut homeostasis and microbiome composition in CKD patients.
Evidence sustaining the benefit of probiotic, prebiotic, and synbiotic supplementation in the management of CKD is mixed. Some studies insinuated that they might be useful by decreasing uremic and inflammatory toxins [99]. They could also improve oxidative stress, as well as lipid profiles in patients with CKD, which are well-known cardiovascular risk factors [100].

Conclusions
In conclusion, patients with CKD had an altered gut microbiome profile, even at early disease stages, as was documented consistently in clinical studies. Moreover, studies reported a shift in gut microbiota composition along with CKD progression, especially in ESKD patients and those requiring RRT. Different abundance at genera and species levels could be used in clinical models to discriminate between healthy individuals and patients with CKD (including those with diabetic nephropathy), with excellent predictive power. In addition to traditional risk factors, ESKD patients with an increased mortality risk could also be identified through gut microbiota analysis. In addition, studies established gut microbiome patterns linked to enhanced inflammatory activity and to a higher risk of peritonitis in patients receiving peritoneal dialysis. Nevertheless, clinical studies with larger sample sizes are required to confirm the association between altered gut microbiota composition and adverse outcomes in CKD patients, including all-cause and cardiovascular mortality. Due to data inconsistency, randomized clinical trials are needed to analyze the effect of different microbiota modulation therapies on gut bacterial composition and adverse end-points in CKD patients.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.