Studies were included if they met the following criteria: (i) Participants: adults (≥18 years) diagnosed with DKD based on established clinical or laboratory criteria, including persistent albuminuria or proteinuria (≥30 mg/24 h or urine albumin-to-creatinine ratio ≥30 mg/g) and/or reduced eGFR (<60 mL/min/1.73 m²) (de Boer et al., 2022); (ii) Intervention: probiotics combination with standard treatment; (iii) Comparison: standard treatment alone. (iv) Outcomes: at least one of the following was reported—renal function indicators [SCR, 24-h urine protein (24 h UP), eGFR, BUN], blood glucose levels [FBG, 2-h postprandial blood glucose (2 h PBG), HbA1c, insulin], blood lipid levels [TG, total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C)], and inflammatory and oxidative markers [high-sensitivity C-reactive protein (hs-CRP), total antioxidant capacity (TAC), malondialdehyde (MDA)]; (v) Study design: randomized controlled trials (RCTs).

Exclusion criteria were: (i) Duplicate publications; (ii) Studies with insufficient data to extract or analyze relevant outcomes; (iii) Participants diagnosed with uremia or ESRD.

A comprehensive search was performed in PubMed, Embase, Web of Science, the Cochrane Library, and Chinese Science Citation Database (CSCD) from their inception to October 1, 2025. The search fields are Topic or Title/Abstract, and the search query is as follows: (Probiotic OR Probiotics OR Synbiotic OR Synbiotics OR Bifidobacterium OR Bifidobacteria OR Bacillus bifida OR Yeast OR Saccharomyces cerevisiae OR Saccharomyces italicus OR Saccharomyces oviformis OR S. cerevisiae OR S. cerevisiae OR Saccharomyces uvarum var. melibiosus OR Candida robusta OR Saccharomyces capensis OR Lactobacillus acidophilus OR Lactobacillus amylovorus OR Lactobacill OR Lactic acid bacteria OR Clostridium butyricum OR Bacillus OR Natto Bacteria OR Streptococcus thermophiles OR Enterococcus) AND (Diabetic nephropathy OR Diabetic nphropathies OR Diabetic Kidney Disease OR Diabetic Kidney Diseases OR Diabetic Glomerulosclerosis OR Intracapillary Glomerulosclerosis OR Kimmelstiel Wilson Disease OR Nodular Glomerulosclerosis OR Kimmelstiel Wilson Syndrome). No restrictions were applied to language or publication status. Reference lists of relevant reviews and included studies were manually screened to identify additional eligible articles.

A systematic approach combining a reference management tool and manual evaluation was employed to screen the literature (Yu et al., 2023). All retrieved records were imported into EndNote for deduplication. Two reviewers independently screened the titles and abstracts for potential eligibility, followed by full-text assessment of the remaining articles. Discrepancies were resolved by discussion or consultation with a third reviewer. The study selection process was documented in a PRISMA flow diagram.

Two reviewers independently extracted data using a pre-designed data collection form (Yu et al., 2024). Extracted information included: (i) basic study characteristics (first author, year of publication, country, study design); (ii) participant characteristics (sample size, age, sex distribution, weight, disease duration); (iii) intervention details (strain of probiotics, dosage, frequency, duration); (iv) funding sources and potential conflicts of interest. Any disagreements were resolved by discussion or adjudication by a third reviewer.

The risk of bias of included RCTs was independently evaluated by two reviewers using the Cochrane Collaboration’s Risk of Bias tool (RoB 1.0) (Higgins et al., 2011). The following seven domains were assessed: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other sources of bias. Each domain was judged as “low risk,” “high risk,” or “unclear risk” of bias. Disagreements between reviewers were resolved by discussion or, if necessary, consultation with a third reviewer.

Meta-analysis was conducted using RevMan (version 5.3) and Stata (version 16.0). Continuous outcomes were expressed as mean differences (MDs) or standardized mean differences (SMDs) with 95% confidence intervals (CIs). Dichotomous outcomes were presented as risk ratios (RRs) with 95% CIs (Li et al., 2025). Statistical heterogeneity among studies was assessed using the I² statistic. An I² > 50% indicated substantial heterogeneity, in which case a random-effects model was applied; otherwise, a fixed-effects model was used (Migliavaca et al., 2022). p < 0.05 was considered statistically significant.

For outcomes with I² > 50%, leave-one-out sensitivity analysis was performed, whereby each study was sequentially omitted to investigate potential sources of heterogeneity and assess the robustness of the pooled results. The stability of the results was judged by observing whether the direction and magnitude of the overall effect size changed materially after the exclusion of any single study. If the pooled effect estimates remained consistent, the results were considered robust (Patsopoulos et al., 2008). Additionally, when the number of included studies for outcomes with high heterogeneity was five or more, subgroup analyses were performed based on factors such as gender, mean age, body weight, HbA1c, SCR, disease duration, probiotic preparations, and treatment duration. These subgroup analyses aimed to investigate the heterogeneity originating from clinical aspects and assess the robustness of the results (Richardson et al., 2019).

Publication bias was assessed using funnel plot symmetry when more than 10 studies were included in a meta-analysis (Jin et al., 2015). When fewer than 10 studies were available, funnel plot inspection was not performed due to limited precision; instead, Egger’s regression test was applied to quantitatively evaluate potential publication bias (Lin and Chu, 2018). A p-value greater than 0.05 in Egger’s test was considered to indicate no potential publication bias.

The certainty of evidence for each outcome was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach (Prasad, 2024). The certainty of evidence was classified into four levels: high, moderate, low, or very low, based on risk of bias, inconsistency, indirectness, imprecision, and publication bias.

A total of 560 records were identified through the electronic databases, including PubMed (n = 109), Embase (n = 158), Cochrane Library (n = 119), Web of Science (n = 151), and CSCD (n = 18). An additional 5 studies were identified from other sources. After removing 206 duplicates, 354 records remained. Screening of titles and abstracts excluded 332 articles, and 22 full-text articles were assessed for eligibility. Among these, 15 articles were excluded (2 for reporting duplicate data and 13 for not meeting the intervention criteria). Finally, 7 RCTs (Jiang et al., 2021; Mafi et al., 2018; Abbasi et al., 2018; Mazruei Arani et al., 2019; Firouzi et al., 2015; Miraghajani et al., 2017; Tang et al., 2020) were included in the meta-analysis. The study selection process is summarized in the PRISMA flow diagram (Figure 2).

A total of 7 randomized controlled trials involving 502 participants were included (Jiang et al., 2021; Mafi et al., 2018; Abbasi et al., 2018; Mazruei Arani et al., 2019; Firouzi et al., 2015; Miraghajani et al., 2017; Tang et al., 2020). Among them, 255 participants received probiotics in addition to standard treatment, while 247 participants received standard treatment alone. Overall, the mean proportion of female participants was 52.0%, with a mean age of 56.4 years and an average body weight of 74.1 kg. Baseline HbA1c levels ranged from 6.8 to 7.9 mmol/L, SCR from 0.8 to 3.3 mg/dL, and disease duration from 6.9 to 9.3 years. Three studies administered single-strain probiotics, whereas four studies used multi-strain probiotic formulations. The intervention duration ranged between 8 and 12 months. Detailed characteristics of the included studies are presented in Table 1.

According to the RoB 1.0 assessment (Figure 3), allocation concealment was rated as having an unclear risk in the studies conducted by Mazruei Arani et al. (2019) and Tang et al. (2020); blinding of participants and personnel was rated as having an unclear risk in the studies conducted by Mazruei Arani et al. (2019), Miraghajani et al. (2017), and Tang et al. (2020); incomplete outcome data were judged to have a high risk in the study conducted by Jiang et al. (2021); selective reporting was rated as having an unclear risk in the study conducted by Firouzi et al. (2015); and other potential sources of bias were rated as having an unclear risk in the studies conducted by Abbasi et al. (2018) and Miraghajani et al. (2017). All other domains were assessed as low risk.

High heterogeneity was observed in eGFR, HbA1c, HDL-C, and MDA, and therefore sensitivity analyses were conducted to explore the potential sources of heterogeneity (Table 2). The results indicated that the heterogeneity in HbA1c was attributable to the study by Mafi et al. (2018), which may be explained by differences in participant ethnicity. After excluding this study, the heterogeneity of HbA1c was substantially reduced and the statistical significance of the effect changed (MD −0.45, 95% CI −0.75 to −0.15, p = 0.003, I² = 0%), suggesting that the HbA1c result was not robust. The heterogeneity in HDL-C was mainly driven by the study of Tang et al. (2020), possibly due to the lack of participant blinding. After excluding this study, the heterogeneity of HDL-C was markedly reduced while the statistical significance remained unchanged (MD 1.48, 95% CI −0.69 to 3.64, p = 0.18, I² = 0%), indicating that the HDL-C result was robust. The heterogeneity in MDA was attributable to the study of Miraghajani et al. (2017), which may be explained by the relatively short intervention duration. After excluding this study, the heterogeneity of MDA was significantly reduced while the statistical significance was preserved (MD −0.68, 95% CI −0.86 to −0.51, p < 0.00001, I² = 47%), indicating that the MDA result was robust. Although the sensitivity analysis did not identify the source of heterogeneity for eGFR, it confirmed the robustness of the pooled result. However, since fewer than five studies were included for each of these outcomes, subgroup analyses were not performed.

Egger’s test indicated no significant publication bias for SCR (p = 0.589), BUN (p = 0.944), FBG (p = 0.937), HbA1c (p = 0.054), TG (p = 0.352), TC (p = 0.340), LDL-C (p = 0.858), HDL-C (p = 0.424), hs-CRP, (p = 0.396) TAC (p = 0.263), and MDA (p = 0.083), while potential publication bias was detected for eGFR (p = 0.034) (Figure 8). Since only two studies were included for 24 h UP, 2 h PBG, and insulin, Egger’s test was not performed for these outcomes.

According to the GRADE approach, the certainty of evidence was rated as moderate for SCR, BUN, FBG, TG, TC, LDL-C, hs-CRP, and TAC; low for 24 h UP, 2 h PBG, HbA1c, insulin, HDL-C, and MDA; and very low for eGFR (Table 3). The strength of recommendation was strong.

DKD is the most severe microvascular complication of diabetes and one of the leading causes of death among diabetic patients (Thomas et al., 2015). With the discovery of the “gut–kidney axis” (Jiang et al., 2025), probiotics have been regarded as a promising complementary therapy for DKD (Zhang et al., 2025; Wang et al., 2025). However, previous clinical trials and meta-analyses have reported inconsistent findings regarding renal and glycemic outcomes, highlighting the controversy in this field. Tarrahi et al. (2022) reported that probiotics significantly reduced FBG and SCR levels in patients with DKD but had no significant effects on HbA1c or BUN. In contrast, Dai et al. (2022) found that probiotics improved multiple renal and glycemic parameters, including reductions in FBG, HbA1c, SCR, and BUN. Notably, both Tarrahi et al. (2022) and Dai et al. (2022) included a clinical trial conducted in uremic patients, which may have introduced additional confounding factors. Furthermore, Tarrahi et al. (2022) only included studies published up to 2019, resulting in temporal limitations.

Therefore, we conducted an updated meta-analysis to more comprehensively and accurately evaluate the effects of probiotics in patients with DKD. Our results demonstrated that probiotic supplementation significantly reduced SCR, BUN, FBG, TG, TC, LDL-C, hs-CRP, and MDA levels in patients with DKD. Notably, no publication bias was detected for these positive outcomes, and except for MDA, the certainty of evidence was rated as moderate, suggesting that these findings are robust and reliable.

Renal function is a key indicator for assessing the progression and prognosis of DKD. The results of this meta-analysis showed that, compared with placebo, probiotic supplementation significantly reduced SCR and BUN levels in patients with DKD, while no significant effects were observed on 24 h UP or eGFR. In previous meta-analyses, Dai et al. (2022) found that probiotics decreased SCR (MD 0.17, 95% CI 0.29 to 0.05) and BUN levels (MD 1.36, 95% CI 2.20 to 0.52) in DKD patients, supporting our findings. However, Tarrahi et al. (2022) reported that probiotics only reduced SCR levels (MD −0.18, 95% CI −0.26 to −0.09), with no significant effect on BUN. AbdelQadir et al. (2020) found no significant effect of probiotics on either SCR or eGFR, presenting conflicting results. These discrepancies may be attributed to differences in participant characteristics and sample sizes. For one, both Tarrahi et al. (2022) and AbdelQadir et al. (2020) included the study by Soleimani et al. (2017), which recruited uremic DKD patients, potentially introducing additional confounding factors that affected the pooled results. In contrast, we excluded the study by Soleimani et al. (2017), thereby reducing the confounding effects of severe renal impairment and dialysis-related metabolic disturbances. For another, Tarrahi et al. (2022) and AbdelQadir et al. (2020) included only four studies when analyzing SCR, which may have limited the statistical power of their analyses. In comparison, our meta-analysis included five RCTs involving 386 participants, enhancing the robustness of the findings. Collectively, these results suggest that probiotics have the potential to reduce SCR and BUN levels in patients with DKD, though their effects on 24 h UP and eGFR remain limited.

Abnormal glucose metabolism is a central driver of the onset and progression of DKD (Stanton, 2021). Our meta-analysis demonstrated that probiotics significantly reduced FBG levels, while no significant effects were observed on 2 h PBG, HbA1c, or insulin levels. Sensitivity analysis revealed that the results for HbA1c were unstable, with heterogeneity largely influenced by the study of Mafi et al. (2018). After excluding this study, HbA1c showed a significant between-group difference (MD −0.45, 95% CI −0.75 to −0.15). We speculate that this variation may be associated with ethnic differences. Among the three included studies reporting HbA1c, only Mafi et al. (2018) recruited Iranian participants, whereas the other two studies (Jiang et al., 2021; Tang et al., 2020) were conducted in Chinese populations. In previous meta-analyses, Tarrahi et al. (2022) found that probiotics significantly reduced FPG levels (MD −19.08, 95% CI −32.16 to −5.99) but had no significant effects on HbA1c or insulin. Similarly, AbdelQadir et al. (2020) also reported no significant impact of probiotics on serum insulin levels (MD −1.99, 95% CI −3.99 to 0.01), supporting our findings. Interestingly, Dai et al. (2022) observed that probiotics not only reduced FBG levels but also significantly lowered HbA1c levels in DKD patients. The inconsistent results regarding HbA1c may stem from additional confounding factors. Unlike Dai et al. (2022), our analysis excluded the study by Soleimani et al. (2017) when pooling HbA1c data, which altered the statistical significance. Given that Soleimani et al. (2017) focused on uremic DKD patients, including this study may have introduced confounding bias and led to a false-positive effect. Overall, current evidence suggests that probiotics effectively reduce FBG levels in patients with DKD, while their effects on 2 h PBG, HbA1c, and insulin levels warrant further investigation.

Lipid metabolism disorders are key pathogenic mechanisms in DKD (Han et al., 2024). Our meta-analysis demonstrated that probiotics significantly reduced TG, TC, and LDL-C levels in DKD patients, while showing no significant effect on HDL-C levels. Sensitivity analysis revealed that the heterogeneity of HDL-C was mainly influenced by the study of Tang et al. (2020). After excluding this study, heterogeneity markedly decreased without altering statistical significance (MD 1.48, 95% CI −0.69 to 3.64), suggesting that the result was robust. However, previous meta-analyses have reported conflicting findings. Dai et al. (2022) found that probiotics significantly decreased TC (MD −6.93, 95% CI −11.67 to −2.19) and LDL-C levels (MD −7.14, 95% CI −11.03 to −3.24) and increased HDL-C levels (MD 2.72, 95% CI 0.47 to 4.97) in DKD patients. In contrast, AbdelQadir et al. (2020) reported that probiotics only reduced LDL-C (MD −6.60, 95% CI −13.11 to −0.10), without significant improvements in TC, TG, or HDL-C. Notably, compared with Dai et al. (2022), our exclusion of the study by Soleimani et al. (2017) yielded a nonsignificant result for HDL-C, suggesting that the previously reported benefit in HDL-C might have been confounded and unstable. Additionally, compared with AbdelQadir et al. (2020), our inclusion of additional clinical trials by Abbasi et al. (2018) and Tang et al. (2020) clarified the beneficial effects of probiotics on TG and TC levels, indicating that the previously negative results may have been due to limited sample sizes. Collectively, the current evidence supports the potential of probiotics to reduce TG, TC, and LDL-C levels in DKD patients, while their effects on HDL-C require further investigation.

The mechanisms underlying the improvement of renal function and glucose–lipid metabolism by probiotics in DKD patients are mainly associated with modulation of gut microbiota composition and its metabolites. Dysbiosis in DKD is characterized by a reduction in beneficial short-chain fatty acid (SCFA)-producing bacteria (Sabatino et al., 2017). Compared with healthy individuals, DKD patients exhibit increased abundance of Proteobacteria and Fusobacteria at the phylum level, and Escherichia-Shigella, Desulfovibrio, and Streptococcus at the genus level, whereas the abundance of Roseburia, Faecalibacterium, Pyramidobacter, Prevotellaceae_UCG-001, and Prevotella_9 is reduced (Zhao et al., 2021). Roseburia spp., Faecalibacterium prausnitzii, and Prevotella are known SCFA producers (Jiang et al., 2017). SCFAs can stimulate glucagon-like peptide-1 secretion and protect against chronic hyperglycemia induced by renal oxidative stress (Everard and Cani, 2014). Further evidence suggests that gut microbiota-derived acetate can modulate glucose metabolism, reduce glycotoxin accumulation, and improve mitochondrial function, thereby alleviating glomerular injury and renal fibrosis (Luo et al., 2022). Additionally, metabolomic studies have revealed that plasma bile acid levels in DKD patients increase progressively and are closely correlated with disease progression (Zhang et al., 2024). Gut microbiota such as Lactobacillus, Clostridium, Bifidobacterium, and Enterococcus can regulate bile acid metabolism and cholesterol secretion, thereby attenuating renal inflammation (Tian et al., 2020).

In addition to disturbances in glucose–lipid metabolism, inflammation and oxidative stress are key drivers of renal injury in DKD (Su et al., 2023; Rana et al., 2022). Our meta-analysis revealed that while probiotics did not significantly affect TAC levels in DKD patients, they significantly reduced hs-CRP and MDA levels, consistent with the findings of Dai et al. (2022). Sensitivity analysis indicated that the high heterogeneity of MDA was mainly associated with the study by Miraghajani et al. (2017) After excluding this study, the heterogeneity of MDA markedly decreased, while statistical significance remained unchanged (MD −0.68, 95% CI −0.86 to −0.51), suggesting that this result was robust. Notably, AbdelQadir et al. (2020) and Bohlouli et al. (2021) both reported that probiotics not only significantly decreased hs-CRP and MDA levels but also increased TAC levels. The discrepancy in TAC results may be attributed to heterogeneity among participants. Specifically, both AbdelQadir et al. (2020) and Bohlouli et al. (2021) included the study by Soleimani et al. (2017), which was conducted in uremic patients and may have exaggerated the difference in TAC between the probiotic and placebo groups. Therefore, although probiotics were found to significantly reduce hs-CRP and MDA levels, their effects on TAC warrant further investigation through large-scale RCTs.

These findings reflect the anti-inflammatory and antioxidant properties of probiotics, which may be mediated by restoring intestinal barrier integrity and preventing lipopolysaccharide (LPS) translocation. On the one hand, gut dysbiosis disrupts intestinal barrier function, promoting the accumulation of inflammatory mediators and uremic toxins such as indoxyl sulfate and trimethylamine N-oxide (Wang et al., 2025; Sun et al., 2022). On the other hand, dysbiosis-induced LPS translocation triggers systemic inflammation, oxidative stress, and immune cell infiltration, leading to renal inflammation and fibrosis (Jiang et al., 2025). Conversely, probiotics may restore intestinal barrier integrity and suppress LPS translocation, thereby mitigating renal inflammation, oxidative stress, and fibrosis to exert renoprotective effects (Wang et al., 2016).

Compared with previous meta-analyses, our study offers several novel findings and methodological improvements. First, unlike earlier analyses by AbdelQadir et al. (2020), Bohlouli et al. (2021), and Tarrahi et al. (2022), our meta-analysis included a broader and more up-to-date range of RCTs, providing a more comprehensive and precise assessment of probiotic efficacy in DKD. Importantly, compared with the meta-analyses by Tarrahi et al. (2022) and Dai et al. (2022), we excluded studies involving uremic patients, such as that by Soleimani et al. (2017), to minimize potential confounding caused by severe renal impairment and dialysis-related metabolic disturbances. This refinement allowed us to more accurately evaluate the true effect of probiotics on DKD rather than on end-stage renal disease. Consequently, our findings revealed that probiotics significantly improved renal function (reducing SCR and BUN levels), glycemic control (lowering FBG), and lipid metabolism (reducing TG, TC, and LDL-C levels), while also exerting anti-inflammatory and antioxidant effects (decreasing hs-CRP and MDA). Moreover, the certainty of evidence for most positive outcomes was moderate and free from publication bias, which enhances the reliability of these conclusions.

From a clinical perspective, this meta-analysis provides several valuable insights. The results support probiotics as a promising complementary therapy for patients with DKD, supplementing standard hypoglycemic and renoprotective treatments. By improving renal biochemical markers, glycemic and lipid metabolism, and attenuating inflammation and oxidative stress, probiotics may help delay DKD progression and improve overall metabolic homeostasis. These findings suggest that probiotics could be particularly beneficial for patients with early or moderate DKD, who may achieve renal and metabolic protection through modulation of the gut–kidney axis. Clinicians may therefore consider probiotic supplementation, particularly strains with proven efficacy in restoring short-chain fatty acid-producing bacteria and enhancing intestinal barrier function, as part of a personalized and multi-target therapeutic strategy. Future large-scale, strain-specific RCTs are warranted to determine the optimal probiotic formulations, dosages, and treatment durations, and to clarify the underlying mechanisms linking gut microbiota modulation with renal outcomes in DKD.

Despite the robustness of our findings, several limitations should be acknowledged. First, all included studies were conducted in Iran, Malaysia, and China, which may limit the external applicability of the results to other populations with different genetic backgrounds, dietary patterns, and lifestyles. Second, although our meta-analysis incorporated the most recent and comprehensive collection of RCTs to date, the total sample size of DKD patients remains relatively small, which may reduce the statistical power and generalizability of the conclusions. Third, heterogeneity was observed in several outcomes, which could be attributed to variations in probiotic strains, dosages, intervention durations, disease stages, and baseline characteristics of participants. Fourth, most included studies did not provide detailed information on dietary habits, medication use, or gut microbiota composition, making it difficult to assess potential confounding factors or to identify the specific microbial alterations responsible for clinical benefits. Fifth, the intervention durations in most trials were relatively short, which may not fully capture the long-term renal and metabolic effects of probiotics. Additionally, publication bias cannot be completely excluded, as studies with negative or null findings are less likely to be published.

Future research should aim to conduct large-scale, high-quality, multicenter RCTs across diverse populations with standardized probiotic strains, dosages, and intervention durations to validate and extend these findings. Further investigations are warranted to explore strain-specific effects and potential synergistic combinations that maximize renal protection in DKD. The integration of metagenomic, metabolomic, and transcriptomic analyses could help elucidate the mechanistic pathways linking gut microbiota modulation to renal and metabolic outcomes. Moreover, future studies should assess the long-term efficacy and safety of probiotics across different stages of DKD and in conjunction with conventional therapies. Such efforts will provide stronger evidence to support the clinical application of probiotics as a safe, accessible, and mechanism-based complementary therapy for DKD.