Glucose Parameters, Inflammation Markers, and Gut Microbiota Changes of Gut Microbiome–Targeted Therapies in Type 2 Diabetes Mellitus: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Abstract

Purpose

This meta-analysis aims to summarize the effects of gut microbiome–targeted therapies (MTTs) on glucometabolic, inflammatory factors and gut microbiota in patients with type 2 diabetes mellitus (T2DM).

Methods

Four databases were searched for randomized controlled trials (RCTs) that included subjects with T2DM who received MTTs. All results were presented as standardized mean difference (SMD) or mean difference (MD) and 95% CIs. In addition, subgroup analyses were performed according to region, type of MTTs, number of probiotic strains, probiotics dose, prebiotics dose, duration of MTTs, mean age, and baseline body mass index.

Results

A total of 54 RCTs were included, encompassing 60 groups and 3390 subjects. Overall, MTTs intervention decreased fasting plasma glucose (MD = −7.97 mg/dL [95% CI = −10.82, −5.12]; P < .00001), 2-hour postprandial blood glucose (MD =  −43.30 mg/dL [95% CI = −75.83, −10.77]; P = .009), fasting insulin (MD = −1.73 uU/mL [95% CI = −2.63, −0.84]; P = .0001), HbA1c (MD = −0.28%, [95% CI = −0.39, −0.17]; P < .00001), and homeostatic model assessment of insulin resistance (MD =−0.53 [95% CI = −0.85, −0.20]; P = .0002). Furthermore, MTTs supplementation reduced high-sensitivity C-reactive protein, tumor necrosis factor alpha, and lipopolysaccharides. Meanwhile, the levels of interleukin-10 were increased. Moreover, the abundance of Actinobacteria, Lactobacillus, and Lactobacillus casei subgroup increased.

Conclusion

MTTs modestly improved glucometabolic parameters, reduced pro-inflammatory cytokines, and enriched beneficial microbes (eg, Actinobacteria, Lactobacillus) in subjects with T2DM. However, heterogeneity and limited long-term data highlight the need for large-scale RCTs.

The continuing worldwide increase in the prevalence of type 2 diabetes mellitus (T2DM) is due to the increasing prevalence of obesity and unhealthy lifestyles (1). With a high disability rate, T2DM has become a major global concern and affects the health and well-being of individuals, families, and societies (2). T2DM is a complex disease involving the interaction of genetic and environmental factors (3, 4). Its etiology is not fully understood. In recent years, the relationship between the gut microbiota (GM) and the incidence of T2DM is rapidly gaining recognition. A metagenome-wide association study showed that patients with T2DM had moderate GM dysbiosis, with a significant reduction in some universal butyrate-producing bacteria and an increase in a variety of opportunistic pathogens (5). An increase in invasive pathogens can degrade intestinal mucus and reach the epithelial barrier. Subsequently, the bacterial component lipopolysaccharide (LPS) binds to the toll-like receptor (TLR) on the surface of intestinal epithelial cells, recruits intestinal immune cells, releases pro-inflammatory cytokines, inhibits the expression of intestinal tight junction proteins, disrupts the intestinal epithelial barrier, promotes the passage of intestinal bacteria through the intestinal plate, and further disrupts the intestinal vascular barrier. Eventually, gut bacteria and their harmful metabolites cross the compromised gut barrier and enter the circulation, causing damage to distant body organs (6). In addition, a growing number of studies have shown that antidiabetic drugs can improve the host's substance and energy metabolism by regulating the structure of GM, thus playing a beneficial role in the treatment of T2DM (7, 8). The meta-analysis by Cao et al highlighted that metformin increased the abundance of Enterobacteriales and Akkermansia muciniphila in T2DM patients (9). A clinical study found that acarbose or vildagliptin monotherapy contributed to the overall and drug-specific GM responses in T2DM patients (10). Glucagon-like peptide-1 (GLP-1) receptor agonists have also been reported to selectively increase short-chain fatty acid–producing bacteria and Bifidobacteria in diabetic male rats (11) to improve glucose-induced insulin secretion (12). A preclinical study (13) showed that dapagliflozin increased the abundance of Proteobacteria to lower glucose. The above results suggest that the remission of diabetes mellitus is associated with the remodulation of GM. Therefore, it is urgent to discern whether a direct modulation of GM would become a promising adjuvant approach for the prevention and amelioration of T2DM.

Supplementation with probiotics, prebiotics, and synbiotics appears to present potential benefits for T2DM in animal models to restore the structure and function of GM. Probiotics are a group of live microorganisms beneficial to the host when consumed in sufficient quantities (14). Many previous studies have shown that probiotics such as L. rhamnosus (15), Bifidobacterium (16), and A. muciniphila (17), have significant glucose-lowering effects in diabetic animal models by significantly increasing the expression of tight junction protein in intestinal tissue, thereby enhancing the intestinal mechanical barrier (18, 19). Prebiotics are non- or low-digested food components (20). In diabetic animal models, investigators have found that the supplementation of prebiotics reduced blood glucose (BG), insulin resistance (IR) (21), and mRNA expression of serum LPS, interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α) (22) by increasing the growth of beneficial bacteria (21) and improving gut barrier function (21, 23). Synbiotics are a combination of live microorganisms and substrates, having pronounced effects on glycemic control and inflammation (24). A study by Yavorov-Dayliev et al (25) showed that synbiotics could increase the abundance of Bifidobacterium lactis and promote the reduction of fasting BG in diet-induced diabetic and obese mice. In addition, fecal microbiota transplantation (FMT), a technique where the fecal microbiota from healthy donors is transplanted into the gastrointestinal tract of patients, has been found to inhibit islet β cell apoptosis and reduce IR in a T2DM mouse model (26, 27).

Microbiome–targeted therapies (MTTs) have been shown to significantly improve metabolic diseases in animals. However, clinical studies have produced conflicting results. A clinical study found that prebiotics caused a significant increase in BG in healthy subjects (28). A previous systematic review concluded the effects of pro-/pre-/synbiotics on glucose homeostasis in diabetes mellitus (29, 30), but only indicators of glucose metabolism or lipid metabolism were analyzed. To date, no meta-analysis has comparatively assessed the efficacy of MTTs on glucose parameters, inflammation markers, and GM changes in T2DM with sufficient depth. Hence, this meta-analysis aims to fill the gap by exploring and assessing the current evidence on the efficacy of MTTs on various biomarkers, such as glycemia, inflammation, and the diversity and composition of GM. The results will contribute to the existing evidence that these methods of GM modulation can complement current therapeutic regimens, develop more rational and scientific formulations, and deepen the understanding of potential health benefits of GM modulation.

Methods

Study Protocol

This meta-analysis was conducted in accordance with the guidance of the Cochrane Handbook for Systematic Reviews of Interventions (31) and was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline (32). The PRISMA Checklist is shown in Supplementary Table S1 (33). The study protocol was registered in the International Prospective Register for Systematic Reviews (PROSPERO) under the number CRD42023480561.

Search Strategy

From the database establishment to November 14, 2023, a comprehensive search was performed in PubMed, Embase, the Cochrane Library, and Web of Science based on Medical Subject Headings (MeSH) terms and corresponding free words. Gray literature was also searched through ClinicalTrials.org and ProQuest Dissertations and Theses. To avoid excluding relevant data, any terms involving the study design were not used. The terms in titles and abstracts are shown in Supplementary Table S2 (33). In February 2024, an updated search of follow-up publications was conducted for study protocols that were identified in the previous search.

Eligibility Criteria

Randomized controlled trials (RCTs) reporting the effects of MTTs on glucose metabolism, inflammation, and GM in T2DM patients were eligible for inclusion. MTTs were assigned as interventions in any of the following 4 categories: probiotics, prebiotics, synbiotics, or FMT. For control studies, the comparator included placebo, usual care, or no intervention. One of the following results was assessed: (i) glucose homeostasis: fasting plasma glucose (FPG), 2 hours postprandial BG (2hPBG), fasting insulin, fasting C-peptide, glycosylated hemoglobin (HbA1c), homeostatic model assessment for insulin resistance (HOMA-IR), homeostatic model assessment for beta-cell function (HOMA-β), or quantitative insulin sensitivity check index (QUICKI); (ii) α-diversity: Shannon index, Simpson diversity indices, Chao1 index, abundance-based coverage estimator (ACE) index, dominance indices, Simpson evenness index, observed operational taxonomic units (OTUs), phylogenetic diversity); (iii) β-diversity: metric, principal coordinates analysis (PCoA), principal component analysis (PCA), multidimensional scaling analysis (MDS); (iv) changes in relative abundance of GM; and (v) systemic inflammation: high-sensitivity C-reactive protein (hs-CRP), TNF-α, interferon-γ (IFN-γ), LPS, IL-1β, IL-6, IL-8, IL-10. There were no restrictions on the duration, publication time, age, sex, ethnicity, and region for included studies. However, this meta-analysis included studies in English only, excluding reviews, case reports, abstracts and proceedings, correspondences, conferences, editorial and nonclinical study papers, animal experiments, and articles with unavailable full text or incomplete data.

Data Extraction

Two investigators (Z.X. and Z.W.B.) independently screened the studies generated from the search based on titles and abstracts to remove those not relevant to this analysis. The same 2 investigators independently read the full reports of included studies to determine which to retain. Any inconsistency was resolved by a third investigator (Z.T.S.) independently. Relevant data below were extracted by 2 independent reviewers (Z.X. and Z.W.B.) using a standardized Excel table: (i) study characteristics (first author’s name, publication year, date of enrollment, inclusion and exclusion criteria, region, study design); (ii) subject characteristics (age, baseline body mass index [BMI], and number of subjects in the intervention and control groups); (iii) intervention characteristics (intervention type, intervention duration, drug formulation, daily dose of nutraceutical and control article/placebo, and number of probiotic strains); and (iv) changes in mean ± SD for continuous variables, as previously defined. Any disagreements were resolved by a third author (Z.T.S.) independently. For studies with multiple time points, only data from the last time point were included. Corresponding authors were contacted as necessary for data not included in the published reports.

Risk Assessment of Bias

Two independent authors (Z.X. and Z.W.B.) evaluated the quality of RCTs using the Cochrane Collaboration's tool for risk of bias (ROB). This tool examines the sources of bias across 7 dimensions, including an assessment of ROB from 6 different aspects: selection bias (SB), implementation bias (IB), measurement bias (MB), follow-up bias (FB), reporting bias (RB), and other bias. Each item listed above contains 3 results: “low risk,” “high risk,” and “unclear.” Less “high risk” indicates better quality.

Data Analysis

The meta-analysis was conducted using Review Manager (RevMan) V.5.4 software and STATA software version 15.0. Given the inevitable clinical heterogeneity due to different doses, duration of MTTs, and the differences in methods for analyzing and reporting GM diversity, the data (≥ 3 studies within a single group) were analyzed using the random-effect method to calculate the pooled effect sizes. Summary effect measures were reported as standardized mean differences (SMDs) or MD with 95% CI. An I-squared (I²) statistic with a cutoff of ≥ 50% was considered as the evidence of between-study heterogeneity (BSH) (34). Based on the World Health Organization regional classification (Europe, Mediterranean, Southeast Asia, Americas, or Western Pacific), intervention type (probiotics, prebiotics, synbiotics, or FMT), number of probiotic strains (multiple strains or single strain), probiotics dose (≥ 10¹⁰ colony-forming units [CFU] or < 10¹⁰ CFU), prebiotics dose (≥ 10 g or < 10 g), study duration (≥ 12 weeks or < 12 weeks), baseline BMI (≥ 28 kg/m² or < 28 kg/m²), mean age (≥ 52 years old or < 52 years old), subgroup analyses (at least 4 studies) were carried out to investigate the possible sources of heterogeneity. In addition, sensitivity analyses were conducted on indicators with more than 10 study groups and an I² ≥ 50%. In each round, one study was excluded to assess the robustness and reliability of the results and identify potential sources of heterogeneity. When more than 10 studies were included, Egger's test and funnel plots were generated to examine the potential publication bias (PB).

Results

Included Studies

The selection process is detailed in Fig. 1. A total of 8933 relevant studies were retrieved from the initial search, including 1818 from PubMed, 3355 from EMBASE, 1216 from Cochrane, and 2545 from Web of Science. After removing duplicates by 2 investigators using EndNote, 5123 studies remained. Following a thorough review of the titles and abstracts, 116 studies were selected based on the inclusion and exclusion criteria. After reading the full text, 54 studies were screened and included in the final meta-analysis.

Figure 1.

Flow diagram of study selection process.

Study Characteristics

The characteristics of the included studies are presented in Table 1. A total of 54 studies (15, 35-87) were included, involving 3390 patients (1740 in intervention groups and 1650 in control groups, respectively). The earliest study was published in 2012, and 27 of the studies included were published in the last 5 years. Of all the studies, 6 (11.11%) were conducted in Europe (Switzerland, Ukraine, Denmark, and the United Kingdom) (57, 60, 61, 66, 70, 79), 30 (55.56%) in the Mediterranean (Iran) (35-40, 43-51, 54, 55, 59, 64, 65, 67, 69, 72-76, 81, 82, 85), 6 (11.11%) in Southeast Asia (India, Indonesia, and Thailand) (41, 62, 63, 77, 80, 83), 3 (5.56%) in the Americas (Brazil and the United States) (53, 71, 84), and 9 (16.67%) in the Western Pacific (15, 42, 52, 56, 58, 68, 78, 86, 87). Most of the subjects did not mention the complications. However, 10 studies had narrow inclusion criteria, such as T2DM with chronic periodontitis (40), T2DM with microalbuminuria (45), T2DM with low-density lipoprotein-cholesterol (LDL-c) levels equal to or greater than 100 mg/dL (65), T2DM with abdominal obesity (66), T2DM with obesity (68), T2DM with grade 3 diabetic foot ulcer (DFU) (67), T2DM with overweight and stable coronary heart disease (CHD) condition (81), T2DM with 2- and 3-vessel coronary heart disease (CHD) (72, 73), and T2DM with prehypertension (80). In all included studies, a total of 29 study administered probiotics supplements (15, 37, 39, 41, 42, 46, 47, 51, 52, 57, 59-67, 69, 72, 73, 77-79, 82-84, 87), 11 study arms prescribed prebiotic supplements (mainly containing fructo-oligosaccharide, galacto-oligosaccharides, or insulin) (36, 43, 44, 48, 49, 50, 53, 54, 70, 75, 76), 13 studies used synbiotic supplements (35, 38, 40, 45, 55, 56, 58, 71, 74, 80-82, 85), and only 2 study arms used FMT (68, 86). Both of the FMT-related studies presented data on co-interventions. Lifestyle modification intervention was applied in one study (68), and metformin in the other (86). Other studies also presented data on co-treatments, including the combination with green pumpkin (39), lactic acid (55), and metformin (42, 62), berberine (87), respectively. Multiple strains of bacteria were administered in a total of 30 studies (50%), while a single strain administered in 15 (25%). Some of these studies had multiple intervention groups comparing probiotics, prebiotic, or synbiotics supplements with placebo, each of such study was considered as a single study in the meta-analysis. The duration of intervention ranged from 4 weeks to 6 months. The median age of subjects was 54.45 years in intervention groups, while the median baseline BMI was 29.20 kg/m². There were only a few reports of side effects due to MTTs with no difference between the control and study groups. Most of the reported side effects were gastrointestinal (Supplementary Fig. S1) (33).

Table 1.

Detailed characteristics of included studies

Author, year	Region	Design	Population	Intervention		Control		Sample size		Duration (week or month)	Age (mean ± SD; y)		BMI (mean ± SD; kg/m2)
				Content	Dose (per day)	Content	Dose (per day)	Intervention	Control		Intervention	Control	Intervention	Control
Akram et al (2015)	Iran	Randomized, double-blind, placebo-controlled trial	T2DM	Synbiotic (pre- and probiotic) tablet	1 tablet	Placebo tablet	1 tablet	22	22	8W	53.45 ± 10.8	54.5 ± 11.0	22.79 ± 2.7	22.47 ± 2.38
Aliasgharzadeh et al (2015)	Iran	Randomized, triple-blind, controlled trial	T2DM	Resistant Dextrin	10 g	Maltodextrin	10 g	30	25	8W	49.2 ± 9.6	49.6 ± 8.4	31.8 ± 4.5	30.8 ± 5.2
Alihosseini et al (2017)	Iran	Randomized, double-blind, placebo-controlled, clinical trial	T2DM	Fermented milk (kefir) containing Streptococcus thermophiles, Lactobacillus casei, L. acidophilus and Bifidobacterium lactis	Lactobacillus casei (1.2-9 × 109 CFU), L. acidophilus (1.8-15 × 109 CFU) and Bifidobacterium lactis (3-48 × 108 CFU)	Conventional fermented milk contained Streptococcus thermophiles and Lactobacillus bulgaricus (dough)	600 mL	30	30	8W	—		28.74 ± 4.75	27.30 ± 3.64
Asemi et al (2013)	Iran	Randomized, double-blinded, controlled clinical trial	T2DM	Synbiotic capsules consisted of freeze-dried strains: L. acidophilus, L. casei, L. rhamnosus, L. bulgaricus, Bifidobacterium breve, B. longum, Streptococcus thermophilus, and FOS with lactose	L. acidophilus (2 × 109 CFU), L. casei (7 × 109 CFU), L. rhamnosus (1.5 × 109 CFU), L. bulgaricus (2 × 108 CFU), Bifidobacterium breve (2 × 1010 CFU), B. longum (7 × 109 CFU), Streptococcus thermophilus (1.5 × 109 CFU) + 100 mg fructo-oligosaccharide	FOS with lactose	1 capsule	27	27	8W	50.51 ± 9.82	52.59 ± 7.14	31.61 ± 6.36	30.17 ± 4.23
Bayat et al (2016)	Iran	Randomized, parallel–group, controlled clinical trial	T2DM	Probiotic yogurt	150 g	Dietary advice	NA	20	20	8W	54.1 ± 9.54	46.95 ± 9.34	28.77 ± 4.59	29.75 ± 4.66
				Green pumpkin plus probiotic yogurt	100 g + 150 g	Green pumpkin	100 g	20	20	8W	53.65 ± 6.99	51.8 ± 2.24	27.98 ± 4.2	28.95 ± 3.34
Bazyar et al (2020)	Iran	Randomized, double-blinded, clinical trial	T2DM with Chronic Periodontitis	Synbiotic capsules containing freeze-dried strains: Lactobacillus acidophilus, L. casei, L. rhamnosus, L. bulgaricus, Bifidobacterium breve, B. longum, Streptococcus thermophilus, and FOS	L. acidophilus (2 × 109 CFU), L. casei (7 × 109 CFU), L. rhamnosus (1.5 × 109 CFU), L. bulgaricus (2 × 108 CFU), Bifidobacterium breve (2 × 109 CFU), B. longum (7 × 109 CFU), Streptococcus thermophilus (1.5 × 109 CFU) + 100 mg FOS	Wheat flour capsule	500 mg	23	24	8W	48.6 ± 5.8	50.1 ± 3.6	24 ± 3.6	25.5 ± 2.7
Chaiyasut et al (2023)	Thailand	Randomized, double-blind, placebo-controlled clinical trial	T2DM	Aluminum-foil sachets containing probiotic Bifidobacterium breve	50 × 109 CFU	Corn starch	10 mg	20	20	12W	63.90 ± 1.44	61.05 ± 1.83	23.80 ± 0.74	25.73 ± 1.27
Chen et al (2023)	China	Randomized, double-blinded, clinical trial	T2DM	Probio-X (Lactobacillus casei Zhang, Bifidobacterium lactis V9, Lactobacillus plantarum P-8, Lactobacillus rhamnosus Probio-M9, and Bifidobacterium lactis Probio-M8) and metformin tablets	3 × 1010 CFU bacteria + 0.75-1.5 g metformin	Maltodextrin	2 g maltodextrin + 0.75-1.5 g metformin	27	21	3M	48.70 ± 11.11	46.90 ± 11.25	—
Dehghan(a) et al (2014)	Iran	Randomized, triple-blind, controlled study	T2DM	Oligofructose-enriched inulin	10 g	Maltodextrin	10 g	27	25	8W	48.4 ± 8.4	48.7 ± 9.7	31.9 ± 4.5	29.9 ± 4.2
Dehghan(b) et al (2014)	Iran	Randomized, triple-blind, controlled study	T2DM	High-performance inulin	10 g	Maltodextrin	10 g	24	25	8W	47.8 ± 10.1	48.7 ± 9.7	31.6 ± 4.1	29.9 ± 4.2
Ebrahimi et al (2017)	Iran	Single-center, randomized, double-blind, placebo-controlled trial	T2DM with microalbuminuria	Synbiotic capsules containing Lactobacillus family, Bifidobacterium family, Streptococcus thermophilus, FOS, and B group vitamins, lactose, malt-dextrin, magnesium saturate, and talc	500 mg	Capsules containing row starch and B group vitamins, lactose, malt-dextrin, magnesium saturate, and talc	500 mg	35	35	9W	58.71 ± 8.20	58.63 ± 8.06	28.13 ± 3.78	27.30 ± 3.81
Ejtahed et al (2012)	Iran	Randomized, double-blinded, controlled clinical trial	T2DM	Probiotic yoghurt containing Lactobacillus bulgaricus, Streptococcus thermophilus, Bifidobacterium lactis Bb12, and L. acidophilus La5	Bifidobacterium lactis Bb12 (5.37-18.12 × 108 CFU) and L. acidophilus La5 (5.55-21.69 × 108 CFU)	Conventional yoghurt containing Lactobacillus bulgaricus, Streptococcus thermophilus	300 g	30	30	6W	50.87 ± 7.68	51.00 ± 7.32	28.95 ± 3.65	29.14 ± 4.30
Eliuz Tipici et al (2023)	Iran	Randomized, placebo-controlled, clinical trial	T2DM	Probiotic drops containing Lactobacillus rhamnosus GG, corn oil, and an emulsifier	1 × 1010 CFU	Corn oil	10 drops	17	17	8W	44.03 ± 6.68	47.52 ± 7.52	34.52 ± 6.43	32.15 ± 5.37
Farhangi et al (2016)	Iran	Randomized, double-blind, placebo-controlled trial	T2DM	Chicory inulin enriched with oligofructose	10 g	Maltodextrin	10 g	27	22	2M	48.07 ± 8.70	48.61 ± 9.16	31.43 ± 3.50	29.98 ± 4.01
Farhangi et al (2018)	Iran	Randomized, triple-blind, placebo-controlled, clinical trial	T2DM	Resistant dextrin supplement (NUTRIOSE®06)	10 g	Maltodextrin	10 g	30	25	8W	49.2 ± 9.60	49.6 ± 8.4	31.8 ± 4.5	30.8 ± 5.2
Farhangi et al (2020)	Iran	Randomized, triple-blind, placebo-controlled trial	T2DM	Resistant dextrin supplement (NUTRIOSE®06)	10 g	Maltodextrin	10 g	32	33	8W	49.5 ± 8.0	48.6 ± 7.9	31.5 ± 4.5	32.0 ± 3.9
Feizollahzadeh et al (2017)	Iran	Randomized, double-blind, parallel-group, controlled trial	T2DM	Probiotic soy milk containing Lactobacillus plantarum A7	4 × 109 CFU	Conventional soy milk	200 mL	20	20	8W	56.90 ± 8.09	53.6 ± 7.16	26.68 ± 3.18	26.58 ± 3.26
Firouzi et al (2017)	Malaysia	Randomized, double-blind, parallel-group, controlled clinical trial	T2DM	Probiotic sachets containing Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus lactis, Bifidobacterium bifidum, Bifidobacterium longum, and Bifidobacterium infantis	3 × 1010 CFU	Organoleptically similar sachets without probiotic	2 sachets	48	53	12W	52.9 ± 9.2	54.2 ± 8.3	29.2 ± 5.6	29.3 ± 5.3
Frias et al (2023)	USA	Randomized, double-blind, placebo-controlled trial	T2DM	Prebiotic shake packets containing of Resistant Starch Blend, Soluble Vegetable Fiber, Oat Beta Glucan, Milk Protein Concentrate, Allulose, Rebaudioside M, High Oleic Sunflower Oil, Xanthan Gum, Guar Gum, Natural Vanilla Flavor, Vitamin-Mineral Blend or shake packets containing of Type 2 Resistant Starch Blend, Soluble Vegetable Fiber, Oat Beta Glucan, Milk Protein Concentrate, Allulose, Rebaudioside M, High Oleic Sunflower Oil, Xanthan Gum, Guar Gum, Natural Chocolate Flavor, Alkalized Cocoa Powder, Vitamin-Mineral Blend	15 g dietary fiber	Shake packets containing maltodextrin, Milk Protein Concentrate, Xanthan Gum, Guar Gum, Natural Chocolate Flavor, Sucralose	49-51 g	75	34	12W	54.3 ± 9.7		35.9 ± 5.8
Gargari et al (2015)	Iran	Randomized, triple-blind, controlled clinical study	T2DM	60% resistant starch type 2	10 g	Maltodextrin	10 g	28	32	8W	49.5 ± 8.0	49.6 ± 8.4	31.5 ± 4.5	30.8 ± 5.2
Ghafouri et al (2019)	Iran	Randomized, double-blinded, controlled clinical trial	T2DM	Synbiotic bread containing betaglucan, Bacillus coagulans, and inulin	9 g beta-glucan + 3 × 108 CFU Bacillus coagulans + 30 g inulin	Bread containing betaglucan	9 g betaglucan (9 g)	25	25	8W	54.92 ± 5.1	54.60 ± 4.15	26.39 ± 2.55	27.04 ± 2.50
				Synbiotic bread containing beta-glucan, Bacillus coagulans, inulin, and lactic acid	9 g beta-glucan + 3 × 108 CFU Bacillus coagulans + 30 g inulin + 12 g lactic acid	Bread containing betaglucan and lactic acid	9 g betaglucan + 12 g lactic acid	25	25	8W	53.88 ± 5.45	55.00 ± 4.85	26.83 ± 2.1	26.33 ± 2.3
Horvath et al (2020)	Austria	Randomized, double-blind, placebo-controlled clinical trial	T2DM	Probiotic sachets containing B. bifidum W23, B. lactis W51, B. lactis W52, L. acidophilus W37, L. casei W56, L. brevis W63, L. salivarius W24, Lc. lactis W58, Lc. lactis W19, maize starch, maltodextrins, vegetable protein, potassium chloride, magnesium sulphate, amylases and manganese sulphate + the prebiotic containing Galacto-oligosaccharides P11 (GOS) and FOS, konjac glucomannan P13 (E425), calcium carbonate (E170), zinc citrate 3-hydrate, vitamin D3 (cholecalciferol) and vitamin B2 (riboflavin) (E101) and a matrix containing maltodextrin, natural elderflower flavouring and Gum Arabic (E414)	1.5 × 1010 CFU bacteria + 8 g prebiotics	Matrix without bacteria + matrix containing maltodextrin, natural elderflower flavouring and Gum Arabic (E414)	6 g + 10 g	12	14	6M	61 ± 7.87	59 + 8.66	33 ± 3.15	34 ± 3.46
Hove et al (2015)	Denmark	Randomized, double-blind, placebo-controlled, prospective 2×2 factorial study	T2DM	Cardi04 yogurt containing Lactobacillus helveticus	300 mL	Artificially acidified milk (nZ18)	300 mL	23	18	12W	58.5 ± 7.7	60.6 ± 5.2	29.2 ± 3.8	27.7 ± 3.3
Hsieh et al (2018)	China	Randomized, double-blind, placebo-controlled, clinical trial	T2DM	Live Lactobacillus reuteri ADR-1	4 × 109 CFU Lactobacillus reuteri	Placebo	2 capsules	22	22	6M	52.32 ± 10.20	55.77 ± 8.55	28.04 ± 4.29	27.53 ± 3.15
Kanazawa et al (2021)	Japan	Randomized controlled study	T2DM	Synbiotic dry powder containing living Lacticaseibacillus paracasei YIT 9029, living Bifidobacterium breve YIT 12272, and galacto-oligosaccharides	≥ 3 × 108  Lacticaseibacillus paracasei YIT 9029 + 3 × 108  Bifidobacterium breve YIT 12272 + 7.5 g galacto-oligosaccharides	Placebo	NA	44	42	24W	61.1 ± 11.0	55.9 ± 10.7	29.5 ± 4.4	29.1 ± 3.7
Khalili et al (2019)	Iran	Randomized, parallel-group, controlled trial	T2DM	Capsules containing Lactobacillus casei	108 CFU Lactobacillus casei	Maltodextrin	1 capsule	20	20	8W	43.95 ± 8.14	45.00 ± 5.37	29.50 ± 3.34	31.94 ± 5.76
Kobyliak et al (2018)	Ukraine	Single-center, double-blind, placebo-controlled, parallel-group trial	T2DM	Probiotc sachets containing alive probiotic strains of Lactobacillus, Lactococcus, Bifidobacterium, Propionibacterium, Acetobacter genera	Lactobacillus + Lactococcus 6 × 1011 CFU/g + 1 × 1011 CFU Bifidobacterium + 3 × 1011 CFU Propionibacterium + 1 × 1011 CFU Acetobacter	Placebo	10  g	31	22	8W	52.23 ± 9.69	57.18 ± 9.66	34.70 ± 7.18	35.65 ± 7.36
Kobyliak et al (2021)	Ukraine	Randomized, placebo-controlled trial	T2DM	Probiotic sachet containing alive microorganisms: Lactobacillus, Bifidobacterium, Lactococcus, Propionibacterium and Acetobacter and smectite gel	Lactobacillus (1.0× 1010 CFU/g), Bifidobacterium (1.0× 1010 CFU), Lactococcus (1.0× 109 CFU), Propionibacterium (1.0× 109 CFU) and Acetobacter (1.0× 106 CFU)	Placebo	10 g	28	27	8W	53.82 ± 9.58	56.93 ± 9.88	31.99 ± 6.02	32.28 ± 6.08
Kumar et al (2022)	India	Randomized prospective observational trial	T2DM	Tab. Metformin and Cap. Probiotics 1 capsule	1000 mg Tab. Metformin + 2 capsule	Tab. Metformin	1000 mg	75	75	12W	50.90 ± 6.24	51.06 ± 5.42	—
Lestari et al (2019)	Indonesia	Randomized, double-blinded, controlled clinical trial	T2DM	Probiotic yogurt containing Lactobacillus acidophilus La-5 and Bifidobacterium lactis BB-12	> 108 CFU bacterial count	Conventional yogurt containing Streptococcus thermophilus and Lactobacillus bulgaricus	100 mL	16	16	4W	56 ± 7	53 ± 10	27.62 ± 4.58	27.74 ± 3.16
Mazloom et al (2013)	Iran	Single-blinded clinical trial	T2DM	Probiotic capsules containing L. acidophilus, L. bulgaricus, L. bifidum, and L. casei	3000 mg	Capsules containing magnesium stearate	3000 mg	16	18	6W	55.4 ± 8	51.8 ± 10.2	27.97 ± 3.81	27.24 ± 2.73
Mirjalili et al (2023)	Iran	Randomized controlled study	T2DM with low-density lipoprotein-cholesterol (LDL-c) equal to or greater than 100 mg/dL	Probiotic yogurt containing live Lactobacillus acidophilus and Bifidobacterium lactis	9.30 × 108 CFU	Conventional yogurt	200 g	36	36	12W	54.5 ± 8.0	58.1 ± 9.8	29.83 ± 4.6	28.65 ± 3.8
Mobini et al (2017)	Sweden	Randomized, double-blind, placebo-controlled trial	T2DM with abdominal obesity	Probiotic stick packs containing low-dose Lactobacillus reuteri DSM 17938	108 CFU	Stick packs containing mildly sweet tasting powder	1 package	15	15	12W	66 ± 6	65 ± 5	30.6 ± 4.5	30.7 ± 4.0
				Probiotic stick packs containing high-dose Lactobacillus reuteri DSM 17938	1010 CFU	Stick packs containing mildly sweet tasting powder	1 package	14	15	12W	64 ± 6	65 ± 5	32.3 ± 3.4	30.7 ± 4.0
Mohseni et al (2017)	Iran	Randomized, double-blind, placebo-controlled trial	T2DM with grade 3 diabetic foot ulcer (DFU)	Probiotic capsules containing Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus Fermentum, and Bifidobacterium bifidum	1 capsule	Placebo	1 capsule	30	30	12W	62.6 ± 9.7	58.5 ± 11.0	26.4 ± 3.0	25.3 ± 3.7
Ostadrahimi et al (2015)	Iran	Randomized, double-blind, placebo-controlled, clinical trial	T2DM	Fermented milk (kefir) containing Streptococcus thermophiles, Lactobacillus casei, L. acidophilus, and Bifidobacterium lactis	Lactobacillus casei (1.2-9.0 × 109 CFU), L. acidophilus (1.8-15 × 109 CFU), and Bifidobacterium lactis (3-48 × 108 CFU)	Conventional fermented milk (dough) containing Streptococcus thermophiles and Lactobacillus bulgaricus	600 mL	30	30	8W	—		28.89 ± 4.77	27.47 ± 3.55
Pedersen et al (2016)	UK	Randomized, double-blind, placebo-controlled parallel study	T2DM	galacto-oligosaccharide mixture	5.5 g	Maltodextrin	5.5 g	14	15	12W	56.7 ± 5.99	58.1 ± 6.58	28.0 ± 4.12	28.4 ± 3.49
Perraudeau et al (2020)	USA	Multicenter, randomized, double-blind, parallel-group, placebo-controlled proof-of-concept trial	T2DM	WBF-0010 synbiotic capsules containing Clostridium beijerinckii, Clostridium butyricum, Bifidobacterium infantis, designated butyrate, inulin, sucrose, trehalose, glycerin, hydroxypropyl methylcellulose, hydroxypropyl methylcellulose phthalate, titanium dioxide, polyvinylpyrrolidone, skim milk, and colloidal silicon dioxide	1.6 × 1010 CFU Clostridium beijerinckii, 3.3 × 109 CFU Clostridium butyricum, 2.0 × 109 CFU Bifidobacterium infantis	WBF-0009 containing colloidal silicon dioxide	6 capsule	21	16	12W	51.2 ± 9.62	53.5 ± 8	33.7 ± 5.96	33.5 ± 6.40
				WBF-0011 synbiotic capsules containing Clostridium beijerinckii, Clostridium butyricum, Akkermansia muciniphila, Eubacterium hallii, Bifidobacterium infantis, designated butyrate, inulin, sucrose, trehalose, glycerin, hydroxypropyl methylcellulose, hydroxypropyl methylcellulose phthalate, titanium dioxide, polyvinylpyrrolidone, skim milk, and colloidal silicon dioxide	1.6 × 1010 CFU Clostridium beijerinckii, 3.3 × 109 CFU Clostridium butyricum, 1.2 × 109 CFU Akkermansia muciniphila, 9 ×  108 CFU Eubacterium hallii, 2 ×  109 CFU Bifidobacterium infantis	WBF-0009 containing colloidal silicon dioxide	6 capsule	21	16	12W	51.8 ± 8.25	53.5 ± 8	31.7 ± 5.04	33.5 ± 6.40
Raygan et al (2018)	Iran	Randomized, double-blind, placebo-controlled trial	T2DM with 2- and 3-vessel CHD	Probiotic capsules containing Bifidobacterium bifidum, Lactobacillus casei, Lactobacillus acidophilus	2 × 109 CFU Bifidobacterium bifidum, 2 × 109 CFU Lactobacillus casei, 2 × 109 CFU Lactobacillus acidophilus	Placebo	1 capsule	30	30	12W	60.7 ± 9.4	61.8 ± 9.8	30.3 ± 5.2	29.3 ± 4.1
Raygan et al (2019)	Iran	Randomized, double-blind, placebo-controlled trial	T2DM with 2- and 3-vessel CHD	Selenium yeast plus probiotic containing Lactobacillus acidophilus, Lactobacillus reuteri, Lactobacillus fermentum, and Bifidobacterium bifidum	200 μg selenium + 8 × 109 CFU probiotic	Placebo	1	27	27	12W	64.8 ± 8.3	62.4 ± 13.1	31.4 ± 5.8	29.9 ± 5.0
Razmpoosh et al (2019)	Iran	Randomized, double-blinded, clinical trial	T2DM	Synbiotic capsules containing freeze-dried strains: Lactobacillus acidophilus, L. casei, L. rhamnosus, L. bulgaricus, Bifidobacterium breve, B. longum, Streptococcus thermophilus, and FOS with lactose	4 × 109 CFU L. acidophilus, L. casei (14 × 109 CFU), L. rhamnosus (3 × 109 CFU), L. bulgaricus (4 × 108 CFU), Bifidobacterium breve (6 × 1010 CFU), B. longum (14 ×  109 CFU), Streptococcus thermophilus (3 × 109 CFU) + 200 mg fructo-oligosaccharide	Capsules containing FOS and magnesium stearate	2 capsules	30	30	6W	58.6 ± 6.5	61.3 ± 5.2	27.7 ± 4.2	27.2 ± 4.2
Roshanravan(a) et al (2017)	Iran	Randomized, double-blind, placebo-controlled, clinical trial	T2DM	Inulin powder and starch capsules	10 g inulin powder + 600 mg starch capsules	Starch capsules and starch powder	600 mg starch capsules + 10 g starch powder	15	15	45d	51.47 ± 6.46	51.73 ± 8.44	30.37 ± 2.82	30.86 ± 5.41
Roshanravan(b) et al (2017)	Iran	Randomized, double-blind, placebo-controlled, clinical trial	T2DM	Inulin powder and starch capsules	10 g inulin powder + 600 mg starch capsules	Starch capsules and starch powder	600 mg starch capsules + 10 g starch powder	15	15	45d	51.47 ± 6.46	51.73 ± 8.44	30.37 ± 2.82	30.86 ± 5.41
Rustanti et al (2022)	Indonesia	Randomized, double-blinded, controlled clinical trial	T2DM	Skim milk powder containing Lactobacillus plantarum Dad-13	1010 CFU	Skim milk powder	1 g	18	18	11W	44.11 ± 3.31	43.44 ± 4.44	25.62 ± 3.15	24.58 ± 2.52
Sato et al (2017)	Japan	Randomized controlled study	T2DM	Lactobacillus casei strain Shirota (LcS)-fermented milk	≥ 4 × 1010 LcS cells	No LcS-fermented milk	80 mL	34	34	16W	64.0 ± 9.2	65.0 ± 8.3	24.2 ± 2.6	24.6 ± 2.6
Savytska et al (2023)	Ukraine	Single-center, randomized, double-blind, placebo-controlled trial	T2DM	Multi-strain probiotic Symbiter containing 14 live probiotic strains of Lactobacillus + Lactococcus, Bifidobacterium, Propionibacterium, and Acetobacter genera	6 × 1011 CFU Lactobacillus + Lactococcus + 1 × 1011 Bifidobacterium + 3 × 1011 Propionibacterium + 1 × 107 Acetobacter genera	Placebo	10 g	34	34	8W	53.82 ± 9.58	56.93 ± 9.88	31.92 ± 4.96	32.57 ± 6.16
Sheth et al (2015)	India	Randomized controlled study	T2DM with prehypertensive	Synbiotic product containing 2 species of Lactobacillus, Bifidobacterium, 1 species of Streptococcus, 1 species of yeast with FOS	1 g	Placebo	1 g	25	10	45d	—		—
SiewCNg et al (2022)	China	Randomized, double-blind, placebo-controlled trial	T2DM with obese	FMT solution with 50 g stool plus lifestyle intervention	100-200 mL	Sham solution plus LSI	100-200 mL	21	20	24W	55.28 ± 15.10	55.0 ± 11.97	31.28 ± 4.61	33.87 ± 3.34
Tajabadi-Ebrahimi et al (2014)	Iran	Randomized, double-blinded, controlled clinical trial	T2DM	Probiotic bread containing L. sporogenes	3 × 108 CFU L. sporogenes	Control bread without probiotic bacteria and prebiotic inulin	120 g	27	27	8W	52.0 ± 7.2	53.4 ± 7.5	29.8 ± 5.7	30.5 ± 4.1
				Synbiotic bread containing the viable and heat-resistant probiotic L. sporogenes and inulin	3 × 108 CFU L. sporogenes + 8.4 g inulin	Control bread without probiotic bacteria and prebiotic inulin	120 g	27	27	8W	51.3 ± 10.4	53.4 ± 7.5	30.8 ± 5.9	30.5 ± 4.1
Tajabadi-Ebrahimi et al (2017)	Iran	Randomized, double-blind, placebo-controlled trial	T2DM with overweight and stable CHD condition	Synbiotic capsules containing 3 probiotic bacteria species Lactobacillus acidophilus, Lactobacillus casei, Bifidobacterium bifidum plus inulin	2 × 109 CFU Bifidobacterium bifidum + 2 × 109 CFU Lactobacillus casei + 2 × 109 CFU Lactobacillus acidophilus + 800 mg inulin	Capsules containing starch	1 capsule	30	30	8W	64.2 ± 12.0	64.0 ± 11.7	29.6 ± 4.6	32.3 ± 6.0
Toejing et al (2021)	Thailand	Randomized, double-blind, placebo-controlled, clinical trial	T2DM	Foil envelope with L. paracasei HII01	50 × 109 CFU L. paracasei HII01	Foil envelope with corn starch	10 mg	18	18	12W	63.50 ± 5.94	61.78 ± 7.73	23.22 ± 2.72	23.05 ± 2.60
Tonucci et al (2017)	Brazil	Randomized, double-blind, parallel-group, placebo-controlled trial	T2DM	Probiotic fermented goat milk containing L. acidophilus La-5 and Bifidobacterium animalis subsp lactis BB-12	109 CFU L. acidophilus La-5 + 109 CFU Bifidobacterium animalis subsp lactis BB-12	Conventional fermented goat containingStreptococcus thermophilus TA-40	120 g	23	22	6W	51.83 ± 6.64	50.95 ± 7.20	27.49 ± 3.97	27.94 ± 4.15
Velayati et al (2021)	Iran	Randomized, double-blind, placebo-controlled trial	T2DM	Synbiotic sachets containing B. coagulans GanedenBC30, Lactobacillus rhamnosus GG, Lactobacillus acidophilus, 500 mg FOS and 0.7% Natural Orange flavour	1011 spores B. coagulans GanedenBC30 + 1010 CFU Lactobacillus rhamnosus GG +109 CFU Lactobacillus acidophilus + 500 mg FOS	Sachets containing starch and 0.7% Natural Orange flavour	2 g	20	23	12W	59.10 ± 9.71	60.39 ± 6.74	27.32 ± 4.34	28.27 ± 2.54
Wu et al (2022)	China	Randomized, controlled, prospective study	T2DM	FMT solution with 50 g stool plus Metformin	200 mL	Metformin	NA	8	12	4W	—		27.27 ± 1.07	27.51 ± 0.89
Zhang et al (2020)	China	Randomized, double-blind, placebo-controlled, clinical trial	T2DM	Multi-strain probiotics sachets containing 9 live, freeze-dried bacteria: Bifidobacterium longum, Bifidobacterium breve, Lactococcus gasseri, Lactobacillus rhamnosus, Lactobacillus salivarius, Lactobacillus crispatus, Lactobacillus plantarum, Lactobacillus fermentum, Lactobacillus casei plus placebo	≥1.5 × 1011 CFU bacteria	Placebo plus placebo	NA	102	103	12W	52.59 ± 10.53	53.65 ± 11.28	25.6 ± 2.96	26.2 ± 3.43
				Multi-strain probiotics sachets containing 9 live, freeze-dried bacteria: Bifidobacterium longum, Bifidobacterium breve, Lactococcus gasseri, Lactobacillus rhamnosus, Lactobacillus salivarius, Lactobacillus crispatus, Lactobacillus plantarum, Lactobacillus fermentum, Lactobacillus casei plus Aberberine	≥1.5 × 1011 CFU bacteria + 1.2 g berberine	Placebo and berberine	NA	106	98	12W	53.5 ± 9.77	51.94 ± 14.30	25.5 ± 2.86	25.7 ± 3.43

Abbreviations: BMI, body mass index; CFU, colony-forming units; CHD, coronary heart disease; FOS, fructo-oligosaccharide.

Risk of Bias and Publication Bias Assessment

The ROB of included studies and the assessment of individual studies are summarized in Fig. 2 and Supplementary Fig. S2 (33), respectively. All studies except for 1 study (58) used a randomization method, and 12 studies (35, 38, 39, 41, 47, 51, 53, 63, 66, 80, 85, 86) did not provide details for the generation of random sequences. All but 8 studies (39, 47, 58, 62, 64, 78, 80, 86) had a low ROB in allocation concealment. In terms of blind assessments, 4 studies (39, 58, 62, 86) were considered high-risk. In 17 studies (36, 42-44, 48, 50, 52, 54-56, 60, 61, 65, 71, 72, 79, 81), outcome assessors were blind to subjects and personnel. With regard to selective result reporting, all studies except for 11 (35, 38, 41, 59, 62-64, 77, 78, 80, 83) were preregistered in a clinical study registry, which might have effectively controlled RB. Two studies (52, 53) were deemed to have a high risk of incomplete outcome data due to a dropout rate exceeding 20%. No other biases were identified.

Figure 2.

Risk of bias graph for included RCTs.

Findings From Meta-Analyses

Effects on Glucose Homeostasis

Fasting plasma glucose

Effects of MTTs on FPG are provided in Fig. 3A and Table 2. FPG was assessed in 53 study groups, including 3053 subjects. The result showed a significant decrease in FPG levels after MTTs intervention in comparison with controls (mean difference [MD] = −7.97 mg/dL [95% CI = −10.82, −5.12]; P < .00001) (Fig. 3A). The heterogeneity was high among these studies (I² = 51%, P < .0001). No meta-analysis indicated evidence of significant BSH. Subgroup analyses revealed significant differences in terms of the Mediterranean region, the intervention of pro/prebiotics, number of probiotics strains, probiotics dose < 10¹⁰ CFU, prebiotics dose ≥ 10 g, study duration, baseline BMI ≥ 28 kg/m², and mean age (Table 2). Most meta-analyses showed no heterogeneity (I² = 50%, P ≥ .05) (Table 2).

Figure 3.

Forest plots of the effect of gut microbiome–targeted therapies on fasting plasma glucose (a), 2-hour postprandial blood glucose (b), fasting insulin (c), HbA1c (d) and HOMA-IR (e).

Table 2.

Pooled mean estimates of random-effects meta-analysis on glycemic markers overall and subgroups analysis

Subgroup	FBG (mg/dl)					PBG (mg/dl)					Fasting C-peptide (ng/ml)					Fasting insulin (uU/ml)
	Study	MD [95% CI]	P for random effect	P for heterogeneity	I 2	Study	MD [95% CI]	P for random effect	P for heterogeneity	I 2	Study	MD [95% CI]	P for random effect	P for heterogeneity	I 2	Study	MD [95% CI]	P for random effect	P for heterogeneity	I 2
Total	53	−7.97 [−10.82, −5.12]	<.00001	<.0001	51%	6	−43.30 [−75.83, −10.77]	.009	<.00001	99%	9	.02 [−0.13, 0.17]	0.77	0.0004	72%	29	−1.73 [−2.63, −0.84]	.0001	<.00001	80%
Region
Europe	6	6.51 [−13.78, 26.80]	.53	.14	39%	—	—	—	—	—	3	−.10 [−.48, .28]	.62	.85	0%	2	−1.85 [−6.89, 3.20]	.47	.57	0%
Mediterranean	28	−14.95 [−18.48, −11.41]	<.00001	.56	0%	1	−5.21 [−80.89, 70.47]	.89	—	—	—	—	—	—	—	18	−1.95 [−2.94, −0.96]	.0001	<.00001	72%
Southeast Asia	6	−9.87 [−20.36, 0.62]	.07	.13	41%	2	−7.51 [−22.33, 7.31]	.32	.08	68%	—	—	—	—	—	1	−4.00 [−8.39, 0.39]	.07	—	—
The Americas	4	0.10 [−14.40, 14.60]	.99	.38	2%	—	—	—	—	—	—	—	—	—	—	1	−0.44 [−4.41, 3.53]	.83	—	—
Western Pacific	10	−0.94 [−2.21, 0.34]	.15	.77	0%	3	−72.48 [−109.04, −35.93]	.0001	<.00001	99%	6	.04 [−0.13, .21]	.64	<.0001	82%	7	−.80 [−3.11, 1.51]	.50	<.00001	87%
Intervention type
Probiotic	29	−7.88 [−12.80, −2.97]	.002	<.00001	67%	3	−37.46 [−79.93, 5.01]	.08	<.00001	100%	6	.02 [-.15, .19]	.83	<.0001	82%	14	-1.78 [-3.14, -0.42]	.01	<.00001	80%
Prebiotic	9	-14.94 [-19.59, -10.29]	<.00001	.84	0%	1	-37.46 [-79.93, 5.01]	.89	—	—	—	—	—	—	—	5	-2.46 [-4.72, -0.19]	.03	.0005	80%
Synbiotic	13	-3.33 [-9.55, 2.89]	.29	.42	0%	1	-17.70 [-34.84, -0.56]	.04	—	—	2	.00 [-.43, .43]	.98	.95	0%	8	-1.53 [-2.98, -0.08]	.04	.19	29%
FMT	2	-19.35 [-44.79, 6.08]	.14	.49	0%	1	-150.66 [-215.17, -86.15]	<.00001	—	—	1	.33 [-.97, 1.63]	.62	—	—	2	-3.33 [-14.06, 7.40]	.54	.80	0%
No. of probiotic strains
Multi strains	26	-4.15 [-7.74, -0.55]	.02	.02	40%	4	-57.48 [-88.69, -26.27]	.0003	<.00001	98%	7	.00 [-.18, .18]	.98	.0001	78%	17	-1.68 [-3.06, -0.31]	.02	<.00001	76%
Single strain	15	-9.83 [-17.03, -2.63]	.007	.24	19%	—	—	—	—	—	2	.10 [-.11, .31]	.36	.95	0%	7	-1.64 [-2.83, -0.45]	.007	.25	24%
Probiotics dose
≥ 1010 CFU	18	-0.05 [-0.12, 0.02]	.17	.03	41%	2	-55.26 [-94.77, -15.75]	.006	<.00001	99%	6	.02 [-.15, .19]	.83	<.0001	82%	9	-.96 [-2.85, .92]	.32	<.00001	85%
< 1010 CFU	16	-9.08 [-14.84, -3.32]	.002	.22	21%	—	—	—	—	—	1	.00 [-.45, .45]	1	—	—	10	-1.69 [-2.64, -.75]	.0004	.32	14%
Prebiotics dose
≥ 10g	10	-14.20 [-18.64, -9.75]	<.00001	.88	0%	1	-5.21 [-80.89, 70.47]	.89	—	—	—	—	—	—	—	6	-1.97 [-3.78, -0.17]	.03	<.0001	81%
< 10g	10	-3.72 [-10.82, 3.38]	.30	.25	21%	1	-17.70 [-34.84, -0.56]	.04	—	—	2	-.20 [-.82, .41]	.51	.70	0%	7	-2.17 [-4.19, -.15]	.04	.26	23%
Study duration
≥ 12 weeks	23	-1.29 [-2.39, -0.19]	.02	.23	17%	3	-37.46 [-79.93, 5.01]	.08	<.00001	100%	6	.04 [-.13, .21]	.67	<.0001	82%	13	-1.78 [-3.64, .08]	.06	<.00001	81%
< 12 weeks	30	-13.24 [-16.86, -9.61]	<.00001	.05	31%	3	-56.24 [-139.70, 27.23]	.19	.0004	87%	3	-.07 [-.45, .31]	.71	.71	0%	17	-1.97 [-2.92, -1.02]	<.0001	.006	53%
Baseline BMI
≥ 28 Kg-m2	33	-12.02 [-17.32, -6.72]	<.00001	<.00001	61%	1	-5.21 [-80.89, 70.47]	.89	—	—	5	-.06 [-.36, .23]	.66	.42	0%	17	-2.26 [-3.43, -1.10]	.0001	<.00001	75%
< 28 Kg-m2	18	-1.61 [-3.70, 0.47]	.13	.32	12%	3	-72.48 [-109.04, -35.93]	.0001	<.00001	99%	4	.05 [-.13, .23]	.59	<.0001	88%	11	-.31 [-1.29, 0.67]	.53	<.00001	77%
Mean age
years ≥ 52y	36	-3.48 [-6.46, -0.51]	.02	.08	26%	3	-72.48 [-109.04, -35.93]	.0001	<.00001	99%	9	.02 [-.13, .17]	.77	.0004	72%	19	-1.21 [-2.52, .09]	.07	<.00001	74%
years < 52y	15	-15.16 [-22.30, -8.02]	<.0001	<.00001	73%	2	-1.91 [-5.83, 2.01]	.34	.93	0%	—	—	—	—	—	8	-1.91 [-3.36, -.45]	.01	<.00001	86%

Subgroup	HbA1c (%)					HOMA-IR					HOMA-β					QUICKI
	Study	MD [95% CI]	P for random effect	P for heterogeneity	I 2	Study	MD [95% CI]	P for random effect	P for heterogeneity	I 2	Study	MD [95% CI]	P for random effect	P for heterogeneity	I 2	Study	MD [95% CI]	P for random effect	P for heterogeneity	I 2
Total	42	-0.28 [-0.39, -0.17]	<.00001	<.00001	76%	31	-0.53 [-0.85, -0.20]	.002	<.00001	76%	11	-1.94 [-4.18, .29]	.09	.08	40%	13	0.00 [-0.01, 0.02]	.52	<.00001	81%
Region
Europe	6	-0.01 [-0.27, 0.25]	.95	.90	0%	3	-0.56 [-2.26, 1.13]	.52	.04	69%	1	8.70 [-31.52, 48.92]	.67	—	—	—	—	—	—	—
Mediterranean	19	-0.37 [-0.55, -0.19]	<.0001	<.00001	74%	17	-0.92 [-1.41, -0.42]	.0003	<.00001	75%	5	-2.95 [-7.24, 1.34]	.18	.06	56%	10	0.00 [-0.02, 0.02]	.80	<.00001	85%
Southeast Asia	4	-1.00 [-1.35, -0.64]	<.00001	.65	0%	1	-0.89 [-2.37, 0.59]	.24	—	—	1	-14.90 [-29.99, 0.19]	.05	—	—	1	0.01 [-0.00, 0.03]	.07	—	—
The Americas	4	-0.48 [-0.80, -0.17]	.002	.67	0%	3	-0.04 [-0.96, 0.88]	.93	.59	0%	—	—	—	—	—	—	—	—	—	—
Western Pacific	9	-0.03 [-0.13, 0.06]	.49	.06	46%	7	-0.14 [-0.79, 0.52]	.68	<.00001	86%	4	-1.59 [-3.99, 0.81]	.19	.57	0%	2	-0.00 [-0.04, 0.03]	.95	.16	50%
Intervention type
Probiotic	22	-0.19 [-0.31, -0.08]	.0008	.0008	56%	16	-0.63 [-1.16, -0.09]	.02	<.00001	83%	6	-0.81 [-2.48, 0.87]	.35	.21	30%	9	0.02 [0.00, 0.03]	.004	.007	62%
Prebiotic	9	-0.56 [-0.70, -0.41]	<.00001	.61	0%	4	-1.04 [-2.40, 0.32]	.14	.004	77%	1	8.70 [-31.52, 48.92]	.67	—	—	1	-0.13 [-0.18, -0.08]	<.00001	—	—
Synbiotic	9	-0.24 [-0.53, 0.05]	.11	<.00001	87%	9	-0.31 [-0.77, 0.15]	.19	.02	57%	3	-9.10 [-15.33, -2.87]	.004	.71	0%	3	0.00 [-0.02, 0.02]	.91	.07	62%
FMT	2	0.32 [-0.70, 1.35]	.54	.71	0%	2	-0.39 [-1.91, 1.12]	.61	.41	0%	1	28.12 [-52.44, 108.68]	.49	—	—	—	—	—	—	—
No. of probiotic strains
Multi strains	19	-0.16 [-0.31, -0.02]	.03	<.00001	78%	19	-0.61 [-1.09, -0.12]	.01	<.00001	75%	8	-1.35 [-3.69, 0.99]	.26	.13	37%	9	0.02 [0.01, 0.02]	<.0001	.09	42%
Single strain	12	-0.27 [-0.35, -0.19]	<.00001	.64	0%	8	-0.44 [-1.02, 0.15]	.14	<.0001	77%	2	-4.23 [-10.54, 2.09]	.19	.14	55%	3	-0.00 [-0.03, 0.03]	.98	.02	75%
Probiotics dose
≥ 1010 CFU	16	-0.09 [-0.20, 0.01]	.08	.03	43%	12	-0.48 [-1.07, 0.12]	.12	<.00001	82%	5	0.00 [-0.14, 0.15]	.96	.50	0%	5	0.00 [-0.05, 0.05]	.94	.13	44%
< 1010 CFU	9	-0.26 [-0.38, -0.13]	<.0001	.21	27%	9	-0.54 [-1.08, 0.00]	.05	<.0001	76%	2	-4.23 [-10.54, 2.09]	.19	.14	55%	4	0.01 [-0.01, 0.03]	.34	<.00001	89%
Prebiotics dose
≥ 10g	10	-0.42 [-0.56, -0.29]	<.00001	.07	44%	5	-0.51 [-1.11, 0.09]	.10	<.0001	86%	—	—	—	—	—	1	-0.13 [-0.18, -0.08]	<.00001	—	—
< 10g	6	-0.17 [-0.79, 0.46]	.60	<.00001	88%	6	-0.78 [-1.58, 0.01]	.05	.23	28%	2	-7.91 [-14.75, -1.07]	.02	.98	0%	3	0.01 [-0.00, 0.02]	.06	.81	0%
Study duration
≥ 12 weeks	17	-0.14 [-0.25, -0.03]	.01	.001	59%	13	-0.26 [-0.77, 0.25]	.32	<.00001	76%	6	-1.86 [-4.23, 0.52]	.13	.45	0%	6	0.02 [0.01, 0.03]	.0006	.03	59%
< 12 weeks	25	-0.40 [-0.58, -0.22]	<.0001	<.00001	72%	18	-0.87 [-1.35, -0.39]	.0004	<.00001	75%	5	-3.02 [-7.41, 1.36]	.18	.05	58%	7	-0.01 [-0.04, 0.01]	.34	<.00001	82%
Baseline BMI
≥ 28 Kg-m2	28	-0.32 [-0.46, -0.19]	<.00001	.04	35%	18	-1.04 [-1.48, -0.59]	<.00001	.02	47%	4	-1.67 [-5.13, 1.80]	.35	.14	46%	8	-0.00 [-0.02, 0.02]	.93	<.00001	88%
< 28 Kg-m2	13	-0.15 [-0.30, -0.01]	.04	<.00001	84%	11	0.08 [-0.27, 0.43]	.65	<.0001	72%	5	-1.94 [-4.32, 0.44]	.11	.45	0%	3	0.02 [0.01, 0.03]	<.00001	.98	0%
Mean age
years ≥ 52y	26	-0.15 [-0.26, -0.04]	.008	<.00001	73%	21	-0.33 [-0.67, 0.02]	.06	<.00001	72%	8	-2.35 [-4.41, -0.30]	.02	.46	0%	7	0.01 [0.00, 0.02]	.04	<.0001	79%
years < 52y	14	-0.52 [-0.65, -0.39]	<.00001	.90	0%	8	-0.81 [-1.64, 0.01]	.05	.0006	73%	1	26.67 [-22.28, 75.62]	.29	—	—	4	-0.04 [-0.15, 0.08]	.51	.003	78%

2-hour postprandial blood glucose

Effects of MTTs on 2hPPG are provided in Fig. 3B and Table 2. The 2hPBG was assessed in 6 studies, including 644 subjects. Data were pooled using a random-effects model. The results revealed that compared with the placebo, the 2hPBG level was lower after probiotic and synbiotic supplementation (MD = −43.30 mg/dL, [95% CI = −75.83, −10.77]; P = .009) (Fig. 3B). The conclusion remained the same with respect to the Western Pacific region, intervention of probiotics and multiple strains, prebiotics dose ≥ 10 g, study duration, baseline BMI < 28 kg/m², and mean age ≥ 52 years (Table 2). The interpretation of most subgroup analyses was limited due to the low number of studies. Most meta-analyses presented significant BSH (I² ≥ 50%, P < .05) (Table 2).

Fasting C-peptide

Effects of MTTs on fasting C-peptide are provided in Supplementary Fig. S3A (33) and Table 2. In the pooled meta-analyses, 9 studies (773 subjects) were included, and the result showed no effect of MTTs on fasting C-peptide levels when compared with the placebo (MD = 0.02 ng/mL [95% CI = −0.13, 0.17], P = .77; I² = 72%, P = .0004) (Supplementary Fig. S3A) (33). The interpretation of most subgroup analyses was limited due to the low number of studies. Most meta-analyses presented significant BSH (I² ≥ 50%, P < .05) (Table 2).

Fasting insulin

Effects of MTTs on fasting insulin are provided in Fig. 3C and Table 2. It was reported by 29 studies. Compared with the placebo, the fasting insulin levels were significantly reduced with MTTs (MD = −1.73uU/mL [95% CI = −2.63, −0.84], P = .0001; I² = 80%, P = .00001) (Fig. 3C). The pooled result remained consistent in all subgroups, even when subgroup analyses were based on the Mediterranean and Western Pacific regions, administration of prebiotics/probiotics and multiple strains, probiotics dose ≥ 10¹⁰ CFU, prebiotics dose ≥ 10 g, study duration ≥12 weeks, BMI, and mean age (Table 2). Most meta-analyses presented significant BSH (I² ≥ 50%, P < .05) (Table 2).

HbA1c

Effects of MTTs on HbA1c are provided in Fig. 3D and Table 2. It was reported by 42 studies (2412 subjects). MTTs were superior in reducing HbA1c than placebo (MD = −0.28% [95% CI = −0.39, −0.17]; P < .00001, I² = 76%, P < .00001) (Fig. 3D). Subgroup analyses showed that the inference was similar in most groups, except for the Europe and Western Pacific regions. This was observed when using synbiotics/FMT and consuming probiotics at a dose ≥ 10¹⁰ CFU and prebiotics at a dose < 10 g (Table 2). Additionally, almost half of the meta-analyses presented significant BSH (I² ≥ 50%, P < .05) (Table 2).

HOMA-IR

Effects of MTTs on HOMA-IR are provided in Fig. 3E and Table 2. It was reported by 31 studies (1781 subjects). The HOMA-IR decreased more after MTT intervention in comparison with the placebo (MD = −0.53 [95% CI = −0.85, −0.20]; P = .0002, I² = 76%, P < .00001) (Fig. 3E). Subgroup analyses showed that the inference was similar in most groups, except for the American regions. This was observed when using FMT and consuming prebiotics at a dose < 10 g (Table 2).

HOMA-β

Effects of MTTs on HOMA-β are provided in Supplementary Fig. S3B (33) and Table 2. It was reported by 11 studies (812 subjects) that there was no significant difference in HOMA-β following MTT intervention in comparison with the placebo (MD = −0.00 [95% CI = −0.15, 0.14], P = .09; I² = 40%, P = .08) (Supplementary Fig. S3B) (33). Subgroup analyses showed that the inference was similar in most groups, except for using FMT, consuming prebiotics at a dose < 10 g and mean age ≥ 52 years (Table 2). Most meta-analyses showed no heterogeneity (I² < 50%, P ≥ .05) (Table 2).

QUICKI

Effects of MTTs on QUICKI are provided in Supplementary Fig. S3C (33) and Table 2. It was reported by 13 studies (692 subjects) that there was no significant difference in QUICKI after MTTs intervention in comparison with the placebo (MD = 0.00 [95% CI = −0.01, 0.02], P = .52; I² = 81%, P < .00001) (Supplementary Fig. S3C) (33). Subgroup analyses revealed that the higher QUICKI level favored intake of probiotics, multiple strains, and a prebiotics dose ≥ 10 g (Table 2). The interpretation of most subgroup analyses was limited due to the lower number of studies. Analyses of BSH showed that the prebiotics dose was the source of heterogeneity (Table 2).

Effects on Inflammation Markers

A variety of inflammation markers were studied, including hs-CRP, TNF-α, IFN-γ, LPS, IL-1β, IL-6, IL-10, and IL-8. MTT intervention led to a slight decrease in hs-CRP levels (SMD [95% CI] = −0.36 [−0.51, −0.21]; P < .00001; I² = 49%, P = .002; n = 29 studies) (Fig. 4A). Subgroup analyses showed that the pooled effect remained consistent across the majority of subgroups, excluding those with intake of a probiotics dose ≥ 10¹⁰ CFU, prebiotics dose < 0 g, and study duration ≥12 weeks (Table 3). The studies with a duration < 12 weeks showed greater and statistically significant reductions than those with a duration ≥ 12 weeks. A reduction of TNF-α was observed after MTT intervention compared to the control group (SMD [95% CI] = −0.86 [−1.40, −0.31]; P = .002; I² = 90%, P = .00001; n = 12 studies) (Fig. 4B). Stratified analyses revealed that TNF-α maintained a decrease in categories of the Mediterranean and Southeast Asia regions, intervention of probiotics, single train probiotics, prebiotics dose ≥ 10 g, study duration < 12 weeks, baseline BMI < 28 kg/m², and mean age of ≥ 52 years old (Table 3). The result revealed a significant decrease in the LPS level after MTT intervention in comparison with controls (SMD [95% CI] = −1.01 [−1.49, −0.53]; P < .00001; I² = 71%, P = .004; n = 6 studies) (Fig. 4C). The interpretation of most subgroup analyses was limited due to the low number of studies (Table 3). MTTs were superior to the controls in increasing IL-10 levels (SMD [95% CI] = 0.62 [0.25, 1.00]; P = .001; I² = 50%, P = .09; n = 5 studies) (Fig. 4D). Subgroup analyses showed that IL-10 levels increased in the group with prebiotics supplementation, probiotics dose ≥ 10¹⁰ CFU, prebiotics dose, study duration < 12 weeks, and age < 52 years. The IL-10 level was significantly increased by a probiotics dose ≥ 10¹⁰ CFU compared with a dose < 10¹⁰ CFU (Table 3). However, compared with the controls, MTT supplementation did not decrease the levels of IFN-γ (SMD [95% CI] = −0.18 [−0.66, 0.30]; P = .46; I² = 70%, P = .02) (Supplementary Fig. S4A) (33), IL-1β (SMD [95% CI] = −0.22 [−0.63, 0.18]; P = .28; I² = 61%, P = .04) (Supplementary Fig. S4B) (33), IL-6 (SMD [95% CI] = −0.21 [−0.45, 0.03]; P = .08; I² = 56%, P = .009) (Supplementary Fig. S4C) (33), and IL-8 (SMD [95% CI] = −0.01 [−0.36, 0.35]; P = .97; I² = 0%, P = .68) (Supplementary Fig. S4D) (33).

Figure 4.

Forest plots of the effect of gut microbiome–targeted therapies on hs-CRP (a), TNF-α (b), LPS (c), and IL-10 (d).

Table 3.

Pooled mean estimates of random-effects meta-analysis on inflammation markers overall and subgroup analysis

Subgroup	hs-CRP					TNF-α					IFN-γ					LPS
	Study	SMD [95% CI]	P value	P for heterogeneity	I 2	Study	SMD [95% CI]	P value	P for heterogeneity	I 2	Study	SMD [95% CI]	P value	P for heterogeneity	I 2	Study	SMD [95% CI]	P value	P for heterogeneity	I 2
Total	29	−0.36 [−0.51, −0.21]	<.00001	.002	49%	12	−0.86 [−1.40, −0.31]	.002	<.00001	90%	4	−0.18 [−0.66, 0.30]	.46	.02	70%	6	−1.01 [−1.49, −0.53]	<.0001	.004	71%
Region
Europe	3	−0.04 [−0.43, 0.35]	.84	.95	0%	3	0.08 [−0.62, 0.77]	.83	.008	80%	2	0.21 [−0.15, 0.56]	.25	.35	0%	—	—	—	—	—
Mediterranean	21	−0.50 [−0.66, −0.34]	<.00001	.06	35%	6	−1.56 [−2.46, −0.65]	.0007	<.00001	91%	2	−0.57 [−0.96, −0.19]	.004	.42	0%	4	−1.28 [−1.71, −0.85]	<.00001	.09	53%
Southeast Asia	1	−0.25 [−0.90, 0.41]	.46	—	—	1	−1.43 [−2.17, −0.69]	.0002	—	—	—	—	—	—	—	1	−0.83 [−1.52, −0.15]	.02	—	—
The Americas	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Western Pacific	4	0.11 [−0.12, 0.33]	.36	.72	0%	2	−0.14 [−0.52, 0.23]	.45	.34	0%	—	—	—	—	—	1	0.14 [−0.63, 0.91]	.72	—	—
Intervention type
Probiotic	16	−0.20 [−0.36, −0.05]	.01	.28	15%	7	−0.88 [−1.75, −0.02]	.05	<.00001	93%	2	0.21 [−0.15, 0.56]	.25	.35	0%	1	−0.83 [−1.52, −0.15]	.02	—	—
Prebiotic	6	−0.67 [−1.10, −0.24]	.002	.005	71%	4	−1.08 [−1.54, −0.61]	<.00001	.05	61%	2	−0.57 [−0.96, −0.19]	.004	.42	0%	4	−1.28 [−1.71, −0.85]	<.00001	.09	53%
Synbiotic	7	−0.39 [−0.65, −0.12]	.004	.14	38%	1	−0.18 [−0.77, 0.41]	.55	—	—	—	—	—	—	—	1	0.14 [−0.63, 0.91]	.72	—	—
FMT	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
No. of probiotic strains
Multi strains	8	−0.31 [−0.58, −0.04]	.02	.03	54%	2	0.20 [−0.85, 1.25]	.71	.004	88%	2	0.21 [−0.15, 0.56]	.25	.35	0%	1	0.14 [−0.63, 0.91]	.72	—	—
Single strain	12	−0.17 [−0.35, −0.00]	.05	.74	0%	5	−1.41 [−2.62, −0.21]	.02	<.00001	94%	—	—	—	—	—	1	−0.83 [−1.52, −0.15]	.02	—	—
Probiotics dose
≥ 1010 CFU	7	−0.15 [−0.46, 0.16]	.34	.06	51%	4	−0.23 [−1.00, 0.55]	.57	<.00001	88%	2	0.21 [−0.15, 0.56]	.25	.35	0%	2	−0.36 [−1.32, 0.59]	.46	.06	71%
< 1010 CFU	10	−0.23 [−0.40, −0.05]	.01	.64	0%	2	−3.16 [−9.07, 2.75]	.29	<.00001	98%	—	—	—	—	—	—	—	—	—	—
Prebiotics dose
≥ 10g	8	−0.65 [−0.97, −0.33]	<.0001	.02	60%	4	−1.08 [−1.54, −0.61]	<.00001	.05	61%	2	−0.57 [−0.96, −0.19]	.004	.42	0%	4	−1.28 [−1.71, −0.85]	<.00001	.09	53%
< 10g	4	−0.34 [−0.76, 0.08]	.11	.05	61%	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Study duration
≥ 12 weeks	12	−0.14 [−0.33, 0.04]	.12	.2	25%	4	−0.45 [−1.01, 0.11]	.11	.01	71%	—	—	—	—	—	2	−0.36 [−1.32, 0.59]	.46	.06	71%
< 12 weeks	17	−0.50 [−0.69, −0.31]	<.00001	.03	44%	8	−1.10 [−1.90, −0.30]	.007	<.00001	93%	4	−0.18 [−0.66, 0.30]	.46	.02	70%	4	−1.28 [−1.71, −0.85]	<.00001	.09	53%
Baseline BMI
≥ 28 Kg-m2	19	−0.36 [−0.56, −0.15]	.0006	.0005	60%	7	−0.59 [−1.18, 0.00]	.05	<.00001	87%	4	−0.18 [−0.66, 0.30]	.46	.02	70%	5	−1.04 [−1.61, −0.47]	.0004	.002	77%
< 28 Kg-m2	10	−0.35 [−0.56, −0.15]	.0006	.29	17%	5	−1.41 [−2.60, −0.21]	.02	<.00001	94%	—	—	—	—	—	1	−0.83 [−1.52, −0.15]	.02	—	—
Mean age
years ≥ 52y	20	−0.24 [−0.39, −0.10]	.001	.16	24%	8	−0.76 [−1.51, −0.02]	.04	<.00001	91%	2	0.21 [−0.15, 0.56]	.25	.35	0%	2	−0.36 [−1.32, 0.59]	.46	.06	71%
years < 52y	9	−0.59 [−0.90, −0.27]	.0003	.007	62%	4	−1.08 [−1.54, −0.61]	<.00001	.05	61%	2	−0.57 [−0.96, −0.19]	.004	.42	0%	4	−1.28 [−1.71, −0.85]	<.00001	.09	53%

Subgroup	IL-1β					IL-6					IL-10
	Study	SMD [95% CI]	P value	P for heterogeneity	I 2	Study	SMD [95% CI]	P value	P for heterogeneity	I 2	Study	SMD [95% CI]	P value	P for heterogeneity	I 2
Total	5	−0.22 [−0.63, 0.18]	.28	.04	61%	12	−0.21 [−0.45, 0.03]	.08	.009	56%	5	0.62 [0.25, 1.00]	.001	.09	50%
Region
Europe	2	0.07 [−0.77, 0.91]	.87	.02	82%	2	−0.05 [−0.40, 0.31]	.8	.33	0%	—	—	—	—	—
Mediterranean	1	−0.43 [−1.00, 0.15]	.15	—	—	6	−0.30 [−0.78, 0.17]	.21	.002	74%	3	0.62 [0.21, 1.03]	.003	.2	38%
Southeast Asia	1	−0.41 [−1.07, 0.25]	.22	—	—	1	−0.46 [−1.12, 0.20]	.18	—	—	1	1.19 [0.48, 1.91]	.001	—	—
The Americas	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Western Pacific	1	−0.52 [−1.13, 0.08]	.09	—	—	3	−0.10 [−0.38, 0.18]	.49	.41	0%	1	0.16 [−0.44, 0.75]	.61	—	—
Intervention type
Probiotic	4	−0.18 [−0.68, 0.32]	.48	.02	68%	7	0.01 [−0.20, 0.22]	.93	.65	0%	2	0.65 [−0.36, 1.67]	.21	.03	79%
Prebiotic	—	—	—	—	—	3	−0.69 [−1.35, −0.03]	.04	.01	77%	3	0.62 [0.21, 1.03]	.003	.2	38%
Synbiotic	1	−0.43 [−1.00, 0.15]	.15	—	—	2	−0.25 [−0.60, 0.09]	.15	.63	0%	—	—	—	—	—
FMT	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
No. of probiotic strains
Multi strains	3	−0.09 [−0.69, 0.52]	.78	.02	74%	4	−0.11 [−0.36, 0.15]	.41	.44	0%	—	—	—	—	—
Single strain	2	−0.47 [−0.92, −0.03]	.04	.81	0%	4	0.01 [−0.28, 0.30]	.93	.41	0%	2	0.65 [−0.36, 1.67]	.21	.03	79%
Probiotics dose
≥ 1010 CFU	4	−0.16 [−0.63, 0.32]	.52	.03	66%	5	−0.02 [−0.26, 0.23]	.9	.42	0%	1	1.19 [0.48, 1.91]	.001	—	—
< 1010 CFU	1	−0.52 [−1.13, 0.08]	.09	—	—	2	−0.20 [−0.54, 0.15]	.26	.37	0%	1	0.16 [−0.44, 0.75]	.61	—	—
Prebiotics dose
≥ 10g	—	—	—	—	—	3	−0.69 [−1.35, −0.03]	.04	.01	77%	3	0.62 [0.21, 1.03]	.003	.2	38%
< 10g	—	—	—	—	—	1	−0.31 [−0.74, 0.11]	.15	—	—	—	—	—	—	—
Study duration
≥ 12 weeks	2	−0.47 [−0.92, −0.03]	.04	.81	0%	4	−0.15 [−0.41, 0.11]	.25	.44	0%	2	0.65 [−0.36, 1.67]	.21	.03	79%
< 12 weeks	3	−0.09 [−0.69, 0.52]	.78	.02	74%	8	−0.24 [−0.60, 0.12]	.19	.002	68%	3	0.62 [0.21, 1.03]	.003	.2	38%
Baseline BMI
≥ 28 Kg-m2	2	0.07 [−0.77, 0.91]	.87	.02	82%	7	−0.32 [−0.68, 0.05]	.09	.003	70%	3	0.62 [0.21, 1.03]	.003	.2	38%
< 28 Kg-m2	3	−0.46 [−0.81, −0.10]	.01	.96	0%	5	−0.04 [−0.30, 0.22]	.78	.66	0%	2	0.65 [−0.36, 1.67]	.21	.03	79%
Mean age
years ≥ 52y	4	−0.18 [−0.68, 0.32]	.48	.02	68%	8	−0.09 [−0.28, 0.09]	.33	.7	0%	2	0.65 [−0.36, 1.67]	.21	.03	79%
years < 52y	1	−0.43 [−1.00, 0.15]	.15	-	-	4	−0.45 [−1.11, 0.20]	.18	.001	81%	3	0.62 [0.21, 1.03]	.003	.2	38%

Abbreviations: BMI, body mass index; CFU, colony-forming units.

Effects on Gut Microbiota

α-Diversity of gut microbiota

All studies used fecal samples for analyses. In 8 studies (41, 42, 53, 56, 58, 77, 86, 87) (Supplementary Table S3) (33), the vast majority of studies found no differences in α-diversity between intervention and control groups. However, some studies noted a decrease in the ACE index and an increase in the Simpson index after the intervention compared to treatment baseline. Only 2 studies provided specific data on the Shannon index. There was no significant difference in the SMD of Shannon Index between MTT intervention groups and the controls (SMD [95% CI] = −0.12 [−0.47, 0.23]; P = .49; I² = 0%, P = .74) ((Fig. 5).

Figure 5.

Forest plots of the effect of gut microbiome–targeted therapies on Shannon Index.

β-Diversity of gut microbiota

Seven studies (41, 42, 53, 56, 77, 86, 87) reported data on β-diversity (Supplementary Table S3) (33). Of these, 2 studies reported results using the weighted or unweighted Unifrac method to calculate distance, 3 studies used the Bray-Curtis method, 1 study used the Euclidean method, and 1 study did not mention the method (Supplementary Table S3) (33). The findings were inconsistent. Only one of the included studies reported significant differences between MTTs group and control group in β-diversity.

Summary of gut microbiota taxa

A total of 14 studies examined the changes in GM between MTT intervention groups and nonintervention groups (Supplementary Table S3) (33) at the phylum, family, genus, or species levels. A high inconsistency among studies was observed. Additionally, a meta-analysis was conducted for the reported GM taxa at the phylum, family, genus, or species levels (Supplementary Fig. S5-S7) (33). At the phylum level, Actinobacteria was significantly enriched (SMD [95% CI] = 1.10 [0.69, 1.50]; P < .00001; I² = 12%, P = .29), while no significant difference was observed in Bacteroidetes, Firmicutes, and Proteobacteria (P values ranged from 0.17 to 0.74). At the family level, the meta-analysis was only performed for Enterobacteriaceae flora, and no significant differences were found. At the genus level, only 2 out of 21 genera showed significant changes. Lactobacillus (SMD [95% CI] = 2.04 [1.11, 2.97]; P < .00001; I² = 82%, P = .02) and Lactobacillus casei subgroup (SMD [95% CI] = 0.44 [0.11, 0.77]; P < .0001; I² = 52%, P = .08) were enriched. At the species level, a meta-analysis was performed for 2 flora (Bacteroides vulgatus and Eubacterium hallii), and neither of them was observed to be significantly different.

Sensitivity Analysis and Publication Bias

Sensitivity analyses resulted in no material difference in SMDs of FPG, 2hPBG, fasting insulin, HbA1c, HOMA-IR, QUICKI, fasting C-peptide, hs-CRP, TNF-α, LPS, IFN-γ, IL-10, Bifidobacterium, Firmicutes, Streptococcus, and Akkermansia muciniphila (Supplementary Fig. S8-S10) (33). For 2hPBG, there was no statistical difference (MD = −27.18 mg/dL [95% CI = −78.48, 24.11]; P = .30; I² = 99%, P < .0001) (Supplementary Fig. S8F) (33) after excluding 2 study groups. However, the sensitivity analysis conducted by the excluding 1 study (36) (Supplementary Fig. S8G) (33) showed a significant increase in QUICKI (MD [95% CI] = 0.01 [0.00, 0.02]; P = .03; I² = 70%, P = .0001). Similarly, it showed a significant increase in fasting C-peptide (MD = 0.13 ng/mL [95% CI = 0.07, 0.19]; P < .0001; I² = 0%, P < .0001) after excluding 1 study group (Supplementary Fig. S8A). For IL-6, a significant decrease in pooled SMD was observed (SMD [95% CI] = −0.25 [−0.50, −0.01]; P = .04; I² = 56%, P = .01) after excluding one study (47) (Supplementary Fig. S9F) (33). Similarly, the analysis of IL-1β showed a decrease after removing 1 study (SMD [95% CI] = −0.22 [−0.63, 0.18]; P = .004; I² = 0%, P = .99) (Supplementary Fig. S9E). PB was assessed using funnel plots (Supplementary Fig. S11) (33) and Egger's test. According to the Cochrane handbook, a minimum of 10 studies is required to draw clear conclusions about PB. Only reports on TNF-α (P = .000) (Supplementary Fig. S11I) (33), hs-CRP (P = .041) (Supplementary Fig. S11H), HOMA-IR (P = .002), and HbA1c (P = .003) (Supplementary Fig. S11D) (33) showed significant PB (31), while studies on other biomarkers did not.

Discussion

Due to the increasing preclinical evidence supporting a strong correlation between GM and T2DM, there has been a growing interest in the use of MTTs for treating T2DM. This meta-analysis includes 60 study comparisons from 54 distinct clinical studies and provides a comprehensive assessment of the effect of 4 common types of gut MTTs: probiotics, prebiotics, synbiotics, and FMT. The analysis covers a wide array of indices, including glucose, insulin homeostasis, systemic inflammation, and GM, among 3390 T2DM study patients. The findings suggest that MTTs are associated with modest improvements in glucose and insulin homeostasis and potential anti-inflammatory effects. Limited but consistent evidence indicates a selective enrichment of certain beneficial bacterial taxa. However, the causal relationship and clinical significance of these changes remain to be fully established (Fig. 6).

Figure 6.

Graphical summary of gut microbiome–targeted therapies on glucose homeostasis, inflammation, and gut microbiota in type 2 diabetes. This figure was created using Figdraw (https://www.figdraw.com/static/index.html).

Regarding glucose homeostasis, the results were consistent with previous meta-analyses but also presented some contrasting trends. The overall changes observed in FPG, 2hPBG, and HbA1c aligned with previous findings in T2DM patients who received pro-/pre-/synbiotics supplementation (88). This suggested that MTTs could improve glucose metabolism. In the RCTs reviewed by this study, microecological modulators showed significantly lowering effects on fasting insulin and HOMA-IR, which is encouraging and in line with prior studies (89-91). However, it was found that MTTs did not bring any significant changes in fasting C-peptide, HOMA-β, and QUICKI. It is noteworthy that all subgroups displayed nonsignificant changes in C-peptide levels regardless of the variable. Previous studies have shown that insulin therapy can restore β cell secretory capacity by eliminating glucolipotoxicity and reducing beta-cell dedifferentiation (92, 93). Due to the heterogeneity among the studies reporting changes in C-peptide levels, it is possible that the lack of change in C-peptide levels may have resulted from different insulin regimens administered to the enrolled patients. Additionally, previous meta-analyses have shown that the effect of microecological agents on QUICKI is diverse (90, 94-96). In our study, the sensitivity analysis showed a significant increase in QUICKI after excluding 1 study (36). Upon careful comparison, the excluded study was found to have included only female patients and used prebiotics for the intervention, which might be an important reason why the results were not robust. It should be noted that HOMA-β, HOMA-IR, fasting C-peptide, fasting insulin, and QUICKI are static index. In the absence of dynamic data, it is difficult to truly reflect the dynamic process of insulin secretion by islet β-cells (97). However, owing to their simplicity, these biomarkers may be more suitable for use in large epidemiological studies to assess treatment effects or investigate the natural history of T2DM (98). Moreover, HOMA and QUICKI have a good correlation with insulin sensitivity/resistance of glucose clamp technique, which are often considered the “gold standard” (99). Therefore, based on the above results, we cautiously suggest that MTTs treatment may be associated with beneficial effects on β-cell function and ameliorate IR.

Another interesting finding is that the reduction observed in HbA1c, the most widely accepted standard for glucose control measurement, was statistically significant. However, a 0.28% HbA1c reduction falls below the 0.3% threshold typically considered clinically meaningful for pharmacotherapy by the U.S. Food and Drug Administration (100). The UK Prospective Diabetes Study (UKPDS) predicts that a 1% reduction in HbA1c over 10 years is associated with a 21% reduction in diabetes-related deaths, a 14% reduction in myocardial infarctions, and a 37% reduction in microvascular complications (101). The incidence rates of these diabetes-related endpoints increase linearly with rising HbA1c levels, with no evidence of a threshold effect. Therefore, an improvement in HbA1c of 0.28% observed over a median intervention period of 8 weeks may indeed be clinically meaningful among individuals with diabetes. Furthermore, the modest effect size underscores the need to contextualize this intervention as complementary rather than a replacement for existing treatments. Future studies should identify patient subgroups most likely to derive clinically meaningful benefit and assess long-term adherence while combining with validated behavioral interventions to evaluate synergistic effects on glycemic variability.

Notably, only 2 studies investigated FMT, both of which had small sample sizes and reported significant heterogeneity. Due to the lack of data, no definitive conclusions can be drawn about the efficacy or safety of FMT. These results should be interpreted with extreme caution, and future research is urgently needed to evaluate FMT as a standalone intervention for T2DM.

The subgroup analysis of most outcome variables related to glucose metabolism showed that the European region, American region, FMT intervention, and prebiotics intervention had no statistical significance on glucose metabolism indicators. This lack of significance may be due to the small number of included studies in corresponding subgroups. However, it is encouraging to note that the probiotics dose < 10¹⁰ CFU, the prebiotics dose ≥ 10 g, and the intervention duration < 12 weeks were found to be optimal for lowering FPG and HbA1c. These findings were consistent with the optimal dose and duration patterns observed in other studies. However, in this meta-analysis of the overall effect of HbA1c, we did not distinguish between studies with a follow-up duration of less than 12 weeks. However, subgroup analyses based on the intervention duration may address the associations that Bock et al were concerned about. Additionally, subgroup analyses revealed that the effectiveness of MTTs on glucose management was influenced by patients' BMI and mean age. In particular, MTTs were found to be more effective in reducing FPG, FINS, and HbA1c levels in younger (< 52 years old) and obese patients with T2DM. This may be due to the alteration of their GM composition. Obese T2DM patients often have an imbalanced GM, which can be restored by administering MTTs. Obesity raises the risk of metabolic disorders, such as IR, hyperglycemia, and hyperlipidemia. MTTs may exert beneficial effects on obesity-induced metabolic disorders by modulating immune responses and altering the production of gut-derived metabolites. One possible explanation for the findings favoring a younger population in this meta-analysis is that older patients generally suffer from comorbidities or take medications that interfere with the results of biotic supplementation.

In addition, this meta-analysis added to multiple inflammatory factors to promote a comprehensive understanding of the gut MTTs for inflammation in T2DM. The conclusion regarding the effects of hs-CRP and TNF-α levels is consistent with the findings from the previous meta-analyses by Zheng et al (102), Zhou et al (103), and Naseri, K et al (104). There was no meta-analysis performed on the effect of gut MTTs on LPS and IL-10. The results of this meta-analysis showed a decrease in LPS levels and an increase in IL-10 levels after the use of MTTs. This might be due to the fact that prebiotics improved intestinal permeability by increasing beneficial bacteria and strengthening intestinal endothelial tight junctions, thereby reducing the serum LPS levels. The supplementation with prebiotics stimulated the Th2-related immune pathway, promoting the secretion of IL-10. The subgroup analysis suggested that FMT had not been reported to affect inflammation markers in T2DM patients. Inflammation indicators were significantly improved by prebiotics supplementation, particularly in subjects who were aged < 52 years with a baseline BMI ≥ 28 kg/m², intervention duration < 12 weeks, and prebiotics dose ≥ 10 g. This result corresponded to the previously mentioned findings regarding glycemic control. The meta-analysis by Liang et al reported that inflammation indicators were significantly improved in subjects who were aged ≤ 55 years with a baseline BMI < 30 kg/m² and an intervention duration > 8 weeks. This suggested that the potential optimal duration (effect on inflammation markers) of MTT intervention might be 8 to 12 weeks. Additionally, it might be more effective in reducing inflammation in diabetic individuals with a BMI between 28 and 30 kg/m².

T2DM is characterized by low-grade metabolic inflammation fueled by GM. Previous studies have reported the therapeutic effect of MTTs in diabetic conditions in terms of moderating gut dysbiosis, reducing inflammation, improving insulin sensitivity, and increasing levels of antioxidants. In this study, a meta-analysis was performed for changes in GM based on limited data. The short intervention duration in most studies may underestimate the sustained effects of MTTs on gut microbiota. Long-term RCTs are critical to assess durability and cumulative benefits. Our analysis revealed that while MTTs intervention led to a significant increase in the relative abundance of specific taxa such as Actinobacteria, Lactobacillus, and the Lactobacillus casei subgroup, overall measures of microbial community diversity (α-diversity) and compositional differences between groups (β-diversity) remained largely unchanged. And the inconsistencies observed in the studies investigating the association between the effects of MTTs on the GM structure and composition in patients with T2DM may originate from multiple factors. Firstly, the inherent complexity and diversity of the GM among individuals may lead to differences in the study results. Differences in study designs, including variations in the duration of interventions, the types of interventions adopted, and the characteristics of participants, may also contribute to the divergent findings. Moreover, methodological differences in the assessment of the GM, such as differences in sequencing technologies or bioinformatics pipelines, may result in different outcomes. Additionally, the intricate interplay between the GM and host physiology involves numerous factors, including diet, genetics, and environmental influences, which further complicates the interpretation of the study results. Hence, due to the small amount of data and the high heterogeneity, the results should be interpreted with caution. The relative abundance of Actinobacteria, Lactobacillus, and Lactobacillus casei increased after MTT intervention. Actinobacteria (such as Streptomyces olivaceus) produce the metabolite acarbose, which is the first antihyperglycemic drug to delay carbohydrate breakdown by inhibiting intestinal α-glucosidase (105). Lactobacillus can improve glucose metabolism, reduce excess reactive oxygen species, alleviate inflammation, and regulate GM, which has been proven to be the potential mechanism in improving T2DM. The Lactobacillus casei group is the most widely studied and applied probiotic species of Lactobacilli. The fimbria-like structure of Lactobacillus casei facilitates binding to mucus-covered mucosal epithelial cells and plays an immunomodulatory role by reducing the expression of pro-inflammatory factors.

The mechanisms of MTTs on T2DM are to regulate the GM composition, to regulate intestinal microbial metabolites, and to improve gut barrier function. MTTs may directly regulate the abundance of Lactobacillus and other beneficial bacteria. The modulation of GM composition promoted the IRS-1-PI3K-AKT pathway and the AMPK-PPARα-CPR1α pathway in the liver and improve tetrathionate metabolism. Although direct evidence from the included studies on microbial functional outputs is limited, the observed increases in bacteria like Lactobacillus, which are known short-chain fatty acid producers, and the reduction in systemic inflammation, are consistent with the proposed mechanisms involving microbial metabolite signaling. Furthermore, MTTs improve the gut barrier function to attenuate metabolic diseases by up-regulating the expression of claudin 1, glucagon-like peptide 1 (GLP-1), IL-10, occludin 1, and ZO-1.

This study has limitations. First, substantial unexplained heterogeneity existed among the included studies. To address this, subgroup analyses were conducted to explore potential sources, including region, intervention type, number of strains, pro/prebiotic dosage, intervention duration, baseline BMI, and mean age. However, the wide diversity of probiotic strains, prebiotic species, and synbiotic compositions across studies precluded meaningful subgroup analyses. Additional heterogeneity sources included dietary habits, physical activity levels, baseline microbiome profiles, proportions of participants receiving different MTTs, and individual genotypes. Dose-response relationships could not be analyzed due to inconsistent dosage units across studies and missing dosage information in some reports, further exacerbating heterogeneity. Second, while most studies exhibited low bias risk, unclear reporting and missing data compromised overall quality and reliability. Ambiguous methodologies hindered accurate bias assessment, and attrition bias from incomplete data threatened result validity and effect estimate precision. Although this analysis included a large number of trials, individual studies often had small sample sizes. FMT studies were particularly limited, precluding subgroup analyses by donor characteristics, delivery routes, or bacterial content–critical factors that may restrict statistical power and generalizability. Regional variation in nutraceutical effects further limited generalizability: significant outcomes were not uniformly observed across populations, with the strongest effects noted in Mediterranean-region studies (Table 2). This may relate to unrestricted access and high consumption of over-the-counter probiotics in some Middle Eastern countries (106). However, pharmacological understanding of next-generation probiotics remains insufficient, leading to practice disparities between community and hospital pharmacists in areas like pro/pre/synbiotic administration, counseling, and storage. These differences pose long-term medico-social challenges and may contribute to regional efficacy variations. Moreover, the majority of included clinical studies originated from Iran, limiting representation of global populations and result extrapolation. Third, the inclusion and search strategy focused exclusively on studies explicitly reporting prebiotic use. Compared to well-characterized probiotics, prebiotic mechanisms of action are poorly understood, potentially causing underrepresentation of prebiotic interventions in diabetic populations (107, 108). Fourth, while HbA1c reflected 12-week post-intervention glycemic changes, most included studies had short durations, attenuating effect sizes and highlighting intervention length as a covariate in subgroup analyses. Few studies adequately measured glucose tolerance and β-cell function—key factors in insulin secretion and sensitivity regulation. Additional limitations include: evidence of publication bias for outcomes like TNF-α, hs-CRP, HOMA-IR, and HbA1c, suggesting underreporting of null/negative results that may inflate observed effects; lack of standardization in microbiota methods, which limits cross-study comparability of microbial outcomes; incomplete and inconsistent adverse event documentation, potentially underestimating risks, especially in vulnerable groups; potential confounding from uncontrolled comorbidities (eg, cardiovascular/kidney disease), medications (eg, statins, anti-inflammatory drugs), and dietary factors—most studies lacked standardized dietary reporting, a known gut microbiome modulator—adding heterogeneity; and limited data on intervention adherence and protocols, despite MTT formulation differences and storage conditions potentially impacting efficacy. Given the paucity of FMT data, well-designed RCTs with larger sample sizes and standardized are essential to evaluate its potential as a microbiome–targeted strategy for T2DM. Comparative studies between FMT and other MTTs would also clarify its clinical utility.

In response to these limitations, we have been cautious in interpreting our results and have emphasized the need for further research. We believe that future studies with larger sample sizes, longer intervention durations, and more standardized methodologies are essential to evaluate MTTs as treatment for T2DM. This is especially relevant given the low cost of MTTs and the relative paucity of major adverse effects (109, 110).

This review has several strengths. First, it is the first systematic review and meta-analysis of RCTs simultaneously evaluating pro-/pre-/synbiotics, and FMT on glucose metabolism, systemic inflammation indices, and GM in T2DM. Furthermore, it included a large number of studies and a large overall sample size with T2DM, which suggests that the conclusions can be considered reliable. Moreover, a subgroup analysis was conducted based on the region, intervention type, study duration, number of probiotic strains, pro/prebiotics dose, baseline BMI, and mean age to determine the best possible dose, best time of intervention, and best intervention modality, and so on.

Conclusions

The results suggest that MTT intervention has a modest but significant benefit in glucose homeostasis, inflammation, and GM in T2DM by reducing FPG, 2hPBG, fasting insulin, HbA1c, HOMA-IR, hs-CRP, TNF-α, and LPS, and increasing IL-10, and the abundance of Actinobacteria, Lactobacillus, and Lactobacillus casei. In particular, individuals with an age ≤ 52 years, baseline BMI ≥ 28 kg/m², and intervention duration < 12 weeks benefited more from the improvement of glycemic and inflammatory cytokines. However, due to considerable heterogeneity and methodological limitations across studies, the evidence supporting the routine use of MTTs as adjuvant therapy remains inconclusive. Well-designed RCTs with longer durations and standardized methodologies are essential for more definitive conclusions.

Abbreviations

2hPBG

2 hours postprandial blood glucose

ACE index

abundance-based coverage estimator index

BG

blood glucose

BMI

body mass index

BSH

between-study heterogeneity

CFU

colony-forming units

FMT

fecal microbiota transplantation

FPG

fasting plasma glucose

GM

gut microbiota

HbA1c

glycosylated hemoglobin

HOMA-β

homeostatic model assessment for beta-cell function

HOMA-IR

homeostatic model assessment for insulin resistance

hs-CRP

high-sensitivity C-reactive protein

IFN-γ

interferon-γ

IL-

interleukin

IR

insulin resistance

LPS

lipopolysaccharide

MD

mean difference

MTT

microbiome-targeted therapy

PB

publication bias

QUICKI

quantitative insulin sensitivity check index

RCT

randomized controlled trial

SMD

standardized mean difference

T2DM

type 2 diabetes

TNF-α

tumor necrosis factor-α

Disclosures

Author contributions

All authors contributed to the study conception and design. Writing—original draft preparation: X.Z.; Writing—review and editing: Y.L. and J.Z.; Conceptualization: X.Z.; Methodology: X.Z. and J.M.; Formal analysis and investigation: X.Z. and W.Z.; Software: X.Z.; Funding acquisition: T.Z.; Supervision: T.Z. and W.K., and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Disclosures

The authors have nothing to disclose.

Ethical Approval

Not applicable

Consent to Participate

Not applicable

Consent to Publish

Not applicable

Funding

This work was supported by the National Natural Science Foundation of China, grant number 82270909 and Key Research and Development Program of Hubei Province, grant number 2020BCB003.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.