The effect of oral synbiotics on the gut microbiota and inflammatory biomarkers in healthy adults: a systematic review and meta-analysis

Abstract

Context

Prior research has explored the effect of synbiotics, the combination of probiotics and prebiotics, on the gut microbiota in clinical populations. However, evidence related to the effect of synbiotics on the gut microbiota in healthy adults has not been reviewed to date.

Objective

A systematic review and meta-analysis was conducted to comprehensively investigate the effect of synbiotics on the gut microbiota and inflammatory markers in populations of healthy adults.

Data Sources

Scopus, PubMed, Web of Science, ScienceDirect, MEDLINE, CINAHL, and The Cochrane Library were systematically searched to retrieve randomized controlled trials examining the primary outcome of gut microbiota or intestinal permeability changes after synbiotic consumption in healthy adults. Secondary outcomes of interest were short-chain fatty acids, inflammatory biomarkers, and gut microbiota diversity.

Data Extraction

Weighted (WMD) or standardized mean difference (SMD) outcome data were pooled in restricted maximum likelihood models using random effects. Twenty-seven articles reporting on 26 studies met the eligibility criteria (n = 1319).

Data Analysis

Meta-analyses of 16 studies showed synbiotics resulted in a significant increase in Lactobacillus cell count (SMD, 0.74; 95% confidence interval [CI], 0.15, 1.33; P = 0.01) and propionate concentration (SMD, 0.22; 95% CI, 0.02, 0.43; P = 0.03) compared with controls. A trend for an increase in Bifidobacterium relative abundance (WMD, 0.97; 95% CI, 0.42, 2.52; P = 0.10) and cell count (SMD, 0.82; 95% CI, 0.13, 1.88; P = 0.06) was seen. No significant differences in α-diversity, acetate, butyrate, zonulin, IL-6, CRP, or endotoxins were observed.

Conclusion

This review demonstrates that synbiotics modulate the gut microbiota by increasing Lactobacillus and propionate across various healthy adult populations, and may result in increased Bifidobacterium. Significant variations in synbiotic type, dose, and duration should be considered as limitations when applying findings to clinical practice.

Systematic Review Registration

PROSPERO no. CRD42021284033.

INTRODUCTION

There is a growing global consumer interest in gut microbiota therapies and their related health benefits. These therapies may include probiotics, prebiotics, or synbiotics, each having different effects on human health. A recent survey of 16 000 participants from 16 countries reported that 48% of respondents consumed probiotics daily, or almost daily, with interests driven by the functional benefits to the gut microbiome and immune system.¹

Probiotics are live bacteria that have demonstrated benefits to human health when consumed in adequate amounts.² Probiotics have been shown to confer strain-specific clinical benefits for multiple conditions, including gastrointestinal symptoms,³ irritable bowel syndrome,⁴ Helicobacter pylori eradication,⁵ infection,^6,7 type 2 diabetes mellitus,⁸ and chronic kidney disease.⁹ Despite their clinical benefits, there is limited evidence to demonstrate that probiotics modulate the gut microbiota. A systematic review by Kristensen et al¹⁰ found no significant effect on any microbiota measurements after probiotic supplementation in 7 randomized controlled trials with healthy adults. Similarly, another review of healthy adults by McFarland et al¹¹ found a limited capacity for probiotics to alter the gut microbiota in only 4 of 29 studies.

Prebiotics are substrates that are selectively utilized by the host gut bacteria to provide a health benefit.¹² The most widely researched prebiotics are inulin, galacto-oligosaccharide (GOS), and fructo-oligosaccharide (FOS), and these have demonstrated the ability to modulate the human gut microbiota in healthy adult populations by increasing the abundance of beneficial bacteria.^13,14

The term “synbiotic” refers to the combination of probiotic and prebiotic therapies.¹⁵ Given the lack of evidence demonstrating significant microbiota modulation from probiotic supplementation alone, and the recent evidence of the effect of prebiotics on bacterial populations, synbiotics have emerged as a potential microbiota-modulating and therapeutic option.¹⁵ To date, the clinical benefits of synbiotics have been explored in many chronic and acute disease populations, including infections,¹⁶ irritable bowel syndrome,¹⁷ critical illness,¹⁸ overweight/obesity,¹⁹ and diabetes.⁸ However, the evidence related to the effect of synbiotics on the gut microbiota has rarely been investigated. Systematic reviews to date have included a limited number of synbiotic intervention trials alongside probiotic and prebiotic interventions, examining the effects on the gut microbiota in adults with chronic conditions.^20–22 In chronic kidney disease, 2 synbiotic interventions showed a significant increase in Bifidobacterium.²¹ In inflammatory bowel disease, a further 2 synbiotic interventions each found contrasting results.²² One included study reported an increase in Bifidobacterium after supplementation with Bifidobacterium longum probiotic and an inulin/oligofructose prebiotic²³; yet, the second study reported no significant difference to Bifidobacterium or Lactobacillus after supplementation with probiotic Bifidobacterium breve and a GOS prebiotic,²⁴ A systematic review of the effect of pro-, pre- and synbiotics on the gut microbiome in major depressive disorder did not include any synbiotic intervention trials that reported on gut microbiota outcomes. ²⁰ Thus, there is a lack of systematic reviews synthesizing the evidence for the effect of synbiotics on the human gut microbiota and, to our knowledge, the effect of synbiotics on the gut microbiota of healthy adult populations has not been systematically reviewed.

Given the global consumer interest in gut microbiota health and supplements, particularly from health-conscious populations,^1,25 it is essential to ascertain to what extent microbiota-modulating therapies are of benefit to healthy adults. A 2020 Australian survey (n = 1265) found that 58.9% of participants were probiotic consumers. These respondents were mostly female, more educated, and reported healthier lifestyle behaviors related to fruit intake, physical activity, and alcohol intake risk compared with respondents who did not take probiotics.²⁵ No data were collected on the respondents’ gut microbiota and the survey only focused on probiotics, and not pre- or synbiotic therapies. An investigation of the effect of microbiota-modulating therapies on the gut microbiota and other associated health benefits in healthy adult populations is timely. Moreover, to the authors’ knowledge, the evidence for the effect of synbiotic therapies on inflammatory biomarkers has not been reviewed, despite the well-known role of the gut microbiota and short-chain fatty acids (SCFAs) on regulation of the immune system.²⁶

Thus, this review aimed to examine the effects of oral synbiotic supplementation on the modulation of the gut microbiota and inflammatory biomarkers in healthy adult populations. The specific research question was as follows: “In healthy adults, what is the impact of synbiotic consumption on the gut microbiota and inflammatory markers?”

METHODS

Study protocol

This was a systematic review and meta-analysis of randomized trials examining the effects of oral synbiotic supplementation on measurements of gut microbiota and clinical biomarkers in healthy adult populations. The systematic review was conducted according to the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions,²⁷ and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.²⁸ The review was first registered in the International Prospective Register of Systematic Reviews on 18 November 2021 (PROSPERO CRD42021284033) with amendments made on 6 December 2022. The PICO (Population, Intervention or exposure, Outcomes) model was used to define the research question (Table 1).

Table 1.

PICO criteria for inclusion and exclusion of studies

Parameter	Criterion
Population	Healthy adults aged 18–65 y
Intervention or exposure	Synbiotic consumption (probiotic + prebiotic), in supplement or food fortified form
Comparison	Placebo or standard diet
Outcomes	Changes in gut microbiota composition or intestinal permeability (primary outcome). Additionally, α-diversity, SCFAs, and inflammatory biomarkers (secondary outcomes)

Abbreviation: SCFA, short-chain fatty acid.

The systematic search was conducted across Scopus, PubMed, Web of Science Core Collection, ScienceDirect, MEDLINE (via EBSCO), CINAHL (via EBSCO), and The Cochrane Library on 24 October 2021 and updated on 30 May 2022. The search strategy used Medical Subject Heading (MeSH) terms, where available, and keywords identified from relevant literature. The full search strategy is presented in Appendix S1 (see the Supporting Information online). No limitations were applied to article language or date of publication. Reference lists of eligible articles were hand-searched for potential studies meeting the eligibility criterion.

Study eligibility

The criteria for eligibility were as follows: (1) randomized or quasi-randomized controlled trials (including parallel and crossover designs), (2) studies conducted with healthy adults aged 18 years or older, (3) oral synbiotic consumption in supplement or fortified food form, (4) placebo as comparator, (5) studies where the effect of synbiotics could be isolated from other interventions including medication or dietary changes, and (6) studies assessing modulation to the gut microbiome composition or intestinal permeability, considered as the primary outcome for this review. Secondary outcome measures included in this review were SCFAs, inflammatory biomarkers, and measures of gut microbiota diversity. Healthy adults were defined as those without a significant clinical disease that impacts the gut microbiota and/or inflammation, including inflammatory bowel disease, viral or bacterial infective states, cancer, gastrointestinal surgery, autoimmune conditions, enteral nutrition, critical illness, neurodegenerative diseases, liver disease, or chronic kidney disease.

The following exclusion criteria were applied: (1) animal or in vitro studies; (2) non-intervention studies including observational and case-control studies; (3) protocols, abstracts, expert opinions, letters, seminars, books, or book chapters; and (4) study populations of children or populations with clinical diseases outlined above. If studies had more than 2 arms, only the comparison of synbiotics to placebo/control arm was considered.

Study selection

After removal of duplicates, articles were imported to Covidence Systematic review software (Veritas Health Innovation)²⁹ for title and abstract screening by 3 independent reviewers (D.J.C., K.C., Y.P.), with any disagreement resolved via consensus. Full-text screening was then performed by 1 reviewer (D.J.C.), with articles of concern discussed among the research team (K.C., Y.P.). Where multiple articles reported results from a single study, the associated articles were linked, as recommended in the Cochrane Handbook for Systematic Reviews of Interventions.²⁷ Reference lists of included articles were manually searched to identify additional relevant articles.

Data extraction

Data extraction was performed by 1 reviewer (D.J.C.) in consultation with the research team (K.L., Y.P., K.C.). Data were extracted from each study for the following: citation, country, study design, population description, sample size, study duration, synbiotic type, dose of synbiotic, and details of control and baseline demographic data (shown in Table 2^30–56). Study findings data were extracted and categorized as microbiota outcomes, SCFAs, diversity and richness, and inflammatory biomarkers (shown in Table 3^30–56). All data are reported as continuous variables. The mean final values or mean change values with respective standard deviations (SDs) (or standard error/95% confidence interval [CI]) were collated for each outcome, where possible. When not available, the mean final values and respective SDs were converted from provided data, as outlined in the Cochrane Handbook for Systematic Reviews of Interventions.²⁷ Where necessary, median and interquartile range values were converted to mean and SD using the formula developed by Wan et al.⁵⁷ Where the published article did not provide adequate information, authors were contacted by email for additional details. Data from intention-to-treat analyses were extracted for use in this meta-analysis, where possible. Where this was not available, data from per-protocol analyses were used and the impact considered in the risk-of-bias assessment.

Table 2.

Characteristics of the included studies according to population subtypes

Reference	Study design	Participants	Sample size, n (n = synbiotic group)	Duration	Probiotic/day	Prebiotic/day	Control/day	Microbiota or intestinal permeability analysis	Inflammatory markers
Healthy adults
Casiraghi et al,55 Italy	Parallel RCT	Healthy adults; 22-47 y	26 (13)	4 wk	109 CFU Lactobacillus acidophilus 74-2 5× and 109 CFU Bifidobacterium lactis 420	9.25 g Inulin	500 mL milk	Culture	—
Palaria et al,30 USA	Crossover RCT	Healthy adults; mean age 30 y	52	3 wk	109–1010 CFU of B. animalis subsp. lactis Bb-12 + yogurt starter blend containing Streptococcus thermophilus and Lactobacillus bulgaricus	1 g Inulin	94 g Milk	qPCR	—
van Zanten et al,31 Denmark	Crossover RCT	Healthy adults; 18-50 y	18	3 wk	109 CFU L. acidophilus NCFM	5 g Cellobiose	5 g Maltodextrin	16s rRNA V3-4; qPCR	—
Roberts et al,32 UK	Parallel RCT	Recreational athletes; mean age 35 y	20 (10)	3 mo	1010 CFU L. acidophilus CUL-60, 1010 CFU L. acidophilus CUL-21, 9.5 × 109 CFU Bifidobacterium bifidum CUL-20, 0.5 × 109 CFU Bifidobacterium animalis subsp. lactis CUL-34	5.6 g FOS	2 g Corn flour	Urinary lactulose and mannitol recovery	Plasma endotoxin
Rajkumar et al,33 India	Parallel RCT	Healthy adults; 20-25 y; BMI 18.5–24.9 kg/m2	30 (15)	6 wk	2 × 109 CFU Lactobacillus salivarius	10 g FOS	Gelatin capsule	Culture	hs-CRP, TNF-α, IL-6, and IL-1β
Valle et al,34 Brazil	Parallel RCT	Healthy adults; 18-22 y; enrolled at Brazil Army Cadet school	80 (40)	1 mo	10.3-log CFU L. acidophilus LA-5, 11.0-log CFU B. animalis BB-12	2.3 g Inulin	60 g Ice cream	16s rRNA V3-4	—
Childs et al,35 UK	Crossover RCT	Healthy adults; 25-65 y; BMI 20-30 kg/m2	44	3 wk	109 CFU B. animalis subsp. lactis Bi-07	8 g XOS	8 g Maltodextrin	qPCR; FISH with 16s rRNA; flow cytometry	Fecal IgA, salivary IgA, IL-6, IL-10, TNF-α, TNF-β
Constipation
Kim et al,36 Korea	Parallel RCT	Adults with Rome III constipation	30 (20)	1 mo	Lacticaseibacillus casei 3451, Lactobacillus plantarum 3501, Lactobacillus rhamnosus 3201, Bifidobacterium lactis 4301, B. breve 4401, Lactococcus lactis 2301 (dose unspecified)	XOS, oat fiber, psyllium husk fiber (dose unspecified)	Glucose anhydrocrystalline, maltodextrin, and corn starch	16s rRNA V3-4	—
Bazzocchi et al,37 Italy	Parallel RCT	18-65 y; Rome III constipation	31 (19)	2 mo	L. plantarum, L. acidophilus, L. rhamnosus, Bifidobacteriumlongum, and Bifidobacteriumbreve (dose unspecified)	5 g Psyllium	Maltodextrin	16s rDNA; denaturing gradient gel electrophoresis; qPCR	—
Bogovič Matijašić et al,38 Slovenia	Parallel RCT	16-65 y; Rome III constipation–predominant IBS	76 (33)	1 mo	6.48 × 109 CFU L. acidophilus La-5 and 9 × 109 CFU B. animalis subsp. lactis BB-12	7.2 g Mixed dietary fiber (inulin and FOS)	Heat-treated fermented milk	Random amplified polymorphic DNA RAPD profiling; PCR; 16s rRNA V4	—
Anzawa et al,39 Japan	Crossover RCT	20-64 y; constipation experienced 3-5 days/wk	60	2 wk	1010 CFU of B. animalis subsp. lactis GCL2505	2 g Inulin	Fermented milk	qPCR	—
Neyrinck et al,40 Belgium	Parallel RCT	50-70 y; Rome III constipation	27 (13)	1 mo	5 × 109 CFU B. animalis lactis Vesalius 002/bag (2 bags for first 5 d, then 1 bag for remaining 25 d)	4.95 g FOS/bag (2 bags for first 5 d, then 1 bag for remaining 25 d)	5 g Maltodextrin	16s rDNA V1-3; plasma intestinal fatty-acid binding protein (iFABP)	IL-6, IL-8, IL-1β, IL-10, IL-17a, TNF-α, IFN-γ, Monocyte chemotactic protein (MCP)-1
Elderly adults
Shioiri et al,41 Japan	Crossover RCT	Healthy adults; mean age 74 y	20	2 wk	3 × 1010 CFU L. casei Shirota	2.5 g GOS	80 mL Fermented milk	Culture	—
Lee et al,56 Korea	Parallel RCT	Healthy females; ≥65 y	51 (35)	1 wk or 4 wk	10 × 109 CFU B. animalis spp. Lactis HY8002, 5 × 109 CFU L. casei HY2782, and 5 × 109 CFU L. plantarum HY7712	8 g Mixed dietary fiber (polydextrose, chicory dietary fiber, lactulose, and wheat, XOS)	Placebo drink (unspecified)	16s rRNA V3-4; zonulin	—
Ouwehand et al,42 Finland	Parallel RCT	Healthy adults; ≥65 y; regularly take NSAIDs	51 (24 completed)	2 wk	2 × 109 CFU/g L. acidophilus NCFM	5 g Lactilol	5 g Sucrose	qPCR	TNF-α and fecal IgA
Macfarlane et al,43 UK	Crossover RCT	Healthy adults; 65-90 y; BMI 18.5–30.0 kg/m2	47	1 mo	4 × 1011 B. longum	12 g Synergy 1 (inulin + FOS)	12 g Maltodextrin	FISH with 16s rRNA	CRP, IgA g/L; IgG g/L; IL-10 + IL-8 + IL-6 + TNF-α
Lehtinen et al,44 UK	Crossover RCT	Healthy adults; >60 y	20 1	3 wk	1 × 109 CFU B. animalis subsp. Lactis Bi-07	8 g GOS	8 g Maltodextrin	FISH with 16s rRNA; qPCR	IFN-γ, IL-1β, IL-6, IL-8, IL-10, TNF-α, and IgA
Costabile et al,45 UK	Crossover RCT	Healthy adults; 60-80 y	40	3 wk	Synbiotic 1: 1 × 1010 CFU L. rhamnosus GG-PB12	6 g Soluble corn fiber	Maltodextrin	16s rRNA V3-4; qPCR	hs-CRP, IL-6, and IL-8
					Synbiotic 2: 1 × 1010 CFU L. rhamnosus GG
Type 2 diabetes
Kassaian et al,46 Iran	Parallel RCT	Adults with prediabetes	80 (40)	6 mo	1.5 × 109 CFU each L. acidophilus, B. lactis, B. bifidum, and B. longum	6 g Inulin	6 g Maltodextrin	qPCR	—
Horvath et al,47 Austria	Parallel RCT	>18 y; diagnosed T2DM; BMI 30-40 kg/m2	41 (21)	6 mo	1.5 × 1010 CFU blend of B. bifidum W23, B. lactis W51, B. lactis W52, L. acidophilus W37, L. casei W56, L. brevis W63, L. salivarius W24, Lactococcuslactis W58, and Lc. lactis W19	10 g Omnilogic Plus (OMNi-BIOTIC®) (GOS + FOS, 8 g active prebiotic)	6 g Starch	16s rDNA V1-2; zonulin; diamine	—
Kanazawa et al,48 Japan	Parallel RCT	30-80 y; T2DM treated with diet and exercise alone	88 (44)	6 mo	3 × 108 Lc. paracasei YIT 9029 and 3 × 108 B. breve YIT 12272	7.5 g GOS	Not stated	qPCR; 16s rRNA V1-2	hs-CRP and IL-6
Overweight adults
Krumbeck et al,49 USA	Parallel RCT	18-65 y; BMI 30-40 kg/m2	58 (39)	3 wk	Synbiotic 1: 1 × 109 CFU Bifidobacterium adolescentis IVS-1	5 g GOS	7 g Lactose	qPCR; 16s rRNA V5-6; urinary lactulose and mannitol	Serum endotoxin
					Synbiotic 2: 1 × 109 CFU B. animalis subsp. lactis BB-12
Hibberd et al,50 USA	Parallel RCT	18-65 y; BMI 28.0–34.9 kg/m2	113 (56)	6 mo	1010 CFU B. animalis subsp. lactis 420	12 g Litesse® Ultra polydextrose (IFF)	12 g Micro-crystalline cellulose	16s rRNA V4; zonulin	hs-CRP
Stenman et al,51 Finland	Parallel RCT	18-65 y; BMI 28.0–34.9 kg/m2	113 (56)	6 mo	1010 CFU B. animalis subsp. lactis 420	12 g Litesse® Ultra polydextrose (IFF)	12 g Micro-crystalline cellulose	Zonulin	hs-CRP, Lipopolysaccharide, Endotoxin, and IL-6
Angelino et al,52 Italy	Parallel RCT	30-65 y; BMI 25-35 kg/m2; healthy sedentary adults	46 (23)	3 mo	Bacillus coagulans BC30 1 × 107 CFU/g pasta (dose unspecified)	Whole-wheat pasta supplemented with barley β-glucan (dose unspecified)	Whole-wheat pasta	qPCR	hs-CRP, IL-6, IL-10, and TNF-α
Asthma and migraines
McLoughlin et al,53 Australia	Crossover RCT	Adults with doctor-diagnosed, stable asthma	17	1 wk	7.5 × 109 CFU L. acidophilus LA-5, 8.75 × 109 CFU L. rhamnosus GG and 8.75 × 109 CFU B. animalis subsp. lactis BB-12	12 g Inulin	12 g Maltodextrin	qPCR; 16s rRNA V4	—
Ghavami et al,54 Iran	Parallel RCT	20-50 y; adult women with migraines diagnosed according to the ICDH-3 criteria.	80 (40)	3 mo	2 × 109 CFU L. casei, L. acidophilus, L. rhamnosus, L. helveticus, L. bulgaricus, L. plantarum, L. gasseri, B. breve, B. longum, B. lactis, B. bifidum, and S. thermophilus	FOS (dose unspecified)	Starch	Zonulin	hs-CRP

Abbreviations: BMI, body mass index; CFU, colony-forming units; CRP, C-reactive protein; FISH, fluorescence in situ hybridization; FOS, fructo-oligosaccharide; GOS, galacto-oligosaccharide; hs-CRP, high-sensitivity C-reactive protein; IFN-γ, interferon-γ; IgA, immunoglobulin A; IgG, immunoglobulin G; IL, interleukin; NSAID, nonsteroidal anti-inflammatory drug; PCR, polymerase chain reaction; qPCR, quantitative polymerase chain reaction; RCT, randomized controlled trial; T2DM, type 2 diabetes mellitus; TNF-α, tumor necrosis factor α; XOS, xylooligosaccharide. 

Table 3.

Microbiota and inflammatory outcomes after synbiotic consumption

Study ID	Microbiota outcomes	SCFA	Richness and diversity	Inflammatory markers
Healthy adults
Casiraghi et al,55 Italy	↑ Bifidobacterium and Lactobacillus	—	—	—
Palaria et al,30 USA	↑ Bifidobacteria (subgroup of participants with low baseline Bifidobacterium) ↓ Clostridium Cluster I (subgroup of participants with high Clostridium Cluster I at baseline)	—	—	—
van Zanten et al,31 Denmark	↔ Phylum taxonomy and F/B ratio ↓ Clostridiales, Desulfovibrionales, Desulfovibrionaceae, Blautia, and Roseburia ↑ Bifidobacteriaceae, Bifidobacterium, Lactobacillus and Paracteroides	↔ Acetate ↔ Propionate ↔ Butyrate	↔ PCoA weighted UniFrac distance	—
Roberts et al,32 UK	↔ 5 h Lactose:mannitol recovery ratio ↔ Lactulose % recovery ↔ Mannitol % recovery	—	—	↓ Plasma Endotoxin ↔ IgG anti-Endotoxin
Rajkumar et al,33 India	↑ Lactobacillus ↓ Escherichia coli and fecal coliform counts	—	—	↓ hs-CRP, TNF-α, IL-6, and IL-1β
Valle et al,34 Brazil	↔ Phylum ↓ Erysipelotrichacea and Lachnoclostridium ↓ Parabacteroides ↑ Bifidobacterium, Lactobacillus ↑ Ruminococcaceae_UCG-00 and RuminococcaceaeUBA1819 ↑ B. animalis	↑ Acetate ↑ Propionate ↑ Butyrate	↔ Shannon Index ↔ Simpson Index PCoA of taxonomic dissimilarity showed clustering with Blautia and Bacteroides at baseline, shifting to Bifidobacterium, Faecalibacterium, and Agathobacter post-intervention.	—
Childs et al,35 UK	↑ Bifidobacterium ↔ B. lactis, total bacteria and Atopobium ATO291 ↔ Bacteroides/Prevotella group ↔ Clostridium clusters I and II (Chis150) ↔ Lactobacillus/Enterococcus subgroup	↔ Acetate ↔ Propionate ↔ Butyrate	—	↔ Fecal IgA and salivary IgA ↔ IL-6, IL-10, TNF-α, and TNF-β
Constipation
Kim et al,36 Korea	↔ Between-group difference for all taxonomic levels; within-group changes below were also seen for placebo ↓ Firmicutes and Ruminococcaceae ↓ Lachnospiraceae and Prevotellaceae ↑ Bacteroidetes	—	↓ OTU, Chao1 Index, and phylogenetic diversity ↔ Shannon Index	—
Bazzocchi et al,37 Italy	↔ L. plantarum, L. acidophilus, and L. rhamnosus ↔ B. longum and B. breve ↔ Bifidobacterium spp. and Lactobacillus	—	—	—
Bogovič Matijašić et al,38 Slovenia	↑ L. acidophilus La-5 ↑ B. animalis subsp. lactis BB12 ↑ L. acidophilus La-5:Lactobacillus ratio ↑ L. acidophilus La-5:all bacteria ratio ↑ B. animalis subsp. lactis BB12:all bacteria ratio ↔ Enterobacteriaceae, Bifidobacterium, Lactobacillus, total bacteria	—	↔ OTU ↔ PCoA of taxonomic dissimilarity	—
Anzawa et al,39 Japan	↑ Total Bifidobacterium ↑ B. lactis, B. longum, and B. adolescentis	—	—	—
Neyrinck et al,40 Belgium	↔ All taxonomic levels ↔ Fecal albumin ↔ Plasma iFABP	—	↔ Simpson Index ↔ Chao1 Index ↔ PCoA Bray-Curtis dissimilarity	↓ IL-6, IL-8, IL-17a, and IFN-γ ↔ IL-1β, IL-10, MCP-1, and TNF-α
Elderly adults
Shioiri et al,41 Japan	↓ Enterobacteriaceae ↑ Bifidobacterium and Lactobacillus	—	—	—
Lee et al,56 Korea	1 week: ↓ Bacteroidetes and Firmicutes ↑ Bifidobacterium and B. longum 4 weeks: ↑ B. animalis and B. pseudolongum ↑ L. Paracasei and L. plantarum	—	—	—
Ouwehand et al,42 Finland	↑ Bifidobacterium and L. acidophilus	—	—	↔ TNF-α and IgA
Macfarlane et al,43 UK	↑ Total Bifidobacterium ↑ Firmicutes and Actinobacteria ↓ Proteobacteria ↔ Total bacteria and Bacteroidetes ↑ B. longum, B. adolescentis, B. angulatum, B. bifidum, B. catenulatum.	↑ Acetate ↑ Butyrate ↔ Propionate ↑ Total SCFAs	—	↓ IL-6, IL-8, and MCP-1 at 2 wk (no change at 4 wk) ↓ TNF-α ↔ CRP, IL-10, IgA, and IgG
Lehtinen et al,44 UK	↔ All microbes assessed	↔ Acetate ↔ Propionate ↔ Butyrate	—	↔ IL-1β, IL-6, IL-8, and IL-10 ↔ IFN-γ, TNF-α, and salivary IgA
Costabile et al,45 UK	Synbiotic 1: ↑ Parabacteroides ↓ Desulfovibrio	—	PCoA of Bray-Curtis dissimilarity showed microbial shifts towards higher Parabacteroides and Ruminococcaceae abundance in both synbiotic groups.	↓ hs-CRP and IL-8 ↑ IL-6
	Synbiotic 2: ↑ Parabacteroides and Ruminococcaceae incertae sedis ↓ Desulfovibrio & Oscillospira			↓ hs-CRP and IL-8 ↑ IL-6
Type 2 diabetes
Kassaian et al,46 Iran	↔ All microbes assessed	—	—	—
Horvath et al,47 Austria	↔ All taxonomic levels ↓ Zonulin at 3 mo (no change at 6 mo)	—	↔ Shannon Index ↔ Chao1 index ↔ PCoA Bray-Curtis dissimilarity	—
Kanazawa et al,48 Japan	↑ Total bacteria and Bifidobacterium ↑ Lactobacillus and Actinobacteriota ↓ Firmicutes and Proteobacteria ↓ Verrucomicrobiota and Bacteroidota ↓ Akkermansia muciniphila	↑ Acetate ↑ Butyrate	↓ Phylogenetic Diversity, OTU and Shannon Index at 12wks (no changes at 24wks)	↔ hs-CRP and IL-6
Overweight adults
Krumbeck et al,49 USA	Synbiotic 1: ↑ Total Bifdobacterium ↑ B. adolescentis IVS-1 ↑ Actinobacteria ↓ Roseburia	—	↔ α-Diversity and β-diversity (measures unspecified)	—
	Synbiotic 2: ↑ Total Bifdobacterium ↑ B. animalis subsp. lactis BB-12 ↑ Actinobacteria	—	—	—
Hibberd et al,50 USA	↓ Zonulin ↑ Bacteroidetes and Parabacteroides ↑ Methanobrevibacter and Akkermansia ↑ Lactobacillus and Bifidobacterium	↔ Acetate ↔ Propionate	↑ Shannon Index ↑ Phylogenetic diversity PCoA of weighted UniFrac distance showed significant clustering by synbiotic.	↓ hs-CRP
Stenman et al,51 Finland	↔ Zonulin	—	—	↔ hs-CRP and IL-6 ↑ LPS
Angelino et al,52 Italy	↑ F. prausnitzii and Prevotella ↓ Bifidobacterium	—	—	↓ hs-CRP ↑ IL-6, IL-10, and TNF-α
Asthma and migraines
McLoughlin et al,53 Australia	↑ Bifidobacterium and Anaerostipes ↑ B. adolescentis and B. longum ↔ Erysipelotrichaceae and Roseburia	↔ Total SCFAs ↔ Acetate ↔ Butyrate ↔ Propionate	↔ Phylogenetic diversity ↔ Shannon Index ↔ Simpsons evenness ↔ PCoA Bray-Curtis dissimilarity	—
Ghavami et al,54 Iran	↓ Zonulin	—	—	↓ hs-CRP

Abbreviations: F/B, Firmicutes/Bacteroidetes; hs-CRP, high-sensitivity C-reactive protein; iFABP, intestinal fatty-acid binding protein; IFN-γ, interferon-γ; IgA, immunoglobulin A; IgG, immunoglobulin G; IL, interleukin; LPS, lipopolysaccharide; OTU, operational taxonomic unit; PCoA, principal coordinates analysis; SCFA, short-chain fatty acid; TNF, tumor necrosis factor.

Statistical analysis

Meta-analyses were performed where 3 or more studies reported on an outcome and where medians/means with SDs could be obtained or calculated. Stata version 17 software (release 17; StataCorp LLC) was used to perform random-effects meta-analyses using the meta esize command for weighted Hedge’s g standardized mean difference (SMD) and meta set command for weighted mean difference (WMD). The random-effects model was chosen to address statistical heterogeneity due to the assumption of real differences in the treatment effect related to the differences in study populations, synbiotic intervention, and duration.⁵⁸ Restricted maximum likelihood to estimate heterogeneity variance was used, as recommended by Langan et al.⁵⁹

The WMD (with 95% CI) was computed from mean change or final values, where possible, and Hedge’s g weighted SMD (with 95% CI) where measurement methods hindered a weighted mean difference calculation. Crossover trials were initially analyzed the same way as parallel trials, using a paired analysis. This approach has been noted as likely to be conservative according to the Cochrane Handbook,²⁷ and was required due to a lack of available data on the results of the first period of crossover trials. Where parallel trials included 2 separate synbiotic interventions, these interventions were included as separate studies and labeled “A” and “B.”⁴⁹ To minimize unit-of-analysis error, the control group population was split between the 2 intervention arms. Paired analyses of crossover studies were performed with correlation coefficients of 0.25, 0.5, and 0.75, as a sensitivity analysis to determine if the paired analyses underweighted the crossover studies.²⁷ In addition, “leave-one-out” sensitivity analyses were conducted to explore the effect of removing each individual study from the overall meta-analysis estimates. For calculating the standard error of the difference for sensitivity analyses, where a crossover trial had a different number of participants who completed the intervention and control due to dropout, the highest population number was considered. The I² test statistic estimated the proportion of total variation that was attributable to between-study heterogeneity, with a score of 30%–60% taken to indicate moderate heterogeneity, 50%–90% substantial heterogeneity, and 75%–100% considerable heterogeneity.²⁷

Quality assessment

The Cochrane Collaboration Risk of Bias Tool 2.0⁶⁰ was used to determine the risk of bias in the randomized controlled trials, with the effect of assignment to intervention considered. Five domains were assessed for risk of bias, including (1) randomization process, (2) deviations from intended interventions, (3) missing outcome data, (4) measurement of outcome, and (5) selection of reported results. An additional domain—period and carryover effects—was assessed in crossover trials. The risk-of-bias assessment was appraised by 1 reviewer (D.J.C.).

The quality of the body of evidence was determined by 3 reviewers independently (D.J.C., K.C., K.L.) using Grading of Recommendations Assessment, Development and Evaluation (GRADE).⁶¹ (GRADEpro GDT: GRADEpro Guideline Development Tool [software]; McMaster University and Evidence Prime, 2022. Available from gradepro.org).

RESULTS

A total of 782 articles were identified from the systematic search strategy, reference lists, and scoping search. After removal of duplicates, 731 articles were screened and 60 full-text articles were reviewed for eligibility. Thirty-two of these articles were excluded, with the most common reasons being that they did not investigate an oral synbiotic in supplement or food form, were not conducted in a healthy adult population, did not include a randomized or quasi-randomized controlled trial, or that the article was a study protocol. This resulted in 27 articles reporting on 26 studies being included in this review (n = 1319) (Fig. 1).⁶² From 17 of these articles reporting on 16 studies, 11 effect sizes were available for inclusion in the meta-analysis.

Figure 1.

Flow diagram of the literature search process. Abbreviation: PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Study characteristics

Characteristics of the included studies are outlined in detail in Table 2. Eighteen studies had a parallel design,^{32–34,36–38,40,42,46–52,54,55} and 9 studies had a crossover design.^{30,31,35,39,41,43–45,53} All included studies were randomized controlled trials. The duration of included studies ranged from 1 week⁵³ to 6 months.^{46–48,50,51} Five studies^{30,32,34,38,50} collected additional outcome data 5 days³⁴ to 1 month⁵⁰ postintervention. Studies were conducted in Australia,⁵³ Italy,^37,52,55 the United States,^30,50 Korea,^36,56 Denmark,³¹ the United Kingdom,^{32,35,43–45} India,³³ Brazil,³⁴ Slovenia,³⁸ Japan,^39,41,48 Belgium,⁴⁰ Finland,^42,51 Iran,^46,54 and Austria.⁴⁷ Studies included participants with constipation,^36–40 healthy adults,^30–35,55 older adults,^41–45,56 as well as those who were overweight or obese,^42,49,50,52 or had type 2 diabetes or prediabetes,^46–48 asthma,⁵³ or chronic migraines.⁵⁴

The synbiotic interventions were either supplemented as a capsule,^{31–33,35–37,39,42–47,49–51,53,54} yogurt,³⁰ ice cream,³⁴ milk beverage,^38,41,55,56 powder,⁴⁰ or pasta.⁵² All probiotic interventions contained Lactobacillus and/or Bifidobacterium, either a single strain^{31,33,35,39–45,49–52} or multi-strain.^{32,34,36–38,46,48,53–56} The most common species in single- and multi-strain probiotics were Lactobacillus acidophilus,^{31,32,34,37,38,46,47,53–55} Lactobacillus rhamnosus,^{36,37,42,45,53,54} and Bifidobacterium animalis.^{30,32,34–36,38–40,44,46,47,49–51,53–56} Four studies included strains not from the Bifidobacterium or Lactobacillus genera in the multi-strain probiotic. These were Lactococcus lactis^36,47 and Streptococcus thermophilis.^30,54 One study included Bacillus coagulans BC30 as the single probiotic strain.⁵² The most common prebiotic interventions were inulin,^{30,34,38,39,43,46,53,55} GOS,^{41,44,47–49} FOS,^{32,33,38,40,43,47,54,56} and xylo-oligosaccharide.^35,36,56

Microbiota analysis techniques included bacterial cell counts (n = 3),^33,41,55 qPCR cell counts (n = 5),^{30,39,42,46,52} 16S rRNA gene sequencing (n = 3),^34,36 intestinal permeability reported as urinary lactulose-to-mannitol ratio (n = 1),³² or zonulin (n = 2),^32,51,54 whereas the majority of studies utilized a combination of methodologies (n = 14).^{31,35,37,38,40,43–45,47–50,53,56} This variation in testing resulted in a combination of quantitative bacterial counts with various measurement units, relative abundance data, and qualitative descriptions of microbiome changes reported in study outcomes. The substantial heterogeneity in outcome data hindered the ability to perform meta-analyses on microbiota outcomes, except for Lactobacillus, Bifidobacterium, zonulin, α-diversity, and SCFAs.

Risk of bias

The risk-of-bias assessment is shown in Fig. 5 and Figs. S5 and S6 (see the Supporting Information online). No studies were considered at low risk of bias. The risk of bias was considered high in 13 studies and as some concerns noted in 14 studies. The high risk of bias was mostly related to concerns with deviations from the intended intervention due to failure to analyze data according to intention-to-treat (domain 2) and missing outcome data (domain 3).

Figure 5.

Risk of bias as a weighted proportion of total studies.

Effect of synbiotics on study outcomes

The number of studies and effect sizes, and the results of each meta-analysis, are shown in Table 4 and Figs. 2–4. Summary data for each study are available in Table S1 (see the Supporting Information online). In line with the Cochrane Handbook, due to the small number of studies included in the meta-analyses, subgroup analyses were not performed.²⁷

Table 4.

Changes in outcomes following synbiotic consumption, compared with control

Outcome	No. of studies and references	Effect sizes, n	Participants, n	Results			Inconsistency
				Effect size (WMD or SMD)	95% CI	P	I2 (%)	P
Bifidobacterium relative abundance (%)	334,49,53	4	146	0.97 WMD	−0.20, 2.14	0.10	93.0	<0.001
Bifidobacterium cell count	831,37,41–43,48,52,55	8	378	0.82 SMD	−0.03, 1.67	0.06	93.1	<0.001
Lactobacillus cell count	431,37,48,55	4	169	0.74 SMD	0.15, 1.33	0.01	69.4	0.02
Zonulin (ng/mL)	347,50,54	3	168	−0.24 WMD	−4.34, 3.86	0.91	90.5	<0.001
Shannon Index	334,48,50	3	218	0.02 WMD	−0.21, 0.26	0.84	86.3	<0.001
Acetate concentration	634,41–43,51,53	6	372	0.27 SMD	−0.02, 0.57	0.07	49.1	0.07
Propionate concentration	634,41–43,51,53	6	372	0.22 SMD	0.02, 0.43	0.03	0.0	0.79
Butyrate concentration	634,41–43,51,53	6	372	0.17 SMD	−0.03, 0.37	0.10	0.0	0.62
IL-6 (pg/mL)	543,45,48,51,52	5	390	−0.76 WMD	−2.21, 0.70	0.31	77.6	0.06
CRP (mg/L)	543,50–52,54	5	378	−0.12 WMD	−0.51, 0.26	0.54	44.6	0.54
Endotoxin (EU/mL)	332,47,50	3	129	−0.00 WMD	−0.03, 0.02	0.79	52.3	0.05

Abbreviations: CI, confidence interval; CRP, C-reactive protein; IL, interleukin; SMD, standardized mean difference; WMD, weighted mean difference.

Figure 2.

Forest plot of weighted standardized mean differences or standardized mean difference and 95% CIs for synbiotics compared with control. (A) Lactobacillus cell count; (B) Bifidobacterium cell count; and (C) Bifidobacterium relative abundance. Abbreviations: CI, confidence interval; REML, restricted maximum likelihood.

Figure 4.

Forest plot of weighted mean difference and 95% CIs for synbiotics compared with control. (A) Interleukin-6; (B) CRP; and (C) endotoxin. Abbreviations: CI, confidence interval; CRP, C-reactive protein; REML, restricted maximum likelihood.

Changes in gut microbiota composition

A narrative synthesis of all studies reveals that the only consistent change to the gut microbiota after a synbiotic intervention, seen across all population subtypes, was an increase in Lactobacillus and/or Bifidobacterium relative abundance or cell count measures. A few changes to bacterial phylum relative abundance were reported, with an increase in Firmicutes seen in 1 study after supplementation with Bifidobacterium for 1 month,⁴³ but a decrease was seen by 2 other studies after supplementation with a probiotic containing members of the Firmicutes phylum, specifically Lactobacillus or Lacticaseibacillus for 1 week or 6 months, respectively.^48,56 These 3 studies displayed some concerns of risk of bias. An increase in Actinobacteria phylum was reported by 3 studies after supplementation with a probiotic containing Bifidobacterium for 3 weeks,⁴⁹ 1 month,⁴³ or 6 months⁴⁸; Proteobacteria⁴³ and Bacteroidetes^48,51,56 abundance decreased after synbiotic intervention in 2 studies. A reduction in bacterial genus or family from the Clostridiales order was seen in 2 studies among a healthy adult population after a 3-week intervention,^30,31 yet Childs et al³⁵ reported no change to this bacterial order among healthy adults after the same duration of intervention. Two of these studies displayed some concerns of risk of bias^31,35 and 1 study displayed a high risk of bias.³⁰

Lactobacillus

A total of 4 studies exploring the effect of synbiotic consumption on Lactobacillus cell count were included in the meta-analysis.^31,37,48,55 Each study included a Lactobacillus strain as the probiotic, with the prebiotic and duration of intervention being inulin for 4 weeks,⁵⁵ cellobiose for 3 weeks,³¹ psyllium for 2 months,³⁷ or GOS for 6 months.⁴⁸ Synbiotic consumption resulted in a statistically significant increase in the Lactobacillus cell count, with an SMD of 0.74 (95% CI: 0.15, 1.33; P = 0.01) (Table 4, Fig. 2). The effect estimate was similar after using different correlation coefficients (0.25, 0.5, and 0.75) (see Table S3 in the Supporting Information online). Sensitivity analyses revealed that excluding Kanazawa et al⁴⁸ and Casiraghi et al⁵⁵ led to a nonsignificant effect size of an SMD of 0.55 (95% CI: -0.15, 1.25; P = 0.13) and an SMD of 0.59 (95% CI: -0.12, 1.29; P = 0.10), respectively (see Fig. S1 in the Supporting Information online).

Bifidobacterium

A total of 11 studies exploring the effect of synbiotic consumption on Bifidobacterium relative abundance or cell count were included in the meta-analysis.^{31,34,37,41–43,48,49,52,53,55} Seven of these studies supplemented with Bifidobacterium with or without Lactobacillus.^{34,37,43,48,49,53,55} The prebiotic and duration of interventions were as follows: inulin for 1 week⁵³ or 4 weeks^34,43,55; cellobiose for 3 weeks³¹; psyllium for 2 months³⁷; GOS for 2 weeks,⁴¹ 3 weeks,⁴⁹ or 6 months⁴⁸; lactilol for 2 weeks⁴²; FOS for 4 weeks⁴³; or β-glucan for 3 months.⁵² The meta-analysis effect estimates showed that synbiotic consumption did not result in significant differences in Bifidobacterium relative abundance or cell count (Table 4, Fig. 2). However, the effect estimate for Bifidobacterium cell count became significant after using correlation coefficients 0.25, 0.5, and 0.75 (see Table S2 in the Supporting Information online). Sensitivity analyses showed substantial study effects when omitting McLoughlin et al⁵³ from Bifidobacterium relative abundance (WMD: 1.47; 95% CI: 0.42 to 2.52; P = 0.006) and omitting Bazzocchi et al³⁷ from Bifidobacterium cell count (SMD: 1.01; 95% CI: 0.13, 1.88; P = 0.024) (see Figs. S2 and S3 in the Supporting Information online).

Change in intestinal permeability

Among the 4 randomized controlled trials assessing zonulin as a marker of intestinal permeability, 3 reported a significant reduction after a 3-month⁵⁴ or 6-month intervention^47,50 and 1 reported no change after a 3-month intervention.⁵¹ Of the studies reporting a significant reduction, 2 displayed a high risk of bias^47,54 and the other reported some concerns of risk of bias.⁵⁰ The study reporting no change displayed some concerns of risk of bias.⁵¹ The meta-analysis effect estimate showed that synbiotic consumption did not result in significant changes in zonulin (Table 4, Fig. 3).

Figure 3.

Forest plot of weighted mean difference or standardized mean difference and 95% CIs for synbiotics compared with control. (A) Zonulin; (B) Shannon Index; (C) acetate concentration: (D) propionate concentration; and (E) butyrate concentration. Abbreviations: CI, confidence interval; REML, restricted maximum likelihood.

Change in α-diversity

An increase in α-diversity, measured as the Shannon Index and Phylogenetic Diversity, was reported by Hibberd et al⁵⁰ after a 6-month intervention. In contrast, 2 studies found a decrease in various α-diversity measurements after a 1-month³⁶ or 6-month⁴⁸ intervention; however, for Kanazawa et al⁴⁸ this reduction was only present at 12 weeks and not 24 weeks. Two of these studies displayed some concerns of risk of bias,^36,50 and 1 displayed a high risk of bias.⁴⁸

The meta-analysis effect estimate showed that synbiotic consumption did not result in significant changes in the Shannon Index (Table 4, Fig. 3). Sensitivity analysis showed substantial changes when omitting Valle et al³⁴ from Shannon Index assessments (WMD: 0.13; 95% CI: 0.01, 0.25; P = 0.031) (see Fig. S5 in the Supporting Information online).

Changes in SCFAs

The meta-analysis effect estimate showed that synbiotic consumption did not result in significant differences in the SCFAs acetate and butyrate (Table 4, Fig. 3). Inconsistencies in the reported SCFA outcomes were seen across all included studies. Five studies reported no significant changes to measured SCFAs after synbiotic supplementation for 1 week,⁵³ 3 weeks,^31,35,44 or 6 months.⁵⁰ Three studies reported significant increases in SCFA production after supplementation for 1 month^34,43 or 6 months.⁴⁸ Of the studies reporting no significant change in SCFAs, 4 displayed some concerns of risk of bias^31,35,50,53 and 1 displayed a high risk of bias.⁴⁴ Of the 3 studies reporting a significant increase in SCFAs, 2 displayed a high risk of bias^34,48 and 1 displayed some concerns of risk of bias.⁴³

Propionate

A total of 6 studies exploring the effect of synbiotic consumption on propionate concentration were included in the meta-analysis.^{34,41–43,51,53} The duration of interventions ranged from 1 week⁵³ to 6 months.⁵¹ Synbiotic consumption resulted in a statistically significant increase in the propionate concentration, with an SMD of 0.22 (95% CI: 0.02, 0.43; P = 0.03) (Table 4, Fig. 3). The effect estimate was no longer significant after using different correlation coefficients (0.25, 0.5, and 0.75) (see Table S3 in the Supporting Information online). Sensitivity analyses revealed that excluding Macfarlane et al⁴³ and Stenman et al⁵¹ led to a nonsignificant effect size of an SMD of 0.18 (95% CI: -0.05, 0.41; P = 0.12) and an SMD of 0.16 (95% CI: -0.08, 0.40; P = 0.18), respectively (see Fig. S7 in the Supporting Information online).

Changes in inflammatory markers

There were significant reductions in proinflammatory cytokines and inflammatory biomarkers reported by 8 studies across 6 population subtypes.^{32,33,40,43,45,50,52,54} The duration of interventions ranged from 3 weeks⁴⁵ to 6 months.⁵⁰ Four of these studies displayed a high risk of bias^32,33,52,54 and 4 displayed some concerns of risk of bias.^40,43,45,50 The meta-analysis estimates showed that synbiotic consumption did not result in a significant difference in interleukin (IL)-6, C-reactive protein (CRP), or endotoxin (Table 4, Fig. 4). However, the effect estimate for IL-6 was significant when using the correlation coefficient of 0.5 (see Table S3 in the Supporting Information online). Sensitivity analyses showed substantial changes when omitting Hibberd et al⁵⁰ from CRP estimates (WMD: −0.28; 95% CI: −0.45, −0.12; P = 0.001) (see Fig. S4 in the Supporting Information online).

Quality of the body of evidence

The quality of the body of evidence was determined using GRADE⁶¹ (see Table S4 in the Supporting Information online). The quality of the body of evidence was “very low” for Bifidobacterium relative abundance, Lactobacillus and zonulin due to downgrades for risk of bias, inconsistency, and imprecision. The quality of the body of evidence was “low” for Bifidobacterium cell count, CRP, and acetate due to downgrades for both risk of bias and inconsistency. The quality of the body of evidence was “moderate” for the Shannon Index, propionate, and butyrate after being downgraded only for risk of bias.

DISCUSSION

To the authors’ knowledge, this systematic review and meta-analysis is the first to investigate the impact of synbiotic intake on the gut microbiota and inflammatory biomarkers in populations of adults with no significant clinical disease impacting the gut microbiota and/or inflammation. The results of the analysis of randomized controlled trials suggest that synbiotic interventions may increase the abundance of Lactobacillus cell count and propionate concentration but likely have no significant effect on zonulin, α-diversity measured as Shannon Index, acetate, butyrate, IL-6, CRP, or endotoxin concentration. When interpreting these findings, it should be noted that approximately half of the included studies displayed a high risk of bias, while some concerns regarding the risk of bias were noted in the remaining studies. The narrative synthesis of all included randomized trials suggests that synbiotics may also increase the abundance of Bifidobacterium measures. The nonsignificant meta-analysis effect estimate for Bifidobacterium cell count should be interpreted with caution, due to a significant effect seen when using correlation coefficients of 0.25, 0.5, and 0.75 for crossover studies, as well as sensitivity analyses revealing a substantial study effect by Bazzocchi et al.³⁷ This study was the only included study with a negative Hedge’s g SMD value (-0.55; 95% CI: -1.28, 0.18), likely explaining why its exclusion resulted in a significant Bifidobacterium cell count estimate of an SMD of 1.01 (95% CI: 0.13, 1.88; P = 0.02).

Given the fact that all probiotic supplements contained single or multiple strains from Lactobacillus and Bifidobacterium genera, it is not surprising to see a consistent increase in these 2 genera. Yet, these changes are not evident when supplementing with probiotics alone. A systematic review examining alterations in fecal microbiota composition among healthy adults found no observed changes to fecal microbiota bacterial profile or measures of α-diversity after probiotic supplementation.¹⁰ The results from this review therefore suggest that synbiotics can modulate the gut microbiota in healthy adult populations, in contrast to the cumulative evidence for probiotic interventions alone. Given that only 5 studies collected data during a follow-up period, the results of this systematic review are unable to identify whether synbiotics offer a long-lasting or only a transient effect on the human gut microbiome. Three of these follow-up studies found that none of the significant differences in the measured gut microbiome outcomes had remained during the follow-up period.^30,34,38 One month after a 6-month intervention period, Hibberd et al⁵⁰ reported an increased abundance of Lactobacillus in the intervention group. Roberts et al³² reported that 6 days after finishing a 3-month synbiotic intervention, participants continued to display a significantly reduced plasma endotoxin level; however, during this time, participants also completed an endurance triathlon event. Probiotic intervention trials have similarly shown inconsistencies regarding the duration of lasting effects on the gut microbiome.^63,64 It is uncertain from findings of this review whether synbiotics offer any advantage over probiotics in this regard.

The microbiota-modulating effects demonstrated by synbiotics in this review have similarly been shown by prebiotic interventions. The prebiotics most supplemented in this review—inulin, GOS, and FOS—have consistently demonstrated the ability to increase Bifidobacterium and Lactobacillus bacteria.^65–67 GOS and FOS, at doses of 5–10 g and 2.5–10 g, respectively, have increased Bifidobacterium populations in healthy adult populations, after as few as 7 days of supplementation.^65–67 Additionally, the systematic review and meta-analysis by So et al⁶⁸ examining the effect of dietary fiber interventions on the gut microbiota in healthy adults reported a significant increase in Lactobacillus and Bifidobacterium compared with placebo or low-fiber comparators. Subgroup analyses identified that prebiotic dietary fibers resulted in a significantly greater increase in Bifidobacterium compared with comparators. The systematic review by Ferro et al⁶⁹ examined the effect of prebiotics, probiotics, and synbiotics on the gut microbiota and health outcomes in infants, and reported an increase of fecal Bifidobacterium measures after prebiotic and synbiotic supplementation, but not after probiotics. These results further suggest that prebiotics, and therefore synbiotics, have a greater capacity to alter the microbiota than probiotic supplementation. Given that prebiotics have demonstrated the ability to modulate the gut microbiota of healthy adults in a similar manner to the synbiotic interventions included in this review,^65–68 it is difficult to ascertain if the microbiota-modulating effects reported in this study are due to the unique synbiotic combination or due to the prebiotics alone.

The human gut microbiota is a stable and resilient ecosystem.⁷⁰ It is therefore not surprising that introducing a bacterial species alone, in the form of probiotics in a transient manner, does not substantially alter the microbiota. Additionally, despite dietary interventions having demonstrated the ability to rapidly alter the human gut microbiota, these alterations revert to the original structure after ceasing dietary interventions.⁷¹ Although probiotics introduce new species, prebiotics when supplemented alone provide a substrate for pre-existing microbial species. Given that all humans possess Bifidobacterium and Lactobacillus genera in their gut microbiota, providing substrates that feed these endogenous bacteria may be a more effective strategy than introducing exogenous bacterial species into a resilient ecosystem. It is widely recognized that Bifidobacterium and Lactobacillus are healthy gut bacteria, having demonstrated key roles in the production of SCFAs, immune system function, maintaining stability and diversity of the gut microbiota, and providing various clinical benefits.^72–74 Interventions to increase these bacterial genera are widely favored, particularly Bifidobacterium due to the demonstrated decrease in this genus throughout adulthood compared with infancy.⁷⁵ Thus, the proliferation of these beneficial bacteria demonstrated in this review is a positive health-enhancing result.

The lack of change in α-diversity seen in this meta-analysis, a measure of the richness and evenness of bacterial species, is not unexpected given the results of previous studies. The reviews by So et al⁶⁸ and Kristensen et al¹⁰ similarly did not find a significant change in α-diversity after prebiotic and probiotic supplementation in healthy adults, respectively. Additional reviews reporting on changes in α-diversity after microbiota-modulating therapies further demonstrate inconsistencies. A recent systematic review²⁰ found no significant change in α-diversity measures after probiotic or synbiotic supplementation in adults with major depressive disorder. In contrast, a meta-analysis reported a significant increase in α-diversity after weight-loss interventions, yet this significant increase was not present when pooling the effect of food-based interventions alone.⁷⁶ Similarly, another systematic review found minimal changes in α-diversity after food-based weight-loss interventions.⁷⁷ These findings align with the discussion by So et al⁶⁸ that current literature suggests that short-term dietary or microbiota interventions are unable to alter microbial diversity,^71,78,79 yet observational studies assessing habitual dietary intake show positive correlations between dietary diversity and microbiota diversity.^80,81 Thus, perhaps assessing changes in α-diversity from short-term studies is not a useful indicator of the overall health of the human gut microbiota, given the stability and resilience of the microbial ecosystem.⁸² Coyte et al⁸³ provide a topical discussion on the interactions between diversity, microbial competition, and microbial stability, demonstrating that microbial diversity does not always correlate with a stable and therefore healthy microbiome. In light of this ecological concept, human gut microbiota research may benefit from taking greater care to consider measures of microbial diversity as indicators of microbiome health.⁸⁴

This review reported no significant reductions to any inflammatory markers after synbiotic supplementation, which is somewhat in contrast to previous literature. Overweight, obesity, prediabetes, and diabetes, collectively referred to as the metabolic syndrome, are associated with a low-grade inflammatory response due to metabolic endotoxemia.⁸⁵ This response is closely related to alterations in the intestinal microbiota.⁸⁶ One meta-analysis found a significant reduction in CRP by an SMD of -0.38 (P = 0.00) after synbiotic or probiotic supplementation in adults with diabetes⁸⁷; however, another meta-analysis found no significant change in any inflammatory markers after probiotic supplements in adults with diabetes.⁸⁸ Anti-inflammatory effects after probiotic, prebiotic, and synbiotic supplementation have additionally been demonstrated in adults who are overweight or obese.^89,90 This review synthesized the data from 7 studies of adults with overweight, obesity, or diabetes,^46–52 with the remaining 20 studies including participants not known to have raised inflammatory markers at baseline. Subgroup analyses were not able to be performed in these 7 studies due to the small number of studies included in each meta-analysis effect estimate. It is likely that changes to inflammatory markers after synbiotic supplementation were not see in this review due to the majority of included studies examining adults not known to have raised inflammatory biomarkers and a relatively small sample size of only 5 studies assessing CRP^{43,50–52,54} and/or IL-6^{43,45,48,51,52} included in the meta-analysis.

Short-chain fatty acids, particularly butyrate, have demonstrated anti-inflammatory effects as butyrate is the main substrate for colonic epithelium barrier cells, thereby preserving the integrity of this barrier.²⁶ In vitro studies have demonstrated that SCFAs can modulate the expression of inflammatory cytokines by colonic epithelial cells.²⁶ Additionally, SCFAs may have a direct impact on immune cells by modulating cell signaling and metabolism, which have effects on the innate and adaptive immune system.²⁶ This review demonstrated a significant increase in propionate SCFAs after synbiotic supplementation, but not acetate or butyrate, and the data were not available to pool total SCFA production. The varied findings of the effect of synbiotics on SCFAs in healthy adults in this review are in line with previous literature. The recent systematic review by Vinelli et al⁹¹ investigating the effect of dietary fiber interventions in healthy adults found that 7 studies reported a significant increase in total SCFAs, while 5 studies reported no changes in total SCFAs. Authors of that review concluded that dietary fibers, including prebiotics, do not produce unequivocal increases in SCFAs in healthy adults.⁹¹

Strengths

All studies included in the meta-analysis were randomized controlled trials and therefore provide the highest degree of evidence. This review also followed a robust methodology including prospective registration of the review protocol and duplicate initial screening of the included articles. Additionally, a comprehensive search was conducted at 2 time points to maximize the likelihood that all eligible trials were included. Considering the heterogeneity in intervention type and population, data were analyzed using a random-effects model to provide a more conservative average effect of the intervention.

Limitations

Various limitations in this review warrant discussion and consideration. Despite all included studies assessing synbiotic interventions, variation in the type of prebiotics and probiotics, doses, and duration resulted in heterogeneity. Studies ranged from 1 week to 6 months in duration, and due to the lack of subgroup analyses, it was not possible to identify an optimal duration for supplementation. Although all included study participants were considered healthy adults who had no significant clinical disease impacting the gut microbiota and/or inflammatory conditions, population characteristics still varied considerably and 6 population subtypes were identified. Moreover, heterogeneity regarding the microbiota measurements and reporting existed among the included studies. Studies with next-generation sequencing, cell culturing, and non-sequencing molecular-based methods were included in this review, resulting in differences in the bacterial identification and specificity.⁹² This substantial heterogeneity, particularly due to synbiotic type and population characteristics, limits the translational and clinical application of the evidence. This systematic review is, therefore, unable to provide evidence for certain synbiotic combinations that offer greater effects on the gut microbiota or inflammation in healthy adult populations.

Additional limitations include the high risk of bias present in 13 included studies, particularly bias related to deviations from the intended intervention and missing outcome data, which may have affected the overall results. Further, due to the small sample sizes, funnel plots to test for publication bias were not conducted. The quality of the body of evidence was downgraded in all assessed outcomes (evaluated using GRADE⁶¹) due to concerns with risk of bias, inconsistency, and imprecision. This further confirms the inappropriateness of the current body of evidence presented in this review for translation into clinical practice.

Recommendations

To reduce the heterogeneity and improve the specificity of results, further reviews pooling data only from next-generation sequencing methods are needed. Additionally, further high-quality randomized controlled trials exploring different doses and durations of synbiotics among homogenous healthy adult populations are needed to identify dose–response relationships and the most effective synbiotic regimens. Further well-designed studies that examine differing lengths of supplementation for the same dose and combination of synbiotics are needed to inform evidence-based clinical recommendations. Consideration also needs to be given to the desired microbiota outcome. Given the complexities of human gut microbiota research involving methodological differences and interdisciplinary research teams, standardized reporting guidelines are necessary to promote consistency and reproducibility. The STORMS (Strengthening The Organisation and Reporting of Microbiome Studies) tool, published in 2021, is a 17-item checklist providing guidance for concise and complete reporting of human microbiome studies.⁹³ Widespread adoption of this reporting guideline may reduce between-study heterogeneity and facilitate future comparative analysis of microbiome research. As outlined in the STORMS tool, it is essential for authors to publish or make available all measured quantitative data from human microbiome clinical trials to ensure that results from eligible studies can be pooled in meta-analyses. To reduce the risk of bias in future clinical trials, intention-to-treat analysis is required, with statistical analytical methods being prespecified in clinical trial registrations or protocols, and the reporting of missing outcome data to improve transparency.

CONCLUSION

This systematic review and meta-analysis of the effects of synbiotics on the gut microbiota and inflammatory biomarkers in healthy adults found evidence for a favorable effect on Lactobacillus cell count and propionate concentration, suggesting modulation of the gut microbiota. However, the nonsignificant differences in Bifidobacterium, α-diversity, zonulin, acetate, and butyrate suggest a limited effect on the gut microbiota composition, structure, and function. The nonsignificant effect estimate for Bifidobacterium measures should be interpreted with caution due to the significant effect size seen with differing correlation coefficients and when excluding 2 studies. Additionally, the nonsignificant differences in CRP, IL-6, and endotoxin concentrations suggest a lack of consistent evidence for the effects of synbiotics on inflammation in healthy adult populations. This is the first review to date to provide evidence for the microbiota-modulating effect of various synbiotic doses and combinations among healthy adult populations, and highlights the need for continued exploration of synbiotics as microbiota-modulating therapies.

Supporting Information

The following Supporting Information is available through the online version of this article at the publisher’s website.

Appendix S1 Full search strategy

Table S1 Summary data for meta-analyses

Table S2 Sensitivity analyses using varying correlation coefficients for cross-over studies

Table S3 Justification for risk of bias judgments

Table S4 GRADE assessment of the quality of the body of evidence

Table S5 PRISMA Checklist

Figure S1 Effect for estimate of Lactobacillus cell count if one study is omitted

Figure S2 Effect for estimate of Bifidobacterium relative abundance (%) if one study is omitted

Figure S3 Effect for estimate of Bifidobacterium cell count if one study is omitted

Figure S4 Effect for estimate of zonulin (ng/mL) if one study is omitted

Figure S5 Effect for estimate of Shannon Index if one study is omitted

Figure S6 Effect for estimate of acetate concentration if one study is omitted

Figure S7 Effect for estimate of propionate concentration if one study is omitted

Figure S8 Effect for estimate of butyrate concentration if one study is omitted

Figure S9 Effect for estimate of interleukin-6 (pg/mL) if one study is omitted

Figure S10 Effect for estimate of CRP (mg/L) if one study is omitted

Figure S11 Effect for estimate of endotoxin (EU/mL) if one study is omitted

Figure S12 Risk of bias assessment summary