Systemic inflammation is associated with the development and progression of a number of chronic health conditions. These include the leading causes of mortality worldwide: ischemic heart disease, dementia, cerebrovascular diseases, cancer, and chronic lower respiratory diseases (1). Systemic inflammation results from the chronic activation of the immune system and release of proinflammatory mediators, such as C-reactive protein (CRP), TNF-α, and ILs (i.e., IL-6 and IL-1β). An inverse relation between dietary fiber intake and biomarkers of systemic inflammation (i.e., IL-6, TNF-α, and CRP) has been demonstrated in numerous studies (2–7). Although the antiinflammatory mechanisms of dietary fiber are yet to be elucidated, short-chain fatty acids (SCFAs) may be a contributing factor (8).
SCFAs (i.e., acetate, butyrate, and propionate) are physiologically active byproducts primarily produced from the fermentation of soluble dietary fiber and resistant starch by commensal bacteria in the colon (9). A number of factors can influence SCFA production, including the fermentation substrate available. For example, oligosaccharide soluble fibers [(i.e., fructo-oligosaccharides (FOS)] produce a higher SCFA yield than longer-chain polysaccharide soluble fibers (i.e., pectin) (10). SCFAs lower the colonic pH, which can promote the growth of beneficial bacteria, such as Lactobacillus and Bifidobacterium. These bacteria are potent SCFA producers and play a role in maintaining healthy immune responses (4, 11). Soluble fibers, such as FOS, that selectively stimulate the growth and activity of commensal bacteria associated with health benefits to the host, are referred to as prebiotics (10). The composition of the colonic microbiome can also be temporarily altered with the use of probiotics. This involves the direct delivery of live bacteria [commonly strains of Bifidobacteriumand Lactobacillus (12)] to the host, either in supplemental form or via functional foods (e.g., yogurt). Synbiotics are a mixture of prebiotics and probiotics and therefore are considered to have synergistic effects.
It is estimated that ≥90% of SCFAs are absorbed from the intestinal lumen, with the majority either metabolized by colonocytes or delivered to the liver via the hepatic portal vein. A small portion of SCFAs (primarily acetate) also enter systemic circulation where they have the capacity to influence cells within peripheral tissues. One proposed effect of SCFAs is the reduction of systemic inflammation (13) through the modulation of molecular signaling pathways, including the activation of G-protein coupled receptors 41 and 43 (8) and the inhibition of histone deacetylase enzymes (14). Previous reviews have examined the metabolic and immunomodulatory properties of synbiotics and selected prebiotics (primarily inulin and FOS) (15, 16). To our knowledge, this is the first systematic review to examine the effect of SCFAs and a wide range of prebiotics and synbiotics on systemic inflammation in humans.
The electronic databases Medline, EMBASE, PubMed, CINAHL (Cumulative Index to Nursing and Allied Health Literature), and Cochrane were searched for English language articles from 1947 to May 2015 with the use of keywords and Medical Subject Headings of the National Library of Medicine. Hand searching of the reference lists of retrieved articles and relevant systematic reviews was conducted, as well as cited reference searches of retrieved articles through the Web of Science database. The Medical Subject Headings search terms included: dietary fibre, dietary fiber, soluble fibre, soluble fiber, indigestible carbohydrates, fermentable carbohydrates, short-chain fatty acids, volatile fatty acids, butyrate, propionate, acetate, prebiotics, synbiotics, inflammatory markers, IL, CRP, TNF-α, and IL-6. See Figure 1 for an example of the search strategy used. The search was conducted again in 2017 to ensure that any relevant articles published after the initial search were identified.FIGURE 1
Example of search strategy with the use of PubMed for studies investigating the effect of short-chain fatty acids, prebiotics, and synbiotics on systemic inflammation in humans. CRP, C-reactive protein; FOS, fructo-oligosaccharide; GOS, galacto-oligosaccharide; scfa, short-chain fatty acid; vfa, volatile fatty acid.
Article inclusion and exclusion criteria
Studies were included if they examined the effects of SCFAs, prebiotics, and synbiotics delivered orally, intravenously, or per rectum (enema) on any biomarker of systemic inflammation in human participants of any age and sex. Randomized controlled trials (RCTs), quasiexperimental studies, cohort studies, case-control studies, before-and-after studies, and observational cross-sectional studies were included. Exclusion criteria were as follows: animal studies, in vitro studies, studies examining the effects of probiotics alone on inflammation, and studies that did not report on systemic inflammation (i.e., inflammation outside the gastrointestinal tract). Systematic reviews, narrative reviews, opinion papers, and case studies were also excluded.
Article appraisal and data extraction
Studies retrieved by the search strategy were independently assessed for relevance to the review by 2 reviewers (RFM and BSB) based on title, abstract, and full text. When there was a disagreement on the inclusion of a study, a third independent reviewer (LGW) was involved. At each stage, the reasons for exclusion were documented. After the full-text appraisal, all included studies were independently assessed by the 2 reviewers (RFM and BSB) for methodologic quality with the use of a standardized critical appraisal checklist designed by the American Dietetic Association (17). This tool incorporates 4 relevance questions that address the applicability of the study findings to practice and 10 validity questions that address scientific rigor. Based on the responses to these questions as determined by the reviewers (RFM and BSB), each study was rated as having negative, positive, or neutral quality. Studies of negative quality (the response to ≥6 validity questions was “no”) were excluded. The level of evidence for each article was determined according to the study design based on the National Health and Medical Research Council of Australia levels of evidence hierarchy (18).
The following data were extracted from included studies with the use of a standardized data extraction tool: country, participant characteristics, study design, sample size, intervention details (SCFA, prebiotic, or synbiotic composition and dose), treatment duration, assessment of compliance to the intervention, assessment of background dietary intake, and outcomes of interest (means and SDs before and after the supplementation period for each group). Outcomes of interest were plasma and serum inflammatory markers (e.g., ILs, TNF-α, and CRP). When SD values for any outcomes of interest were not reported, they were calculated from the reported SEs or 95% CIs. Included studies were categorized according to the intervention described to assess the evidence relating to SCFAs, prebiotics, and synbiotics.
Meta-analysis was performed with the use of Review Manager (RevMan, version 5.3, Nordic Cochrane Centre). Heterogeneity was examined with the use of the χ2 test (P < 0.1 considered to indicate significant heterogeneity) and the I2 parameter [with 30–60% indicating moderate, 50–90% indicating substantial, and 75–100% indicating considerable heterogeneity (19)]. When considerable heterogeneity was identified, subgroup analyses were conducted to investigate possible contributing factors. Because the studies were considered heterogeneous in relation to the type and dosage of prebiotic or synbiotic supplementation, treatment duration, and study population (i.e., disease status and country of origin), the random effects meta-analysis model was applied to all meta-analyses. The inverse-variance statistical method was used, and the standardized mean difference [(SMD) effect size] and corresponding 95% CIs were calculated. The Cochrane Handbook for Systematic Reviews of Interventions was used to determine whether it was appropriate to include data from crossover studies in the meta-analyses (19). Crossover studies were excluded from the meta-analysis if insufficient data were available to eliminate the possibility of carry-over effects, results from paired analyses were not reported, or data were not reported in a suitable form (i.e., individual participant data or within-patient differences) to allow paired analysis to be approximated. Ex vivo trials were also excluded due to their inherent differences to in vivo studies.
A total of 8609 articles were identified, with 2744 excluded because they were duplicates (Figure 2). The titles of the remaining 5865 articles were reviewed, with 843 articles (∼14%) retrieved for abstract appraisal. Abstracts from 150 articles met the inclusion criteria and the full texts were retrieved for further review. Data extraction and assessment of methodologic quality were performed on the 75 articles that met the review criteria, of which 7 were excluded due to negative methodologic quality (20–26).FIGURE 2
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart of articles for inclusion in a systematic review of the effect of short-chain fatty acids, prebiotics, and synbiotics on systemic inflammation.
Description of included studies
Of the 68 included studies, 65 (95%) were RCTs (n = 18 crossover trials), 2 were observational cross-sectional studies, and 1 was a before-and-after intervention study. The publication year ranged from 2004 to 2017. Twenty-nine studies (43%) were from Europe (27–55), 20 (30%) were from Asia (56–75), 16 (25%) were from America (2, 76–90), and 3 (5%) were from Australia (91–93). The majority of studies (97%) were conducted in adults (≥18 y). In 55 (81%) studies, participants had existing inflammatory conditions; these were most commonly type 2 diabetes mellitus, hypercholesterolemia, cancer, inflammatory bowel diseases, liver diseases, and overweight or obesity.
Of the 68 included studies, 5 (7%) used SCFAs (27, 28, 54, 55, 76) (Table 1), 29 (44%) used prebiotics (2, 29–39, 52, 58, 62–64, 74, 75, 77–86) (Table 2), 26 used synbiotics (39%) (40–49, 56, 57, 59–61, 65, 67–73, 89, 90, 93) (Table 3), and 8 (12%) compared the effect of prebiotic supplementation with that of synbiotic supplementation (50, 51, 53, 66, 87, 88, 91, 92) (Table 4). Intervention duration ranged from 1 h to 24 wk (median: 6 wk). In 42 intervention studies (65%), participants were instructed to maintain their usual diet habits; however, only 32 studies (50%) reported measuring background diet intake before and after the intervention. Of these studies, 23 reported no significant difference in dietary intake between the study groups at baseline and at the end of the study, and the other 9 did not report the findings of the background diet analysis.
Summary of included studies examining the effect of SCFAs on systemic inflammation1
|Study (country) (ref)||Design/evidence level||Quality2||SCFA, daily dose||Control, daily dose||Participants, n||Duration||Effect of intervention on inflammation|
|Freeland and Wolever 2010 (Canada) (76)||RCT (x-over)/II||∅||Acetate, 60 mmol enema or 20 mmol i.v.||Normal saline, 300 mL enema or 100 mL i.v.||Hyperinsulinemia, 6||1 h||↓ TNF-α|
|Hamer et al. 2009 (Netherlands) (27)||RCT (x-over)/II||+||Butyrate, 60 mL enema (100 mM)||NaCl solution, 60 mL (140 mM) enema||Healthy, 16||2 wk||↔ CRP|
|Hamer et al. 2010 (Netherlands) (28)||RCT/II||∅||Butyrate, 60 mL enema (100 mM)||NaCl solution, 60 mL (140 mM) enema||UC, 35||20 d||↔ CRP|
|Canfora et al. 2017 (Netherlands) (54)||RCT (x-over)/II||+||SCFA mixtures high in acetate, butyrate, and propionate, 200 mL enema (200 mM)||NaCl, 200 mL enema (40 mmol NaCl)||Overweight or obese, 12||1 d/arm; 5 d w/o||↓ Fasting IL-1β (acetate vs. propionate)|
|van der Beek et al. 2016 (Netherlands) (55)||RCT (x-over)/II||+||Acetate, via distal or proximal feeding catheter (100 or 180 mM in 120 mL 0.9% NaCl)||0.9% NaCl, 120 mL via distal or proximal feeding catheter||Overweight or obese, 6||3 d/arm; 7 d w/o||↔ TNF-α, IL-6, IL-8|
CRP, C-reactive protein; i.v. intravenous; RCT, randomized controlled trial; ref, reference; SCFA, short-chain fatty acid; UC, ulcerative colitis; w/o, washout; x-over, crossover study design; ∅, neutral study quality; +, positive study quality; ↓, decrease, ↑, increase, ↔, no change.2
Methodologic study quality was determined with the use of the American Dietetic Association critical appraisal checklist.View Large
Summary of included studies examining the effect of prebiotics on systemic inflammation1
|Study (country) (ref)||Design/evidence level2||Quality3||Participants, n||Prebiotic, daily dose, g||Control, daily dose, g||Duration||Effect of intervention on inflammation|
|Studies on oligosaccharide prebiotics|
|Dehghan et al. 2014a (Iran) (63)||RCT/II||∅||T2DM, 49||INU, 10||MDX, 10||8 wk||↓ CRP|
|Dehghan 2014b (Iran) (64)||RCT/II||∅||T2DM, 52||OF-enriched INU, 10||MDX, 10||8 wk||↓ TNF-α and IL-6|
|Dehghan et al. 2016 (Iran) (62)||RCT/II||+||T2DM, 49||OF-enriched INU, 10||MDX, 10||8 wk||↓ IL-12 and IFN-y|
|Lecerf et al. 2012 (France) (32)||RCT/II||∅||Healthy, 40||INU-XOS, 6.64||MDX, 6.64||4 wk||↓ IL-1β|
|Morel 2015 (France) (33)||RCT/II||+||Overweight, 88||α-GOS, 12||Dried glucose syrup, 12||2 wk||↓ CRP|
|van den Berg et al. 2013 (Netherlands) (36)||RCT/II||+||Preterm infants, 113||GOS/FOS, 2.25||MDX, 2.2||27 d||↓ TNF-α, IFN-γ, and IL-1β|
|Vulevic et al. 2008 (United Kingdom) (37)||RCT (x-over)/II||∅||Healthy, 82||β-GOS, 5.5||MDX, 5.5||10 wk/arm;||↓ TNF-α, IL-6, and IL-1β|
|4 wk w/o|
|Vulevic et al. 2013 (United Kingdom) (38)||RCT (x-over)/II||∅||Overweight, 90||β-GOS, 5.5||MDX, 5.5||12 wk/arm;||↓ CRP|
|4 wk w/o|
|Benjamin et al. 2011 (United Kingdom) (29)||RCT/II||+||CD, 41||FOS, 15||MDX, 15||4 wk||↔ Any biomarkers|
|Vaisman et al. 2010 (Israel) (74)||RCT/II||∅||Acute diarrhea, 42 (aged 9 mo to 2 y)||Oligosaccharides and pectin, 6||MDX, 6||12 d||↔ Any biomarkers|
|Clarke et al. 2016 (Canada) (77)||RCT (x-over)/II||+||Healthy, 60||β2-1 fructan (50:50 mixture of INU and short-chain oligosaccharides), 15||MDX, 15||28 d/arm;||↑ TNF-α and IL-6|
|2 wk w/o|
|Vulevic et al. 2015 (United Kingdom) (39)||RCT (x-over)/II||+||Healthy, 80||β-GOS, 5.5||MDX, 5.5||10 wk/arm;||↑CRP|
|4 wk w/o|
|Williams 2016 (United Kingdom) (52)||RCT (x-over)/II||+||Asthma, 10; healthy, 8||B-GOS, 5.5||MDX, 5.5||3 wk/arm; 2 wk w/o||↓ TNF-α and CRP|
|Studies examining polysaccharide prebiotics|
|Smith et al. 2008 (United States) (83)||RCT/II||+||Hypercholesterol, 90||β-glucan, 6||6 wk||↓ CRP|
|Xie et al. 2015 (China) (75)||RCT/II||∅||Kidney disease, 83||SF (NS), 20||Starch, 20||6 wk||↓ CRP, IL-6, and IL-8|
|Brouns et al. 2012 (Netherlands) (30)||RCT (x-over)/II||∅||Hypercholesterol, 108||Pectin, 6||Cellulose, 6||3 wk/arm;||↔ Any biomarkers|
|≥1 wk w/o|
|Dall’Alba et al. 2013 (Brazil) (78)||RCT/II||+||T2DM, 44||PHGG, 10||No supplement||6 wk||↔ Any biomarkers|
|King et al. 2008 (United States) (79)||RCT/II||+||Overweight, 87||Psyllium, 14||No supplement||12 wk||↔ Any biomarkers|
|Nieman et al. 2008 (United States) (80)||RCT/II||+||Healthy, 36||β-glucan, 5.6||Cornstarch, 5.6||2 wk||↔ Any biomarkers|
|Queenan et al. 2007 (United States) (82)||RCT/II||∅||Hypercholesterol, 75||Oat β-glucan, 6||Dextrose, 6||6 wk||↔ Any biomarkers|
|Salas-Salvadó et al. 2008 (Spain) (34)||RCT/II||+||Overweight, 113||Plantago ovata seed husks, 9; glucomannan, 3||Microcrystalline cellulose, 3||16 wk||↔ Any biomarkers|
|Theuwissen et al. 2009 (Netherlands) (35)||RCT (x-over)/II||∅||Hypercholesterol, 42||Oat β-glucan, 4.8||Control fiber (NS), 4.8||4 wk/arm;||↔ Any biomarkers|
|2 wk w/o|
|Wood et al. 2006 (United States) (86)||RCT/II||∅||Overweight, 29||Glucomannan, 3||MDX, 3||12 wk||↔ Any biomarkers|
|Studies examining resistant starch|
|Aliasgharzadeh et al. 2015 (Iran) (58)||RCT/II||+||T2DM, 55||Resistant dextrin, 10||MDX, 10||8 wk||↓ TNF-α and IL-6|
|Penn-Marshall et al. 2010 (United States) (81)||RCT (x-over)/II||∅||Prediabetes, 34||RS bread, 12.39||Control bread, no added RS||6 wk/arm;||↔ Any biomarkers|
|2 wk w/o|
|Stewart et al. 2010 (United States) (84)||RCT (x-over)/II||∅||Healthy, 36||Pullulan, RS, SF dextrin, soluble corn fiber, 12||MDX, 12||2 wk/arm;||↔ Any biomarkers|
|3 wk w/o|
|Johansson-Persson et al. 2014 (Finland) (31)||RCT (x-over)/II||+||Hypercholesterol, 25||HF diet, SF, 10.7||LF diet, SF, 2.5||5 wk/arm;||↓ CRP|
|3 wk w/o|
|Villaseñor et al. 2011 (United States) (85)||Cross-sectional/IV||+||Breast cancer, 40||N/A||N/A||N/A||No association|
|Ma et al. 2008 (United States) (2)||Cross-sectional/IV||∅||Postmenopausal, 1958||N/A||N/A||N/A||Inverse association|
B-GOS, Bimuno–galato-oligosaccharide; CD, Crohn disease; CRP, C-reactive protein; FOS, fructo-oligosaccharide; GOS, galacto-oligosaccahride; HF, high fiber; IFN-γ, interferon-γ INU; inulin; LF, low fiber; MDX, maltodextrin; N/A, not applicable; NS, not specified; OF, oligofructose; PHGG, partially hydrolyzed guar gum; RCT, randomized controlled trial; ref, reference; RS, resistant starch; SF, soluble fiber; T2DM, type 2 diabetes mellitus; w/o, washout; x-over, crossover study design; XOS, xylo-oligosaccharide; ∅, neutral study quality; +, positive study quality. ↓, decrease, ↑, increase, ↔, no change.2
Evidence levels are defined by the National Health and Medical Research Council.3
Methodologic study quality was determined with the use of the American Dietetic Association critical appraisal checklist.View Large
Summary of included studies examining the effect of synbiotics on systemic inflammation1
|Study (country) (ref)||Study design/evidence level2||Quality3||Population, n||Intervention, daily dose||Control, daily dose||Duration||Effect of intervention on inflammation|
|Abbas et al. 2014 (Pakistan) (56)||RCT/II||+||IBS, 72||Saccharomyces bourardiicapsule (750 mg) + psyllium, 5 g||Placebo capsule + psyllium, 5 g||12 wk||↓ IL-8 and TNF-α|
|Akram et al. 2015 (Iran) (57)||RCT/II||∅||T2DM, 44||1 synbiotic capsule (dose NS)||Placebo capsule||8 wk||↓ CRP, IL-6, and TNF-α|
|Amati et al. 2010 (Italy) (40)||Before-and-after/IV||∅||Elderly (aged >66 y), 10||2 × 107 CFU LGG + OF (dose NS)||—||4 wk||↑IL-6, IL-8, and IL-1β|
|Anderson et al. 2004 (United Kingdom) (41)||RCT/II||+||Elective surgery, 137||12 × 109 CFU Lactobacillus acidophilus, Lactobacillus bulgaricus, Bifidobacterium lactis, Streptococcus thermophilus + 32 g OF||Placebo capsule + 32 g sucrose||1–2 wk||↔ Any biomarkers|
|Asemi et al. 2014 (Iran) (59)||RCT (x-over)/II||+||T2DM, 124||27 × 107 CFU Lactobacillus sporogenes + 1.08 g INU||Placebo supplement||6 wk/arm||↓ CRP|
|3 wk w/o|
|Asemi et al. 2013 (Iran) (60)||RCT/II||+||T2DM, 54||2 × 109 CFU L. acidophilus, 7 × 109 CFU Lactobacillus casei, 1.5 × 109CFU Lactobacillus rhamnosus, 2 × 108 CFU L. bulgaricus, 2 × 1010 CFU Bifidobacterium breve, 7 × 109CFU Bifidobacterium longum, 1.5 × 109 CFU S. thermophilus + 100 mg FOS||Placebo supplement||8 wk||↓ CRP|
|Asgharian et al. 2016 (Iran) (61)||RCT/II||∅||NAFLD, 74||1 × 500 mg capsule [L casei, L. acidophilus, L. rhamnosus, L. bulgaricus, B. breve, B. longum, S. thermophilus + FOS (dose NS)]||Placebo capsule (120 mg starch)||8 wk||↔ Any biomarkers|
|Eslamparast et al. 2014 (Iran) (65)||RCT/II||+||NAFLD, 52||4 × 108 CFU of 7 bacterial strains (L. casei, L. rhamnosus, S. thermophilus, B. breve, L. acidophilus, B. longum, L. bulgaricus) + FOS (dose NS)||MDX capsule||28 d||↓ CRP and TNF-α|
|Mofidi et al. 2017 (Iran) (68)||RCT/II||+||50||4 × 108 CFU L. casei, L. rhamnosus, S. thermophilus, B. breve, L. acidophilus, B. longum, L. bulgaricus + 125 mg FOS||MDX capsule||28 wk||↓ CRP|
|Federico et al. 2009 (Italy) (42)||RCT/II||∅||UC, 18||10 × 109 CFU Lactobacillus paracasei + 1 g XOS and 6 g INU||Placebo supplement (starch)||8 wk||↓ IL-6 and IL-8|
|Fernandes et al. 2016 (Brazil) (89)||RCT/II||∅||Obesity, 6||109 CFU of each: L. paracasei, L. rhamnosus, L. acidophilus, B. lactis + 6 g FOS||MDX, 6 g||15 d||↔ Any biomarkers|
|Giamarellos-Bourbouli et al. 2009 (Greece) (43)||RCT/II||∅||Multiple organ injury, 72||12-g sachet: 1011 CFU each of: Pediococcus pentosaceus, Leuconostoc mesenteroides, L. paracasei, Lactobacillus plantarum + INU, β-glucan, pectin, and RS||Placebo sachet, 12 g||15 d||↓ CRP|
|Kelishadi et al. 2014 (Iran) (67)||RCT/II||+||Overweight/obese, 55||2 × 108 CFU of: L. casei, B. breve, L. rhamnosus, B. longum, S. thermophilus, L. acidophilus, L. bulgaricus + FOS (dose NS)||MDX capsule||8 wk||↔ Any biomarkers|
|Macfarlane et al. 2013 (United Kingdom) (44)||RCT (x-over)/II||+||Healthy, 43||4 × 1011 CFU B. longum + 12 g prebiotic (INU + OF)||Starch capsule + 12 g MDX||4 wk/arm||↓ TNF-α, IL-6, and IL-8|
|4 wk w/o|
|Malaguarnera et al. 2012 (Italy) (45)||RCT/II||+||NASH, 66||B. longum + 2.5 g FOS||Placebo supplement||24 wk||↓ CRP and TNF-α|
|Neto et al. 2013 (Brazil) (90)||RCT/II||∅||Healthy, 17||108–109 CFU of each: L. paracasei, L. rhamnosus, L. acidophilus, B. lactis + 6 g FOS||Placebo supplement, 6 g MDX||12 wk||↔ Any biomarkers|
|Nova et al. 2011 (Spain) (46)||RCT/II||+||Healthy, 36||2.4 × 109 CFU of 5 bacterial strains (L. acidophilus, Bifidobacterium animalis, Lactobacillus delbrueckii, S. thermophilus, L. paracasei) + 1.4 g FOS||3 placebo capsules (sucrose, talcum powder, and stearic acid magnesium salt)||6 wk||↔ Any biomarkers|
|Rajkumar et al. 2015 (India) (69)||RCT/II||∅||Healthy, 30||2 × 109 CFU Lactobacillus salivarius +10 g FOS||Gelatin capsule||6 wk||↓ CRP, TNF-α, and IL-1β|
|Riordan et al. 2007 (Australia) (93)||RCT/II||∅||Cirrhosis, 30||1010 CFU of each: P. pentosaceus, L. plantarum, L. mesenteroides, L. paracasei, + 2.5 g of each: β-glucan, INU, pectin, RS||Placebo sachet, crystalline cellulose||1 wk||↑ IL-6 and TNF-α|
|Roller et al. 2007 (Ireland) (47)||RCT/II||+||Colon cancer, 34; polypectomized, 40||1010 CFU L. rhamnosus, 1010 CFU of B. lactis + 10 g prebiotic (INU and OF)||MDX capsules, MDX sachet, 10 g||12 wk||↔ Any biomarkers|
|Sugawara et al. 2006 (Japan) (70)||RCT/II||∅||Biliary cancer, 81||3 × 108 CFU L. casei, 3 × 108CFU B. breve + 15 g GOS||—||4 wk||↓ IL-6|
|Taghizadeh and Asemi 2014 (Iran) (71)||RCT/II||+||Pregnant, 52||Synbiotic food containing: 18 × 107 CFU L. sporogenes + 0.72 g INU||Same food, no synbiotics||9 wk||↔ Any biomarkers|
|Tajadadi-Ebrahimi et al. 2014 (Iran) (72)||RCT/II||+||T2DM, 54||Bread containing: 12 × 109 CFU L. sporogenes + 8.4 g INU||Same bread, no synbiotics||8 wk||↔ Any biomarkers|
|Usami et al. 2011 (Japan) (73)||RCT/II||∅||Hepatic cancer, 61||3 × 108 CFU B. breve, 3 × 108CFU L. caseiShirota + 15 g GOS||No synbiotics||2 wk||↓ CRP and IL-6|
|van De Pol et al. 2011 (Netherlands) (48)||RCT/II||+||Asthma and HDM allergy, 26||Food supplement with: 2 × 1010CFU B. breve + 14.4 g scGOS, 1.6 g lcFOS||Food supplement with MDX||4 wk||↓ IL-5|
|van der Aa et al. 2012 (Netherlands) (49)||RCT/II||+||Atopic dermatitis (infants aged <7 mo), 90||1.3 × 109 CFU B. breve + 0.8 g (90% scGOS, 10% lcFOS)/100 mL formula||No synbiotics||12 wk||↔ Any biomarkers|
CFU, colony-forming unit; CRP, C-reactive protein; FOS, fructo-oligosaccharides; GOS, galacto-oligosaccharide; HDM, house dust mite; IBS, irritable bowel syndrome; INU, inulin; lc, long chain; LGG, Lactobacillus rhamnosus GG; MDX, maltodextrin; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; NS, not specified; OF, oligofructose; RCT, randomized controlled trial; ref, reference; RS, resistant starch; sc, short chain; T2DM, type 2 diabetes mellitus; UC, ulcerative colitis; XOS, xylo-oligosaccharide; x-over, crossover study design; ∅, neutral study quality; +, positive study quality; ↓, decrease; ↑, increase, ↔, no change.2
Evidence levels are defined by the National Health and Medical Research Council.3
Methodological study quality was determined with the use of the American Dietetic Association critical appraisal checklist.View Large
Summary of included studies comparing the effect of prebiotics and synbiotics on systemic inflammation1
|Study (country) (ref)||Design/evidence level||Quality2||Population, n||Synbiotic, daily dose||Prebiotic, daily dose, g||Duration||Effect on inflammation|
|Childs et al. 2014 (United Kingdom) (50)||RCT (x-over)/II||+||Healthy, 120||109 CFU Bifidobacterium lactis + 8 g XOS||XOS, 8||3 wk/arm||↔ Any biomarkers|
|4 wk w/o|
|Fujimori et al. 2009 (Japan) (66)||RCT/II||∅||UC, 22||2 × 109 CFU Bifidobacterium longum + 8 g psyllium||Psyllium, 8||4 wk||↓ CRP (synbiotic group only)|
|González-Hernández 2012 (Mexico) (87)||RCT/II||+||HIV, 15||Lactobacillus rhamnosus and B. lactis at 109CFU/mL + 10 g FOS||FOS, 10||16 wk||↓ IL-6|
|Horvat et al. 2010 (Slovenia) (51)||RCT/II||+||Colorectal cancer, 68||2 × 1010 CFU each of Pediococcus pentosaceus, Leuconostoc mesenteroides, Lactobacillus paracasei, Lactobacillus plantarum + 20 g prebiotic (5 g each: β-glucan, RS, INU, pectin)||β-glucan. RS, INU, pectin, 5 each||3 d||↑ IL-6 (symbiotic group only)|
|Schunter et al. 2012 (United States) (88)||RCT/II||+||HIV, 27||1010 CFU each of P. pentosaceus, L. mesenteroides, L. paracasei, L. plantarum + 10 g prebiotic (2.5 g each: β-glucan, INU, pectin, and RS)||β-glucan, INU, pectin, and RS, 2.5 each||4 wk||↔ Any biomarkers|
|West et al. 2012 (Australia) (91)||RCT/II||+||Healthy, 25||13.8 × 108 CFU each of L. paracasei, Lactobacillus acidophilus, and L. rhamnosus GG, 18 × 108 CFU Bifidobacterium animalis + 270 mg Raftiline and 30 mg Raftilose||Acacia powder, 0.348||3 wk||Synbiotics limited IL-6 ↑ by 50% relative to prebiotics|
|Worthley et al. 2009 (Australia) (92)||RCT (x-over)/II||∅||Colorectal cancer, 36||5 × 109 CFU B. lactis + 12.5 g RS||RS, 12.5||4 wk/arm||↔ Any biomarkers|
|Krebs 2016 (Slovenia) (53)||RCT/II||+||Colorectal cancer, 73||2 × 1011 CFU each of P. pentosaceus, L. mesenteroides, L. paracasei, L. plantarum + 20 g prebiotic (5 g each of β-glucan, INU, pectin, RS).||β-glucan, INU, pectin, RS, 5 each||3 d||↔ Any biomarkers|
CFU, colony-forming unit; CRP, C-reactive protein; FOS, fructo-oligosaccharides; INU, inulin; RCT, randomized controlled trial; ref, reference; RS, resistant starch; UC, ulcerative colitis; w/o, washout; XOS, xylo-oligosaccharides; x-over, crossover study design; ∅, neutral study quality; +, positive study quality; ↓, decrease, ↑, increase, ↔, no change.2
Methodologic study quality was determined with the use of the American Dietetic Association critical appraisal checklist.3
Manufactured by Chr. Hansen A/S.View Large
Approximately 50% of the studies had positive methodologic quality. These studies were methodologically strengthened by their use of random allocation to the intervention and control group or treatment sequence (crossover trials), double-blinding, and comparability of study groups. Factors that limited the methodologic quality of the 29 studies rated as neutral included insufficient detail provided regarding the intervention protocol, the measurement of potential intervening factors (e.g., background diet intake), and the comparability of study groups (i.e., limited or lack of subject characteristics reported). A biomarker of inflammation was a primary outcome in 45 (66%) of the included studies.
Effect of SCFAs on systemic inflammation
The characteristics of the 5 included SCFA studies are presented in Table 1 (27, 28, 54, 55, 76). Canfora et al. (54) examined the effect of 3 SCFA mixtures (high acetate, high propionate, and high butyrate), delivered via enema, on systemic inflammation during fasting and postprandial conditions. Although fasting IL-1β was significantly decreased by the high acetate mixture compared with the high propionate mixture, there was no significant difference compared with placebo or in IL-1β postprandially between treatments (54). Furthermore, changes in other proinflammatory cytokines (TNF-α, IL-6, and IL-8) did not significantly differ between treatments in either the fasting or postprandial conditions (54). The study by van der Beek et al. (55) reported lower fasting plasma TNF-α concentrations after distal administration of acetate compared with placebo, however this did not reach significance. Two studies by Hamer et al. (27, 28) reported no significant change in plasma CRP concentration following the delivery of the SCFA butyrate via enema. However, Freeland and Wolever (76) reported a significant decrease in plasma TNF-α concentration following administration of the SCFA acetate both via enema and intravenously, with no significant difference between administration routes (76).
Dietary intake was similar between study groups at baseline and follow-up in all of the studies. Overall, a significant decrease in inflammation was observed in 40% of the included SCFA studies. Meta-analysis was not possible due to small study numbers and heterogeneity in study design.
Effect of prebiotics on systemic inflammation
Characteristics of the 29 included prebiotic studies are presented in Table 2, of which 14 (48%) showed a decrease in inflammation. Of the 13 studies investigating the effect of an oligosaccharide prebiotic (29, 32, 33, 36–39, 52, 62–64, 74, 77), 9 (69%) reported a significant decrease in ≥1 biomarker of systemic inflammation (primarily TNF-α, IL-6, CRP, or interferon-γ) compared with control. An increase in inflammation (CRP, TNF-α, and IL-6) following oligosaccharide supplementation was reported in 2 studies, both of which were crossover studies conducted in healthy populations (39, 77). No change in inflammation was reported in the remaining 2 studies that used oligosaccharide prebiotics, which were conducted in populations with a gastrointestinal tract condition (Crohn disease and acute diarrhea) (29, 74).
Two (20%) of the 10 polysaccharide prebiotic studies (30, 34, 35, 75, 78–80, 82, 83, 86) reported a significant decrease in inflammation compared with control or baseline (49, 65), with the remaining 8 studies reporting no anti-inflammatory effect. Of the 2 studies investigating the effect of resistant starch supplementation (58, 81), 1 reported a significant decrease in inflammation (TNF-α and IL-6) (58). A significant decrease in CRP was reported following a high- (10.7 g) compared with low-soluble fiber (2.5 g) diet (31), with an inverse association between soluble fiber intake and systemic inflammation (IL-6: P = 0.004 and TNF-α-R2: P = 0.02) observed in the cross-sectional study by Ma et al. (2).
Meta-analyses were performed to examine the effect of prebiotic supplementation on CRP (n = 7), IL-6 (n = 6), and TNF-α (n = 4). Results of the meta-analysis indicate that prebiotics significantly decrease CRP concentration compared with placebo or control (SMD: −0.60; 95% CI: −0.98, −0.23; I 2= 64%; P = 0.002) (Figure 3). Subgroup analyses by prebiotic fiber type provided evidence for a decrease in CRP concentrations with oligosaccharides (n = 3; SMD: −0.49, 95% CI −0.83, −0.15, I2 = 69%, P = 0.04), but no effect was observed with polysaccharide supplementation (SMD: −0.38; 95% CI: −1.05, 0.28; I2 = 75%; P = 0.26) (Table 5). There was no effect of prebiotics on IL-6 (SMD: −0.35; 95% CI: −0.84, 0.13; I2 = 75%; P = 0.15) (Figure 4) or TNF-α (SMD: −0.49; 95% CI: −1.20, 0.22; I2 = 84%; P = 0.18) (Figure 5). Investigation into the effects of prebiotics on other inflammatory biomarkers by meta-analysis was not possible due to small study numbers.FIGURE 3
Forest plot of randomized controlled trials investigating the effect of prebiotic supplementation on circulating C-reactive protein concentration, subgrouped by disease. Pooled effect estimates (diamonds) for C-reactive protein are shown. Values are standardized mean differences with 95% CIs determined with the use of generic IV random-effects models. Heterogeneity was quantified by I2 at a significance of P < 0.10. IV, inverse variance; Std., standardized; T2DM, type 2 diabetes mellitus.
Additional subgroup meta-analysis of prebiotic studies1
|Outcome||N||Experiment/control, n/n||SMD (95% CI)||I2, % (P)||P|
|Oligosaccharide||3||61/66||−0.49 (−0.83, −0.15)||69 (0.04)||0.04|
|Polysaccharide||3||81/79||−0.38 (−1.05, 0.28)||75 (0.02)||0.26|
|Oligosaccharide||3||60/60||−0.13 (−0.61, 0.34)||33 (0.23)||0.58|
|Polysaccharide||4||100/96||−0.41 (−1.23, 0.40)||78 (0.0001)||0.32|
|Oligosaccharide||2||57/55||−0.30 (−0.92, 0.31)||63 (0.10)||0.34|
|Polysaccharide||2||53/59||−0.25 (−2.16, 1.66)||95 (<0.0001)||0.80|
Values are SMDs with 95% CIs determined with the use of generic inverse-variance random-effects models. Heterogeneity was quantified by I2 at a significance of P < 0.10. CRP, C-reactive protein; SMD, standardized mean difference.View LargeFIGURE 4
Forest plot of randomized controlled trials investigating the effect of prebiotic supplementation on circulating IL-6 concentration, subgrouped by disease status. Pooled effect estimates (diamonds) for IL-6 are shown. Values are standardized mean differences with 95% CIs determined with the use of generic IV random-effects models. Heterogeneity was quantified by I2 at a significance of P < 0.10. IV, inverse variance; Std., standardized; T2DM, type 2 diabetes mellitus.
Forest plot of randomized controlled trials investigating the effect of prebiotic supplementation on circulating TNF-α concentration, subgrouped by disease status. Pooled effect estimates (diamonds) for TNF-α are shown. Values are standardized mean differences with 95% CIs determined with the use of generic IV random-effects models. Heterogeneity was quantified by I2 at a significance of P < 0.10. IV, inverse variance; Std., standardized; T2DM, type 2 diabetes mellitus.
Effect of synbiotics on systemic inflammation
The characteristics of the 26 included synbiotic studies are presented in Table 3. The mean daily prebiotic dose was 8.7 g (range: 0.1–32 g), and the median daily probiotic dose was 2.2 × 109 CFU (range: 2 × 107 − 4 × 1011 CFU). The majority of studies used a probiotic supplement comprised primarily of Lactobacillus or Bifidobacteriumcombined with a single oligosaccharide prebiotic. Fourteen studies (53%) reported a significant decrease in ≥1 biomarker of inflammation (primarily CRP, TNF-α, or IL-6) following synbiotic supplementation (42–45, 48, 56, 57, 59, 60, 65, 68–70, 73). However, 10 studies reported no significant change in inflammation. A significant increase in inflammation was reported by Amati et al. (40) (IL-6, IL-8, and IL-1β) and Riordan et al. (93) (IL-6 and TNF-α).
Meta-analyses were conducted to examine the effect of synbiotics on CRP (n = 11), IL-6 (n = 5), and TNF-α (n = 6). The results indicate that CRP and TNF-α concentrations are lower following synbiotic supplementation than with placebo [SMD: −0.40; 95% CI: −0.73, −0.06; I2 = 78%; P = 0.02 (Figure 6) and SMD: −0.90; 95% CI: −1.50, −0.30; I2 = 78%; P = 0.003 (Figure 7), respectively]. Although there was significant heterogeneity between studies in both analyses (I2 = 76%; P = 0.002 and I2= 78%; P = 0.0003, respectively), a subgroup analysis of the studies reporting on CRP suggested that disease is a possible source of the heterogeneity. There was no significant difference in IL-6 between synbiotic and placebo supplementation [SMD: −21; 95% CI: −0.71, 0.33; I2 = 71; P = 0.45 (Figure 8)]. Investigation into the effects of synbiotics on other inflammatory biomarkers by meta-analysis was not possible due to small study numbers.FIGURE 6
Forest plot of randomized controlled trials investigating the effect of synbiotic supplementation on circulating C-reactive protein, subgrouped by disease. Pooled effect estimates (diamonds) for C-reactive protein are shown. Values are standard mean differences with 95% CIs determined with the use of generic IV random-effects models. Heterogeneity was quantified by I2 at a significance of P < 0.10. IV, inverse variance; Std., standardized; T2DM, type 2 diabetes mellitus.
Forest plot of randomized controlled trials investigating the effect of synbiotic supplementation on circulating TNF-α, subgrouped by disease status. Pooled effect estimates (diamonds) for TNF-α are shown. Values are standard mean differences with 95% CIs determined with the use of generic IV random-effects models. Heterogeneity was quantified by I2 at a significance of P < 0.10. IV, inverse variance; Std., standardized.
Forest plot of randomized controlled trials investigating the effect of synbiotic supplementation on circulating IL-6 concentration, subgrouped by disease status. Pooled effect estimates (diamonds) for IL-6 are shown. Values are standardized mean differences with 95% CIs determined with the use of generic IV random-effects models. Heterogeneity was quantified by I2 at a significance of P < 0.10. IV, inverse variance; Std., standardized.
Effects of prebiotics versus synbiotics on systemic inflammation
The characteristics of the 8 RCTs of prebiotic and synbiotic supplementation are presented in Table 4 (50, 51, 53, 66, 87, 88, 91, 92). The mean daily prebiotic dose was 10 g (range: 0.57–20 g), and the median daily probiotic dose was 5 × 109 CFU (range: 1 × 109–8 × 1010 CFU). Fujimori et al. (66) and González-Hernández et al. (87) reported a significant decrease in systemic inflammation (CRP and IL-6, respectively) in the synbiotic group only. West et al. (91) observed an increase in proinflammatory IL-16 following both prebiotic and synbiotic supplementation; however, synbiotic supplementation was found to be less inflammatory, leading to an increase that was only 50% of the increase following prebiotic supplementation. Four studies, 2 of which were conducted in healthy participants, reported no significant change in inflammation in either the prebiotic or synbiotic treatment arm (50, 53, 88, 92).
This review examined evidence for the effect of SCFAs, prebiotics, and synbiotics on systemic inflammation in healthy populations, diabetes, overweight and obesity, kidney disease, cancer, liver disease, and bowel diseases. Approximately half of the included studies reported a significant decrease in ≥1 systemic inflammatory biomarker. Meta-analyses show that prebiotic and synbiotic supplementation are associated with decreased systemic inflammation, including CRP, IL-6, and TNF-α, although the association was stronger with certain supplement types (particularly oligosaccharides).
Various methods can be used to deliver SCFAs. In animal studies, large oral doses of SCFA can reduce systemic inflammation (94). In humans, SCFAs can be administered by enemas and tablets designed to release SCFAs into the colon (95). In this review, 5 studies directly delivered SCFAs. One study, which used an acetate enema, reported a significant decrease in systemic inflammation (76), with another study reporting a trend toward a decrease in systemic inflammation compared with control following the delivery of acetate to the distal colon (55). Acetate is the primary SCFA to enter circulation, and thus has the most potential to exert systemic anti-inflammatory effects (96). In contrast, butyrate is primarily absorbed by colonocytes (96), which may explain why there were no significant changes in inflammation in studies that used butyrate (27, 28, 54).
Circulating SCFA concentrations in humans can also be increased via gut fermentation of prebiotic soluble fibers (9). In a previous review of prebiotic supplementation by Kellow et al. (15), 3 of 4 studies reported significant reductions in CRP in overweight and obese adults and women with type 2 diabetes mellitus compared with controls. However, a pooled analysis of these studies (n = 181 participants) indicated a nonsignificant decrease in CRP (15). This review also reported conflicting results for the effect of prebiotics on TNF-α and ILs (15). Similarly, 50% of prebiotic studies included in this current review reported a significant decrease in ≥1 inflammatory marker, with pooled analysis showing no significant effect on TNF-α or IL-6. In contrast to the previous review, our pooled analysis of 7 prebiotic supplementation studies (n = 172) indicated a significant decrease in CRP. Heterogeneity between studies included in the previous meta-analysis and our review in regards to supplement formulation and dosage, intervention duration, and study population may explain these different findings.
Subgroup meta-analyses by prebiotic fiber type demonstrated strong evidence that oligosaccharide supplementation reduces CRP concentrations, whereas there was no effect with polysaccharide supplementation. Studies suggest that short-chain substrates with lower degrees of polymerization, such as oligosaccharides (e.g., FOS, GOS, and inulin), are more rapidly fermented and produce a greater SCFA yield than molecules with higher degrees of polymerization (10). Although SCFA production was not measured in the majority of the included studies, the type of prebiotic fiber used and resulting SCFA concentrations may explain the divergent effects on inflammation. In addition to substrate type, other factors can influence SCFA production, including gut transit time (97), the composition of the colonic microbiota, and the site of substrate fermentation (96). Most SCFA production occurs in the proximal colon, where substrate availability and bacterial density is the highest and decreases distally (9). Luminal pH is modulated by SCFA concentrations, which subsequently influences the types of SCFA-producing bacteria present (98). The lower pH in the proximal colon favors butyrate-producing bacteria, and as pH increases distally, acetate- and propionate-producing bacteria become dominant (98). Future studies investigating prebiotic supplementation that include the measurement of circulating SCFA concentrations are warranted.
Supplementation with synbiotics is hypothesized to have a greater effect on systemic inflammation than prebiotics alone due to their superior ability to increase SCFA-producing bacteria numbers, as well as providing substrates for fermentation (99). Yet, there is conflicting evidence in the literature regarding their systemic anti-inflammatory effects. Heterogeneity was also observed in this review, where 50% of synbiotic studies reported a significant decrease in systemic inflammation. Furthermore, only 43% of included studies reported greater anti-inflammatory effects with synbitoic than with prebiotic supplementation. Nonetheless, meta-analyses indicate that synbiotic supplementation significantly reduces CRP and TNF-α concentrations. Supplement formulation and dosage is likely to have an effect on whether anti-inflammatory effects are observed, which may explain the conflicting evidence.
Although the anti-inflammatory effects observed following prebiotic and synbiotic supplementation may be attributable to the production of SCFAs, other mechanisms may be involved. Prebiotic and synbiotic supplementation have both been shown to stimulate the growth of beneficial bacteria in the colon (10). Certain bacteria (e.g., Lactobacillus species) indirectly regulate inflammation through maintaining and repairing epithelial barriers, which subsequently reduces the impact of proinflammatory stimuli, such as LPS (100). Commensal bacteria can also increase the synthesis of antimicrobial peptides involved in inflammation resolution pathways (100, 101). Furthermore, specific bacterial species and their metabolic products have a direct influence on proinflammatory signaling pathways (e.g., nuclear factor-κ B) by acting as ligands for innate immune system receptors (e.g., Toll-like receptors) (102). Another potential anti-inflammatory mechanism of gut bacteria is via the modulation of the differentiation and activity of immune cells, such as dendritic cells, promoting the production of cytokines, such as IL-10 (102).
This systematic review has a number of limitations. Primarily, heterogeneity between studies in regards to supplement formulation, dosage, study duration, and systemic inflammatory outcome variables limited the number of studies included in the meta-analyses. This also inhibited the ability to perform meta-regression to investigate the reasons for heterogeneity. Furthermore, background dietary nutrient intake can influence changes in systemic inflammation biomarkers, making it difficult to assess the effect of supplementation alone, and should be considered in the design of future studies. Although most studies advised participants to maintain their usual dietary habits, only 36 (∼57%) studies measured background dietary intake. It is possible that changes or group differences in other nutrients (e.g., soluble fiber, proinflammatory nutrients, such as saturated fat, or antiinflammatory nutrients, such as vitamin A, C, E, lycopene, lutein, and omega-3) may have also influenced the changes in inflammation reported in these studies. The measurement of background dietary intake in future studies is warranted. Despite these limitations, to our knowledge, this is the first systematic review to comprehensively examine the available evidence on the effect of SCFA, prebiotics, and synbiotics on systemic inflammation in humans. Furthermore, the majority of studies included in this review (95%) were classified as level II evidence as per the National Health and Medical Research Council evidence hierarchy, thus strengthening the findings of the review.
In summary, this review has demonstrated that there is promising evidence supporting the anti-inflammatory benefits of synbiotics and prebiotics, in particular oligosaccharides, in humans. However, due to the heterogeneity between studies, it is difficult to determine the most beneficial supplement dosage and formulation as well as intervention duration. Further research is needed to confirm the association between SCFA, prebiotics and synbiotics, and systemic inflammation and to elucidate the responsible mechanisms.
We thank the statistical analysis support of Heather Powell.
The authors’ responsibilities were as follows—RFM, BSB, KJB, and LGW: designed the research; RFM and BSB: conducted the research; RFM: analyzed the data and wrote the manuscript; RFM, BSB, MEJ, and LGW: had primary responsibility for the final content; and all authors: read and approved the final manuscript. None of the authors reported a conflict of interest related to the study.