Effects of probiotic supplementation on the gut microbiota composition of adults: a systematic review of randomized clinical trials

Abstract Researchers have associated the therapeutic potential of probiotics with its ability to modulate gut microbiota, which is considered an “invisible organ” of the human body. The present study investigates the effects of probiotic supplementation on the gut microbiota composition of adults. The authors conducted a systematic review of the literature published in six different databases. The search followed PRISMA guidelines and aimed to identify randomized clinical trials on probiotic supplementation. All relevant publications indexed up to May 28, 2021, were retrieved. Then, the authors defined the inclusion and exclusion criteria. Two independent reviewers performed study screening, data extraction, and quality assessment. A total of 2,404 publications were retrieved, and eight studies met the eligibility criteria. The included randomized clinical trials were published between 2015 to 2020. The worldwide studies included adults aged from 18 to 79 years, most of whom were women (66.5%). Only one of the included studies observed significant effects on fecal microbiota composition in the relative abundance of Bacteroidetes and Firmicutes phyla in comparison with the placebo treatment. Overall, this systematic review could not draw consistent conclusions on the effects of probiotic supplementation on the gut microbiota composition of adults.


INTRODUCTION
The human gastrointestinal tract is home to more than 100 trillion different microorganisms, including bacteria, fungi, protozoa, and viruses.These microorganisms act in syntrophy, forming a complex and dynamic microbial population, i.e., the gut microbiota (Gill et al. 2006, Honda & Littman 2012).The gastrointestinal tract harbors more than 1,000 different bacterial species (Qin et al. 2010), mainly belonging to Bacillus spp., Bacteroides spp., Bifidobacterium spp., Clostridium spp., Enterococcus spp., Lactobacillus spp., Prevotella spp., and Ruminicoccus spp.(Arumugam et al. 2011).The four dominant phyla, Actinobacteria, Bacteroidetes, Firmicutes, and Proteobacteria, account for more than 90% of the human gut microbiota (Hugon et al. 2015).The abundance and diversity of prokaryotic cells vary throughout the gastrointestinal tract, from the mouth to the anus (Arumugam et al. 2011).However, it was reported that Bacteroidetes and Firmicutes are dominant bacteria in the gut of healthy adults (Sheehan et al. 2015, Andoh 2016).
Researchers consider gut microbiota as an "invisible organ" of the human body (Chen et al. 2020, Li et al. 2020), which plays essential functions in maintaining the host healthy (Mohr et al. 2020).These include: controlling epithelial cell proliferation and differentiation (Sekirov et al. 2010); strengthening gut integrity or shaping the intestinal epithelium (Natividad & Verdu 2013); acting on nutrient uptake, vitamin synthesis, and energy harvesting (Ursell et al. 2012); protecting against pathogens (Bäumler & Sperandio 2016); affecting behavioral and neurological functions (Wahlström et al. 2016, Zheng et al. 2019); and changing drug absorption, toxicity, metabolism, and bioavailability (Li et al. 2020).The intestinal microbiota also assists in metabolic functions and improves the host immune system (Chen et al. 2020).However, factors such as ethnicity, age, birth method, body mass index, antibiotic use, lifestyle, exercise frequency, and feeding habits shape the gut microbiota and influence important differences among people (Rinninella et al. 2019, Mohr et al. 2020).
Studies have increasingly highlighted the importance of the gut microbiota in human health and disease development in the last few years.Many researchers address new interventions to modulate this microbiota (Quigley 2019).In this context, the modulation of the gut microbial community is a promising treatment target for many diseases related to gut dysbiosis.Efficient methods to restore microbial balance include the consumption of prebiotics, probiotics, and fecal microbiota transplantation (Hasan & Yang 2019).
Probiotics are living microorganisms that, when administered in adequate amounts, confer health benefits to the host (Hill et al. 2014).Thus, regular consumption of probiotics can modulate immune responses and metabolic processes (Quigley 2019) as well as antioxidant and antiinflammatory effects, with these microorganisms acting as gut microbial modulators (Neto et al. 2018).Notwithstanding, the mechanisms or metabolic pathways through which probiotic supplementation benefits human health are not yet well established (Neto et al. 2018, Quigley 2019).
This systemic review of randomized clinical trials discusses the effect of multi-strain probiotic supplementation (Lactobacillus spp.and Bifidobacterium spp.) on the gut microbiota composition of adults.Aiming to evaluate the therapeutic efficiency of these probiotics, the authors analyzed changes on the relative abundance of the main intestinal bacterial phyla (Actinobacteria, Bacteroidetes, Euryarchaeota, Firmicutes, Fusobacteria, Proteobacteria, Synergistetes, and Verrucomicrobia).The review protocol was registered in the PROSPERO database (CRD42020177567).

Literature search
This systematic review of the current literature followed the PRISMA statement guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) (Moher et al. 2009).Two authors (APM and EP) independently searched for articles published until May 28, 2021, in the databases Embase, Lilacs, PubMed, ScienceDirect, SCOPUS, and SpringerLink.These searches aimed at identifying studies that reported the effects of probiotic supplementation on the gut microbiota composition of adults.The keywords used as search strategy were: "probiotics" AND "Lactobacillus" AND "Bifidobacterium" AND ("gut microbiota" OR "fecal microbiota") AND human AND "clinical trial".There were no restrictions on language or year of publication.The titles and abstracts of the selected articles were assessed.Articles whose titles appeared to meet the inclusion criteria or lacked sufficient information for exclusion were read in full.The reference lists of selected records were also searched to include potentially relevant additional studies.

Inclusion and exclusion criteria
Inclusion criteria were: (1) randomized control trials; (2) use of probiotics as intervention method and placebo as control; (3) strains of Lactobacillus spp.and Bifidobacterium spp. as probiotic supplementation; (4) report of changes in the gut microbiota composition of adults; and (5) assessment of differences in the relative abundance of prokaryotic cells at phylum level between groups or before-after supplementation.Exclusion criteria were: case reports, reviews, systematic reviews, metaanalyses, duplicate publications, in vitro experimental trials, children and adolescent studies, animal trials, prebiotic and symbiotic studies, research assessing probiotic supplementation with a single strain of Lactobacillus spp. or Bifidobacterium spp., and studies investigating the survival of probiotic strains only.Studies with multiple comparisons between probiotic and placebo groups were considered for evaluation.

Data extraction
Two authors independently collected the data (APM and EM), and discrepancies were solved by reevaluation of the original record by both authors.The authors extracted the data directly from the articles using a standardized form, including the following information from each clinical trial: author, year of publication, original country, study design, study population, sample size (numbers of case and control subjects), gender and mean age of patients, probiotic supplementation (strains, intake form, dosage, and intervention time), and data on microbiota (including intervention effect, the phyla and genera of bacteria detected, and the methodology used for microbiological assessment).The studies included the measurement of the effects of probiotic supplementation on the overall gut microbiota composition of adults, recording major alterations in the relative abundance of the phyla Bacteroidetes and Firmicutes.

Quality assessment
The same two authors who performed data extraction assessed the methodological quality of the clinical trials.For that, the authors used the Jadad score, with a five-point scale evaluating three traits: randomization quality, blinding quality, and reasons for withdrawal/dropout (Jadad et al. 1996).The included studies were of medium or high quality, with low quality studies being excluded from the selection process.

Risk of bias assessment
Two authors (APM and EP) used the Cochrane Collaboration's tool for assessing risk of bias in each randomized trial included (Higgins et al. 2011).The investigation included the following items: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and any other sources of bias.Risk of bias in each trial was classified as high, moderate (unclear), or low.Studies with high risk of bias were excluded from the selection process.

Data analyses
The included studies reported the relative abundance of prokaryotic cells belonging to the phyla Actinobacteria, Bacteroidetes, Euryarchaeota, Firmicutes, Fusobacteria, P r o t e o b a c t e r i a , S y n e r g i s t e t e s , a n d Verrucomicrobia in the gut microbiota of adults.The selected clinical trials were identified and examined as a systematic review rather than a meta-analysis due to the high heterogeneity of study designs and methods.
T h e re l a t i v e a b u n d a n c e o f t h e previously cited phyla was extracted using WebPlotDigitizer version 4.5 (https://automeris.io/WebPlotDigitizer).This variable was used to identify differences between pre-and after-probiotic or placebo consumption by discriminant analysis (PAST3 software -Hammer et al. 2001).

Study selection
The search in the electronic databases included articles published up to May 28, 2021.From the predesigned search strategy, the authors identified 2,404 relevant publications.The first clean strategy eliminated 126 publications since they were duplicate records.After that, from the analysis of titles and abstracts, 2,147 publications were excluded because they were reviews, case reports, conferences and abstracts, design studies, animal trials, children and adolescent studies, in vitro experimental trials, prebiotic and symbiotic studies, or studies on probiotic supplementation with Bacillus spp.and/or Saccharomyces spp.strains.Then, 132 publications were selected for full-text analysis, of which 124 were excluded for the following reasons: being nonrandomized clinical trials; using a combination of probiotics and prebiotics; not identifying total fecal microbiota; investigating only the survival of the probiotic strain.Finally, eight studies achieved all inclusion criteria and were selected for this systematic review (Table I). Figure 1 summarizes the screening strategy.

Selected studies
The selected articles were published from 2015 to 2020.Table I identifies

Quality assessment
Table II shows the results of the methodological quality of the clinical trials.According to the Jadad score (Jadad et al. 1996), six studies (Botelho et al. 2020, Ford et al. 2020, Liu et al. 2020, Song et al. 2020, Francavilla et al. 2019, Martoni et

Risk of bias assessment
The analysis of the bias risk of the selected clinical trials followed the methodology described in Cochrane Collaboration's tool (Higgins et al. 2011) (Figure 2).Four studies (50.0%) showed low risk of bias in all evaluated criteria.Two studies (25.0%) had one unclear risk of bias each.One study (12.5%) presented two risks of bias, and another presented one risk of bias and one unclear risk.Random sequence generation was the main risk of bias, followed by selective reporting (reporting bias).All studies reported lost data or patients who did not attend followup appointments.Overall, all studies presented low risk of bias.Publication bias assessment  through funnel plot analysis was not performed since the number of included studies was less than ten (Cooper et al. 2019).Principal coordinate analysis using the UniFrac metric showed similar dissimilarity among the treatments.Furthermore, the gut microbiome profile did not differ at three distinct taxonomic levels (phyla, family, and genus).Accordingly, the relative abundance of bacteria belonging to the phyla Bacteroidetes and Firmicutes was similar in the treatments.Liu et al. (2020) showed that regular probiotic supplementation with strains B. longum NQ1501, L. acidophilus YIT2004, and E. faecalis YIT0072 did not alter bacterial richness and α-diversity (Chao1 and Shannon indices) in the gut microbiome.The authors compared placebo and probiotic supplementation after six months of intervention.However, probiotic supplementation reduced the relative abundance of Firmicutes and increased the abundance of Bacteroidetes.Probiotics enhanced the proportion of Enterococcaceae (phylum Firmicutes) and reduced that of Peptostreptococcaceae and Ruminococcaceae (phylum Firmicutes).These three families were remarkable in the use of probiotics versus placebo.Serum albumin, inflammatory markers, endothelial activation markers, SF-36 (shortform patient-reported health survey), and GSRS (Gastrointestinal Symptom Rating Scale) scores did not differ between probiotic and placebo treatments.Song et al. (2020) also observed that probiotic supplementation (B.breve CBTBR3 and L. plantarum CBT LP3) did not influence Shannon and Simpson indexes (α-diversity).Nonmetric multidimensional scaling (NMDS) based on the Bray-Curtis distance matrix showed no obvious separation between the probiotic and placebo groups.Furthermore, probiotic supplementation did not change the relative abundance of the top five most abundant phyla (Bacteroidetes, Firmicutes, Proteobacteria, Actinobacteria, and Fusobacteria) and of the most abundant families (Alcaligenaceae, Bacteroidaceae,

B i f d o b a c t e r i a c e a e , C l o s t r i d i a c e a e , Enterobacteriaceae, Fusobacteriaceae, La c h n o s p i ra ce a e , Pa ra p re v o te l l a ce a e , Po r p hy ro m o n a d a ce a e
, Pre v o te l l a ce a e , Rikenellaceae, Rumminococcaceae, and Veillonellaceae) in each group.However, twelve weeks of probiotic intervention ameliorated obesity-related markers such as waist circumference, total fat area, visceral fat, and biochemical markers such as triglyceride, total cholesterol, high-density lipoprotein cholesterol, glucose, and insulin.Wu et al. (2020) showed that, after two weeks of probiotic supplementation with strains of B. longum, L. acidophilus, and E. faecalis, the number of operational taxonomic units (OTUs) of the probiotic group was significantly higher, increasing gut microbiota richness (Chao 1 index).The Shannon index of the gut microecological structure was similar among the placebo and probiotic groups (after 7 and 14 days of treatment).Noteworthy, the most abundant phyla did not differ from each other: Bacteroidetes, Firmicutes, Proteobacteria, and Fusobacteria.The phyla Acidobacteria,  Describe as randomized* 1 1 1 1 1 1 1 1 Randomization method described and appropriate** Double-blind method described and appropriate** Jadad score 5 5 5 5 3 5 5 3 * A study receives a score of 1 for "yes" and 0 for "no".** A study receives a score of 0 if no description is given, 1 if the method is described and appropriate, and -1 if the method is described but inappropriate.
Fusobacteria, Cyanobacteria, Deferribacteres, Euryarchaeota, and Synergistetes were present at low proportions.In turn, Bacteroidetes and Firmicutes accounted for more than 90% of all bacterial flora at phylum level.Notwithstanding, probiotic intervention can reduce the incidence of gut dysbiosis and antibiotic-induced diarrhea, and the prophylactic use of BIFICO can stabilize the gut microbiota.Francavilla et al. ( 2019) showed that probiotic supplementation (L.casei LMG 101/37 P-17504, L. plantarum CECT 4528, B. animalis subsp.lactis Bi1 LMG P-17502, B. breve Bbr8 LMG P-17501, and B. breve Bl10 LMG P-17500) did not change total prokaryotic community richness (rarefaction curves and Chao1 estimates) and α-diversity (Shannon index).Moreover, β-diversity did not show clear separation for the gut microbiome composition of placebo and probiotic groups (inferred by principal coordinate analysis using the UniFrac metric).The relative abundance of the phyla Bacteroidetes, Firmicutes, Proteobacteria, and Fusobacteria did not differ, neither did metabolically active bacteria among the treatments.However, the abundance of Actinobacteria OTUs was higher in the probiotic treatment than in the placebo treatment.In addition, IBS-SSS (Irritable Bowel Syndrome -Severity Scoring System) and GSRS scores decreased after probiotic supplementation.Martoni et al. (2019) showed that the probiotic group presented OTU richness and Shannon index (α-diversity) values similar to baseline values after four weeks of probiotic supplementation with L. acidophilus, B. animalis subsp.lactis, B. longum, and B. bifidum.In contrast, OTU richness was lower in the placebo group than in the probiotic supplementation group.The relative abundances of the phyla Bacteroidetes, Firmicutes, and Actinobacteria were similar over time in both groups.Patient assessment of constipation symptom score was similar between groups (regarding β-diversity, samples did not cluster or separate according to time point in either group).
The probiotic supplementation (B.breve CNCM I-4035 and L. rhamnosus CNCM I-4036) performed by Plaza-Díaz et al. (2015) did not change the α-diversity (Shannon indices) of the gut microbiome in comparison with the placebo group.Bacteroidetes and Firmicutes were the most abundant phyla in fecal compositions of the placebo and probiotic groups.However, the relative abundance of these phyla did not differ between groups after 30 days of intervention.
Discriminant analysis was performed with 307 samples from the eight selected studies (Figure 3).The samples were divided into four treatments: pre-and post probiotic/placebo consumption.Joint analysis of all data from the eight selected studies confirms the individual conclusion: the pattern of prokaryotic community did not differ significantly at phylum level.However, all selected studies identify health benefits and encourage probiotic consumption.Thus, the health mechanism activated by gut microbiota remains unclear.

DISCUSSION
This study is an updated comprehensive systematic review on the effects of probiotic supplementation with Lactobacillus spp.and Bifidobacterium spp.strains on the gut microbiota composition of adults.The final review included eight randomized clinical trials.Only the study performed by Liu et al. (2020) observed changes in gut microbiota composition after intervention with probiotic strains.These authors reported that probiotic supplementation reduced the relative abundance of bacteria belonging to the phylum Firmicutes after six months of treatment with a mixture of E. faecalis, B. longum, and L. acidophilus.The results suggest that these bacterial strains act as probiotics and help to restore the intestinal flora balance of uremic patients undergoing hemodialysis, consequently modulating their gut microbiota.Another factor to be considered is the supplementation time used by Liu et al. (2020).These authors used probiotic supplementation for six months.It could explain the data obtained by these authors.In addition, Liu et al. (2020) evaluated patients with chronic kidney disease subjected to a restrictive diet.This diet could also be a determining factor in the data observed.
The clinical trials investigated by this systematic review did not provide consistent evidence to conclude that probiotic supplementation affects microbial community composition in adults (Figure 3).The high heterogeneity in the selected studies can be due to: (I) very restrictive criteria for choosing the studies; (II) trials with small number of participants, with the smallest sample having nine participants; (III) short period of intervention, ranging from 2 to 4 weeks; and (IV) trials with different multi-strain probiotic interventions (see Table I).
Gut microbiota alterations may require a long time, typically eight weeks (Dethlefsen et al. 2008, McFarland 2008), and less than half of the selected studies used an appropriate period of analysis.Studies that failed to detect microbiota alterations could present positive results by increasing the experimental time.In addition, all the clinical trials used different multi-strain probiotic interventions, and only the study of Ford et al. (2020) examined the effects of a diet associated with probiotic supplementation.Efficacy of probiotic supplementation is straindependent (Derrien & Vlieg 2015) and varies according to feeding habits and the baseline microbiota structure (Albenberg & Wu 2014, David et al. 2014, Maier et al. 2017).Thus, to observe alterations in the gut microbiome, clinical trials must analyze each probiotic strain alone before evaluating multi-strain formulations.An increase in the number of participants is also essential to visualize these differences.All participants selected for randomized clinical trials were adults with widely distinct characteristics.Plaza-Díaz et al. (2015) (Koren et al. 2012) interfere directly in the gut microbiome composition and in the potential health effects of probiotic supplementation.In addition to these factors, health status influences the composition of the intestinal microbiota and may be a determining factor for the high heterogeneity observed among the studies.Kristensen et al. (2016) found results similar to those of the present study.These authors evaluated seven randomized clinical trials which investigate alterations in the composition of healthy human fecal microbiota by highthroughput molecular approaches.Only one study reported significant changes in the overall structure of the fecal bacterial community in terms of β-diversity (compositional dissimilarity) after probiotic intervention in comparison with the placebo group.According to the authors, this demonstrates the lack of consistent evidence to conclude whether there is a positive effect of probiotics on fecal microbiota composition in healthy adults.
McFarland (2014) published a systematic review on the effects of probiotic intervention on human gut microbiota.Their study selected three groups: model A (restoration) assayed patients with healthy and undisturbed microbiota and then assayed postdisruptive event and probiotic therapy; model B (alteration) assayed patients with pre-existing disrupted microbiota and then post probiotic therapy; and model C (no dysbiosis) assayed volunteers with undisturbed microbiota and absence of disruptive event.All groups received probiotic therapy.From the inclusion criteria, the authors selected a total of 63 clinical trials.In these, 83% of the probiotic treatments using model A restored the microbiota, 56% of treatments using model B improved the microbiota, and only 21% of treatments using model C had some effect on the microbiota.The authors mention the need for more studies to conclude which probiotic strains have a beneficial impact on gut microbiota.The search for evidence that links probiotic supplementation efficacy to its ability to modulate the gut microbiome is an important tool to prove the healthy action of probiotics.

Limitations
Systematic literature reviews can present some limitations.In the case of this review, the number of published randomized clinical trials that evaluate the effects of probiotics on the fecal microbiota composition at the phylum level is still limited.Selected studies used supplementation with multi-strain probiotics formulated with both Lactobacillus spp.and Bifidobacterium spp..In this sense, studies that used probiotic supplementation with Bacillus spp., Enterococcus spp., Lactococcus spp., or Saccharomyces spp.were excluded.Finally, six databases and a wide variety of terms have been used for the literature review; however, some articles may have not been identified.To minimize the impact of this selection strategy, the references used in selected articles were searched and added to the literature review.Therefore, the need of studies that group and critically revise the effects of probiotic supplementation on the gut microbiota composition reinforces the importance of this review, regardless of quantitative synthesis.

CONCLUSIONS
Overall, this systematic review was not able to identify consistent effects caused by probiotic supplementation on the gut microbiome of adults.However, it was possible to identify that is necessary to improve the time of probiotic supplementation for identify the modulation of gut microbiome, improving the efficacy of this type of therapeutics.Thus, the present study can be a helpful reference for future research because highlight the importance of carefully planning experiments to obtain specific sample characteristics (number, health status, age, diet) that improve the probability of identifying the effects of probiotics in gut microbiome.
al. 2019) presented high quality, and two (Wu et al. 2020, Plaza-Díaz et al. 2015) presented medium quality.All studies were designed as randomized clinical trials.Wu et al. (2020) used an openlabel design, while other studies used a doubleblind design.Ford et al. (2020) used a cross-over design, and other studies used a parallel design.

Figure
Figure 1.Flow chart of the screening process for inclusion of studies in the systematic review.

Figure
Figure 2. Risk of bias in the included clinical trials.

Figure 3 .
Figure 3. Discriminant Analysis performed at phylum level with 307 samples extracted from the eight selected studies.

Table I .
Main characteristics of the included clinical trials.Participant characteristics regarding the number of individuals in the probiotic group/total number of individuals included in the study.Participant characteristics (% women and age range of participants in years) are based on the total number of individuals included in the study.CFU, colony-forming units.The study by Plaza-Díaz et al. (2015) had two intervention arms not relevant for the present study, where the number of participants are approximately two-thirds of the total number of participants included in the study. *

Table II .
Methodological quality assessment in the included clinical trials.