The use of probiotics and prebiotics in decolonizing pathogenic bacteria from the gut; a systematic review and meta-analysis of clinical outcomes

ABSTRACT

Repeated exposure to antibiotics and changes in the diet and environment shift the gut microbial diversity and composition, making the host susceptible to pathogenic infection. The emergence and ongoing spread of AMR pathogens is a challenging public health issue. Recent evidence showed that probiotics and prebiotics may play a role in decolonizing drug-resistant pathogens by enhancing the colonization resistance in the gut. This review aims to analyze available evidence from human-controlled trials to determine the effect size of probiotic interventions in decolonizing AMR pathogenic bacteria from the gut. We further studied the effects of prebiotics in human and animal studies. PubMed, Embase, Web of Science, Scopus, and CINAHL were used to collect articles. The random-effects model meta-analysis was used to pool the data. GRADE Pro and Cochrane collaboration tools were used to assess the bias and quality of evidence. Out of 1395 citations, 29 RCTs were eligible, involving 2871 subjects who underwent either probiotics or placebo treatment to decolonize AMR pathogens. The persistence of pathogenic bacteria after treatment was 22%(probiotics) and 30.8%(placebo). The pooled odds ratio was 0.59(95% CI:0.43–0.81), favoring probiotics with moderate certainty (p = 0.0001) and low heterogeneity (I² = 49.2%, p = 0.0001). The funnel plot showed no asymmetry in the study distribution (Kendall’sTau = −1.06, p = 0.445). In subgroup, C. difficile showed the highest decolonization (82.4%) in probiotics group. Lactobacillus-based probiotics and Saccharomyces boulardii decolonize 71% and 77% of pathogens effectively. The types of probiotics (p < 0.018) and pathogens (p < 0.02) significantly moderate the outcome of decolonization, whereas the dosages and regions of the studies were insignificant (p < 0.05). Prebiotics reduced the pathogens from 30% to 80% of initial challenges. Moderate certainty of evidence suggests that probiotics and prebiotics may decolonize pathogens through modulation of gut diversity. However, more clinical outcomes are required on particular strains to confirm the decolonization of the pathogens. Protocol registration: PROSPERO (ID = CRD42021276045).

GRAPHICAL ABSTRACT

KEYWORDS:

• Probiotics
• prebiotics
• gut microbiota
• alpha diversity
• decolonization

Introduction

The human intestinal microbiota is a complex microbial ecosystem that colonizes the gastrointestinal system and comprises approximately 500 to 1000 bacterial species,¹ besides viruses, archaea, protozoa, and other eukaryotes. It is a dynamic ecosystem affected by diet, lifestyle and stress.² This microbial community modulates human health, including host metabolism, physiology, nutrition, and immune response.² Besides, the host microbiota is essential for fostering resistance to pathogens.^3,4 Repeated exposure to antibiotics and changes in diet and environment can cause an imbalance in gut microbial diversity and composition, termed gut dysbiosis.

Gut dysbiosis makes the host susceptible to pathogenic colonization, as the compromised microbiota cannot offer resistance against colonization by exogenous antibiotic-resistant or pathogenic bacteria.³ The gastrointestinal tract can, therefore, function as a reservoir to harbor a range of bacterial pathogens as they colonize and multiply to high densities. This gives rise to two important concerns. Firstly, individuals heavily colonized with these pathogens act as reservoirs, contributing to the ongoing persistence of pathogens and facilitating person-to-person transmission. Secondly, colonization can lead to potentially fatal infections, particularly in individuals with impaired immune systems.⁵

Modulating the gut microbiota can be an innovative and effective strategy to defend against colonization of antimicrobial-resistant or infectious enteric pathogens.⁶ However, conventional therapies do not provide reliable decolonization outcomes.⁷ Research has been conducted using prebiotics and probiotics to decolonize AMR pathogens as a public health demand by increasing bacterial diversity in the gut.

Probiotics are defined as live nonpathogenic microorganisms administered to improve microbial diversity in the gut. Prebiotics are compounds that do not contain microorganisms but provide health benefits by stimulating the growth of beneficial microorganisms in the gut.⁸ Probiotics have the potential to eliminate or reduce pathogens in the gut through various mechanisms, including competitive exclusion, enhancing intestinal colonization, improving intestinal barrier function, and regulating immune properties.⁹ Prebiotics help to balance the gut flora by promoting the growth of beneficial microbes. Digestion of prebiotics produces SCFAs and decreases the gut lumen’s pH, which helps avoid pathogen colonization.¹⁰ In addition, the anti-adherence properties of prebiotics can help reduce pathogenic microorganisms in the gut.¹¹

While alternative approaches like antibiotics, fecal microbiota transplantation (FMT), phage therapy, and CRISPR-Cas9 systems have also been used to decolonize pathogens, they all have inherent limitations.¹² Although evidence suggests that probiotics and prebiotics can positively impact pathogenic decolonization,^9,11-14 it would be necessary to understand the magnitude of these effects and their generalizability since some studies failed to demonstrate successful decolonization or reduction of the carriage of drug-resistant pathogens.^15,16 To address this variability in clinical efficacy, we have conducted a systematic review and meta-analysis of randomized clinical trials (RCTs) on probiotics and prebiotics to decolonize or eradicate pathogenic bacteria from the human gut microbiota.

Methodology

Protocol development and registration

The systematic review protocol was developed following the recommendations of the PRISMA-P 2015 guidelines¹⁷ (Supplement Table S1). The review protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) and is available at https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42021276045.

Systematic search strategy

A comprehensive search was conducted using databases including PubMed, SCOPUS, Web of Science (SCI), CINAHL (EBSCO), Embase (via Ovid), and MEDLINE (via Ovid), and Google Scholar to extract data for this review and meta-analysis. The details of search strategies used in different databases can be found in Supplement Table S2a-f. There was no time limit for searching the data, which was completed to that available in February 2023. Additional articles from the reference lists of identified publications were also added to the COVIDENCE platform for screening when appropriate.

Eligibility criteria

The PICOS (Population, Intervention, Comparator, Outcomes, and Study Design) approach was used to guide the determination of the eligibility criteria.¹⁸ This systematic review analyses the effect of two interventions, probiotics and prebiotics, in decolonizing pathogenic bacteria from the gut. We considered probiotics intervention studies with or without concomitant conventional treatment (as an experimental group) and compared them with the placebo to evaluate the outcomes. Similarly, in the case of the efficacy of prebiotics to decolonize or reduce the pathogen’s load, we considered the studies that used prebiotics (treatment group) compared with the placebo, antimicrobial agent, or control group (no treatment). The primary outcome evaluated for the probiotic-treated cohorts was the decolonization of pathogens from the gut within one month (30 days) or the end of the study, whichever is earliest, confirmed by PCR or plate culture method. However, complete eradication and CFU reduction were also considered outcomes of interest for the prebiotics.

For the effect of the probiotics cohort, we included randomized control trials (RCTs) and other clinical studies with a distinct control arm compared with a placebo administered for any length of time among people to eradicate pathogens from the gut. Tables 1 and 2 show the details of the inclusion and exclusion criteria.

Table 1. Inclusion and exclusion criteria for the Probiotics study.

PICOS Criteria	Inclusion criteria	Exclusion criteria
Population	Adult population (> 18 years). Confirmed colonization of pathogenic(, including extended-spectrum ß-lactamase (ESBL)-producing Enterobacteriaceae, multidrug-resistant (MDR) Escherichia coli, MDR gram-negative Bacillus, Staphylococcus aureus, Clostridium difficile, Helicobacter pyroli, Vancomycin-resistant Enterococci (VRE), Multi-drug resistant Pseudomonas aeruginosa and Carbapenemase-producing Klebsiella pneumonia.) bacteria through PCR, plate culture methods, 16S rRNA sequencing, breath tests, etc.	Patients who had repeated or stubborn C. difficile infections (other studies solely focus on this issue). Studies on VRE and C. difficile that have already been reviewed.^Citation19,Citation20 Children (Age <18 years) (since the gut microbiota of children is still evolving and does not resemble the adults.^Citation21 People infected with pathogens other than bacteria.
Intervention	Probiotics were concomitant with conventional therapy (Table in decolonizing pathogens. Distinct types of probiotics with no restriction of dosages based on bacteria or yeast Different types of probiotics and prebiotics delivery materials were considered, such as capsules, tablets, sachet powder, yoghurt, milk drinks, etc. Instead of focusing on a specific type, we included all the delivery methods since the study is still rare.	We excluded all the studies other than probiotics to decolonize pathogens. Focusing on any drugs other than conventional therapy concomitant with probiotics.
Comparison	Probiotics with or without conventional medicine (experimental group) compared with only traditional therapy or a placebo (control group).	Studies with no control or placebo group
Outcome	The primary outcome was decolonizing pathogens from the gut within one month (30 days) or the end of the study, whichever is earliest, confirmed by PCR or plate culture method.	Not applicable
Study design	Randomized control trials (RCTs) and other clinical studies with a distinct control arm comparing combined therapy with (Placebo). Limited to original articles.	Observation studies, case reports, experimental studies, pilot studies, etc.

Table 2. Inclusion and exclusion criteria for the Prebiotics study.

PICOS Criteria	Inclusion criteria	Exclusion criteria
Population	Adult population (>18 years) Any other animals, including mice, chicken, etc., are also included since only probiotics studies on humans are scarce. Confirmed colonization of pathogenic bacteria (Salmonella enteritidis, S. typhimurium, S.enterica, Acinetobacter, Clostridium perfringens, Helicobacter, etc.) through PCR, plate culture methods, 16S rRNA sequencing, breath tests, etc.	Children (Age <18 years) as the gut microbiota of children do not resemble the adults. People or animals infected with pathogens other than bacteria.
Intervention	Prebiotics or a combination of prebiotics with no restriction of dose and types. Different types of prebiotic delivery materials were considered.	All the studies other than prebiotics to decolonize pathogens were excluded. Focusing on any drugs other than prebiotics.
Comparison	Prebiotic treatment was compared with the baseline pathogenic load in the gut.	Studies with no comparison.
Outcome	The primary outcome was decolonizing or reducing pathogens compared to the initial challenge confirmed by PCR or plate culture.	Not applicable
Study design	Due to the scarcity of human-based RCTs for the efficacy of prebiotic treated cohort, we considered clinical control trials and case studies on animals or humans where prebiotics or a combination of prebiotics are used to decolonize or reduce pathogenic burdens (CFU reduction). Limited to original articles.	Observation studies, case reports, experimental studies, pilot studies, etc.

Quality assessment and reduction of risk of bias to assess the eligibility of the articles

The web platform https://www.covidence.org was used to conduct this assessment. All citations from the databases were imported into the software. All authors reached a consensus to include articles from Google Scholar. NR and NB conducted preliminary abstract screening and inclusion of articles for this review. Then NR, NB, MT, and PD conducted a full-text review of the chosen studies to reduce bias. Any discrepancies were reviewed with the senior author MI.

We applied the Cochrane Collaboration tool to assess the risk of bias of the included RCTs (https://www.bmj.com/content/343/bmj.d5928##) via the RevMan 5.4 version. The risk of bias parameters encompassed randomization (selection bias), allocation concealment (selection bias), blinding (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), group similarity baseline (selecting bias), compliance (performance bias), co-intervention (performance bias) and others. Each potential bias was rated by assigning uncertain, high, or low risk.

Data extraction

We created a spreadsheet for data abstraction using Excel 2022 (Microsoft Office 365, Microsoft Corporation, Redmond, Washington, U.S.A.) The details are found in Supplement Tables S3 and S4. A final column was added for the reviewer’s judgment of whether the study was accepted or rejected.

Data synthesis and analysis

All RCTs and clinical trials with a distinct control arm for the efficacy of probiotic(s) underwent a meta-analysis for the primary outcome (Decolonization of pathogenic bacteria at the end of the intervention or 1-month time point). Meta-analysis was conducted using RevMan 5.4, Statistical Packages for Social Science (SPSS) 28, USA, and Jamovi²² (version 2.3, Sydney, Australia) following the generic inverse-variance approach of the random-effects model. The combined measures of the estimate were illustrated by creating a forest plot. Subgroup analyses were conducted for different pathogenic species, probiotic(s) groups, and regions to understand treatment efficacy and study heterogeneity. Any moderators affecting the outcome were evaluated by Jamovi.²² The meta-regression function of the Open meta-analyst was used to assess the effect of probiotics dose on the outcome.²³ Statistical heterogeneity between the studies included in the meta-analysis was described using I² statistic. A funnel plot was generated to determine the publication bias and asymmetry of the studies. Since meta-analysis should not rely on a single test to quantify publication bias, we analyzed bias using rank correlation and regression of funnel plot asymmetry and Eggar’s and Begg’s tests. Eggar’s and Begg’s regression tests detect publication bias more frequently than others.^24,25 Prebiotic studies did not have a similar control arm and decolonization objectives. Thus, a narrative synthesis of data from the articles involved in decolonizing or reducing pathogenic burdens by prebiotic interventions was performed. We also conducted a narrative synthesis of data from the study with metagenomic analysis.

Results

Study selection

The overall process of study selection is summarized in a PRISMA flow diagram (Figure 1). Forty articles were eligible for detailed data abstraction and discussion. Twenty-nine RCTs focused on probiotic treatments were selected for quantitative meta-analysis, whereas 15 articles were chosen for qualitative data analysis (10 prebiotics and 5 probiotics).

Figure 1. PRISMA flow diagram showing the systematic review process with the bibliometric assessment, including article attrition and study selection.

Characteristics of probiotic studies

A total of 29 RCTs with probiotic(s) treatment were included for quantitative meta-analysis. A detailed summary of these studies is listed in Supplement Table S5. A total of 6 RCTs on probiotics intervention with metagenomic analysis were chosen for qualitative analysis (Supplement Table S6). The probiotics and their dosages to decolonize pathogens varied among the studies. The maximum dose of probiotics was 2 × 10¹¹ CFU/day (E. coli nissle), whereas the minimum dose was 1 × 10⁷ CFU/day (non-toxigenic C. difficile). Most studies decolonized bacteria using 10¹⁰ CFU/day (Table 3).

Table 3. Types and dosages of probiotics.

Reference	Type of probiotics	Dosage (CFU/day)	Formulation ofprobiotic
Alberda 2018^Citation26	Lactobacillus casei sp. paracasei CNCM I-1518	2 x 10^10	yogurt drink,
Barker 2017^Citation27	L. acidophilus NCFM, ATCC 700,396; L. paracasei Lpc-37, ATCC SD5275; Bifidobacterium lactis Bi-07, ATCC SC5220;	1.7 x 10^10	Capsule
Dall 2019^Citation28	L. rhamnosus GG	6 x 10^9	Tablet
De Regt 2015^Citation29	B. bifidum, B. lactis, Enterococcus faecium, L. acidophilus, L. paracasei, L. plantarum, L. rhamnosus, and L. salivarius	1 x 10^9	Powder
Doron 2015^Citation30	L. rhamnosus GG	2 x 10^10	Capsule
Eggers 2018^Citation31	L. rhamnosus HN001	1 x 10^10	Capsule
Gerding 2015^Citation32	Nontoxigenic Clostridium difficile strain M3 (VP20621; NTCD-M3	1 x 10^7	Liquid drink
He 2022^Citation33	Lyophilized Saccharomyces boulardii	1 x 10^9	Powder
Klarin 2008^Citation34	L. plantarum 299v	8x 10^8	Capsule
Kwon 2015^Citation35	L. rhamnosus GG	1 x 10^10	Capsule
Ljungquist 2020^Citation13	Probiotic Vivomixx®, a mixture of 8 different living bacterial strains	9 x 10^11	Sachet
Manely 2007^Citation36	L. rhamnosus GG (LGG)	1 x 10^9	Yogurt
Mc Farland 1994^Citation37	S. boulardii	3 × 10^10	Capsule
Nouvenne 2015^Citation38	Not declared	2.4 × 10^10	Capsule
Rauseo 2022^Citation39	L. rhamnosus GG (LGG)	2 x 10^10	Capsule
Rubin 2022^Citation40	L. rhamnosus GG	3 × 10^10	Capsule
Salomao 2016^Citation41	L. bulgaricus and L. rhamnosus	4 x 10^10	Suspension
Sullivan 2004^Citation42	L. paracasei ssp. paracasei F19	1 x 10^10	Milk Powder
Lawrence 2005^Citation43	Lyophilized S. boulardii	1 x 10^9	Capsule
Surawicz 2000^Citation44	L. rhamnosus GG	1 x 10^9	Capsules
Szachta 2011^Citation45	L. rhamnosus GG	3 x 10^9	Sachet powder
Tang 2021^Citation46	E. faecium and Bacillus subtilis	1.3 x 10^9	Capsule
Tannok 2011^Citation16	Escherichia coli Nissle 1917	2 x 10^11	Capsule
Viazis 2022^Citation47	L. acidophilus LA-5, L. plantarum, Bifidobacterium lactis BB-12, and S. boulardii.	5.5 x 10^9	Capsule
Warrak 2014^Citation48	L. rhamnosus HN001	1 x 10^9	Capsule
Wieers 2021^Citation49	L. acidophilus NCFM, L. paracasei Lpc-37, B. lactis Bl-04, and B. lactis Bi-07 (Bactiol duo®)	1.4 x 10^10	Capsule
Wullt 2003^Citation50	L. plantarum 299v	5 x 10^10	Fruit drink
Yang 2021^Citation51	L. reuteri DSM17648	1 x 10^10	Sachet
OH 2015^Citation52	Streptococcus faecium and Bacillus subtilis	9 x 10^10	Capsule

Efficacy of the probiotics on the primary outcome (decolonization of pathogenic bacteria)

All 29 studies (2871 patients) included in the meta-analysis evaluated the efficacy of probiotics with or without traditional treatment over traditional treatment alone (or placebo) for decolonizing pathogenic bacteria at the end of the trial or at a 1-month time-point. The persistence of pathogenic bacteria after treatment was 361 (23.24%) of the 1353 patients in the probiotic(s)-treated group (decolonization rate 77%) and 498 (32.8%) of the 1518 patients in the placebo group (decolonization rate 67.2%) respectively. The pooled odds ratio (OR) for total events (persistence of bacteria upon treatment) was 0.59, indicating that pathogenic bacteria persistence was significantly lower in the probiotic(s) group than placebo (p = 0.0001, CI:0.43–0.81). Acceptable heterogeneity was observed in the 29 clinical trials (I² = 49%, Figure 2).

Figure 2. Forest plot for the intervention group (probiotics) vs. placebo for the decolonization success at the end of the study or at a 1-month time point. The I² values are interpreted as low, moderate, and high levels of heterogeneity, where I² = 0–49%, I²= 50–75%, and I² > 75%.^19,53 A 5% significance level (p<0.05) was used to determine the significant difference between the treatment efficacies. The events consist of the persistence of bacteria.

Subgroup analysis for the effect of probiotics with or without conventional medicine

The effects of probiotics concurrent with or without conventional medicine in decolonizing pathogens have been depicted in Figure 3. The probiotics with traditional medicine or antibiotics (20 assigned by hospital depending on diseases, total 21 studies) decolonized pathogens from 818 (78.65%) of 1040 subjects in the treatment group vs. 919 (72.19%) of 1273 subjects in the placebo group (OR = 0.51, p = 0.001). The observed statistical heterogeneity is low (I² = 52%, p = 0.002). Besides, the probiotics without antibiotics or traditional treatment (8 studies), decolonized pathogens from 174 (56%) of 313 intervene subjects vs. 101 (41%) of 245 placebo subjects (OR = 0.79, p = 0.44) with the statistical heterogeneity I² = 43%, p = 0.11. However, the observed subgroup difference estimate (chi²  = 1.46, p = 0.23) is not significant.

Figure 3. Forest plot of the sub-group analysis for the efficacy of probiotics concomitant with or without conventional medicine. The I² values are interpreted as low, moderate, and high levels of heterogeneity, where I² = 0–49%, I²= 50–75%, and I²> 75%, respectively.^19,53 A 5% significance level (p<0.05) was used to determine the significant difference between the treatment efficacies. The events consist of the persistence of bacteria.

Subgroup analysis for the types of probiotics

Figure 4 depicts the probiotics group-wise effects on the outcome of decolonizing pathogens from the intestine and statistical heterogeneity among probiotics studies. Lactobacillus-based probiotics (16 studies) decolonized pathogens from 168 (62%) of 442 patients vs. 212 (53%) of 451 patients in the placebo group (OR = 0.70, p = 0.04) with statistical heterogeneity I² = 39%, p = 0.06. In five studies using Bifidobacterium-based mixed probiotics, 413 (80%) of 514 subjects on probiotics vs. 574 (78%) of 733 subjects on placebo decolonized pathogens, and no significant differences were demonstrated between the two groups (OR:0.83, p = 0.23, I² = 48%). S. boulardii significantly decolonized pathogens (76.3%) compared to placebo (57.9%) (OR = 0.43, p = 0.0009) with minimal statistical heterogeneity (I² = 43%, p = 0.05).

Figure 4. Forest plot of the sub-group analysis for the efficacy of specific probiotics Lactobacillus, Mixed probiotics, S. boulardii, and other probiotics vs placebo. The I² values are interpreted as low, moderate, and high levels of heterogeneity, where I² = 0–49%, I²= 50–75%, and I²> 75%, respectively.^19,53 A 5% significance level (p<0.05) was used to determine the significant difference between the treatment efficacies. The events consist of the persistence of bacteria.

Subgroup analysis for the types of bacterial pathogen

Figure 5 shows the statistical heterogeneity and treatment efficacy of the studies of decolonized primary gut colonizers. C. difficile decolonization without recurrent infection was achieved in 206 (76.9%) of 268 patients who received combined treatment (Probiotic + conventional treatment) and in 94 (49.5%) of the 190 in placebo (conventional treatment alone) (OR = 0.29, p < 0.00001). Importantly, no heterogeneity was observed between the 8 studies on C. difficile (I² = 0, p = 0.98). Probiotics intervention showed no advantage in treatment efficacy over placebo for MDR-Enterobacteriaceae in 5 RCTs involving 228 subjects. The pooled odds ratio was 0.64 (95% CI: 0.19–2.16) with moderate heterogeneity (I2 = 61). The probiotics intervention shows a statistically significant benefit on decolonizing Vancomycin-resistant Enterococcus (VRE) (4-RCTs, n = 145, pooled OR 0.17, p = 0.08, 95% CI:0.02–1.25, I² = 81) and Helicobacter pylori (5-RCTs, n = 1167, pooled OR:0.68, p = 0.04, 95% CI:0.48–0.97, low or no heterogeneity I²=0).

Figure 5. Forest plots for sub-group analysis of the efficacy of the intervention (Probiotics) vs. Placebo on C. difficile, MDR-Enterobacteriaceae, VRE, and H. pylori. The I² values are interpreted as low, moderate, and high levels of heterogeneity, where I² = 0–49%, I²= 50–75%, and I²> 75%, respectively.^19,53 A 5% significance level (p<0.05) was used to determine the significant difference between the treatment efficacies. The events consist of the persistence of bacteria.

Subgroup analysis based on region

We further clustered the studies by continents where they were studied, according to North America (n = 9), Europe (n = 12), Asia (n = 3), and others* (n = 4), to observe differences in the treatment efficacy (Supplement Figure S1 and S2). No significant differences (p = 0.43) were found among regions in decolonizing pathogenic bacteria. Studies conducted in the USA effectively (p=0.008) eradicated pathogenic colonizers compared to placebo (pooled OR: 0.56, 95% CI:0.36–0.85, low heterogeneity, I² = 16%), and that conducted in Europe (OR: 0.71, = 0.02, 95% CI:0.53–0.94, moderate heterogeneity, I² = 53%). The Others* cluster showed excellent effectiveness in decolonizing bacteria from the gut, where pooled effect size, OR was 0.27, CI:0.14–0.53, and p = 0.0002. Though the study in other clusters was conducted in different countries, the observed heterogeneity, I² was 0%.

Moderation effects of the different covariates of probiotics intervention

Table 4 shows the moderation effects of probiotics types, pathogens, clusters, and the dose of probiotics (CFU/day) on the primary outcome. The types of probiotics (p = 0.018) and pathogenic gut colonizers (p=0.003) significantly predict the effect size of decolonization. However, other covariates, regions (p=0.842), and the dose of probiotics (p = 0.737) could not control the outcome of pathogen decolonization. The regression models for probiotics in decolonizing pathogens are shown in Figure 6. The QQ plot was used to examine the RCT distribution before running the regression model (Supplement Figure S3). The distribution of studies met the assumption for meta-regressions conducted in open-meta analyst¹⁷ and Jamovi.¹⁶

Figure 6. Results of the meta-regression of the covariates, dosages of probiotics, and types of probiotics and types of bacteria on the decolonization of pathogens.

Table 4. Moderation effect of the covariates of probiotic intervention, types of probiotics, pathogens, regions, and dosages of probiotics on the effect size of pathogenic decolonization.

Moderator (k* =29)	Estimates	Standard error (SE)	p-value
Types of probiotics	−0.0486	0.0265	0.018
Types of pathogens	0.0514	0.0174	0.003
Region of studies	0.00648	0.0326	0.842
Probiotics dose	0.0109	0.0325	0.737

Risk of bias assessment

Eighteen of 25 RCTs possess minimal or no elements of bias^{13,16,27,30-32,34,36,37,40-42,44,46,48-51} (Supplement Figure S4). We selected 3 prospective cohort studies, two^39,47 of which retain outstanding quality concerning predilection, whereas 1 study⁴² owns moderate features. A case-control study²⁶ also confirmed its strength concerning bias reduction. There are 11 studies^{13,30,34,36,37,39,41,46,47,49,51} that thoroughly addressed all 10 possible risks of bias. At the same time, between two to seven studies pose a high risk of bias random sequence generation,^29,45 allocation concealment,³⁵ blinding of participants,^{26,28,33,35,45} blinding of outcome assessment,²⁸ incomplete outcome data,³⁵ and selective reporting^31,43 respectively.

All elements of bias were less than 5% except blinding of participants and personnel (performance bias) and blinding of outcomes (detection bias). However, the performance and detection biases did not exceed 25%, as shown in the summary graph (Supplement Figure S5).

Publication bias

There was no indication of potential publication bias influencing observed aggregated results. A funnel plot was created by RevMan 5.4 to determine the asymmetry of the 29 RCTs (Figure 7). The funnel plot shows a symmetrical distribution. According to the Egger test, the publication bias and asymmetry between the studies were insignificant (Table 5). The Begg’s rank correlation (p=0.445) and regression tests (0.272) for funnel plot asymmetry were also not statistically significant at a 5% significance level. Cook’s distances were examined to determine any outliers or influential studies that could affect the outcomes (Supplement Figure S6). Though there are flagged studies among covariates, no studies were found to be extreme outliers or influencers of outcomes (Cook’s distance < 1).

Figure 7. Funnel plot of comparison: efficacy of pathogenic bacteria decolonization.

Table 5. Publication bias summary.

Egger’s test
Estimates of measure		−0.7512
Significance level		p = 0.2573
Begg’s test (Rank Correlation Test for Funnel Plot Asymmetry)
Kendall’s Tau estimates	−0.1138
Significance level	p = 0.4148
Regression Test for Funnel Plot Asymmetry
Z	−1.098
Significance level	p = 0.272

Quality of evidence through GRADEPro analysis

The quality of evidence of the studies was analyzed using the Cochrane review tool GRADEpro with the relative and absolute effects of probiotics intervention to decolonize pathogens. Table 6 shows that the certainty of overall probiotics (n = 29) effects was moderate (⨁⨁⨁◯), which reveals the effectiveness of probiotics in decolonizing pathogens. However, certainty was increased from moderate to high in the subgroup analysis for the probiotics and types of pathogens. The certainty of the evidence for prebiotics trials (n = 10) was low due to not having a control arm and randomization, as GRADEpro considers these criteria to judge the certainty of evidence. GRADEpro analysis of evidence for all subgroups is presented in Supplement Table S7.

Table 6. GRADEpro summary of findings for assessing the quality of evidence as assessed separately by two independent reviewers.

Certainty assessment							No. of patients		Effect		Certainty	Importance
No. of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Pathogens decolonization by probiotics	placebo	Relative (95% CI)	Absolute (95% CI)
Overall probiotic effect (follow-up: mean 1 month; assessed with placebo)
29	RCTs	not serious	serious^a	not serious	not serious	none	982/1343 (73.1%)	1011/1508 (67.0%)	OR 0.59 (0.42 to 0.81)	125 fewer per 1,000 (from 210 fewer to 48 fewer)	⨁⨁⨁◯ Moderate	Critical
Groupwise probiotic effect - Lactobacillus (follow-up: range 4 weeks to 6 weeks; assessed with placebo)
15	RCTs	not serious	serious^a	not serious	not serious	none	254/422 (60.2%)	239/451 (53.0%)	OR 0.72 (0.50 to 1.04)	82 fewer per 1,000 (from 169 fewer to 10 more)	⨁⨁⨁◯ Moderate	Important
Groupwise probiotic effect - Mix probiotics (follow-up: range 4 weeks to 6 weeks; assessed with placebo)
5	RCTs	not serious	serious^a	not serious	not serious	none	413/514 (80.4%)	574/759 (75.6%)	OR 0.83 (0.59 to 1.16)	36 fewer per 1,000 (from 110 fewer to 26 more)	⨁⨁⨁◯ Moderate	Important
Groupwise probiotic effect - Saccharomyces boulardii (follow-up: range 4 weeks to 6 weeks; assessed with placebo)
4	RCTs	not serious	not serious	not serious	not serious	none	122/160 (76.3%)	95/164 (57.9%)	OR 0.34 (0.23 to 0.49)	260 fewer per 1,000 (from 339 fewer to 176 fewer)	⨁⨁⨁⨁ High	Important
Subgroup analysis based on types of bacteria - Clostridium difficile (follow-up: mean 1 month; assessed with placebo)
8	RCTs	not serious	not serious	serious^b	not serious	none	206/268 (76.9%)	94/190 (49.5%)	OR 0.25 (0.17 to 0.36)	298 fewer per 1,000 (from 352 fewer to 234 fewer)	⨁⨁⨁◯ Moderate	Critical
Subgroup analysis based on types of bacteria - MDR Enterobacteriaceae (follow-up: mean 1 month; assessed with placebo)
5	RCTs	not serious	not serious	serious^b	not serious	none	34/116 (29.3%)	27/112 (24.1%)	OR 0.39 (0.19 to 0.80)	131 fewer per 1,000 (from 184 fewer to 38 fewer)	⨁⨁⨁◯ Moderate	Critical
Subgroup analysis based on types of bacteria - Vancomycin-resistant Enterococcus (VRE) (follow-up: mean 1 month; assessed with placebo)
4	RCTs	not serious	serious^a	not serious	not serious	none	43/70 (61.4%)	23/75 (30.7%)	OR 0.40 (0.12 to 1.34)	156 fewer per 1,000 (from 256 fewer to 65 more)	⨁⨁⨁◯ Moderate	Critical
Subgroup analysis based on types of bacteria - Helicobacter pylori (follow-up: mean 1 month; assessed with placebo)
4	RCTs	not serious	serious^a	not serious	not serious	none	521/583 (89.4%)	498/584 (85.3%)	OR 0.69 (0.49 to 0.98)	53 fewer per 1,000 (from 113 fewer to 3 fewer)	⨁⨁⨁◯ Moderate	Critical
Prebiotics in reducing or eradicating pathogens (follow-up: range 3 weeks to 4 weeks; assessed with baseline)
10	Observational studies	serious^c	serious^d	serious^e	not serious	none	Some groups were missing the control or placebo group for comparison of intervention. There was no defined event in the intervention and control group. Therefore, no meta-analysis was done to make a comparison.				⨁◯◯◯ Very low	Important

CI: confidence interval; OR: odds ratio.

Explanations

a. Some studies showed that probiotics have excellent results in decolonising multi-drug resistant pathogens, whereas, in some studies, a placebo works best.

b. The treatment was different. The same probiotics were not used in all the studies to decolonise the pathogens.

c. Some studies did not have a control group; therefore, it is difficult to conclude the results.

d. No meta-analysis was done to determine heterogeneity.

e. No patient and control group can compare the data for a direct assessment.

Grading of evidence

High certainty: We are highly confident that the true effect lies closer to that of the estimate of the effect

Moderate certainty: We are moderately confident that the true effect is likely to be close to the estimate of the effect. However, there is the possibility of a difference between the true effect and measures of estimate.

Low certainty:We are not confident that the true effect will likely be close to the effect size. There might be a substantial difference between the true effect and the estimate of the effect.

Very low certainty: The true effect is likely to differ substantially from the mesure estimate (effect size).

Sensitivity analysis

Post-hoc sensitivity analyses were conducted, keeping aside the high risk of bias studies according to JBI guidelines (Supplement Figure S7). We further analyze post-hoc sensitivity by excluding the most weighted (influential) studies (Supplement Figure S8). The intervention effect size for all included studies was 0.59 (favoring probiotics). However, the pooled intervention effect sizes (OR) of the probiotics intervention were 0.63 (p = 0.01) and 0.54 (p = 0.0001) after eliminating the potential high risk of biased and comparatively larger weighted studies, respectively. No significant differences were observed between the adjusted OR and overall OR (0.59) (p > 0.05, t-test of SPSS). Besides, heterogeneity (I² = 43%) was noticeably changed after eliminating biased and weighted studies from the analysis. A Trial Sequence Analysis (TSA) graph of 29 RCTs was constructed to confirm the effectiveness of probiotics intervention, which also shows that RCTs favor the probiotics intervention for decolonizing pathogens (Supplement Figure S9).

Characteristics of the selected prebiotic studies

A total of ten studies investigating prebiotic treatments for the decolonization or reduction of pathogenic bacterial burden from the intestine were included and are summarized in Supplement Table S8. The prebiotics used to reduce the burden of pathogens by the selected studies are Fructooligosaccharide (FOS), Xylooligosaccharide (XOS), Galactooligosaccharide (GOS), Xylooligosaccharide and mannose-oligosaccharide (MOS), Mannan-oligosaccharide, Beta-glucan and galactoglucomannan oligosaccharide (GGM). The dosages of the prebiotics varied with the types of prebiotics and the challenged microorganisms. The dosage of fructooligosaccharides ranged from 5% for low doses and 25% for high doses.^54–56 Mannan-oligosaccharide was used as a 2 g/kg basal diet for chickens.⁵⁷ In comparison, a dose of Galacto-oligosaccharides (GOS) 0.2 g/kg body weight was used to reduce the burden of Helicobacter and Clostridium.⁵⁸ However, the xylooligosaccharide dose was 1.2 gm mixed with 150 gm rice for the human trial.⁵⁹ Besides, Rajani et al.⁶⁰ mentioned that the galactoglucomannan oligosaccharides (GGMs) dose in their study was up to 0.3%. The prebiotics were mixed with drinking water or rice porridge for consumption (Table 7).

Table 7. Types and dosages of prebiotics.

Reference	Type of prebiotics	Dosage/day	Study subjects
Lin et al., 2016^Citation59	Xylooligosaccharides (XOS)	1.2 gm	Human
Mao et al., 2018^Citation56	Fructooligosaccharide (FOS)	5–25% of the regular diet	Rat
Adhikari et al., 2018^Citation54	Fructooligosaccharide (FOS)	1% of the regular diet	Poultry
Monteagudo-Mera et al., 2016^Citation58	Galactooligosaccharides (GOS)	1% of the regular diet	Mice
Ribeiro et al., 2007^Citation57	Mannan oligosaccharide – Bio Mos	2.5 mg	Poultry
Pourabedin et al., 2017^Citation61	Xylooligosaccharides and mannose-oligosaccharides	2 gm	Poultry
Spring et al., 2000^Citation62	Mannanoligosaccharides (MOS)	4,000 ppm of dietary MOS	Poultry
Lowry et al., 2005^Citation63	Beta-glucan	Not mentioned	Poultry
Rajani et al., 2016^Citation60	Galactoglucomannan oligosaccharides (GGMs)	0.3% of the regular diet	Poultry
Matsumoto et al., 2017^Citation55	Fructooligosaccharide (FOS)	5% of the regular diet	Mice

Decolonisation potential of prebiotics

The prebiotic studies showed the potential to eradicate pathogens from the gut. All studies used oligosaccharides to reduce the colonization of pathogens, though they have high heterogeneity in types and dosages of prebiotics. One study was conducted on humans to evaluate the efficacy of prebiotics on gut microbiota. The study showed that xylooligosaccharides reduced the Clostridium perfringens content from 4log to 2log CFU/g stool. The probiotics also reduced Coliform content, but the difference in CFU content is insignificant compared to the control group.⁵⁹ The frequently used prebiotic was fructooligosaccharide (FOS), which reduced the fecal shedding of Salmonella enterica significantly in the treatment group (10² CFU/g) compared to the control (10⁵ CFU/g) (initial challenge 10⁸ CFU).⁵⁴ FOS reduced Clostridium subcluster XIVa from 6% to 2.7% in the treatment group, whereas it increased from 5.9% to 26.3% in the no-treatment group.⁵⁵

Mao et al.⁵⁶ mentioned that FOS decreased Actinobacteria from 0.3% to 4.2% at a 95% confidence level (p < 0.05) in the treatment group. It also reduces Proteobacteria and Verrucomicrobia at different levels. Monteagudo-Mera et al.⁵⁸ showed that galacto-oligosaccharide (GOS) effectively reduced the abundance of Bacteroidales, Helicobacter, and Clostridium. According to Pourabedin et al.,⁶¹ mannose-oligosaccharide (MOS) was more effective than xylooligosaccharide (XOS) in reducing pathogens. MOS reduced the burden of Salmonella enteritidis by 1.6logs, whereas XOS reduced only 1log compared to the placebo group.⁶¹ Beta-glucan also reduced the burden of Salmonella enterica from 50% of subjects in the treatment group to 6% in the control group.⁶³ Spring et al.⁶² showed that mannan-oligosaccharide had reduced cecal S. typhimurium concentrations on Day 10 in the treatment group, which significantly differed from the control group (5.40 vs. 4.01 log CFU/g; p < 0.05). The highest rate of colonization reduction of S. typhimurium was observed in the chicks that received 0.2% GGM consisting of 2logs reduction⁶⁰ (Supplement Table S8).

Changes in microbial diversity in the gut

We narratively synthesize the metagenomic data of included studies to assess alpha diversity’s shifting in gut microbiota after having probiotics intervention.^46,51,64 According to the Shannon and Chao indexes, the bacterial diversity of the studies with metagenomic analysis had shifted after probiotic treatment (Table 8). Oh et al.⁵² and Yang et al.⁵¹ showed that probiotic intervention significantly increased the alpha diversity of bacteria in the gut compared with placebo. Conversely, the alpha diversity in the probiotics group was decreased compared to placebo in the study conducted by Chen et al.⁶⁵ However, the change in diversity was insignificant. Tang et al.⁴⁶ and Yang et al.⁵¹ reported a statistically significant shift in beta diversity in the probiotics group compared to the placebo. Piewngam et al.⁶⁴ also altered but did not reach statistical significance.

Table 8. Metagenomic analysis outcome on alpha diversity for probiotic intervention.

Reference					Metagenomic analysis outcome (probiotics vs placebo)
	Colonized bacteria	Types of probiotics and dose	Antibiotics/conventional treatment	Trial period (days) Probiotics +antibiotics	Status of Alpha diversity	p-value	Status of Beta diversity	p-value
Chen et al., 2021^Citation65	H. pylori	L. acidophilus and L. rhamnosus (one sachet containing 3 × 10^Citation9 CFU per sachet, twice a day for 28 days)	No antibiotics, only a placebo	28	Alpha diversity changed significantly	Shannon index (p= 0.097), Chao index (p = 0.05)	No significant difference was observed	PERMANOVA (p = 0.76)
Oh et al., 2016^Citation52	H. pylori	Streptococcus faecium 9x10^Citation8 CFU/day, Bacillus subtilis 1x10^Citation8 CFU/day	500 mg of clarithromycin, 1 g amoxicillin, and 30 mg of lansoprazole	14+14	Alpha diversity increased in the probiotics group	Shannon diversity (p<0.05)	Not analyzed
Piewngam et al.,2018^Citation64	S. aureus	Bacillus subtilis spores	No antibiotics	14	Alpha diversity did not change.	Shannon index (p value not mentioned) Chao index, p value not mentioned	Beta diversity changed but not significant
Yang et al., 2021^Citation51	H. pylori	Non-viable L. reuteri DSM17648 L. reuteri DSM17648 (1 × 1010/day)	Standard triple therapy (esomeprazole 20 mg, amoxicillin 1 g, and clarithromycin 500 mg)	28 +14	Alpha diversity changed significantly	Chao index (p = 0.001), Shannon index (p =0.001)	Beta diversity significantly increased in the probiotics group	(p = 0.002)
Tang et al., 2021^Citation46	H. pylori	Enterococcus faecium 4.5 x 10^Citation8 and B. subtilis 5.0 x 10^Citation7 (three times a day for four weeks)	14-day BQT (esomeprazole 20 mg, amoxicillin 1000 mg, furazolidone 100 mg, bismuth potassium citrate 220 mg, all given twice daily	28 +14	Alpha diversity increased after the treatment	Shannon diversity (p > 0.05) is not significant	Significant difference observed at 4 weeks	(p = 0.03)

Microbiome composition changes after probiotic treatments

The qualitative synthesis of the available data from the included studies revealed that the probiotics changed the gut microbiome compositions (Table 9). OH et al.⁵² revealed that after 14 days of probiotics treatment, the drug-resistant Klebsiella, Citrobacter, Pseudomonas, and Escherichia were reduced significantly in the probiotics group. In contrast, Klebsiella and Citrobacter were increased in the placebo group. As both arms received the combination therapy with antibiotics and proton pump inhibitor for treating H. pylori infection, the total Operational Taxonomic Units (OTUs) were reduced in both groups. However, the reduction was significantly higher in the placebo group.⁵² Chen et al.⁶⁵ mentioned that the Lactobacillus intervention changed the phyla and genera of bacteria in the gut significantly. After 28 days of treatment, Firmicutes and Actinobacteria were reduced in the probiotics group and increased in the placebo group. He et al.³³ showed that probiotics increased beneficial bacteria, including Bacteroides, Bifidobacterium, and Lactobacillus, but decreased Enterococcus and Enterobacter in the gut, and the reverse occurred in the gut of the placebo group. Yang et al.⁵¹ mentioned no change in the gut microbiota composition in the probiotics and placebo groups. However, some changes were observed in the phyla, as Bacteroides, Firmicutes, and Actinobacteria were increased in the probiotics group, while only Proteobacteria was increased in the placebo group.⁵¹ According to Tang et al.,⁴⁶ Fecalibacterium and Bacteroides were increased in the probiotics group but decreased in the placebo group. However, the changes observed in Shigella, Klebsiella, Streptococcus, Veillonella, Roseburia, Phascolaretobacterium, and Blautia were the same for the probiotic and placebo groups.⁴⁶ Piewngam et al.⁶⁴ also presented the same scenario as He et al.,³³ stating that after the probiotic treatment, the Firmicutes and Actinobacteria were increased, and Bacteroidetes and Proteobacteria were decreased in the probiotics group. In contrast, the reverse was noted in the placebo group.

Table 9. Metagenomic analysis outcome on the microbiome composition for probiotic intervention.

Study	Probiotics Treatment	Changes in the phyla after treatment		Changes in the genera after treatment
		Probiotics group	Placebo	Probiotics (Genera increased)	Probiotics (Genera decreased)	Placebo (Genera increased)	Placebo (Genera decreased)
Chen et al., 2021^Citation65	L. acidophilus and L. rhamnosus (one sachet containing 3 × 10^Citation9 CFU per sachet, twice a day for 28 days)	Firmicutes reduced, Actinobacteria and Bacteroidetes reduced but not significant.	Firmicutes and Bacteroides increased, Proteobacteria and Actinobacteria decreased	Stackia Eubacterium Mitsuokella Xylanophilum Pseudoflavonifactor	Prevotelleceae Peptococcus Enterococcus Paludecola Selimonas	Succinatimonas Gordonibacter Sangubacteroids Anaerostipes Erysipeletoclostridium Lactobacillus Hydrogenoanaerobacterium UCG-011	Pseudomonas Oxalobacter Roultella Dielma Frisingicoccus Lachanospiraceae Mitsuokella Hungatella
OH 2015^Citation52	Streptococcus faecium 9x10^Citation8 CFU/day, Bacillus subtilis 1x10^Citation8 CFU/day for 14 days	Firmicutes were reduced, Proteobacteria were increased The number of OTUs reduced from 231 to 186	Proteobacteria were increased, Firmicutes increased The number of OTUs reduced from 219 to 120		Drug-resistant Citrobacter, Klebsiella, PseudomonasEscherichia	Drug-resistant Klebsiella Citrobacter
He et al., 2022^Citation33	Saccharomyces boulardii (14 days)	Not analyzed	Not analyzed	Bifidobacterium Lactobacillus Bacteroides	Enterococcus Enterobacter	Enterococcus Enterobacter	Bifidobacterium Lactobacillus Bacteroids
Yang et al., 2021^Citation51	Non-viable L. reuteri DSM17648 L. reuteri DSM17648 (1 × 1010/day) (28 days)	Firmicutes Actinobacteria Bacteroides increased and Proteobacteria decreased	Only proteobacteria increased	Fecalibacterium Bifidobacterium Subdoligranulum Prevotella Collinsella Fuscicatenibacter Baceroids	Blautia Shigella Escherichia	Blautia Shigella Escherichia	Fecalibacterium Bifidobacterium Subdoligranulum Prevotella Collinsella Fuscicatenibacter Bacteroides
Tang et al., 2021^Citation46	Enterococcus faecium 4.5 x 10^Citation8 and B. subtilis 5.0 x 10^Citation7 (three times a day for 28 days)	Proteobacteria increased. Actinobacteria, Firmicutes, Bacteroidetes, and Fusobacteria remained the same.	Proteobacteria reduced, and Other phyla remained the same.	Shigella, Klebsiella, Streptococcus, Veillonella, Bacteroides, Faecalibacterium,	Roseburia, Phascolarctobacterium, Blautia	Shigella, Klebsiella, Streptococcus, Veillonella	Bacteroides, Faecalibacterium Roseburia, Phascolarctobacterium, Blautia
Piewngam et al., 018^Citation64	Bacillus subtillis spores	Actinobacteria, Firmicutes increased and Bacteroidetes, Proteobacteria decreased	Proteobacteria and Bacteroidetes increased and Firmicutes, Actinobacteria decreased	Not analyse	Not analyse

Discussion

The influence of the gut microbiota on the host physiology is well recognized. The increasing prevalence of multiple drug-resistant pathogens mandates the need to explore alternatives to antibiotics, particularly as antibiotics have long-term effects on gut microbiota and human health,^66–68 and may contribute to increasing undesirable microbial communities and resistant pathogens.^14,69,70 Besides, antibiotics do not consistently decolonize multi-drug-resistant bacteria. It may reduce healthy bacteria together with the pathogens from the gut.^69,71 Probiotics or prebiotics with antibiotics treatment may improve long-term effects on health and microbial diversity.^9,12,72 However, probiotics and prebiotics have received little attention as therapies for eliminating or decolonizing harmful bacteria. Results from this meta-analysis suggest that probiotics may decolonize pathogens from the intestinal tract irrespective of the types of bacteria. The results also revealed that all types of probiotics used in these studies could be effective against pathogens.

The use of antibiotics to decolonize pathogens is a common practice worldwide. Recent studies have established that antibiotics disturb the gut and alter the microbial composition, affecting normal gut homeostasis.^{14,68,70,71,73-76} Therefore, probiotics are used concomitantly with antibiotics to reduce long-term antibiotic effects on the gut microbiome ,^77-79 On the other hand, without antibiotics, probiotics showed successful decolonization of pathogens, which agrees with the results of recent studies.^13,28,32 Thus, considering the shortcomings of antibiotics, probiotics could be an alternative way of decolonizing potential pathogens.

However, differences in effect size are observed in the subgroup analysis (Figure 4). The highest decolonization rate was observed with S. boulardii, followed by other* probiotics (including E. coli Nissle 1971, E. faecium, and nontoxigenic C. difficile). S. boulardii was especially effective on CDI during or following antibiotics.^80,81 In fact, probiotic use was most effective in accelerating the decolonization rate of C. difficile, followed by MDR Enterobacteriaceae and vancomycin-resistant Enterococci (VRE) (Figure 5). This could be due to the varying efficacy of specific probiotics against particular pathogenic colonizers. The overall heterogeneity among the RCTs is minimal. These results are consistent with previous systematic reviews and meta-analyses conducted by Penumetcha et al.²⁰ and Wang et al.¹⁹ However, it was significantly reduced when we divided them into subgroups according to types of bacteria, probiotics or regions as these grouping collated the RCTs with the same characteristics.

We observed that the types of probiotics and pathogens significantly affect the results. Decolonization by probiotics depends on many factors, which may include competitive inhibition and exclusion of pathogens for adhesion sites, production of antimicrobial agents, improvement of immune response, and strengthening of the intestinal barrier.^82,83 However, other covariates, including probiotics doses and region of studies, did not moderate effect size (p > 0.05). There were no substantial differences in doses among most probiotics (from 10⁹ to 10¹⁰ CFU/day), so suggesting that the minimum quantities of probiotics should be given as 10⁹ CFU/day to have positive effects. These results are consistent with the study conducted by Christensen et al.⁸⁴ However, despite having proven benefits, certain issues such as the probiotics quality, supply chain, single or multiple-strains formula, and CFU quantitation required for decolonization should be further resolved through clinical research.

Although the changes in gut microbial diversity due to probiotic intervention are under dispute, some studies established that probiotics could shift the alpha or beta diversity in the gut.^46,65 Éliás et al.⁸⁵ stated that probiotics are ineffective at preventing antibiotic-induced low-diversity dysbiosis in the gut. Washburn et al.⁸⁶ also studied the effect of Bifidobacterium infantis commercial probiotics on gut microbiota diversity. They concluded that using a single-species probiotic has no appreciable impact on gut bacteria diversity in healthy individuals. Conversely, Muwonge et al.showed that probiotics could significantly change the diversity of the microbiome of an individual recovering from enteric pathogens.⁸⁷

Antibiotics are well known to disrupt the microbial composition and negatively impact the host’s health. Probiotic intervention modulates through minimizing the gut microbiota alteration and increasing the gut microbiome diversity. Even though some studies gave variable results on bacterial diversities,^46,51,52,65 in cases of antibiotics as eradication therapy, probiotics helped to establish a beneficial bacterial profile.⁴⁶ Grazul et al. stated that probiotics did not successfully colonize the intestine after antibiotic treatments and did not seem to alter the overall diversity of the gut microbiota.⁸⁸ However, probiotics effectively suppress the growth of Enterobacteriaceae, including harmful bacteria like Shigella and Escherichia, while promoting the growth of beneficial bacteria from the Anaerotruncus genus.⁸⁸ Therefore, supplementation of probiotics is a safe and effective way for the recovery of healthy gut microbiota.

Our analysis showed that probiotics might help shift the diminished alpha and beta diversity caused by antibiotic exposure. Although the RCTs with metagenomics data were limited, some studies showed reverse results, consistent with the meta-analysis, which demonstrated the effect of probiotics supplementation during antibiotics treatment.⁸⁵ Probiotics can alter the microbiome composition and increase the beneficial bacteria, increasing their functional activity and helping remove pathogens or harmful bacteria.⁸⁹ Our study also shows that probiotics changed the gut compositions, increasing beneficial bacteria and decreasing antibiotic-resistant bacteria, including Klebsiella, Citrobacter, Enterococcus, Enterobacter etc. Antibiotics-induced decrease in Bacteroidetes: Firmicutes (B: F) ratio is associated with several metabolic disorders.^90,91 Probiotics can balance the ratio and normalize the dysbiotic gut,⁹² as our results show Proteobacteria may be a microbiological hallmark of several illnesses, including inflammatory bowel disease and metabolic problems, which may make the host susceptible to infection by pathogens.⁹³ The included studies in our analysis showed that the Proteobacteria was decreased in the probiotics group and increased in the placebo group.

We did not assess bias for the prebiotics study selected for narrative reviews since these studies did not have any control arm. There is a scarcity of prebiotics intervention studies conducted on human subjects. The dosages of prebiotics were not consistent among the studies. About 1.2 to 2.5 g of oligosaccharides was used to decolonize the pathogens. However, a beta-glucan-supplemented diet of up to 40 mg/kg of chicken significantly up-regulated innate immunity and significantly reduced the S. enterica organ invasion.⁶³ Though the exact mechanisms are yet to be elucidated, the authors speculated that the purified beta-glucan helps to prime the heterophils to phagocytize and kill invasive S. enterica by releasing oxygen-free radicals.⁶³ Another possible mechanism could be the dectin-1 receptor-mediated event against the invading enteropathogens.^94,95 The dectin-1 receptor on the surface of mammalian macrophages and neutrophils is specific to beta-glucan and activated once it binds with beta-glucan.⁹⁶

The impact of prebiotics on the gut microbiome and their mechanism of action are subjects of increasing interest and have become a wider topic of debate. Prebiotics selectively influence the growth and activity of probiotics or friendly bacteria and help in -homeostasis in the gut.^10,97,98 In =homeostatic conditions, certain Gram-negative commensal bacteria stimulate antibody production, such as IgA and IgG, identifying Gram-negative bacteria surface antigens and contributing to host defense against symbionts and pathogens.⁹⁹ Besides, the intestinal flora breaks them down and selectively ferments them to create specific secondary metabolites.⁹⁸ These metabolites are then absorbed by the intestinal epithelium or go to the liver via the portal vein, where they can improve the function of the intestinal barrier, regulate immunity, and resist pathogens.^98,99

It has already been established that probiotics use prebiotics as their food source. However, it raised a burning question of whether pathogenic bacteria use prebiotics as their food to grow. Very few studies have explored this issue. According to recent research, inulin and FOS promote the growth of Klebsiella, a bacterium linked to increased intestinal permeability.¹⁰⁰ Valyshev (2000)¹⁰¹ stated that inulin increased the growth of Klebsiella, whereas it restricted E. coli and Salmonella. According to the literature review, prebiotics may enhance the growth of certain pathogens, but this is far exceeded by the rate at which they increase friendly bacteria.^9,102

Research on the impact of prebiotics on the fecal microbiota of healthy human volunteers revealed that Lactobacilli and Bifidobacterium proliferated, with a slight increase of Enterobacteriaceae (mainly Klebsiella),^103-104,105 Recent research also showed that prebiotics appear to primarily increase the amount of friendly bacteria in the gut, especially Lactobacillus and Bifidobacteria.^97,106,107 It should be noted that although this study was conducted in vitro, it is always preferable to have results verified by human clinical trials. Prebiotics increase the ratio of “good” to “bad” bacteria. Still, the precise rate at which various genus and species in the gut alter in response to varying kinds and dosages of prebiotics is unknown. Moreover, more research is needed to confirm that the pathogenic load does not reach an unhealthy level due to prebiotic application.

Conclusions

Probiotics and prebiotics may help to decolonize pathogens from the gut or increase the effectiveness of conventional medicine to decolonize multidrug-resistant pathogens. The certainty of the evidence of probiotic intervention in the subgroup was moderate to high. Though the overall effect size favored probiotic intervention, more clinical trials are warranted to make a robust conclusion. Prebiotics help maintain a healthy gut by influencing the growth of friendly bacteria, but human clinical trials are required to determine whether they enhance the growth of pathogens. Dosage, type of probiotic(s), and pathogens for decolonization are important factors influencing the effect size, and thus, maintaining uniformity in the properties or composition of probiotics and prebiotics, dosages, and categories of patient inclusion for conducting RCTs is necessary.

Author Contributions

Conceptualization, MI; methodology, NR and NB; writing – original draft preparation, NR, NB, MT, PD; writing – Review and editing, MI, MCFT, SHW; supervision, MI; project administration. All authors have read and agreed to the published version of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19490976.2024.2356279

Additional information

Funding

We thank the Health and Medical Research Fund, Food and Health Bureau, HKSAR for the support (Project no. HMRF 18170082 and CID_CUHK_C, both PI to MI). The funding body was not involved in the study’s design, data collection, analysis, or interpretation of data or in writing the manuscript.