Effects of postbiotics on chronic diarrhea in young adults: a randomized, double-blind, placebo-controlled crossover trial assessing clinical symptoms, gut microbiota, and metabolite profiles

ABSTRACT

Chronic diarrhea has a considerable impact on quality of life. This randomized, double-blind, placebo-controlled crossover intervention trial was conducted with 69 participants (36 in Group A, 33 in Group B), aiming to investigate the potential of postbiotics in alleviating diarrhea-associated symptoms. Participants received postbiotic Probio-Eco® and placebo for 21 days each in alternating order, with a 14-day washout period between interventions. The results showed that postbiotic intake resulted in significant improvements in Bristol stool scale score, defecation frequency, urgency, and anxiety. Moreover, the postbiotic intervention increased beneficial intestinal bacteria, including Dysosmobacter welbionis and Faecalibacterium prausnitzii, while reducing potential pathogens like Megamonas funiformis. The levels of gut Microviridae notably increased. Non-targeted metabolomics analysis revealed postbiotic-driven enrichment of beneficial metabolites, including α-linolenic acid and p-methoxycinnamic acid, and reduction of diarrhea-associated metabolites, including theophylline, piperine, capsaicin, and phenylalanine. Targeted metabolomics confirmed a significant increase in fecal butyric acid after postbiotic intervention. The levels of aromatic amino acids, phenylalanine and tryptophan, and their related metabolites, 5-hydroxytryptophan and kynurenine, decreased after the postbiotic intervention, suggesting diarrhea alleviation was through modulating the tryptophan-5-hydroxytryptamine and tryptophan-kynurenine pathways. Additionally, chenodeoxycholic acid, a diarrhea-linked primary bile acid, decreased substantially. In conclusion, postbiotics have shown promise in relieving chronic diarrhea.

1. Introduction

Diarrhea is a condition characterized by increased water content in the stool, resulting in softening of the stool and frequent bowel movements. It is caused by damage to ion transport channel proteins, channels, and physical and chemical barriers in the intestine, which disrupts water and electrolyte transport in the digestive tract.¹ Diarrhea is classified into three categories based on duration: acute (up to 14 days), persistent (14–30 days) and chronic (over 30 days).² According to the guidelines of the British Society of Gastroenterology on investigating chronic diarrhea in adults (3rd edition), chronic diarrhea is defined as a persistent deviation from normal stool consistency (types 5 to 7 on the Bristol stool chart) with increased frequency lasting over four weeks.³ While the prevalence of chronic diarrhea in the general population is estimated to be around 4%–5%,⁴ diagnosing the specific underlying pathological conditions associated with it can be challenging in many cases.

The human intestinal tract harbors a diverse assortment of microorganisms, including bacteria, fungi, protozoa, and viruses. These microbes establish a symbiotic relationship with the host, fostering colonic homeostasis and effectively inhibiting the colonization of harmful pathogens. The gut microbiota also actively contributes to fundamental physiological functions like nutrient absorption and immune regulation.⁵ Emerging evidence indicates that an imbalanced gut microbiota heightens the risk of infections by various pathogens and contributes to the development of numerous diseases, including diarrhea, irritable bowel syndrome (IBS), colorectal cancer, metabolic disorders, and obesity.⁶ Probiotics are widely recognized for their potential health benefits, including the alleviation of different types of diarrhea.⁷ Probiotics inhibit the colonization of harmful bacteria through nutrient competition and the production of antibacterial compounds.⁸ They also help maintain the balance of the intestinal microbial community,⁷ boost the immune system,⁹ reinforce the defense of the intestinal barrier,¹⁰ and produce beneficial metabolites.¹¹ However, the use of Probiotics in vulnerable populations raises potential safety concerns, and ensuring their viability presents a substantial challenge. Consequently, there is a growing interest in exploring alternatives to Probiotics.

Postbiotics are formulations composed of various bioactive compounds with diverse mechanisms of action.¹² They encompass inactivated microbial cells or cellular components, making them more stable, while offering health benefits similar to Probiotics. For example, heat-treated Lacticaseibacillus paracasei could alleviate diarrhea, significantly reducing its frequency.¹³ Postbiotics, such as heat-treated Lactobacillus LB, have shown promise in mitigating antibiotic-associated diarrhea in patients receiving antibiotics.¹⁴ However, the precise mechanisms underlying the beneficial effects of postbiotics and their impact on human health remain largely unclear in many cases. These mechanisms may act independently or in combination, sometimes resembling known Probiotic mechanisms of action. Postbiotics have garnered increasing attention due to their perceived advantages and potential applications as safer alternatives to Probiotics in improving health and managing diseases. Exploring their specific beneficial mechanisms and interactions with local gut commensals is also of great interest, as it paves the way for designing novel targeted postbiotic therapies.

The objective of this work was to conduct a randomized, placebo-controlled crossover trial to evaluate the therapeutic effects of postbiotics on chronic diarrhea. The study assessed stool frequency and consistency, quality of life, along with changes in fecal microbiota and biomarkers associated with metabolic health and function in subjects with chronic diarrhea. The long-term goal of this study is to present empirical evidence supporting the use of postbiotics as a potentially effective alternative treatment for chronic diarrhea and lay a foundation for a deeper understanding of the underlying mechanisms that contribute to their efficacy.

2. Materials and methods

2.1. Preparation of postbiotic Probio-Eco®

Probio-Eco® is a pill-form Probiotic product manufactured by Jiangzhong Pharmaceutical Company Limited (Nanchang, China). The product was made by thoroughly mixing soy powder, skim milk, and sodium citrate with water for 15 min and heated to 55–60°C. The mixture was homogenized at 18–20 MPa with the GEA Lab Homogenizer PandaPLUS 1000 (Germany Engineering Alliance Group, Dusseldorf, Germany), followed by sterilization at 95°C for 60 min. Afterward, the mixture was cooled to 35°C, and three Probiotic strains (Lacticaseibacillus paracasei Zhang, Lactiplantibacillus plantarum p-8, and Bifidobacterium animalis subsp. lactis V9) were inoculated for fermentation at a constant temperature of 35°C. The fermentation process ended when the pH value reached 4.5–4.6. Subsequently, the fermentation broth was heated to 75°C for 15 min, followed by aseptic homogenization (25–30 MPa). Probio-Eco® dried powder was obtained by spray-drying the fermentation broth using a SZ-2000 spray dryer (Shanghai Shunzhi Instrument Manufacturing Co., Ltd., Shanghai, China) with an inlet and outlet air temperature of 210°C and 88°C, respectively. The powder was compressed into triangular tablets with a ZP-17D rotary tablet press. The placebo, a control for the postbiotic product, was made of soy powder, skim milk, and sodium citrate. Prior analysis confirmed that this placebo material did not contain any functional molecules, such as short-chain fatty acids (SCFAs). It was designed to have a similar physical appearance and taste to the postbiotic product.

Both Probio-Eco® and the placebo underwent rigorous testing to ensure compliance with strict standards for contaminants (National standard for food safety – Limits of Contaminants in foods (GB 2762–2022)), mycotoxins (National standard for food safety – Limits of mycotoxins in foods (GB 2761–2017)), and microbial limits (National Food Safety Standard Milk Powder (GB 19,644–2010)). The moisture content of both products was maintained below 4.0% by using a moisture analyzer (Ohaus Instrument Co., Ltd., Shanghai, China). The water activity was kept below 0.30 by using a water activity meter (Huake Instrument Co., Ltd., Wuxi, China).

2.2. Trial design

The study was a double-blind, randomized, placebo-controlled crossover intervention trial conducted at the Inner Mongolia Agricultural University in Inner Mongolia, China. Participants were randomly assigned to receive two treatments in a reciprocal sequence: the postbiotic Probio-Eco® and a placebo. Before the interventions, a 7-day observation period was implemented. During the initial treatment phase, half of the participants were randomly assigned to receive either postbiotic Probio-Eco® or the placebo. The allocation was then reversed during the second treatment phase. Each treatment period lasted for 21 days, followed by a 14-day washout period. Participants were instructed to refrain from consuming any post- or Probiotic products throughout the observation and washout periods. To ensure blinding, the placebo was carefully designed to closely resemble the postbiotic Probio-Eco® in both appearance and taste. Both the postbiotic and placebo products were unlabeled and indistinguishable pastel yellow tablets, each weighing 0.6 g. Participants consumed 12 tablets of Probio-Eco® or the placebo daily.

Each participant provided a fecal sample before starting either the Probio-Eco® or placebo treatment. After 21 days of supplementation and before starting the second arm of the study following the washout period, another fecal sample was collected. The sample process was repeated upon completing the crossover, resulting in a total collection of four fecal samples per participant.

2.3. Sample size calculation

The minimum sample sizes required for this study were calculated using the PASS statistical software (ncss.com/software/pass.; NCSS, LLC., Kaysville, UT, USA). The sample size determination was based on the assumption of a 2-point difference in stool frequency between the pre- and post-postbiotic intervention periods, with α error and power set as 0.025 and 90%, respectively. Using these parameters, the calculated minimum sample size was 23 participants per group. To account for an anticipated 20% loss to follow-up, the target enrollment was increased to 29 participants per group, resulting in a total of 58 participants for the entire crossover trial.

2.4. Study cohort

The study participants underwent a rigorous screening process based on predefined inclusion and exclusion criteria. A total of 69 adults between the ages of 18 and 35 met the eligibility criteria and were included in the study.

Randomization was performed using computer-generated lists, and participants were randomly assigned to either Group A (postbiotic followed by placebo) or Group B (placebo followed by postbiotic). The randomization process was conducted by a statistician who had no contact with the participants, ensuring impartiality. All individuals involved in this study, including participants, doctors, and researchers, remained unaware of the randomization sequences until the end of the study.

2.4.1. Inclusion criteria

The study enrolled men and women aged 18 to 65 who suffered from chronic diarrhea. Participants had to experience persistent diarrhea for at least six months, with loose or watery stool (Bristol type 5, 6, or 7) occurring in at least 25% of defecation over the past three months. For individuals between the ages of 18 (exclusive) to 50 (inclusive), the results of the stool test (including occult blood) obtained during the screening process must be normal or medically insignificant as judged by the investigators. Individuals between the ages of 50 (exclusive) and 65 (inclusive) were eligible for inclusion if they had undergone a colonoscopy at a tertiary or higher-level hospital within the past six months, showing either normal or clinically irrelevant results as determined by the investigators. Furthermore, participants had to submit a signed informed consent prior to the start of the study to be included in the trial.

2.4.2. Exclusion criteria

Participants were excluded if they had: 1) a personal or family history of colon cancer, celiac disease, or inflammatory bowel disease; 2) received a diagnosis of confirmed intestinal organic diseases via colonoscopy; 3) planned to become pregnant or father a child in the near future, or were currently pregnant or breastfeeding; 4) known allergies to Probiotics, postbiotics or the placebo; 5) taken antibiotics, Probiotics, postbiotics within the past two weeks; 6) taken antianxiety, antidepressant, or other psychotropic drugs within the past month; 7) engaged in long-term use of medications for diarrhea; 8) a history of severe diseases, such as myocardial infarction, cerebral infarction, or malignant tumors; 9) major mental illnesses, an inability to control their actions, or a lack of cooperation; 10) illiteracy, difficulty in comprehending, or an inability to independently sign the informed consent form.

2.5. Self-administered questionnaire

Participants were requested to maintain a daily diary, which served as input for the weekly questionnaire. This questionnaire gathered information on several aspects, including defecation frequency, Bristol stool scale, and urgency, any adverse events experienced, changes in daily habits (such as physical activity, smoking, alcohol consumption, and fluid intake), compliance with the study protocol, and medication intake.

The primary outcome measure was the defecation frequency. Secondary outcomes included the effects of postbiotics on various diarrhea-related parameters, such as the Bristol stool scale score, defecation urgency, mental health, gut microbiome profiles, and gut metabolite profiles. Clinical efficacy was determined by calculating the proportion of subjects whose defecation frequency fell within the range of “0 < defecation times/day ≤2” following the postbiotic intervention.

Stool consistency was assessed using the Bristol Stool Form Scale, an ordinal scale that ranges from type 1 (the hardest) to type 7 (the softest), with scores of 1 to 7, respectively. Defecation frequency was scored as follows: 1 point for once a day. Defecation urgency was graded on a scale of: 0 (no urgency), 1 (slight urgency), 2 (sudden need for immediate defecation), and 3 (fecal incontinence). The Depression, Anxiety, and Stress Scales 21 (DASS-21) was used to measure mental health through three self-report scales. A higher score indicates greater severity of depression, anxiety, and stress symptoms.

2.6. Metagenomic DNA extraction for shotgun sequencing

Metagenomic DNA was extracted from fecal samples using the QIAamp Fast DNA stool mini kit (Qiagen GmbH, Hilden, Germany). DNA concentration and quality of the extracted DNA were evaluated using a Nanodrop spectrophotometer, the Qubit double-stranded DNA assay kit, and a Qubit 2.0 fluorometer (Life Technologies, Carlsbad, CA, USA). Furthermore, DNA integrity was assessed using 1% agarose gel electrophoresis.

All qualified samples (a total of 276 samples; n = 36 and 33 in Groups A and B, respectively; four time point samples per subject) were subjected to shotgun metagenomic sequencing on an Illumina NovaSeq instrument (Illumina Inc., San Diego, CA, USA). The NEBNext Ultra DNA library prep kit (New England BioLabs, Ipswich, MA, USA) was used following the manufacturer’s instruction to construct the library. KneadData v0.7.5 (accessible at http://huttenhower.sph.harvard.edu/kneaddata) was used to filter out any low-quality sequences.

2.7. Read assembly, contig binning, genome dereplication, and bacterial taxonomic annotation

The quality-controlled sequences were assembled into contigs using MEGAHIT.¹⁵ Contigs with a minimum length of 2,000 bp were selected for binning through the VAMB tool with default options. The completeness and contamination levels of the metagenome-assembled genomes (MAGs) were assessed using CheckM (https://github.com/Ecogenomics/CheckM).¹⁶ High-quality genomes were defined as those with over 80% completeness and less than 5% contamination. Then, clustering was performed on these genomes. From each replicate set, the most representative genome was chosen using dRep.¹⁷ These representative genomes were used to extract species-level genome bins (SGBs) with parameters, -pa 0.95 and -sa 0.95, yielding a total of 304 SGBs.¹⁷ These SGBs were annotated using the NCBI nonredundant nucleotide sequence database. Predicted genes were BLASTp-searched against the UniProt Knowledgebase (UniProtKB; release 2020.11) using the DIAMOND tool with default settings. The relative abundance of each SGB was quantified using CheckM.¹⁸ The average SGB content in each contig was calculated and normalized as reads per kilobase per million (RPKM) using CoverM (https://github.com/wwood/CoverM). Next, sample diversity was calculated by using two R packages (vegan and optparse) based on the SGB abundance expressed in RPKM.

2.8. Bacteriophage contig identification and abundance analysis

First, potential bacteriophage features were identified from contigs exceeding 1,000 bp in length using both VIBRANT¹⁹ and Checkv.²⁰ Next, candidate bacteriophage sequences meeting the following criteria were selected: greater than 5,000 bp in size, exhibited a 95% or greater nucleotide identity, and had a sequence coverage of at least 80% when compared against CD-HIT (https://github.com/weizhongli/cdhit) to identify viral operational taxonomic units (vOTUs). This selection process yielded 145,589 vOTUs. To assess their novelty, these vOTUs were compared against the Metagenomic Gut Virus catalog (2021).²¹ Finally, the CoverM contig pipeline (https://github.com/wwood/CoverM) was utilized to calculate the average abundance of these vOTUs across the viral population using the following parameters: –min-read-percent-identity 0.95, –min-read-aligned-percent 0.5, –proper-pairs-only, and – exclude-supplementary.

2.9. Targeted quantitative metabolomics analysis

All chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA). Ultrapure water was generated using a Milli-Q water purification system (Millipore, Molsheim, France). Acetonitrile, methanol, and formic acid were of high-performance liquid chromatography grade.

2.9.1. Sample preparation for determining SCFAs

Fecal and postbiotic Probio-Eco® samples (0.1 g) were weighed and transferred into 2 mL sterile homogenizer tubes containing zirconium beads (3.0 mm diameter). Next, a 50% methanol-water solution (2 mL) was added to the samples, followed by homogenization using an OMNI Bead Ruptor homogenizer (55 Hz, 1 min). After homogenization, the samples were centrifuged (4°C, 12000 × g, 10 min). Then, supernatants (40 μL) were mixed thoroughly with 3-nitrophenylhydrazine (200 mM, 20 μL) and N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide-6% pyridine solution (120 mM, 20 μL). The mixtures were incubated for 30 min in a 40°C water bath. To terminate the reaction, a 10% acetonitrile-water solution (1.92 mL) was added. Samples were then centrifuged (4°C, 12000 × g, 10 min). Supernatants (500 μL) were filtered through a 0.22 μm filter and transferred to appropriate sample vials for further analyses.

2.9.2. Sample preparation for determining bile acids (BAs), amino acids (AAs), and other metabolites

Fecal samples (0.1 g) were weighed and transferred into sterile homogenizer tubes containing zirconium beads (3.0 mm diameter). Next, a 50% methanol-water solution (2 mL) was added to the samples, followed by homogenization using an OMNI Bead Ruptor homogenizer (55 Hz, 1 min). After homogenization, the samples were centrifuged (4°C, 12000 × g, 10 min). Supernatants (500 μL) were filtered through a 0.22 μm filter and transferred to appropriate sample vials for further analyses.

2.9.3. Targeted quantitative metabolomics analysis by ultra-high-performance liquid chromatography coupled with a triple quadrupole mass spectrometry (UPLC-QqQ-ms/ms)

Fecal and postbiotic Probio-Eco® samples were analyzed using targeted quantitative metabolomics with a UPLC-QqQ-MS/MS system (SCIEX Exion LC coupled to a SCIEX QTRAP 6500+; AB SCIEX, Foster City, CA, USA). This analysis utilized a Kinetex EVO C18 column (2.1 mm × 100 mm, 1.7 μm, Phenomenex, Torrance, CA, USA). The UPLC conditions were: autoinjector temperature, 4°C; column incubator temperature, 40°C; mobile phase A, ultrapure water containing 0.1% formic acid; mobile phase B, methanol containing 0.1% formic acid. The gradient elution was programmed as follows: 0.0–10.0 min, 5.0% B; 1.0–6.5 min, 5.0–100.0% B; 6.5–9.5 min, 100.0–100.0% B. 9.5–11.0 min, 100.0%-5.0% B: 11.0–12.0 min 5.0% B. The flow rate was 0.4 mL/min with an injection volume of 1 μL. The mass spectrometry parameters were set as follows: curtain gas at 25 psi; ion source gas 1 and 2 at 50 psi; source temperature at 550°C; ion spray voltage in positive and negative modes at 5,500 V and −4,500 V, respectively; nitrogen as the carrier gas. Mass spectrometry scanning was conducted using the multiple reaction monitoring scanning mode.

2.10. Non-targeted qualitative metabolomics

Fecal and postbiotic Probio-Eco® samples (20 mg) were weighed and transferred into clean centrifuge tubes. A methanol-water internal standard consisting of 70% methanol and 30% water (400 μL) was added to each sample. Afterward, the samples were vortexed for 3 min, followed by sonication in an ice water bath for 10 min. Following sonication, a further vortexing step for 1 min was performed before allowing the samples to stand at −20°C for 30 min. Finally, the samples were centrifuged at 12,000 rpm for 10 min (4 °C). Sample supernatants (200 μL) were carefully transferred to new centrifuge tubes for another round of centrifugation at 12,000 rpm for 3 min. Afterward, aliquots of 200 μL of sample supernatants were transferred into injection bottles for subsequent analysis. Throughout the instrumental analysis, we injected a quality control (QC) sample after every set of 15 samples. The quality control sample was created by combining identical volumes of each sample extract. The application of the QC sample ensured the instrumental stability and experimental reproducibility over the liquid chromatography-mass spectrometry (LC-MS) analysis.

The analysis was conducted using a UPLC system with a Waters ACQUITY UPLC HSS T3 C18 column (1.8 µm, 2.1 mm × 100 mm; Waters Corporation, Milford, MA, USA) operated at a column temperature of 40°C. The flow rate was set at 0.4 mL/min, and the injection volume was 2 μL. The solvent system consisted of water (0.1% formic acid) and acetonitrile (0.1% formic acid). The gradient program was as follows: 95:5 (v/v) at 0 min, 10:90 (v/v) at 11.0 min, 10:90 (v/v) at 12.0 min, 95:5 (v/v) at 12.1 min, and 95:5 (v/v) at 14.0 min.

For data analysis, the original LC-MS data file was converted into mzML format by ProteoWizard software. Peak extraction, peak alignment, and retention time correction were performed using the XCMS program. The peak area was corrected using the “SVR” method. Peaks with a detection rate lower than 50% in each group of samples were discarded. Metabolic identification information was then obtained by searching a self-built database in our laboratory and the METLIN database (https://metlin.scripps.edu/).

2.11. Statistical analyses

The non-targeted metabolomics analysis data were imported into MetaboAnalyst 5.0 (https://www.metaboanalyst.ca) and R software (version 4.1.2) for further processing and multidimensional statistical analysis. All other statistical analyses were performed using R software (version 4.1.2). Wilcoxon rank-sum tests were used to evaluate the differences in the abundance of SGBs, phages, metabolites, diarrhea indicators, and other variables between groups. All P-values were corrected for multiple testing using the Benjamini-Hochberg procedure, and a corrected P-value <0.05 was considered statistically significant. All graphical representations were generated using the R software in conjunction with Adobe Illustrator.

3. Results

3.1. Metabolites in Probio-Eco®

The metabolites in Probio-Eco® were analyzed using targeted and non-targeted metabolomics approaches, resulting in the identification of 30 compounds (Table S1), including 10 organic acids, five SCFAs, and 15 health-promoting bioactive molecules.

3.2. Participant flow chart

The flow of participants in the randomized, double-blind crossover trial is shown in Figure 1(a). Initially, 72 participants were recruited for this study and were randomized to participate in the intervention study. They were further divided into two groups, with 36 participants in each group. Following the completion of the first intervention phase, three participants (all from Group B) withdrew from the study due to antibiotic consumption. As a result, a total of 69 individuals (36 and 33 in Groups A and B, respectively) completed the study. Importantly, no adverse effects related to the interventions described in the study protocol were observed.

Figure 1.

Study design and changes in psychological state and diarrhea symptoms during the intervention study. (a) The study was a randomized double-blind placebo-controlled crossover trial. Participants were randomly assigned to receive two 21-day treatments in a reciprocal sequence: the postbiotic Probio-Eco® and a placebo, with a 14-day washout period in between. Participants were randomly assigned to either group a (postbiotic-placebo) or group B (placebo-postbiotic), and 69 participants completed the study (n = 36 and 33 in groups A and B, respectively). Fecal samples were collected from each participant before and after Probio-Eco® and placebo treatments, resulting in a total of four fecal samples per participant. Participants also kept a daily diary, providing information for the weekly questionnaire. Longitudinal changes in (b) the distribution of participants reporting different degrees of diarrhea symptoms, including defecation frequency, the Bristol stool scale score, and defecation urgency (shown in the left, middle, and right panels, respectively) and (c) the psychological state of subjects reflected by anxiety, stress, and depression scores (shown in the left, middle, and right panels, respectively) in groups A and B. (d) The violin plots show the cross-sectional analysis results of diarrhea symptom and psychological state scores between groups A and B during the first phase of treatment when group a received the postbiotic and group B received the placebo. The Wilcoxon rank-sum test was used to evaluate statistical differences between groups, and the resultant p-values are shown.

3.3. Subjects’ baseline characteristics

The baseline characteristics of all participants are shown in Table S2. No significant differences were observed between the two groups in terms of age, gender, body mass index, smoking status, or previous treatments for intestinal diarrhea (p > .05), suggesting that the randomization process effectively balanced these demographic and clinical factors across the groups, minimizing potential confounding effects.

3.4. Postbiotic administration improved diarrhea-associated clinical parameters

During the postbiotic Probio-Eco® ingestion period, both groups exhibited changes in clinical parameters, such as defecation frequency, the Bristol stool scale, and defecation urgency. In Group A, this period spanned from day 0 to day 21, while in Group B, it was from day 35 to day 56 (Figure 1(b)). Analysis of the effect of postbiotics on defecation frequency demonstrated a significant increase in the proportion of individuals transitioning from “2 < defecation times/day ≤4” to “0 < defecation times/day ≤2” after the postbiotic consumption. In Group A, 33 subjects reported a defecation frequency of “0 < defecation frequency/day ≤2” following the ProBio-Eco® intervention, while in Group B, 29 subjects exhibited the same frequency after the intervention. Thus, the overall diarrhea improvement rate for diarrhea was 89.86%. A comparison to the baseline revealed a tendency for an increase in the proportion of individuals with “normal consistency” in both groups following the postbiotic intervention. Additionally, the impact of postbiotics on defecation urgency indicated an increase in the proportion of individuals with “no urgency” in both groups after the postbiotics intervention. These findings, in conjunction with the results in Figure S1, suggest significant effects of postbiotics on the Bristol stool scale, defecation frequency, and defecation urgency in subjects (p < .05), with Group A exhibiting a more favorable response.

On the other hand, differential clinical responses were also observed during the placebo ingestion period. In Group A, this period spanned from day 35 to day 56, while in Group B, it was from day 0 to day 21. Analysis of the changes in clinical parameters found significant improvements during the placebo phase only in Group A (p < .05), but not in Group B, showing no significant differences compared to baseline levels (p > .05). This observation suggests a carry-over effect of the postbiotic intervention, with the impact of the initial intervention persisting into the second period.

Additionally, this study assessed the psychological state of the subjects, specifically focusing on anxiety, stress, and depression (Figure 1(c)). The results indicated a significant effect of postbiotics on the subjects’ levels of anxiety (p < .05), but not on stress or depression (p > .05).

To further eliminate the effect of time on the postbiotic effect, cross-sectional analyses were conducted at the same time point. Due to the presence of a carry-over effect, only data collected from the first phase of the intervention trial (when Group A received the postbiotic and Group B received the placebo) were analyzed. The comparison of clinical parameters and the psychological state of the subjects in the postbiotic and placebo groups revealed no significant differences in defecation frequency, Bristol stool scale, defecation urgency, anxiety, stress, and depression between the two groups at baseline (day 0; p > .05). However, after 21 days of postbiotics intervention, significant improvements were observed in frequency scores, Bristol stool scale scores, defecation urgency, and anxiety (p < .05), while there was no significant effect on stress and depression (p > .05; Figure 1(d)). These findings collectively suggest that postbiotics administration alleviated subjects’ diarrhea symptoms and improved their overall quality of life.

3.5. Postbiotic effects on the gut microbiota diversity and composition

To assess the changes in the gut microbiota under different treatments, metagenomic sequencing was conducted on 276 fecal samples. We first analyzed the changes in fecal microbial diversity to understand the impact of postbiotic administration on the gut microbiota. Our analysis revealed no significant differences in Shannon and Simpson diversity between groups or time points (p > .05; Figure 2(a)). Subsequently, a non-metric multidimensional scaling analysis was performed to visualize the changes in the fecal microbiota structure. The symbols representing the four subgroups showed a high degree of overlap on the non-metric multidimensional scaling score plots (Figure 2(b)), aligning with the non-significant results obtained from the analysis of similarity (p > .05; Table S3). These findings collectively indicate that there were no apparent differences in overall microbiota diversity and structural composition between groups at any time point.

Figure 2.

Fecal microbiota features of the subjects and their correlation with clinical indicators of diarrhea. (a) Shannon and Simpson diversity indexes and (b) non-metric multidimensional scaling (NMDS) analysis of the fecal microbiota of groups A and B. (c) K-means clustering analysis was conducted to investigate abundance changes in species-level genome bins (SGBs) in groups A and B during the intervention trial. The analysis grouped consistently changing SGBs into eight clusters (shown in the line charts). The x-axis represents the time points of the intervention trial, while the y-axis shows the standardized Z-scores of each SGB. The number of SGBs assigned to a particular cluster is written next to the corresponding cluster number. The standardized Z-score changes of an SGB are shown by the thin lines, while the thick lines represent the average change trends for all SGBs in that cluster. The red frames highlight postbiotic-responsive clusters of SGBs, which were exclusively influenced by the postbiotic intervention and not the placebo. (d, e) boxplots and heatmaps show the changes in abundance of postbiotic-responsive SGBs found by K-means cluster analysis and their correlations with the clinical indicators of diarrhea. The Wilcoxon rank-sum test was used to evaluate abundance differences in SGBs between time points, and significant p-values are shown (cut-off level: p < .05). The color scale in each heatmap represents correlation strength, positive and negative correlations are represented by red and purple, respectively. Significant correlations are represented by asterisks (*: p < .05; **: p < .01; ***: p < .001).

To determine if the diarrhea symptom improvements were associated with the changes in subjects’ fecal microbiota after the postbiotic intervention, we next analyzed the group-based postinterventional changes in the fecal microbiota to identify responsive SGBs exhibiting differential abundance exclusively following postbiotic intervention but not placebo treatment. We found 21 and 16 differentially abundant SGBs in Groups A and B, respectively (Table S4). We then identified postbiotic-responsive clusters from these SGBs by K-mean clustering analysis, which assigned SGBs showing consistent changes into the same cluster. In Group A, there were four postbiotic-responsive clusters, involving 11 SGBs after the postbiotic intervention (five increased SGBs: Faecalibacterium sp900765105, Lawsonibacter sp000177015, Fournierella sp002160145, Pseudoruminococcus massiliensis, and Ruminococcus_C sp000980705; six decreased SGBs: Succinivibrio sp000431835, Amedibacterium intestinale, Duodenibacillus sp900538905, Fusobacterium_A necrogenes, Fusobacterium_A sp900555845, and Ruminococcus_B gnavus; p < .05 in all cases; Figure 2(c,d), Table S4). In Group B, there were three postbiotic-responsive clusters, involving 10 SGBs (seven increased SGBs: Alistipes finegoldii, Bifidobacterium aerophilum, Angelakisella sp900547385, Dialister hominis, Dysosmobacter sp001916835, Dysosmobacter welbionis, and Lawsonibacter sp900066825; three decreased SGBs: Alistipes dispar, Bifidobacterium adolescentis, and Megamonas funiformis; p < .05 in all cases; Figure 2(c,d), Table S4).

Additionally, a cross-sectional comparison between the fecal microbiota of Groups A and B at days 0 and 21 was conducted (Figure S2). No significant differences were identified in α diversity (p > .05, Kruskal-Wallis test; Figure S2(a)) and β diversity (p > .05, PERMANOVA; Figure S2(b)) between the two groups at any time point. Eight differentially abundant SGBs exhibited significant differences only at day 21 but not at baseline (Figure S2(c); p < .05 in all cases). Notably, Angelakisella sp900547385, Sutterella sp900770335, Faecalibacterium prausnitzii, Faecalibacterium prausnitzii_C, and Faecalibacterium prausnitzii_A were significantly enriched in Group A compared to Group B at day 21 (following the postbiotic treatment), while Metalysinibacillus saudimassiliensis, Blautia_A obeum, and Holdemania filiformis displayed the opposite trend.

3.6. Multivariable association between clinical indicators and subjects’ gut bacteria

Next, we further investigated the relationship between subjects’ gut bacteria and clinical indicators of diarrhea, uncovering several interesting correlations between individual fecal microbes and various clinical parameters (Figure 2(e)). Notably, Faecalibacterium sp900765105 (r = −0.38, p < .01), Fournierella sp002160145 (r = −0.31, p < .05), Pseudoruminococcus massiliensis (r = −0.41, p < .01), and Ruminococcus_C sp000980705 (r = −0.34, p < .01) showed significant negative correlations with the Bristol stool scale. Defecation urgency score displayed a significant positive correlation with Fusobacterium_A sp900555845 (r = 0.26, p < .05), but exhibited a significant negative correlation with Lawsonibacter sp000177015 (r = −0.39, p < .01). Dialister hominis showed a significant positive correlation with defecation frequency (r = 0.48, p < .05). Additionally, Ruminococcus_B gnavus (r = 0.27, p < .05), Succinivibrio sp000431835 (r = 0.26, p < .05), and Bifidobacterium aerophilum (r = 0.41, p < .05) showed significant positive correlations with anxiety.

3.7. Postbiotic effects on the gut phageome

Bacteriophages are viruses that infect bacterial hosts, exerting substantial effects on bacterial population dynamics and diversity across various ecosystems. Moreover, it is becoming increasingly recognized that the composition of the gut phageome can profoundly influence host health.²²

Our comprehensive search for bacteriophages within the dataset yielded a total of 25,224 unique vOTUs. These vOTUs were annotated based on the Metagenomic Gut Virus catalog, assigning 60.93% to known bacteriophage families. Among them, 555 (3.61%) were classified as complete genomes, 3,033 (19.74%) as high-quality genomes, and 11,780 (76.65%) as medium-quality genomes. To assess the impact of postbiotic administration on the overall bacteriophage community, we conducted group-based non-metric multidimensional scaling analyses on the gut phageome. The results revealed non-significant longitudinal changes in the structure of the gut phageome for both Groups A and B (p > .05; Figure 3(a)). These findings suggest that postbiotic administration had no significant effect on the overall bacteriophage community. Furthermore, no significant differences were detected in the α diversity of the gut phageome between Groups A and B (p > .05; Figure 3(b)). This lack of significant differences persisted in the cross-sectional comparison of the gut phageome α diversity between Groups A and B at days 0 and 21 (Figure 3(c)). Therefore, postbiotic administration did not result in any substantial changes in the α diversity of the gut phageome.

Figure 3.

Fecal phageome features of the subjects and their correlation with the bacterial microbiota. (a) Non-metric multidimensional scaling (NMDS) analysis and (b) the Shannon diversity index of the fecal phageome of groups A and B at different time points during the intervention study. (c) Cross-sectional comparison of the Shannon diversity index of the fecal phageome of groups A and B at days 0 and 21. Group a received the postbiotic between days 0 and 21, while group B received the placebo between days 0 and 21. (d) Family-level taxonomic distribution of the phage metagenome. (e) Longitudinal and cross-sectional comparisons of the post-interventional changes in the abundance of fecal Microviridae. (f) A strong positive Spearman’s rank correlation was observed between the Shannon diversity index of the fecal bacterial microbiota and phageome in both groups. (g) Procrustes analysis of the fecal bacterial species-level genome bins (SGBs) and bacteriophages of the groups A and B at different time points revealed a positive cooperativity between the gut bacteria and bacteriophages.

To identify differentially abundant bacteriophages associated with the postbiotic intervention, a comparative analysis was performed on the fecal phageome of Groups A and B at various time points. In total, 12 bacteriophage families were identified in gut phageome of all subjects, including Adenoviridae, Caulimoviridae, Circoviridae, crAss-phage, Herelleviridae, Inoviridae, Metaviridae, Microviridae, Myoviridae, Papillomaviridae, Podoviridae, and Siphoviridae. The overall composition of bacteriophage profiles exhibited a high degree of similarity across different subgroups, with Siphoviridae, Myoviridae, Microviridae, and crAss-phage being the most dominant families (Figure 3(d)). Notably, the abundance of Microviridae significantly increased following the postbiotic intervention, as observed in both cross-sectional and longitudinal analyses (p < .05; Figure 3(e)).

Then, we analyzed the correlation between the fecal bacterial microbiota and phageome, revealing their strong positive correlation in both Group A (r = 0.94, p < .001) and Group B (r = 0.96, p < .00; Figure 3(f)). These findings highlight the interconnectedness and mutual influence between the bacterial and phage components of the gut microbiota. Consistent results were obtained from the Procrustes analysis, indicating a high level of cooperativity between the bacterial and phage microbiota (Figure 3(g)). This analysis suggests that the gut bacterial and bacteriophage communities interact and shape the composition of each other.

3.8. Postbiotic administration increased fecal SCFAs levels

A comprehensive qualitative and quantitative analysis of the fecal SCFAs levels was performed because of the high levels of SCFAs found in the administered postbiotic and the presence of butyrate-producing species among the postbiotic-responsive bacteria in our dataset. Fecal samples were analyzed for acetic acid, propionic acid, butyric acid, hexanoic acid, and pentanoic acid. The results demonstrated a notable increase in all five SCFAs following postbiotic intervention, with butyric acid exhibiting a particularly significant rise (p < .05; Figure 4(a), S3).

Figure 4.

Identification of differential metabolites and metabolite biomarkers of diarrhea. (a) Boxplots show the results of targeted metabolite analysis for butyric acid. P-values were generated by the Wilcoxon rank-sum test. (b) Orthogonal partial least squares-discriminant analysis (OPLS-DA) of fecal metabolomes after the postbiotic intervention. The upper and middle panels show the results of longitudinal comparison between day 0 and day 21 in groups A and B, respectively, while the lower panel shows the intergroup comparison at day 21. Group a received the postbiotic from days 0 to 21 and the placebo from days 35 to 56, while group B underwent the two interventions in the reverse order. Chromatography was performed in both positive and negative ion modes. Heatmaps show (c) the changes in the six significant postbiotic-responsive metabolites in groups A and B and (d) their associations with diarrhea-related clinical indicators. The color scales in (c) represent the metabolite abundance, from more (green) to less (pink); (d) represent the correlation strength, from positive (red) to negative (purple). Asterisks represent significant correlations (*: p < .05; **: p < .01; ***: p < .001). (e) The Venn network plot illustrates the common and unique postbiotic-responsive metabolites identified in longitudinal (intragroup but at different time points) and cross-sectional (intergroup, at day 21) comparisons of pre-/post- or with/without postbiotic intervention. Each dot represents one metabolite. Pink dots represent metabolites exclusively identified in one pairwise comparison. Tan, light sea green, and purple dots represent metabolites common to two pairwise comparisons. Slate gray dots represent metabolites commonly shared by the three pairwise comparisons. (f) Boxplots show the results of targeted metabolite analysis of α-linolenic acid and p-methoxycinnamic acid. P-values were generated by the Wilcoxon rank-sum test. (g) Receiver operating characteristic curves evaluate the predictive accuracy of six different metabolites (left panel) and their combined analysis as a biomarker set using multiple logistic regression analysis (right panel). The area under the curve (AUC) value reflects the discriminatory ability of the model in distinguishing between normal individuals and subjects suffering from chronic diarrhea (AUC <0.7 and above 0.8 represent poor and good discriminatory ability, respectively).

3.9. Postbiotic administration modulated subjects’ fecal metabolome

The gut microbiome and its bacterial metabolites can impact human metabolism and contribute to the development of metabolic disorders. Herein, we employed non-targeted metabolomics to investigate alterations in gut bacterial metabolites following postbiotics treatment, with the application of a proper QC sample to ensure instrumental stability. The metabolomics data were analyzed by principal component analysis, and symbols representing the QC samples were closely clustered on the score plots of both positive and negative ion modes, suggesting a good instrumental and experimental stability, thereby affirming the data reliability (Figure S4(a)).

To identify differential metabolites associated with the postbiotic intervention, we performed orthogonal partial least squares-discriminant analyses (OPLS-DA; Figure 4(b)). The OPLS-DA plots exhibited clear clustering patterns for the time-course analysis (Group A: day 0 versus day 21; Group B: day 35 versus day 56) and the cross-sectional analysis (Group A: day 21 versus Group B: day 21). These clustering patterns suggest that postbiotic administration might have impacted on the gut metabolite composition. The validity of our models was supported by significant results from the permutation test (1000 permutations; p < .001). Moreover, the constructed models were evaluated using the R2 and Q2 values, ranging from 0.752–0.972 and 0.547–0.930, respectively; Table S5). These values indicate an acceptable level of model fitness to the original data and consistency between the original and cross-validation predicted data. Postbiotic intervention-associated differential metabolites were identified by Student’s t-tests and the variable importance in projection (VIP) scores generated by OPLS-DA (cutoff level of p < 0.05 and VIP score > 1; Figure S4(b)). For the time-course analysis, a total of 48 (Table S6) and 43 (Table S7) differential metabolites were identified in Groups A and B, respectively. For the cross-sectional analysis, 24 differential metabolites (11 increased and 13 decreased following the postbiotic intervention; Table S8) were identified at day 21 between Groups A and B. These metabolites became significantly differential only after the postbiotic intervention but not at baseline. Interestingly, six differential metabolites were consistently detected in both groups, which were further analyzed semi-quantitatively, revealing a significant enrichment in α-linolenic acid and p-methoxycinnamic acid, as well as a significant reduction in Phe, capsaicin, theophylline, and piperine after the postbiotic intervention (Figure 4(c), S5). We further explored the relationship between these six differential metabolites and clinical indicators of diarrhea in each group (Figure 4(d)). In Group A, we observed a significant positive correlation between piperine and the Bristol stool scale (r = 0.45, p < .001), while α-linolenic acid (r = −0.33, p < .05) and p-methoxycinnamic acid (r = −0.37, p < .01) were negatively correlated with the Bristol stool scale. Furthermore, theophylline (r = 0.35, p < .01), piperine (r = 0.34, p < .01), and capsaicin (r = 0.28, p < .05) displayed a significant positive correlation with defecation urgency. In Group B, theophylline exhibited a significant positive correlation with both the Bristol stool scale (r = 0.41, p < .05) and defecation frequency (r = 0.42, p < .05). No significant correlation was found between any of the six metabolites and the levels of anxiety, stress, and depression in either group (p > .05).

Notably, both our time-course and cross-sectional analyses revealed significant increases in α-linolenic acid and p-methoxycinnamic acid following the postbiotic intervention, compared to baseline or between groups (p < .05; Figure 4(e)). We further validated their presence and confirmed their post-interventional increases in the fecal samples using LC-MS (p < .05; Figure 4(f)).

3.10. Receiver operating characteristic curve analysis of differential metabolites

To identify potential biomarkers for chronic diarrhea, we constructed receiver operating characteristic curves and calculated the area under the curve to evaluate the predictive accuracy of the six differential metabolites in distinguishing between normal individuals and subjects suffering from diarrhea (Figure 4(g)). The area under the curve values for α-linolenic acid, p-methoxycinnamic acid, Phe, capsaicin, theophylline, and piperine were below 0.7, suggesting poor discrimination. However, a subsequent multiple logistic regression analysis was performed to reevaluate the combined discriminatory ability of these metabolites. Remarkably, the analysis yielded a highly favorable outcome, with an AUC value of 0.822, indicating good discriminatory ability.²³

3.11. Postbiotic administration modulated the fecal AA profile differentially

Since AA levels are potentially associated with gut inflammation and colonic health, and Phe was identified as one of the differential metabolites following postbiotic intake, we further investigated the post-interventional changes in the fecal AA profile, focusing on 22 AAs. Among these, arginine, isoleucine, leucine, and proline were not detectable in the samples. Among the remaining 18 AAs, kynurenine (Kyn), Phe, threonine (Thr), tryptophan (Trp), and aspartic acid exhibited significant changes in Group A after postbiotic intervention (p < .05; Figure 5(a), S6). In Group B, Phe, Try, and valine displayed significant alterations following postbiotic intervention (p < .05; Figure 5(a), S6). Notably, both Groups A and B showed a significant decrease in Phe and Try levels after the postbiotic intervention, which corresponds to the reduced Phe level observed in non-targeted metabolomics analysis. Considering the significant role of Trp and its metabolites, such as 5-hydroxytryptophan (5-HTP) and 5-hydroxytryptamine (5-HT), in gastrointestinal physiology and mood regulation, we also quantified the fecal levels of 5-HTP and 5-HT by using LC-MS. The analysis revealed a significant decline in 5-hydroxytryptophan levels in Group A after the postbiotic intervention (p < .05; Figure 5(b)).

Figure 5.

Changes in fecal amino acid and bile acid levels after postbiotic intervention. Group a received the postbiotic intervention from day 0 to 21, while group B received it from day 35 to 56. The boxplots illustrate changes in (a) significant postbiotic-responsive amino acids; (b) 5-hydroxytryptophan and 5-hydroxytryptamine levels; and (c) significant postbiotic-responsive bile acids. The Wilcoxon rank-sum test was used to evaluate significant differences between time points, with the corresponding p-values shown (cut-off: p < .05).

3.12. Postbiotic administration regulated the gut BAs metabolism

Malabsorption of BAs is common in patients with chronic diarrhea exhibiting functional symptoms. To assess this, we quantified 14 BAs in the fecal samples (Figure 5(c), S7). Our analysis revealed significant changes in dehydrocholic acid and chenodeoxycholic acid (CDCA) in both groups after the postbiotic intervention. Specifically, dehydrocholic acid exhibited a significant increase, while CDCA displayed a significant decrease (p < .05). Among the remaining BAs, only deoxycholic acid showed a significant increase in Group A (p < .05), while taurocholate acid decreased significantly in Group B (p < .05). No significant changes were observed in the other BAs.

4. Discussion

Chronic diarrhea poses a widespread challenge within the general population, negatively impacting patients’ quality of life, productivity, and financial well-being.²⁴ While Probiotics have demonstrated potential in alleviating clinical symptoms associated with chronic diarrhea,²⁵ emerging evidence highlights the diverse clinical benefits of postbiotics in managing this condition. Postbiotics offer a promising therapeutic approach that may serve as a safer alternative to Probiotics.²⁶ However, there is a lack of studies integrating multi-omics analysis to comprehensively understand the relieving effects of specific postbiotics on diarrhea, particularly concerning their influence on host gut microbiota and metabolomic profiles. To bridge this knowledge gap, we conducted a randomized, double-blind, placebo-controlled crossover intervention trial to explore the relieving effects and plausible beneficial mechanisms of postbiotics on chronic diarrhea.

Organic acids, SCFAs, and other bioactive molecules with potential health benefits were detected in ProBio-Eco®. Previous studies have demonstrated that SCFAs can serve as reliable diagnostic biomarkers for diarrhea diseases.²⁷ Propionic acid, butyric acid, and valeric acid have been implicated in regulating intestinal motility through various mechanisms such as stimulating mucosal receptors, acting on the intestinal tract and/or the vagus nerve, or directly modulating colonic smooth muscle function.²⁷ Organic acids have been shown to inhibit the growth of pathogenic intestinal bacteria by impacting the intestinal pH, consequently diminishing the presence of harmful bacterial metabolites.²⁸ Notably, lactic acid, citric acid, and malic acid have been documented for their efficacy in preventing diarrhea.^29,30 In addition, 3-indole acrylic acid can enhance the production of anti-inflammatory interleukin-10 by macrophages and exert its anti-inflammatory effects by suppressing the expression of the cytokines interleukin-1β and interleukin-6 by peripheral blood mononuclear cells.^31,32 The presence of these metabolites in ProBio-Eco® strongly supports its use as a therapeutic option for relieving chronic diarrhea in the context of this study.

We observed intriguing alleviation effects on diarrhea symptoms following postbiotic administration. This was evidenced by significant reductions in defecation frequency, Bristol stool score, and defecation urgency during the postbiotic phase, compared to baseline, in both groups. However, in Group A (postbiotic followed by placebo), during the second intervention phase (placebo), there was also a noteworthy improvement in these parameters compared to the baseline. This improvement can be attributed to the carry-over effects of the postbiotics from the initial phase. A previous study reported the high efficacy of heat-treated Lactobacillus LB in improving chronic diarrhea and its associated clinical symptoms, surpassing the effectiveness of live lactobacilli. These findings align with the results of our current study.³³ Similarly, the administration of inactivated Lactobacillus LB along with fermented culture medium resulted in a significant reduction in the number of weekly stools and a notable improvement in symptoms such as abdominal pain, bloating, and overall quality of life in patients with IBS.³⁴ Furthermore, well-controlled studies have demonstrated the efficacy of heat-treated Lacticaseibacillus paracasei in reducing the frequency of diarrhea episodes and effectively preventing diarrhea when compared to a placebo.^13,35 Additionally, postbiotics have shown effectiveness in alleviating anxiety in individuals with diarrhea, although there were no notable improvements observed in terms of stress or depression. In a 24-week double-blind, placebo-controlled, parallel-group clinical trial involving 41 men and 19 women, participants consumed daily tablets containing heat-inactivated Lactobacillus gasseri CP2305 or placebo. Mental and physical states were assessed using questionnaires, revealing that the postbiotic formulation effectively reduced anxiety and sleep disturbance compared to the placebo.³⁶ The findings of this study contribute to the expanding body of literature advocating the beneficial use of postbiotics in relieving symptoms related to diarrhea.

The role of gut microbiota in individuals with bowel dysfunction has gained significant attention in the past decade. Irrespective of the specific etiological factors involved, disruptions in intestinal transit have consistently been associated with alterations in gut dysbiosis.³⁷ Microbial communities, including intestinal bacteria and bacteriophages, are increasingly recognized as an intricate “organ” with profound implications for human health and disease. Research has revealed that postbiotics can have direct and indirect influences on the composition and functionality of the human gut microbiota, thereby enhancing their potential therapeutic benefits.³⁸ We thus conducted an extensive fecal metagenomics analysis to deepen our understanding of how the gut microbiota contributes to the relief of diarrhea following postbiotic administration. The analysis of α and β diversity in the gut bacteriome and phageome of the subjects did not reveal any significant within- and between-group differences in their structure or composition. However, our Procrustes analysis results did indicate a notable correlation between the richness and diversity of the gut microbiome and those of the phageome.

To identify subtle yet meaningful changes in the gut microbiome that may play a part in symptom improvement, we focused on finding the specific postbiotic-responsive SGBs. Our analysis revealed several SGBs that were substantially enriched only following postbiotic intervention but not at baseline. These SGBs include Faecalibacterium sp900765105, Faecalibacterium prausnitzii, Faecalibacterium prausnitzii_C, Faecalibacterium prausnitzii_A, Ruminococcus_C sp000980705, and Dysosmobacter welbionis. Faecalibacterium has been reported to have a negative correlation with the severity of IBS symptoms in humans.^39,40 Faecalibacterium prausnitzii is recognized as one of the primary producers of butyrate in the intestines. Our fecal metabolomics analysis consistently detected a significant increase in butyrate after postbiotic intervention. Butyrate plays a crucial role in gut physiology and overall host well-being.⁴¹ It serves as the primary energy source for colonocytes and is an anti-inflammatory metabolite that protects against colorectal cancer and inflammatory bowel disease.^42,43 Oral administration of live cultures of Faecalibacterium prausnitzii reduced the incidence of severe diarrhea and related mortality rates.⁴⁴ Ruminococcus bromii, a member of the Ruminococcus genus, is considered a keystone species in the degradation of non-digestible carbohydrates, including resistant starch, in the human colon, and enhancing resistant starch fermentation can help reverse infectious diarrhea.⁴⁵ Dysosmobacter welbionis, a recently discovered human commensal bacterium, has been found to prevent diet-induced obesity and metabolic disorders in mice.⁴⁶ However, further investigation is needed to determine its role in diarrhea. Moreover, our data found significant negative correlations between the Bristol stool scale score with the abundance of Faecalibacterium sp900765105 and Ruminococcus_C sp000980705, suggesting that the enrichment of these taxa in the gut is associated with a reduction in the severity of diarrhea. On the other hand, we observed a reduction in certain potentially harmful gut bacteria following postbiotic intervention, such as Megamonas funiformis, Succinivibrio sp000431835, Fusobacterium_A necrogenes, and Fusobacterium_A sp900555845. Megamonas funiformis is a well-known opportunistic pathogen associated with inflammatory bowel disease⁴⁷ and colorectal cancer.⁴⁸ Succinivibrio has been found to be associated with gastrointestinal dysfunction, including diarrhea and abdominal pain. Notably, an increase in intestinal succinate production, primarily by the Succinivibrio genus, has been associated with antibiotic-induced diarrhea.⁴⁹ Furthermore, we observed a positive correlation between Fusobacterium_A sp900555845 and defecation urgency.

In addition to changes in the fecal bacteriome, our study revealed a significant increase in Microviridae abundance following the postbiotic intervention. Microviridae and Caudovirales are important components of a healthy human virome, and diminished levels of these phage taxa were observed in infection-related diarrheal children compared to healthy controls.⁵⁰ Our results align with a previous work reporting notable alterations in the intestinal bacteriomes and viromes of patients with recurrent Clostridioides difficile infection. It was reported that a healthier gut bacteriome and phageome were restored cooperatively through fecal microbiota transplantation, characterized by an increase in Microviridae and a decrease in Proteobacteria.⁵¹ These findings suggest that Microviridae may play a regulatory role in reshaping the intestinal microbiota. Then, we performed a comprehensive fecal metabolomics analysis to identify gut metabolites related to postbiotic intervention. Interestingly, six differential metabolites were identified in both Groups A and B. The postbiotic intervention resulted in marked reductions in Phe, capsaicin, theophylline, and piperine, while substantial enrichments of α-linolenic acid and p-methoxycinnamic acid. It is worth noting that some of these differential metabolites have more direct links with diarrhea. For example, Phe is a lipophilic and aromatic amino acid. The fecal extracts of mice with IBS showed higher levels of Phe compared to those of healthy control mice. This dysregulation in Phe metabolism is associated with the energy metabolism disorder induced by stress in IBS mice, which may contribute to the development of diarrhea.⁵² Additionally, Phe could increase membrane permeability, contributing to mucosal barrier dysfunction and the pathophysiology of IBS.^53,54 Another study demonstrated that Phe-producing bacteria could enhance the production of colonic aromatic trace amines, such as phenylethylamine and tryptamine, and thus 5-HT, stimulating gastrointestinal transit and aggravating diarrheal IBS.⁵⁵ Moreover, high levels of capsaicin,⁵⁶ theophylline,⁵⁷ and piperine⁵⁸ have been reported to be associated with the onset of diarrhea. Thus, the significant decreases in these fecal metabolites following postbiotic intervention may contribute to the relief of diarrhea symptoms. The other two fecal differential metabolites appear to have a less direct link with diarrhea, but they generally play beneficial roles in promoting colonic health. Dietary supplementation with α-linolenic has been shown to alleviate colonic inflammation in experimental colitis rats.⁵⁹ It also promotes the proliferation of beneficial Probiotics such as lactobacilli and bifidobacteria while inhibiting potentially harmful ones like enterococci and Escherichia coli .⁶⁰ On the other hand, p-methoxycinnamic has antidiabetic properties and is utilized as a functional food ingredient to prevent and improve various chronic diseases. It also exerts anticancer, antimicrobial, hepato-, and neuroprotective effects.⁶¹ Notably, both metabolites were significantly enriched in the postbiotic group in the cross-sectional comparison.

As Phe was identified as one of the significantly diminished metabolites in our non-targeted metabolomics analysis following postbiotic intervention in both Groups A and B, strongly reflecting modulation of AA metabolism might be involved in the improvement of diarrheal symptoms. Further metabolomics analysis was implemented to monitor post-interventional changes in the fecal AA profile. Particularly, the two aromatic AAs, Phe and Trp, along with their intermediate metabolites, Kyn and tyrosine, play central roles in inflammation-induced metabolism and are important immune- and neurometabolic biomarkers that reflect the host body and gut health status.⁶² Interestingly, besides Phe, the fecal Trp levels also showed significant reduction following postbiotics intervention in both Groups A and B. Trp metabolism has been found to be linked to the severity of inflammatory bowel disease,⁶³ and the fecal samples from patients suffering from diarrheal IBS have been found to have elevated Trp levels.⁶⁴ We observed a significant reduction in the level of 5-HTP in Group A, suggesting that Trp was involved in the Trp-5-HT pathway. Furthermore, the fecal levels of Kyn and Thr were diminished after postbiotics intervention in Group A. This finding is consistent with a previous work reporting the observation of significantly higher levels of Thr, Trp, and Phe in fecal samples from patients suffering from diarrhea-predominant IBS compared to healthy controls.⁶⁵ As Kyn can induce oxidative stress, anxiety, and other neurotoxic effects,⁶⁶ the reduced fecal Kyn level observed in our study may have contributed to the mitigated anxiety level following the postbiotic intervention. It is worth noting that a substantial portion of human Trp intake is typically metabolized through the Kyn pathway, which plays a pivotal role in Trp conversion.⁶⁷ This observation suggests that postbiotic intake may affect Trp metabolism via the Trp-Kyn pathway and alleviate symptoms of diarrhea.⁶⁸

The gut microbiota plays a crucial role in BA biotransformation, affecting the size and composition of the BA pool. Conversely, an imbalance in BAs can result in an excessive influx of BAs into the colon, disrupting electrolyte balance and intestinal peristalsis, thereby causing diarrhea. This study observed a significant increase in CDCA, one of the most abundant primary BAs in humans, following postbiotic intervention. Clinical investigations have shown that excessive production or incomplete absorption of primary BAs in the intestinal tract may cause abdominal pain, heightened colonic motility, and fluid secretion in patients.^69,70 Furthermore, individuals with diarrhea-predominant IBS have been found to exhibit elevated fecal levels of primary BAs, especially CDCA, compared to healthy controls.⁷¹ Moreover, IBS-D patients with higher fecal BA excretion exhibit a greater proportion of CDCA in their stools compared to those with lower BAs excretion.⁷² In conclusion, the results of our metagenomics and metabolomics analyses provide strong evidence supporting the idea that the alleviation of diarrhea symptoms through postbiotics is intricately linked to the observed changes in the intestinal microbiota and metabolomic profiles (Figure 6).

Figure 6.

Proposed role of postbiotics-associated host gut microbiota/metabolomic changes in relieving diarrhea. The gut microbiome produces short-chain fatty acids, such as butyrate, directly stimulating trp hydroxylase 1 and resulting in the synthesis and secretion of 5-HT by intestinal enterochromaffin cells. 5-HT interacts with receptors from neurons in the enteric nervous system to modulate gastrointestinal (GI) motility. The vagus nerve senses 5-HT and connects the GI tract to the nucleus of the solitary tract and the dorsal raphe nucleus, where most 5-HT neurons reside. These areas then interact with brain networks that regulate emotions and mood. Trp and Kyn can cross the blood-brain barrier, with Kyn potentially inducing oxidative stress, anxiety, and other neurotoxic effects. Butyrate has also been associated with enhanced neuronal excitability in the enteric nervous system and contractile responses of intestinal smooth muscle. Excessive production of CDCA, phenylalanine, piperine, theophylline, and capsaicin in the intestinal tract may stimulate GI and colonic motility. Dysosmobacter welbionis, p-methoxycinnamic acid, and α-linolenic acid appear to have a less direct link with diarrhea, but they generally play beneficial roles in promoting colonic health. CDCA: chenodeoxycholic acid; 5-HT: 5-hydroxytryptamine; 5-HTP: 5-hydroxytryptophan; Kyn: kynurenine; try: tryptophan.

This study had several strengths. Firstly, the crossover design allowed each participant to serve as their own control. Participants received both the active and placebo treatments in a randomized order, mitigating the impact of individual differences and enhancing the precision and validity of the results. Secondly, the study demonstrated that the participants did not experience any adverse effects from the postbiotic intervention. However, it is important to acknowledge certain the limitations of this study. The 14-day wash-out period between intervention phases may have been relatively short, potentially impacting the establishment of uniform baseline data. In addition, the study was partially conducted during the coronavirus outbreak, which limited the duration, sample size, and diversity of the trial. Consequently, the number of participants was limited, precluding the inclusion of a separate healthy control group. This study primarily focused on the young adult cohort without preexisting chronic diarrhea or associated medications/clinical treatments. While chronic diarrhea is more commonly observed in elderly individuals, it is often complicated by other factors, such as concurrent medications or clinical conditions, which can influence study outcomes in postbiotic or probiotic research. This deliberate choice to investigate a specific population seeking alternative approaches for symptom improvement may limit the generalizability of the conclusions to other age groups, such as younger children or the elderly. Furthermore, while meticulous measures were taken to document the participants’ daily eating patterns, their food intake was not strictly controlled. Future studies should strive to encompass a broader range of participants and consider additional contributing factors to provide a more comprehensive understanding of the effects of postbiotic interventions on diarrhea across different age groups and populations. Although this study identified some differential fecal microbes and metabolites following the postbiotic intervention, the exact mechanisms by which postbiotics mitigated diarrhea remain unclear. While 5-HT has been thought to play a crucial role in regulating gut motility,⁷³ the current results did not detect significant post-interventional changes in fecal 5-HT levels. The anti-diarrheal effects were likely due to other mechanisms, such as enhancing the gut mucosal barrier and regulating gut mucosal immunity and inflammation,⁷⁴ as indicated by the observed gut metagenomics and metabolomics changes. Further validation of the current findings would require in-depth mechanistic studies, such as the use of diarrheal germ-free animal models and fecal microbiota transplantation.

5. Conclusion

In conclusion, this study underscores the therapeutic potential of postbiotics in managing chronic diarrhea. The observed benefits include significant improvements in bowel frequency, the Bristol stool scale score, and defecation urgency. Furthermore, postbiotics play a pivotal role in modulating the gut microbiota and metabolite profiles. These compelling results highlight the promising prospects of postbiotics as an efficacious treatment for chronic diarrhea, fostering the development and utilization of functional postbiotic products in clinical practice.

Funding Statement

Special Funds for International Science and Technology Cooperation of China [2018YFE0123500]; Science and Technology Major Projects of Inner Mongolia Autonomous Region [2021ZD0014].

Disclosure statement

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

Author contributions

Shuai Guo: conceptualization, data curation, formal analysis, writing – original draft preparation. Teng Ma: investigation, methodology. Lai-Yu Kwok: Writing – review & editing, Resources. Keyu Quan: funding acquisition, writing – review & editing. Bohai Li: conceptualization, data curation. Huan Wang: conceptualization. Heping Zhang: conceptualization, Methodology. Bilige Menghe, Yongfu Chen: supervision, funding acquisition, writing – review & editing.

Data availability statement

This study received approval from the Ethics Board of Inner Mongolia People’s Hospital (No. 20210004) and adhered to the principles outlined in the Declaration of Helsinki. The study was registered on the Chinese Clinical Trial Registry (http://www.chictr.org.cn) with the registration number ChiCTR2100054376. Sequencing data are available in China National GeneBank (CNGB; https://db.cngb.org/cnsa/; accession number: CNP0005157). Prior to participating in the trial, all participants provided informed consent.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19490976.2024.2395092