Efficacy in bowel movement and change of gut microbiota on adult functional constipation patients treated with probiotics-containing products: a systematic review and meta-analysis

Abstract

Objectives

This study aimed to pool the efficacy in bowel movement and explore the change of gut microbiota on adult functional constipated patients after probiotics-containing products treatment.

Design

Systematic review and meta-analysis.

Data sources

PubMed, Cochrane Library for published studies and ClinicalTrials.gov for ‘grey’ researches were independently investigated for randomised controlled trials up to November 2022.

Eligibility criteria, data extraction and synthesis

The intervention was probiotics-containing product, either probiotics or synbiotics, while the control was placebo. The risk of bias was conducted. The efficacy in bowel movement was indicated by stool frequency, stool consistency and Patient Assessment of Constipation Symptom (PAC-SYM), while the change of gut microbiota was reviewed through α diversity, β diversity, change/difference in relative abundance and so on. The subgroup analysis, sensitivity analysis and random-effect meta-regression were conducted to explore the heterogeneity. The Grading of Recommendations Assessment Development and Evaluation was conducted to grade the quality of evidence.

Results

17 studies, comprising 1256 participants, were included with perfect agreements between two researchers (kappa statistic=0.797). Compared with placebo, probiotics-containing products significantly increased the stool frequency (weighted mean difference, WMD 0.93, 95% CI 0.47 to 1.40, p=0.000, I^²=84.5%, ‘low’), improved the stool consistency (WMD 0.38, 95% CI 0.05 to 0.70, p=0.023, I²=81.6%, ‘very low’) and reduced the PAC-SYM (WMD −0.28, 95% CI: −0.45 to −0.11, p=0.001, I²=55.7%, ‘very low’). In subgroup analysis, synbiotics was superior to probiotics to increase stool frequency. Probiotics-containing products might not affect α or β diversity, but would increase the relative abundance of specific strain.

Conclusions

Probiotics-containing products, significantly increased stool frequency, improved stool consistency, and alleviated functional constipation symptoms. They increased the relative abundance of specific strain. More high-quality head-to-head randomised controlled trials are needed.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• We only included randomised controlled trials that recruited adult functional constipation patients without defecation-impairing comorbidities and reported the efficacy in bowel movement or the change of gut microbiota.

• We pooled the efficacy in bowel movement including stool frequency, stool consistency and Patient Assessment of Constipation Symptom, and qualitatively summarised the change of gut microbiota including α diversity, β diversity, change/difference in relative abundance, etc.

• We explored sources of heterogeneity through subgroup analysis, sensitivity analysis and random-effect meta-regression, and we graded the quality of pooled results based on Grading of Recommendations Assessment Development and Evaluation methodology.

• In eligible studies, the change of gut microbiota was researched through the comparison between pretreatment and post-treatment stage in the intervention group, or between the intervention group and the control group in post-treatment stage.

Introduction

Functional constipation (FC) is a common functional gastrointestinal disorder around the world, with a series of symptoms including reduced bowel movements (<3 times/week), hard/lumpy stool, incomplete evacuation, defecation straining, anorectal obstruction or blockage, and manual manoeuvre to defecation for at least 6 months.^{1 2} Approximately, the global average prevalence of FC is 10.1% based on the Rome IV criteria.³ Patients with FC often suffer impaired quality of life and insufferable financial burden.⁴ Osmotic laxatives, stimulative laxatives, prokinetics and secretagogues have constituted the main proportion of constipation-relieving drugs. However, long-term use of osmotic laxatives, such as lactulose, often presents impaired efficacy, while stimulative laxatives often cause persistent side effects, especially melanosis coli. And the high cost of new prokinetics and secretagogues makes them unaffordable for many patients.⁵

Gut microbiota, which comprises a complex community of micro-organisms inhabiting the gastrointestinal tract, including bacteria, viruses, fungi and other micro-organisms, plays a crucial role in gut motility.⁶ Several studies have found that gut microbiota could participate in the development of FC by regulating the enteric nervous system, immune system and intestinal microenvironment through metabolites including bile acids, short-chain fatty acids, 5-hydroxytryptamine and methane.⁷ Moreover, characteristic of lower abundance of Bifidobacterium and Lactobacilli, the gut flora in FC patients is significantly distinctive from that in healthy controls.⁸ Probiotics are ‘live micro-organisms which when administered in adequate amounts confer a health benefit on the host’,⁹ prebiotics are seen as substrates for probiotics inhabiting the host’s gastrointestinal tract,¹⁰ while synbiotics are defined as ‘a mixture comprising live micro-organisms and substrate(s) selectively used by host micro-organisms that confers a health benefit on the host’.^{7 11} Recently, lots of researches have focused on the efficacy of probiotics-containing products on patients. But a wide variety of probiotic-containing products are available, including single-strain formulations, multistrains formulations and some with added active ingredients such as prebiotics. These products vary in terms of strains, dose, dosage forms and even storage methods. These inconsistencies might contribute to the contradiction in the efficacy of probiotic-containing products and the confusion regarding changes in gut microbiota among many studies.

A meta-analysis by Miller et al found that probiotics significantly reduced gastrointestinal transit time, suggesting the promising future of probiotics to promote intestinal motility and increase the frequency of bowel movements, but this paper included many small-sample randomised controlled trials (RCTs) whose participants were not FC patients.¹² Another meta-analysis showed that probiotics significantly increased stool frequency and softened stool consistency, and the subgroup analysis indicated that multistrains products were more effective than single-strain products.¹³ But the paper included some researches which recruited participants diagnosed with constipation-dominated irritable bowel syndrome (IBS-C), increasing the clinical inconsistency and then weakening the reliability of the conclusions. Furthermore, these studies are very old and an update in this field is needed to re-evaluate the efficacy of probiotics-containing products, as also recommended by the Cochrane Handbook.¹⁴ As mentioned above, multistrains probiotics demonstrated superior efficacy compared with single-strain ones; therefore, it was worth investigating whether synbiotics conferred even greater efficacy than probiotics alone. And there is still a lack of systematic review which qualitatively summarise the impact of probiotics-containing products on microbiota in FC patients.

Hence, we conducted this systematic review and meta-analysis to quantitatively calculate the efficacy of probiotics-containing products, including probiotics or synbiotics, on FC comprising stool frequency, stool consistency and Patient Assessment of Constipation Symptom (PAC-SYM), and to qualitatively explore the change of gut microbiota. Also, we compared the efficacy between probiotics and synbiotics, and investigated the impact of covariates on efficacy

Methods

Search strategy

The objective of this study was to systematically review RCTs that calculated the efficacy and described the change of gut microbiota after probiotics-containing products treatment. Two researchers (FD and MH) independently investigated electronic databases consisting of PubMed, Cochrane Library for published studies and ClinicalTrials.gov for ‘grey’ researches up to November 2022 by using search strategies as follows: (constipation) and (probiotics or synbiotics or microbiome) (online supplemental table 1). This meta-analysis had been registered in PROSPERO (registration ID: CRD42022381368).

Eligibility and exclusion criteria

We mainly identified parallel or cross-over RCTs that met the following eligibility criteria: (1) Adult participants diagnosed by FC through explicit criteria (eg: Rome criteria); (2) Intervention was probiotics-containing product including probiotics or synbiotics; (3) Control was placebo and (4) Outcomes included stool frequency, stool consistency, PAC-SYM and change of gut microbiota.

The following exclusion criteria were applied to this study: (1) irrelevant studies or duplicate literature and (2) conference abstracts, reviews, articles, letters, book chapters or case report(s).

Study selection and data extraction

Two researchers (FD and MH) removed duplicates through EndNote V.20, and then independently screened the titles and abstracts for eligibility. References from relevant systematic reviews and all eligible researches were also screened. Disagreements were resolved by discussion between these two researchers (FD and MH), or judged by the third researcher (YD). Two researchers (FD and MH) extracted data independently.

Two researchers (FD and MH) extracted items included basic information (author, publication year, country, diagnosis criteria, participants number, age, sex, body mass index (BMI), species and strains of probiotics, additional active prebiotics and duration) in Excel. Disagreements between these two researchers (FD and MH) were resolved by the third researcher (YD). Kappa value was used to assess the degree of agreements. The impact of probiotics-containing products on gut microbiota in these researches would be extracted to conduct qualitative analysis. Two researchers (FD and MH) extracted continual data including mean, SD and total number. Missing data would be handled by contacting with study investigators or acquired from the figure in the paper (through https://apps.automeris.io/wpd/).

Quality assessment and outcomes synthesis/summary

RevMan V.5.4 was used to assess the risk of bias of included studies through the Cochrane Collaboration’s tool, including randomisation, treatment allocation, blinding, incomplete outcome data and selective reporting. Disagreements would be resolved by discussion or judged by the third researcher (YD).

Stata V.17 was used for data synthesis. Weighted mean difference (WMD) or standard mean difference (SMD) was calculated for outcome data, including stool frequency, stool consistency and PAC-SYM. A random or fixed effect model was applied based on the meta-analysis heterogeneity. Effect size was presented as WMD or SMD with 95% CI, a p<0.05 indicates statistical significance. Heterogeneity was calculated quantitatively mainly through I² statistics. I²<25%, 25%≤I²<75% and I²≥75% indicated low, moderate and high heterogeneity, respectively. When I²>25%, possible sources for heterogeneity would be investigated through the following pathway: subgroup analysis, sensitivity analysis, random-effect meta-regression. A funnel plot would be conducted to detect the small-sample effect when more than 10 studies were included.

Grading of the pooled results was conducted according to Grading of Recommendations Assessment Development and Evaluation (GRADE) methodology through GRADE profile software. Risk of bias, inconsistency, indirectness, imprecision and potential publication bias were comprehensively assessed.¹⁵

Change of gut microbiota was qualitatively summarised through the following pathway: α diversity, β diversity, change/difference in relative abundance, etc. Change in relative abundance referred to the change of relative abundance of specific microflora between pretreatment and post-treatment stage in one group (mainly in the intervention group), while difference in relative abundance referred to the difference in specific flora between the intervention group and the control group after treatment.

Patient and public involvement

No patients were involved.

Results

Study characteristics

In this meta-analysis, as depicted in the flow diagram (figure 1), the researchers identified 1285 studies from electronic databases and 180 registration trials from ClinicalTrials.gov, and then screened studies after removing duplicates and assessed studies for eligibility. Finally, a total of 17 studies, consisting of 1256 participants, were included with perfect agreements between two researchers (kappa statistic=0.797).^16–32 Among these researches 11 were probiotics RCTs, the others were synbiotics RCTs. Probiotics included single-strain products and multiple strain products, while additional active prebiotics included xylooligosaccharide fructooligosaccharide and psyllium husks (table 1 and (online supplemental table 4)). The basic information about these studies was listed in table 1. The risk of bias of included researches was shown in online supplemental figure S1. All of these studies were randomised, double-blind, parallel controlled trials and the drop-out rate of each study was <20%. Two studies were only included for qualitative analysis about the change of gut microbiota.^{22 23}

Figure 1.

PRISMA flow diagram. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Table 1.

Characteristics of RCTs included in the meta-analysis

Study Country	Design criteria	N (pro/con) female%(pro/con)	Age (pro/con) BMI (pro/con)	Microbiota genus Dose/(109/CFU/day)	Additional active ingredient(s)	Duration/week(s)
Probiotics
Zhang, 2022 China32	R.D.P. Ro. IV	59/58 72.9%/60.3%	45.80 (12.03)/43.07 (9.36) na	L. 20	–	1
Wang, 2022 China30	R.D.P. Ro. IV	53/50 79.2%/92.0%	51.7 (12.8)/47.6 (13.6) 23.1 (2.64)/23.9 (2.48)	B. 2	–	4
Ito, 2022 Japan20	R.D.P. constipation	23/21 81.0%/94.1%	36.6 (11.3)/39.4 (8.9) 21.7 (3.4)/21.8 (3.6)	B. 10	–	2
Kang, 2021 Korea21	R.D.P. Ro. III	40/40 87.5%/87.5%	44.4 (2.2)/45.3 (1.8) 22.2 (0.4)/23.2 (0.5)	Ba. 1	–	8
Madempudi, 2020 India24	R.D.P. Ro. III	50/50 32%/50%	42.54 (12.16)/45.30 (11.33) na	Ba. 2	–	4
Dimidi, 2019 UK17	R.D.P. Ro. III	37/38 92%/92%	35(12)/31(10) 23.6 (3.1)/22.8 (2.6)	B. 1.5×1010 CFU/day or ≥8×109 CFU/day	–	4
Tanaka, 2015 Japan29	R.D.P. <5 t/w	18/20 100%/100%	40.8 (7.6)/42.8 (7.7) na	B. 15	–	8
Ojetti, 2014 Italy28	R.D.P. Ro. III	20/20 55%/65%	34.5 (16)/36.7 (14) na	L. 0.2	–	4
Mazlyn, 2013 Malaysia27	R.D.P. Ro. II	47/43 89.4%/83.7%	31.8 (9.4)/31.7 (9.4) 22.3 (3.4)/22.9 (3.2)	L. ≥30	–	4
Kondo, 2013 Japan23	R.D.P. constipation	34/32 76.5%/71.9%	85.8 (7.3)/82.7 (9.5) na	B. 50	–	16
Yang, 2008 China31	R.D.P. < 5 t/w	63/63 100%/100%	46.4 (9.8)/46.4 (6.7) na	B., S. and L. 12.5, 1.2	–	2
Synbiotics
Kim, 2021 Korea22	R.D.P. Ro. III	20/10 95%/100%	50 (9.14)/40.70 (12.76) 23.06 (3.62)/22.02 (2.10)	L., B. and La. na	Xylooligosaccharide	4
Martoni, 2019 Canada26	R.D.P. Ro. III	48/46 70.8%/80.4%	44.0 (11.3)/42.9 (13.8) 26.0 (3.6)/26.6 (4.5)	L. and B. 15	FOS	4
Cudmore, 2017 Ireland16	R.D.P. Ro. III	35/34 94.3%/91.2%	45.09 (15.09)/42.21 (12.88) na	L. and B. 1.2	psyllium husk, inulin	4
Ding, 2016 China18	R.D.P. Ro. III	48/45 62.5%/64.4%	47.2 (10.7)/48.3 (11.3) 22.6 (1.1)/22.8 (1.1)	En., B. and L. 0.1	pectin	12
Magro, 2014 Brazil25	R.D.P. Agachan	26/21 92.3%/90.5%	31.5 (7.1)/32.7 (7.3) 28.2 (5.9)/26.8 (4.8)	L. and B. 2	polydextrose	2
Fateh, 2011 Iran19	R.D.P. Ro. III	31/29 0%/0%	23(4)/22.62 (3.96) na	B., L. and S. 0.2	FOS	4

B, Bifidobacterium; Ba, Bacillus; BMI, body mass index; En, Enterococci; FOS, fructooligosaccharide; L, Lactobacillus; La, Lactococcus ; na, not available; pro/con, probiotics-containing product/control; RCT, randomised controlled trial; R.D.P, randomised double-blind parallel; Ro, Rome; S, Streptococcus; t/w, time(s)/week.

Efficacy

Stool frequency

Fifteen studies were included in the pool-analysis for stool frequency.16–21 24–32 Among these studies, two studies measured stool frequency by the criteria called ‘CSBM’^{17 26} and one by ‘SBM’,³⁰ while the remaining didn’t describe specific measurement methodology of stool defecation. However, the pooled result of baseline stool frequency of included studies, that presented non-significant difference between the intervention group and control group (WMD −0.02, 95% CI −0.10 to 0.05, p=0.523) with low heterogeneity (I²=0), indicated the reasonability to pool these using ‘WMD’ (shown in online supplemental figure S2A). Overall, the probiotics-containing products significantly increased the stool frequency by 0.93 time per week compare with placebo (WMD 0.93, 95% CI 0.47 to 1.40, p=0.000) with high heterogeneity (I²=84.5%), as shown in figure 2A. The evidence of small-sample effect was not found (Egger’s test: p=0.554) as shown in online supplemental figure S2B. By subgroup analysis, we found that both probiotics and synbiotics significantly improved bowel movements, while synbiotics (WMD 1.25, 95% CI 0.59 to 1.90, p=0.000) was superior to probiotics (WMD 0.83, 95% CI 0.26 to 1.41, p=0.005) on the promotion of stool frequency (shown in figure 2A). In addition, when compared with multistains subgroup (WMD 1.38, 95% CI 0.94 to 1.82, p=0.000), Lactobacillus subgroup (WMD 0.74, 95% CI −0.17 to 1.65, p=0.111) and Bifidobacterium subgroup (WMD 0.17, 95% CI −0.25 to 0.58, p=0.431), Bacillus subgroup (WMD 1.67, 95% CI −0.65 to 3.99, p=0.159) might had the best medical response, although it was non-significant (shown in online supplemental figure S2C). Through the sensitivity analysis, the removal of each research from the meta-analysis resulted in the WMDs ranging from 0.89 to 1.16 (all p<0.05, (online supplemental figure S2D). Through the random-effect meta-regression analysis, we found that the proportion of gender (slope −2.218, 95% CI −4.05 to −0.385, p=0.021) significantly influenced the pooled efficacy, indicating that the increase in stool frequency became less significant as the proportion of females increased, while age (from 23 to 51.7 years old), the duration of treatment (from 1 to 8 weeks), geographical position(Asian, Europe and America), diagnostic criteria (Rome III, IV, etc), dose of strains (from 0.1 to 30×10⁹ (colony-forming units) CFU/day) had non-significant influence (shown in online supplemental table 5).

Figure 2.

Pooled estimates of the efficacy of probiotics-containing products on FC. (A) Forest plot for data pooling of stool frequency, (B) Forest plot for data pooling of stool consistency, (C) Forest plot for data pooling of PAC-SYM. FC, functional constipation; PAC-SYM, Patient Assessment of Constipation Symptom.

Stool consistency

Six studies were included in the meta-analysis for stool consistency measured by Bristol Stool Scale.17 18 24 26 27 30 Two studies measured stool consistency in three grades and thus were excluded from the quantitative analysis.^{31 32} Overall, probiotics-containing products significantly improved the stool consistency by 0.38 compared with placebo (WMD 0.38, 95% CI 0.05 to 0.70, p=0.023) with high heterogeneity (I²=81.6%), as shown in figure 2B. In synbiotics versus probiotics subgroup, synbiotics subgroup (WMD 0.52, 95% CI −0.63 to 1.67, p=0.377) might be more effective than probiotics subgroup (WMD 0.30, 95% CI 0.03 to o.57, p=0.028), though there was no significance in the synbiotics subgroup (online supplemental figure S3A). In genus subgroup, multistains subgroup (WMD 0.52, 95% CI: −0.63 to 1.67, p=0.377) showed the most remarkable but non-significant efficacy as shown in online supplemental figure S3B. Through the sensitivity analysis, the removal of each research from the meta-analysis resulted in the WMDs ranging from 0.24 to 0.44 (all p<0.05, (online supplemental figure S3C). Through the random-effect meta-regression analysis, we found that age (from 31.8 to 51.7 years old), proportion of gender (from 41% to 92%), geographical position (Asian, Europe and America), diagnostic criteria (Rome III and IV), dose of strains (from 0.1 to 30×10⁹ CFU/day) had non-significant influence (shown in online supplemental table 5).

Patient Assessment of Constipation Symptom

Five studies were included in the meta-analysis for PAC-SYM.17–19 26 30 Overall, probiotics-containing products significantly reduced the PAC-SYM by 0.28 compared with placebo (WMD −0.28, 95% CI −0.45 to −0.11, p=0.001) with moderate heterogeneity (I²=55.7%), as shown in figure 2C. In synbiotics versus probiotics subgroup, synbiotics subgroup (WMD −0.35, 95% CI −0.52 to −0.17, p=0.000) was more effective than probiotics subgroup (WMD −0.18, 95% CI −0.54 to 0.18, p=0.362) (online supplemental figure S4A). In genus subgroup, multistains subgroup aslo showed better remarkable efficacy than Bifidobacterium subgroup as shown in online supplemental figure S4B. Through the sensitivity analysis, the removal of each research from the meta-analysis resulted in the WMDs ranging from −0.35 to −0.21 (all p<0.05, (online supplemental figure S4C). Through the random-effect meta-regression analysis, we found that age (from 23 to 51.7 years old), proportion of gender, geographical position (Asian, Europe and America), diagnostic criteria (Rome III and IV), dose of strains (from 0.1 to 15×10⁹ CFU/day) had non-significant influence (shown in (online supplemental table 5).

GRADE summary of evidence

According to GRADE criteria, as seen in figure 3, the quality of evidence of efficacy of probiotics-containing products on FC were ‘low’ (for stool frequency), ‘very low’ (for stool consistency) and ‘very low’ (for PAC-SYM), respectively. Generally, the inconsistency caused by moderate or high heterogeneity and the imprecision caused by limited trails numbers were responsible for the degradation.

Figure 3.

GRADE summary of pooled results. FC, functional constipation; GRADE, Grading of Recommendations Assessment Development and Evaluation; PAC-SYM, Patient Assessment of Constipation Symptom.

Change of gut microbiota

Eight studies were included in the qualitative review, as seen in table 2, to explore the influence of probiotics-containing products on gut microbiota.17 21–23 26 29 30 32 The sequencing methods included 16S rRNA sequencing,21 22 26 30 32 whole genome sequencing,³² metagenomic analysis²¹ and quantitative PCR (qPCR).^{17 23 29} The analysis methodology included α diversity analysis,21 22 26 30 β diversity analysis,21 26 30 32 analysis of relative abundance,17 21 23 26 29 30 analysis of Clusters of Orthologous Groups,^{26 32} LEfSe analysis²⁶ and analysis of clinical factor correlation.^{22 26}

Table 2.

Change of gut microbiota

Study	Sequencing	Active ingredient(s)	Change of gut microbiota
Zhang 202232	16 s rRNA sequencing and whole genome sequencing	Species and strains: L. plantarum Lp3a	β-diversity: NS between pre-in and post-in probiotics group. COG: secondary bile acid biosynthesis (ko00121, 4 u).
Wang 202230	16 s rRNA sequencing	Species and strains: B. bifidum CCFM16	α-diversity: NS between postprobiotics and postplacebo, NS between preprobiotics and postprobiotics. β-diversity: by Aitchison distance: SS between pre-in and post-in probiotics group, while NS between pre-in and post-in placebo group; by Bray-Curtis dissimilarity: NS between pre-in and post-in probiotics group. Relative abundance of Firmicutes and Bacteroidetes were altered by CCFM16 treatment. B. bifidum CCFM16 increased the Shannon and Simpson α-diversity indices of the bifidobacterial species.
Kang 202121	metagenomic analysis and 16 s rRNA sequencing	Species and strains: Ba. coagulans SNZ 1969	α-diversity: NS between postprobiotics and postplacebo. β-diversity: NS between postprobiotics and postplacebo. LEfSe: Bacillales (p=0.007) and Lactobacillales (p=0.014) was significantly abundant in postprobiotics compare with postplacebo, while Cloacibacillus (p=0.047) was significantly diminished in postprobiotics. Relative abundance of Ba. coagulans was significant abundant after probiotics treatment.
Kim 202122	16 s rRNA sequencing	Species and strains: L. casei IDCC 3451, L. plantarum IDCC 3501, L. rhamnosus IDCC 3201, B. lactis IDCC 4301, B. breve IDCC 4401, La. lactis IDCC2301 additional ingredient: xylooligosaccharide	α-diversity: by OTUs, Chao1 index, or phylogenetic diversity: NS between pre-in and post-in placebo group, while significantly decreased in the postprobiotics; by Shannon index, NS between pre-in and post-in probiotics and placebo group, respectively. Correlation between clinical and microbiota data: in bowel-activity improvement, Ruminococcaceae and Lachnospiraceae showed a positive correlation, Prevotellaceae and Bacteroidaceae showed a negative correlation.
Martoni 201926	16 s rRNA sequencing	Species and strains: L. acidophilus DDS-1, B. animalis subsp. lactis UABla-12, B. longum UABl-14 and B. bifidum UABb-10, additional ingredient: fructooligosaccharide, magnesium stearate and silica	α-diversity: by OUTs and Shannon diversity: NS between pre-in and post-in probiotics group. Relative abundance: Ruminococcaceae increased and Tenericutes decreased in postprobiotics compared with preprobiotics after adjusting for multiple comparison testing; B. animalis subsp. lactis (p<0.001), B. bifidum (p=0.001), B. longum (p<0.001) and L. acidophilus (p=0.003) increased in postprobiotics compared with preprobiotics β-diversity: samples did not cluster or separate according to time point in either group COG: NS in KEGG pathway abundances between pre-in and post-in probiotics groups. Correlation between clinical and microbiota data: sex (p=0.012) and Bristol Stool Scale score (p=0.001) contributed to overall β diversity
Dimidi 201917	qPCR	Species and strains: B. lactis NCC2818	Relative abundance: NS in microbiota concentrations between probiotic and placebo group in pre or post
Tanaka 201529	qPCR	Species and strains: B. lactis GCL2505	Relative abundance: bifidobacteria increased after probiotics treatment
Kondo 201323	qPCR	Species and strains: B. longum BB536	Relative abundance: bifidobacteria increased after treatment more in post-probiotics than postplacebo

. B, Bifidobacterium; Ba, Bacillus; COG, clusters of orthologous groups; ko, KEGG orthology; L, Lactobacillus; La, Lactococcus; NS, non-significance; OUT, operational taxonomic unit; qPCR, quantitative PCR; rRNA, ribosomal RNA; SS, significance; u, unigenes.

Two studies compared the difference in α diversity between probiotics and placebo groups after treatment, Wang et al found non-significant difference in observed features, shannon, chao1 and simpson diversity indices,³⁰ while Kang et al indicated non-significant difference in shannon and chao1 diversity indices.²¹ Kim et al found significant decrease in OTUs and Chao1 in experimental group after treatment, suggesting the decline in α diversity.²² In β diversity, Zhang et al revealed no difference between the pretreatment and post-treatment faecal samples in experimental group at the phylum and genus level,³² Wang et al found that there was no significant difference between pretreatment and post-treatment faecal samples in experimental group when measured by Bray-Curtis dissimilarity rather than Aitchison distance.³⁰ Kondo et al and Tanaka et al found that the relative abundance of bifidobacteria significantly increased after probiotics-containing products treatment,^{23 29} Kang et al showed that Bacillus coagulans was significant more abundant after probiotics treatment,²¹ Wang et al revealed that the relative abundance of Firmicutes and Bacteroidetes was altered by CCFM16 treatment,³⁰ however, Dimidi et al found no difference in microbiota concentrations between probiotics group and placebo group at end of the intervention.¹⁷

Discussion

The systematic review and meta-analysis was an update of the efficacy as well as a summarise of gut microbiota profile for the first time. In efficacy, our pooled results showed, compared with placebo, that probiotics-containing products significantly increased stool frequency, improved stool consistency and decreased constipation symptoms measured by PAC-SYM. We also reviewed the change of microbiota after treatment through α diversity, β diversity, change in relative abundance, etc.

Considered as promising ingredients to alleviate constipation, controversial results between different RCTs of probiotics-containing products suggested the desire for valid quantitative review. Compared with previous meta-analysis studies,12 13 33–37 we did not just update eligible RCTs, but more importantly, we made a series of improvements. We only included studies recruiting adult patients diagnosed with FC with the exclusion of RCTs whose participants mixed FC with IBS-C, or studies that included constipated patients with defecation-impairing comorbidities (such as diabetes). We also graded the quantitative results by GRADE methodology in order to help clinicians more comprehensively understand the pooled efficacy based on current researches.

Patients with FC often experience decreased bowel movement frequency accompanying delayed gastrointestinal transit time.^{1 38} Zhang et al showed that probiotics administration significantly increased stool frequency and shortened the gastrointestinal transit time in FC patients, indicating the favourite treatment for slow transit constipation.¹³ While our study focused on the efficacy of probiotics-containing products (probiotics and synbiotics) on bowel movement frequency and revealed positive results, by subgroup analysis, our pooled results, with similarity to previous researches,^{12 13} indicated that Bacillus-containing formula might had the best but non-significant efficacy, while multistrains formula had satisfactory and significant efficacy. Moreover, we found that synbiotics had superior efficiency over probiotics in stool frequency, stool consistency (non-significance in synbiotics subgroup) and PAC-SYM (shown in figure 2A, online supplemental figure S3A and S4A). In fact, another active ingredient in the product, prebiotics, also beneficially affected bowel function and relieve constipation. Fructooligosaccharides could be fermented by gut microflora accompanying the production of short-chain fatty acid which stimulate bowel movements. One RCT by Meksawan et al showed that fructooligosaccharides significantly increased frequency of defecation and improve stool consistency compared with placebo.³⁹ Besides the prebiotic benefit, psyllium also had the capability to retain water to increase stool content, and thereby, improve the motility of intestinal tract.⁴⁰ Through random-effect meta-regression, the influence of covariate on pooled efficacy was explored. As a result, age, treatment duration and dose of probiotics showed non-significant correlation with pooled efficacy, consistent with the studies by Zhang et al.¹³ But we found that the proportion of gender significantly correlated with the pooled efficacy in stool frequency. Obesity and diet greatly influence the composition of the gut microbiota, which might affect the efficacy,⁷ however, in many of included studies, there was a lack of BMI data or dietary information record. Therefore, we did not take BMI and dietary factors into the analysis of random-effect meta-regression.

Hard stool seriously aggravates FC patient’s defecation symptoms, prolongs the defecation time and even induces haemorrhoids.^{41 42} Our meta-analysis revealed that probiotics-containing administration significantly softens stools, consistent with the results of previous studies.^{13 36} Similarly, we also found that probiotics-containing administration reduce the scores of PAC-SYM more effectively compared with placebo. However, the relatively high heterogeneity and the relatively limited eligible studies degraded the quality as ‘very low’, hence, more high-quality RCTs are still needed for further verification.

Constipated patients suffered structural alteration on gut flora compared with healthy controls. Khalif et al found that the concentration of Bifidobacterium and Lactobacillus in constipated patients significantly decreased, while potentially pathogenic micro-organism increased.⁴³ Zhu et al observed significant decreased abundance in Prevotella in constipated children compared with healthy controls.⁴⁴ In addition, slow transit constipated patients had higher prevalence of methanogenic flora, which decreased smooth muscle peristaltic velocity, than normal transit constipated patients or controls.^{44 45}

Supplementation of probiotics is an active attempt to reverse intestinal structural imbalance in patients with constipation. Bifidobacterium and Lactobacillus are essential microbes for normal intestinal function. These micro-organisms ferment indigestible foods and produce metabolites that promote gut motility, such as short-chain fatty acids and serotonin.^{46 47} There are a variety of probiotics-containing products, including single-strain products and multistrains products, containing Bifidobacterium, Lactobacillus or specific strains indirectly promoting the growth of Bifidobacterium, etc. The work by Zhang et al indicated the priority of multi-strains treatment over single-species probiotics treatment in efficacy,¹³ while Miller et al found, by subgroup analysis, that treatment containing Lactobacillus and Bifidobacterium had significant higher improvement than treatment containing Lactobacillus or Bifidobacterium only.³⁵

In order to further understand the effects of probiotics-containing products on intestinal flora, we systematically reviewed the relevant researches. In α diversity, three of four studies showed that there was no significant change/difference,^{21 26 30} while in β diversity, three of four studies observed similar negative results,^{21 26 32} indicating that short-term supplementations of probiotics-containing products might not affect the stability in α or β diversity of intestinal flora. A systematic review, including seven RCTs summarised by Kristensen et al, also demonstrated a similar negative efficacy of probiotics supplement on faecal microbiota composition in healthy participants.⁴⁸ By contrast, Zhang et al observed that the intestinal flora in FC patients showed significant increasement in α diversity as well as alteration in β diversity after faecal microbiota transplantation.⁴⁹ However, this did not mean the uselessness of probiotics-containing products, in the opposite, the minimal effect of probiotics-containing products on α or β diversity of intestinal flora might imply the safety. More comprehensively, five of six studies revealed that probiotics-containing products significantly increased the relative abundance of specific gut micro-organism which corresponded with the probiotics-containing supplement.21 23 26 29 30 While the studies by Dimidi et al found no significant difference in stool microbiota concentration including Bifidobacterium spp between the probiotic and placebo group before or after probiotics treatment.¹⁷

Synbiotics combined probiotics with prebiotics, improving the survival of formula probiotics as well as the reproduction of intestinal probiotics.⁵⁰ In our included studies, the most popular probiotics ingredients were Bifidobacterium and Lactobacillus, while the most popular prebiotics ingredients were inulin, fructooligosaccharide and psyllium.⁵¹ The change of gut microbiota was researched through the comparison between pretreatment and post-treatment stage in the intervention group, or between the intervention group and the control group in post-treatment stage. There is still a lack of appropriate methodology to measure the synergistic effects of probiotics combined with prebiotics on gut microbiota, compared with probiotics alone.

This meta-analysis has some limitations. The efficacy of treatment might be misestimated due to modest sample sizes and high heterogeneity of pooled results among the pooled results, additionally, the absence of clear protocols increases the possibility of publication bias. Obesity and diet influence the composition of the gut microbiota, which might affect the pooled results. But relevant data are unavailable in many studies, so it is recommended that future clinical trials pay attention to covariates that influence the gut microbiota to obtain more reliable conclusions. Moreover, there is no unified measurement methodology for the impact of probiotics-containing products on intestinal flora, and the protocols lack corresponding specific description. Many researchers were hesitant to publish negative data, perceiving them as seemingly inconsequential, which might cause our results unreliable or even misleading. However, negative data could provide an objective assessment of efficacy and offer valuable insights into various factors, such as dosages, duration of treatment. Through comprehensive analysis, we would gain a better understanding of situations where the treatment might not be effective and how we could improve. Although publishing negative data may not be popular, they are indispensable for ensuring the accuracy and comprehensiveness of Meta-analysis. Hence, we encourage future research community to actively report and use negative data.

Conclusion

In summary, probiotics-containing products are effective to increase stool frequency, improve stool consistency and alleviate constipated symptoms scaled by PAC-SYM. Bacillus-containing formula exhibits promising but non-significant efficacy, and requires further validation; the multistrains formula displays satisfactory and credible efficacy, while synbiotics demonstrates superior efficacy compared with probiotics. In addition, probiotics-containing products would increase the relative abundance of specific strain. Due to the quality of included studies, the evidence ought to be used with caution. More high-quality head-to-head RCTs are needed to further verify the relative efficacy of different administration. Moreover, future researches should pay attention to the influence of covariable such as BMI, dietary habits, medication history, combination of probiotics and treatment duration, and use better statistical methods to measure change of gut microbiota.

Data availability statement

Data are available on reasonable request.

Ethics statements

Patient consent for publication

Not applicable.