Inulin-induced improvements on bowel habit and gut microbiota in adults with functional constipation: findings of a randomized, double-blind, placebo-controlled study

Abstract

Background

Functional constipation is a common disorder of the gut-brain interaction characterized by infrequent bowel movements and hard stools, which substantially affects patients’ quality of life. Supplementation with gut microbiome-targeted prebiotics is a promising non-pharmacological alternative to current treatments.

Methods

In a randomized, double-blind, placebo-controlled, cross-over trial, we investigated the effect of four-week daily 12 g chicory inulin intake in 39 individuals, with functional constipation according to Rome III criteria. We assessed stool frequency and consistency, constipation-related quality of life (PAC-QOL) and symptoms (PAC-SYM), and microbiota composition through 16S rRNA gene sequencing of fecal samples.

Results

After inulin intake, we observed larger changes in stool frequency, abdominal symptoms, and particularly social and emotional well-being related to quality of life, compared to placebo. Additionally, relative abundances of putative butyrate-producing Anaerostipes spp. and Coprococcus 1 spp. were higher. Further investigation, however, pointed to a carry-over whereby half of the participants receiving inulin in the first period had the largest improvements in all outcomes during inulin intake. Moreover, these same participants had higher baseline relative abundances of butyrate-producing Faecalibacterium spp. and Roseburia spp., and lower baseline relative abundances of Bifidobacterium spp. In turn, these participants’ microbiota responded more strongly to inulin intake through a stronger increase in the relative abundance of Bifidobacterium spp. and Anaerostipes spp. To address this apparent carry-over induced bias, we analyzed the first period alone as a parallel trial which supported the observed positive effect of inulin, also affirming its established bifidogenic effect.

Conclusions

Daily intake of 12 g inulin improves functional constipation by increasing stool frequency, positively affecting abdominal symptoms and well-being, and modulating the gut microbiota towards higher relative abundances of butyrate-producing genera.

Trial registration

Ethical approval was obtained from the Cork Research Ethics Committee of the Cork Teaching Hospitals (Reference ECM 4 (v) 01/09/15), and the trial design was retrospectively registered at ClinicalTrials.gov (NCT05447481; 07/2022).

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-025-04409-6.

Background

Functional constipation is a disorder affecting individuals of all ages. It is characterized by infrequent bowel movements and difficult and/or painful passage of (hard) stools that have no organic or medication-related cause [1, 2]. Depending on the criteria applied, the estimated prevalence ranges from one in ten to eight in ten adults reporting constipation-related symptoms [3–6], with higher rates consistently observed in women [7, 8]. The most widely accepted defining criteria of functional constipation is the Rome Foundation’s diagnostic framework, most recently updated as the Rome IV criteria, which defines functional constipation as meeting at least two of several symptoms, including straining, hard stools, incomplete evacuation, anorectal obstruction, manual maneuvers to aid defecation, fewer than three bowel movements per week, and the absence of loose stools without laxatives. Symptoms must persist for at least three months, with onset at least six months before diagnosis, and should not meet the criteria for irritable bowel syndrome [1, 9, 10]. In 2016, functional constipation was reclassified as a disorder of gut-brain interaction (DGBI) to highlight the bidirectional connection between the gut and brain, mediated through anatomical, endocrine, and immune pathways [1, 11]. Diagnosis is based on reported symptoms, including their duration, frequency, and interference with daily life [2, 11]. Like other DGBI, the exact cause of functional constipation remains unclear but is likely multifactorial, involving psychological stress, lifestyle factors, genetic predisposition, motility disturbances, and the gut microbiome [4].

The gut microbiome is a diverse community of bacteria, viruses, fungi, protozoa, and archaea, also referred to as the gut microbiota, along with their collective functions, residing in the human gut and unique to each individual [12]. Alterations in gut microbiota composition have been associated with hard stools and slow intestinal transit times [13, 14]. Notably, individuals diagnosed with functional constipation according to the Rome III criteria, were shown to have reduced relative abundances of butyrate-producing taxa in their fecal samples compared to healthy controls [15]. This reduction has also been observed in colonic tissue samples from individuals with severe constipation requiring surgical intervention [16] and in fecal samples from individuals with irritable bowel syndrome, another DGBI [17, 18]. Additionally, osmotic laxatives, a common pharmacological treatment for functional constipation, have been associated with reduced butyrate-producing taxa [19]. Interestingly, butyrate-producing taxa often rely on cross-feeding from fiber-fermentation products from other gut bacteria [20]. Since low fiber consumption is linked to an increased incidence of constipation [8], the limited availability of fermentable dietary fiber in the colon may contribute to the condition.

Increasing dietary fiber intake has been associated with significant cost savings in medical expenses for constipation, potentially exceeding $12 billion annually in the USA [21]. Moreover, as one-third of patients express dissatisfaction with pharmacological treatments, seeking non-pharmacological alternatives [22], dietary fibers offer an appealing solution. Dietary fibers or osmotic polymers are believed to aid bowel function by holding water in the colon to soften stools, as seen with psyllium and polyethylene glycol (PEG), which serve only minimally as fermentable substrates for gut microbiota [23]. In contrast, fermentable fibers, metabolized by gut microbes, have the dual benefit of influencing both bowel function and gut microbiota composition. Certain fermentable fibers, known as prebiotics, are of particular interest as they can selectively modulate bacterial taxa and provide health benefits [24, 25]. One such well-established prebiotic is (native) inulin, which promotes the growth of bifidobacteria [26–28] and is approved by the European Food Safety Authority (EFSA) for maintaining normal defecation by increasing stool frequency [29]. In addition, inulin may alleviate constipation symptoms [30], and also impact aspects of brain function, such as cognition [31], food decision-making [32], or neurological biomarkers in Parkinson’s disease [33]. Thus, inulin offers a promising non-pharmacological alternative to managing functional constipation by modulating the gut microbiota.

Here, we aimed to investigate the effect of a four-week daily intake of 12 g inulin versus placebo on stool frequency and stool consistency, constipation-related quality of life and symptoms, resort to laxatives, physical activity, and fecal microbiota composition in adults with functional constipation.

Materials and methods

Study design

This study was designed as a randomized, double-blind, placebo-controlled, cross-over trial conducted in adults with functional constipation, in which we compared a four-week period of daily intake of 12 g of inulin against a four-week period of placebo intake (maltodextrin), separated by a four-week wash-out period in between the two intervention periods. The primary outcome was the change in stool frequency after treatment compared to baseline. Ethical approval was obtained from the Cork Research Ethics Committee of the Cork Teaching Hospitals (Reference ECM 4 (v) 01/09/15), and the trial design was registered at ClinicalTrials.gov (NCT05447481). The study was conducted from 2015 to 2016 at Atlantia Clinical Trials, Cork, Ireland, and was executed in accordance with the principles of the Declaration of Helsinki and in compliance with the International Council for Harmonization Good Clinical Practice.

Participants

Study participants were recruited through the clinical research organization’s database, general practitioners’ offices, hospital clinics, and adverts in local newspapers in the surroundings of Cork, Ireland. Written informed consent was obtained from all study participants prior to their participation. To be eligible for the study, participants had to be between 18 and 75 years old and diagnosed with functional constipation according to the Rome III criteria [34]. Individuals were excluded if they had hypersensitivity to any of the test product components or an acute or chronic, unstable, and untreated disease or any condition that contra-indicated entry to the study. Additionally, a history of laxative use, drug and/or alcohol abuse, and intake of any probiotic or prebiotic product or supplement in the two weeks before the screening visit resulted in exclusion from the study. Lastly, women who were pregnant, lactating, or wished to become pregnant during the study were not included. Participants were evaluated to be in good general health, as determined by the investigator, and asked to continue their normal diet while avoiding pro- or prebiotic products/supplements or dietary fiber supplements for the duration of the study.

Study procedure

Potential participants were screened for eligibility, and eligible individuals entered a two-week run-in period, after which they were randomized into either the treatment-placebo sequence or the placebo-treatment sequence. Participants received as treatment 12 g inulin (Frutafit^® HD native chicory inulin, Sensus B.V., Roosendaal, the Netherlands) and as placebo maltodextrin (MD20, Avebe, Foxhol, the Netherlands) for four weeks. The first intervention period (either treatment or placebo) was followed by a four-week wash-out period before participants crossed over into the second intervention period (Fig. 1). The intervention products were consumed in two doses of 6 g per day, except for the first three days of each intervention period when participants only took one dose of 6 g per day. Digestible maltodextrin was chosen as the placebo as it is easily broken down in the upper gastrointestinal tract by human endogenous enzymes [35], and therefore is not expected to affect the lower gut (colonic) microbiota. The study products were similar in flavor, appearance, and packaging, and both participants and research personnel involved in this study were blinded to ensure true allocation concealment. The study products were provided in dark plastic 120 ml bottles. Participants were instructed to add 60 ml of water to the bottle and shake until the study product had dissolved. As a marker of compliance, participants were asked to return all used and unused bottles, and compliance was calculated by counting the actual number of empty bottles returned and dividing it by the expected number of bottles used. Each participant received 60 bottles, with 53 as the total expected number to be used to be 100% compliant. Participants received all bottles of the respective intervention product (inulin or placebo) on day 0 for each four-week intervention period (period 1 or 2). During each intervention period, participants filled in a daily bowel diary in which all outcomes were assessed, including stool frequency, stool consistency, and constipation complaints according to Rome III criteria. At the end of each intervention period, participants also completed three questionnaires covering aspects of constipation-related quality of life and symptoms as well as physical activity. Finally, participants were asked to provide a fecal sample at the end of each intervention period (Fig. 1), in order to assess the gut microbiota composition. Other measurements were blood pressure, heart rate, and temperatures and anthropometric measures such as weight, height, and body mass index (BMI), as well as medication use, which were all taken before the start of the run-in and at the end of each study period (run-in, intervention period 1, wash-out, intervention period 2). Any (serious) adverse events or symptoms that were observed by the investigator or spontaneously reported by study subjects were systematically recorded.

Fig. 1.

Study design. Overview of the randomized, placebo-controlled cross-over study design and study measurements

Outcome assessments

The primary outcome measure was defined as the change in the number of bowel movements per week from a two-week baseline to the final two weeks of inulin intake, compared to the change from a two-week baseline to the final two weeks of placebo intake. Secondary outcomes included differences between changes after inulin compared to placebo intake in the following parameters: (1) stool consistency, measured using the Bristol Stool Form Scale (BSFS); (2) quality of life related to constipation, assessed by Patient Assessment of Constipation – Quality of Life (PAC-QOL) scores; (3) constipation-related symptoms, assessed by Patient Assessment of Constipation – Symptoms (PAC-SYM) scores; (4) use of laxatives; (5) physical activity, determined using the International Physical Activity Questionnaire (IPAQ); and (6) fecal microbiota composition.

Stool frequency, stool consistency, and resort to laxatives

Stool frequency, stool consistency, Rome III criteria, and resort to laxatives were measured by using daily bowel habit diaries. Participants completed daily bowel habit diaries, reporting the number of bowel movements and the consistency of the stool using the BSFS.

Patient assessment of constipation – quality of life and symptoms

The PAC-QOL is a retrospective questionnaire with a recall period of two weeks, comprising 28 items scored on a 5-point Likert scale from 0 to 4, and designed to assess the patient-reported quality of life and the impact of constipation symptoms [36]. It generates an overall score and four subscores, including worries/concerns (11 items), physical discomfort (four items), psychosocial discomfort (eight items), and satisfaction (five items). Similarly, the PAC-SYM is a retrospective questionnaire with a two-week recall period, consisting of 12 items (scored on a 5-point Likert scale from 0 to 4) assessing the severity of patient-reported symptoms [37]. This tool also produces an overall score and three subscores, focusing on abdominal symptoms (four items), rectal symptoms (three items), and stool symptoms (five items). In both the PAC-SYM and PAC-QOL, lower scores on the overall score or subscore indicate less severe symptoms and a higher quality of life, respectively.

Physical activity

The IPAQ is a retrospective questionnaire with a recall period of seven days, comprising 27 questions to evaluate health-related physical activity [38]. The questions address types of vigorous and moderate physical activity and time spent thereon and are grouped into five parts (job-related, transportation, housework/house maintenance, recreation/sport/leisure-time physical activity and time spent sitting). Outcomes are summarized either in three categories of physical activity, as implemented here, which are “high” (one hour of moderate-intensity physical activity or more per day), “moderate” (half an hour of moderate-intensity physical activity on most days) and “low” (not meeting the other categories’ criteria) or can be further converted taking energy requirements into account.

Sample size calculation and randomization

To detect a minimal difference of one bowel movement per week between inulin treatment and placebo with a standard deviation (SD) of 2, a total sample size of 39 participants was required. The sample size was calculated using a power of β = 80% and a significance level of α = 5% based on a two-sided Wilcoxon non-parametric test and takes a drop-out rate of 10% into account. The participants were randomized equally into the treatment-placebo or placebo-treatment arm by an independent statistician, using the uniform random number function in SPSS (IBM Corp. IBM SPSS Statistics for Windows. Armonk, NY: IBM Corp). After an initial blinded data analysis for internal use, the data was unblinded, and the data analysis was performed unblinded.

Fecal DNA extraction, 16S rRNA gene amplification and sequencing

During the study, participants were asked to provide a fecal sample during the run-in period and at the end of each intervention period and the wash-out (Fig. 1). Samples were collected at home, stored in home freezers, and then transferred to the study center in cooler bags with a frozen ice pack. The samples were then stored at −20°C before being processed. DNA was extracted from fecal samples using a repeated bead-beating step and the Maxwell^® 16 instrument (Promega, Leiden, The Netherlands). First, 0.25 g of fecal material was added to a bead-beating tube with 700 µl Stool Transport and Recovery (STAR) buffer, 0.5 g of sterilized zirconia beads (0.1 mm), and five glass beads (2.5 mm). These tubes, containing the fecal sample, were bead-beaten three times (60 s × 5.5 ms) and incubated for 15 min at 95 °C at 300 rpm. Samples were then centrifuged for 5 min at 4 °C, and 14,000 g. The resulting supernatants were transferred to sterile tubes, and the procedure was repeated on the remaining pellets using 300 µl STAR buffer. Both supernatants were pooled, and DNA purification was performed with a customized kit (AS1220; Promega) using 250 µl of the final supernatant pool. DNA was eluted in 50 µl of DNAse- and RNAse-free water, and its concentration was measured using a DS-11 FX + Spectrophotometer/Fluorometer (DeNovix Inc., Wilmington, USA). The V4 region of the 16S ribosomal RNA (rRNA) gene was amplified in duplicate PCR reactions for each sample in a total reaction volume of 50 µl. The primers used were 515F (5’-GTGTGYCAGCMGCCGCGGTAA-3’) [39] and 806R (5’-CCGGACTACNVGGGTWTCTAAT-3’) [40]. The master mix contained 1 µl of a unique barcoded primer, 515F-n and 806R-n (10 µM stock concentration), 1 µl dNTPs mixture (200 µM), 0.5 µl Phusion Green Hot Start II High-Fidelity DNA Polymerase (2 U/µl; Thermo Scientific, Landsmeer, The Netherlands), 10 µl 5× Phusion Green HF Buffer, and 36.5 µl DNAse- and RNAse-free water. The amplification program included 30 s of an initial denaturation step at 98 °C, followed by 25 cycles of denaturation at 98 °C for 10 s, annealing at 50 °C for 10 s, elongation at 72 °C for 10 s, and a final extension step at 72 °C for 7 min. The PCR product was visualized in 1% agarose gel (~ 290 bp) and purified with CleanPCR kit (CleanNA, Alphen aan den Rijn, The Netherlands). The concentration of the purified PCR product was measured with Qubit dsDNA BR Assay Kit (Invitrogen, California, USA), and 200 ng of microbial DNA from each sample was pooled for the creation of the final amplicon library, which was sequenced (150 bp, paired-end) on the Illumina HiSeq 2000 platform (GATC Biotech, Constance, Germany). For quality control purposes, two in-house assembled DNA mock communities were included in the library and compared to their theoretical composition (MC3 and MC4; [41]). Additionally, a negative control of the DNA extraction and purification procedure and a water blank were included. The 16S rRNA gene amplicon sequences have been submitted to the European Nucleotide Archive under accession number PRJEB98644.

Data analysis

Data was analyzed according to the assigned sequence allocation at baseline, and available data from participants who dropped out or provided incomplete data were included in the analysis.

Assessment of clinical outcomes

Clinical outcomes were analyzed using R (version 4.0 or higher) [42], and the packages tidyverse [43], ggstatsplot [44], ggplot2 [45], ggpubr [46], lme4 [47], lmerTest [48], emmeans [49], rstatix [50] and tableone [51] were used. Outcomes recorded in the bowel habit diaries, being stool frequency, stool consistency, and resort to laxatives, were summarized over the last two weeks of each period (run-in, period 1, wash-out, and period 2) and averaged to reflect measurements per week (e.g., stool frequency as bowel movements per week). Total scores and subscores of the PAC-QOL and PAC-SYM were calculated as previously described [36, 37] and covered the same two-week periods as the bowel habit outcomes. Changes in all these outcomes were then assessed by calculating the difference between the average of each intervention period (period 1 or period 2; post-intake levels) and that of each respective baseline (run-in for period 1 and wash-out for period 2; pre-intake levels). Normality was checked using QQ-plot, and outcomes were summarized using parametric descriptive statistics, except stool frequency, which was also summarized using nonparametric statistics. Physical activity was categorized as previously described [38] into three groups (“low,” “moderate,” and “high”) and analyzed as categorical data. All other outcomes (blood pressure, heart rate, anthropometry) were summarized at baseline using descriptive statistics. Inference of statistical significance was made at α = 0.05.

The main analysis assessed whether the change during treatment (inulin post-intake versus pre-intake) differed from that during placebo intake (maltodextrin), here referred to as intervention difference, and calculated using linear mixed modeling. Time (pre versus post) and intervention (inulin versus placebo) were included as fixed effects, and individual was included as random effect. The change was modeled as a fixed effect by the interaction term intervention*time. There is ongoing debate on how to address carry-over effects, given the lack of a robust test and its possible concealment within other effects [52–54]. Here, we modeled apparent carry-over (whether the effect of the intervention is impacted by the period) by including the interaction term intervention*period (period 1 versus period 2), which we found to improve the models assessed by the Akaike Information Criterion (AIC). Additionally, we analyzed the first period alone, treating it as a parallel trial [52, 53], a common approach known to increase the risk of type 1 errors (false-positives) [52]. Acknowledging the limitations of this approach, we incorporated it to function as a robustness check with respect to apparent carry-over and reduced fecal samples in period 2. Further exploration of potential reasons for carry-over was based on per-sequence-analysis (treatment-placebo versus placebo-treatment) and by investigating differences between subjects in the treatment-placebo sequences who were apparently affected by carry-over (stool frequency of >3x/week in period 2) versus those who were not (stool frequency of < = 3x/week in period 2). Data are presented as estimated marginal means (EMM) with their standard errors (SEM), unless stated otherwise.

Assessment of fecal microbiota outcomes

For fecal microbiota composition data analysis, data filtering and taxonomy assignment of the 16S rRNA gene amplicon sequences were performed using the NG-Tax pipeline with default settings [55]. A table based on Amplicon Sequence Variants (ASVs) was created for each sample and taxonomically annotated using the SILVA database release 138.1 (release date: August 27, 2020) [56]. Low abundance ASVs were discarded, using a minimum relative abundance threshold of 0.1% as described previously [41, 55]. Fecal microbiota composition and changes in individual taxa were analyzed using the microViz [57] and the microbiome [58] packages, which use the phyloseq [59] and the vegan [60] packages. Moreover, the mare package [61] was used, which relies on the vegan package, the MASS [62] and glmmADMB [63, 64] packages. Microbiota composition was compared by calculating β-diversity based on Bray-Curtis dissimilarity, a measure of between-sample variation. Differences in β-diversity were assessed using permutational analysis of variance (PERMANOVA). Differences in relative abundance of individual taxa at pre- and post-intake and changes for both the cross-over as well as the analysis of subjects apparently affected by carry-over and analysis of the first period only were calculated as implemented in mare, and details of such analyses have been previously described elsewhere [65]. In short, the mare package includes the individual as random factor in the repeated measure analysis and offers the possibility of analyzing data using zero-inflated negative binominal models as well as models excluding samples where respective taxa are not observed (non-zero models). Multiple testing was adjusted using the false discovery rate (FDR), and results with q-values ≤ 0.1 were considered statistically significant.

Results

Participant characteristics

Forty participants were randomly assigned to the study, with 20 each allocated to the treatment-placebo and placebo-treatment sequences, respectively (Fig. 2A). Participants were primarily females, on average 37 years old, and all fulfilled the Rome III criteria at run-in, having two defecations per week and an average BSFS score of less than two at run-in (Fig. 2B). One participant in the placebo-treatment sequence dropped out after the wash-out period; a second was excluded due to unaccounted symptom improvements. Five participants provided incomplete bowel habit data, but all participants completed the PAC-QOL, PAC-SYM, and IPAQ questionnaires, and all available data was included in the analysis. Thirty-eight of the 40 participants provided a fecal sample at baseline, 38 after period 1, 29 provided a fecal sample after wash-out, and 15 after period 2. Compliance was excellent, with all participants exceeding the minimum product intake of 80% (range 92.5% − 113.2%). None of the participants used laxatives.

Fig. 2.

Participants in the cross-over study. A Consort statement flow diagram (B) Baseline characteristics. BMI, Body Mass Index; BSFS, Bristol Stool Form Scale; IPAQ, International Physical Activity Questionnaire; PAC-QOL, Patient Assessment of Constipation-Quality of Life; PAC-SYM, Patient Assessment of Constipation-Symptoms. Data are presented as mean ± SD or median [range] if not specified otherwise; p-values represent outcomes of corresponding parametric or nonparametric statistical testing

Safety data

During the intervention, 16 of 40 participants reported possibly intervention-related adverse events. These were limited and aligned with expectations, i.e., gastrointestinal symptoms from increased dietary fiber intake. The most common was flatulence (inulin: n = 8), followed by bloating (inulin: n = 5, placebo: n = 1) and cramps (inulin: n = 4, placebo: n = 1). No action was required, and all events were resolved/resolving at the last visit; there were no discontinuations due to adverse effects. No differences at baseline were detected between participants in the treatment-placebo sequence and those in the placebo-treatment sequence (Fig. 2B). There were no serious adverse events.

Cross-over analysis of bowel habit changes

First, we compared changes in bowel habits between inulin and placebo intake (Table 1 and Supplementary Table 1). Stool frequency increased by 1.43 (0.19) defecations per week with inulin and 0.90 (0.19) with placebo (estimated marginal means (SEM)). The difference between interventions was statistically significant, with 0.53 (0.26) defecations per week (p = 0.046, Table 1) for inulin versus placebo. On average, participants defecated 3.44 (0.15) times per week with inulin and 3.15 (0.15) times with placebo (Table 1). However, we found a statistically significant intervention*period interaction for stool frequency (p = 0.014, Supplementary Table 1), suggesting that carry-over effects could not be excluded for either outcome. Stool consistency increased similarly after both inulin and placebo by about half a BSFS unit, resulting in comparable final post-intake BSFS scores (Table 1).

Table 1.

Changes in stool frequency, stool consistency, PAC-QOL, PAC-SYM and physical activity. Pre, and post-intake levels and changes are reported as estimated marginal means (SEM) if not otherwise indicated. In addition, for intervention differences between inulin and placebo also the 95%standard confidence interval (CI) and related p-values are given

	Inulin			Placebo			Intervention difference Inulin vs. placebo
	Pre-intake	Post-intake	Change	Pre-intake	Post-intake	Change	Difference	95%CI	p-value
Stool frequency (x/week) a	2.01 (0.15)	3.44 (0.15)	1.43 (0.19)	2.25 (0.15)	3.15 (0.15)	0.90 (0.19)	0.53 (0.26)	[0.02, 1.04]	0.046*
Stool consistency (BSFS)	1.95 (0.12)	2.57 (0.12)	0.62 (0.12)	2.03 (0.12)	2.55 (0.12)	0.52 (0.12)	0.09 (0.17)	[−0.25, 0.43]	0.600*
PAC-QOL
Total score	2.01 (0.11)	1.39 (0.11)	−0.61 (0.10)	1.81 (0.11)	1.60 (0.11)	−0.21 (0.10)	−0.41 (0.15)	[−0.69, −0.12]	0.007
Physical discomfort	2.29 (0.14)	1.49 (0.14)	−0.80 (0.13)	2.03 (0.14)	1.66 (0.14)	−0.37 (0.13)	−0.43 (0.19)	[−0.08, −0.07]	0.023
Psychosocial discomfort	1.44 (0.14)	0.94 (0.14)	−0.50 (0.12)	1.23 (0.14)	1.21 (0.14)	−0.02 (0.12)	−0.48 (0.17)	[−0.81, −0.14]	0.007
Worries and concerns	1.85 (0.13)	1.28 (0.13)	−0.56 (0.11)	1.67 (0.13)	1.51 (0.13)	−0.16 (0.10)	−0.40 (0.15)	[−0.69, −0.11]	0.009
Satisfaction	3.17 (0.13)	2.28 (0.13)	−0.89 (0.16)	3.05 (0.13)	2.44 (0.13)	−0.62 (0.16)	−0.28 (0.22)	[−0.71, 0.16]	0.218*
PAC-SYM
Total score	1.96 (0.11)	1.34 (0.11)	−0.61 (0.11)	1.72 (0.11)	1.45 (0.11)	−0.27 (0.11)	−0.35 (0.15)	[−0.64 −0.06]	0.022
Abdominal symptoms	2.08 (0.13)	1.46 (0.14)	−0.62 (0.14)	1.73 (0.13)	1.58 (0.13)	−0.15 (0.14)	−0.47 (0.19)	[−0.84, −0.09]	0.017
Rectal symptoms	1.25 (0.12)	0.73 (0.12)	−0.52 (0.12)	1.22 (0.12)	0.97 (0.12)	−0.25 (0.12)	−0.27 (0.17)	[−0.60, 0.05]	0.104
Stool related symptoms	2.26 (0.12)	1.33 (0.13)	−0.94 (0.13)	2.04 (0.12)	1.41 (0.12)	−0.64 (0.13)	−0.30 (0.19)	[−0.66, 0.06]	0.110
IPAQ b
“low”	9 (23.7%)	11 (28.9%)	-	8 (20.5%)	8 (21.1%)	-	-		0.432**
“moderate”	21 (55.3%)	18 (47.4%)	-	22 (56.4%)	16 (42.1%)	-
“high”	8 (21.1%)	9 (23.7%)	-	9 (23.1%)	14 (36.8%)	-

BSFS Bristol Stool Form Scale, IPAQ International physical activity questionnaire, PAC-QOL Patient assessment of constipation – quality of life, PAC-SYM Patient assessment of constipation – symptoms

* p-value biased due to contribution of intervention *period interaction term

** represents the p-value for the difference in post-intake counts

^a Stool frequency values expressed as median (IQR) were for inulin: pre-intake = 2.0 (1.5, 2.5), post-intake = 3.5 (3.0, 4.0), and change = 1.5 (0.5, 2.5), and for placebo: pre-intake = 2.0 (2.0, 2.5), post-intake = 3.0 (2.5, 4.0) and change = 1.0 (0.5, 1.5). The median [95%CI] difference between changes after inulin versus placebo intake was 0.75 [0.00, 1.50] and p = 0.042

^b Values are expressed as count (%)

Cross-over analysis of patient assessment of constipation – quality of life and symptoms and physical activity

Next, we analyzed the outcomes of constipation-related quality of life and symptoms questionnaires (Table 1 and Supplementary Table 1). Improvements in these scores are indicated by a negative change, with − 0.5 for PAC-QOL and − 0.6 for PAC-SYM being the minimum important differences [36, 66]. The total PAC-QOL score decreased by −0.61 (0.10) after inulin intake, which was larger than the − 0.21 (0.10) decrease after placebo, resulting in a statistically significant intervention difference of −0.41 (0.15) (p = 0.007, Table 1) for inulin versus placebo. Notably, the three PAC-QOL subscores, physical discomfort, psychological discomfort, and worries discomfort, improved statistically significantly more on inulin versus placebo (Table 1). The most remarkable decreases were observed for psychosocial discomfort and worries discomfort during inulin intake, while the placebo had no effect on these outcomes (Table 1). No differences between interventions were observed for satisfaction (the fourth subscore), but we identified a statistically significant intervention*period interaction (p = 0.002, Supplementary Table 1), which was not present for the other PAC-QOL scores.

Similar to the PAC-QOL outcomes, the total PAC-SYM score decreased more after inulin intake (−0.61 (0.11)) compared to placebo (−0.27 (0.11)), with a statistically significant intervention difference of −0.35 (0.15) (p = 0.022, Table 1) for inulin versus placebo. This was mainly due to a statistically significant decrease in abdominal symptoms with inulin (−0.62 (0.14) versus with placebo − 0.15 (0.14); p = 0.017, Table 1), despite the reported adverse events of inulin intake-related flatulence and bloating. Rectal and stool-related symptoms also decreased after inulin intake but did not differ statistically significantly from placebo intake (Table 1). None of the outcomes had a statistically significant intervention*period interaction term.

Physical activity levels, as assessed by IPAQ scores, remained consistent during inulin intake (pre- versus post-intake). However, during placebo intake, a slightly higher proportion of participants engaged in high-intensity physical activities (Table 1).

Changes in fecal microbiota and their relation with clinical outcomes

To investigate whether the gut microbiota changed and to determine if there may be associations with the clinical outcomes, we analyzed fecal microbiota composition using 16S rRNA gene amplicon sequencing (Fig. 2A). We observed no differences in overall microbiota composition after inulin compared to placebo intake, assessed by Principal Coordinate analysis (PCoA) based on β-diversity using Bray-Curtis dissimilarity at the genus level (1% of variation explained, PERMANOVA p = 0.88; Fig. 3A, Supplementary Table 2). However, similar to the clinical changes, we observed a statistically significant intervention*period interaction in the PCoA, which explained 9% of the variation (PERMANOVA p = 0.001; Fig. 3A), indicating an apparent carry-over effect was also reflected in the fecal microbiota.

Fig. 3.

Fecal microbiota composition, changes in individual taxa, and correlation with clinical outcomes. A Fecal microbiota composition at the genus level was visualized and assessed by principal coordinate analysis using β-diversity based on Bray-Curtis dissimilarity after inulin intake compared to after placebo intake. The difference between interventions, periods, and the contribution of the intervention*period interaction was tested using PERMANOVA, and corresponding p-values are depicted in the upper left of the figure. B Genera identified to significantly differ post-intake between inulin and placebo. C-F Bifidobacterium spp. relative levels and PAC-SYM rectal symptoms after (C) inulin and (D) placebo intake, as well as the correlation between Coprococcus spp. relative levels and stool frequency after (E) inulin and (F) placebo intake. The black dotted lines and Spearman correlation coefficients depict the overall correlation outcomes (both sequences combined). The colored lines and correlation coefficients depict the correlation outcomes per respective sequence in each intervention intake

Assessing individual genera and changes in their relative abundances, Bifidobacterium spp. was observed to be the most abundant genus across all samples (Fig. 3B). Bifidobacterium spp. relative abundance increased from 16.4% to 18.4% during inulin intake, compared to an increase from 14.4% to 15.1% during placebo intake, meaning post-intake levels were 1.2-fold higher for inulin versus placebo (p = 0.56; q = 0.987), though not statistically significant. Even more notable, relative abundances of Anaerostipes spp. and Coprococcus 1 spp. were 1.75-fold and 1.59-fold higher after inulin intake than placebo, respectively, based on samples where these taxa were present (Coprococcus 1 spp: 99/127 samples, 34/39 participants; Anaerostipes spp.: 110/127 samples, 37/39 participants; Fig. 3B).

Furthermore, two less abundant genera (< 1% relative abundance) changed differently after inulin compared to placebo. These were Ruminiclostridium 5 spp. (inulin: 0.75-fold from 0.34% to 0.25%; placebo: 1.74-fold from 0.28% to 0.48%) and Ruminiclostridium 9 spp. (inulin: 0.33-fold from 0.05% to 0.02%; placebo: 1.15-fold from 0.04% to 0.04%; Supplementary Fig. 1). Several other genera of the Ruminococcaceae family also decreased after inulin, but not placebo (Supplementary Table 2). However, these differences were not statistically significant between interventions.

Building on these observations, we assessed potential relationships between the relative abundance of bacterial taxa and clinical outcomes. After inulin intake Bifidobacterium spp. correlated moderately with PAC-SYM rectal symptoms (ρ = −0.36, p = 0.04; Fig. 3C), indicating that higher relative abundances of bifidobacteria were associated with lower rectal symptom scores, while this relationship was very weak after placebo intake (ρ = −0.13, p = 0.50; Fig. 3D). Also, after inulin intake Coprococcus 1 spp. relative abundances correlated moderately with stool frequency (ρ = 0.40, p = 0.02; Fig. 3E), indicating higher relative abundances of Coprococcus 1 spp. were associated with higher stool frequency outcomes, whereas this relationship was absent after placebo intake (ρ = 0.08, p = 0.68; Fig. 3F).

Analysis of the run-in and first period as parallel trial

To address the robustness of the findings in light of the apparent carry-over indicated in the cross-over analysis and the limited number of available fecal samples for microbiota analysis for the second period, we leveraged the first part of the cross-over (run-in & period 1) by analyzing it as a parallel trial (Fig. 4 and Supplementary Table 3). This analysis supported the improvements in stool frequency, PAC-QOL, and PAC-SYM total and subscores, with a larger magnitude of the observed intervention differences (Supplementary Table 3). For instance, the decrease in PAC-QOL score after inulin versus placebo was larger with 0.61 (0.16) (p = 0.003), and hence exceeded the minimum important difference of 0.5 [36]. Moreover, the analysis showed that inulin increased stool consistency by 0.83 (0.16) BSFS units, which was double that of placebo (0.40 (0.16) units), and also led to significantly higher final BSFS scores (inulin: 2.75 (0.14) versus placebo: 2.23 (0.14); p = 0.011). Also, the PAC-QOL satisfaction subscore improved significantly after inulin versus placebo intake, contrary to the cross-over analysis indicating carry-over (Supplementary Table 3).

Fig. 4.

Individual changes in bowel function, constipation quality of life and symptoms: cross-over vs. first-period analysis. A-E Cross-over analysis. Individual changes for (A) stool frequency, (B) stool consistency, (C) patient assessment of constipation quality of life (PAC-QOL) total score and (D) satisfaction subscore, and (E) symptoms (PAC-SYM) total score. F-J First-period analysis. Changes based on the analysis of the run-in and period 1 as a parallel trial for (F) stool frequency, (G) stool consistency, (H) PAC-QOL total score and (I) satisfaction subscore and (J) PAC-SYM total score. For the cross-over analysis and the first-period analysis, the estimated marginal means for the change after inulin or placebo are reported inside the figures. Additionally, the estimated marginal means (EMM) of the difference between inulin and placebo intake are reported on top of the figures together with the 95% confidence interval (CI), and related p-values

In line with the clinical outcomes, the analysis supported the inulin-induced effect on the fecal microbiota (Supplementary Figs. 2 and 3). Bifidobacterium spp. showed a larger increase (p = 0.013, q = 0.192; Supplementary Fig. 3A) after inulin (run-in: 14.8%, period 1: 19.2%; fold-change: 1.3) compared to placebo (run-in: 12.8%, period 1: 10.9%; fold-difference: 0.8), though not statistically significant. Also, the increase in Anaerostipes spp. was higher with 2.5-fold higher relative abundance than placebo (p = 0.001, q = 0.025), while Akkermansia spp. relative abundances showed the opposite behavior, skewed by one individual (Supplementary Fig. 3B). In summary, the analysis of the run-in and first period supported the observed improvements induced by inulin while also highlighting differences with larger observed effects in the clinical and fecal microbiota changes.

Differences between intervention sequences

As we observed a significant intervention*period interaction indicative of apparent carry-over for several outcomes, we investigated how this may be reflected in differences between intervention sequences (Table 2). Both sequences began with comparable run-in values, yet distinct differences emerged during the wash-out phase. Expectedly, these differences were caused by participants who received inulin first (treatment-placebo sequence) and, despite the four-week wash-out period, experienced lasting improvements in stool frequency, consistency, PAC-QOL, PAC-SYM scores (Table 2). These improvements were not linked to changes in physical activity (Table 2). Instead, we observed that the composition of the fecal microbiota, assessed by PCoA of β-diversity, did not return to its run-in composition during wash-out, while that of the placebo-treatment sequence did (Supplementary Fig. 4). Collectively, these differences between intervention sequences suggested a lasting improvement by inulin in the treatment-placebo sequence that persisted beyond the four-week wash-out and was detectable in both sustained clinical bowel habit and fecal microbiota changes.

Table 2.

Stool frequency, consistency and PAC-QOL, and SYM, and physical activity per sequence. Descriptive statistics of each outcome are reported as mean (SEM) if not indicated otherwise

	Treatment-Placebo				Placebo-Treatment
	Run-in	Period 1	Wash-out	Period 2	Run-in	Period 1	Wash-out	Period 2
Stool frequency (x/week) a	2.00 [2.00, 2.12]	4.00 [2.75, 4.50]	2.25 [2.00, 3.00]	3.50 [2.50, 4.50]	2.00 [2.00, 2.25]	2.50 [2.25, 3.25]	2.00 [1.50, 2.50]	3.25 [3.00, 3.50]
Stool consistency (BSFS)	1.92 (0.12)	2.75 (0.18)	2.19 (0.22)	2.88 (0.27)	1.82 (0.11)	2.23 (0.14)	1.97 (0.16)	2.33 (0.16)
PAC-QOL
Total score	2.05 (0.14)	1.26 (0.14)	1.74 (0.17)	1.50 (0.21)	1.89 (0.14)	1.71 (0.15)	1.97 (0.15)	1.57 (0.15)
Physical discomfort	2.37 (0.12)	1.36 (0.16)	1.81 (0.19)	1.55 (0.22)	2.24 (0.19)	1.76 (0.22)	2.21 (0.23)	1.69 (0.21)
Psychosocial discomfort	1.58 (0.19)	0.93 (0.13)	1.25 (0.21)	1.28 (0.25)	1.21 (0.17)	1.14 (0.19)	1.29 (0.21)	1.00 (0.16)
Worries and concerns	1.84 (0.17)	1.16 (0.17)	1.62 (0.20)	1.46 (0.24)	1.72 (0.16)	1.55 (0.17)	1.85 (0.18)	1.46 (0.16)
Satisfaction	3.18 (0.13)	1.95 (0.22)	2.75 (0.17)	1.94 (0.24)	3.34 (0.11)	2.94 (0.14)	3.15 (0.13)	2.60 (0.21)
PAC-SYM
Total score	1.98 (0.12)	1.25 (0.13)	1.54 (0.17)	1.35 (0.19)	1.90 (0.14)	1.56 (0.17)	1.93 (0.16)	1.47 (0.17)
Abdominal symptoms	2.13 (0.12)	1.34 (0.15)	1.52 (0.20)	1.51 (0.21)	1.94 (0.18)	1.64 (0.20)	2.02 (0.22)	1.63 (0.22)
Rectal symptoms	1.25 (0.16)	0.59 (0.12)	0.93 (0.20)	0.82 (0.19)	1.52 (0.15)	1.11 (0.17)	1.26 (0.17)	0.88 (0.16)
Stool related symptoms	2.29 (0.18)	1.58 (0.17)	1.96 (0.20)	1.07 (0.13)	2.12 (0.17)	1.75 (0.20)	2.24 (0.19)	1.09 (0.14)
IPAQ b
“low”	5 (25%)	7 (36.8%)	4 (20%)	6 (30%)	4 (21.1%)	2 (11.1%)	4 (22.2%)	4 (21.1%)
“moderate”	12 (60%)	7 (36.8%)	10 (50%)	7 (35%)	12 (63.2%)	9 (50%)	9 (50%)	11 (57.9%)
“high”	3 (15%)	5 (26.3%)	6 (30%)	7 (35%)	3 (15.8%)	7 (38.9%)	5 (27.8%)	4 (21.1%)

BSFS Bristol Stool Form Scale, IPAQ International physical activity questionnaire, PAC-QOL Patient assessment of constipation – quality of life, PAC-SYM Patient assessment of constipation – symptoms 

^a Values are expressed as median [IQR]

^b Values are expressed as count (%)

Investigating clinical and gut microbiota differences linked to carry-over

To better comprehend the carry-over in the treatment-placebo sequence, we first assessed individual responses (Fig. 5A). We observed a remarkable dichotomy in the primary outcome stool frequency following the wash-out, with half of the participants reporting more than three bowel movements weekly in the second period (placebo intake). Guided by this observation, we investigated whether these subjects (> 3 bowel movements period 2) already showed differences during run-in and in their response to inulin (period 1) compared to the other individuals (< 3 bowel movements period 2) in that sequence. While run-in stool frequency, consistency and PAC-QoL values were similar across participants, these subjects had lower PAC-SYM scores initially and during inulin intake (Supplementary Table 4). Moreover, they had a slightly stronger response to inulin with regard to stool frequency and PAC-SYM scores. Consequently, their improvements persisted throughout the wash-out period, while for the other individuals in this sequence, values returned to run-in levels (Supplementary Table 4).

Fig. 5.

Differences in fecal microbiota associated with apparent carry-over in the treatment-placebo sequence. A Dichotomous stool frequency outcomes following the wash-out during period 2 in participants in the treatment-placebo sequence highlighting how they were apparently affected or unaffected by carry-over. B Run-in common genera (>1% abundance in 50% of the samples) in participants in treatment-placebo sequence apparently affected or unaffected by carryover. C Gut microbiota composition at genus level visualized and assessed by principal coordinate analysis using β-diversity based on Bray-Curtis dissimilarity at run-in, period (1 inulin intake), wash-out and period 2 (placebo intake) for participants in the treatment-placebo sequence apparently affected or unaffected by carryover. The difference between groups was tested using PERMANOVA and corresponding p-values are depicted in the lower part of the figures. Note the decreasing number of samples over time, attributed to the decreasing availability of fecal samples from all participants. D Differences in relative abundance of selected genera between participants apparently affected or unaffected by carry-over. These genera were selected based on their observed inulin-induced modulation in this cross-over trial (Bifidobacterium and Anaerostipes spp.), or because they have been linked to constipation or slow transit time (Methanobrevibacter spp. [14, 67]), or because they were found to statistically significantly differ (Faecalibacterium, Roseburia and Subdoligranulum spp.; see also Supplementary Fig. 5) between subjects apparently affected or unaffected by carry-over

Using the available fecal microbiota data (Fig. 5 and Supplementary Fig. 5), we investigated underlying gut microbiota differences between subjects apparently affected and unaffected by carry-over. PCoA of β-diversity using Bray-Curtis dissimilarity already indicated slight differences at run-in, which became even more apparent in the wash-out samples (Fig. 5C). At run-in, subjects apparently affected by carry-over had statistically significantly higher relative abundances of the butyrate-producers Faecalibacterium spp. and Roseburia spp., a difference that also re-emerged during wash-out (Fig. 5D and Supplementary Fig. 5A-B). Interestingly, those subjects also had lower Bifidobacterium spp. relative levels at run-in than those not affected (0.6-fold difference), and their bifidobacteria appeared to be more responsive to inulin intake. In affected subjects, Bifidobacterium spp. increased by 1.8-fold (from 11.7% to 21.7%), exceeding levels in subjects not affected by carry-over (18.7%). After the four-week wash-out, Bifidobacterium spp. relative abundances returned to run-in levels again. Similarly, Anaerostipes spp. relative abundances appeared to be more responsive in affected subjects and increased by 2.4-fold in subjects affected by carry-over (from 1.4% to 3.4%) compared to 1.7-fold in those not affected (from 1.3% to 2.1%), even though baseline relative abundances did not differ between groups. During wash-out, Anaerostipes spp. relative abundances remained elevated in subjects affected by carry-over (2.0%) while returning to run-in levels in subjects unaffected by carry-over (1.4%).

Additionally, levels of Subdoligranulum spp., another putative butyrate-producer, remained highly elevated during wash-out in subjects affected by carry-over (Fig. 5D and Supplementary Fig. 5B). This genus constituted 10.4% of the fecal microbiota, which was reflected in a 2.4-fold increase over run-in (4.4%) and a 1.7-fold increase over period 1 (inulin intake, 6.1%). In contrast, in subjects not affected by carry-over, Subdoligranulum spp. relative abundances decreased again after inulin intake (run-in: 3.2%; inulin: 5.2%; wash-out: 2.5%). Finally, it is worth noting that also Methanobrevibacter spp., methanogenic archaea reportedly linked to constipation [67] and slow gastrointestinal transit [14], was consistently lower throughout the whole study in subjects affected by carry-over versus those not affected. Collectively, our observations suggest that the apparent carry-over effects are related to baseline microbiota differences, including lower but more inulin-responsive Bifidobacterium spp. and higher relative abundances of butyrate-producing taxa that remained notably elevated after a four-week wash-out.

Discussion

Here, we assessed the impact of four-week daily 12 g inulin intake on stool frequency, stool consistency, constipation-related quality of life, symptoms, and physical activity, as well as fecal microbiota composition in 39 individuals with functional constipation according to Rome III criteria. In the cross-over analysis of the predominantly female study population, we observed that inulin had a positive impact on stool frequency, which improved from once every three to four days to once every second day despite the lack of observed changes in stool consistency. This was concurrent with improvements in constipation-related quality of life and symptoms, notable for the decrease in worries, concerns, psychosocial discomfort, and abdominal symptoms alongside an increase in the relative abundance of genera of known butyrate-producing bacteria. After detecting apparent carry-over in several outcomes, we analyzed the first period only, supporting the positive effect of inulin on bowel habits, quality of life, symptoms, and fecal microbiota.

Our results align with previous studies showing the positive effects of inulin on stool frequency, extending these findings to a clinical application for functional constipation. However, in the clinical context not only an objective outcome such as stool frequency or consistency is of importance. Recent guidelines stress that also the perceived impact of symptoms on a patient’s life - specifically their “bothersomeness”- plays a critical role in diagnosing DGBI [11]. In this study, we evaluated the impact of inulin on quality of life and symptom interference using validated constipation questionnaires. Both the PAC-QOL and PAC-SYM scores showed improvements exceeding the minimum important difference in both the cross-over and first-period analyses [36, 66]. These findings suggest that inulin could be a viable non-pharmacological alternative for managing functional constipation. Notably, despite reported initial adverse events of flatulence and bloating due to increasing inulin fiber, we observed no worsening of PAC-SYM scores, and instead, inulin led to a reduction in abdominal symptoms.

This study was designed as a cross-over trial to account for individual responses to the treatment and placebo, which were expected to be affected by the gut microbiota’s ability to influence the response to dietary stimuli [68, 69]. The adoption of the cross-over design was underpinned by previous cross-over trials that used inulin at a similar dosage (12 g/day), albeit in healthy individuals without impaired bowel function [70], or lower dosages (3 g/day or 7 g/day) in healthy individuals with low fiber intake [71]. However, we could not rule out carry-over bias for stool frequency, stool consistency, PAC-QOL satisfaction, and PAC-SYM rectal and stool-related symptom subscores as observed by the considerable contribution of the intervention*period interaction term for these outcomes. Following recommendations [54], we explored possible causes of carry-over to gain insights from this trial and better understand part of the observed effect while acknowledging inherent limitations. Based on the per-sequence analysis, we observed that carry-over, defined as the effect of the first period persisting into the second period [54], was expectedly caused by subjects receiving inulin first. Notably, these subjects’ fecal microbiota composition did not return to their run-in composition before the start of period 2, indicating a lasting effect of inulin exceeding the four-week wash-out period. Although no guidelines exist on the appropriate wash-out length for human gut microbiota studies [72], a four-week wash-out was estimated to be sufficient, as the summarized evidence of several studies has suggested that the effect of inulin-type fructans on the fecal microbiota progressively disappears after one to two weeks [73]. Our study underscores the need for careful evaluation of cross-over designs and wash-out periods in future research including gut microbiota modulation by dietary fibers. Cross-over trials are useful for highlighting individual differences, but their suitability for disorders like DGBI should be reconsidered. Past recommendations for studying pharmacological treatments in these conditions have favored parallel designs, given the variability of symptoms over time in both adults and children [74–76]. We believe these same considerations should apply to non-pharmacological treatments targeting the gut microbiota.

We observed that improvements in bowel habit outcome were accompanied by an increase in the relative abundances of Bifidobacterium spp. as well as Anaerostipes spp. and Coprococcus 1 spp [20, 77]. Specifically, higher relative abundances of Coprococcus 1 spp. correlated with increased stool frequency, while greater Bifidobacterium spp. relative abundances were linked to reduced PAC-SYM rectal symptom scores. These findings suggest a potential connection between gut microbiota modulation and bowel function, particularly involving taxa related to butyrate production. Inulin is known for its bifidogenic effect, meaning it specifically promotes the growth of bifidobacteria. Bifidobacterium spp. metabolize inulin into acetate and lactate, which can then serve as substrates for butyrate production by other microorganisms through cross-feeding. Anaerostipes spp. is such a taxon capable of lactate-to-butyrate conversion [78, 79], while Coprococcus spp., aside from butyrate, primarily produce propionate [77, 80]. Butyrate-producing bacteria are often reduced in constipated individuals, including those with irritable bowel syndrome with constipation (IBS-C) [17], and butyrate itself plays a well-documented role in gut health by serving as energy source for colonocytes, supporting gut barrier integrity [81], and potentially enhancing motility [82, 83]. Previous studies have associated butyrate-producing taxa, such as Faecalibacterium, Coprococcus, and Roseburia spp., with aspects of improved bowel function, including faster transit [84] and increased fecal water content [85]. Additionally, there was an observed decrease in Ruminococcaceae members, which are typically elevated in individuals with constipation [15, 16, 86]. Although our study did not assess other bowel habit markers, such as fecal metabolites, or water content, an acknowledged limitation, our findings demonstrate that inulin supplementation promotes a beneficial intestinal microbiota composition that contributes to measurable improvements in functional constipation.

These findings highlight the role of taxa responsible for butyrate production in bowel function improvements, prompting further investigation into their potential contribution to carry-over effects. Leveraging available fecal samples for gut microbiota analyses, we explored whether baseline microbiota composition was associated with these effects. Notably, subjects more affected by carry-over had higher initial relative abundances of butyrate producers (Faecalibacterium spp. and Roseburia spp.) and reported lower constipation-related symptoms despite similar run-in stool frequency, consistency, and PAC-QOL scores, suggesting a more favorable gut microbial environment from the onset. Additionally, these subjects had lower Bifidobacterium spp. relative abundances at baseline, yet responded more strongly to inulin supplementation through a stronger increase in Bifidobacterium spp. and Anaerostipes spp., relative abundances, along with a sustained increase in the butyrate-producer Subdoligranulum spp. that persisted through wash-out. This suggests that baseline Bifidobacterium spp. relative abundances and their increased responsiveness may have influenced cross-feeding to Anaerostipes spp. and potentially other butyrate producers, contributing to stronger and sustained clinical responses to inulin. This aligns with prior research showing greater metabolic responsiveness in individuals with lower baseline Bifidobacterium relative abundances when supplemented with inulin-type fructans [87, 88]. It should be noted that only 70% of participants provided fecal samples during wash-out, limiting our ability to conclusively link lasting intestinal microbiota changes to carry-over effects. Nonetheless, the observed differences in butyrate-producing bacteria, combined with previous research linking such taxa with bowel habit improvements, underscore how the intestinal microbiota composition may mediate treatment responses. Moving forward, the initial intestinal microbiota may, in the future, be considered for more personalized treatment strategies for functional constipation.

A key outcome following inulin intake, seemingly unaffected by carry-over effects, was the improvement in PAC-QOL scores, particularly in the psychosocial discomfort and worries and concerns subscores. These improvements were consistent across both the cross-over and first-period analyses and absent during placebo intake, indicating a genuine effect of inulin. Notably, run-in scores for these subscores were slightly higher than reference values, reflecting a greater impact on participants’ quality of life [36]. The psychosocial discomfort subscore encompasses aspects of the condition’s interference with daily life, such as embarrassment, worries related to toilet visits, and considerations regarding routine, appetite, and food intake [36]. The worries and concerns subscore encompasses aspects regarding emotional well-being, including irritability, upset, obsession with the condition, stress, impaired self-confidence, and feelings of control over one’s body and condition [36]. While improved stool frequency certainly contributed to social and emotional well-being, it is unlikely to be the sole driver, given that stool frequency also improved during placebo intake. Instead, the improvement in these subscores may be partly attributed to inulin-induced reductions in abdominal symptoms, including discomfort, bloating, and cramps, as measured with the PAC-SYM. Collectively, the findings suggest that inulin’s modulation of the intestinal environment positively influences the social and emotional aspects of the gut-brain interaction in functional constipation. Inulin-derived butyrate may play a role in these effects, as seen in animal models where butyrate induced neuroplastic changes in the enteric nervous system and influenced vagus nerve activity [89, 90]. Additionally, initial studies have demonstrated inulin’s beneficial impact on neuroimmunological and gut barrier functions in relation to constipation-related depression and anxiety [91]. Future studies on DGBI should integrate assessments of common bowel habit markers like stool frequency and consistency with selected intestinal microbial, neuroimmunological, and barrier function biomarkers (e.g., zonulin [92]) to more comprehensively understand their impact on quality of life, particularly emotional and social well-being.

Our study has several limitations, extending beyond the study design and the lack of fecal samples, microbial metabolites, or other aspects of bowel habits, as previously addressed. First, our study population was primarily female, which raises questions about the generalizability of the results and precludes the inclusion of sex as a covariate or effect modifier, although functional constipation is more common in women [7, 8]). Additionally, we did not assess habitual dietary intake, which may account for some of the baseline variations observed between subjects who were more or less impacted by carry-over effects [93]. Furthermore, stool frequency and consistency were analyzed using Gaussian linear mixed models for interpretability and comparability, acknowledging that this approach does not fully capture their count and ordinal nature, respectively. Finally, due to the strong clinical focus on improvement of functional constipation and the design of the study prior to the re-definition of functional gastrointestinal disorders as disorders of the gut-brain interaction, other assessments of the gut-brain axis were not included at the time of the design.

Despite the challenges posed by the apparent carry-over effects of inulin on statistical evaluation, its lasting impact could offer unique benefits. Unlike pharmacological treatments such as laxatives, which provide quick but transient relief, inulin’s gradual influence on the gut environment might lead to more sustained and broad-reaching improvements. Inulin, as a dietary component, may influence multiple aspects of the intestinal environment simultaneously, including microbiota composition, metabolite production, gut barrier, and immune function. Unlike pharmacological treatments, such as laxatives, these changes are likely to accumulate over time. Therefore, while the effects of inulin may not be immediately apparent, its long-term benefits could become more pronounced with extended use and offer benefits beyond those of laxatives. Future research should explore how inulin’s enduring effects are reflected in specific fecal microbiota signatures and other biomarkers relevant to DGBI, shedding light on its potential for sustained therapeutic impact.

Conclusion

In conclusion, our cross-over study confirmed that inulin is a promising non-pharmacological treatment alternative. It improved constipation-related quality of life, social-emotional well-being as well as abdominal symptoms, and promoted gut health by increasing stool frequency and modifying the fecal microbiota in functional constipation.

Abbreviations

AIC

Akaike Information Criterion

ASVs

Amplicon Sequence Variants

BMI

Body Mass Index

BSFS

Bristol Stool Form Scale

CI

Confidence Interval

DGBI

Disorder of the Gut-Brain Interaction

EFSA

European Food Safety Authority

EMM

Estimated Marginal Mean

FDR

False-Discovery Rate

IBS-C

Irritable bowel syndrome with constipation

IPAQ

International Physical Activity Questionnaire

IQR

Interquartile Range

PAC-QOL

Patient Assessment of Constipation – Quality of Life

PAC-SYM

Patient Assessment of Constipation – Symptoms

PEG

Polyethylene Glycol

PERMANOVA

Permutational Multivariate Analysis of Variance

PCoA

Principal Coordinate Analysis

rRNA

Ribosomal RNA

SD

Standard Deviation

SEM

Standard Error of Mean

STAR

Stool Transport and Recovery

Authors’ contributions

DM, VD, EEV – conception and design; AD, CAMW - data acquisition; M-LP, CAMW, SCCvdZ - data analysis; M-LP, CAMW, SCCvdZ, CB, MB, HS data interpretation; CAMW, M-LP – drafting manuscript; M-LP, SCCvdZ, HS – editing manuscript; M-LP, CAMW, SCCvdZ, VD, AD, DM, CB, EEV, MB, HS - final review and approval.

Funding

This work was financially supported by Sensus (Royal Cosun). Sensus had no role in the final decision about the inclusion or exclusion of records, data extraction, data analysis, data interpretation and risk assessment.

Data availability

The 16S rRNA gene sequencing data are deposited at the European Nucleotide Archive (ENA) under the accession number PRJEB98644. All additional data supporting this study is available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Cork Research Ethics Committee of the Cork Teaching Hospitals (Reference ECM 4 (v) 01/09/15), and the trial was executed in accordance with the principles of the Declaration of Helsinki and in compliance with the International Council for Harmonization Good Clinical Practice. All participants in the study provided written informed consent prior to their participation.

Consent for publication

Not applicable.

Competing interests

Elaine Vaughn and Veerle Dam are current employees, and Diederick Meyer is a former employee of Sensus B.V. (Royal Consun). All other authors declare to have no conflict of interest.