The Effect of Fiber Supplementation on Chronic Constipation in Adults: An Updated Systematic Review and Meta-Analysis of Randomized Controlled Trials

ABSTRACT Background Chronic constipation is a prevalent disorder that remains challenging to treat. Studies suggest increasing fiber intake may improve symptoms, although recommendations on the fiber type, dose, and treatment duration are unclear. Objectives We investigated the effects of fiber supplementation on stool output, gut transit time, symptoms, and quality of life in adults with chronic constipation via a systematic review and meta-analysis of randomized controlled trials (RCTs). Methods Studies were identified using electronic databases, backward citation, and hand searches of abstracts. RCTs reporting administration of fiber supplementation in adults with chronic constipation were included. Risks of bias (RoB) was assessed with the Cochrane RoB 2.0 tool. Results were synthesized using risk ratios (RRs), mean differences, or standardized mean differences (SMDs) and 95% CIs using a random-effects model. Results Sixteen RCTs with 1251 participants were included. Overall, 311 of 473 (66%) participants responded to fiber treatment and 134 of 329 (41%) responded to control treatment [RR: 1.48 (95% CI: 1.17, 1.88; P = 0.001); I2 = 57% (P = 0.007)], with psyllium and pectin having significant effects. A higher response to treatment was apparent in fiber groups compared to control groups irrespective of the treatment duration, but only with higher fiber doses (>10 g/d). Fiber increased stool frequency [SMD: 0.72 (95% CI: 0.36, 1.08; P = 0.0001); I2 = 86% (P < 0.00001)]; psyllium and pectin had significant effects, and improvement was apparent only with higher fiber doses and greater treatment durations (≥4 weeks). Fiber improved stool consistency (SMD: 0.32; 95% CI: 0.18, 0.46; P < 0.0001), particularly with higher fiber doses. Flatulence was higher in fiber groups compared to control groups(SMD: 0.80; 95% CI: 0.47, 1.13; P < 0.00001). Conclusions Fiber supplementation is effective at improving constipation. Particularly, psyllium, doses >10 g/d and treatment durations of at least 4 weeks appear optimal, though caution is needed when interpreting the results due to considerable heterogeneity. These findings provide promising evidence on the optimal type and regime of fiber supplementation, which could be used to standardize recommendations to patients. The protocol for this review is registered at PROSPERO as CRD42020191404.


Introduction
Chronic constipation is a common gastrointestinal disorder with an estimated prevalence of 12% in adults (1). It is characterized by predominant symptoms of infrequent, difficult, or incomplete defecation (2). Chronic constipation impacts quality of life, with impairments in social functioning and mental health (3). In the United States, each year there are more than 2.5 million visits to medical centers where constipation is the leading diagnosis, with annual direct costs of the condition ranging from $1912 to $11,991 per patient (4)(5)(6).
The management of chronic constipation can be challenging. Recommendations include lifestyle advice (e.g., increasing dietary fiber), followed by laxatives (7). However, half of people with constipation are dissatisfied with the current treatment options, primarily because of a lack of effectiveness and side effects (8). This highlights the need for better access to management strategies that can safely and effectively improve constipation symptoms.
Fiber supplementation is indicated as first-line management for chronic constipation in British, American, and European guidelines (7,(9)(10)(11). Fiber encompasses all carbohydrates that are neither digested nor absorbed in the small intestine and have a degree of polymerization of 3 or more monomeric units, plus lignin (12). This includes prebiotic fibers, which are substrates that are selectively utilized by host microorganisms, conferring a health benefit (13). Soluble, viscous fibers can influence stool bulking directly through water retention in the colon, resulting in softer stools. Insoluble, nonviscous fibers can cause mechanical stimulation of the gut mucosa that accelerates gut transit time (GTT) (14)(15)(16). Fermentable fibers may increase stool bulk indirectly through fermentation byproducts that influence contractile activity in the gut and increase microbial biomass through changes in the gut microbiome (14,17). However, fermentation also results in the production of gas, which can cause side effects, such as flatulence or bloating.
A meta-analysis published in 2016 indicated beneficial effects of fiber supplementation at improving stool frequency and consistency in adults with chronic constipation, although its consumption led to side effects, such as flatulence (18). The review included only 7 randomized controlled trials (RCTs), limiting the interpretation of the overall effectiveness and subgroup analyses based upon the types of fiber, doses, and treatment durations, rendering its application to clinical practice challenging. Many studies have since been performed evaluating a range of types and doses of fiber and measuring additional outcomes. The aim of this systematic review and meta-analysis was to investigate the effects of fiber supplementation on response to treatment, stool output, GTT, symptoms, quality of life, adverse events, and compliance in adults with chronic constipation.

Methods
This systematic review and meta-analysis was conducted following the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions (19) and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) updated guidelines (20). The eligibility criteria, search strategy, and methods of screening, data extraction, and data analysis were specified in advance and detailed in a protocol published on PROSPERO (CRD42020191404).

Eligibility criteria
The eligibility criteria were developed using a Patient, Intervention, Comparators, Outcome, and Study Design (PICOS) approach and are outlined in Table 1. Briefly, the inclusion criteria were any RCTs reporting the effects of fiber supplementation in adults with chronic constipation that measured constipation outcomes.

Search strategy
Studies were identified through a systematic search of electronic databases and a clinical trials register, a hand search of conference abstracts, and a back search of reference lists of eligible studies and relevant review papers.
The following databases were searched for eligible studies: MEDLINE (1946

Selection process
References were imported into a reference manager for assessment of eligibility (EndNote X9; Thomson Reuters). Following automatic and manual removal of duplicates, 2 reviewers (AvdS, CD) independently screened the titles and abstracts and then the full-text articles against the predefined inclusion and exclusion criteria. Disagreements were resolved by a third reviewer (ED).

Data collection process
Two reviewers (AvdS, CD) independently extracted data from eligible studies onto a standardized form. The data extracted included the characteristics of the trial participants, the intervention, the comparator group, the outcomes measured, and the study design (Table 1). Extracted data were compared and discrepancies were resolved. Where a full paper, abstract, or trial registration provided insufficient data, the authors were contacted to provide additional information. When trial reporting allowed, data were extracted as intention-to-treat analyses. For dichotomous data, dropouts were assumed to be treatment failures. If this information was not clear in the paper, an analysis on all participants with reported evaluable data was undertaken. Where a study reported multiple measurement methods for an outcome [e.g., Patient Assessment of Constipation Symptoms (PAC-SYM) and Cleveland Clinic Constipation Score (CCCS) (21,22)], the data for the most frequently reported method across studies were selected for the meta-analysis. When necessary, data were estimated from figures using Plot Digitizer Version 2.6.9, as recommended by the Cochrane Handbook (19). Name of study product and ingredients; fiber type, degree of polymerization, form and dose; and schedule and duration of intervention

Comparators
Studies using an appropriate placebo as a control that allowed the effect of the fiber alone to be isolated. Where the fiber intervention was a fortified food or drink, an appropriate comparator was the same food or drink without the fiber. Studies that contained multiple study arms were included if the fiber and control arms could be isolated. Control interventions could contain a negligible amount of fiber (<0.5 g/d using the Englyst method or <0.67 g/d using the Association of Analytical Chemists method).
Type, form, dose, schedule, and duration of comparator Outcomes Studies reporting either dichotomous or continuous data on treatment success; stool frequency, consistency, and weight; whole and regional gut transit time; gastrointestinal symptoms (integrative symptom scores, frequency and severity of individual symptoms); quality of life; laxative use; adverse events; and compliance.
Outcomes; method of measurement; baseline, midpoint and endpoint values or change from baseline; and details of adverse events and compliance Study design Randomized controlled trials with ≥2 study groups where it was possible to extract data on the fiber and control interventions. Parallel-group and cross-over studies with a washout period of ≥2 weeks were eligible. Cross-over studies without an adequate washout period were only considered eligible if the data from the first period could be extracted to reduce the risk of a carryover effect.
Study design; washout period duration; intention-to-treat analysis; number of excluded participants and reasons for exclusion; number of centers; randomization method, allocation concealment, and blinding; and funding source, funder involvement, and conflicts of interest

Risk of bias assessment
Risks of bias (RoB) for eligible studies was assessed independently by 2 reviewers (AvdS, CD) (23). The Cochrane RoB 2.0 tool assesses 5 domains consisting of bias arising from: randomization; deviations from intended interventions; missing outcome data; outcome measurements; and selection of the reported results. Judgements were categorized as having a low risk, some concerns, or a high risk, and reasoning for each judgement was recorded. The highest level of bias across the individual domains determined the overall RoB for each outcome of each included study. Disagreements were resolved through discussion with a third reviewer (ED). Information in available clinical trial registrations and protocols was compared against each final publication to ensure that prespecified outcomes matched those in the final publication.

Data synthesis
A meta-analysis was performed when data for the same outcome from 2 or more studies were obtained, using Review Manager Version 5.4 (The Cochrane Collaboration, 2020). When a study reported multiple timepoints, endpoint data were used for the meta-analysis. Dichotomous outcomes were expressed as risk ratios (RRs) and 95% CIs. Mean differences (MDs) were calculated for continuous outcomes that were measured using the same tool and reported in the same units, or where the same unit could be calculated through direct conversion (e.g., bowel movements per day converted to per week). Standardized mean differences (SMDs) were calculated for continuous outcomes that were measured or reported differently. Means and SDs, sample sizes, and P values were used for the analysis. Where medians and ranges were reported, means and SDs were estimated using guidance from Wan et al. (24). For studies that had multiple intervention arms administering different fiber doses, each dose was compared to the control separately (21,22,25), and the sample size of the control group was divided by the number of intervention arms to reduce the unit-of-analysis error (19).
Meta-analyses were performed using a random-effects model. Statistical heterogeneity was assessed using the χ 2 test and quantified using the I 2 statistic. Thresholds of 50% and 75% were considered to represent substantial and considerable heterogeneity, respectively. Subgroup analyses were performed to investigate heterogeneity and explore the effects of fiber type, prebiotic status, treatment dose, and treatment duration. For subgroup analyses, a P value < 0.1 was considered to be statistically significant (26). To assess for publication bias, funnel plots were generated where a meta-analysis included ≥10 studies and symmetry was identified by visual inspection (19).

Outcomes
The outcomes of the meta-analysis are reported in Table 3.
Forest plots for subgroup analyses on fiber type, prebiotic status, doses, and treatment durations for each outcome are found in Supplemental Figures 1-22.

Stool consistency
Stool consistency was measured in 11 studies; however, 1 study did not report data, and these data were not obtained on request (36). Ten studies including 918 participants were included in the meta-analysis (21, 22, 25, 28-30, 33-35, 37). The methods of measuring stool consistency varied, with most studies using the Bristol Stool Form Scale or a modified version of it. Fiber significantly improved stool consistency compared to control (SMD: 0.32; 95% CI: 0.18, 0.46; P < 0.0001), and   Values are for the whole study population unless groups are specified. 1 Data were meta-analyzed using a random effects model. Statistical heterogeneity was assessed using the χ 2 test and quantified using the I 2 statistic. GTT, gut transit time; MD, mean difference; PAC-SYM, Patient Assessment of Constipation Symptoms; PAC-QoL, Patient Assessment of Constipation Quality of Life; Refs, references; RR, risk ratio; SMD, standardized mean difference.
Regional GTT was reported in 2 studies (21,29). Fiber had no significant impact on right, left, or rectosigmoid transit time compared to control (Supplemental Figure 7).

FIGURE 4
Forest plot of stool consistency in randomized controlled trials comparing fiber with control groups in adults with chronic constipation (n = 918). Values were calculated as standardized mean differences (95% CIs) using a random-effects model. IV, inverse variance.
A sensitivity analysis was conducted for studies that measured and reported integrative symptom scores using the PAC-SYM questionnaire, and a MD was calculated. In 3 studies including 428 participants (21,22,25), fiber had no impact on PAC-SYM global scores [MD: −0.10 scale points (95% CI: −0.31, 0.11 scale points; P = 0.35); I 2 = 48% (P = 0.07); a lower score denotes less severe symptoms]. In addition, fiber had no impact on the abdominal, rectal, or stool subscales of PAC-SYM (Supplemental Figures 9-12). All 3 studies measured 1 other integrative symptom score-2 measured the CCCS (21, 22) and 1 measured the Bowel Function Index (25)-and reported no difference between intervention and control groups.

Individual gastrointestinal symptoms
The effect of fiber supplementation on the severity of individual symptoms was widely reported in various studies (Supplemental Figures 13-17).
The frequency of individual symptoms was less commonly reported amongst studies, and suitable data could not be obtained for more than 1 study; therefore, a meta-analysis was not performed.
Several other individual symptoms were reported in the same way by only 1 study each and, therefore, meta-analyses were not possible. These outcomes were occurrence of hard stools (31,35,40), pain with defecation (31,28,29), sensation of anal obstruction (31,35,40), use of manual maneuvers (31,35,40), defecation difficulties (33,38), ease of passage (37), time spent on the toilet (38), and borborygmi (33). Of note, in 1 study administering an unspecified type of fiber, the frequency of sensation of anal obstruction was significantly lower in the fiber group compared to the control group (P = 0.032) (40). In a study administering pectin, the time spent on the toilet was significantly lower in the fiber group compared to the control group(P < 0.05) (38).

Quality of life
Two studies including 296 participants reported overall quality of life (QoL) measurements using the Patient Assessment of Constipation Quality of Life (PAC-QoL) global score (21,25). Both studies administered polydextrose and found that it did not impact QoL scores (MD: −0.04 scale points; 95% CI: −0.19, 0.10 scale points; P = 0.54; a higher score indicates more severe), and heterogeneity was not significant (I 2 = 37%; P = 0.17). Fiber doses higher or lower than 10 g/d had no impact on PAC-QoL. Greater treatment durations (≥4 weeks) significantly improved global scores [MD: −0.25 scale points (95% CI: −0.48, −0.03 scale points; P = 0.03); I 2 = 0% (P = 0.66)], while shorter durations (<4 weeks) did not, with a significant subgroup effect (P = 0.03). However, these data derive from 2 intervention groups of the same study (Supplemental Figure 18). Fiber had no impact on QoL subscales for satisfaction, physical discomfort, worries and concerns, or psychosocial discomfort (Supplemental Figures  19-22).

Laxative use
Two studies measured laxative use (21,38); however, data were only reported for 1 study. In this study, pectin supplementation significantly reduced the number of days participants took laxatives compared to control [mean: 1.4 days (SD: 1.0 days) compared with 1.9 days (SD: 1.2 days), respectively; P < 0.01] (38).

Adverse events
Adverse events were reported in 12 studies including 1089 participants (21, 22, 25, 28-31, 34, 37-40). In 1 study administering polydextrose, 18 events were considered to be related to the intervention and 10 events were considered to be related to the control treatment. These were all mild to moderate, except for 1 event in the polydextrose group that was severe (nausea) (25). Another study administering polydextrose reported events such as abdominal pain, but there was no link identified between treatment groups and the occurrence of adverse events (21). One study administering psyllium reported that 2 participants (18%) in the intervention group had abdominal pain, compared to none in the control group (28). Another study administering psyllium reported that 3 participants withdrew from the intervention group and 5 from the control group, for reasons related to the reported treatment side effects (30). A study administering wheat bran reported participant withdrawal due to "unbearable" abdominal pain (39). Two studies reported that while adverse events did occur, none interfered with the trial process (34,40). Finally, 2 studies administering psyllium (29,31), 1 administering pectin (38), and 2 administering GOS (22,37) reported no serious adverse events associated with the fiber products.

Compliance
Compliance was measured in 7 studies (21,25,30,34,36,39,40), but reported in only 3 (25,39,40). Of those that counted returned sachets, 1 study reported >98% compliance (25) and another reported >95% compliance in each group (40). One study reported that compliance was equally satisfactory in both groups; however, no data were provided (39). Another 2 of the 7 studies reported the numbers of participants excluded for noncompliance only: in the first RCT, 3 participants were excluded in the fiber group and none were excluded in the control group (21); and in the other, 4 participants were excluded, but the group they belonged to was not dislosed (34).

Risk of bias at the outcome and study levels
No study was at low RoB across all domains. Bias due to the randomization process was judged to have low risk in 3 studies (25,31,34) and some concerns in 13 studies due to inadequate details on the methods used in the generation of the randomization sequence or for concealment of allocation (21, 22, 27-30, 32, 33, 35-40). Bias due to deviations from the intended interventions was judged to have low risk in 6 studies (21,22,25,29,37,38), some concerns in 2 studies (28,32), and high risk in 8 studies (30, 31, 33-36, 39, 40); as although all but 2 studies were double-blinded, many did not follow an intention-to-treat analysis. Bias across other domains varied at the outcome level (Supplemental Figures 1-22A).

Discussion
This systematic review and meta-analysis shows that fiber supplementation is effective in the management of chronic constipation and, most importantly, identifies the optimal types of fiber, doses, and treatment durations. Psyllium supplements and fiber doses greater than 10 g/d were most effective at improving response to treatment, stool output, and straining, while treatment durations of 4 weeks or more were optimal for improving stool frequency and whole GTT.
Response to treatment was higher following fiber supplementation, with fiber increasing the likelihood of a response by 48% compared to control treatment. The type (psyllium) and doses (>10 g/d) of fiber were crucial in achieving response to treatment. Notably, just over half of the 16 included studies measured response to treatment as a binary outcome, despite recommendations by the Rome committee, and definitions of response were inconsistent across studies (41). The placebo response rate was high (41%), which is a common feature in functional bowel disorders, although the rate is higher here than in a recent meta-analysis of 73 RCTs in irritable bowel syndrome (34%) (42).
This systematic review identified fiber as having a moderate to large effect in increasing stool frequency. Psyllium led to an increase of 3 bowel movements/week, indicating that psyllium is as effective or even more effective than osmotic and stimulant laxatives, which increase stool frequencies by 2.5 bowel movements/week (43). Since normal stool frequency is 3 to 21 bowel movements/week, an increase of 3 bowel movements/week is clinically significant and could normalize stool frequency in constipation (44). Thus, psyllium may be an appropriate first-line intervention in chronic constipation prior to pharmacological therapies. Pectin also increased stool frequency; however, only 1 study contributed to this subgroup analysis.
Stool consistency was moderately improved upon fiber supplementation. Psyllium had the greatest effect on this outcome, with inulin-type fructans also modestly softening stool consistency. The effects of psyllium on stool frequency and consistency are expected, as this fiber has a high water-holding capacity that is resistant to fermentation and forms a viscoelastic substance in the gastrointestinal tract, thus softening stools (14,45,46). The beneficial effect of inulin-type fructans on stool consistency in chronic constipation is in line with the findings of a previous meta-analysis (47). The fermentation of inulin leads to microbiota modifications and an increase in SCFA production, resulting in an increased bacterial mass, stimulation of the enteric nervous system, and a higher water content of digesta (17). Funnel plot asymmetry was detected for stool consistency, and publication bias is likely, as 1 study did not report findings (36). Another explanation could be that the study with the largest effect size used an nonvalidated measurement tool, potentially inflating the effect size (37).
An improvement in whole GTT by 18 hours was found only when fiber was administered for a greater duration. The study that exhibited the largest effect used wheat bran (39). It is likely that the decrease in GTT with wheat bran supplementation results from modification of the gut microbiota composition and its metabolites, ultimately regulating the intestinal contractile activity (48).
Severity of straining improved with fiber supplementation, with psyllium and high fiber doses being particularly effective. However, integrative symptom scores were not affected, nor were severity of incomplete evacuation, abdominal pain or discomfort, or bloating. This suggests that fiber supplementation may be more suitable to specific cohorts of individuals who experience certain symptoms, such as infrequent bowel movements, hard stools, and straining, but not other symptoms, such as incomplete evacuation or bloating.
QoL scores were measured in studies administering polydextrose. However, QoL scores improved by only 0.25 points when polydextrose was administered for 4 weeks, which is considerably less improvement than the recommended minimal clinically meaningful difference of a 0.5-to 1-point change in PAC-QoL; therefore, it is unlikely that this will have a noticeable impact on patients' QoL scores (49,50). Furthermore, polydextrose was not effective at improving other patientimportant outcomes, such as stool output or symptom scores.
In general, fiber was a safe intervention, although adverse events were reported. Severity of flatulence was significantly higher with fiber compared to control treatment. This is an expected side effect, as 2 of the 3 included studies, contributing to 89.9% of weight in the meta-analysis, were from inulin-type fructans, which are highly fermentable fibers that lead to microbiotamediated gas production (45). This finding agrees with a previous meta-analysis showing that fiber increases flatulence (18). One study reported abdominal pain leading to withdrawal from a participant receiving wheat bran. These adverse effects highlight the importance of selecting the appropriate type of fiber and dose regime, both for optimum effectiveness and for the prevention of adverse effects. Psyllium supplements should be used in favor of wheat bran, and individuals should gradually increase their fiber intakes to prevent sudden symptom exacerbation (9).
The findings of this review have important implications. National guidelines for chronic constipation suggest increasing fiber intake, but there is a lack of detail on the types, doses, and durations of fiber supplementation (7,(9)(10)(11). This may result in nonstandardized recommendations in clinical care, with patients left to choose the type and regime of fiber supplements, likely leading to dissatisfaction (8,51). This review provides important evidence on the optimal fiber supplementation, which can support clinicians in providing standardized and effective recommendations to patients with chronic constipation. A previous systematic review showed a dose-dependent response for fiber doses >15 g/d (18). The current review shows that a smaller cutoff of >10 g/d is still effective; this dose may be more realistic to ingest and more tolerable than doses >15 g/d, potentially increasing compliance and satisfaction with treatment.
This is the largest systematic review and meta-analysis to investigate the effect of fiber supplementation in chronic constipation. It identified more than double the number of trials than the most recent systematic review (18), thus allowing for more robust subgroup analyses of specific fibers, doses, and treatment durations. This review strictly adhered to recommendations from the Cochrane Handbook and PRISMA guidelines. Efforts were made to search gray literature with no language restrictions, reducing publication bias. Limitations of this review include significant heterogeneity amongst outcomes, explained by the types of fiber and differences in the methods used to measure outcomes. While included studies provided further insight into the effects of certain fiber types in chronic constipation (e.g., psyllium), for other types (e.g., pectin), only single studies contributed to the subanalyses; therefore, how effective these fiber types might be remains unclear. No studies in this review had a low risk of bias.
In conclusion, this meta-analysis provides evidence that psyllium is the most efficacious investigated fiber at providing constipation relief, with improvements in stool frequency and severity of straining, which highlights psyllium's potential to be used as a first-line strategy for the management of constipation. Doses >10 g/d and durations of at least 4 weeks appear to be optimal for several constipation symptoms. However, there was a low number of studies and substantial heterogeneity between studies in the subgroup analyses. This limits confidence in the interpretations of the findings of the subgroup analyses regarding the types and regimes of fiber supplementation (26). It was also found that fiber may lead to increased flatulence, highlighting the need to address this with patients and apply strategies to mitigate this. These findings provide evidence for the optimal types and regimes of fiber supplementation that could impact clinical care and recommendations to patients, ultimately improving their care and treatment response.
We thank Costanza Stocchi, Patricia Sanz Morales, Miao Yun and Dr Curie Kim for their assistance with translation of foreign language articles. We also thank Dr Peter Duncan, Dr Arthur Ouwehand, Dr Stephan Theis and Xiushan Dong for providing additional information on their studies.
The authors' responsibilities were as follows -AvdS, KW, ED: designed the systematic review protocol and interpreted the data; AvdS: conducted the search and the meta-analysis and wrote the initial manuscript; AvdS, CD: independently reviewed studies against the inclusion and exclusion criteria, extracted data, and performed the risk-of-bias assessment; CD, KW, ED: critically revised the manuscript; and all authors: edited the manuscript and read and approved the final manuscript. ED has received an education grant from Alpro; has received research funding from the British Dietetic Association, Almond Board of California, the International Nut and Dried Fruit Council, and Nestec Ltd.; and has served as a consultant for Puratos. KW has received research funding from government bodies, including the National Institute of Health Research and Medical Research Council; has received funding from charities, including Crohn's and ColitisUK, ForCrohns, The Leona M. and Harry B. Helmsley Charitable Trust, and the Kenneth Rainin Foundation; has received funding from industrial sources, including the Almond Board of California, Clasado Biosciences, Danone, and the International Dried Fruit and Nut Council; and is the coinventor of a mobile application to support people following dietary restrictions (FoodMaestro) and biomarkers to detect irritable bowel syndrome and markers of response. All other authors report no conflicts of interest.

Data Availability
Data described in the manuscript, code book, and analytic code will be made available upon request pending application and approval by the corresponding author.