Clinical trial: the effects of a trans-galactooligosaccharide prebiotic on faecal microbiota and symptoms in irritable bowel syndrome



This article is corrected by:

  1. Errata: Corrigenda Volume 34, Issue 11-12, 1353–1354, Article first published online: 16 November 2011

Prof. D. B. A. Silk, Departments of Academic Surgery, Imperial College Healthcare NHS Trust, London, UK


Background  Gut microflora-mucosal interactions may be involved in the pathogenesis of irritable bowel syndrome (IBS).

Aim  To investigate the efficacy of a novel prebiotic trans-galactooligosaccharide in changing the colonic microflora and improve the symptoms in IBS sufferers.

Methods  In all, 44 patients with Rome II positive IBS completed a 12-week single centre parallel crossover controlled clinical trial. Patients were randomized to receive either 3.5 g/d prebiotic, 7 g/d prebiotic or 7 g/d placebo. IBS symptoms were monitored weekly and scored according to a 7-point Likert scale. Changes in faecal microflora, stool frequency and form (Bristol stool scale) subjective global assessment (SGA), anxiety and depression and QOL scores were also monitored.

Results  The prebiotic significantly enhanced faecal bifidobacteria (3.5 g/d P < 0.005; 7 g/d P < 0.001). Placebo was without effect on the clinical parameters monitored, while the prebiotic at 3.5 g/d significantly changed stool consistency (P < 0.05), improved flatulence (P < 0.05) bloating (P < 0.05), composite score of symptoms (P < 0.05) and SGA (P < 0.05). The prebiotic at 7 g/d significantly improved SGA (P < 0.05) and anxiety scores (P < 0.05).

Conclusion  The galactooligosaccharide acted as a prebiotic in specifically stimulating gut bifidobacteria in IBS patients and is effective in alleviating symptoms. These findings suggest that the prebiotic has potential as a therapeutic agent in IBS.


Irritable bowel syndrome (IBS) is the most common functional gastrointestinal disorder. Depending on the criteria used to define IBS, its prevalence in the general adult population is up to 20%.1, 2 In many sufferers, it is a chronic disorder3 and patients experience significant impairment in health related quality of life.4 The contribution of IBS to annual healthcare costs is high5 and patients also lose significant amounts of time off work.3, 6 The diagnosis of IBS has importance therefore in both human and economic terms.7 IBS continues to represent a significant therapeutic challenge and based on what is known of its pathophysiology, multidisciplinary approaches have been proposed.7, 8 Current interest has focused on the role of gut microflora-mucosa interactions linked to inflammatory and immune processes that are likely to contribute towards enteric neuromuscular dysfunction in these patients.9, 10 The development of IBS following bacteriologically proven gastroenteritis,11–15 the presence of low grade mucosal inflammation,15, 16 endoendocrine cell hyperplasia,15, 16 abnormal fermentation17, 18 and evidence of small bowel bacterial overgrowth in IBS19 are some of the observations that have pointed to a pivotal role for indigenous microbiota in the pathogenesis of IBS. Probiotics are defined as live microbes that confer a significant benefit to the host.20 The results of early studies of probiotic therapy in IBS produced conflicting results.7 Methodological differences and variability of probiotic strains were factors that contributed to the initial scepticism of this therapeutic approach. In a recent and important pilot study, treatment with Bifidobacterium infantis 35624 but not Lactobacillus salivarius UC4331 resulted in symptom improvement and correction of a pro-inflammatory, Th-1 IL-10/IL-12 ratio.21 This study was followed by a large scale trial that confirmed the benefits of therapy with Bifidobacterium infantis 35624 in primary care female IBS patients.22 Subsequently, the beneficial effects of another Bifidobacterium probiotic strain (Bifidobacterium animalis DN-173010) have been shown in IBS adult subjects with predominant constipation.23

Probiotics suffer the disadvantage of degradation during transit through the gastrointestinal tract and the need for stringent quality controls during and after manufacture. In contrast, prebiotics are nondigestible food ingredients that beneficially affect the host by selectively stimulating the growth and/or activity of bacteria already resident in the colon.24 As such, they exert positive changes in the resident gut microbiota.

In response to the findings pointing to the benefits of Bifidobacterium probiotic strains, a novel prebiotic, composed of a trans-galactooligosaccharide mixture B-GOS has been synthesized (Clasado, Milton Keynes, UK). Its capacity to act as a bifidogenic prebiotic was recently confirmed in healthy humans.25 Using this compound, we report here the results of a pilot study undertaken to investigate the efficacy of a prebiotic as therapy in patients with IBS.

Patients and methods

Study population

A total of 42 subjects were required for an 80% probability that the study would detect, at a two sided 5% significance level, an effect on the colonic bifidobacterial population if the true change in the bifidobacterial population was 0.495 times the standard deviation. Patients were recruited from gastroenterology clinics at Central Middlesex Hospital (CMH) and by direct advertisement in the Irritable Bowel Syndrome Network publication. Individuals aged between 18 and 80 years who satisfied Rome II criteria for diagnosis of IBS26 and in whom organic gastrointestinal disease, including inflammatory bowel disease, had been excluded were considered for inclusion in the study. Patients were excluded if they had a functional disorder of the upper gastrointestinal tract for which treatment had not been stable for the preceding three months. All patients were excluded if they had abnormal haematological and biochemical indices and excluded, if they had abnormal findings on barium enema or colonoscopy undertaken within the previous 5 years.

Patients were also excluded, if they had been ingesting products containing prebiotics or probiotics in the 2 weeks preceding entry into the trial. Ethics Committee approval was received for the study which has been registered with ISRCTN (ISRCTN54052375).

Trial protocol

This was a single centre, parallel, patient blinded, randomized cross over controlled trial. Each potentially eligible patient was invited to attend CMH to discuss the study in detail (Visit 1). All patients were given an information sheet and allowed 72 h to decide whether or not to take part (Figure 1). Prior to enrolment in the study, patients were required to read, sign and date a consent form. General practitioners of patients recruited from the IBS Network were contacted to confirm suitability. All patients were evaluated by a full review of clinical history (Visit 2) and based on clinical symptoms, the patients were categorized to one of the following groups (Figure 1); diarrhoea predominant (IBS-D), constipation predominant (IBS-C) or alternating (IBS-A) the latter patients having symptoms of both diarrhoea and constipation. Eligible subjects then entered a 2-week baseline period (Visit 3). At this visit, the investigator recorded symptoms and completed a subjective global assessment (SGA) on the Case Report Form (CRF). At the end of the baseline period, bowel symptoms were re-assessed to confirm that IBS symptoms remained unchanged. During this time and throughout the rest of the study, subjects were instructed not to take any IBS treatment. Patients fulfilling any one of the exclusion criteria were excluded from the study at this stage. At the end of the baseline period, subjects were randomized to one of three treatment groups. Randomization was performed by stratified block randomization within each IBS symptom subgroup. Trial design consisted of a 2-week ‘baseline’ period followed by 2 treatment periods of 4 weeks each, separated by a 2-week ‘wash out’ phase.

Figure 1.

 Summary of trial procedures.

During the first treatment period, patients were asked to ingest once daily, before breakfast, a drink containing either 3.5 g (Group I) or 7.0 g (Groups II and III) placebo (chocolate or banana flavoured). After the 2 week wash out period, patients were then asked to drink once daily, before breakfast, either 3.5 g (Group I), 7.0 g (Group II) B-GOS or 7.0 g placebo (Group III) (chocolate or banana flavoured). At 7 day intervals during the 3 month study period, patients were contacted over the telephone and symptoms and SGA of relief were recorded. Compliance was assessed by direct questioning every week over the telephone.

Prebiotic preparation

The trans-galactooligosaccharide mixture used in this study was produced from the activity of galactosyltransferases from Bifidobacterium bifidum NCIMB 41171 on lactose.25 The final product was in dry powder form consisting of (wt:wt) 48% galactooligosaccharides with a degree of polymerization between 2–5, 22% lactose 18% glucose and 12% galactose. Based on previous studies, the bifidogenic properties of B-GOS have been established for daily intake levels of 3.5 and 7.0 g25 and as the aim of the intervention was to increase the bifidobacterial numbers in the colonic microfora, those intake levels were chosen for this study as well. Placebo (maltodextrins DE 20) and B-GOS samples were packaged in a double blind fashion and placed in boxes, A (3.5 g chocolate or banana flavoured placebo), B (7.0 g chocolate or banana flavoured placebo), C (3.5 g chocolate or banana flavoured B-GOS), D (7.0 g chocolate or banana B-GOS) and delivered to the study centre. Subjects were asked to reconstitute sachets immediately before consumption by mixing the powder with water.


Throughout the entire study, subjects were contacted on a weekly basis and symptoms recorded on the CRF. Bowel movement frequency was recorded as number per day and consistency was scored on a 7-point scale and evaluated using the Bristol stool scale.27 The following 3 cardinal IBS symptom clusters were assessed:

  • (i) abdominal pain or discomfort.
  • (ii) bloating or distension.
  • (iii) bowel movement difficulty.

The latter could reflect difficulty with evacuation (e.g. straining or sense of incomplete evacuation) or urgency. Each symptom was evaluated using an ordinal scale (Likert scale: maximum score 7).28 A composite score that comprised the sums of the three cardinal symptoms (pain/discomfort, bloating/distension and the bowel movement difficulty score) was also calculated for each patient (maximum score 21). The subjective global assessment (SGA) of relief was recorded at weekly intervals during the course of the study. SGA were scored from 1–5 (1 = completely relieved, 2 = considerably relieved, 3 = somewhat relieved, 4 = unchanged and 5 = worse).

Stool samples for faecal flora analysis were obtained at baseline, at the end of both treatments and during the washout period. Patients collected samples in a vessel provided. Each faecal sample was transferred to a vial and 1:1 (weight:weight) glycerol was added. Samples were stored at −20 °C until required for analysis.

Bacterial enumeration

Faecal homogenates were subjected to fluorescent in situ hybridization using synthetic oligonucleotide probes, targeting specific regions of the 16S rRNA molecule and labelled with fluorescent dye Cy3 as described by Rycroft et al.29 Briefly, samples were fixed overnight (4 °C) in 4% (w/v) paraformaldehyde. Fixed cells were centrifuged at 1500 g for 5 min and washed twice in 1 mL filtered PBS. The washed cells were re-suspended in 150 μL PBS and stored in ethanol (1:1 v/v) at −20 °C. Following overnight hybridization with each probe, the fixed cells were washed and vacuum filtered (2 μm polycarbonate isopore membrane filter, Millipore UK Ltd, Watford, UK). They were then mounted onto a glass slide of 20 μm of slowfade (Molecular Probes, Leiden, The Netherlands) and enumerated using the Fluor 100 lens (Eclipse 400 Epifluorescent Microscope, Nikon, Kingston-upon-Thames, UK). The probes used were Bif164 for Bifidobacterium genus.30 His150 for Clostridium perfringenshistolyticum subgroup,31 Bac303 for Bacteriodes-Prevotella32 Lab158 for Lactobacillus – Enterococcus spp.,33 Erec482 for Eubacterium rectale/C.coccoides group,31 Ec1531 for E. coli34 and Srb687 for Desulfovibrio spp..35 4,6-diamidino-2-phenylindole (DAPI) was used for the enumeration of total cell counts. All probes were provided by MWG-Biotech (London, UK).

Quality of life and psychological assessments

The impact of IBS symptoms on quality of life was assessed by administration of an IBS specific questionnaire developed and validated by Drossman et al.36 at each study phase. The IBS-36 questionnaire scores symptoms on a 7-point Likert scale where 0 = never and 6 = always. The highest possible score is 216 and the lowest 0. Anxiety and depression were also measured using the Hospital Anxiety and Depression (HAD) scale at each study phase. This scale detects states of anxiety and depression. The authors provided score ranges for ‘non-case’, ‘doubtful’ cases, and ‘definite’ case of anxiety and depression.37 With either of the HAD subscales, a score above 11 indicated definite clinically significant anxiety or depression.

Dietary intake was assessed and recorded for 5 days at each trial phase throughout the study. Each patient was instructed not to alter his/her diet during the treatment. Fibre intake including soluble and insoluble fibre (measured in grams) was calculated using a dietary programme.38

Safety assessments

Safety assessment consisted of monitoring and recording all adverse events. This information was reported by the patient or discovered by investigator questioning. Each adverse event was described by:

  • (i) duration (start and end dates).
  • (ii) severity grade (mild, moderate, severe).
  • (iii) relationship to the study treatment (suspected, not suspected).
  • (iv) action(s) taken.

Statistical methods

All statistical tests were performed using spss version 15.0 (SPSS Inc., Chicago, IL, USA) and a value of P < 0.05 was taken to indicate statistical significance. All data were analysed by an analysis of variance (anova) model of repeated measurements taking into account the crossover design. In the anova model, volunteer, treatment period (Baseline week 2, end of Placebo week 6 and end of treatment week 12) and treatment were introduced as fixed effects and subject measurements as random effect. Significant differences between treatments were determined by Tukey’s Honestly Significant Difference test.

Role of the funding source

The funding source, Clasado Limited, had no role in study design, collection, analysis and interpretation of the data or in the writing of the report. The decisions to submit the paper for publication was made by the authors.



A total of 153 subjects were invited to participate in the study. A total of 60 subjects consented and were enrolled into the study. Sixteen of the enrolled patients did not complete the study. Figure 2 summarizes the reasons for patient dropouts. Forty four subjects completed all phases of the study.

Figure 2.

 Flow diagram of participants.

Baseline characteristics

Among the 44 subjects, 28 were female and 16 were male and as shown in Table 1, there was a female predominance in each treatment group. Subjects averaged 54 years in age (range, 20–79 years).

Table 1.   Gender and age of IBS patients recruited into the study
Gender and ageIBS-D (n = 23)IBS-C (n = 9)IBS-A (n = 12)
Female (F)9F14M8F1M11F1M
Male (M)
Age mean444742615661
(Standard Deviation)(13.8)(12.6)(9.8) (12.0) 

Categorization of subjects at baseline indicated that 23 (52%) had IBS-D, 9 (21%) IBS-C and 12 (27%) IBS-A. The gender and age of the patients are shown in Table 1. Sixteen of 23 patients with IBS-D were randomized to receive prebiotic following placebo (3.5 g n = 8; 7.0 g n = 8). Seven patients were randomized to receive placebo for both treatment periods. Seven of 9 patients in the IBS-C group received prebiotic (3.5 g n = 4; 7.0 g n = 3). Two patients were randomized to the placebo group. Seven of 12 patients in the IBS-A group received prebiotic (3.5 g n = 4; 7.0 g n = 3) and five were randomized to the placebo group.


Total bacterial counts and individual bacterial groups were determined by FISH. As shown in Table 2, there were no differences in the numbers of bacteria in any of the bacterial groups at the beginning of the placebo and prebiotic treatment periods.

Table 2.   Log10 numbers of the monitored bacterial groups in the faecal microflora of the volunteers at the beginning of each treatment
 Group I (n = 16)Group II (n = 14)Group III (n = 14)
Placebo/Prebiotic 3.5 gPlacebo/Prebiotic 7.0 gPlacebo/Placebo
Bacterial groupPlacebo (week 6)Prebiotic (week 12)Placebo (week 6)Prebiotic (week 12)Placebo (week 6)Placebo (week 12)
Total10.51 ± 0.3410.59 ± 0.2410.50 ± 0.2510.56 ± 0.2610.51 ± 0.3110.46 ± 0.35
Bifidobacterium spp.9.00 ± 0.329.10 ± 0.338.96 ± 0.319.03 ± 0.428.97 ± 0.338.94 ± 0.31
Lactobacillus-Enterococcus spp.8.46 ± 0.258.57 ± 0.238.36 ± 0.228.48 ± 0.228.52 ± 0.198.44 ± 0.28
Clostridium perfringens-hystolyticum subgroup8.52 ± 0.298.60 ± 0.298.56 ± 0.288.58 ± 0.328.62 ± 0.368.56 ± 0.31
Eubacterium rectale/C.coccoides9.54 ± 0.319.59 ± 0.249.51 ± 0.299.51 ± 0.369.57 ± 0.339.47 ± 0.29
Bacteroides-Prevotella spp.9.44 ± 0.329.52 ± 0.259.41 ± 0.219.51 ± 0.239.45 ± 0.299.41 ± 0.32
E.coli7.62 ± 0.297.73 ± 0.287.5 ± 0.297.63 ± 0.227.58 ± 0.327.63 ± 0.27
Desulfovibrio spp.7.84 ± 0.307.93 ± 0.217.78 ± 0.317.82 ± 0.237.86 ± 0.307.88 ± 0.25

The relative proportions of each bacterial group assessed at baseline and at the end of each 4-week treatment period taking into account changes in total bacterial numbers are shown in Table 3. With one exception, group proportions at the beginning of the placebo treatment periods were similar to the group proportions at the beginning of the prebiotic treatment periods. In group 3 patients (placebo/placebo), E. rectale/C. cocciodes proportions were lower (P < 0.05) at the start of the second placebo treatment period compared with the first.

Table 3.   Bacterial proportion of the monitored faecal microflora components during the treatments used for IBS patients
Bacterial groupGroup I (n = 16)Group II (n = 14)Group III (n = 14)
Placebo/Prebiotic 3.5 gPlacebo/Prebiotic 7.0 gPlacebo/Placebo
  1. b, baseline; e: end of 4 week treatment.

  2. Values are means ± s.d.

  3. * Indicates difference from beginning of treatment (P < 0.05).

  4. ** Indicate difference from beginning of treatment (P < 0.005).

  5. *** Indicate differences from the beginning of the treatment (P < 0.001). (paired t-tests).

  6. Mean values (±s.d.) with unlike superscript letters within each row were significantly different (P < 0.05).

  7. The monitored bacterial group ratios were analysed by anova using three fixed effect (treatment, period of treatment, sequence of treatment).

Bifidobacterium sppb3.08 ± 0.313.25 ± 0.512.90 ± 0.423.01 ± 0.382.85 ± 0.542.99 ± 0.45
e3.42 ± 0.295.51 ± 0.43**,b2.94 ± 0.38a7.48 ± 0.59***,c2.90 ± 0.672.95 ± 0.27
Lactobacillus-Enterococcus sppb0.89 ± 0.120.95 ± 0.170.74 ± 0.570.85 ± 0.751.00 ± 0.150.95 ± 0.44
e0.99 ± 0.111.19 ± 0.100.81 ± 0.771.16 ± 0.670.96 ± 0.231.04 ± 0.33
Clostridium perfringens-hystolyticum subgroupb1.02 ± 0.101.03 ± 0.081.15 ± 0.131.07 ± 0.181.26 ± 0.161.24 ± 0.29
e0.95 ± 0.28a,b0.71 ± 0.15b1.19 ± 0.28a0.76 ± 0.09*,b1.01 ± 0.17a1.38 ± 0.35a
Eubacterium rectale/C.coccoidesb10.79 ± 1.21a,b9.95 ± 0.76a,b10.19 ± 1.42a,b9.07 ± 0.74a11.47 ± 1.32b10.07 ± 1.03a,b
e10.32 ± 0.99a13.12 ± 1.21**,b9.32 ± 1.01a10.16 ± 0.96a10.94 ± 0.89a11.91 ± 0.78a
Bacteroides-Prevotella spp.b8.57 ± 0.958.56 ± 0.068.16 ± 0.638.93 ± 0.938.51 ± 0.648.95 ± 0.72
e8.59 ± 0.69a8.03 ± 0.73a8.67 ± 0.59a5.75 ± 0.75**,b8.01 ± 0.59a9.50 ± 0.67a
E.colib0.13 ± 0.030.14 ± 0.020.10 ± 0.010.12 ± 0.020.12 ± 0.020.15 ± 0.02
e0.16 ± 0.02a0.14 ± 0.01a0.11 ± 0.02a,b0.09 ± 0.02b0.13 ± 0.01a0.15 ± 0.02a
Desulfovibrio sppb0.21 ± 0.030.22 ± 0.030.19 ± 0.020.18 ± 0.010.22 ± 0.020.26 ± 0.01
e0.24 ± 0.03a0.24 ± 0.04a0.18 ± 0.01b0.16 ± 0.01b0.24 ± 0.02a0.27 ± 0.02a

Four weeks treatment with placebo did not change proportions of bacteria in any of the groups assessed. In contrast, prebiotic treatment resulted in significantly higher proportions of bifidobacterium spp. (P < 0.05 and P < 0.001, for 3.5 and 7.0 g/d doses respectively). There were significantly lower proportions of C.perfringens subgroup histolyticum (P < 0.05) and Bacteroides/Prevotella spp. (P < 0.005) after administration of 7.0 g/d prebiotic and higher proportions of E.rectale/C.coccoides after administration of 3.5 g/d (P < 0.005).

Compared to the effects of 4 weeks treatment with placebo, the prebiotic treatment had a clear and significant effect on the relative proportions of Bifidobacterium spp. after administration of both 3.5 (P < 0.005) and 7 g (P < 0.001) and it was also significantly (P < 0.05) higher with the higher dose of a prebiotic. The higher dose of the prebiotic also resulted in the lower proportion of C.perfringens subgroup histolyticum and Bacteroides-Prevotella spp. (P < 0.05 in both cases) when compared with the placebo in the same treatment group and a higher proportion of E.rectale/C.coccoides spp. was achieved with the lower prebiotic dose (P < 0.05).

Response to treatment (Table 4)

Table 4.   Mean scores for efficacy variables during the treatments used for the IBS patients
  Group I (n = 16)Group II (n = 14)Group III (n = 14)
Placebo/Prebiotic 3.5 gPlacebo/Prebiotic 7.0 gPlacebo/Placebo
  1. b, Baseline; e, end of 4 week treatment.

  2. Values are means ± s.d.

  3. * Indicates different from beginning of treatment (P < 0.05). (Matched pairs t-tests).

  4. Mean values (±s.d.) with unlike superscript letters within each row were significantly different (P < 0.05).

  5. The recorded scores were analysed by anova using three fixed effect (volunteer, treatment, period of treatment)

Defecation No/db2.7 ± 1.32.1 ± 1.12.4 ± 1.31.6 ± 1.12.0 ± 1.12.5 ± 0.9
e2.1 ± 1.31.8 ± 1.11.9 ± 1.11.8 ± 1.42.1 ± 0.92.0 ± 1.1
Stool consistencyb3.5 ± 1.44.4 ± 1.44.0 ± 1.14.6 ± 1.14.6 ± 1.14.3 ± 0.8
e4.1 ± 1.4*,b3.8 ± 1.0*,b,a3.4 ± 1.3a4.3 ± 1.1b4.3 ± 1.1b4.1 ± 1.1b
Flatulenceb1.8 ± 0.8b2.0 ± 0.62.1 ± 0.8b1.6 ± 0.9a2.0 ± 0.6b1.9 ± 0.7a
e1.8 ± 0.7b1.3 ± 0.6*a1.5 ± 0.9*a,b1.5 ± 1.0a,b1.7 ± 0.7b1.7 ± 0.7*b
Abdominal Painb4.2 ± 2.03.5 ± 2.23.6 ± 2.62.9 ± 2.53.0 ± 2.33.5 ± 2.1
e2.4 ± 2.1*a,b1.9 ± 1.7a3.1 ± 2.5a,b2.6 ± 2.1a3.6 ± 1.5a,b3.4 ± 1.9a,b
Bloatingb4.4 ± 15b4.1 ± 1.3b3.9 ± 1.3a,b3.4 ± 1.5a3.9 ± 1.4a,b2.2 ± 2.3a,b
e3.9 ± 1.1b2.8 ± 0.9*a3.8 ± 1.3b4.1 ± 1.2b*3.8 ± 1.4b3.5 ± 1.6b
Incomplete evacuation/urgency/strainingb2.7 ± 2.02.4 ± 2.53.2 ± 2.83.6 ± 2.22.7 ± 2.52.7 ± 2.3
e2.0 ± 1.91.8 ± 1.73.6 ± 2.32.6 ± 2.22.9 ± 2.62.6 ± 2.2
Likert scaleb11.1 ± 5.3b9.9 ± 6.2a10.2 ± 6.3b8.2 ± 5.7a10.4 ± 4.6b9.4 ± 4.2a
e7.9 ± 4.2*b,c6.2 ± 4.3*a9.4 ± 5.6c8.1 ± 4.6b10.1 ± 4.4c7.7 ± 5.4*c
SGAb4.3 ± 0.54.2 ± 0.54.2 ± 0.64.1 ± 0.84.1 ± 0.64.0 ± 0.7
e4.0 ± 0.7b3.1 ± 0.8*a4.1 ± 0.8b3.6 ± 0.9*a,b4.1 ± 0.7b3.5 ± 0.5b
HAD (A)b7.8 ± 4.07.9 ± 4.210.6 ± 4.39.7 ± 4.38.1 ± 4.39.0 ± 5.0
e7.5 ± 4.4b8.1 ± 3.8b9.7 ± 4.3a7.8 ± 4.6*b9.1 ± 5.5b8.5 ± 5.3b
HAD (D)b5.8 ± 3.95.3 ± 2.77.9 ± 3.36.7 ± 5.86.9 ± 4.16.9 ± 4.8
e5.4 ± 3.6a5.1 ± 2.8a7.4 ± 3.6b,c5.8 ± 3.4a,b6.0 ± 4.2b6.6 ± 4.1b,c
QOLb92.8 ± 44.2b81.4 ± 40.9a,b97.4 ± 31.4b77.1 ± 39.8a93.4 ± 48.1b97.1 ± 46.4b
e89.4 ± 41.4b,c79.8 ± 40.4a,b91.1 ± 34.8b,c74.3 ± 41.6a98.5 ± 48.1c97.0 ± 49.2c

The clinical and psychological parameters were similar at the beginning of the placebo and prebiotic treatment periods. In group I patients, treatment with placebo resulted in a significant change in stool consistency (P < 0.05) and an improvement in abdominal pain (P < 0.05). In group II patients, treatment with placebo resulted in a significant improvement in flatulence (P < 0.05). Treatment with placebo (3.5 g/d, group I; 7.0 g/d group II; 7.0 g/d group II) was without effect on any of the other clinical parameters monitored and on none of the psychological parameters. Treatment of group I patients with 3.5 g/d prebiotic resulted in a significant change in stool consistency (P < 0.05), flatulence (P < 0.05) bloating (P < 0.05), Composite Likert scale (P < 0.05) and SGA (P < 0.05). Treatment of group II patients with 7.0 g/d prebiotic resulted in a significant improvement in SGA (P < 0.05) and anxiety (HAD[A]) scores (P < 0.05).

Compared with the effects of 4 weeks treatment with placebo, prebiotic treatment (3.5 g/d) had a significantly greater effect on flatulence (P < 0.05), bloating (P < 0.05), Composite Likert Scale (P < 0.05) and SGA (P < 0.05). Compared with the effects of a 4-week treatment with placebo, prebiotic treatment (7.0 g/d) had a significantly greater effect on composite Likert Scale (P < 0.05), anxiety scores (P < 0.05) and quality of life (P < 0.05).

Fibre intake

Mean daily fibre intakes during the treatment periods are shown in Table 5. During the course of all treatment periods in the three groups of patients, fibre intake remained unchanged. Neither placebo nor prebiotic treatment was associated with any significant change in fibre intake.

Table 5.   Mean daily fibre intakes during the treatments used for the IBS patients (g/day)
 Group I (n = 16)Group II (n = 14)Group III (n = 14)
Placebo/Prebiotic 3.5 gPlacebo/Prebiotic 7.0 gPlacebo/Placebo
  1. b, baseline; e, end of 4 week treatment.

  2. Values are means ± s.d.

Fibreb13.1 ± 4.0613.8 ± 6.649.4 ± 3.4610.5 ± 4.7310.9 ± 5.0410.5 ± 4.66
e13.7 ± 5.614.3 ± 6.7410.6 ± 4.899.9 ± 3.758.6 ± 3.299.11 ± 3.82
Soluble Fibreb4.6 ± 1.444.5 ± 1.323.2 ± 1.383.5 ± 1.613.9 ± 2.103.6 ± 2.15
e4.3 ± 1.234.3 ± 1.683.6 ± 1.783.4 ± 1.562.9 ± 1.273.4 ± 2.02
Insoluble Fibreb3.0 ± 1.592.6 ± 1.161.9 ± 1.012.2 ± 1.512.3 ± 1.782.3 ± 1.63
e2.9 ± 1.262.7 ± 1.662.3 ± 1.262.0 ± 0.932.3 ± 1.712.0 ± 1.29

Adverse events

Three patients reported adverse events. One (IBS-D) reported an increase in severity of diarrhoea while taking 3.5 g placebo. Symptoms were described as moderate and the patient withdrew from the study. Two patients reported nausea (1 patient 7.0 g placebo; 1 patient 3.5 g prebiotic), described as mild events; the nausea lasted a mean of two days.


The present study is the first to have been undertaken to investigate the potential efficacy of a prebiotic in the management of patients with IBS. The prebiotic consisting of a galactooligosaccharide mixture has been confirmed as being selective towards the beneficial genus Bifidobacterium25 the capacity of which to improve lower gut health is recognized.39 As such, it seemed to have a potential as a therapeutic agent in IBS as three recent studies have shown the beneficial effects of probiotic bifidobacteria in the management of IBS.21–23 The test strains were shown not only to improve abdominal symptoms21–23 but also provided an insight into possible mechanisms of action in that one study demonstrated a pro-inflammatory THI-like cytokine IL-10/IL-12 ratio in the peripheral blood of the IBS patients that was normalized by Bifidobacterium infantis treatment.21 Importantly, in a more recent study, treatment with fermented milk containing Bifidobacterium animalis resulted in a significant improvement in health related quality of life.23

Orally applied prebiotics have also been associated with immune-modulatory effects on parameters of the innate and specific immunity, mostly in animal experiments or in vitro. As prebiotics are fermented to various degree by the intestinal beneficial bacteria (e.g. bifidobacteria, lactobacilli), the immune-modulatory properties have mainly been attributed to a microflora dependent effect. However, specific prebiotics have also been reported to bind to specific receptors on cells of the immune system suggesting that a direct interaction between prebiotics and the host is in place.40 One of the important findings of our study is that the prebiotic was shown to have a bifidobacterial enhancing effect in the IBS patients at both doses administered, with levels of bifidobacteria becoming similar to those of healthy humans.25 In addition, at the 3.5 g daily dose, the numbers of Eubacterium rectale/C.coccoides were increased and at the 7.0 g daily dose, those of Clostridium perfringens subgroup histolyticum and Bacteroides-Prevotella Spp. were reduced. Whether IBS is accompanied by quantitative or qualitative changes in the bacterial flora of the large intestine remains a contentious issue. Using older microbiological culture techniques, a qualitative alteration in the faecal flora was found41–43 as was increased colonic fermentation in some patients.17, 18 Recently, it has been reported that there are higher numbers of bacteroides and lower numbers of bifidobacteria in the mucosa associated microflora of IBS sufferers.44 Further studies using the more advanced analytical technique such as used here are clearly required to characterize in more detail possible quantitative or qualitative changes in colonic bacterial flora of different subgroups of IBS patients.

Treatment with the prebiotic, which induced qualitative changes in the faecal flora, was associated with significant changes in stool consistency, flatulence, composite scores (abdominal pain/discomfort, bloating/distension and bowel movement difficulty) as well as subjective global assessment values in our patients. Our results are thus reminiscent of the beneficial clinical effects obtained when treating IBS patients with Bifidobacterium spp. containing probiotics.21–23 The lower flatulence production seen in the prebiotic treatment arm may have been due to the increased bifidobacterial load, as these are not a gas producing genus of gut bacteria.45 The IBD-D patients were the only subgroup to have clinically significant anxiety (HAD (A)>11). In these, anxiety scores were significantly reduced during treatment with 7.0 g/d prebiotic and in these patients, quality of life also significantly improved.

One of the objectives of this study was to try and obtain some information about what might constitute an optimum dose of prebiotic. Towards this end, the bacteriological data suggest that the 7.0 g rather than 3.5 g daily dose had the best prebiotic effect on the faecal microflora. On the other hand, from the clinical standpoint, the 3.5 g/d dose was overall more effective than the 7.0 g/d dose, although both doses showed improvements. There is no clear explanation yet to explain those differences. It is of some interest that in the dose comparison of the probiotic Bifidobacterium infantis 35624 in IBS patients, a dose of 1 x 108 CFU provided the greatest benefit, whereas the higher 1 × 1010 CFU dose was without effect as compared to placebo.22 This lack of effect was ascribed to the difficulties in ‘packaging’ the higher dose of probiotic.22, 46 The possibility might exist that in both that probiotic study22 and in the present prebiotic study, the lack of effects of the higher doses were real ones.

Both doses were well tolerated and no adverse side effects were reported by any of the patients. The study provides no information, however, as to how long treatment might have to be continued nor information as to how long the alterations in bacteriological milieu might last after discontinuation of treatment. It was because of the known carry over effect of prebiotic treatments on the faecal bacteriology in cross over studies, that our patients always received the placebo treatment before the probiotic.

IBS is a very common disorder1, 2 with many patients suffering chronic abdominal symptoms3 and poor quality of life.4 Treatment remains problematic for many patients, particularly since the recent withdrawal of clinically proven receptor active drugs.15

Although the results of this study point to the need for multicentre, randomized, controlled, clinical trials of large numbers of patients to assess more formally the efficacy of prebiotic interventions in IBS patients, our data do support the concept that dietary intervention directed towards gut microbiota modulation represents a significant step forward in therapy.


Declaration of personal interest: G. Tzortzis is an employee of Clasado Ltd. Declaration of funding interests: This study was funded in full by Clasado Ltd under contract. However, the funding source, Clasado Ltd, had no role in study design, collection, analysis and interpretation of the data or in the writing of the report. The decision to submit the paper for publication was made by the authors.