Irritable bowel syndrome (IBS) and chronic idiopathic constipation

CLINICAL AND SYSTEMATIC S nature publishing group 1547 see related editorial on page 1563 CME Efficacy of Prebiotics, Probiotics, and Synbiotics in Irritable Bowel Syndrome and Chronic Idiopathic Constipation: Systematic Review and Meta-analysis Alexander C. Ford, MBChB, MD 1, 2, Eamonn M.M. Quigley, MD, FRCP, FACP, FACG, FRCP 3, Brian E. Lacy, MD, PhD 4, Anthony J. Lembo, MD 5, Yuri A. Saito, MD, MPH 6, Lawrence R. Schiller, MD, MSHS, RFF, FACG, AGAF 7, Edy E. Soffer, MD 8, Brennan M.R. Spiegel, MD, MSHS, RFF, FACG, AGAF 9 and Paul Moayyedi, MBChB, PhD, MPH, FACG 10 OBJECTIVES: METHODS: RESULTS: Irritable bowel syndrome (IBS) and chronic idiopathic constipation (CIC) are functional bowel disorders. Evidence suggests that disturbance in the gastrointestinal microbiota may be implicated in both conditions. We performed a systematic review and meta-analysis to examine the efficacy of prebiotics, probiotics, and synbiotics in IBS and CIC. MEDLINE, EMBASE, and the Cochrane Controlled Trials Register were searched (up to December 2013). Randomized controlled trials (RCTs) recruiting adults with IBS or CIC, which compared prebiotics, probiotics, or synbiotics with placebo or no therapy, were eligible. Dichotomous symptom data were pooled to obtain a relative risk (RR) of remaining symptomatic after therapy, with a 95 % confidence interval (CI). Continuous data were pooled using a standardized or weighted mean difference with a 95 % CI. The search strategy identified 3,216 citations. Forty-three RCTs were eligible for inclusion. The RR of IBS symptoms persisting with probiotics vs. placebo was 0.79 (95 % CI 0.70 0.89). Probiotics had beneficial effects on global IBS, abdominal pain, bloating, and flatulence scores. Data for prebiotics and synbiotics in IBS were sparse. Probiotics appeared to have beneficial effects in CIC (mean increase in number of stools per week = 1.49; 95 % CI = 1.02 1.96), but there were only two RCTs. Synbiotics also appeared beneficial (RR of failure to respond to therapy = 0.78; 95 % CI 0.67 0.92). Again, trials for prebiotics were few in number, and no definite conclusions could be drawn. CONCLUSIONS: Probiotics are effective treatments for IBS, although which individual species and strains are the most beneficial remains unclear. Further evidence is required before the role of prebiotics or synbiotics in IBS is known. The efficacy of all three therapies in CIC is also uncertain. SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg Am J Gastroenterol 2014; 109:1547 1561; doi: 10.1038/ajg.2014.202; published online 29 July 2014 INTRODUCTION Irritable bowel syndrome (IBS) and chronic idiopathic constipation (CIC) are functional bowel disorders. Both conditions are common, with a prevalence of between 5 and 20 % in the general population, depending on the criteria used to define their presence ( 1,2 ). IBS and CIC are more common in women ( 2,3 ), and IBS occurs more frequently in younger individuals ( 1 ), whereas the prevalence of CIC increases with age ( 2 ). Evidence for any effect of socioeconomic status is uncertain. The cause of these functional bowel disorders remains unclear, but 1 Leeds Gastroenterology Institute, St James s University Hospital, Leeds, UK ; 2 Leeds Institute of Biomedical and Clinical Sciences, University of Leeds, Leeds, UK ; 3 Division of Gastroenterology and Hepatology, Department of Medicine, Houston Methodist Hospital, Houston, Texas, USA ; 4 Dartmouth-Hitchcock Medical Center, Gastroenterology, One Medical Center Drive, Lebanon, New Hampshire, USA ; 5 The Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA ; 6 Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota, USA ; 7 Digestive Health Associates of Texas, Baylor University Medical Center, Dallas, Texas, USA ; 8 Division of Gastroenterology at Cedars-Sinai, University of Southern California, Los Angeles, California, USA ; 9 Department of Gastroenterology, VA Greater Los Angeles Healthcare System, Los Angeles, California, USA ; 10 Gastroenterology Division, McMaster University, Health Sciences Center, Hamilton, Ontario, Canada. Correspondence: Alexander C. Ford, MBChB, MD, Leeds Gastroenterology Institute, St. James s University Hospital, Room 125, 4th Floor, Bexley Wing, Beckett Street, Leeds LS9 7TF UK. E-mail: alexf12399@yahoo.com Received 25 February 2014; accepted 30 April 2014

1548 Ford et al. Box 1. Eligibility criteria Randomized controlled trials Adults (participants aged > 16 years) Diagnosis of IBS or CIC based on either a clinician s opinion, or meeting specific diagnostic criteria (Manning, Kruis score, Rome I, II, or III), supplemented by negative investigations where trials deemed this necessary. Compared prebiotics, probiotics, or synbiotics with placebo. Minimum duration of therapy 7 days. Minimum duration of follow-up 7 days. Dichotomous assessment of response to therapy in terms of effect on global IBS symptoms or abdominal pain following therapy, or response to therapy for CIC, or continuous data in the form of effect on IBS symptom scores at study end or effect on mean number of stools per week for CIC (Preferably patient-reported, but if this was not available then as assessed by a physician or questionnaire data). visceral hypersensitivity ( 4,5 ), disturbances in gastrointestinal (GI) flora ( 6,7 ), and chronic immune activation leading to a low-grade mucosal inflammation ( 8 11 ) have all been implicated in the pathogenesis of IBS and CIC. Effective pharmacological therapies for IBS and CIC exist ( 12 15 ), but the duration of most treatment trials is < 6 months, meaning that no therapy has been proven to alter the natural history of either condition in the long term. In addition, many patients do not respond to, or become dissatisfied with, conventional therapies ( 16,17 ). Modulating the GI flora, as a means of improving symptoms, may therefore be an attractive treatment option. Probiotics, which are live or attenuated microorganisms that may have beneficial effects in humans, have been widely studied in IBS ( 18 ). Some probiotics appear to have anti-infl ammatory properties ( 19 ), or the ability to modulate visceral hypersensitivity ( 20,21 ). Prebiotics are ingredients in food that remain undigested, and which may stimulate either the growth or the activity of bacteria that are also beneficial to human health. Examples of prebiotics include fructo-oligosaccharides and inulin. Synbiotics are combinations of prebiotics and probiotics, with a potentially synergistic action. In our previous systematic review and meta-analysis ( 18 ), conducted to inform the American College of Gastroenterology s monograph on the management of IBS ( 22 ), we summarized all available evidence for the use of probiotics in IBS. At this time, there was limited information concerning the efficacy of either prebiotics or synbiotics in IBS, or the benefit of any of these three treatments in CIC. However, in the intervening 4 years, there has been a considerable amount of evidence published. We have therefore re-examined this issue. METHODS Search strategy and study selection We updated our previous systematic review and meta-analysis examining the efficacy of probiotics in IBS ( 18 ), and also performed a search for studies that reported on the effectiveness of prebiotics and synbiotics in IBS, as well as prebiotics, probiotics, and synbiotics in CIC. A search of the medical literature was conducted using MEDLINE (1946 to December 2013), EMBASE and EMBASE Classic (1947 to December 2013), and the Cochrane central register of controlled trials. Randomized placebo-controlled trials examining the effect of prebiotics, probiotics, and synbiotics in adult patients (over the age of 16 years) with IBS or CIC were eligible for inclusion ( Box 1 ). Duration of therapy had to be at least 7 days. The diagnosis of IBS or CIC could be based on either a physician s opinion or symptom-based diagnostic criteria, supplemented by the results of investigations to exclude organic disease, where studies deemed this necessary. Subjects were required to be followed up for at least 1 week, and studies had to report response to therapy as either dichotomous or continuous data. For IBS, dichotomous assessment could be in the form of either an assessment of global symptom cure or improvement, or abdominal pain cure or improvement, after completion of therapy: preferably as reported by the patient, but if this was not recorded then as documented by the investigator or via questionnaire data. Continuous data of interest were the effect of therapy on IBS symptom scores at the end of study. For CIC, outcomes of interest were a dichotomous assessment of overall response to therapy, or continuous data in the form of mean number of stools per week during therapy, preferably as reported by the patient, but if this was not recorded then as documented by the investigator or via questionnaire data. In studies that included patients with IBS or CIC among patients with other functional GI disorders, or studies that did not report these types of dichotomous or continuous data, but were otherwise eligible for inclusion in the systematic review, we attempted to contact the original investigators in order to obtain further information. The literature search was performed as part of a broader exercise to inform the update of the American College of Gastroenterology s (ACG) monograph on the management of IBS and CIC. Specifically, studies on IBS were identified with the terms irritable bowel syndrome and functional diseases, colon (b ot h as medical subject heading (MeSH) and free text terms), and IBS, spastic colon, irritable colon, or functional adj5 bowel (as f ree text ter ms). Studies on CIC were identified with constipation or gastrointestinal transit (b ot h as medical subjec t he adings (MeSH) and f ree text terms), or functional constipation, idiopathic constipation, and chronic constipation, or slow transit (as f ree text ter ms). These were The American Journal of GASTROENTEROLOGY VOLUME 109 OCTOBER 2014 www.amjgastro.com

Prebiotics, probiotics, and synbiotics in IBS and CIC 1549 Box 2. Data extraction methodology Outcome of interest: improvement in global IBS symptoms preferable, if this was not reported then improvement in abdominal pain. Reporting of outcomes: patient-reported was preferable, if this was not available then investigator-reported. Time of assessment: upon completion of therapy. Denominator used: true intention-to-treat analysis, if this was not available then all evaluable patients. Cutoff used for dichotomization: any improvement in global IBS symptoms or abdominal pain for Likert-type scales. Any response to therapy for CIC. combined using the set operator AND with studies identified with the following terms: Saccharomyces, Lactobacillus, Bifidobacterium, Escherichia coli, probiotics, synbiotics, or prebiotics (both as MeSH and free text terms). There were no language restrictions, and abstracts of the papers identified by the initial search were evaluated by the lead reviewer for appropriateness to the study question, with all potentially relevant papers obtained and evaluated in detail. Foreign language papers were translated where necessary. Abstract books of conference proceedings between 2001 and 2013 were hand-searched to identify potentially eligible studies published only in abstract form. The bibliographies of all identified relevant studies were used to perform a recursive search of the literature. Articles were independently assessed by two reviewers using predesigned eligibility forms, according to the prospectively defined eligibility criteria. Any disagreement between investigators was resolved by consensus. Outcome assessment The primary outcomes assessed were the effect of prebiotics, probiotics, or synbiotics compared with placebo on global IBS symptoms or abdominal pain after cessation of therapy, or effect on overall response to therapy for CIC. Secondary outcomes included the effect of prebiotics, probiotics, or synbiotics on global IBS symptom scores and individual IBS symptom scores at the end of study, including abdominal pain, bloating, urgency, or flatulence, or effect on mean number of stools per week for CIC. We also examined numbers of adverse events as a result of prebiotics, probiotics, or synbiotics. Data extraction All data were extracted independently by two reviewers on to a Microsoft Excel spreadsheet (XP professional edition; Microsoft, Redmond, WA) as dichotomous outcomes (global IBS symptoms persistent or unimproved, or abdominal pain persistent or unimproved, or response or no response to therapy for CIC; Box 2 ), or mean symptom scores or mean number of stools per week at study end, along with a s.d.. In addition, the following clinical data were extracted for each trial: setting (primary, secondary, or tertiary care-based), number of centers, country of origin, prebiotic, probiotic, or synbiotic used (including strain and species where applicable), duration of therapy, total number of adverse events reported, criteria used to define IBS or CIC, primary outcome measure used to define symptom improvement or cure following therapy, duration of follow-up, proportion of female patients, and proportion of patients according to predominant stool pattern for IBS. Data were extracted as intention-to-treat analyses, with all dropouts assumed to be treatment failures, wherever trial reporting allowed this. Assessment of risk of bias This was performed independently by two investigators, with disagreements resolved by discussion. Risk of bias was assessed as described in the Cochrane handbook ( 23 ), by recording the method used to generate the randomization schedule and conceal allocation, whether blinding was implemented, what proportion of patients completed follow-up, whether an intention-to-treat analysis was extractable, and whether there was evidence of selective reporting of outcomes. Data synthesis and statistical analysis D at a w e re p o o l e d u s i n g a r a n d om - effects model ( 24 ), to give a more conservative estimate of the effect of prebiotics, probiotics, or synbiotics, allowing for any heterogeneity between studies. The impact of prebiotics, probiotics, or synbiotics in IBS was expressed as a relative risk (RR) of global IBS symptoms or abdominal pain persisting with intervention compared with control, with 95 % confidence intervals (CIs), or a standardized mean difference (SMD) in global and individual IBS symptom scores at study end, with 95 % CIs. The impact of prebiotics, probiotics, or synbiotics in CIC was expressed as a RR of no response with intervention compared with control, with 95 % CIs, or a weighted mean difference in the mean number of stools per week at the end of study, with 95 % CIs. Adverse events data were also summarized with RRs. The number needed to treat (NNT) and the number needed to harm, with 95 % CIs, were calculated from the reciprocal of the risk difference of the metaanalysis. Heterogeneity b etween studies was ass ess ed using b ot h t he I 2 statistic with a cutoff of 50 % and t he χ 2 -test wit h a P value < 0.10, used to define a significant degree of heterogeneity ( 25 ). In cases in which the degree of statistical heterogeneity was greater than this between trial results in this meta-analysis, possible explanations were investigated using sensitivity analyses according to the type of prebiotic, probiotics, or synbiotic used, trial setting, criteria used to define IBS or CIC, whether the method of randomization or

1550 Ford et al. concealment of allocation were reported, level of blinding, and risk of bias of included trials. We compared individual RRs between these analyses using the Cochran s Q statistic. These were exploratory analyses only, and may explain some of the observed variability, but the results should be interpreted with caution. R e v i e w Man a ge r ve rs i on 5. 1. 4 ( R e v Man for Wi n d ow s 2 0 0 8, the Nordic Cochrane Centre, Copenhagen, Denmark) and StatsDirect version 2.7.7 (StatsDirect, Sale, Cheshire, UK) were used to generate Forest plots of pooled RRs, risk differences, and SMDs for primary and secondary outcomes with 95 % CIs, as well as funnel plots. The latter were assessed for evidence of asymmetry, and therefore possible publication bias or other small study effects, using the Egger test ( 26 ), if there were sufficient ( 1 0 ) e l i g i bl e s tu d i e s i n clu d e d i n t h e m e t a - an a ly s i s, in line with recent recommendations ( 27 ). RESULTS The search strategy generated a total of 3,216 citations, of which 73 published articles appeared to be relevant, and were retrieved for further assessment ( Figure 1 ). Of these, 30 articles were excluded for various reasons, leaving 43 eligible articles. Agreement between reviewers for the assessment of trial eligibility was excellent (kappa statistic = 0.93). We identified only one randomized controlled trial (RCT) that evaluated the prebiotic trans-galacto-oligosaccharide in IBS ( 28 ). This study was excluded from further analysis, as the data were not extractable. We identified 35 RCTs of probiotics in IBS ( 19,29 62 ), and two studies of synbiotics ( 63,64 ). There was only one trial of prebiotics in CIC ( 65 ), three of probiotics ( 66 68 ), and two of synbiotics ( 69,70 ). Efficacy and safety of probiotics in IBS The 35 RCTs of probiotics in IBS involved 3,452 patients ( 19,29 62 ). The proportion of women in the trials ranged between 26 and 100 %. Fourteen trials were at a low risk of bias ( 30,35,36,41,43, 45 47,49,50,57 60 ), with the remainder being unclear. Nineteen trials used a combination of probiotics, eight Lactobacillus, three Bifidobacterium, t wo E. coli, one Streptococcus, one Saccharomyces, and one either Lactobacillus or Bifidobacterium. Detailed characteristics of included RCTs are provided in Table 1. Efficacy of probiotics in the treatment of IBS: effect on persistence of symptoms There were 23 RCTs comparing probiotics with placebo for the treatment of IBS ( 30,31,33 41,43,44,46,49,51,53 55,57 60 ), evaluating 2,575 patients, which gave outcomes as a dichotomous variable. Overall, 777 (55.8 % ) of 1,392 patients assigned to probiotics reported persistent or unimproved IBS symptoms following therapy, compared with 865 (73.1 % ) of 1,183 allocated to placebo. The RR of IBS symptoms persisting or remaining unimproved after treatment with probiotics vs. placebo was 0.79 (95 % CI 0.70 0.89), with statistically significant heterogeneity detected between studies ( I 2 = 7 2 %, P < 0. 0 0 1 ; Figure 2 ). There was no statistically significant asymmetry detected in the funnel Studies identified in literature search (n = 3216) Studies retrieved for evaluation (n = 73) Eligible studies (n = 43): Synbiotics in IBS (n = 2) Prebiotics in IBS (n = 0) Probiotics in IBS (n = 35) Combination probiotics = 19 Lactobacillus = 8 Bifidobacterium = 3 Escherichia = 2 Bifidobacterium or Lactobacillus = 1 Saccharomyces = 1 Streptococcus = 1 Synbiotics in CIC = 2 Prebiotics in CIC = 1 Probiotics in CIC = 3 Excluded (title and abstract revealed not appropriate) (n = 3143) Excluded (n = 30) because: No extractable data reported = 9 Not the intervention of interest = 5 No placebo arm = 4 Dual publication = 3 Not randomized = 3 Review article = 2 Crossover study with no extractable data = 3 Enrolled healthy women = 1 Figure 1. Flow diagram of assessment of studies identifi ed in the updated systematic review and meta-analysis. plot (Egger test, P = 0.36), to suggest publication bias or other small study effects. The NNT with probiotics was 7 (95 % CI 4 12.5). When only the 12 RCTs at a low risk of bias that provided dichotomous data were considered in the analysis, the effect of probiotics was still statistically significant (RR of persistent or unimproved symptoms = 0.82; 95 % CI 0.69 0.98). Combination probiotics were assessed in 12 RCTs ( 30,31,33,35,37,43,44,46,54,55,57,59 ), comprising 1,197 patients, with a significant effect on symptoms (RR = 0.81; 95 % CI 0.67 0.98; Figure 2 ), but with significant heterogeneity between studies ( I 2 = 72 %, P < 0.001). The NNT with combination probiotics was 8 (95 % CI 4 50). In terms of the different combinations tested, three trials used the same combination of L. paracasei ssp paracas ei F19, L. acidophilus L a5, and B. lactis Bb12 in 269 patients ( 30,57,59 ), with no benefit over placebo (RR = 0.92; 95 % CI 0.76 1.11). The American Journal of GASTROENTEROLOGY VOLUME 109 OCTOBER 2014 www.amjgastro.com

Prebiotics, probiotics, and synbiotics in IBS and CIC 1551 Table 1. Characteristics of randomized controlled trials of probiotics vs. placebo in irritable bowel syndrome Study Country and recruitment Criteria used to define symptom improvement following therapy Sample size ( % female) and diagnostic criteria for IBS Probiotic used and duration of therapy Methodology Gade ( 40 ) Denmark, primary care IBS symptoms signifi cantly improved 54 (78), Manning Paraghurt (Streptococcus faecium ) for 4 weeks Method of randomization is stated. Method of concealment of allocation is not stated. Doubleblind. No other IBS medications allowed. Nobaek ( 53 ) Sweden, population based > 1.5 improvement in VAS scale for abdominal pain, and continuous scale for IBS symptoms 60 (69), Rome I Rose hip drink (400 ml) containing Lactobacillus plantarum DSM 9843 (5 10 7 c.f.u.) o.d. for 4 weeks allocation are not stated. Double-blind. Unclear if other IBS medications are allowed. Niedzielin ( 51 ) Poland, primary care Any improvement in IBS symptoms 40 (80), clinical diagnosis Fruit drink (200 ml) containing L. plantarum 299V (5 10 7 c.f.u. / ml) b.i.d. for 4 weeks allocation are not stated. Double-blind. No other IBS medications allowed. Kim ( 46 ) USA, secondary care Satisfactory relief of IBS symptoms for 50 % of weeks, and continuous scale for IBS symptoms 25 (72), Rome II One packet containing VSL # 3 (225 billion bacteria / packet) b.i.d. for 8 weeks allocation are stated. Double-blind. No other IBS medications allowed apart from antidepressants. Kajander ( 44 ) Finland, advertising Relief of IBS symptoms, and continuous scale for IBS symptoms 103 (76), Rome I and II One capsule containing L. rhamnosus GG, L. rhamnosus Lc705, Propionibacterium freudenreichii, and Bifidobacterium breve Bb99 (8 9 10 9 c.f.u. / capsule) o.d. for 6 months Method of randomization is stated. Method of concealment of allocation is not stated. Double-blind. Other IBS medications are allowed. Kim ( 47 ) USA, secondary care and advertising Continuous scale for IBS symptoms 48 (94), Rome II One packet containing VSL # 3 (450 billion bacteria / packet) b.i.d. for 4 8 weeks allocation stated. Double-blind. No other IBS medications allowed apart from antidepressants. Niv ( 52 ) Israel, secondary care Continuous scale for IBS symptoms 54 (67), Rome II One tablet containing L. reuteri ATCC 55730 (1 10 8 c.f.u. / tablet) q.i.d. for 2 weeks then b.i.d., thereafter for 6 months allocation are not stated. Double-blind. Other IBS medications are allowed. O Mahony ( 19 ) Eire, secondary care and advertising Continuous scale for IBS symptoms 80 (64), Rome II Malted drink containing L. salivarius UCC4331 (1 10 10 live bacteria / drink) or B. infantis 35624 (1 10 10 live bacteria / drink) o.d. for 8 weeks Method of randomization is stated. Method of concealment of allocation is not stated. Double-blind. No other IBS medications allowed. Kim ( 46 ) Korea, secondary care Continuous scale for IBS symptoms 40 (26), clinical diagnosis Medilac DS ( Bacillus subtilis and S. faecium ) for 4 weeks allocation are not stated. Blinding is not stated. Unclear if other IBS medications are allowed. Simren ( 56 ) Sweden, advertising Continuous scale for IBS symptoms 66 (63), Rome II Rose hip drink (400 ml) containing L. plantarum DSM 9843 (5 10 7 c.f.u. / ml) o.d. for 6 weeks allocation are not stated. Double-blind. Unclear if other IBS medications are allowed. Whorwell ( 60 ) UK, primary care Adequate relief of IBS symptoms, and continuous scale for IBS symptoms 362 (100), Rome II One capsule containing B. infantis 35624 (1 10 6 live bacteria / capsule, 1 10 8 live bacteria / capsule, or 1 10 10 live bacteria / capsule) o.d. for 4 weeks allocation are stated. Double-blind. Loperamide for diarrhea and bisocodyl for constipation are allowed. Guyonnet ( 42 ) France, primary care Continuous scale for IBS symptoms 274 (75), Rome II Fermented milk (125 g) containing B. animalis DN173 010 (1.25 10 10 c.f.u. / 125 g) S. thermophilus (1.2 10 9 c.f.u. / 125 g) and L. bulgaricus (1.2 10 9 c.f.u. / 125 g) b.i.d. for 6 weeks allocation are not stated. Double-blind. Other IBS medications are allowed (not fi ber or fermented dairy products). Drouault-Holowacz ( 35 ) France, not stated Satisfactory relief of global IBS symptoms, and continuous scale for IBS symptoms 106 (76), Rome II One sachet containing B. longum LA 101, L. acidophilus LA 102, L. lactis LA 103, and S. thermophilus LA 104 (1 10 10 c.f.u. / sachet) o.d. for 4 weeks allocation are stated. Double-blind. Unclear if other IBS medications are allowed. Table continued on following page

1552 Ford et al. Table 1. Continued Study Country and recruitment Criteria used to define symptom improvement following therapy Sample size ( % female) and diagnostic criteria for IBS Probiotic used and duration of therapy Methodology Enck ( 37 ) Germany, primary care 50 % improvement in IBS global symptoms 297 (49), Kruis score Autolysate of cells and cell fragments of Enterococcus faecalis DSM16440 and Escherichia coli DSM17252 (3.0-9.0 10 7 c.f.u. / 1.5 ml) 0.75 ml t.i.d. for 1 week, then 1.5 ml t.i.d. for weeks 2 and 3, then 2.25 ml t.i.d. for weeks 3 8 allocation are not stated. Double-blind. Other IBS medications are allowed (not antispasmodics). Kajander ( 45 ) Finland, primary care Continuous scale for IBS symptoms 86 (93), Rome II Milk-based drink (1.2 dl) containing L. rhamnosus GG ATCC 53103, L. rhamnosus Lc705 DSM 7061, P. freudenreichii, and B. animalis Bb12 DSM 15954 (1 10 7 c.f.u. / ml) o.d. for 20 weeks allocation are stated. Double-blind. Other IBS medications are allowed. Sinn ( 58 ) Korea, secondary care Any improvement in abdominal pain measured on a 10-point Likert scale 40 (65), Rome III One capsule containing L. acidophilus SDC 2012 and 2013 (2 10 9 c.f.u. / ml) b.i.d. for 4 weeks allocation are stated. Double-blind. No other IBS medications allowed. Zeng ( 62 ) China, tertiary care Continuous scale for IBS symptoms 29 (34), Rome II Fermented milk (200 ml) containing S. thermophilus (1 10 8 c.f.u. / ml), L. bulgaricus (1 10 7 c.f.u. / ml), L. acidophilus (1 10 7 c.f.u. / ml), and B. longum (1 10 7 c.f.u. / ml) b.i.d. for 4 weeks allocation are not stated. Patient-blinded. No other IBS medications allowed. Agrawal ( 29 ) UK, tertiary care Continuous scale for IBS symptoms 34 (100), Rome III Fermented milk (125 g) containing B. animalis DN173 010 (1.25 10 10 c.f.u. / 125 g) S. thermophilus (1.2 10 9 c.f.u. / 125 g) and L. bulgaricus (1.2 10 9 c.f.u. / 125 g) b.i.d. for 4 weeks allocation are not stated. Double-blind. Other IBS medications are allowed. Enck ( 38 ) Germany, primary care 50 % Improvement in IBS global symptoms 298 (56), Kruis score E. coli DSM17252 (1.5 4.5 10 7 c.f.u. / ml) 0.75 ml drops t.i.d. for 1 week, then 1.5 ml t.i.d. for weeks 2 8 allocation are not stated. Double-blind. Other IBS medications are allowed (not antispasmodics). Hong ( 43 ) Korea, tertiary care 50 % Improvement in IBS global symptoms 70 (33), Rome III One sachet containing B. bifidum BGN4, B. lactis AD011, L. acidophilus AD031, and L. casei IBS041 (20 billion bacteria / sachet) b.i.d. for 8 weeks allocation are stated. Double-blind. No other IBS medications allowed. Williams ( 61 ) UK, tertiary care Continuous scale for IBS symptoms 52 (87), Rome II One capsule containing L. acidophilus CUL- 60 NCIMB 30157 and CUL-21 NCIMB 30156, B. bifidum CUL-20 NCIMB 30153, and B. lactis CUL-34 NCIMB 30172 (2.5 10 10 c.f.u. / capsule) o.d. for 8 weeks allocation are not stated. Double-blind. No other IBS medications allowed. Simren ( 57 ) Sweden, tertiary care Adequate relief of global IBS symptoms for at least 50 % of weeks, and continuous scale for IBS symptoms 74 (70), Rome II Fermented milk (200 ml) containing L. paracasei ssp paracasei F19, L. acidophilus La5, and B. lactis Bb12 (5 10 7 c.f.u. / ml) b.i.d. for 8 weeks allocation are stated. Double-blind. Other IBS medications are allowed. Choi ( 32 ) Korea, tertiary care Continuous scale for IBS symptoms 90 (51), Rome II Two capsules containing Saccharomyces boulardii (2 10 11 live cells) b.i.d. for 4 weeks Method of randomization is stated. Method of concealment of allocation is not stated. Double-blind. No other IBS medications allowed. Guglielmetti ( 41 ) Germany, not stated Improvement in average weekly global IBS symptom score of 1 or more for 50 % of weeks 122 (67), Rome III One capsule containing B. bifidum MIMBb75 (1 10 9 c.f.u. / capsule) o.d. for 4 weeks allocation are stated. Double-blind. No other IBS medications allowed. Table continued on following page The American Journal of GASTROENTEROLOGY VOLUME 109 OCTOBER 2014 www.amjgastro.com

Prebiotics, probiotics, and synbiotics in IBS and CIC 1553 Table 1. Continued Study Country and recruitment Criteria used to define symptom improvement following therapy Sample size ( % female) and diagnostic criteria for IBS Probiotic used and duration of therapy Methodology Michail ( 50 ) USA, tertiary care Continuous scale for IBS symptoms 24 (67), Rome III VSL # 3 (900 billion bacteria / day) for 8 weeks allocation are stated. Double-blind. Unclear if other IBS medications are allowed. Ringel-Kulka ( 54 ) USA, population based 50-point decrease in IBS symptom severity score 33, Rome III One pill containing L. acidophilus NCFM and B. lactis Bi-07 (1 10 11 c.f.u. / pill) b.i.d. for 8 weeks Method of randomization is not stated. Method of concealment of allocation is stated. Double-blind. Other IBS medications are allowed. Sondergaard ( 59 ) Denmark and Sweden, primary and secondary care Adequate relief of global IBS symptoms, and continuous scale for IBS symptoms 64 (75), Rome II Fermented milk (250 ml) containing L. paracasei ssp paracasei F19, L. acidophilus La5, and B. lactis Bb12 (5 10 7 c.f.u. / ml) b.i.d. for 8 weeks allocation are stated. Double-blind. Unclear if other IBS medications are allowed. Cha ( 31 ) Korea, tertiary care Adequate relief of global IBS symptoms for at least 50 % of weeks, and continuous scale for IBS symptoms 50 (48), Rome III One capsule containing L. acidophilus KCTC11906BP, L. plantarum KCTC11867BP, L. rhamnosus KCTC11868BP, B. breve KCTC11858BP, B. lactis KCTC11903BP, B. longum KCTC11860BP, and S. thermophilus KCTC11870BP (5 billion bacteria / capsule) b.i.d. for 8 weeks Method of randomization is stated. Method of concealment of allocation is not stated. Doubleblind. No other IBS medications allowed. Cui ( 33 ) China, tertiary care Improvement in abdominal pain frequency 60 (70), Rome III Two capsules containing B. longum and L. acidophilus t.i.d. for 4 weeks allocation are not stated. Double-blind. No other IBS medications allowed. Dapoigny ( 34 ) France, tertiary care Reduction in IBS symptom severity score of 50 % or more 52 (70), Rome III Three capsules containing L. casei rhamnosus LCR35 (2 10 8 c.f.u. / capsule) o.d. for 4 weeks allocation are not stated. Double-blind. No other IBS medications allowed. Ducrotte ( 36 ) India, primary care Patients rated treatment effi cacy as excellent, and continuous scale for IBS symptoms 214 (29), Rome III One capsule containing L. plantarum LP299V DSM 9843 (10 billion c.f.u. / capsule) o.d. for 4 weeks allocation are stated. Double-blind. Unclear if other IBS medications are allowed. Farup ( 39 ) Norway, secondary care Satisfactory relief of symptoms (data obtained from the authors), and continuous scale for IBS symptoms 16 (69), Rome II One capsule containing L. plantarum MF1298 (10 10 c.f.u. / capsule) o.d. for 3 weeks allocation are stated. Double-blind. Unclear if other IBS medications are allowed. Kruis ( 49 ) Germany, tertiary care Patients reported very much satisfi ed or a little satisfi ed with treatment 120 (77), Rome II One capsule containing E. coli Nissle 1917 (2.5-25 10 9 c.f.u. / capsule) o.d. for 4 days then b.i.d. for 12 weeks allocation are stated. Double-blind. Other IBS medications are allowed (not laxatives or anti-diarrheals). Begtrup ( 30 ) Denmark, primary care Adequate relief of global IBS symptoms for at least 50 % of the time, and continuous scale for IBS symptoms 131 (74), Rome III Four capsules containing L. paracasei ssp paracasei F19, L. acidophilus La5, and B. lactis Bb12 (1.3 10 10 c.f.u. / capsule) o.d. for 6 months allocation are stated. Double-blind. Unclear if other IBS medications are allowed. Roberts ( 55 ) England, primary care Adequate relief of global IBS symptoms 184 (83), Rome III One pot containing B. lactis I-2494 (previously known as DN173 010) (1.25 10 10 c.f.u. / pot), S. thermophilus I-1630 (1.2 10 9 c.f.u. / pot), and L. bulgaricus I-1632 and I-1519 (1.2 10 9 c.f.u. / pot) b.i.d. for 12 weeks allocation are stated. Double-blind. Other IBS medications are allowed. ATCC, American Type Culture Collection; c.f.u, colony-forming units; IBS, irritable bowel syndrome; o.d., once daily; VAS, visual analog scale. a Does not meet Food and Agricultural Organization definition of a probiotic.

1554 Ford et al. Probiotics Control Risk ratio Risk ratio Study or subgroup Events Total Events Total Weight M-H, random, 95% Cl Year M-H, random, 95% Cl 1.1.1 Combination Kim, 2003 8 12 8 13 2.6% 1.08 [0.60, 1.95] 2003 Kajander, 2005 21 52 34 51 4.1% 0.61 [0.41, 0.89] 2005 Enck, 2008 47 149 92 148 5.3% 0.51 [0.39, 0.66] 2008 Drouault-Holowacz, 2008 33 53 31 53 4.9% 1.06 [0.78, 1.45] 2008 Hong, 2009 16 36 17 34 3.1% 0.89 [0.54, 1.46] 2009 Simren, 2010 23 37 27 37 4.8% 0.85 [0.62, 1.17] 2010 Ringel-Kulka, 2011 11 17 9 16 2.7% 1.15 [0.66, 2.01] 2011 Sondergaard, 2011 25 32 23 32 5.1% 1.09 [0.82, 1.44] 2011 Cha, 2012 13 25 22 25 3.9% 0.59 [0.39, 0.88] 2012 Cui, 2012 13 37 16 23 3.0% 0.51 [0.30, 0.84] 2012 Roberts, 2013 70 92 67 92 6.4% 1.04 [0.88, 1.24] 2013 Begtrup, 2013 32 67 38 64 4.7% 0.80 [0.58, 1.11] 2013 Subtotal (95% Cl) 609 588 50.6% 0.81 [0.67, 0.98] Total events 312 384 Heterogeneity: τ 2 = 0.07; χ 2 = 39.53, df = 11 (P < 0.0001); I 2 = 72% Test for overall effect: Z = 2.15 (P = 0.03) 1.1.2 Lactobacillus Nobaek, 2000 21 30 25 30 5.1% 0.84 [0.63, 1.12] 2000 Niedzielin, 2001 11 20 17 20 3.6% 0.65 [0.42, 1.00] 2001 Sinn, 2008 4 20 13 20 1.3% 0.31 [0.12, 0.78] 2008 Dapoigny, 2012 19 26 16 26 4.1% 1.19 [0.81, 1.74] 2012 Ducrotte, 2012 61 108 105 106 6.4% 0.57 [0.48, 0.67] 2012 Farup, 2012 6 9 3 7 1.2% 1.56 [0.59, 4.11] 2012 Subtotal, (95% Cl) 213 209 21.7% 0.75 [0.54, 1.04] Total events 122 179 Heterogeneity: τ 2 = 0.10; χ 2 = 20.20, df = 5 (P = 0.001); I 2 = 75% Test for overall effect: Z = 1.72 (P = 0.08) 1.1.3 Bifidobacterium Whorwell, 2006 143 270 54 92 6.0% 0.90 [0.74, 1.11] 2006 Guglielmetti, 2011 26 60 49 62 4.8% 0.55 [0.40, 0.75] 2011 Subtotal (95% Cl) 330 154 10.8% 0.71 [0.44,? 1.16] Total events 169 103 Heterogeneity: τ 2 = 0.11; χ 2 = 6.72, df = 1 (P = 0.010); I 2 = 85% Test for overall effect: Z = 1.35 (P = 0.18) 1.1.4 Escherichia Enck, 2009 121 148 143 150 7.1% 0.86 [0.79, 0.93] 2009 Kruis, 2012 33 60 37 60 4.9% 0.89 [0.66, 1.21] 2012 Subtotal (95% Cl) 208 210 12.1% 0.86 [0.79, 0.93] Total events 154 180 Heterogeneity: τ 2 = 0.00; χ 2 = 0.08, df = 1 (P = 0.78); I 2 = 0% Test for overall effect: Z = 3.65 (P = 0.0003) 1.1.5 Streptococcus Gade, 1989 20 32 19 22 4.8% 0.72 [0.53, 0.99] 1989 Subtotal (95% Cl) 32 22 4.8% 0.72 [0.53, 0.99] Total events 20 19 Heterogeneity: Not applicable Test for overall effect: Z = 2.01 (P = 0.04) Total (95% Cl) 1392 1183 100.0% 0.79 [0.70, 0.89] Total events 777 865 Heterogeneity: τ 2 = 0.05; χ 2 = 78.60, df = 22 (P < 0.00001); I 2 = 72% Test for overall effect: Z = 3.95 (P < 0.0001) Test for subqroup differences: χ 2 = 2.16, df = 4 (P = 0.71), I 2 = 0% 0.1 0.2 0.5 1 2 5 10 Favors probiotics Favors control Figure 2. Forest plot of randomized controlled trials of probiotics vs. placebo in irritable bowel syndrome (IBS): effect on persistence of symptoms. Lactobacillus was used in six trials (422 patients) ( 34,36,39,51, 53,58 ), with no clear benefit detected over placebo (RR = 0.75; 95 % CI 0.54 1.04), again with significant heterogeneity between studies ( I 2 = 75 %, P = 0.001). However, when only the three RCTs that used L. plantarum DSM 9843 were considered in the analysis ( 36,51,53 ), which contained 314 subjects, the RR of symptoms persisting was significantly lower with active therapy (0.67; 95 % CI 0.51 0.87), although the significant heterogeneity observed persisted ( I 2 = 63 %, P = 0.07). Bifidobacterium was studied in two RCTs (484 patients) ( 41,60 ), with no benefit over placebo (RR = 0.71; 95 % CI 0.44 1.16). Escherichia was assessed in two trials (418 patients) ( 38,49 ), with a benefit detected compared with placebo (RR = 0.86; 95 % CI 0.79 0.93), although only significantly so in the trial of E. coli DSM 17252 ( 38 ). Finally, S. faecium was used in one trial recruiting 54 patients, and appeared to be superior to placebo (RR = 0.72; 95 % CI 0.53 0.99 ( 40 )). Efficacy of probiotics in the treatment of IBS: effect on Global IBS or abdominal pain scores There were 24 separate trials ( 19,29 32,35,36,39,41,42,44 48, 50,52,53,56,57,59 62 ), making 25 comparisons, comprising 2001 patients that reported the effect of probiotics on global IBS or abdominal pain scores. There was a statistically significant effect of probiotics in reducing global symptoms or abdominal pain (SMD = 0. 2 5 ; 9 5 % C I 0.36 to 0.14) with no significant heterogeneity ( I 2 = 27 %, P = 0.11; Figure 3 ). There were six trials The American Journal of GASTROENTEROLOGY VOLUME 109 OCTOBER 2014 www.amjgastro.com

Prebiotics, probiotics, and synbiotics in IBS and CIC 1555 Probiotics Control Std. mean difference Std. mean difference Study or subgroup Mean s.d. Total Mean s.d. Total Weight IV, random, 95% Cl Year IV, random, 95% Cl 1.2.1 Combination Kim, 2005 Kajander, 2005 101.8 20.4 79.72 11 13.88 41 26.8 99.72 86.81 10 13.88 40 1.5% 4.6% 0.02 [ 0.83, 0.88] 0.46 [ 0.90, 0.02] Kim, 2005 102.4 47.03 24 125.3 52.79 24 3.1% 0.45 [ 1.02, 0.12] 2005 Kim, 2006 1.6 1.6 17 1.8 2.1 17 2.4% 0.10 [ 0.78, 0.57] 2006 Guyonnet, 2007 5.07 1.14 135 5.22 1.26 132 9.3% 0.12 [ 0.36, 0.12] 2007 Zeng, 2008 7.64 1.24 14 9.18 1.48 15 1.8% 1.09 [ 1.88, 0.30] 2008 Drouault Holowacz, 2008 2.71 2.16 48 3.34 2.24 52 5.4% 0.28 [ 0.68, 0.11] 2008 Kajander, 2008 24 16.73 43 30 18.01 43 4.8% 0.34 [ 0.77, 0.08] 2008 Williams, 2009 150.23 101.96 28 172 99.51 24 3.3% 0.21 [ 0.76, 0.33] 2009 Agrawal, 2009 2.9 0.9 17 3.7 0.9 15 2.0% 0.87 [ 1.60, 0.14] 2009 Simren, 2010 206 113 33 228 125 34 4.1% 0.18 [ 0.66, 0.30] 2010 Sondergaard, 2011 176 138 27 206 124 25 3.3% 0.22 [ 0.77, 0.32] 2011 Michail, 2011 1.5 0.3 15 1.7 0.8 9 1.6% 0.36 [ 1.19, 0.48] 2011 Cha, 2012 1.56 1.21 24 1.97 1.65 23 3.1% 0.28 [ 0.85, 0.30] 2012 Begtrup, 2013 2.9 1.1 54 2.8 1 44 5.3% 0.09 [ 0.30, 0.49] 2013 Subtotal (95% Cl) 531 507 55.8% 0.24 [ 0.37, 0.12] Heterogeneity: τ 2 = 0.00; χ 2 = 13.30, df = 14 (P = 0.50); l 2 = 0% Test for overall effect: Z = 3.90 (P < 0.0001) 2005 2005 1.2.2 Lactobacillus Nobaek, 2000 3.9 1 25 4.26 1.67 27 3.3% 0.26 [ 0.80, 0.29] 2000 Niv, 2005 270 139 21 230 139 18 2.6% 0.28 [ 0.35, 0.91] 2005 O'Mahony, 2005 5.25 2.8 26 5.68 2.8 25 3.3% 0.15 [ 0.70, 0.40] 2005 Simren, 2006 279 129 29 245 118 29 3.6% 0.27 [ 0.25, 0.79] 2006 Farup, 2012 6.18 1.83 9 5.61 1.31 7 1.2% 0.33 [ 0.67, 1.33] 2012 Ducrotte, 2012 0.68 0.53 105 0.92 0.57 99 8.2% 0.43 [ 0.71, 0.16] 2012 Subtotal (95% Cl) 215 205 22.2% 0.08 [ 0.38, 0.21] Heterogeneity: τ 2 = 0.06; χ 2 = 9.30, df = 5 (P =0.10); l 2 = 46% Test for overall effect: Z = 0.54 (P = 0.59) 1.2.3 Bifidobacterium O'Mahony, 2005 3.7 2.88 24 5.68 2.8 25 3.0% 0.69 [ 1.26, 0.11] 2005 Whorwell, 2006 2.01 0.82 250 2.09 0.89 80 9.0% 0.10 [ 0.35, 0.16] 2006 Guglielmetti, 2011 2.07 0.85 60 2.63 0.74 62 5.9% 0.70 [ 1.07, 0.33] 2011 Subtotal (95% Cl) 334 167 18.0% 0.46 [ 0.92, 0.00] Heterogeneity: τ 2 = 0.12; χ 2 = 8.70, df = 2 (P = 0.01); l 2 = 77% Test for overall effect: Z = 1.97 (P = 0.05) 1.2.4 Saccharomyces Choi, 2011 1.2 0.8 34 1.3 0.8 Subtotal (95% Cl) 34 33 33 4.1% 4.1% 0.12 [ 0.60, 0.36] 0.12 [ 0.60, 0.36] 2011 Heterogeneity: Not applicable Test for overall effect: Z = 0.51 (P = 0.61) Total (95% Cl) 1114 912 100.0% 0.25 [ 0.36, 0.14] Heterogeneity: τ 2 = 0.02; χ 2 = 32.74, df = 24 (P = 0.11); l 2 = 27% Test for overall effect: Z = 4.44 (P < 0.00001) Test for subgroup difference: χ 2 = 2.18, df = 3 (P = 0.54), l 2 = 0% 4 2 0 2 4 Favors probiotics Favors control Figure 3. Forest plot of randomized controlled trials of probiotics vs. placebo in irritable bowel syndrome (IBS): effect on global symptom or abdominal pain scores. (420 patients) that evaluated Lactobacillus ( 19,36,39,52,53,56 ), and three trials (501 patients) that investigated Bifidobacterium ( 19,41,60 ), and neither were statistically significantly more efficacious than placebo ( Figure 3 ), although there was a trend toward a benefit for the latter (SMD 0.46; 95 % CI 0.92 to 0, P = 0.05). When only the three trials that used L. plantarum DSM 9843 were considered in the analysis, there was no benefit in 314 patients (SMD = 0.18; 95 % CI 0.60 to 0.25) ( 36,53,56 ). Similarly, when only the two trials that used B. infantis 35624 were included in the analysis there was no benefit in 379 patients (SMD = 0.33; 95 % CI 0.90 to 0.24) ( 9,60 ). There were 15 trials ( 29 31,35,42,44 48,50,57,59,61,62 ), evaluating 1,038 patients, using combinations of probiotics that did suggest a significant improvement in IBS symptom scores with active treatment (SMD 0.24; 95 % CI 0.37 to 0.12; Figure 3 ). When specific combinations were studied, three trials used VSL # 3 in 93 patients, with no significant benefit over placebo (SMD 0.32; 95 % CI 0.73 to 0.10) ( 46,47,50 ); three trials used a combination of L. paracasei ssp paracas ei F19, L. acidophilus L a5, and B. lactis Bb12 in 217 patients with no benefit over placebo (SMD = 0.07; 95 % CI 0.34 to 0.20) ( 30,57,59 ); and two trials used a combination of B. lactis DN-173 010, S. thermophilus, and L. bulgaricus

1556 Ford et al. Probiotics Control Std. mean difference Study or Subgroup Mean s.d. Total Mean s.d. Total Weight IV, random, 95% Cl Year 1.3.1 Combination Kim, 2003 22.32 21.93 11 27.3 24.42 10 1.9% 0.21 [ 1.07, 0.65] 2003 Kim, 2005 32.24 18.05 24 35.8 15.27 24 4.1% 0.21 [ 0.78, 0.36] 2005 Kajander, 2005 5.1 4.25 41 6.7 4.19 40 6.3% 0.38 [ 0.82, 0.06] 2005 Kim, 2006 2.2 1.44 17 2.58 0.86 17 3.0% 0.31 [ 0.99, 0.36] 2006 Guyonnet, 2007 3.13 1.12 135 3.06 1.17 132 15.1% 0.06 [ 0.18, 0.30] 2007 Zeng, 2008 32.1 4.53 14 29.67 3.91 15 2.5% 0.56 [ 0.18, 1.30] 2008 Williams, 2009 25.88 25.05 28 32.05 29.64 24 4.4% 0.22 [ 0.77, 0.32] 2009 Agrawal, 2009 3.2 1.2 17 3.9 0.9 15 2.7% 0.64 [ 1.35, 0.08] 2009 Simren, 2010 41 33 33 43 30 34 5.5% 0.06 [ 0.54, 0.42] 2010 Sondergaard, 2011 33.3 33.3 27 41.1 30.2 25 4.4% 0.24 [ 0.79, 0.30] 2011 Michail, 2011 1.6 0.3 15 1.5 0.9 9 2.0% 0.16 [ 0.67, 0.99] 2011 Cha, 2012 1.91 1.48 24 2.41 2.2 23 4.0% 0.26 [ 0.84, 0.31] 2012 Begtrup, 2013 3.7 1.4 54 3.5 1.3 44 7.4% 0.15 [ 0.25, 0.54] 2013 Subtotal (95% Cl) 440 412 63.3% 0.08 [ 0.21, 0.06] Heterogeneity: τ 2 =0.00; χ 2 = 11.53, df = 12 (P = 048); l 2 = 0% Test for overall effect: Z = 1.11 (P = 0.27) Std. mean difference IV, random, 95% Cl 1.3.2 Lactobacillus O'Mahony, 2005 15.32 12.44 26 17.04 15.7 25 4.3% 0.12 [ 0.67, 0.43] 2005 Subtotal (95% Cl) 26 25 4.3% 0.12 [ 0.67, 0.43] Heterogeneity: Not applicable Test for overall effect: Z = 0.43 (P = 0.67) 1.3.3 Bifidobacteriun O'Mahony, 2005 Whorwell, 2006 Guglielmetti, 2011 Subtotal (95% Cl) 11.66 1.94 2.09 11.51 0.91 0.88 24 250 60 334 17.04 1.96 2.61 15.7 0.89 0.93 Heterogeneity: τ 2 = 0.08; χ 2 = 6.29, df = 2 (P = 0.04); l 2 = 68% Test for overall effect: Z = 1.53 (P = 0.13) 25 80 62 167 4.1% 14.3% 8.6% 27.0% 0.38 [ 0.95, 0.18] 0.02 [ 0.27, 0.23] 0.57 [ 0.93, 0.21] 0.30 [ 0.68, 0.09] 2005 2006 2011 1.3.4 Saccharomyces Choi, 2011 Subtotal (95% Cl) 1.7 1.3 34 34 Heterogeneity: Not applicable Test for overall effect: Z = 1.48 (P = 1.14) 2.2 1.4 33 33 5.4% 5.4% 0.37 [ 0.85, 0.12] 0.37 [ 0.85, 0.12] 2011 Total (95% Cl) 834 Heterogeneity: τ 2 = 0.01; χ 2 = 20.19, df = 17 (P = 0.26); l 2 = 16% Test for overall effect: Z = 2.46 (P = 0.01) Test for subgroup differences: χ 2 = 2.22, df = 3 (P = 0.53), l 2 = 0% 637 100.0% 0.15 [ 0.27, 0.03] 4 2 0 2 4 Favors probiotics Favors control Figure 4. Forest plot of randomized controlled trials of probiotics vs. placebo in irritable bowel syndrome (IBS): effect on bloating scores. in 299 patients, again with no significant benefit over placebo (SMD = 0.41; 95 % CI 1.12 to 0.30) ( 29,42 ). Efficacy of probiotics in the treatment of IBS: effect on individual symptom scores There were 17 separate trials ( 19,29 32,41,42,44,46 48,50,57, 59 62 ), making 18 comparisons, and containing 1,446 patients, that reported the effect of probiotics on bloating symptom scores. Overall, bloating scores were significantly reduced with probiotics (SMD = 0. 1 5 ; 9 5 % C I 0.27 to 0. 0 3 ; Figure 4 ), with no significant heterogeneity between individual study results ( I 2 = 16 %, P = 0.26). Ten trials reported continuous data for the effect of probiotics on flatulence symptom scores in 741 patients ( 29,31,32,44, 46 48,53,60,62 ). Flatulence scores were significantly lower with probiotics compared with placebo (SMD = 0.23; 95 % CI 0.38 to 0.07; Figure 5 ), wit h no significant heterogeneity detected ( I 2 = 0 %, P = 0.63). Finally, six RCTs rep or ted t he effect of probiotics on urgency symptom scores in 635 patients ( 31,32,41,46,47,60 ). There was no apparent benefit detected for probiotics, in terms of effect on symptoms of urgency (SMD = 0.10; 95 % CI 0.30 to 0.10), with no significant heterogeneity detected ( I 2 = 20 %, P = 0.28). Adverse events with probiotics Total adverse events were reported by 24 RCTs ( 30 32, 34, 36 38, 40 47,49 53,57,58,61,62 ), containing 2,407 patients. Overall, 201 (16.5 % ) of 1,215 patients allocated to probiotics experienced any adverse event, compared with 164 (13.8 % ) of 1,192 patients assigned to placebo. The RR of experiencing any adverse event The American Journal of GASTROENTEROLOGY VOLUME 109 OCTOBER 2014 www.amjgastro.com

Prebiotics, probiotics, and synbiotics in IBS and CIC 1557 Study or subgroup Probiotics Mean s.d. Control Total Mean s.d. Total Std. mean difference Weight IV, random, 95% Cl 1.4.1 Combination Kim, 2003 27.63 27 11 28.35 28.13 10 3.3% 0.03 [ 0.88, 0.83] 2003 Kajander, 2005 8.2 4.74 41 9.5 4.68 40 12.4% 0.27 [ 0.71, 0.16] 2005 Kim, 2005 31.4 17.35 24 39.63 16.2 24 7.2% 0.48 [ 1.06, 0.09] 2005 Kim, 2006 6 5.36 17 6.6 7.01 17 5.3% 0.09 [ 0.77, 0.58] 2006 Zeng, 2008 32.5 8.11 14 37.62 7.76 15 4.3% 0.63 [ 1.38, 0.12] 2008 Agrawal, 2009 3.1 1.1 17 3.4 1.1 15 4.9% 0.27 [ 0.96, 0.43] 2009 Cha, 2012 2.38 2.25 24 3.02 2.16 23 7.2% 0.29 [ 0.86, 0.29] 2012 Subtotal (95% Cl) 148 144 44.5% 0.30 [ 0.53, 0.07] Heterogeneity: τ 2 = 0.00; χ 2 = 1.90 df = 6 (P = 0.93); I 2 = 0% Test for overall effect: Z = 2.56 (P = 0.01) Year Std. mean difference IV, random, 95% Cl 1.4.2 Lactobacillus Nobaek, 2000 Subtotal (95% Cl) 3.4 1 Heterogeneity: Not applicable Test for overall effect: Z = 2.38 (P = 0.02) 25 25 4.48 1.94 27 27 7.6% 7.6% 0.68 [ 1.24, 0.12] 0.68 [ 1.24, 0.12] 2000 1.4.3 Bifidobacterium Whorwell, 2006 Subtotal (95% Cl) 2.01 0.82 250 250 Heterogeneity: Not appplicable Test for overall effect: Z = 0.29 (P = 0.78) 2.04 0.8 80 80 37.6% 37.6% 0.04 [ 0.29, 0.22] 0.04 [ 0.29, 0.22] 2006 1.4.4 Saccharomyces Choi, 2011 Subtotal (95% Cl) 1.7 1.4 Heterogeneity: Not appplicable Test for overall effect: Z = 1.15 (P = 0.25) 34 34 2.1 1.4 33 33 10.3% 10.3% 0.28 [ 0.76, 0.20] 0.28 [ 0.76, 0.20] 2011 Total (95% Cl) 457 284 Heterogeneity: τ 2 = 0.00; χ 2 = 7.08, df = 9 (P = 0.63); I 2 = 0% Test for overall effect: Z = 2.91 (P = 0.004) Test for subgroup differences: χ 2 = 5.18, df = 3 (P = 0.16), I 2 = 42.1% 100.0% 0.23 [ 0.38, 0.07] 4 2 0 2 4 Favors probiotics Favors control Figure 5. Forest plot of randomized controlled trials of probiotics vs. placebo in irritable bowel syndrome (IBS): effect on fl atulence scores. was significantly higher with probiotics (1.21; 95 % CI 1.02 1.44). The number needed to harm with probiotics was 35 (95 % CI 16 362). Efficacy and safety of synbiotics in IBS The two RCTs of synbiotics in IBS recruited a total of 198 patients ( 63,64 ). The first RCT was a single-blind RCT conducted in Italy, recruiting 68 patients with Rome II-defined IBS, and it used a combination of L. acidophilus a n d L. helveticus, w it h Bifidobacterium species, in a vitamin and phytoextract-enriched medium for 12 weeks ( 63 ). The second RCT, conducted in South Korea, used B. lactis i n c ombi n at i on w it h a c a c i a fiber for 8 weeks, in 130 patients who met the Rome III criteria for IBS ( 64 ). This doubleblind trial was at an unclear risk of bias owing to failure to report the method used to conceal treatment allocation. Only one trial reported dichotomous data. There were 7 (20.6 % ) of 34 patients assigned to synbiotics with persistent symptoms, compared with 30 (88.2 % ) of 34 patients assigned to control therapy ( P < 0.01) ( 63 ). Both trials assessed IBS symptoms on a continuous scale in 185 patients ( 63,64 ). There was no statistically signifi cant effect of synbiotics in reducing symptoms, even though both trials were individually positive, owing to significant heterogeneity between studies (SMD = 1. 7 3 ; 9 5 % C I 3. 7 3 t o 0. 2 7, I 2 = 96 %, P = 0. 0 9 ; Supplementary Figure 1 online). Adverse events were reported in both studies, and no significant events occurred in either treatment arm. Efficacy and safety of prebiotics in CIC The trial of prebiotics vs. placebo in CIC recruited 60 female volunteers ( 65 ). Patients were randomized to receive 3 weeks of a 15 g per day mixture of inulin and partially hydrolyzed guar gum or placebo. The trial was at an unclear risk of bias, as the methods used to randomize and conceal treatment allocation were not reported. There was no difference in satisfaction in relief of constipation in the prebiotic group (9 (32.1 % ) of 28 satisfied) vs. the placebo group (10 (31.3 % ) of 32), and the mean number of bowel movements per week was also not statistically different (5.95 ± 2.50 in the prebiotic group, compared with 6.70 ± 3.83 in the control group). Adverse events data were not reported. Efficacy and safety of probiotics in CIC We identified three trials evaluating probiotics in 245 CIC patients ( 66 68 ). Two trials were at an unclear risk of bias ( 66,68 ),