Inflammatory bowel disease and the gastrointestinal microbiota

The causes of IBD have not been fully elucidated, but are thought to involve both genetic and environmental factors. More recently, a number of observations have suggested that the intestinal microbiota may drive intestinal inflammation. First, experimental animal models of IBD remain healthy when reared in a germ-free environment and develop inflammation only after colonisation with commensal bacteria (Onderdonk et al. Reference Onderdonk, Franklin and Cisneros1981; Sellon et al. Reference Sellon, Tonkonogy, Schultz, Dieleman, Grenther, Balish, Rennick and Sartor1998). Second, higher concentrations of bacteroides and enterobacteria are found in regions of inflamed mucosa in CD (Swidsinski et al. Reference Swidsinski, Ladhoff, Pernthaler, Swidsinski, Loening-Baucke, Ortner, Weber, Hoffmann, Schreiber, Dietel and Lochs2002). Third, in both CD (Harper et al. Reference Harper, Lee, Kettlewell, Bennett and Jewell1985; D'Haens et al. Reference D'Haens, Geboes, Peeters, Baert, Penninckx and Rutgeerts1998) and pouchitis (de Silva et al. Reference de Silva, Millard, Soper, Kettlewell, Mortensen and Jewell1991) disease activity deteriorates when the faecal stream is surgically restored to previously-defunctioned intestine, implying that the lumen contents are necessary for the development of inflammation. Finally, the association between ileal CD and mutations in the bacterial-sensing caspase-activating recruitment domain 15 gene implicate bacterial recognition as a central pathogenic process (Ogura et al. Reference Ogura, Bonen, Inohara, Nicolae, Chen and Ramos2001).

In contrast, there is also evidence that the commensal microbiota protect the mucosa from inflammation by decreasing intestinal permeability, increasing epithelial defence mechanisms and promoting an immunoregulatory acquired immune response (Madsen et al. Reference Madsen, Doyle, Jewell, Tavernini and Fedorak1999; Macpherson & Uhr, Reference Macpherson and Uhr2004; Rakoff-Nahoum et al. Reference Rakoff-Nahoum, Paglino, Eslami-Varzaneh, Edberg and Medzhitov2004). The mechanisms by which commensal microbiota interact with the gut immune system have been the focus of recent interest. There is evidence that metabolites produced by bacteria may interact directly with gut epithelial cells to enhance mucosal integrity (Venkatraman et al. Reference Venkatraman, Ramakrishna, Pulimood, Patra and Murthy2000; Stempelj et al. Reference Stempelj, Kedinger, Augenlicht and Klampfer2007). In addition, commensal bacteria may compete with pro-inflammatory species to decrease access to the mucosa (Lee et al. Reference Lee, Puong, Ouwehand and Salminen2003). Finally, microbial signature molecules, termed pathogen-associated molecular patterns, can interact with pattern recognition receptors such as toll-like receptors expressed on many host cells, including both epithelial cells and lamina propria dendritic cells (Stagg et al. Reference Stagg, Hart, Knight and Kamm2003).

Despite this evidence for a microbial influence in pathogenesis, drugs conventionally used to induce and maintain remission in IBD, such as aminosalicylates (5-aminosalicylic acid drugs), corticosteroids, immunosuppressants (azathioprine and methotrexate) and infliximab (a chimeric monoclonal antibody), primarily act to suppress the enhanced immune response and do not affect the microbiota. Antibiotic-mediated microbial manipulation has some efficacy, especially in active CD and pouchitis, but cannot be used to maintain remission because of lack of long-term efficacy and side effects (Sutherland et al. Reference Sutherland, Singleton, Sessions, Hanauer, Krawitt and Rankin1991; Borgaonkar et al. Reference Borgaonkar, MacIntosh, Fardy and Simms2000). None of these treatments is effective in all patients and all treatments are associated with major side effects. Therapies such as prebiotics and probiotics that selectively manipulate the gastrointestinal microbiota present a novel treatment option with a low side-effect burden.

Probiotics are living microbes that when introduced to the gut are capable of benefiting the host, and include organisms such as lactobacilli, bifidobacteria, Gram-positive cocci, enterococci and yeast species such as Saccharomyces boulardii (Fuller, Reference Fuller1989). Prebiotics are selectively fermented short-chain carbohydrates that allow specific changes both in the composition and/or activity of the gastrointestinal microbiota that confers health benefit, and include fructo-oligosaccharides and galacto-oligosaccharides (Gibson & Roberfroid, Reference Gibson and Roberfroid1995). Combinations of probiotics and prebiotics are termed synbiotics. Evidence from animal models of IBD indicates that probiotics can alter the intestinal microbiota and ameliorate disease (Sartor, Reference Sartor2004). Similarly, prebiotics have been shown to enhance lumenal immunoregulatory bacteria, reduce the activity of pro-inflammatory transcription factors (e.g. NF-κB) and attenuate inflammation (Holma et al. Reference Holma, Juvonen, Asmawi, Vapaatalo and Korpela2002; Cavin et al. Reference Cavin, Delannoy, Malnoe, Debefve, Touche, Courtois and Schilter2005). In addition, prebiotics produce SCFA, such as acetate and butyrate (Probert et al. Reference Probert, Apajalahti, Rautonen, Stowell and Gibson2004), that inhibit mucosal inflammation, impacting on both epithelial and dendritic cell function (Kinoshita et al. Reference Kinoshita, Suzuki and Saito2002; Millard et al. Reference Millard, Mertes, Ittelet, Villard, Jeannesson and Bernard2002).

The present review aims to assess the evidence that these observations from in vitro experiments and animal models can be put into clinical practice. The evidence presented relates to trials that have assessed these therapies in terms of clinical outcomes in IBD, and the data extracted and presented are the clinical end points in each study.

Probiotics and prebiotics in pouchitis

The most compelling evidence for the use of probiotics in IBD comes from randomised double-blind placebo-controlled trials of VSL#3 (a mixture of four species of lactobacilli, three species of bifidobacteria and Streptococcus thermophilus; Vsl Pharmaceuticals Inc., Gaithersburg, MD, USA) in patients with pouchitis (Table 1). Gionchetti et al. (Reference Gionchetti, Rizzello, Venturi, Brigidi, Matteuzzi, Bazzocchi, Piggoli, Miglioli and Campieri2000) have assessed the efficacy of VSL#3 as a maintenance treatment in forty patients with chronic relapsing pouchitis after antibiotic-induced remission. After 4 months fewer relapses were found to occur in the intervention group than in the control group. Moreover, all patients were subsequently found to relapse 3 months after cessation of VSL#3. The same group (Gionchetti et al. Reference Gionchetti, Rizzello, Helwig, Venturi, Lammers, Brigidi, Vitali, Piggoli, Miglioli and Campieri2003) have assessed VSL#3 in the primary prevention of pouchitis in forty patients following surgery. The incidence of pouchitis was found to be reduced and the quality of life improved in the VSL#3-treated group compared with the placebo group at the end of the intervention period; these beneficial effects were associated with faecal colonisation with the probiotics. Finally, a further study (Mimura et al. Reference Mimura, Rizzello, Helwig, Poggioli, Schreiber, Talbot, Nicholls, Gionchetti, Campieri and Kamm2004) has confirmed the effectiveness of VSL#3 as maintenance therapy in patients with recurrent or chronic pouchitis. Importantly, these studies demonstrate high relapse rates of pouchitis in the placebo group, perhaps reflecting intensive patient follow-up resulting in the detection of relapse that may have otherwise remained subclinical.

Table 1.

Clinical outcomes of published studies investigating the efficacy of probiotics and prebiotics in the treatment of pouchitis

R, randomised; DB, double blind; CO, cross-over study; OL, open label; VSL#3, a mixture of four species of lactobacilli, three species of bifidobacteria and Streptococcus thermophilus; LGG, Lactobacillus rhamnosus GG; FOS, fructo-oligosaccharide; Pl, placebo; HC, historical controls; CFU, colony-forming units; NR, not reported.

In contrast, Shen et al. (Reference Shen, Brzezinski, Fazio, Remzi, Achkar, Bennett, Sherman and Lashner2005) have reported no significant benefit of VSL#3 in maintaining antibiotic-induced remission in thirty-one patients. After 8 months twenty-three patients had discontinued VSL#3 because of symptom recurrence and two had discontinued because of side effects. In the six patients who remained on VSL#3 endoscopic evidence of mild to moderate pouchitis was found to persist even though symptom scores were not significantly different from baseline. Although the fact that patients purchased, stored and administered the VSL#3 themselves may have resulted in lower compliance, this methodology may more faithfully reproduce patients' use of probiotics outside clinical trials.

Trials of other probiotics in the management of pouchitis have yielded mixed results. One observational study of patients receiving Lactobacillus rhamnosus GG (LGG) after pouch formation (Gosselink et al. Reference Gosselink, Schouten, van Lieshout, Hop, Laman and Ruseler-van Embden2004) has reported a lower rate of pouchitis than in historical controls. However, Kuisma et al. (Reference Kuisma, Mentula, Jarvinen, Kahri, Saxelin and Farkkila2003) have found no difference in mean pouchitis scores between placebo and LGG-treated groups at the end of a 3-month study period, despite an increase in faecal lactobacilli:anaerobes. Finally, a reduction in endoscopic and clinical disease activity associated with an increase in faecal probiotic species has been demonstrated (Laake et al. Reference Laake, Bjorneklett, Aamodt, Aabakken, Jacobsen, Bakka and Vatn2005) in fifty-one patients with pouchitis after surgery for ulcerative colitis who consumed a fermented milk containing lactobacilli and bifidobacteria. These studies suggest that therapeutic benefit requires a combination of probiotic species (as with VSL#3) or that the component(s) responsible for the anti-inflammatory effect in combination preparations have specific properties that monotherapy probiotics do not. This concept supports the use of prebiotics that increase concentrations of several commensal immunoregulatory bacteria.

Two studies have examined the use of prebiotics or synbotics in pouchitis. Ten patients with antibiotic-refractory or antibiotic-dependent pouchitis treated with a combination of LGG and fructo-oligosaccharide in an open-label study were reported to experience complete clinical and endoscopic remission (Friedman & George, Reference Friedman and George2000). Welters et al. (Reference Welters, Heineman, Thunnissen, van den Bogaard, Soeters and Baeten2002) have reported that intake of inulin is associated with a decrease in histological and endoscopy scores in twenty patients in a cross-over study with a 4-week wash-out period between 3-week study phases. No patient had overt clinical pouchitis at enrolment and no change in clinical activity scores was detected, although overall pouchitis disease activity index was found to be significantly reduced during the inulin period (P=0·01). Prebiotic use was shown to be associated with a reduction in the faecal concentration of Bacteroides fragilis (P=0·02), but had no effect on lactobacilli or bifidobacteria. In addition, inulin was found to increase faecal butyrate concentrations (P=0·01) and reduce pH, with no effect on secondary bile acid concentration.

Probiotics and prebiotics as maintenance therapy in ulcerative colitis

Many trials have examined the use of probiotics to maintain remission in patients with UC. Three controlled trials (Kruis et al. Reference Kruis, Schutz, Fric, Fixa, Judmaier and Stolte1997, Reference Kruis, Fric, Pokrotnieks, Lukas, Fixa and Kascak2004; Rembacken et al. Reference Rembacken, Snelling, Hawkey, Chalmers and Axon1999) have demonstrated that Escherichia coli Nissle 1917 has similar efficacy to conventional 5-aminosalicylic acid (mesalazine) treatment, but the evidence to support the use of other probiotic species is less persuasive (Table 2). Kruis et al. (Reference Kruis, Schutz, Fric, Fixa, Judmaier and Stolte1997) have demonstrated equivalence of mesalazine and E. coli Nissle 1917 in patients with UC in remission over 12 weeks. However, studying the maintenance of remission over a short follow-up period is of limited clinical applicability. Rembacken et al. (Reference Rembacken, Snelling, Hawkey, Chalmers and Axon1999) have treated patients for a 12-month period and have also demonstrated equivalence of mesalazine v. E. coli Nissle 1917. Remission was induced using standard medical treatment (including topical and systemic corticosteroids) alongside the study intervention. No significant difference was found in the time to achieve remission or mean duration of remission, although the lack of a placebo group may make it difficult to distinguish a small difference in effectiveness from the effect of potent doses of steroid. Finally, no significant difference was found in the rate of relapse in a study comparing E. coli Nissle 1917 with mesalazine in 327 patients using no concomitant medication (Kruis et al. Reference Kruis, Fric, Pokrotnieks, Lukas, Fixa and Kascak2004). Probiotic therapies such as E. coli Nissle 1917 that have equivalent efficacy but fewer side effects than conventional drug therapies present an appealing alternative for patients and physicians.

Table 2.

Clinical outcomes of published studies investigating the efficacy of probiotics in maintenance of remission in ulcerative colitis

R, randomised; DB, double blind; DD, double dummy; OL, open label; E. coli N, Escherichia coli Nissle 1917; CFU, colony-forming units; VSL#3, a mixture of four species of lactobacilli, three species of bifidobacteria and Streptococcus thermophilus; LGG, Lactobacillus rhamnosus GG; L. salivarius, Lactobacillus salivarius; B. infantis, Bifidobacterium infantis; Pl, placebo; NR, not reported; SE, significant equivalence.

Several studies examining the use of lactobacilli or bifidobacteria as maintenance treatment in UC have demonstrated conflicting results. In patients randomised to three open-label arms (LGG v. mesalazine v. LGG plus mesalazine) it was demonstrated (Zocco et al. Reference Zocco, dal Verme, Cremonini, Piscaglia, Nista and Candelli2006) that both the lactobacillus arms experience a prolonged relapse-free time compared with 5-aminosalicylic acid treatment alone, despite no significant differences in overall relapse rates and clinical, histological or endoscopy scores. In a study of thirty patients with UC in medically-induced remission (Cui et al. Reference Cui, Chen, Wang, Yang, Cun, Wu, Liu, Dan, Jian and Chen2004) those receiving supplements of bifidobacteria were found to have fewer relapses during the 2-month follow-up period than those receiving placebo (20% v. 93% respectively). In addition, those receiving the probiotic were found to have increases in the concentration of faecal lactobacilli and bifidobacteria, a reduction in the expression of the pro-inflammatory transcription factor NF-κB and enhancement of anti-inflammatory cytokine release. Similarly, Ishikawa et al. (Reference Ishikawa, Akedo, Umesaki, Tanaka, Imaoka and Otani2003) have demonstrated a reduction in the number of disease exacerbations in a group of Japanese patients receiving fermented milk containing Bifidobacterium breve, Bifidobacterium bifidum and Lactobacillus acidophilus compared with placebo. However, this clinical benefit was not found to be associated with an increase in steroid-free remission or endoscopic improvement in disease activity. In contrast, a recently presented study (Shanahan et al. Reference Shanahan, Guarner, von Wright, Vilpponen-Salmela, O'Donoghue and Kiely2006) comparing Lactobacillus salivarius with Bifidobacterium infantis or placebo in 157 patients with UC in medically-induced remission has demonstrated no significant benefit of either of the probiotics when compared with placebo. Finally, treatment with VSL#3 to maintain remission in UC (Venturi et al. Reference Venturi, Gionchetti, Rizzello, Johansson, Zucconi, Brigidi, Matteuzzi and Campieri1999) was found to result in an increase in faecal concentrations of VSL#3 bacteria and a reduction in faecal pH. At the end of the study only four of twenty patients had experienced relapse.

There are few data on the efficacy of prebiotics or synbiotics in the maintenance of remission in UC.

Probiotics and prebiotics to treat active ulcerative colitis

Four diverse studies have examined the use of probiotics to treat active UC (Table 3). Tursi et al. (Reference Tursi, Brandimarte, Giorgetti, Forti, Modeo and Gigliobianco2004) have compared low-dose balsalazide+VSL#3 v. medium-dose balsalazide alone v. mesalazine alone in patients with mild- to moderately-active UC. Remission rates and time to remission were found to be shortest in the combination group. These results suggest that probiotics may augment the effect of 5-aminosalicylic acid drugs, although the balsalazide doses used in this study were lower than those generally used in clinical practice. In a second study in which clinical benefit of VSL#3 was also demonstrated (Bibiloni et al. Reference Bibiloni, Fedorak, Tannock, Madsen, Gionchetti and Campieri2005) VSL#3-specific organisms were detected in only three of eleven patients after microbiological analysis of mucosal biopsies. This finding suggests that the clinical effect observed was not the result of a direct influence of VSL#3-derived organisms at the mucosa.

Table 3.

Clinical outcomes of published studies investigating the efficacy of probiotics in induction of remission in patients with active ulcerative colitis

OL, open label; R, randomised; DB, double blind; VSL#3, a mixture of four species of lactobacilli, three species of bifidobacteria and Streptococcus thermophilus; CFU, colony-forming units; B. longum, Bifidobacterium longum; E. coli N, Escherichia coli Nissle 1917; DB, double blind; Pl, placebo; NR, not reported.

Efficacy of direct delivery of the probiotic to the colon with E. coli Nissle 1917 enemas in left-sided UC has been demonstrated (Matthes et al. Reference Matthes, Krummenerl, Giensch, Wolff and Schulze2006). A significant dose-dependent response was found and the time to remission was also reported to be shortest in the group receiving the highest dose of enema. A fourth probiotic was studied by Kato et al. (Reference Kato, Mizuno, Umesaki, Ishii, Sugitani, Imaoka, Otsuka, Hasunuma, Kurihara, Iwasaki and Arakawa2004), who treated active UC with fermented milk (B. breve strain Yakult, B. bifidum strain Yakult and a L. acidophillus strain). Although no difference was found in either remission rates or histological and endoscopy scores between the groups, the clinical activity scores were found to be lower in the treatment group than in the placebo group at the end of the study period.

A series of small open-label studies have evaluated a prebiotic germinated barley foodstuff (a glutamine-rich protein and dietary fibre, mainly hemicellulose) and have found evidence for therapeutic efficacy. For example, when the germinated barley foodstuff was administered to ten patients with mild- to moderately-active UC (Mitsuyama et al. Reference Mitsuyama, Saiki, Kanauchi, Iwanaga, Tomiyasu and Nishiyama1998) a significant decrease was found in mean clinical activity score after 4 weeks of treatment (P<0·05) associated with increases in faecal SCFA concentration.

A pilot study of eighteen patients in a randomised double-blind double-dummy study over 1 month (Furrie et al. Reference Furrie, Macfarlane, Kennedy, Cummings, Walsh, O'Neil and Macfarlane2005) has suggested that synbiotics may have a role in reducing inflammation in UC. A non-significant trend in favour of the combination of oligofructose and inulin and a probiotic (Bifidobacterium longum) was observed on histological and sigmoidoscopic assessment but clinical measures were not found to be different. However, mucosal-associated bifidobacteria concentrations were found to increase in association with a decrease in both pro-inflammatory cytokines such as TNFα and IL-1α and antimicrobial human β-defensin peptides. It is disappointing that the observed alterations in mucosal cytokine balance did not translate into more clinically-important changes in symptoms or endoscopic appearance. It may be that this rapid change in cytokine expression would give rise to a more clinically-relevant change if treatment were sustained for a longer period.

Probiotics and prebiotics to maintain remission in Crohn's disease

Evidence for the use of probiotics as maintenance therapy in CD is not persuasive, with only two studies reporting positive results (Table 4). A study comparing S. boulardii+the antibiotic mesalazine with mesalazine alone (Guslandi et al. Reference Guslandi, Mezzi, Sorghi and Testoni2000) has shown fewer relapses in the former group in patients with medically-induced remission of CD. Campieri et al. (Reference Campieri, Rizzello, Venturi, Gilberto, Ugolini, Helwig, Amadini, Romboli and Gionchetti2000) have shown benefit of VSL#3 in preventing post-operative recurrence in forty patients randomised to 3 months of rifaximin followed by 9 months of VSL#3 or to 12 months of mesalazine.

Table 4.

Clinical outcomes of published studies investigating the efficacy of probiotics and prebiotics in maintenance of remission in patients with Crohn's disease

R, randomised; OL, open label; DB, double blind; S. boulardii, Saccharomyces boulardii; VSL#3, a mixture of four species of lactobacilli, three species of bifidobacteria and Streptococcus thermophilus; LGG, Lactobacillus rhamnosus GG; CFU, colony-forming units; L. johnsonii, Lactobacillus johnsonii; Pl, placebo; NR, not reported.

* Pediacoccus pentoseceus, Lactobacillus raffinolactis, Lactobacillus paracasei susp paracasei 19, Lactobacillus plantarum, 2·5 g β-glucans, 2·5 g inulin, 2·5 g pectin, 2·5 g resistant starch.

Conversely, three studies using LGG (Prantera et al. Reference Prantera, Scribano, Falasco, Andreoli and Luzi2002; Schultz et al. Reference Schultz, Timmer, Herfarth, Sartor, Vanderhoof and Rath2004; Bousvaros et al. Reference Bousvaros, Guandalini, Baldassano, Botelho, Evans and Ferry2005) have not confirmed the effectiveness of this probiotic as a maintenance strategy after surgically- or medically-induced remission. No significant differences in clinical or endoscopic recurrence rates were found between patients receiving LGG or placebo to maintain surgical remission (Prantera et al. Reference Prantera, Scribano, Falasco, Andreoli and Luzi2002). Similarly, Schultz et al. (Reference Schultz, Timmer, Herfarth, Sartor, Vanderhoof and Rath2004) have shown no significant effect of LGG in maintaining remission induced with antibiotics and steroids. Finally, no significant difference in time to relapse was found between probiotics or placebo (inulin) in a paediatric population with medically-induced remission (Bousvaros et al. Reference Bousvaros, Guandalini, Baldassano, Botelho, Evans and Ferry2005). In this study few patients were found to have high concentrations of faecal lactobacilli (7% and 20% in the intervention and placebo groups respectively). The study was terminated early because of lack of effect and poor recruitment. It is important to note that the placebo (inulin) is a prebiotic that could therefore mask any difference in efficacy between the groups. It could even be argued that the non-significant trend towards fewer relapses in the placebo group could relate to the higher dose of inulin administered, although this explanation is unlikely at the low doses employed.

Finally, two randomised double-blind placebo-controlled studies (Marteau et al. Reference Marteau, Lemann, Seksik, Laharie, Colombel and Bouhnik2006; Van Gossum et al. Reference Van Gossum, Dewit, Louis, de Hertogh, Baert, Fontaine, DeVos, Enslen, Paintin and Franchimont2007) have reported no effect of Lactobacillus johnsonii LA1 in preventing recurrence in patients with CD in surgically-induced remission. Marteau et al. (Reference Marteau, Lemann, Seksik, Laharie, Colombel and Bouhnik2006) have found no significant difference in endoscopic recurrence in comparison with placebo 6 months after surgery. Likewise, Van Gossum et al. (Reference Van Gossum, Dewit, Louis, de Hertogh, Baert, Fontaine, DeVos, Enslen, Paintin and Franchimont2007) have reported no significant difference from placebo in endoscopy scores or clinical relapse rates after 3 months of treatment.

The use of prebiotic therapies in the maintenance of remission in CD has not been studied extensively. In patients with CD who have undergone ileal resection treatment with Synbiotic 2000 (Pediacoccus pentoseceus, Lactobacillus raffinolactis, Lactobacillus paracasei susp paracasei 19, Lactobacillus plantarum, 2·5 g β-glucans, 2·5 g inulin, 2·5 g pectin, 2·5 g resistant starch) was reported to have no impact on maintaining remission when compared with placebo (Chermesh et al. Reference Chermesh, Tamir, Reshef, Chowers, Suissa, Katz, Gelber, Halpern, Bengmark and Eliakim2007). It is notable that the investigators had intended to enrol twice as many patients, but were unable to do so because of reluctance of patients and physicians to risk exposure to the placebo arm.

Probiotics and prebiotics to treat active Crohn's disease

Two studies have evaluated probiotics in patients with active CD (Malchow, Reference Malchow1997; Gupta et al. Reference Gupta, Andrew, Kirschner and Guandalini2000), but neither study has demonstrated convincing efficacy, in part be-cause of concomitant drug prescription and small numbers of patients (Table 5). Malchow (Reference Malchow1997) has compared E. coli Nissle 1917 with placebo in patients with active CD who also received the steroid prednisolone, 60 mg daily for 11 weeks, to induce remission and then continued with probiotics or placebo during a 12-month maintenance phase. The proportion of patients achieving remission and time to remission was found to be similar in the two groups, suggesting that the remission was mostly attributable to steroid use. In the maintenance phase there was a non-significant trend towards decreased relapse rates in the E. coli group v. the placebo group. Gupta et al. (Reference Gupta, Andrew, Kirschner and Guandalini2000) have examined the effects of LGG in four children with steroid refractory CD, two of whom were taking concomitant metronidazole. Colonisation with LGG and a sustained reduction in disease activity associated with improved intestinal permeability was demonstrated in all patients, suggesting a mechanism for the therapeutic effect.

Table 5.

Clinical outcomes of published studies investigating the efficacy of probiotics and prebiotics in induction of remission in patients with active Crohn's disease

R, randomised; DB, double blind; OL, open label; E. coli N, Escherichia coli Nissle 1917; LGG, Lactobacillus rhamnosus GG; CFU, colony-forming units; FOS, fructo-oligosaccharide; Pl, placebo.

Prebiotics have also been investigated in the treatment of CD. Lindsay et al. (Reference Lindsay, Whelan, Stagg, Gobin, Al-Hassi, Rayment, Kamm, Knight and Forbes2006) have found that fructo-oligosaccharide reduces disease activity in active CD. As in healthy controls (Langlands et al. Reference Langlands, Hopkins, Coleman and Cummings2004), fructo-oligosaccharide was shown to increase faecal bifidobacteria concentrations, although no corresponding change in mucosal-associated bifidobacteria was found. However, a significant increase was found in mucosal bifidobacteria in those patients entering remission. Evidence was found of an immunoregulatory shift in dendritic cell activity, with an increase in the proportion of IL-10-positive dendritic cells (P=0·06), and enhanced toll-like receptor 2 and toll-like receptor 4 expression (P=0·086 and P<0·001 respectively). A placebo-controlled trial to confirm the benefit of fructo-oligosaccharide in active CD is underway.

Limitations to research on probiotics and probiotics in inflammatory bowel disease

Interpretation of the current literature on probiotics and prebiotics in IBD is undermined by marked heterogeneity between trials frequently precluding meta-analysis. Variables include the choice of probiotic or prebiotic studied and their dose, the trial design and the outcome measures reported. The populations studied differ, with some trials enrolling patients with active disease and others studying maintenance of remission induced by antibiotics, conventional drug treatment or surgery. Few studies stratify for disease distribution, and none have reported the prevalence of genotypes that may impact on the response to therapies manipulating the microbiota. Most studies have enrolled small numbers of patients, which limits statistical power and is of particular relevance given the high placebo response rates seen in clinical trials of IBD (Su et al. Reference Su, Lichtenstein, Krok, Brensinger and Lewis2004, Reference Su, Lewis, Goldberg, Brensinger and Lichtenstein2007). Finally, no study has included details of the patients' diet, which could exert a marked influence on the efficacy of microbial therapies.

The precise mechanisms of action of probiotics and prebiotics remain to be elucidated. For example, it is not clear whether intestinal colonisation with probiotic species is necessary, as some studies demonstrating clinical efficacy have not demonstrated such colonisation (Bibiloni et al. Reference Bibiloni, Fedorak, Tannock, Madsen, Gionchetti and Campieri2005). Further research is required to determine whether prebiotics exert their effect via increased numbers of bacteria that are recognised by intestinal immune cells or via effects of SCFA on the mucosa. In vitro studies examining the effect of prebiotics in enterocyte models may reveal direct effects on the mucosa. Investigation of changes in immune cell phenotype on exposure to prebiotics and/or bacteria may elucidate such relationships.