Review article: probiotics in gastrointestinal and liver diseases


  • Conflicts of interest:
    The authors have declared no conflicts of interest.

  • This article appeared as part of a supplement sponsored by Nycomed bv.

Dr D. Jonkers, Div. Gastroenterology-Hepatology, University Hospital Maastricht, PO Box 5800, 6202 AZ Maastricht, The Netherlands.


Background  Probiotics, defined as live micro-organisms with beneficial effects for the host, are widely applied in gastrointestinal and liver diseases.

Aim and method  To review the available evidence of clinical trials on probiotics in gastrointestinal and liver diseases, with a major focus on irritable bowel syndrome, inflammatory bowel disease, pancreatitis and chronic liver diseases.

Results  Evidence for the therapeutic or preventive application of particular probiotic strains is available for antibiotic-associated diarrhoea, rota-virus-associated diarrhoea and pouchitis. Results are encouraging for irritable bowel syndrome, ulcerative colitis and for reducing side effects by Helicobacter pylori eradication therapies, but are less clear for Crohn’s disease, lactose intolerance and constipation. In general, for most of these patient groups, more placebo-controlled methodologically well-designed studies that pay attention to both clinical outcome and mechanistic aspects are required. The application in liver disease and pancreatitis is promising, but more human trials have to be awaited. Possible mechanisms of probiotics include modulation of the intestinal microbiota and the immune system, but different bacterial may have different effects.

Conclusion  Further insight into disease entities and the functioning of probiotic strains is required to be able to select disease-specific strains, which have to be tested in well-designed placebo-controlled studies.


The possible role of the intestinal microbiota and the therapeutic or preventive application of probiotics in several disorders receives increasing attention.

The intestinal microbiota is a complex ecological system consisting of at least 500 different bacterial species with anatomical variations in composition and numbers:1 in the stomach, small numbers (0–103 colony forming units per mL) are found, predominantly consisting of lactobacilli, streptococci, staphylococci, enterobacteriaceae and yeasts. These small numbers are mainly because of the low intragastric pH. Subsequently, there is a quantitative and qualitative increase from 0–105 cfu per gram in the duodenum, to 108 cfu per gram in the ileum and 1010–1012 cfu per gram in the colon because of the neutral intestinal pH, a slower transit time and the availability of nutrients. In the colon more than 99% of the micro-organisms are strictly anaerobic, such as bifidobacteria, Bacteroides spp., Clostridium spp., Eubacterium spp., Fusobacterium spp. and peptostreptococci.1, 2

The indigenous intestinal microbiota develops early in life, starting at birth, and is influenced by mode of delivery, initial diet and early contact with the environment.2 After weaning, a unique adult flora develops which resembles the adult flora by the second year of life.3 Both Zoetendal et al.4 and Eckburg et al.5 have demonstrated that the individual faecal microbiota is relatively stable over time but differs between individuals and can be considered as a bacterial fingerprint.

Apart from studying individual bacterial groups, the microbiota can be regarded as a highly metabolic entity. It has several important functions for the host including the conversion of (pro)carcinogenic substances, the production of vitamins, degradation of bile acids and digestion of nutrients. Especially saccharolytic fermentation of unabsorbed and indigestible carbohydrates by intestinal bacteria, occurring mostly in the proximal colon, are important because of the production of short chain fatty acids (i.e. acetate, propionate and butyrate).6 Butyrate, a major energy source for intestinal epithelial cells, affects cell proliferation and differentiation, increases mucus secretion and decreases inflammation.7 Proteolytic bacterial fermentation usually takes place in the more distal colon, where carbohydrates are no longer available, and results in the production of toxic compounds such as ammonia, phenols, cresols and paracresols.8

The colonization of the intestine by commensal bacteria is also important for the development and functioning of the immune system. For example, in germ-free raised animals, fewer lymphocytes, plasma cells and mononuclear cells as well as a reduced development of lymphoid structures are found. In addition, lower IgA and mucin production has been observed.1, 9 Finally, together with the intestinal mucosa, the endogenous intestinal flora forms an important barrier against pathogens by the mechanism of colonization resistance.

Several of the functions of the intestinal microbiota may be beneficially influenced by probiotics. Probiotics are defined as ‘mono- or mixed cultures of live micro-organisms which, when applied to animal or man, beneficially affect the host by improving the properties of the indigenous flora’.10 They are mostly given as fermented milk or in a freeze-dried form and often include Bifidobacterium spp. or Lactobacillus spp.; Saccharomyces boulardii and Escherichia coli Nissle 1917 have also frequently been applied.

Lactic acid bacteria commonly occur in fermented foods such as green olives, sausage and sauerkraut. In a report from the European Workshop on the safety aspects of lactic acid bacteria, Adams11 concluded that the long history of consumption, the available epidemiological evidence, clinical trials and toxicity studies do not indicate a health risk for ingestion of lactic acid bacteria. In a large Finnish study,12 only 8 of 3317 sepsis cases were caused by lactic acid bacteria, but this never was a dairy or ingested strain. However, limited safety data are available for severely ill patients, and caution is especially required using yeast strains in these patients. In addition, attention should be paid to the absence of transferable antibiotic resistance in probiotic bacterial strains.

To exert a possible effect, probiotic bacteria have to survive passage through the gastrointestinal (GI) tract containing gastric juice, bile and pancreatic juice. Furthermore, adhesion to the intestinal mucosa is considered to be a prerequisite for the interaction with the immune system and this is more likely to occur with host-specific strains. Finally, the probiotic products have to have a good taste, smell and an acceptable shelf-life.

A study in healthy volunteers with Lactobacillus plantarum 299v found a significant increase of faecal numbers of this bacterium already 1 week after daily ingestion, remaining on a constant level during prolonged (i.e. 4 weeks) consumption, but decreasing significantly towards pretreatment levels within 1 week after cessation of intake.13 These results show that probiotic bacteria are not able to colonize the intestine permanently.

In line with the probiotic concept, prebiotics are often applied. Prebiotics are defined as ‘non-digestible food-ingredients that beneficially affect the host by selectively stimulating the growth and/or activity of one or a limited number of bacteria in the colon and thus improve host health’.14 Synbiotics consist of a combination of prebiotics and probiotics.

The mechanisms of action of probiotic bacteria are thought to result from modification of the composition of the indigenous intestinal microbiota and its metabolic activity, prevention of overgrowth and colonization of pathogens, and stimulation of the immune system.15 Probiotics have been studied in several disease entities, especially when intestinal bacteria are thought to be involved in their pathogenesis. In this review, the available evidence of clinical trials on probiotics in GI and liver diseases will be summarized with a major focus on irritable bowel syndrome (IBS), inflammatory bowel disease, pancreatitis and chronic liver diseases.

Irritable Bowel Syndrome

Irritable bowel syndrome is a functional GI disorder, characterized by symptoms such as abdominal pain, bloating, diarrhoea or constipation. The patient group is rather heterogeneous and a possible role of the intestinal microbiota is suggested in at least the subgroup of postinfectious diarrhoea in which the development of IBS is associated with prior enteric infections and/or antibiotic use.16–18 A few studies have examined the faecal flora of IBS patients and found a decrease in E. coli, lactobacilli and bifidobacteria and an increase in other aerobes in comparison with healthy volunteers.19, 20 Furthermore, a mal-fermentation, associated with an overproduction of hydrogen, has been observed in IBS patients17, 21 and some studies report a low-grade mucosal inflammation in unselected and postinfectious diarrhoea-predominant IBS patients.22, 23

Probiotics are applied in IBS as they may change the composition and metabolic activity of the intestinal microbiota such as the production of short chain fatty acids and gas. Furthermore, they may affect gut motility and exert an anti-inflammatory effect.

An overview of studies performed with probiotics in IBS is given in Table 1. Three placebo-controlled studies have been performed with L. plantarum 299v given for 4 weeks to mixed patient populations with IBS. Nobaek et al24 included 52 patients defined by the Rome criteria and found a significant reduction in abdominal pain and flatulence but no difference in bloating. No concurrent medication was taken and after 12 months’ follow-up, the overall GI function was significantly better in the probiotic compared with the placebo group. A significant decrease in abdominal pain and better overall GI function was also found by Niedzielin et al.25 in 40 patients. A small study with 12 untreated IBS patients could not find a significant effect.26 Also no effect could be demonstrated by Lactobacillus GG vs. placebo in bloating-positive IBS patients.27

Table 1.   Studies with probiotic bacteria in patients with irritable bowel syndrome
ReferenceNumber (IBS criteria)TreatmentDurationEffect
Open studies
 Andrews and Borody1603318 bacterial strainsOnce25 improved
 Hunter et al.16128E. faecium PR8812 weeks19 improved
Controlled studies
 Gade and Thorn16254S. faecium vs. placebo4 weeksSign. improvement in abdominal symptoms
 Halpern et al.3118L. acidophilus vs. placebo, crossover6 weeksSign. clinical improvement
 O’Sullivan and O’Morain2719 (Rome)L. GG vs. placebo, crossover8 weeksNo effect
 Nobaek et al.2460 (Rome)L. plantarum vs. placebo4 weeksDecrease in pain and flatulence, no difference in bloating
 Niedzielin et al.2540 (Manning)L. plantarum vs. placebo4 weeksDecrease pain and overall GI function
 Sen et al.2612 (Rome)L. plantarum vs. placebo, crossover4 weeksNo effect
 Kim et al.2825 (Rome II)VSL#3 vs. placebo8 weeksSign. reduction in abdominal bloating
 Saggioro3070 (Rome II)L. plantarum B. breve vs. L. plantarum L.acidophilus vs. placebo4 weeksSign. decrease in pain and IBD-symptom score
 O’Mahony et al.3277 (Rome II)L. salivarius vs. B. infantis vs. placebo8 weeksB. infantis: sign. improvement symptoms and IL-10/IL-12 ratio
 Niv et al.16339 (Rome II)L. reuteri vs. placebo6 monthsNo effect
 Kim et al.2948 (Rome II)VSL#3 vs. placebo4–8 weeksDecrease flatulence and transit time
 Bittner et al.16425 (Rome II)multispecies synbiotic vs. placebo2 weeksSign. reduction in scores for general ill feelings, nausea, indigestion and flatulence

Using the multispecies product VSL#3 vs. placebo in 25 diarrhoea-predominant patients, no difference in intestinal transit time or bowel function but a significant reduction in abdominal bloating was observed.28 In a second study,29 the same author included 48 bloating-predominant IBS patients and found a significant decrease in flatulence and transit time but no difference in abdominal pain and bloating. A rather large study has been performed by Saggioro,30 who treated 70 patients with L. plantarum plus Bifidobacterium breve (n = 24), L. plantarum plus Lactobacillus acidophilus (n = 26) or placebo (n = 20) for 4 weeks and found significant decreases in pain and the IBD symptom score in both probiotic-treated groups. A significant clinical improvement was also found in 18 patients after treatment with L. acidophilus vs. placebo, studied with a crossover design.31

In 2005, the 8 weeks’ consumption of Lactobacillus salivarius was compared with Bifidobacterium infantis and placebo in 77 IBS patients defined by the Rome II criteria.32 Although L. salivarius resulted in a significantly better bowel movement measured by the Bristol stool scale in comparison with placebo, the B. infantis group had significant better scores for abdominal pain, bloating, bowel movement and overall GI function compared with both the L. salivarius and the placebo group. In addition, this was the first study in IBS patients, paying attention to clinical outcome as well as immune-modulating effects of probiotic bacteria. The patients had a significantly lower IL-10 (anti-inflammatory cytokine)/IL-12 (pro-inflammatory cytokine) ratio compared with that in healthy controls; this ratio improved significantly by B. infantis but not by L. salivarius.

The majority of IBS studies were placebo-controlled and nine of 12 controlled studies showed beneficial effects. However, the number of dropouts in these studies was high (20–30%), and the role of concurrent medication and diet were not always taken into account. In addition, mostly subjective outcome parameters were used and in several studies only some of the clinical scores measured improved significantly. Apart from these methodological limitations, the inclusion of heterogeneous IBS sub-populations makes it difficult to draw definite conclusions from these studies.

The findings by O’Mahony et al.32 support previous observations that different probiotic bacteria can have different effects. As subgroups of IBS patients (diarrhoea-, constipation- or bloating-predominant) will benefit from different mechanistic actions, further studies should focus on interventions in well-defined subgroups of IBS patients, taking into account possible confounding factors.

Inflammatory Bowel Disease

Crohn’s disease (CD) and ulcerative colitis (UC) are chronic inflammatory bowel diseases (IBD) characterized by unpredictable periods of active and inactive disease. Evidence suggests that (luminal) bacteria, loss of oral tolerance and an impairment of the mucosal barrier function play a role in the pathogenesis of IBD.33

Oral tolerance is defined as the induction of a state of systemic immune non-responsiveness to orally administered antigens such as intestinal bacteria, after an initial antigen challenge.34 An exaggerated T-cell response causing mucosal hyper-responsiveness is thought to be an important factor driving inflammation in IBD.33, 34 T-cells from the lamina propria of CD patients respond with a Th1 polarized response;34 in UC patients, a preferential expression of Th2 cytokines IL-4 and IL-5 has been described.35 In addition, reduced numbers of regulating T-cells which produce IL-10 and/or transforming growth factor-β are found in both UC and CD.35 Finally, mutations in the NOD2/CARD15 gene (involved in the intracellular recognition of bacterial components) are associated with CD,34 and increased expression of toll-like receptors (trans-membrane receptors that recognize conserved structures of micro-organisms) is found in both UC and CD.36, 37

A role of intestinal bacteria in the development of inflammation is supported by several animal studies. Models for colitis such as IL-10 knock-out mice, HLA B27 transgenic mice and chemically treated rats do not develop inflammation when raised under germ-free conditions, but introduction of a ‘normal intestinal microbiota’ results in the development of colitis.38–40 A possible role of intestinal bacteria in inflammation is further supported by findings that intestinal inflammation is often present in anatomical areas with high bacterial numbers.41 Furthermore, diversion of the faecal stream42–44 and washing out the gut lumen45 resulted in a decreased disease activity in CD patients and some studies show a beneficial effect of antibiotics in CD and UC patients.46

Apart from controversial results on the aetiological role of Mycobacterium paratuberculosis, Listeria monocytogenes and paramyxoviruses in CD,47 no specific micro-organism has yet been described as a possible causal factor in IBD. However, a change in the bacterial composition of both the faecal and mucosal microbiota has been found. Although the results are not always consistent, more aerobes (especially E. coli) and anaerobes (especially Bacteroides spp.) and lower bifidobacteria and lactobacilli counts have been found in faecal samples of IBD patients in comparison with healthy volunteers.48–52 In addition, using molecular methods, several groups have found increased numbers of bacteria in both inflamed and non-inflamed tissue samples of UC and CD patients in comparison with non-IBD controls.53–56 Especially, an increase in E. coli, clostridia and Bacteroides spp. with a decrease in lactic acid bacteria has been reported.54, 56 Finally, sulphate-reducing bacteria are more often found in biopsy samples of UC patients compared with healthy volunteers.57

Whether the changes in the intestinal bacterial composition of IBD patients are a cause or a consequence of the inflammation is not yet elucidated, but possible manipulation of the indigenous intestinal microbiota as well as the immune system by probiotics is the rationale for studying the effect of probiotics in IBD.

The first evidence for the use of probiotics in IBD comes from animal models of colitis. For example, Lactobacillus reuteri and L. plantarum were found to reduce the severity of methotrexate-induced and acetic acid-induced colitis.58, 59 A reduction in the development of colitis was demonstrated in IL-10 knock-out mice treated with Lactobacillus spp.60–62

In humans, most convincing evidence is available for pouchitis patients. Pouchitis is a complication after ileal-anal pouch anastomosis in UC patients. Pouchitis is characterized by an increase in aerobes, an increase or decrease in anaerobes, and a decrease in short chain fatty acid concentrations in faecal samples.63 Small but well-designed studies have been performed with the multispecies probiotic VSL#3 consisting of four lactobacilli (L. casei, L. plantarum, L. acidophilus, L. delbrueckii spp. bulgaricus), three bifidobacteria (B. longum, B. breve, B. infantis) and Streptococcus salivarius spp. thermophilus. In two double-blind placebo-controlled studies, 40 patients64 and 36 patients65 with chronic pouchitis were randomized to receive either VSL#3 or placebo. Relapse rates were significantly lower, being 15% vs. 100% after 9 months’ probiotic vs. placebo treatment64 and 15% vs. 94% after 12 months’ probiotic vs. placebo treatment.65 In the latter study, quality of life had also improved significantly. The development of pouchitis in 40 patients with a newly formed ileo-anal pouch was also found to be prevented by VSL#3 (10%) in comparison with placebo (40%) during an observation period of 1 year.66

A mixture of L. acidophilus and B. lactis has been tested by the group of Laake et al. in three uncontrolled trials.67–69 In their first study, they found a significant reduction in endoscopic but not in histological scores after 4 weeks’ treatment of 10 patients operated with an ileal-anal pouch anastomosis.68 Subsequently, they studied 41 pouch-patients for 4 weeks post-surgery and found a significant reduction in the pouch disease activity index,69 a finding that was once more confirmed in 51 operated pouch patients.67

Thirty-nine patients in whom an ileal pouch was created, started with daily intake of L. GG or placebo for 3 years postoperatively:70 the first episode of pouchitis was observed less frequently in the L.GG group (7%) compared with the placebo group (29%).

The effect of L. GG vs. placebo was also studied in 20 patients with a previous history of pouchitis and endoscopic inflammation. Although the faecal ratio of lactobacilli to total anaerobes increased significantly, no clinical or endoscopic response was found after 3 months’ probiotic consumption.71

In 1989, Bennet72 was the first to describe the effect of intestinal bacteria on disease outcome in UC. He successfully induced remission in an active UC patient by giving an antibiotic and a faecal enema with a ‘normal colon flora’.

Subsequent trials in UC often focused on maintenance treatment and are mostly uncontrolled (Table 2). Three trials have been performed comparing E. coli Nissle 1917 vs. mesalazine and reported similar relapse rates73–75 studying relatively high numbers of patients. However, mesalazine doses in two of these studies are generally considered to be rather low (500 mg thrice daily). A placebo-controlled trial has been performed with bifido-fermented milk in 21 inactive UC patients.76 After a 12-month treatment period, a reduction in exacerbations was found after bifido-fermented milk (27%) vs. placebo (90%).

Table 2.   Studies with probiotic bacteria in patients with ulcerative colitis
ReferenceNumber + disease activityTreatmentDurationEffect
Active disease
 Bennet and Brinkman 721 activeAntibiotic + faecal enema of colon floraOnceInduced remission
 Kordecki779 activeL. plantarum?6 of 9 (%) in remission
 Guslandi et al.7825 mod. activeS. boulardii + mesalazine4 weeks71% in remission
 Borody et al.1656 activeEnema with faecal flora5 daysComplete reversal of symptoms
 Kato et al.8020 mod. activeBifido-fermented milk or placebo12 weeksDecrease endoscopic and histol. score
 Tursi et al.8190 mod. activeVSL#3 + balsalazide vs. 2 conventional ther.8 weeksFaster remission (4 vs. 7.5 and 13 days)
 Bibiloni et al.7934 mod. activeVSL#36 weeks77% sign. decrease in disease activity index
 Furrie et al.8218 activeB. longum + FOS/inuline vs. placebo4 weeksSign. lower endosc. score and decrease inflammatory mediators
Inactive disease
 Kruis et al.73120 inactiveE. coli vs. mesalazine12 weeksSimilar relapse rate (16% vs. 11%)
 Rembacken et al.74116 inactiveE. coli vs. mesalazine12 monthsSimilar relapse rate (67% vs. 73%)
 Venturi et al.16620 inactiveVSL#312 months75% remission
 Ishikawa et al.7621 inactiveBifido-fermented milk vs. placebo12 months27% vs. 90% exacerbation
 Kruis et al.75327 inactiveE. coli vs. mesalazine12 monthsSimilar relapse rate (36% vs. 34%)

Uncontrolled studies in active UC patients have shown 67%, 71% and 77% clinical response or induction of remission with L. plantarum (n = 9),77S. boulardii plus mesalazine (n = 25)78 and VSL#3 (n = 34)79 respectively. Bifidobacteria-fermented milk vs. placebo for 12 weeks resulted in a significantly lower endoscopic and histological score in active UC patients,80 and VSL#3 resulted in a faster remission (4.5 days) in comparison with a medium dose of balsalazide (7.5 days) or mesalazine (13 days).81 Furrie et al82 were the first to study the effect of a synbiotic in UC on both clinical outcome and inflammatory mediators. In this small placebo-controlled pilot study, 18 active UC patients were randomized to receive a combination of B. longum with a fructo-oligosaccharide/inuline mix or placebo. Although only eight patients in the probiotic group and six in the placebo group completed the study, they found a significant improvement in the endoscopic score, as well as in human β defensin 2, 3 and 4, tumour necrosis factor-α (TNF-α) and IL-1α levels.

The number of studies published with CD patients is limited and results are conflicting (Table 3). Malchow83 studied 28 active CD patients and found a decreased relapse rate with E. coli Nissle1917 (33%) vs. placebo (64%) when combined with prednisolon. Matilla-Sandholm et al.84 reported the results of an open study with 20 active patients treated with L. salivarius UCC35624 for 10 days and found a significant improvement in quality of life but could not induce remission. Whether 10 days is long enough to induce remission is debatable. Guslandi et al.85 studied 32 inactive CD patients and found a decreased relapse rate (33% vs. 64%) using S. boulardii plus mesalazine (2 g/day) vs. mesalazine alone (3 g/day). Finally, Prantera et al.86 did not find a beneficial effect of L. GG compared with placebo and reported even more severe endoscopic recurrence in the probiotic group.

Table 3.   Studies with probiotic bacteria in patients with Crohn’s disease
ReferenceNumber +disease activityTreatmentDurationEffect
Active disease
 Malchow8328 activePrednisolon + E.coli vs. placebo12 monthsDecrease relapse rate (33% vs.64%)
 Matilla-Sandholm et al.8420 activeL.  salivarius UCC3562410 daysNo remission, increase quality of life
Inactive disease
 Guslandi16710 inactiveS. boulardii +  mesalazine6 monthsOnly 1 relapse
 Guslandi et al.8532 inactiveS. boulardii + mesal. vs. mesalazine6 monthsDecrease relapse rate (6% vs. 38%)
 Prantera et al.8632 inactive (operated)L. GG vs. placebo12 monthsSimilar relapse (7% vs. 11%) but more end. recur. by L.GG (60% vs. 35%)

Evidence from well-designed studies is available for the efficacy of VSL#3 to maintain remission in pouch patients. The evidence for efficacy in UC and CD is still not clear. Although, results in UC patients are encouraging, only a limited number of studies are available for CD and in both disease entities well-designed placebo-controlled studies are often lacking. Confounders such as dietary habits and use of stable or concurrent medication are often not taken into account and apart from the studies with E. coli Nissle in UC patients, generally small numbers of patients are studied. However, the circumstantial evidence warrants further well-designed studies with sufficient patient numbers where attention should be paid to clinical outcome as well as to mechanistic aspects.

Apart from bacteria, the probiotic concept has recently also been applied to helminths. Intestinal helminths are known to induce a Th2 and T regulatory cells and to down-regulate the Th1 response.35 This knowledge in combination with the epidemiological observation that the prevalence of IBD is inversely related to the prevalence of helminthic parasites resulted in the application of Trichuris suis ova for the treatment of IBD. A pilot study treated four patients with active CD and three patients with active UC with 2500 T. suis ova once every 2 weeks for a period of 12 weeks. The application was found to be safe and to induce remission in three of four CD patients and all UC patients.87 Subsequently, an open label study with T. suis ova induced remission in 21 of 29 active CD patients.88 In a placebo-controlled study with 54 active UC patients, clinical improvement of the disease activity was found in 43% of the ova-treated and 17% of the placebo-treated subjects.89


Many studies have been performed on the effect of probiotics in antibiotic-associated diarrhoea (AAD) and acute diarrhoea. AAD occurs in 4–30% of antibiotic-treated patients and is caused by a disturbance of the intestinal microbiota, direct tissue damage or modulation of the immune system.90–92Clostridium difficile is found in 10–25% of AAD patients and can cause pseudo-membranous colitis.90, 91

Two meta-analyses by D’Souza et al.93 and Cremonini et al.94 have shown a significant benefit for probiotics in the prevention of AAD with an OR of 0.37 and RR of 0.40 respectively. Furthermore, Hawrelak et al.95 published a systematic review on the effect of L. GG for the prevention of AAD and concluded that four of six placebo-controlled trials showed a significant reduction in the risk of AAD. Szajewska et al.96 performed a meta-analysis on five controlled trials with S. boulardii and found a significant RR of 0.43 for the risk of AAD.

In a recently published large meta-analysis for the prevention of AAD and Clostridium difficile-associated diarrhoea, 25 and six randomized controlled trials were included.97 Probiotics significantly reduced the risk of both AAD (RR 0.43) and C. difficile-associated diarrhoea (RR 0.59). The meta-analyses were then repeated stratifying for the probiotic strains used: S. boulardii, L. GG and probiotic mixtures significantly reduced the development of AAD and only S. boulardii was effective for C. difficile-associated diarrhoea.97

Most studies in acute diarrhoea have focused on rota-virus-associated diarrhoea in children. In placebo-controlled studies, L. GG was found to decrease the duration of this diarrhoea with 1–2 days in combination with an increase in IgA and a decrease in rota-virus shedding.98–101 No such effect was found for L. acidophilus, L. rhamnosis, L. delbrueckii and S. thermophilus.102–104 Van Niel et al.105 conducted a meta-analysis on the effect of lactobacilli in acute infectious diarrhoea in children and found a significant reduction in the duration of diarrhoea with 0.7 days.

Results on traveller’s diarrhoea are inconsistent and vary with the probiotic strain and country of destination tested.106, 107

In conclusion, placebo-controlled trials have repeatedly shown a beneficial effect of probiotics in the prevention of AAD and for the treatment of acute diarrhoea in children but results depend on the probiotic strain used.

Helicobacter pylori Eradication, Constipation and Lactose Malabsorption

Probiotics have also been tested as eradication therapy for H. pylori. In vitro, specific lactobacilli and bifidobacteria have been found to exert bactericidal effects against H. pylori, probably by the production of bacteriocins or organic acids.108 Several studies, mostly not placebo-controlled, have examined the use of probiotics alone in H. pylori-positive patients and found no or only a small effect on eradication of H. pylori.109 However, small but significant effects have occasionally been observed on H. pylori density and inflammatory cell scores in gastric biopsies.110

A second group of studies have focused on the effect of probiotics on H. pylori eradication and side effects in patients treated with a standard H. pylori eradication regimen. All of them are placebo-controlled. Only a few studies have reported a significant higher eradication rate after probiotic vs. placebo treatment,111–113 but most other studies could not find a beneficial effect on H. pylori prevalence.114–120 However, in many of the studies a significant decrease in adverse effects, such as diarrhoea, nausea or unpleasant taste, have been observed.112, 115, 117–120 Future trials should address the type of patients and probiotic strain to be used.

As probiotics may affect GI transit time, a limited number of placebo-controlled studies have been performed to evaluate the effect of probiotics in patients with constipation. Beneficial effects were found on self-reported severity of constipation in 70 chronically constipated adults treated with L. casei shirota or placebo for 4 weeks.121 A small beneficial effect of a 24% increase in defecation frequency has been observed after L. rhamnosis intake in combination with Propionibacterium freudenreichii vs. placebo.122 No effect on bowel movements could be demonstrated in 48 children treated with lactulose plus L. GG vs. lactulose plus placebo for 12 weeks123 or in constipated elderly patients treated with L. GG or placebo.124

Yoghurt and probiotic bacteria are also often given to improve lactose digestion and to reduce symptoms of intolerance in lactose malabsorbers, as fermented milk products have been observed to be tolerated better than milk.15 Beneficial effects can be caused by microbial β-galactosidase, a delayed GI transit time, a reduced sensitivity to symptoms and positive effects on intestinal functions and the colonic microbiota.125 An improvement in parameters for lactose maldigestion has been found by some authors,126–128 but not by others.129–131 Recently Levri et al.132 reviewed controlled studies only and concluded that probiotic supplementation in general did not alleviate symptoms and signs of lactose intolerance but some evidence was found that specific strains might be effective.

In conclusion, studies with probiotics in H. pylori-positive patients demonstrate a significant reduction in side effects but do not have a major impact on eradication. Results of studies in patients with lactose malabsorption and constipation are inconclusive.

Chronic Liver Diseases

Patients with chronic liver diseases such as alcoholic hepatitis and cirrhosis are at risk of developing ascitis and even spontaneous bacterial peritonitis (SBP). SBP is associated with a 1-year recurrence rate of 40–70% and a high mortality of up to 30–80% after 1 year follow-up.133 The increased risk of infections is thought to be caused by intestinal bacterial overgrowth, increased translocation of bacteria or bacterial products and failures in immune defence. At least small intestinal bacterial overgrowth, a change in the composition of the gut microbiota, increased bacterial translocation and increased intestinal permeability have been demonstrated in cirrhotic patients.133 Although prophylaxis using norflaxacin is found to be (cost)effective in reducing the incidence of (recurrent) SBP, the incidence of resistant micro-organisms is increasing following this treatment.133 As non-antibiotic alternative, the effect of pro-and synbiotics has been studied in these patients.

Decreased bacterial translocation and a reduction in small intestinal bacterial overgrowth has been found using L. johnsonii134 but not by L. GG135 in rats with experimental cirrhosis. A human trial has been performed with a combination of four lactic acid bacteria and one type of fermentable fibres. This combination was found to reduce the viable counts of potentially pathogenic bacteria in the small intestine significantly.136 Rayes et al.137 compared selective bowel decontamination by antibiotics with L. plantarum 299 plus a fermentable fibre preparation in liver transplant recipients, showing a significant lower rate of postoperative bacterial infections (13% vs. 48%). This result was confirmed in a randomized, double-blind study where a mixture of lactic acid bacteria and fibres was compared with fibres only.138Postoperative bacterial infections were significantly reduced in the bacteria plus fibre group vs. the fibre group (3% vs. 48%).

In an animal model for non-alcoholic fatty liver disease (NAFLD), VSL#3 added to the diet vs. placebo was found to improve liver histology, to reduce hepatic total fatty acid content and to decrease ALT levels.139 This multispecies probiotic product was subsequently studied in 22 patients with NAFLD, 20 patients with alcoholic liver cirrhosis and 33 HCV-positive patients with chronic hepatitis.140 Results differed between the patient groups: malondialdehyde and 4-hydroxynonenal (markers of lipid peroxidation) improved significantly in the NAFLD and alcoholic liver cirrhosis patients while TNF-α, IL-6, and IL-10 improved significantly only in the last group. No effect was observed in the HCV-positive patients.


In line with the possible mechanisms for infectious complications in cirrhosis, bacterial translocation is also thought to be a major cause of infectious complications in the late phase of severe acute pancreatitis resulting in severe sepsis and multiple organ failure. This bacterial translocation can be caused by bacterial overgrowth and/or a disturbed intestinal motility, mucosal barrier or an overproduction of pro-inflammatory cytokines141, 142 and may be prevented by probiotics.

To date only one study has been published on the effect of probiotics on pancreatitis in humans. Using a combination of a L. plantarum 299 with a fibre preparation in patients with non-biliary pancreatitis, Olah et al.143 found significantly less infected pancreatic necroses with live lactobacilli (5%, n = 22) compared with heat-killed lactobacilli (30%, n = 23). However, this small study suffered from some methodological limitations such as lack of intention to treat analysis, fewer patients than required included and concurrent use of antibiotics in a subgroup of patients.

Other studies on the effect of probiotics in pancreatitis are based on animal models. Lactobacillus plantarum had a significantly reduced bacterial translocation when compared with placebo and with sham operation in rats with experimental acute pancreatitis induced by isolation and ligation of the biliopancreatic duct.144 In another animal model for acute necrotizing pancreatitis induced by 3% sodium taurocholate, the effect of S. boulardii alone and in combination with two different antibiotics was evaluated.145 The probiotic alone was found to decrease bacterial translocation and a significant lower histopathological score was found for the combination of S. boulardii in combination with ciprofloxacin in comparison with sham feeding, ciprofloxacin alone, meropenem alone or the combination of S. boulardii with meropenen. Marotta et al.146 found that pretreatment with a synbiotic, but not metronidazole, prevented endotoxin and bacterial translocation as well as liver damage in rats with acute pancreatitis and heavy alcohol consumption.

A multispecies probiotic product for acute pancreatitis has been designed based on stimulation of anti-inflammatory cytokine production, stimulation of GI motility and competitive inhibition of opportunistic infections. It has been tested in a rat model for severe acute pancreatitis. The probiotic was given 5 days before and 7 days after induction of pancreatitis and resulted in significantly less overgrowth with E. coli in the small intestine, significantly less bacterial translocation in the mesenteric lymph nodes, liver and spleen, and a lower mortality compared with placebo.147 Based on these results, a large randomized, double-blind placebo-controlled trial on the effect of this multispecies probiotic in pancreatitis is currently being performed by the Dutch Acute Pancreatitis Study group.141

Which Probiotic to Choose?

Possible mechanisms of action of probiotic bacteria are related to inhibition of potential pathogenic bacteria, modulation of the indigenous microbiota, improvement in the mucosal barrier function or stimulation of the immune system. With regard to pathogen exclusion, probiotic bacteria can produce antibacterial substances, block adhesion sites and be competitive for nutrients.34, 36 An enhancement of the barrier function is mostly studied in animal models or in vitro using cell lines. For example Madsen et al.,60 found an increase in epithelial barrier function after VSL#3 supplementation in IL-10-deficient mice. A decreased intestinal permeability has also been demonstrated for L. GG in rat pups.148Comparing two Lactobacilli spp. in rats with hemorrhagic shock, Luyer et al.149 found a reduction in plasma endotoxin levels and bacterial translocation after pretreatment with L. rhamnosus but not after pre-treatment with L. fermentum.

A major mechanistic effect of probiotic bacteria is thought to be their influence on the immune system. Not only enhanced phagocytic activity and NK cell activity but also stimulation of the adaptive immune response have been demonstrated for several probiotic strains.3 For example, Borruel et al.150 found ex vivo an induction of IL-10 and reduction in INF and TNF release by L. casei and L. bulgaricus in tissue samples of CD patients. Others have demonstrated increased local and plasma IgA levels after administration of L. GG to children with CD.151In vitro, different lactobacilli have been shown to exert different patterns of dendritic cell activation152 and different expression of proinflammatory and anti-inflammatory cytokines.153

Apart from the in vitro results, lessons from animal models36 and results from human trials32, 102 teach that not all probiotic strains have a similar effect. Therefore, in future, selection of strains for disease-specific indications is needed.154 Although selection criteria have been given by a working group of the WHO,155 including for example antimicrobial production, competition for nutrients, competitive exclusion of pathogen binding and modulation of the immune system, further mechanistic insight into disease entities as well as the functioning of the intestinal microbial ecological system and their interaction with the host is warranted.

Apart from particular bacterial strains, multispecies products are also increasingly applied. As different bacterial strains can have different effects, in a multispecies product these combined effects may act complementarily or even synergistically,156 but antagonistic effects cannot always be excluded.

Until now, most definitions of probiotics are based on ‘live micro-organisms’.10, 155, 157 However, evidence is growing that non-viable strains, secreted metabolites or even DNA moieties of probiotic bacteria can also exert a beneficial effect. Both DNA and secreted proteins of VSL#3 strains were found to inhibit nuclear factor κB in epithelial cells.158 Non-Non-viable irradiated bacteria or the purified DNA of VSL#3 suppressed experimental colitis, an effect which was found to be mediated by TLR9.159 In contrast, non-viable strains were not always found to exert beneficial effects as demonstrated by Olah et al.143 and Rayes et al.137 Probably, the mechanism of action of a particular strain in a specific host situation will determine whether a strain has to be viable to exert its effects.


The intestinal microbiota is a very complex ecosystem, which has several important functions for the host. Probiotic bacteria are increasingly used in several GI and liver diseases, especially when intestinal bacteria might be involved in their pathogenesis. Evidence for the therapeutic or preventive application of particular probiotic strains is available for AAD, rota-virus-associated diarrhoea and pouchitis. Results are encouraging for IBS, UC and for reducing side effects by H. pylori eradication therapies but are less clear for CD, lactose intolerance and constipation. In general, for most of these patient groups, more placebo-controlled methodologically well designed studies paying attention to both clinical outcome and mechanistic aspects are required. The application of probiotics in pancreatitis and chronic liver disease is promising but is mainly based on animal data and (more) human trials have to be awaited.

As different probiotic strains can have different effects, further insight into disease entities and into the functioning of probiotic strains is required to be able to select well-characterized strains with specific health benefits.