Meta-analysis: Lactobacillus rhamnosus GG for abdominal pain-related functional gastrointestinal disorders in childhood

Authors


  • As part of AP&T's peer-review process, a technical check of this meta-analysis was performed by Dr. P. Collins.

Dr H. Szajewska, Department of Paediatrics, The Medical University of Warsaw, Dzialdowska 1, Warsaw 01-184, Poland.
E-mail: hania@ipgate.pl

Summary

Background  A lack of reliable treatments for abdominal pain-related functional gastrointestinal disorders prompts interest in new therapies.

Aim  To evaluate systematically the effect of Lactobacillus rhamnosus GG (LGG) for treating abdominal pain-related functional gastrointestinal disorders in children.

Methods  MEDLINE, EMBASE, CINAHL, the Cochrane Library, trial registries and proceedings of major meetings were searched for randomised controlled trials (RCTs) evaluating LGG supplementation in children with abdominal pain-related functional gastrointestinal disorders based on the Rome II or Rome III criteria. Risk of bias was assessed for generation of the allocation sequence, allocation concealment, blinding and follow-up.

Results  Compared with placebo, LGG supplementation was associated with a significantly higher rate of treatment responders (defined as no pain or a decrease in pain intensity) in the overall population with abdominal pain-related functional gastrointestinal disorders (three RCTs, = 290; risk ratio, RR 1.31, 95% CI 1.08–1.59, number needed to treat, NNT 7, 95% CI 4–22) and in the irritable bowel syndrome (IBS) subgroup (three RCTs, = 167; RR 1.70, 95% CI 1.27–2.27, NNT 4, 95% CI 3–8). However, no difference was found in the rate of treatment responders between children with functional abdominal pain or functional dyspepsia who received placebo or LGG. The intensity of pain was significantly reduced in the overall study population and in the IBS subgroup. The frequency of pain was significantly reduced in the IBS subgroup only.

Conclusion  The use of Lactobacillus rhamnosus GG moderately increases treatment success in children with abdominal pain-related functional gastrointestinal disorders, particularly among children with IBS.

Introduction

According to the Rome III criteria,1, 2 abdominal pain-related functional gastrointestinal disorders (FGD) in children may be categorised as functional dyspepsia (FD), irritable bowel syndrome (IBS), abdominal migraine and functional abdominal pain (FAP). Because of their obscure pathophysiology, management of abdominal pain-related FGD remains difficult, prompting interest in new and safe treatment options.

Recently, probiotics, defined as ‘live microorganisms which when administered in adequate amounts confer a health benefit on the host’,3 have been proposed as treatment for FGD. While the exact mechanisms by which probiotics may exert their actions in patients with FGD are not fully understood, several mechanisms have been suggested. These include changes in colonic fermentation by displacement of gas-producing, bile salt-deconjugating bacteria strains, down-regulation of the local proinflammatory response, control of intestinal motor functions, and reducing visceral hypersensitivity by inducing the expression of u-opioid and cannabinoid receptors in human intestinal epithelial cells.4

Previously, four meta-analyses5–8 and one systematic review9 evaluated the effects of probiotics for the treatment of IBS, primarily in the adult population and reached slightly different conclusions. For adults, despite the equivocal results, the American College of Gastroenterology recently concluded that ‘in single organism studies, lactobacilli do not appear effective; bifidobacteria and certain combinations of probiotics demonstrate some efficacy.’10 Regarding the paediatric population, a Cochrane systematic review11 (search date: December 2006) concluded that there is no evidence that lactobacillus supplementation is effective in the management of children with recurrent abdominal pain. New studies12, 13 have been published since this meta-analysis, prompting interest in re-evaluating the role of probiotics in the management of abdominal pain-related FGD in children. With this systematic review, we aimed to update and synthesise the available randomised controlled clinical trial evidence of the likely effects of Lactobacillus rhamnosus GG (LGG) compared with placebo or no intervention in children affected by abdominal pain-related FGD. In contrast to previous meta-analyses, whether performed in adults or in children, we focused on a single probiotic microorganism. This is because the beneficial effects of probiotics seem to be strain specific, thus, pooling data on different strains may result in misleading conclusions. The choice of the probiotic LGG was determined by the fact that it is available and commonly used in many countries, and it has demonstrated efficacy in the treatment of some gastrointestinal conditions in children.14, 15

Methods

Criteria for considering studies for this review

All relevant randomised controlled trials (RCTs) that compared the effects of LGG supplementation with the effects of placebo or no supplementation for treating abdominal pain-related FGD in children were considered for this review.

Search methods for identification of studies

The Cochrane Central Register of Controlled Trials (CENTRAL, the Cochrane Library, Issues 4, 2010), MEDLINE (1966–2010) and EMBASE (1980–2010) were systematically searched up to December 2010. In addition, we searched two trial registries (ClinicalTrials.gov, http://www.clinicaltrials.gov, and EU Clinical Trials Register, http://www.clinicaltrialsregister.eu) and proceedings from major scientific gastrointestinal meetings such as ESPGHAN, NASPGHAN, UEGW and DDW published in the last 3 years. The reference lists of identified studies and key review articles, including previously published reviews, were also searched. In all cases, the searches were for studies that assessed the effects of probiotic supplementation on abdominal pain-related FGD in children up to 18 years of age. No language restrictions were imposed. The search strategy included the use of a validated filter for identifying RCTs, which was combined with a topic-specific strategy using the following PubMed MeSH terms: (‘probiotics’[MeSH Terms] OR ‘probiotics’[All Fields] OR ‘probiotic’[All Fields]) OR (‘lactobacillus’[MeSH Terms] OR ‘lactobacillus’[All Fields]) OR (LGG [All Fields]) OR (‘lactobacillus rhamnosus’[MeSH Terms] OR (‘lactobacillus rhamnosus’[All Fields]) AND (recurrent[All Fields] AND (‘abdominal pain’[MeSH Terms] OR (‘abdominal’[All Fields] AND ‘pain’[All Fields]) OR ‘abdominal pain’[All Fields])) OR (functional[All Fields] AND (‘abdominal pain’[MeSH Terms] OR (‘abdominal’[All Fields] AND ‘pain’[All Fields]) OR ‘abdominal pain’[All Fields])) OR (‘irritable bowel syndrome’[MeSH Terms] OR (‘irritable’[All Fields] AND ‘bowel’[All Fields] AND ‘syndrome’[All Fields]) OR ‘irritable bowel syndrome’[All Fields]) OR (functional[All Fields] AND (‘dyspepsia’[MeSH Terms] OR ‘dyspepsia’[All Fields]) OR ((‘abdomen’[MeSH Terms] OR ‘abdomen’[All Fields] OR ‘abdominal’[All Fields]) AND pain-related[All Fields] AND (‘disease’[MeSH Terms] OR ‘disease’[All Fields] OR ‘disorders’[All Fields])) OR ‘Rome criteria’[All Fields]) AND (‘child’[MeSH Terms] OR ‘child’[All Fields] OR ‘children’[All Fields]) OR (‘adolescent’[MeSH Terms] OR ‘adolescent’[All Fields]). Two of the reviewers (PD and AH) searched the databases independently and screened bibliographies of retrieved studies and recent review articles. The primary outcome measure was the rate of responders to the treatment (defined as no pain or a clinically meaningful decrease in pain intensity). The secondary outcome measures were the frequency and intensity/severity of abdominal pain reported by patients and adverse events.

Selection of studies

We excluded studies if the title and abstract were not relevant; however, we obtained papers for all potentially relevant studies if the abstract contained insufficient information to warrant exclusion. All areas of disagreement were discussed by the researchers to achieve a consensus.

Data extraction and management

Data from each study were extracted by all of the reviewers using standardised data extraction forms. The data sought included baseline characteristics of the participants (age and diagnostic criteria), the LGG dose and duration of the intervention, follow-up duration, the nature of the placebo and all outcome measures with their definitions, as reported by the authors. After extraction, all data were compared to minimise the possibility of errors.

Assessment of risk of bias in included studies

The reviewers independently, but without being blinded to the authors or journal, assessed the risk of bias in the studies that met the inclusion criteria. We used the Cochrane Collaboration’s tool for assessing risk of bias, which includes the following criteria: adequacy of sequence generation, allocation concealment, and blinding of participants, personnel and outcome assessors; and incomplete outcome data are addressed. In all cases, an answer of ‘yes’ indicates a low risk of bias and an answer of ‘no’ indicates a high risk of bias.16

Measures of treatment effect

The dichotomous outcomes, the results for individual studies and pooled statistics are reported as the risk ratio (RR) between the experimental and control groups with 95% CI. The standardised mean difference (SMD) between the treatment and control groups was selected to represent the difference in continuous outcomes (with 95% CI). For all outcomes, subgroup analyses based on the diagnosis (i.e., IBS, FAP and FD) were performed.

Dealing with missing data

We assessed pooled data using available case analysis, i.e., an analysis in which data are analysed for every participant for whom the outcome was obtained, rather than intention-to-treat analysis with imputation.17

Assessment of heterogeneity

Heterogeneity was quantified by χ2 and I2, which can be interpreted as the percentage of the total variation between studies that is attributable to heterogeneity rather than to chance. A value of 0% indicates no observed heterogeneity and larger values show increasing heterogeneity. If heterogeneity was not revealed, we present results of only the fixed effects model. If there was substantial heterogeneity (over 50%), all analyses were based on the random effects model if it was still considered appropriate to pool the data.

Data synthesis (Statistical methods)

The data were analysed using Review Manager (revman Version 5.1, The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark). The number needed to treat (NNT), with a 95% CI, was calculated using statsdirect statistical software (version 2,7,8; StatsDirect Ltd., England, UK). The weights given to each study were based on the inverse of the variance.

Results

Included studies

The literature search initially yielded 45 articles, of which three RCTs met the inclusion criteria (Figure 1).12, 18, 19 All were published in English. These trials randomised a total of 290 patients. Table 1 summarises the characteristics of the included studies. All three studies enrolled patients with IBS,12, 18, 19 two RCTs enrolled patients with FAP,12, 19 and one RCT also enrolled children with FD.19 In all of the studies, the diagnoses were made according to the Rome II criteria. The sample size ranged from 64 to 141 participants. In all of the studies, LGG was compared with placebo. The daily dose of LGG ranged from 109 CFU twice daily18 to 3 × 109 CFU twice daily,12, 19 and supplementation lasted for 4 weeks,19 6 weeks18 or 8 weeks.12 Two RCTs were undertaken in the two European countries Poland (one RCT)19 and Italy (one RCT)12 and one RCT was conducted in the United States.18

Figure 1.

 Identification process for eligible trials.

Table 1.   Characteristics of included studies
StudyParticipantsDiagnostic criteriaLGG (dose)ComparisonPrimary outcomeSecondary outcomes
  1. CFU, colony forming units; IBS, irritable bowel syndrome; FAP, functional abdominal pain; FD, functional dyspepsia; LGG, Lactobacillus rhamnosus GG; VAS, visual analogue scale.

Bausserman et al. 200518Children & adolescents (mean age: 12 years, range: 6–17 years)Rome II criteria for IBS1010 CFU, twice daily, for 6 weeksPlacebo (inulin)Change in the abdominal pain severity scoreNumber of responders vs. nonresponders in each group and changes in the symptoms on the Gastrointestinal Symptom Rating Scale (15-item GSRS) by syndrome. Responders were classified as patients with a decrease in abdominal pain severity (1 point or more on the 4-point Likert scale)
Francavilla et al. 201012Children (mean age: 6.4 years, range: 5–14 years)Rome II criteria for IBS or FAP3 × 109 CFU, twice daily, for 8 weeksPlacebo (inert powder)Change in abdominal pain (frequency/severity) according to the VAS score from baseline to the end of the treatment period(i) A decrease of at least 50% in the number of episodes and intensity of pain (treatment success); (ii) A decrease in the perception of children’s pain according to their parents; (iii) Modification of intestinal permeability
Gawronska et al. 200719Children (mean age: 11.6 years, range: 6–16 years)Rome II criteria for IBS, FAP, FD3 × 109 CFU, twice daily, for 4 weeksPlacebo (maltodextrin)Treatment success defined as no pain (a relaxed face, score of 0, on the Faces Pain Scale) at the end of the interventionImprovements defined as a change in: (i) the Faces Pain Scale by at least 2 faces scores; (ii) self-reported severity of pain during the preceding week measured on the Faces Pain Scale; (iii) self-reported frequency of pain during the preceding week; (iv) use of medication for abdominal pain; and (v) school absenteeism because of abdominal pain

Risk of bias in included trials

Table 2 shows the results of the methodological quality assessment. All of the studies had good methodological quality. The main limitation of the study by Bausserman et al. was incomplete outcome data (more than 20% lost to follow-up).18

Table 2.   Methodological quality summary: review authors’ judgments about each methodological quality item for each included study
Study IDAdequate sequence generation?Allocation concealment?Blinding?Incomplete data addressed?
  1. In all cases, an answer of ‘yes’ indicates a low risk of bias and an answer of ‘no’ indicates a high risk of bias.

Bausserman et al. 200518YesYesYesNo
50/64 (78%)
Gawronska et al. 200719YesYesYesYes
104/104 (100%)
Francavilla et al. 201012YesYesYesYes
136/141 (96%)

Heterogeneity

No heterogeneity for the primary outcome was found (I2 = 0%). For the secondary outcomes, significant heterogeneity (I2 ≥ 50%) was found for the overall study population for the frequency of pain (χ2 = 9.15, = 0.002, I2 = 89%) and the intensity of pain (χ2 = 2.21, = 0.14, I2 = 55%). In all cases, the observed statistical heterogeneity was not judged to be clinically relevant (i.e., studies consistently reported results in the same direction with clinically insignificant differences between the studies). However, there were too few studies to adequately determine heterogeneity.20

Effects of intervention

Primary outcome.  All three RCTs reported responders to the treatment or treatment success defined as no pain or a decrease in pain intensity (Figure 2). One RCT19 reported both treatment success (defined as no pain) and improvement of symptoms (defined as a change in the Faces Pain Scale by at least 2 faces scores). For this review, data related to those two outcomes were combined into a single outcome.

Figure 2.

 Primary outcome: effect of Lactobacillus GG on responder rates.

For the overall study population with abdominal pain-related FGD, compared with placebo, LGG supplementation was associated with a significantly higher rate of responders to the treatment (defined as no pain or a decrease in pain intensity)(three RCTs, = 290, RR 1.31, 95% CI 1.08–1.59, NNT 7, 95% CI 4–22). For a subgroup of children with IBS, those in the LGG group were more likely to respond to the treatment than those in the placebo group (three RCTs, = 167, RR 1.70, 95% CI 1.27–2.27, NNT 4, 95% CI 3–8). For the FAP group (two RCTs, = 103, RR 1.08, 95% CI 0.77–1.50), as well as for the FD group (one RCT, = 20, RR 0.83, 95% CI 0.37–1.85), we found no evidence that LGG supplementation influenced the treatment response in these children.

Secondary outcomes

Frequency of pain.  Only for a subgroup of children with IBS, the frequency of pain was reduced in those in the LGG group compared with those in the placebo group (two RCTs, = 117; SMD −1.04, 95% CI −1.43 to −0.65). For the overall study population, as well as for the subgroups of children with FAP or FD, there were no differences in the frequency of pain between the groups that received placebo or LGG (Figure 3).

Figure 3.

 Secondary outcome: effect of Lactobacillus GG on frequency of pain.

Intensity/severity of pain.  Compared with placebo, the use of LGG was associated with a significant decrease in the perception of pain intensity in the overall study population with abdominal pain-related FGD (two RCTs, = 240; SMD −0.44, 95% CI −0.82 to −0.05). Similarly, there was a reduction in pain intensity in the subgroup of children with IBS who received LGG compared with placebo (two RCTs, = 117; SMD −0.60, 95% CI −0.97 to −0.23), but not in children with FAP and FD (Figure 4).

Figure 4.

 Secondary outcome: effect of Lactobacillus GG on intensity/severity of pain.

Adverse effects.  The LGG was well tolerated and no adverse effects were reported.

Discussion

Summary of main results

This meta-analysis provides a summary of current knowledge regarding the effects of a single probiotic microorganism, LGG, in children affected by abdominal pain-related FGD. With the limited evidence available, we found that the use of LGG increased the proportion of responders to the treatment (defined as no pain or a decrease in pain intensity) in children with abdominal pain-related FGD, particularly among children with IBS. Additionally, LGG reduced the frequency and intensity of pain, again particularly among children with IBS. Of note, although positive and statistically significant, the effects were clinically modest.

Overall completeness and applicability of evidence

Our search included three major databases, with no language restrictions. The searching, extraction of the data and assessment of the validity of the studies were performed independently by two reviewers to decrease the likelihood of reviewer error or bias. However, we cannot fully exclude the possibility of publication bias, which is an important threat to the validity of systematic reviews and is difficult to combat except through the registration of all RCTs. We did not perform a statistical test for the detection of publication bias, as we are aware that these tests have very low power in the meta-analysis of only a few trials.21 One strength of our review, which distinguishes it from other reviews, is that it focuses on only one probiotic microorganism. This is because probiotic supplementation is not a homogeneous intervention. Pooling data from different genera, species, strains and doses of probiotics may result in misleading conclusions.

Quality of evidence

Any systematic review is only as good as the constituent studies. In general, the included RCTs were of sound methodology with adequate randomization, allocation concealment, blinding and follow-up. One potential limitation of the review is that it included a small number of trials with a small sample size, particularly when subgroups of patients with specific diagnoses were evaluated. However, to increase power is one of the reasons why a meta-analysis is performed within a systematic review.

In all of the included trials, the probiotic intervention lasted not less than 4 weeks, which is in line with the Rome Foundation document providing guidance for the design of treatment trials in patients with FGD.22 However, there is a lack of data on the long-term effects of LGG treatment for abdominal pain-related FGD. Only one study12 analysed children for 8 weeks after cessation of the therapy.

According to the published recommendations,22 the primary outcome measure for treatment studies in patients with FGD should be defined as ‘adequate relief’ or ‘satisfactory relief’. However, studies assessing this outcome measure in the paediatric population are lacking. To our knowledge, no outcome measures for FGD are sufficiently validated to be recommended unequivocally for use as the primary outcome measure.22 The studies included in our review used different scales and different definitions in assessing treatment success and treatment response. While in two trials, validated tools (Faces Pain Scale and Visual Analogue Scale) were used, the definition of treatment success was subjective.12, 19 In the remaining study,18 the investigators used a psychometric scale that was not validated for the paediatric population.

All three RCTs were placebo controlled, which is considered an essential requirement for interventional studies of FGD. However, there were differences in the choice of placebo. Two of the three RCTs used inulin and maltodextrin as their placebos.18, 19 As these are potential prebiotics, it could have affected the outcome. However, this was not observed in our analysis. Still, the choice of the placebo should be taken into account when designing future studies.

Agreement and disagreement with other studies or reviews

Compared with the previously published Cochrane review,11 which concluded that there is no evidence that lactobacillus supplementation is effective in the management of children with recurrent abdominal pain, our review focuses on a single probiotic microorganism and also includes recent data. Thus, our results more precisely define the effects of LGG. In adults, as discussed in the Introduction, some meta-analyses6–8 have shown that some probiotics are clinically more effective than placebo in the treatment of IBS. However, there are no data on LGG given as a single probiotic.

Implications for practice

The results of our systematic review provide preliminary evidence that LGG may be useful for treating children with abdominal pain-related FGD, particularly in a subgroup of children with IBS. While it is too soon to recommend its routine use in clinical practice, given the lack of effective therapy for abdominal pain-related FGD and the generally good safety profile of LGG in an otherwise healthy population, this therapeutic option could be discussed with patients and/or caregivers.

Implications for research

LGG seems to be a good candidate for a large multicentre trial in children with abdominal pain-related FGD, properly categorised into IBS, FAP and FD. As a result of the relapsing nature of the disease and that effects of probiotics may endure only for as long as the organism is administered, the treatment duration and follow-up period after cessation of treatment should allow for the adequate assessment of short- and long-term effects of LGG. There is also a strong need for a standardised, properly validated outcome assessments scale. As a result of the fact that the exact mechanism of action of LGG is not known, experimental studies are desirable.

Acknowledgements

Declaration of personal interests: H. Szajewska and A. Horvath served as speakers for Dicopharm, the manufacturer of Lactobacillus GG. AH and HS initially conceptualised this study. AH and PD were responsible for data collection. All authors were responsible for data analysis, interpretation and preparation of the report. All authors contributed to the writing of this manuscript and agreed upon the final version. Declaration of funding interests: This study was funded in full by The Medical University of Warsaw.

Ancillary