Intervention Review

You have free access to this content

Probiotics for treating persistent diarrhoea in children

  1. Guillermo Bernaola Aponte1,*,
  2. Carlos Alfonso Bada Mancilla2,
  3. Nilton Yhuri Carreazo3,
  4. Raúl Alberto Rojas Galarza4

Editorial Group: Cochrane Infectious Diseases Group

Published Online: 20 AUG 2013

Assessed as up-to-date: 13 DEC 2012

DOI: 10.1002/14651858.CD007401.pub3


How to Cite

Bernaola Aponte G, Bada Mancilla CA, Carreazo NY, Rojas Galarza RA. Probiotics for treating persistent diarrhoea in children. Cochrane Database of Systematic Reviews 2013, Issue 8. Art. No.: CD007401. DOI: 10.1002/14651858.CD007401.pub3.

Author Information

  1. 1

    Suárez Angamos Hospital, Lima, Peru

  2. 2

    Emergencias Pediátricas Hospital, Lima, Peru

  3. 3

    Universidad Peruana de Ciencias Aplicadas, Escuela de Medicina, Lima, Peru

  4. 4

    Instituto Nacional de Salud del Niño (Children's Hospital of Lima), Lima, Peru

*Guillermo Bernaola Aponte, Suárez Angamos Hospital, Angamos Este Avenue 261, Miraflores, Lima, 18, Peru. guiber0307@hotmail.com. guiber37@yahoo.es.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 20 AUG 2013

SEARCH

 
Characteristics of included studies [ordered by year of study]
Touhami 1992

MethodsStudy design: RCT; Duration: 3 years, from January 1987 to February 1990.


ParticipantsInclusion criteria: children aged 3 to 36 months with 3 or more liquid stools per day for > 13 days but < 29 days; and fed predominantly or exclusively milk formula at the time of onset of the diarrhoea.

Exclusion criteria: exclusive breast feeding; gross blood in the stools; additional disease requiring antimicrobial treatment; severe clinical malnutrition (weight/height < 70% NCHS);failure to give informed consent.

Number completing study: 36/40 (90%) in probiotic group (1 withdrawn by the parents, 4 developed urinary tract infection, 1 developed bronchiolitis); 35/38 (92.1%) in placebo group (3 developed urinary tract infection).


Interventions(1) Yogurt formula prepared by fermentation with streptococcus thermophilus and lactobacillus bulgaricus (4 x 108 CFU/Litre of yogurt).

(2) Non-fermented milk formula.

Intervention started after a 2-day observation period. The repartition of milk or yogurt varied with age as follow: between 3 and 6 months 120 mL/kg four times a day; 7 and 12 months 90 mL/kg three times a day; 13 and 18 months 60 mL/kg twice a day, during 5 days. ORS solution was offered according to WHO recommendations.


Outcomes(1) Treatment failure defined as a 5% loss of body weight in 24 hours, or if liquid stools were still present at the end of the 5-day period.

(2) Total number of stools.

(3) Stool volume.

No comment regarding adverse events.

Resolution of diarrhoea defined as the last liquid or semi-liquid stool evacuated before two formed stools in 12 hours, or no passage of stools for 12 hours before a formed stool.


NotesStudy location: Algeria

Stool analyses: Rotavirus 8, Campylobacters jejuni 3, EPEC 3, and Giardia Lamblia 2.

14 children were enrolled in error.

Source of funding: WHO and INSERM.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskInvestigators describe use of random permutation of fixed length, separately for boys and girls.

Allocation concealment (selection bias)High riskInformation obtained from trial author.

Blinding (performance bias and detection bias)
Duration of diarrhoea
High risk"...non-blinded clinical trial that lasted for 5 days."

Blinding (performance bias and detection bias)
Duration of vomiting
High risk"...non-blinded clinical trial that lasted for 5 days."

Blinding (performance bias and detection bias)
Stool frequency by day of treatment
High risk"...non-blinded clinical trial that lasted for 5 days."

Blinding (performance bias and detection bias)
Frequency of vomiting by day of treatment
High risk"...non-blinded clinical trial that lasted for 5 days."

Blinding (performance bias and detection bias)
Duration of hospital stay
High risk"...non-blinded clinical trial that lasted for 5 days."

Incomplete outcome data (attrition bias)
All outcomes
High risk6/40 missing outcome data from yogurt group; 3/38 missing outcome data from milk group.

The proportion of missing outcomes could induce clinically relevant bias in intervention effect estimate.

Selective reporting (reporting bias)Unclear riskInsufficient information.

Other biasUnclear riskThe incorrect enrolment of 14 children can introduce baseline imbalance.

Castañeda 1995

MethodsStudy design: RCT; Duration: no available.


ParticipantsInclusion criteria: diarrhoea for 3 or 4 weeks; persistence of diarrhoea after antibiotic treatment.

Exclusion criteria: not stated.

Number completing study: 20/20 (100%) in probiotic group and 20/20 (100%) in placebo group.


Interventions(1) Saccharomyces boulardii (5 x 109/250 mg) diluted with water or juice, twice daily for 30 days.

(2) Identical placebo, but without any active ingredient, at the same dose.

Timing of start of administration not stated.


Outcomes(1) Number of stools on a daily basis and assigned to categories of 1-3, 4-6, 7-10, and more than 10 per day.

(2) Degree of reversal of histological changes (partial, subtotal, total atrophy).

No adverse events observed.


NotesStudy location: Cuba.

Stool analyses (probiotic group/placebo group): Giardia lamblia 17/18, Shigella 2/2.

Duration of diarrhoea not measured.

Source of funding: no available.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low risk"...randomly assigned to treatment with the active product or with placebo on the basis of a table of random numbers".

Allocation concealment (selection bias)Unclear riskInsufficient information.

Blinding (performance bias and detection bias)
Duration of diarrhoea
Unclear riskInsufficient information to know who were blinded.

Blinding (performance bias and detection bias)
Duration of vomiting
Unclear riskInsufficient information to know who were blinded.

Blinding (performance bias and detection bias)
Stool frequency by day of treatment
Unclear riskInsufficient information to know who were blinded.

Blinding (performance bias and detection bias)
Frequency of vomiting by day of treatment
Unclear riskInsufficient information to know who were blinded.

Blinding (performance bias and detection bias)
Duration of hospital stay
Unclear riskInsufficient information to know who were blinded.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo missing outcome data.

Selective reporting (reporting bias)Unclear riskInsufficient information.

Other biasUnclear riskInsufficient information to assess the baseline balance.

Gaón 2003

MethodsStudy design: RCT; Duration: 2 years, from January 1996 to April 1998.


ParticipantsInclusion criteria: a history of frequent loose stools (>3 per day) for the last consecutive 14 days or more; not being breast fed; no history of allergy to cow's milk; no history of treatment with antimicrobials or antidiarrhoeal agents within the preceding 7 days; ability to take oral food.

Exclusion criteria: presence of concurrent systemic illness; a weight for age < 60% of the value for the 50 th percentile according to the tables of the NCHS; dehydration of more than 10% of body weight (severe).

Number completing study: 30/31 (96.8%) in saccharomyces group (1 developed urinary tract infection); 30/31 (96.8%) in lactobacillus group (1 developed urinary tract infection); and 29/31 (93.5%) in control group (2 developed vomiting).


Interventions(1) Pasteurized cow's milk containing Saccharomyces boulardii (1010 - 1012 CFU/g), 175 g twice a day for a 5 day period.

(2) Pasteurized cow's milk containing Lactobacillus casei and Lactobacillus acidophilus sp (1010 - 1012 CFU/g), 175 g twice a day for a 5 day period.

(3) Identical placebo (pasteurised cow milk).

Intervention started after the initial correction of dehydration. After rehydration and until diarrhoea stopped, ORS solution was offered according to WHO recommendations.


Outcomes(1) Duration of diarrhoea.

(2) Stool frequency by day of treatment.

(3) Proportion of participants with diarrhoea at end of intervention (day 5).

(4) Duration of symptoms.

Resolution of diarrhoea defined as an unformed stool follow by two stools that retains its shape and does not stick to the container (formed stool).

Results for rotavirus subgroup also presented.

No adverse events observed.


NotesStudy location: Argentine

Stool analyses (saccharomyces group/ lactobacillus group/ placebo group): Rotavirus 9/8/7; E. coli, Salmonella, Shigella 4/3/4.

Patients with malnutrition 6/30 (20%) in saccharomyces group, 7/30 (23.3%) in lactobacillus group, and 6/29 (20.7%) in placebo group.

Stool frequency by day of treatment derived from graph.

Source of funding: Sancor Cooperativas Unidas and CONICET.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandomised, but no further details given.

Allocation concealment (selection bias)Unclear riskThe method of concealment is not describe in sufficient detail.

Blinding (performance bias and detection bias)
Duration of diarrhoea
Unclear riskInsufficient information to know who were blinded.

Blinding (performance bias and detection bias)
Duration of vomiting
Unclear riskInsufficient information to know who were blinded.

Blinding (performance bias and detection bias)
Stool frequency by day of treatment
Unclear riskInsufficient information to know who were blinded.

Blinding (performance bias and detection bias)
Frequency of vomiting by day of treatment
Unclear riskInsufficient information to know who were blinded.

Blinding (performance bias and detection bias)
Duration of hospital stay
Unclear riskInsufficient information to know who were blinded.

Incomplete outcome data (attrition bias)
All outcomes
Low risk2/62 missing outcome data from intervention group (both due to vomiting); 2/31 missing outcome data from control group (both due to urinary tract infection).

Selective reporting (reporting bias)High riskThe pre-specified treatment failure outcome was not reported.

Other biasUnclear riskSmall number of patients were recruited in the three study groups.

Basu 2007

MethodsStudy design: RCT; Duration: 2 years, from January 2003 to December 2004.


ParticipantsInclusion criteria: history of diarrhoea persisting for 14 days or more without any remission in between; stool pH < 5.5; stool reducing substance > 1%.

Exclusion criteria: presence of any systemic illness other than diarrhoea on admission; development of any systemic complication of diarrhoea during hospital stay; failure to give informed consent.

Number completing study: 117/125 (93.6%) in probiotic group (1 developed septicaemia, 1 developed renal failure, 4 withdrew consent, 2 discharged on request) and 118/128 in control group (3 developed septicaemia, 3 withdrew consent, 4 discharged on request).


Interventions(1) Lactobacillus GG powder containing 60 million cells was dissolved in 100 ml of ORS twice daily for a minimum period of 7 days or till diarrhoea stopped.

(2) ORS twice daily for the same duration.

Intervention started after the initial correction of dehydration. All participants with positive stool culture received antibiotics. All participants were given lactose-free diet; however, breast fed children continue breastfeeding.


Outcomes(1) Duration of diarrhoea.

(2) Duration of vomiting.

(3) Stool frequency by day of treatment.

(4) Frequency of vomiting by day of treatment.

(5) Duration of hospital stay.

No complication could be documented.


NotesStudy location: India

Stool analyses (probiotic group/placebo group): E. Coli 12/10; Shigella spp 9/7; C. difficile 6/8; E. histolytica 7/9; G. lamblia 5/8; mixed infections 6/3.

Patients with PEM (body weight < 80% expected)107/117 (91.5%) in the probiotic group and 105/118 (89%) in placebo group.

Source of funding: no available.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low risk"Randomization was performed by computer generated random numbers"

Allocation concealment (selection bias)Low risk"...by opaque and sealed envelops."

Blinding (performance bias and detection bias)
Duration of diarrhoea
Low riskBlinding of participants and key study personnel ensured.

Blinding (performance bias and detection bias)
Duration of vomiting
Low riskBlinding of participants and key study personnel ensured.

Blinding (performance bias and detection bias)
Stool frequency by day of treatment
Low riskBlinding of participants and key study personnel ensured.

Blinding (performance bias and detection bias)
Frequency of vomiting by day of treatment
Low riskBlinding of participants and key study personnel ensured.

Blinding (performance bias and detection bias)
Duration of hospital stay
Low riskBlinding of participants and key study personnel ensured.

Incomplete outcome data (attrition bias)
All outcomes
Low risk8/125 missing outcome data from intervention group; 10/128 missing outcome data from control group.

Reasons similar across groups.

Selective reporting (reporting bias)Low riskPublished report include all expected outcomes.

Other biasLow riskThe study appears to be free of other sources of bias.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Balli 1992Included children with chronic non-specific diarrhoea only.

Boudraa 1990Preliminary report of included trial.

Chouraqui 1995Unpublished data; unable to scrutinize.

Dinleyici 2011Included children with gastrointestinal symptoms, but not all children had diarrhoea

Locascio 2002No mention of randomization.

Roggero 1990Included children with chronic non specific diarrhoea only.

 
Comparison 1. Probiotic versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 duration of diarrhoea2324Mean Difference (IV, Fixed, 95% CI)-4.02 [-4.61, -3.43]

 
Summary of findings for the main comparison. Probiotic compared to placebo for treating children with persistent diarrhoea

Probiotic compared to placebo for treating children with persistent diarrhoea

Patient or population: Children with persistent diarrhoea
Settings:
Intervention: Probiotic
Comparison: Placebo

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

PlaceboProbiotic

Duration of diarrhoea
days
The mean duration of diarrhoea ranged across control groups from
8.5-9.2 days
The mean Duration of diarrhoea in the intervention groups was
4.02 lower
(4.61 to 3.43 lower)
324
(2 studies5)
⊕⊕⊕⊝
moderate1,2,3,4

Stool frequency on day 5See commentSee commentNot estimable327
(2 studies5)
⊕⊕⊝⊝
low1,3,6,7
Both studies showed a benefit with probiotics, however the size of the benefit was very different in the two trials so the data were not pooled

Hospital stay
days
The mean hospital stay in the control groups was
15.5 days
The mean Hospital stay in the intervention groups was
8.2 lower
(8.6 to 7.8 lower)
235
(1 study12)
⊕⊕⊕⊝
moderate8,9,10,11

Death from any cause - not measuredSee commentSee commentNot estimable-See commentNot estimable

Weight-for-age z score - not measuredSee commentSee commentNot estimable-See commentNot estimable

*The basis for the assumed risk (eg the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1 No serious study limitations: Basu 2007 adequately concealed allocation and blinded both participants and study staff to be considered at low risk of bias. As this was the larger study contributing 86% of the data to the meta-analysis we did not downgrade for study limitations. Gaon 2003 did not adequately describe the study methodology.
2 No serious inconsistency: There was no statistical heterogeneity. I2 = 0%
3 Serious indirectness: Only two studies, one from India and one from Argentina have assessed this comparison. Before the result can be confidently generalised to all situations further studies may be necessary. Basu 2007 included paediatric patients (mean age 4.2 yrs) who had had diarrhoea for 14 consecutive days, a stool pH <5.5 and reducing substances <1%. 90% of participants had weight for age <80% of expected. The intervention was ORS plus lactobacillus vs ORS alone. Gaon 2003 included children age 6-24 months, and excluded breastfed infants, cow's milk allergies, and children with <60% of the 50th percentile for weight, or >10% dehydration. The intervention was cow's milk plus lactobacillus or S. boulardii vs cow's milk alone.
4 No serious imprecision: The 95% CI around the pooled effect is narrow. Even the lower limit suggests a clinically important reduction in the duration of diarrhoea by 3 days.
5 Gaon 2003: A hospital-based study of 93 children in Argentina, and Basu 2007: A hospital-based study of 125 children in India
6 Serious inconsistency: There is substantial heterogeneity between the two trials: I2 test for heterogeneity 80%. Both trials showed a benefit with probiotics, however the size of this effect was much larger in Basu 2007. The heterogeneity may be due to differences in the trial methodology, the characteristics of the study population or the nature of the different interventions. Due to the limited number of studies we were unable to further investigate this.
7 No serious imprecision: Not downgraded for imprecision. Data not pooled due to the substantial heterogeneity.
8 No serious limitations: Basu 2007 adequately concealed allocation and blinded both participants and study staff to be considered at low risk of bias.
9 No serious inconsistency: Not applicable as only one trial
10 Serious indirectness: Only one study has reported this outcome. Before the result can be confidently generalised to all situations further studies may be necessary.
11 No serious imprecision: Both limits of the 95% CI suggest a clinically important reduction in hospital stay in the participants given the intervention.
12 Basu 2007
 
Table 1. Detailed search strategies

Search setCIDG SRaCENTRALMEDLINEbEMBASEbLILACSb

1diarrheadiarrheadiarrheadiarrheadiarrhea

2diarrhoeadiarrhoeadiarrhoeadiarrhoeadiarrhoea

31 or 21 or 21 or 21 or 21 or 2

4probiotic*lactobacill*lactobacill*lactobacill$probiotic$

5lactobacill*lactococc*lactococc*lactococc$lactobacill$

6Bifidobacter*Bifidobacter*Bifidobacter*Bifidobacter$Bifidobacter$

74 or 5 or 6Enterococc*Enterococc*Enterococc$4 or 5 or 6

83 and 7Streptococc*Streptococc*Streptococc$3 and 7

9child*SaccharomycesSaccharomycesSaccharomyceschild$

10infant*4-9/OR4-9/OR4-9/ORInfant$

11pediatr*3 and 103 and 103 and 10pediatr$

129 or 10 or 11child*child*child$9 or 10 or 11

138 and 12Infant*Infant*Infant$8 and 12

14pediatr*pediatr*pediatr$

1512 or 13 or 1412 or 13 or 1412 or 13 or 14

1611 and 1511 and 1511 and 15

 aCochrane Infectious Diseases Group Specialized Register.
bSearch terms used in combination with the search strategy for retrieving trials developed by The Cochrane Collaboration (Lefebvre 2008); upper case: MeSH or EMTREE heading; lower case: free text term.