Probiotics for treating acute infectious diarrhoea

 

Characteristics of included studies [ordered by study ID]

---

Methods	Randomized controlled trial; 1 centre Duration: 1 year (January -December 2003)
Participants	Inclusion criteria: inpatients; infants and children with ≥ 3 watery stools/day without visible blood or mucus (duration not stated); < 10 white blood cells/high power field and no red cells, mucus flakes and bacteria on stool microscopy; negative hanging drop preparation; negative bacterial stool culture. Exclusion criteria: systemic illness other than diarrhoea on admission; systemic complication of diarrhoea during hospital stay; failure to give informed consent. Number completing study: 323/330 (97.9%) in the probiotic group (3 participants had electrolyte imbalance, 2 had septicaemia, 2 withdrew consent); 323/332 (97.3%) in control group (3 participants had electrolyte imbalance, 2 had septicaemia, 2 withdrew consent, 1 was discharged, 1 died).
Interventions	Live L. rhamnosus GG (120 x 10^6 CFU/day for 7 days) ORF Dehydration was corrected using oral rehydration fluid (ORF) following WHO guidelines
Outcomes	Frequency of diarrhoea Duration of diarrhoea (time to 2 consecutive soft or formed stools or no stool for 12 consecutive hours) Duration of vomiting Length of hospital stay No adverse events attributed to probiotic.
Notes	Study location: India (high child and adult mortality) Cause of diarrhoea: bacterial diarrhoea excluded. Rotavirus identified in 241 (74.6%) probiotic and 249 (77.1%) control group. Nutritional status: most participants malnourished: probiotic group; 198/323 moderately malnourished, 31/323 severely malnourished; control group; 185/323 moderately malnourished, 33/323 severely malnourished. Hydration status: all participants dehydrated: probiotic group: 48 mild, 173 moderate, 102 severe dehydration; control group: 51 mild, 168 moderate, 104 severe dehydration. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	computer randomization
Allocation concealment?	Yes	concealed in envelopes
Blinding? All outcomes	Yes	double blind
Incomplete outcome data addressed? All outcomes	Yes	Follow-up ≥90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 1 year (period not stated)
Participants	Inclusion criteria: inpatients; infants and children with ≥3 watery stools/day, without macroscopic blood or mucus, white cells < 10 high power field and absent red blood cells, mucus flakes and bacteria on stool microscopy, negative hanging drop preparation and negative bacterial stool culture. Exclusion criteria: symptoms of illness other than diarrhoea; development of any systemic complication of diarrhoea during hospitalization; failure to give informed consent. Number completing the study: probiotic group: 186/196 (94.9%; withdrawals: 5 electrolyte imbalance, 3 septicaemia, 2 withdrew consent); placebo group: 185/196 (94.4%; withdrawals: 4 electrolyte imbalance, 3 septicaemia, 2 withdrew consent; 1 discharged on request; 1 died).
Interventions	Live L. rhamnosus GG 2 x 10^10 CFU/day for minimum 7 days or until diarrhoea stopped (data not extracted for meta-analysis) Live L. rhamnosus GG 2 x 10^12 CFU/day for minimum 7 days or until diarrhoea stopped (data extracted for meta-analysis ORF Interventions started after initial rehydration and stabilization.
Outcomes	Frequency of diarrhoea by day Average duration of diarrhoea Average duration of vomiting Average duration of IV therapy Average duration of hospital stay No adverse events attributed to probiotic.
Notes	Study location: India (high child and adult mortality) Cause of diarrhoea: bacterial diarrhoea excluded. Rotavirus identified in 106 (57.0%) probiotic and 102 (55.1%) control group. Nutritional status: severe malnutrition in 17 (9.1%) probiotic and 12 (6.5%) control group; mild/moderate malnutrition in 102 (54.8%) probiotic and 100 (54.1%) control group. Hydration status: severe dehydration in 35 (18.8%) probiotic and 39 (21.1%) control group; mild/moderate dehydration in 121 (65.1%) probiotic and 122 (66.0%) control group. Source of funding not stated but no authors had a financial arrangement regarding this study
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Computer-generated random numbers
Allocation concealment?	Yes	Opaque, sealed envelopes
Blinding? All outcomes	Yes	Interventions prepared by pharmacy; packets of similar appearance
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 2 centres. Duration: 16 months
Participants	Inclusion criteria: inpatients; malnourished boys (weight for height < 80% NCHS median) with diarrhoea (≥ 5 liquid stools in preceding 24 hours) for ≤ 96 hours. Nearly all children were dehydrated (48/49 milk group and 43/47 yogurt group). Exclusion criteria: females; severe non-gastrointestinal illness; gross blood in the stools; exclusive breast-feeding. Number completing study: 47/49 (95.9%) in probiotic group (2 withdrawn because cholera in stool cultures); 49/53 (92.5%) in control group (2 withdrawn because cholera in stool cultures and 2 left against medical advice).
Interventions	Yogurt formula (Lactogen-2, Nestle India Ltd; after fermentation with 90 g S. thermophilus and Lactobacillus bulgaricus standard starter (International Yoghurt Manufacturers Club, Paris) 120 mL/kg/day for at least 72 hours) added to milk formula Non-fermented Lactogen-2 Given after 8 hours initial observation. All participants received rehydration fluids (IV if stool > 4 g/kg/hour), IV cephalosporin and gentamicin, and fed with rice lentil oil gruel.
Outcomes	Proportion recovered at 48 hours and 72 hours (defined as 2 consecutive formed stools, ≤3 stools in 24 hours of which at least 2 were formed, or no stool for 12 hours) Median duration of diarrhoea Treatment failures (episode of diarrhoea after 72 hours or stool weight > 150 g/kg on any day) No comment regarding adverse events.
Notes	Study location: India (high child and adult mortality). Cause of diarrhoea: excluded if gross bloody stools. Nutritional status: all malnourished boys (weight for height < 80% NCHS median); mean weight for length and length for age (% NHCS median) similar in both groups. Hydration status: Nearly all children were dehydrated: 43/47 (91.5%) probiotic and 48/49 (98.0%) control group. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	randomisation list
Allocation concealment?	Unclear	not stated
Blinding? All outcomes	Unclear	probably open study
Incomplete outcome data addressed? All outcomes	Yes	Follow-up ≥ 90% in both groups

---

---

Methods	Randomized trial; probably open study; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with acute watery diarrhoea of mild to moderate severity Exclusion criteria: Severe intercurrent illness; severe diarrhoea and dehydration requiring admission and IV rehydration; temperature > 38.5°C; anti-diarrhoeals or antibiotics in last 24 hours; severe malnutrition Number completing study: 50/50 (100%) in probiotic group; 50/50 (100%) in control group.
Interventions	S. boulardii (500mg/day for 5 days) ORF and nutritional support only Timing of interventions not stated.
Outcomes	Stoppage of diarrhoea (not defined) Weight gain Daily stool frequency and consistency Tolerance and acceptability of intervention No adverse events attributed to probiotic.
Notes	Study location: Pakistan (high child and adult mortality) Cause of diarrhoea: Rotavirus identified in 8 (16.0%) probiotic and 10 (20.0%) control group. Bacterial diarrhoea identified in 13 (26.0%) probiotic and 6 (12.0%) control group. Nutritional status: severe malnutrition excluded; no further data presented Hydration status: severe dehydration excluded; no further data presented Source of funding: supported by Laboratoires Biocedex (France); Hilton Pharma (Pvt.) Ltd. Pakistan
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Randomized controlled trial but methods not described
Allocation concealment?	Unclear	Methods not described
Blinding? All outcomes	No	No placebo; probably open study
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; well-nourished children aged 3-24 months with watery diarrhoea < 5 day duration and > 3 watery stools in previous 24 hours. All children were dehydrated, including some with severe dehydration. Exclusion criteria: exclusive breast feeding, history of allergy to cow's milk, severe malnutrition (weight or height < 70% or oedema) Number completing study: 49/56 (87.5%) in probiotic group (3 with urinary tract infection and 1 with bronchopneumonia withdrawn, others withdrawn by parents) and 48/56 (85.7%) in non-probiotic group (2 with urinary tract infection, 1 with amebiasis withdrawn and 1 failed to attend for follow up, others withdrawn by parents). Reasons for withdrawal by parents not stated. Diarrhoea outcomes reported for all randomized children.
Interventions	Infant formula (Enapal-Sopad, Nestlé, Courbevoie, France) fermented with L. bulgaricus and S. thermophilus (Yalacta, Caen, France; total 2 x 10^8 CFU/g). Infant formula acidified with lactic acid to match pH of fermented formula 180 mL/kg/day of either fermented or non-fermented infant formula given after initial oral rehydration. All infants also received other foods.
Outcomes	Weight gain Cessation of diarrhoea (defined as last liquid or semi-liquid stool before 2 formed stools). Means and 95% CIs stated Food and liquid intake Frequency of vomiting similar in both groups. No other comment regarding adverse events.
Notes	Study location: Algeria (high child and adult mortality). Cause of diarrhoea: rotavirus identified in 25/56 (44.6%) probiotic and 26/56 (46.4%) in control group. No bacterial pathogens isolated. Nutritional status: all well-nourished Hydration status: all dehydrated; severe dehydration in 5 (8.9%) in the probiotic and 4 (7.1%) in the control group. Reduced duration of diarrhoea in the probiotic compared with non-probiotic group observed only in children with reducing substances in stools. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	No	Stated as double blind but mothers able to distinguish fermented from non-fermented infant formula
Incomplete outcome data addressed? All outcomes	No	Follow up < 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre. Duration: 3 years
Participants	Inclusion criteria: inpatients; young children with acute diarrhoea (definition not stated; 3/4 had diarrhoea < 3 days); weight loss of at least 5%. Exclusion criteria: any treatment that could have affected diarrhoea during hospitalization. Number completing study: 38/38 (100%) in probiotic group and 33/33 (100%) in control group.
Interventions	Killed L. acidophilus (LB strain, Lacteol Forte, France; 1 sachet thrice daily for first 24 hours, then 1 sachet daily for next 3 days) Placebo (no details provided; same regimen) Loperamide Timing of start of administration not stated. All young infants were given Pregestimil, and older children were given an anti-diarrhoeal diet.
Outcomes	Time to first normal stool Failure defined as no improvement by the end of day 2 (clinical criteria) No adverse events attributed to probiotic.
Notes	Study location: France (very low child and adult mortality). Cause of diarrhoea: 18% all participants had positive stool cultures and 49% positive virology tests (no further details given). Nutritional status: no data presented. Hydration status: all dehydrated with weight loss of at least 5%. Results presented for oral rehydration group only and all children. Resolution of diarrhoea in killed L. acidophilus group similar for rotavirus positive and negative participants. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Random number table stratified in groups of 18
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Not described
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre. Duration: not stated
Participants	Inclusion criteria: inpatients; adults with acute enteritis (diarrhoea, fever, vomiting, nausea, abdominal pain with or without toxicity; duration not stated). Exclusion criteria: typhoid cases. Number completing study: stool cultures available after randomization; participants with Salmonella typhi withdrawn (number not stated); for non-typhoid participants, results presented for 25/25 (100%) in probiotic group and 24/24 (100%) in control group.
Interventions	Enterococcus LAB SF68 (Bioflorin; ≥75 x 10^6 lyophilized bacteria tds for 10 days) Placebo Timing of start of administration not stated.
Outcomes	Proportion of participants with diarrhoea by day of treatment Resolution of diarrhoea defined as 2 or less formed stools/day and no abdominal pain or fever. No adverse events attributed to probiotic.
Notes	Study location: Italy (very low child and adult mortality). Cause of diarrhoea: non-typhoid. Bacterial stool culture (probiotic group/placebo group): Salmonella 4/3; enteropathogenic E. coli 18/20; other enteropathogen 1/3. Nutritional status: no data presented. Hydration status: no data presented. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre. Duration: not stated
Participants	Inclusion criteria: inpatients; adults with acute febrile enteritis (duration of diarrhoea not stated). Exclusion criteria: typhoid cases. Number completing study: 10/10 (100%) in the probiotic group and 11/11 (100%) in the control group.
Interventions	Enterococcus LAB SF68 (Bioflorin; ≥75 x 10^6 lyophilized bacteria thrice daily for at least 10 days) Placebo Intervention started after initial treatment with chloramphenicol (all participants) and after stool culture results available.
Outcomes	Proportion of participants with diarrhoea by day of treatment (definition for recovery from diarrhoea not stated). No adverse events attributed to probiotic.
Notes	Study location: Italy (very low child and adult mortality). Cause of diarrhoea: non-typhoid. Nutritional status: no data presented. Hydration status: no data presented. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Randomization list
Allocation concealment?	No	Not described
Blinding? All outcomes	No	Not described
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 2 centres. Duration: not stated
Participants	Inclusion criteria: inpatients and outpatients; adults with acute diarrhoea (>= 3 watery or loose stools in last 24 hours). Exclusion criteria: diarrhoea > 3 days; blood in faeces; faecal leukocytes; temperature > 39 °C; friable and haemorrhagic mucosa in rectosigmoid; history of chronic diarrhoea; polyps; colon cancer; Crohn's disease; ulcerative colitis; malabsorption; use of antidiarrhoeals or antibiotics in past 7 days; severe diarrhoea (dehydration with weight loss >10%); associated major diseases. Number completing study: 93/105 (88.6%) in probiotic group (4 violated protocol, 5 did not comply with study medications, 3 lost to follow up) and 92/106 (86.8%) in control group (5 violated protocol, 7 did not comply with study medications, 2 lost to follow up).
Interventions	Enterococcus strain SF68, lyophilized (Bioflorin; 75 x10^6 CFU thrice daily for ≥5 days) Placebo Started on day of presentation.
Outcomes	Number of participants with diarrhoea by day of treatment Mean stool frequency by day of treatment Diarrhoea resolved when stool frequency < 3/day and semisolid or solid and no associated symptoms. No adverse events attributed to probiotic.
Notes	Study location: Belgium (very low child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded. Bacterial diarrhoea identified in 12 (11.4%) in the probiotic and 16 (15.1%) in the control group. Nutritional status: no data presented Hydration status: > 10% dehydration excluded; no further data presented. Highly significant reduction in duration of diarrhoea in the probiotic group confirmed by an intention-to-treat analysis, which included the excluded participants as non-recovered on day 7 (but no data shown). Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Randomization by central computer
Allocation concealment?	Yes	Randomization by central computer
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	No	< 90% follow-up in probiotic and placebo groups

---

---

Methods	Randomized controlled trial; 6 centres Duration: 12 months, October 1999 to September 2000
Participants	Inclusion criteria: outpatients; infants and children aged 3 to 36 months with >2 loose or liquid stools/day for <48 hours. Exclusion criteria; malnutrition, severe dehydration; coexisting acute systemic illness (meningitis, sepsis, pneumonia), immunodeficiency; underlying severe chronic disease; cystic fibrosis; food allergy or other chronic GI diseases; use of probiotics in the previous 3 weeks; antibiotics or any other antidiarrhoeal medication in the previous 3 weeks; poor compliance (< 4 doses of the study medication administered). Number completing study: 95/100 in the probiotic group (2 did not receive the allocated intervention, 1 faster remission, 1 worsening symptoms, 1 poor compliance); 88/92 in the control group (1 did not receive the allocated intervention, 1 worsening symptoms, 1 contracted pneumonia, 1 had coeliac disease).
Interventions	Live Lactoacillus casei rhamnosus GG (Dicoflor 60; 12 x 10^9 CFU/day for 5 days) Placebo, no details given but same appearance as active intervention. Intervention started within 48 hours of admission. ORF given for 3-6 hours after admission, lactose-containing formula milk or cow's milk according to age.
Outcomes	Diarrhoea duration (time of the last loose or liquid stool preceding a normal stool) Number and consistency (scoring system) of stools/day recorded by parents Vomiting Fever (> 37.5°C) Number of hospital admissions 1 patient with poor compliance in the probiotic group; 31 and 34 participants had vomiting in the probiotic and placebo groups, respectively. No adverse events attributed to probiotic.
Notes	Study location: Italy (very low child and adult mortality). Cause of diarrhoea: stool culture in only few participants; no data presented. Nutritional status: malnutrition excluded Hydration status: severe dehydration excluded; no other data presented. Source of funding: none Single blind trial. Parents instructed to buy probiotic preparation. This study also allocated children to 4 other probiotic groups: 1) S. boulardii It 5 × 10^9 live organisms daily (Codex) for 5 days; 2) Bacillus clausii O/C84, N/R84, T84, SIN84 (Enterogermina) 10^9 CFU bd for 5 days; 3) a combination of L. delbrueckii var bulgaricus LMG-P17550 10^9 CFU daily, L. acidophilus LMG-P 17549 109 CFU daily, S. thermophilus LMG-P 17503 10^9 CFU daily, B. bifidum LMG-P 17500 5 × 10^8 CFU daily (Lactogermina) for 5 d; 4) Enterococcus faecium SF 68 (Bioflorin) 7.5×10^7 CFU daily for 5 days and compared each of the probiotic groups with the single control group. Mean duration of diarrhoea and mean stool frequency on day 2 and 3 were significantly shorter than in the control group for intervention groups 1 and 3. These outcomes were similar to the control group for the other probiotic groups. To avoid a unit-of-analysis error as a result of the multiple comparisons between the intervention groups and the single control group, we elected to include data for the L. GG group only in this review. We selected L. GG because this was the probiotic most frequently evaluated in acute infectious diarrhoea and we wished to maximize the body of evidence. We rejected the alternative approach of pooling the data from all of the different probiotic intervention groups into a single group because this would not be helpful in selecting a specific probiotic intervention for use in clinical practice.
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Computer-generated randomization list allocation in blocks of 6
Allocation concealment?	Yes	Concealed until treatment assigned
Blinding? All outcomes	Yes	Blinded third-party blind assessor
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients and outpatients; children with non-bloody diarrhoea (not defined) of less than 5 days duration. Exclusion criteria: antimicrobials in the last 72 hours; concomitant illness; severe malnutrition; antidiarrhoeal drugs; immunocompromised. Participants completing study: 35/35 (100%) in probiotic group and 35/35 (100%) in control group.
Interventions	L. acidophilus and L. bifidus (Infloran Berna; dose and duration not stated). Placebo (no details given; unclear whether or not placebo was identical to probiotic). No details of when interventions started.
Outcomes	Resolution of diarrhoea (defined as no passage of stool for 12 hours or 2 consecutive formed stools). Assessed in outpatients by phoning the parents. No adverse events attributed to probiotic.
Notes	Unpublished data. Study location: Philippines (low child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded. Nutritional status: severe malnutrition excluded; no other data presented. Hydration status: overall, 42 children had some dehydration (none severe) and 28 had no dehydration Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Unclear whether placebo identical to probiotic
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 11 months, 1April 1988 to 15 March 1989
Participants	Inclusion criteria: unclear whether inpatients or outpatients, or both; children aged 3 months to 3 years with acute (duration not stated) non-bloody diarrhoea; no dehydration; no concomitant illness; no antibiotics or drugs affecting gut motility. Number completing study: unclear how many participants randomized; participants who deteriorated, developed concomitant illness, and needed other drugs, or who wished to withdraw were excluded from the analysis (details not given).
Interventions	S. boulardii (live Saccharomyces cerevisiae Hansen CBS 5926; 600 mg/day; duration not stated) Glucose placebo (diluted in 5 mL cold water). No details of when interventions started.
Outcomes	Number of stools per day First day stools formed Side effects Cure defined as < 4 stools in 24 hours and absence of liquid stools. No adverse events attributed to probiotic.
Notes	Study location: Mexico (low child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded Nutritional status: all well nourished. Hydration status: dehydration excluded. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Random table
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Unclear whether placebo was identical to the probiotic
Incomplete outcome data addressed? All outcomes	Unclear	Unclear how many participants were randomized at beginning of study

---

---

Methods	Intervention study; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with sudden, recent onset of watery diarrhoea (not defined) of variable importance with or without fever and vomiting. Exclusion criteria: dehydration >10% needing IV rehydration; bloody or purulent stools; fever >39°C; associated pathology. Number completing study: 19/19 (100%) probiotic group and 19/19 (100%) control group.
Interventions	Live S. boulardii (500 mg/day for 5 days) ORF When the probiotic was administered was not stated.
Outcomes	Mean number of stools, mean stool weight and carmine red transit time on days 1 and 4. Stool consistency on day 4. Stool frequency on day 4 was lower in the probiotic than the control group (n = 19; mean 2.1 [SD 0.9] versus n = 19; 3.4 [1.9] respectively). The reduction in stool frequency from baseline was statistically significantly greater in the probiotic than control group (P < 0.01). No adverse events attributed to probiotic.
Notes	Location: France (very low child and adult mortality) Cause of diarrhoea: bloody or purulent stools excluded; pathogenic bacteria isolated from 9 children in the probiotic and 6 in the control group. Nutritional status: no data presented. Hydration status: dehydration > 10% needing IV rehydration excluded; Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	No	Infants allocated alternately to the two groups as enrolled in trial
Allocation concealment?	No	Alternate allocation
Blinding? All outcomes	No	No placebo; open study
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 22 months; February 2006 to November 2007
Participants	Inclusion criteria: inpatients; children aged 3 months to 6 years with acute diarrhoea defined as 3 or more loose or liquid stools per day of less than 72 hours duration. Exclusion criteria: immunodeficiency, severe abdominal distension with risk of bowel perforation, severe infection or sepsis, history with gastrointestinal tract surgery, probiotics use in the preceding 1 week. Number completing study: 304 children enrolled and 293 were included in the analysis (150 in the probiotic and 143 in the control group). Overall, 7 children discontinued medication and 4 were lost to follow up; group allocation unclear.
Interventions	Live Bacillus mesentericus, Enterococcus faecalis, and Clostridium butyricum (Bio-three; 2.5 x 10^7 CFU/kg/d) for 7 days Starch powder of identical appearance to probiotic preparation When interventions started not stated.
Outcomes	Duration of diarrhoea (time from inclusion into the study until the first normal stool was passed) No. of diarrhoea episodes Mean stool frequency on days 2 and 3 Diarrhoea lasting ≥ 3 days Duration of fever Duration of vomiting Appetite/intake score Abdominal pain episodes Length of hospital stay Duration of diarrhoea also reported for children with rotavirus diarrhoea and those with bacterial diarrhoea No adverse events attributed to probiotic.
Notes	Study location: Taiwan (low child and adult mortality). Cause of diarrhoea: 47 (31.3%) of children in probiotic and 44 (30.8%) in control group had rotavirus in stools. Norovirus and adenovirus also identified. 27 (18.0%) children in probiotic and 30 (20.0%) in the control group had bacteria in stools (either Salmonella enterica or Campylobacter jejuni). Nutritional status: no data presented Hydration status: no data presented Source of funding: The study was supported in part by a grant from Chang Gung Memorial Hospital research project grant XMRPG440021, Northern Taiwan. First author was contacted and asked to clarify that children who had received antibiotics before recruitment were included that children with blood in stools were included whether they could provide outcome results separately for rotavirus diarrhoea hydration status nutritional status
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; boys, age 1 to 24 months with acute diarrhoea (3 or more watery or loose stools per 24 hours during at least one 24 hour period in the 72 hours before admission) with moderate dehydration or severe dehydration after correction by rapid IV fluids. Exclusion criteria: systemic infections requiring antibiotics, severe malnutrition (weight for age < 65% of NCHS standards), bloody diarrhoea. Number completing study: 61/61 (100%) in the probiotic group and 63/63 (100%) in the control group.
Interventions	L. casei subspecies rhamnosus 10 x10^9 CFU/day inulin 320mg/day Interventions started after correction of severe dehydration if required
Outcomes	Duration of diarrhoea (cessation of diarrhoea defined as passage or 2 formed or semi-formed stools or no stools for 24 hours). Note: SDs quoted for mean duration of diarrhoea in each group appeared small in comparison with other trials. Authors contacted and clarification awaited. Diarrhoea lasting 3 or more days Diarrhoea lasting 4 or more days. 24 hour and total stool output Unscheduled IV fluids Vomiting during first 24 hours after randomization Hyponatraemia at 24 hours after randomization No comment regarding adverse events.
Notes	Study location: Brazil (low child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded; 52% of children in the probiotic and 48% in the control group had rotavirus in stools; no data shown for outcomes in rotavirus diarrhoea although stated as "no significant difference" between groups. Nutritional status: severe malnutrition excluded; median WHZ score -1.13 (IQR −1.63 to −0.43) in control and -1.22 (−1.87 to −0.62) in probiotic group. Hydration status: all dehydrated; moderate or severe dehydration in 92% in the probiotic and 94% in the control group. Source of funding: the study was supported in part by a grant from Pronex/CNPq (661086/1998-4), Brazil.
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Randomization code
Allocation concealment?	No	Sequential administration
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with acute infectious diarrhoea who had failed oral rehydration. Exclusion criteria: bloody stools; coexisting disease that might influence the course of diarrhoea. Number completing study: 50/50 (100%) in the probiotic group and 50/50 (100%) in the control group.
Interventions	Live L. rhamnosus 50 ml/kg/day of ORF containing 5 x 10^12 organisms/200 mL Live L. rhamnosus (dose unclear) ORF Interventions started after rapid IV rehydration
Outcomes	Duration of treatment No. stools during the whole treatment period No. stools on a typical day of treatment No specific comment regarding adverse events.
Notes	Study location: Poland (low child, low adult mortality). Cause of diarrhoea: bloody diarrhoea excluded; 28/50 in the probiotic and 30/50 in the control group had rotavirus diarrhoea. Nutritional status: no data presented Hydration status: no data presented Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not stated
Allocation concealment?	Unclear	Not stated
Blinding? All outcomes	Unclear	Not stated
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; unclear whether single or multi-centre. Duration: not stated
Participants	Inclusion criteria: unclear whether inpatients or outpatients, or both; children with acute enteritis (duration and definition not given). Exclusion criteria: none stated. Number completing study: 21/21 (100%) in the probiotic group and 18/18 (100%) in the control group.
Interventions	Live Streptococcus faecium (S. faecium 68; 75 x10^6 bacteria thrice daily for 7 days) Placebo (details not given). When interventions started not stated.
Outcomes	Number of participants with < 2 stools/day. Formed, yellow/brown stools without mucus. No abdominal pains vomiting or fever for the whole day. No adverse events attributed to probiotic.
Notes	Study location: Switzerland (very low child and adult mortality). Cause of diarrhoea: 7 participants in each group had positive stool cultures for bacteria. Nutritional status: no data presented Hydration status: no data presented S. faecium 68 also appeared to promote recovery from abdominal pains, fever, and vomiting. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Unclear whether placebo identical to probiotic
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: February 2005 to February 2007
Participants	Inclusion criteria: inpatients; infants and children with watery diarrhoea (defined as watery stools) <72 hours duration due to rotavirus infection, parental consent. Exclusion criteria: systemic infection, chronic disease, body weight <60% NCHS standard, vomiting, need for antibiotics. Number completing study: 113/113 (100%) in the probiotic group and 111/111 (100%) in the control group. Six children did not complete the study; no group allocation or reasons given.
Interventions	L. acidophilus, L. paracasei, L. bulgaricus, L. plantarum, B. breve, B. infantis, B. longum, S. thermophilus (VSL#3; body weight < 5 kg: 180 billion organisms/day; body weight 5-10 kg: 360 x10^9 organisms/day for 4 days). Placebo (details not given although placed in identical sachets) When interventions started not stated.
Outcomes	Number stools/day; duration diarrhoea; IV fluid requirement; ORF requirement. No adverse effects attributed to probiotic.
Notes	Study location: India (high child and high adult mortality) Cause of diarrhoea: all rotavirus Nutritional status: severe malnutrition excluded; statement that "malnutrition status similar in two groups" Hydration status: dehydration status similar in two groups at baseline but no data presented. Source of funding: supported by grant from VSL
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	identical sachets
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 150 hospitals Duration not stated
Participants	Inclusion criteria: acute diarrhoeal disorder; diarrhoea defined as ≥ 3 not formed stools/day; duration not stated Exclusion criteria: not stated Number participants recruited at baseline not reported. 534 patients in the placebo group and 540 in the probiotic group completed the study.
Interventions	Enterococcus SF 68 (Bioflorin; 3 caps/day for 7 days) Placebo (not details given) When interventions started not stated.
Outcomes	Duration of diarrhoea (only statistical analysis reported; no raw data) No adverse effects attributed to probiotic.
Notes	Study location: Italy (very low child and adult mortality) Cause of diarrhoea: no data presented Nutritional status: no data presented Hydration status: no data presented Source of funding: not stated Probiotic also evaluated in antibiotic-associated diarrhoea
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Random allocation; no details reported
Allocation concealment?	Unclear	No details reported
Blinding? All outcomes	Unclear	No details regarding placebo reported.
Incomplete outcome data addressed? All outcomes	Unclear	Number participants recruited not reported

---

---

Methods	Randomized controlled trial; 1 centre Duration not stated
Participants	Inclusion criteria: inpatients; children with acute rotavirus diarrhoea Exclusion criteria: not stated Number completing study: overall, 64/70 (91.4%) completed study. Number in each intervention group not stated.
Interventions	ORF + S. boulardii ORF + L. acidophilus, L. rhamnosus, B. longum, S. boulardii ORF only When interventions started not stated.
Outcomes	Duration of diarrhoea Duration of fever Duration of vomiting Duration of hospitalization No comment regarding adverse events.
Notes	Study location: Chile (low child and adult mortality) Cause of diarrhoea: all rotavirus Nutritional status: no data presented Hydration status: no data presented Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Not described
Incomplete outcome data addressed? All outcomes	Unclear	Number children in each intervention group not stated

---

---

Methods	Randomized controlled trial; multi-centre Duration: 1 year, 1996
Participants	Inclusion criteria: inpatients and outpatients; infants and children with > 4 liquid or semi-liquid stools/day for 1 to 5 days. Exclusion criteria: previous probiotic usage; underlying chronic untreated small bowel disease; inflammatory bowel disease; any underlying chronic disease or immunosuppressive disease or treatment. Number completing study: 287 forms (269 participants) of total of 323 forms (88.9%) received at the coordinating centre were analysed (36 incomplete data or not compliant with protocol); unclear whether withdrawals occurred at participating centres.
Interventions	L. GG (ATC 53103, ≥10 x 10^9 CFU/250 ml) with ORF ORF with placebo Interventions added to ORF and started at recruitment.
Outcomes	Number of treatment failures (need for IV fluids) Mean duration of diarrhoea (time to last recorded fluid stool) Weight gain Proportion of children with diarrhoea longer than 7 days Mean stool frequency by day of treatment (SDs not given) Mean hospital stay Some outcomes also reported for rotavirus, bacterial, and no organism-isolated subgroups. No comment regarding adverse events.
Notes	Study locations: Poland (low child and adult mortality), Egypt (high child and high adult mortality), Croatia, Italy, Slovenia, The Netherlands, Greece, Israel, United Kingdom, Portugal (all very low child and very low adult mortality). Cause of diarrhoea: rotavirus (56 probiotic/45 placebo); bacteria (35/34); parasites (7/6); no pathogen (45/54). 10 (6.8) probiotic and 15 (10.7) control group had bloody diarrhoea. Nutritional status: no data presented. Hydration status: severe dehydration in 1 (0.7) probiotic and 1 (0.7) control group; mild/moderate dehydration in 107 (72.7%) probiotic and 96 (68.2%) control group. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Yes	Code broken at end of study
Blinding? All outcomes	Unclear	Identical placebo
Incomplete outcome data addressed? All outcomes	Unclear	Unclear whether withdrawals occurred at participating centres; also 36/323 (11.2%) participant data forms received at the co-ordinating centre were not analysed as incomplete and/or not compliant with protocol.

---

---

Methods	Randomized controlled trial; 1 centre Duration: 3 months, November 1995 to January 1996
Participants	Inclusion criteria: consecutive outpatients attending 3 family physicians; infants and children with ≥ 3 watery stools/day of < 48 hours duration. Exclusion criteria: antibiotic treatment in preceding 3 weeks, breastfeeding, and weight:height ratio < 5th percentile. Number completing study: 52/52 (100%) in probiotic group and 48/48 (100%) in control group.
Interventions	Lyophilized L. casei strain GG (Dicloflor 30; 6 x 10^9 million CFU/day for maximum 5 days) re-suspended in milk or formula feed ORF only Interventions started after 6 hours of ORF.
Outcomes	Mean duration of diarrhoea (time to last loose or liquid stool assessed by mothers) Results for rotavirus subgroup also presented. No comment regarding adverse events.
Notes	Study location: Italy (very low child and adult mortality). Cause of diarrhoea: Rotavirus identified in 30 (57.7%) probiotic and 31 (64.6%) control group. Nutritional status: weight:height ratio < 5th percentile excluded. Hydration status: all had mild to moderate dehydration. The study author clarified that Figure 1 in the published article reports the mean and standard error for the duration of diarrhoea; SDs derived from graph. We also extracted data from Canani 1997 (abstract), which also reports standard errors. Probiotic also reduced prevalence of rotavirus in stools on day 6. Source of funding: Ministero della Sanità, AIDS Project (9205.30)
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Random number table
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	No	Open study
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial - randomization according to odd and even participant numbers; three centres Duration: 2 months
Participants	Inclusion criteria: outpatients; children aged 6 months to 5 years with acute watery diarrhoea of mild or moderate severity (not defined), suitable for ambulatory treatment. Exclusion criteria: anti-diarrhoeals or antibiotics before admission, grade III malnutrition, bloody diarrhoea, needed IV rehydration, diarrhoea for >14 days. Number completing study: 51/54 (94%) probiotic group and 50/54 (93%) control group.
Interventions	Lyophilized S. boulardii (500 mg/day for 6 days) standard treatment (oral rehydration and feeds) Unclear whether researchers and participants able to distinguish between interventions.
Outcomes	Frequency and consistency (loose vs. formed) of stools Duration of illness (definition of end of diarrhoea not stated). Tolerance of treatment No adverse events attributed to probiotic.
Notes	Study location: Pakistan (high child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded; stool analysis not done. Nutritional status: grade III malnutrition excluded Hydration status: participants who needed IV rehydration excluded.. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	No	Alternate allocation
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Probably open study; no placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 11 centres Duration: 3 months, February to April 2005
Participants	Inclusion criteria: outpatients; infants and toddlers < 4 years with > 3 watery or loose and non-bloody stools /day for ≤ 3 days. Exclusion criteria: > 5% dehydration; intake of E. coli Nissle 1917 in last 3 months; intake of food supplements or drugs which contain living microorganisms or their metabolic products or components within 7 days prior to enrolment or during the trial; other antidiarrhoeal drugs; breast-feeding, premature birth; severe or chronic disease of the bowel or severe concomitant diseases. Antibiotics stated as exclusion criteria but some children included. Number completing study: 54/55 (98.2%) probiotic group and 45/58 (93.8%) control group. Reason for withdrawals in both groups stated as intervention no longer suitable or required other treatment.
Interventions	Live E. coli strain Nissle 1917 (Mutaflor suspension; 100-300 x10^6 organisms/day according to age) Placebo
Outcomes	Number of stools, stool consistency, admixture of blood or mucus Frequency of vomiting, abdominal pain and cramps Fluid intake, concomitant medication and general state of health for up to 10 days Diarrhea resolution: reduction in stool frequency to < 3 watery or loose stools in 24 hours over a period of at least 2 consecutive days. Adverse effects: 1 had rhinitis and 1 had abdominal cramps in the probiotic group. 2 had acute otitis media in the placebo group. 1 participant with poor compliance in the placebo group. No adverse events attributed to probiotic.
Notes	Study location: Ukraine, Russia (low child, high adult mortality) Cause of diarrhoea: bloody diarrhoea excluded; 16/55 (29.1%) probiotic and 19/58 (32.8%) control group had viral diarrhoea. Bacterial pathogens isolated from 9/55 (16.4%) probiotic and 4/58 (6.8%) control group. Nutritional status: most children well nourished. Hydration status: > 5% dehydration excluded; 0/55 probiotic and 1/58 control children had mild dehydration. Better outcomes in probiotic than placebo for abdominal pain (28/30 vs. 24/33) and abdominal cramps (17/18 vs. 21/26). Parents reported slightly better tolerance of probiotic than placebo, although investigators noted no difference. Authors supplied data regarding SDs for diarrhoea duration. Source of funding: ARDEYPHARM provided verum and placebo medications and reimbursed study-related expenses
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Computer-generated randomly permuted blocks of 4
Allocation concealment?	Yes	Sequence concealed from parents and researchers
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 11 centres Duration: 3 months, February to April 2005
Participants	Inclusion criteria: inpatients; > 3 loose or watery stools without blood / 24 hours for > 4 days and < 14 days; moderate dehydration (5-10% loss of body weight). Exclusion criteria: other severe organic or infectious disease; participation in another trial; intake of trial preparation in the past 3 months; intake of probiotic preparations within the past 7 days; antibiotics or antidiarrhoeals; severe dehydration (>10% weight loss); weight <5th percentile; growth faltering; breast-feeding; preterm birth. Number completing study: 72/75 (96.0%) probiotic group (trial intervention no longer suitable/different treatment needed - 2; personal reasons - 1); 59/76 (77.6%) control group (trial intervention no longer suitable/different treatment needed - 11; personal reasons - 5f; intolerable adverse event - 1).
Interventions	Escherichia coli strain Nissle 1917 (Mutaflor Suspension, Germany; participants received 100-300 x10^6 organisms/day according to age) Placebo - Identical suspension
Outcomes	Resolution of diarrhoea (<=3 watery or loose stools/24 hours for 4 consecutive days) Clinical improvement General state of health Adverse events Tolerance of intervention 1 participant in the probiotic group had a mild hypersensitivity reaction which was assessed as possibly related to the intervention. In the control group, 1 participant had vomiting, 1 abdominal pain, 1 dermatitis and 1 withdrawn because of influenza. Authors commented that the probiotic was safe and well tolerated.
Notes	Study location: Ukraine, Russia (low child, high adult mortality) Cause of diarrhoea: bloody diarrhoea excluded; 12 (16.0) probiotic and 15 (21.1) control group had viral diarrhoea. Bacterial pathogens isolated from 15 (20.0) probiotic and 19 (25.0) control group. Nutritional status: weight < 5th percentile and growth faltering excluded; 2 (2.7) probiotic and 3 (3.9) controls had mild/moderate malnutrition. Hydration status: all had moderate dehydration (5-10% loss of body weight). Fewer children with dehydration at the end of the study in the probiotic than the placebo group. General state of health improved to a greater extent in the probiotic than the placebo group. Significantly fewer children with diarrhoea > 21 days in the probiotic than the placebo group. At the end of the study the rates of mucus in stool, abdominal cramps, and abdominal pain were all lower in the probiotic group. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Computer-generated randomly permuted blocks of 4
Allocation concealment?	Yes	Study personnel and participants blinded to treatment assignment for the duration of the study
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	No	Follow up < 90% in placebo group

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; uncomplicated acute diarrhoea (not defined) and mild dehydration. Exclusion criteria: fever; malnutrition; bloody stools. Number completing study: 25/25 (100%) probiotic group; 25/25 (100%) control group.
Interventions	S. boulardii (200 mg every 8 hours for 5 days) Placebo
Outcomes	Stool frequency Persistence of diarrhoea No adverse events attributed to probiotic.
Notes	Study location: Mexico (low child and adult mortality) Cause of diarrhoea: bloody diarrhoea excluded Nutritional status: malnutrition (not defined) excluded. Hydration status: all had mild dehydration. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	No	Not described
Blinding? All outcomes	Unclear	Not described
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; multi-centre Duration: not stated
Participants	Inclusion criteria: outpatients attending general practitioners, gastroenterologists, and internal physicians; adults with acute diarrhoea (> 3 liquid stools in last 24 hours; in great majority duration 2 days or less; 1 participant in the placebo group had diarrhoea for >10 days). Exclusion criteria: chronic diarrhoea; blood in stools; drug-induced diarrhoea; antimicrobial treatment; inflammatory bowel disease. Number completing study: 92/107 (86.0%) randomized participants completed study (1 took additional drugs, 14 < 3 liquid stools at presentation). 3 participants dropped out (2 probiotic, 1 placebo) because intervention not effective; results included in analysis.
Interventions	S. boulardii (Perenterol; 600 mg/day for 2 days then 300 mg/day on days 3 to 7) Placebo Interventions started at presentation.
Outcomes	Mean stool frequency on days 1, 3, and 8 Score derived from stool frequency and consistency No adverse events attributed to probiotic.
Notes	Study location: Germany (very low child and adult mortality). Cause of diarrhoea: Stool analyses in first 50 participants only: 2 had rotavirus and 3 Salmonella Nutritional status: all well nourished. Hydration status: no data presented. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	identical placebo
Incomplete outcome data addressed? All outcomes	No	< 90% in probiotic and placebo groups

---

---

Methods	Randomized controlled trial - participants alternately assigned to the probiotic or control group on hospital admission; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children aged 3 months to 10 years; acute watery diarrhoea of duration < 7 days. Exclusion criteria: fever > 38°C; severely dehydrated; macroscopic blood in the stools; intake of antifungals; existing severe malnutrition. Number completing the study: 50 (100%) probiotic group, 50 (100%) control group.
Interventions	S. boulardii (500 mg/day for 5 days) ORF according to WHO protocol Interventions started on admission.
Outcomes	Mean duration of diarrhoea (diarrhoea resolution: <3 stools/day or solid stools only) Stool frequency Consistency of stools No adverse events attributed to probiotic.
Notes	Study location: Myanmar (high child and high adult mortality) Cause of diarrhoea: bloody diarrhoea excluded Nutritional status: severe malnutrition excluded, no other data presented Hydration status: severe dehydration excluded, no other data presented SDs for the duration of diarrhoea were not reported. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	No	Alternate allocation
Allocation concealment?	No	Alternate allocation
Blinding? All outcomes	No	Probably open study; no placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with > 3 watery stools/day for < 7 days and stools positive for rotavirus. Average dehydration about5% in both groups. Exclusion criteria: not stated. Number completing study: 21/21 (100%) in probiotic group and 21/21 (100%) in control group.
Interventions	Live L. casei strain GG (2 x 10^10 CFU/day for 5 days) No probiotic Interventions started after 6 hours ORF.
Outcomes	Mean weight gain Mean duration of diarrhoea (definition for recovery from diarrhoea not stated) Proportion of participants with diarrhoea by day of treatment No comment regarding adverse events.
Notes	Study location: Finland (very low child and adult mortality). Cause of diarrhoea: all rotavirus diarrhoea. Nutritional status: all well nourished Hydration status: mean dehydration about 5% in both groups. Source of funding: Academy of Finland and the Foundation for Nutrition Research (Finland).
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not defined
Allocation concealment?	Unclear	Not defined
Blinding? All outcomes	No	Open study; no placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 12 centres Duration: not stated
Participants	Inclusion criteria: inpatients and outpatients; age 1month to 3 years; acute diarrhoea (3 or more liquid stools in 12 hours or single liquid or semi-solid stool with mucus or blood, or both, for 5 days or less). Exclusion criteria: antibiotics or probiotics in last 5 days; chronic diseases of small or large intestine (eg coeliac, cow milk protein allergy, inflammatory bowel disease), immunosuppression, phenylketonuria Number completing study: 45/45 (100%) probiotic and 52/52 (100%) placebo
Interventions	Live L. GG ATCC 53103 (10^10 organisms in 250 mL ORF). ORF administered at 100 mL/kg over first 4 hours. Then either IV fluids or 10-15 mL/kg ORF per liquid/semi-solid stool. ORF with placebo. Start time for administration unclear.
Outcomes	Stool frequency, character Volume and length of use of ORF Duration of diarrhoea (until 2 consecutive normal stools) Use of antibiotics after recruitment No comment regarding adverse events.
Notes	Study location: Europe, Egypt, Africa, and single site (Montevideo) in S. America (variable child and adult mortality) Cause of diarrhoea: bacterial pathogens: probiotic group 29 (64.4%) and placebo group 37 (71.2%); rotavirus: probiotic group 18 (40.0%) and placebo group 21 (40.4%); parasites: probiotic group 2 (4.4%) and placebo group 4 (7.7%); no pathogens identified: probiotic group 11 (24.4%) and placebo group 14 (26.9%). Nutritional status: 15 (33.3%) in the probiotic and 20 (38.5%) in the control group had at least some malnutrition. Hydration status: mild/moderate dehydration in 15 (33.3%) probiotic and 17 (32.7%) control group. Severer dehydration in 0 in the probiotic and 2 (3.8%) control group. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	No	Alternate allocation
Allocation concealment?	No	Alternate allocation
Blinding? All outcomes	Unclear	Not described
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 19 months, April 2001 to September 2002
Participants	Inclusion criteria: inpatients; aged 6 months to 12 years with acute diarrhoea (not defined) Exclusion criteria: systemic infection; encephalopathy; convulsions; use of pharmaceutical probiotics Number completing study: 1/49 (2.0%) in the probiotic group and 3/53 (5.7%) controls left before the completion of the study.
Interventions	Tyndalized (heat-killed) Lactobacilus acidophilus (Lactrol, Raptakos; 15 x 10^9 bacteria/day for 3 days) Placebo (puffed rice powder) Interventions started on admission. All children received ORF, feeding and IV fluids if needed
Outcomes	Duration of diarrhoea (time to first of 3 consecutive semi-formed stools or to last loose stool before gap of no stools for 12 hours). SDs stated for mean duration of diarrhoea in each group appear to be too small, resulting in excessive weight in forest plots. SDs calculated from 95% CI stated in text. Treatment failure (diarrhoea persisting >72 hours, ORF >8L if < 5 years and > 10L if > 5 years, > 200mL/kg IV fluid required) Time to rehydration Duration of hospital stay Weight gain No comment regarding adverse events.
Notes	Study location: India (high child and adult mortality) Cause of diarrhoea: overall, 14/22 (63.6%) children tested were rotavirus positive and 8/98 (8.2%) has a positive culture for cholera. Nutritional status: most children were stunted and some had wasting. Hydration status: 19 in (39.6%) in the probiotic and 15 (30.0%) in the control group had severe dehydration. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Described as "simple randomisation done by a non-departmental colleague"
Allocation concealment?	Yes	Investigators blinded to group allocation
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow-up > 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 18 months, April 2006 to September 2007
Participants	Inclusion criteria: inpatients; infants and children aged 6 to 36 months with acute non-bloody, non-bacterial diarrhoea (not defined) of less than 2 days' duration and moderate dehydration Exclusion criteria: severe dehydration, antibiotic consumption, severe vomiting, convulsion, inflammatory cells in stool samples Number completing study: 32/34 (94.1%) probiotic and 30/34 (88.2%) placebo; participants excluded because of poor compliance.
Interventions	Live L. acidophilus 3 x 10^9 and Bifidobacterium bifidum 3 x 10^9 /day for 5 days (Infloran; Laboratorio Farmaceutico SIT S.r.I., Mede, Pavia, Italy) in 5–10 mL of water placebo (maltodextran) Start time for administration not stated. All children received IV fluid therapy, oral rehydration solution, and mother’s milk in breast-feeding infants, or complementary food according to the patient's age.
Outcomes	Duration of diarrhoea Reduction in defecation frequency Weight gain Duration of hospital admission No adverse events attributed to probiotic.
Notes	Study location: Iran (low child and adult mortality) Cause of diarrhoea: non-bloody, non-bacterial diarrhoea (not defined) Nutritional status: not stated. Hydration status: all had moderate dehydration; severe dehydration excluded. Source of funding: grant from the Vice Chancellery for Research, Mashad University of Medical Sciences.
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Random number table sequence
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	placebo sachets matched for size, shape, and volume of contents; physicians, nurses and parents were blinded to the treatment protocol.
Incomplete outcome data addressed? All outcomes	No	Follow up < 90% in placebo group

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: unclear whether inpatients or outpatients, or both; age < 24 months with acute rotavirus diarrhoea (> 3 loose or watery stools/24 hours lasting < 48 hours prior to inclusion). Exclusion criteria: not stated. Number completing study: 33/33 (100%) in probiotic group and 30/30 (100%) in control group.
Interventions	Live Bifidobacterium ruminatum (2 x 10^9 CFU/day for 5 days) Placebo Timing of administration not stated.
Outcomes	Duration of diarrhoea (definition for recovery from diarrhoea not stated.) Risk of diarrhoea lasting > 72 hours. No adverse events attributed to probiotic.
Notes	Study location: Poland (low child and adult mortality). Cause of diarrhoea: all rotavirus diarrhoea. Nutritional status: no data presented. Hydration status: dehydration status similar in both group; no other data presented. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 3 centres Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with 3 or more loose stools within 24h period of < 72 hours duration Exclusion criteria: history of acute diarrhoea within 14 days preceding the inclusion in the study; antibiotic treatment; received probiotic up to 7 days before the participation in the study; exclusively breast fed; chronic alimentary disease; diagnosis of malabsorption; lack of parental consent; lack of diarrhoea. Number completing study: 86/87 (98.9%) probiotic group and 87/89 (97.8%) placebo group.
Interventions	L. acidophilus, B. bifidum, L. bulgaricus (3.2 x 10^9 CFU/day for 5 days) identical placebo (no details given) Interventions administered from recruitment.
Outcomes	Duration of diarrhoea (defined as time to last loose stool) Duration of diarrhoea in rotavirus positive children Diarrhoea severity Vomiting Weight gain Duration of hospital stay Mean duration of diarrhoea reported for children with rotavirus diarrhoea. No adverse effects attributed to probiotic.
Notes	Study location: Poland (low child and adult mortality) Cause of diarrhoea: rotavirus identified in 31 (37.3%) probiotic and 22 (26.8%) placebo group. Bacterial pathogens identified in 6 (7.2%) probiotic and 14 (17%) placebo group. Nutritional status: no data presented. Hydration status: no data presented. Source of funding: interventions provided by Allergon, Sweden
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	No	Allocated according to order of presentation
Allocation concealment?	No	Allocated according to order of presentation
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; aged 3 months to 7 years with acute diarrhoea (liquid, mucous, or bloody stools passed at least twice as frequently as usual for ≥ 24 hours and < 7 days) Exclusion criteria: chronic disease; malnutrition; use of antibiotics, antidiarrhoeal or other drugs influencing gut motility Number completing study: probiotic group 100/115 (87.0%; 10 required antibiotics, 5 non-compliant); control group 100/117 (85.5%; 13 required antibiotics, 4 non-compliant)
Interventions	S. boulardii (250mg/d given with water or juice for 5 days) placebo (no details given) Interventions administered from admission. All children received ORF, normal food for age and IV fluids as required
Outcomes	Number stools/day and number watery stools/day Duration diarrhoea (time to first normal stool) Duration vomiting Duration fever Duration hospital stay 1 child had meteorism (group allocation unclear). No adverse events attributed to probiotic.
Notes	Study location: Turkey (low child and adult mortality). Cause of diarrhoea: 39 (39.0%) children in probiotic group and 44 (44.0%) controls had rotavirus diarrhoea. Overall, bacterial pathogens were isolated in 9 and parasites in 11 children. Nutritional status: malnutrition excluded; no other data presented. Hydration status: severe or moderate dehydration in 3 (3.0%) probiotic and 5 (5.0%) control group; mild/moderate dehydration in 17 (17.0%) probiotic and 24 (24.0%) control group. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	No	Follow up < 90% in both groups

---

---

Methods	Randomized controlled trial, non-blinded; 1 centre Duration: 6 months, October 1999 to March 2000
Participants	Inclusion criteria: inpatients; consecutive admissions aged 6-60 months; diarrhoea < 5 days and > 3 watery stools in last 24 hours. Average dehydration about 5% in both groups. Exclusion criteria: bloody stools, antidiarrhoeal or antiperistaltic drugs; children receiving lactose-free, protein hydrolysated formula for malabsorptive disorder; compromised immune system. Number completing study: 50/50 (100%) probiotic and 50/50 (100%) control group.
Interventions	Lyophilized L. acidophilus and Bifidobacteria infantis (Infloran Berna; 3 x 10^9 of each organism/day for 4 days) No additional treatment All children had IV fluids because of vomiting. Interventions administered after initial fluid therapy.
Outcomes	Stool frequency by day of intervention Duration of diarrhoea (time until the last watery stool) Recovery rate on day 2 No comment regarding adverse effects.
Notes	Study location: Taiwan (low child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded. Nutritional status: no data presented. Hydration status: % average dehydration 4.3 (SD 1.5) in probiotic and 4.0 (1.4) in control group. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Probably open study; no placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children aged ≤24 months; > 4 liquid stools/24 hours of < 72 hours duration. Exclusion criteria: rotavirus diarrhoea Number completing study: 42/42 (100%) probiotic and 38/38 (100%) control group.
Interventions	Heat-killed L. acidophilus strain LB (loading dose of 2 sachets, followed by 1 sachet every 12 hours. 1 sachet contained 10^10 CFU plus 160 mg of spent culture medium) Placebo sachet containing sucrose, ferrous oxides, silicic acid, and banana and orange flavouring All sachets diluted in ORF. Every admitted child was given at least 100 mL/kg of ORF.
Outcomes	Duration of diarrhoea (time to passage of first normal stool) Number whose diarrhoea stopped within 4 days. No adverse events attributed to probiotic.
Notes	Study location: Ecuador (high child and high adult mortality) Cause of diarrhoea: rotavirus diarrhoea excluded; bloody diarrhoea included. Nutritional status: no data presented Hydration status: severe dehydration in 0 probiotic and 1 (2.6%) control group; mild/moderate dehydration in 4 (10.5%) probiotic and 7 (23.3%) control group. Source of funding: Laboratoire du Lacte´ol (Houdan, France) provided strain LB and batches of lyophilized, heat-killed LB bacteria plus their culture medium to Dr Servin and Lactéol Fort sachets and placebo sachets to Dr Sarrazin-Davila.
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	No	Sequential allocation
Allocation concealment?	No	Sequential allocation
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with severe acute diarrhoea (defined as 1 watery or mucous stool or 3 or more loose stools daily for > 24 hours). Exclusion criteria: moderate or severe malnutrition; total or partial breast feeding; diarrhoea > 48 hours; need for antibiotic treatment; allergy to cow's milk; gastrointestinal or other chronic pathologies. 12/212 (5.7%; 3 study groups) withdrawn after recruitment as they did not match the age criteria. Number completing study: 70/70 (100%) probiotic and 71/71 (100%) control group.
Interventions	Live B. lactis Bb12 (10^9 CFU/g milk powder) and S. thermophilus TH4 (5 x 10^8 CFU/g milk powder) administered until 24 hours after diarrhoea ended Same probiotic preparation in a lower dose; not included in this review Milk-based, lactose-free formula Interventions administered after oral or parenteral rehydration.
Outcomes	Stool frequency and consistency daily until day 7 Diarrhoea duration (end of episodes defined as first formed stool if followed by 2 consecutive non-watery stools or 12 hours without evacuation) Failure of treatment No specific comment regarding adverse effects.
Notes	Study location: China; low child and adult mortality Cause of diarrhoea: rotavirus diarrhoea occurred in 87% and bacterial diarrhoea in 13% in both groups. Nutritional status: moderate or severe malnutrition excluded. Hydration status: no data presented. Source of funding:not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Reported as double blind but methods of blinding not described
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with diarrhoea (> 3 stools per day, watery or taking the shape of the container); duration not stated. Exclusion criteria: parents refused consent, children living outside the municipal area, bloody diarrhoea, severe dehydration, shock, inability to take and retain oral feeds, suspected systemic infection. Number completing study: 105/111 (94.6%) probiotic and 105/118 (89.0%) control group; children withdrawn as they did not complete allocated treatment.
Interventions	Live L. rhamnosus GG (1 x 10^6 - 10^9 bacteria/day; Culturelle; Amerifit Brands, Cromwell, CT, USAt) Identical placebo (crystalline micro cellulose) Start of interventions not stated
Outcomes	Duration of diarrhoea Number of stools on days 3, 6, and 10 Difference in number of stool in the same patient at presentation and on days 3, 6, and 10 Relative risk of diarrhoea continuing on day 3 No comment regarding adverse effects
Notes	Study location: India; high child and adult mortality Cause of diarrhoea: rotavirus identified in 29/105 (27.6%) probiotic and 25/105 (23.8%) in control group. Bloody diarrhoea excluded but 30/105 (28.6%) in probiotic and 30/105 (28.6%) in control group had white blood cells in stools and, overall, 10 children had bacterial diarrhoea. Nutritional status: no data presented. Hydration status: severe dehydration excluded; mild/moderate dehydration in 18 (17.1%) probiotic and 23 (21.9%) control group. Source of funding: partly by the International Development Fund of the John Nuveen Centre for International Affairs, University of Illinois, Chicago, USA
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Computer-generated randomization
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical capsules
Incomplete outcome data addressed? All outcomes	No	Follow up < 90% in placebo group

---

---

Methods	Randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with acute rotavirus diarrhoea (stool frequency and consistency not stated) of duration of ≤ 3 days. Exclusion criteria: infectious diarrhoea other than rotaviral diarrhoea;  serum sodium > 155 mmol/L or <130 mmol/L; history of malabsorption, respiratory or systemic infections Number completing study: 40/40 (100%) probiotic and 40/40 (100%) control group.
Interventions	Bifilac (species of bacteria not mentioned; information from manufacturers, Streptococcus faecalis T-110 30 million bacteria, Clostridium butyricum TO-A 2 million bacteria, Bacillus mesentericus TO-A 1 million bacteria, Lactobacillus sporogenes 50 million bacteria. Total of 249 x 10^6 bacteria/day for < 14 days). Placebo (no details given) When interventions started not stated
Outcomes	Frequency of diarrhoea Duration of diarrhoea Amount of IV fluid given Amount of ORF given Rotavirus shedding. No adverse effects attributed to the probiotic.
Notes	Study location: India; high child and adult mortality Cause of diarrhoea: all rotavirus diarrhoea. Nutritional status: no data presented. Hydration status: no data presented. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Reported as double blind but no details given
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 1 year, 16 January 1998 to 15 January 1999
Participants	Inclusion criteria: inpatients; infants and children with non-bloody diarrhoea (characteristics not stated) for < 5 days. Exclusion criteria: antibiotics in last 72 hours; antidiarrhoeal drugs; other illness; severe malnutrition; compromised immune system, severe electrolyte disturbance and dehydration. Number completing study: 47/47 (100%) in probiotic group and 47/47 (100%) in placebo group.
Interventions	Live L. acidophilius and L. bifidus (Infloran berna; 3 x 10^9 of each organism/day) Placebo When interventions started not stated.
Outcomes	Mean duration of diarrhoea (diarrhoea improved when no stool for 12 hours or 2 consecutive formed stools) Proportion of participants with diarrhoea by day of treatment Duration of hospital stay No adverse events attributed to probiotic.
Notes	Study location: Philippines (low child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded. Nutritional status: severe malnutrition excluded; no other data presented. Hydration status: dehydration excluded. Unpublished data. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Random number table
Allocation concealment?	Yes	Randomization by independent person
Blinding? All outcomes	Yes	Administration of interventions by independent person
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1centre Duration: 6 months, October 2004 to March 2005
Participants	Inclusion criteria: inpatients and outpatients; previously healthy children; aged 6 months to 10 years; acute diarrhoea (not defined). Exclusion criteria: severe systemic infection or sepsis; chronic disease; previous antibiotics; anti-diarrhoeal drugs; primary/secondary immune deficiency. Number completing study: 16/16 (100%) for the probiotic group. 11/11(100%) for the control group.
Interventions	S. boulardii (500 mg/day in 5 mL of water for 7 days) Placebo Start of intervention unclear.
Outcomes	Number, characteristics and frequency of stools; Blood tests (blood count and lymphocyte subsets, C-reactive protein, blood smear, complement, immunoglobulins and cytokines). No adverse events attributed to probiotic.
Notes	Study location: Turkey (low child, low adult mortality) Cause of diarrhoea: 1 (6.3%) child in probiotic and 0 in control group had bacterial diarrhoea. Nutritional status: mild/moderate malnutrition in 2 (12.5%) in the probiotic and 1 (9.1%) in the control group. Hydration status: severe dehydration in 1 (6.3%) in the probiotic and 0 in the control group; mild/moderate dehydration in 3 (18.8%) in the probiotic and 2 (18.2%) in the control group. Source of funding: Sanofi-Aventis (Paris, France) provided laboratory reagents and a commercial preparation of S. boulardii
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 6 weeks, July to mid-August 1993.
Participants	Inclusion criteria: inpatients; infants and children with > 3 watery stools in last 24 hours and diarrhoea for < 14 days. . Mean (SD) weight for age z score -1.15 (0.95) in the probiotic group and -1.8 (1.4) in the placebo group. Exclusion criteria: exclusive breast-feeding; septicaemia. Number completing study: 20/20 (100%) in probiotic group and 19/19 (100%) in placebo group. However, data extractable for subset with watery diarrhoea only: 14/20 (70%) in probiotic group and 12/19 (63.2%) in placebo group. No data for children with bloody stools presented.
Interventions	Live L. GG (10^9 - 10^10 CFU bd for 2 days) Placebo Interventions started after 6 hours ORF.
Outcomes	Mean duration of diarrhoea (time to last watery stool) Mean stool frequency on days 1 and 2 Vomiting occurred in 1 child in the placebo group. No adverse events attributed to probiotic.
Notes	Study location: Thailand (low child and adult mortality). Cause of diarrhoea: bloody stools in 6 children in probiotic and 7 in placebo group. All negative for parasites and cryptosporidium; 2 rotavirus and 1 astrovirus patients in the probiotic group and 5 rotavirus patients in the control group Nutritional status: no data presented Hydration status: severe dehydration in 2 (10%) in the probiotic and 4 (21%) in the control group; mild/moderate dehydration in all remaining children. Source of funding: Scientific Hospital Supplies, UK, provided the probiotic
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 4 months, September to December 2002
Participants	Inclusion criteria: inpatients; aged 6 to 24 months, breast fed with increased frequency, fluidity and volume of faeces of duration less than 4 days and moderate dehydration. Exclusion criteria: mucoid or bloody stools; oral feeding contra-indicated or intolerance; pneumonia; septicaemia; malnutrition; severe dehydration; stool culture positive for bacteria; recent intake of yogurt; poor compliance with yogurt intervention Number completing study: 3/43 (7.0%) withdrew from probiotic and 3/43 (7.0%) from control group all due to poor compliance with management
Interventions	Pasteurized cow's milk yogurt (L. bulgaricus 50,000 organisms/mL and S. thermophilus 50,000 organisms/mL; 15mL/kg/day yogurt or more) Control group received standard treatment Interventions administered from admission to discharge. All infants received ORF, IV fluids, complementary feeds
Outcomes	Duration of hospital admission Weight gain Reduction in diarrhoea frequency (communication from authors: achievement of previous defecation habit) Number of stools on days 2 and 3 of intervention No comment regarding adverse effects.
Notes	Study location: Pakistan (high child and adult mortality). Cause of diarrhoea: no data presented Nutritional status: malnutrition excluded. Hydration status: all had moderate dehydration. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	No	No placebo; probably open study
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 12 months; May 2005 to May 2006
Participants	Inclusion criteria: inpatients; infants and children with 3 or more watery stools/day for less than 48 hours and clinical dehydration. Exclusion criteria: bloody stools; hypovolaemic shock; acute systemic illness; antibiotic or anti-diarrhoeal medication. 18/178 children withdrawn mainly because of parent non-compliance; likely to have been withdrawn before recruitment. Number completing study: 40/40 (100%) in the probiotic group and 40/40 (100%) in the placebo group.
Interventions	Children randomized to one of 4 groups:  A, yogurt fermented with L. acidophilus, B, L. acidophilus supplement, C, conventional yogurt and D, placebo. Groups B and D selected for review. L. acidophilus (10 x 10^9 CFU/day; duration of treatment not stated; unclear if live or killed). Placebo (no details given) Start of administration not stated.
Outcomes	Weight change Duration of hospital stay Stool frequency on days 1, 2 and 3 Signs and symptoms on day 3 No adverse effects attributed to probiotic.
Notes	Study location: Iran; low child and adult mortality Cause of diarrhoea: bloody diarrhoea excluded; no bacteria or parasites identified in stool samples. Nutritional status: severe malnutrition excluded. Hydration status: severe dehydration in 1/40 (2.5%) probiotic and 2/40 (5%); all the rest had mild/moderate dehydration. Source of funding: supported by a grant from Tabriz Medical University
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Restricted randomization using random permuted blocks
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Not described
Incomplete outcome data addressed? All outcomes	Yes	Follow up ≥ 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 2 months, July and August 1993
Participants	Inclusion criteria: inpatients; undernourished infants and children with > 3 watery stools in last 24 hours for < 14 days duration and at least moderate dehydration. Exclusion criteria: severe malnutrition; septicaemia. Number completing study: 36/40 participants; 4 withdrawals (2 diagnosed with cholera, 1 developed pneumonia, 1 refused anything by mouth). Results presented for 19/21 (90.5%) in the probiotic group and 17/19 (89.5%) in the placebo group.
Interventions	Live L. GG (2 x 10^11-12 CFU/day for 2 days) Placebo Interventions started after 4 to 6 hours ORF.
Outcomes	Stool frequency on days 1 and 2. Frequency of vomiting on days 1 and 2. Weight gain. Outcomes for watery (non-bloody) diarrhoea also presented: mean (SD) stool frequency day 2 for probiotic (n = 16) versus placebo (n = 16) was 4.4 (2.0) versus 6.6 (4.2), P = < 0.05, and persistent diarrhoea at 48 hours in 5 (31%) versus 12 (75%) patients, P = < 0.01. Definition of persistent diarrhoea not stated. Less vomiting in the probiotic group; myoclonic jerks occurred in one child in each group. No adverse events attributed to probiotic.
Notes	Study location: Pakistan (high child and adult mortality). Cause of diarrhoea: bloody diarrhoea included. Nutritional status: all had mild/moderate malnutrition; severe malnutrition excluded. Hydration status: severe dehydration in 14 (66.7) probiotic and 7 (37) control group; all the rest had moderate dehydration. Duration of diarrhoea not measured (many children discharged before stool character had changed). Source of funding: Scientific Hospital Supplies, UK, provided the probiotic
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not defined
Allocation concealment?	Unclear	Not defined
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Approximately 90% follow up in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 21 months, June 2002 to March 2004.
Participants	Inclusion criteria: inpatients; Aboriginal children aged 4 months to 2 years with acute diarrhoea defined as ≥ 3 loose stools during 24 hours before presentation and duration < 7 days and able to tolerate ORF. Exclusion criteria: oxygen required during the study period; chronic cardiac, renal or respiratory disease; previous gastrointestinal surgery; proven sucrose intolerance; suspected on known immunodeficiency; received probiotic before enrolment; younger than 4 months. Number completing study: 201 assessed for eligibility; 103 refused participation and 28 failed to consent. Probiotic arm: 4 discharged before intervention, 1 parental withdrawal, 33 completed study. Control arm: 1 parental withdrawal, 31 completed study.
Interventions	Live L. casei strain GG (>15 x 10^9 CFU/day for 3 days) Identical placebo (no details given) Interventions administered within 24 hours of admission.
Outcomes	Small intestinal absorption capacity Diarrhoea duration (defined as time to last loose stool in which fewer than 3 loose stools occurred within a 24 hour period) Diarrhoeal frequency Total stool output Proportion of subjects with diarrhoea on days 3 and 4 Change in body weight on days 1 and 4 Total ORF and IV fluid required Safety and tolerability No adverse events attributed to probiotic.
Notes	Study location: Australia (very low child and low adult mortality). However, this study recruited Aboriginal children who commonly had co-morbidities such as pneumonia and malnutrition related to poverty and social disadvantage in the top end of the Northern Territory. Therefore, data not included in analysis according to country mortality strata. Cause of diarrhoea: bacterial pathogens identified in 4 (12%) probiotic and 4 (13%) in the control group; rotavirus identified in 11 (33%) in the probiotic and 6 (19%) in the control group; parasites identified in 2 (6%) probiotic and 2 (6%) control group. Nutritional status: mild/moderate malnutrition common amongst participants; no other data presented. Hydration status: severe dehydration in 0 probiotic and 1 (3.2%) in the control group; all the rest had mild/moderate dehydration. Source of funding: Project supported by Australian National Health and Medical Research Council
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Computer-generated block randomization
Allocation concealment?	Yes	Randomization by independent research institute; allocation concealed until recruitment, data collection and analyses were completed
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	No	Follow up < 90% in probiotic group

---

---

Methods	Randomized controlled trial; 2 centres Duration: 6 months, December 1998 to May 1999
Participants	Inclusion: inpatients; children aged 6 to 36 months with 2 or more consecutive loose stools in 24 hours and a duration no more than 7 days. Exclusion criteria: underlying chronic disease or antibiotics prescribed during the study period. Number completing study: 86 children enrolled, of whom 69 (80.2%) completed the study; exclusions were made after randomization because antibiotics were prescribed (3 in the control group and 2 in the probiotic group), rapid recovery before intervention started (3 in the control group and one in the probiotic group), non-compliance with the protocol (4 in the control group and 4 in the probiotic group).
Interventions	Live L. rhamnosus 19070-2 and L. reuteri DSM 12246 (2 x 10^9 CFU of each organism/day for 5 days) Identical placebo (skimmed milk powder and dextrose anhydrate) Interventions started as soon as possible after randomization and did not await rehydration.
Outcomes	Duration of diarrhoea (time from treatment start to appearance of first normal stool as recorded by parents). Persistence of diarrhoea at end of intervention (day 5). No adverse events attributed to probiotic.
Notes	Study location: Denmark (very low child and adult mortality). Cause of diarrhoea: overall, rotavirus was the only pathogen in 40 (58%) children; 6 children had rotavirus and a bacterial pathogen was identified; in addition, Campylobacter jejuni was isolated in 3 children and Salmonella typhimurium in 1 child. Nutritional status: no data presented. Hydration status: no severe dehydration; mild/moderate dehydration in 5 (16.7%) in the probiotic and 17 (43.6%) in the control group. The probiotics appeared to reduce significantly the duration of diarrhoea in children treated within 60 hours of the onset of diarrhoea. Hospital stay was shorter in the probiotic group than the controls (mean 1.6 (SD 1.0) versus 2.7 (SD 2.0) respectively; P = 0.02). The probiotics also appeared to reduce significantly the number of children excreting rotavirus in stools on day 5. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	No	Follow up < 90% in both groups

---

---

Methods	Randomized controlled trial; 19 day-care centres Duration: 6 months, December 1998 to May 1999
Participants	Inclusion criteria: outpatients; children aged 6 to 36 months with 2 or more consecutive loose stools in 24 hours as assessed by parents and with a duration no more than 7 days. Exclusion criteria: underlying chronic disease; antibiotics prescribed during study period. Number completing study: 50 children enrolled, of whom 43 (86%) participants completed the study. Exclusions were because of hospitalization with excessive vomiting and moderate dehydration (2 in the placebo group and 3 in the probiotic group), 1 because antibiotics were prescribed (placebo group), 1 non-compliant with protocol (placebo group).
Interventions	Live L.rhamnosus 19070-2 and L. reuteri DSM 12246 (2 x 10^9 CFU of each organism/day for 5 days) Identical placebo Interventions started as soon as possible after randomization.
Outcomes	Duration of diarrhoea (time from treatment start to appearance of first normal stool as recorded by parents). Persistence of diarrhoea at end of intervention (day 5). One participant in the probiotic group complained of constipation (no stools passed from day 3 for 10 days). No adverse events attributed to probiotic.
Notes	Study location: Denmark (very low child and adult mortality). Cause of diarrhoea: overall, rotavirus was the only pathogen in 25 children, 2 had rotavirus and a bacterial pathogen identified, 2 had an infection with C. jejuni and Salmonella typhimurium. Nutritional status: no data presented. Hydration status: mild/moderate dehydration in 3 patients (12.5%) in the probiotic and 4 (13.8%) in the control group; no severe dehydration. The probiotics appeared to reduce significantly the duration of diarrhoea in children treated within 60 hours of the onset of diarrhoea. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	No	Follow up < 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 23 months, February 2001 to December 2002
Participants	Inclusion criteria: inpatients; boys aged 4 to 24 months of age; acute watery diarrhoea (> 4 liquid stools during 24 hours) of < 48 hours duration. Exclusion criteria: severe malnutrition (< 65% weight for age by the standard of the National Centre for Health Statistics (NCHS)); systemic infection requiring antimicrobial therapy; bloody diarrhoea; spot sample of stool revealed V. cholerae by dark-field microscopy; antibiotic treatment in the preceding 2 weeks Number completing study: 112/115 (97.4%) in the probiotic group (3 withdrawn by parents) and 115/115 (100.0%) in the control group.
Interventions	Live Lactobacillus paracasei strain ST11 (10^10 CFU/day for 5 days) Placebo (whey-protein and skimmed-milk powder blend) Interventions started after enrolment. All children received ORF and continued feeding, including breast milk if breast fed.
Outcomes	Stool output and frequency Oral rehydration solution intake Daily excretion of rotavirus No comment regarding adverse outcomes.
Notes	Study location: Bangladesh (high child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded. Rotavirus detected in 78 (69.6%) in the probiotic and 73 (63.5%) in the placebo group; V. cholera detected in 14 (12.2%) in the probiotic and 16 (13.9%) in the placebo group. Nutritional status: severe malnutrition (weight < 65% weight for age of NCHS standard) excluded no further data presented. Hydration status: mild/moderate dehydration in 54 (47.0%) in the probiotic and 65 (56.5%) in the control group. Source of funding: Nestle Research provided L. paracasei. Research supported by the Swedish Agency for Research in Developing Countries , the Karolinska Institute (Stockholm, Sweden), and Nestlé Research Centre (Lausanne, Switzerland).
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Randomly permutated blocks
Allocation concealment?	Yes	Randomization code generated by an independent statistician
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up >90% in both groups

---

---

Methods	Randomized trial; 1 centre Duration: 1 year; November 1992 to October 1993
Participants	Inclusion criteria: inpatients; infants and children with ≥2 loose stools for 1 to 3 days or haemorrhagic colitis, fever ≥ 38.0^oC,or second stage dehydration, or both. All had shigellosis. Exclusion criteria: not stated Number completing study: 13/13 (100%) children in the probiotic group and 12/12 (100%) in the control group.
Interventions	L. casei strain GG (10^10-11 CFU/day for either 5 or 10 days) + trimethoprim-sulfamethoxazole (36 mg/kg/day for 5 days) trimethoprim-sulfamethoxazole (36 mg/kg/day for 5 days) When interventions started was not stated.
Outcomes	Number cured (defined as < 2 loose stools/24 hours without additional clinical symptoms for at least 3 days) Duration of diarrhoea Duration of hospital stay No comment regarding adverse effects
Notes	Study location: Estonia (low child, high adult mortality) Cause of diarrhoea: all shigellosis. 9 (69.2%) in the probiotic and 4 (33.3%) in the controls had bloody diarrhoea. Nutritional status: no data presented. Hydration status: no data presented. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	No	No placebo; probably open study
Incomplete outcome data addressed? All outcomes	Yes	Follow up > 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 1 year, 1 April 1994 to 31 May 1995
Participants	Inclusion criteria: inpatients; infants and children with ≥ 1 watery stool in the last 24 hours and diarrhoea for < 5 days. Exclusion criteria: not stated. Number completing study: 123/214 (57%) eligible children admitted during the study period enrolled; no reasons given for those not recruited. A total of 59/59 (100%) children allocated to the probiotic group and 64/64 (100%) in the placebo group completed the trial.
Interventions	Live L. strain GG (American-type culture collection 53 103; 10^10 CFU/day as a dried powder for 5 days) Placebo Interventions started with oral rehydration solution. All participants with positive stool cultures received antibiotics. Effect of isotonic versus hypotonic oral rehydration solution also assessed.
Outcomes	Duration of diarrhoea (defined as last appearance of watery stools) Weight gain Duration of hospital stay No comment regarding adverse events.
Notes	Study location: Russia (low child and high adult mortality). Cause of diarrhoea: rotavirus identified in 13 (22.0%) in the probiotic and 21 (32.8%) in the control group. Bacterial diarrhoea identified in 11 (18.7%) in the probiotic and 15 (23.4%) in the control group. Nutritional status: no data presented. Hydration status: mean dehydration ˜4% in both groups. Among children with rotavirus diarrhoea, the probiotic (n = 13) reduced the number of watery stools compared with the placebo (n = 21; P = 0.02, but no data given). No beneficial effect of the probiotic was seen in those with bacterial diarrhoea (probiotic (n = 11) and placebo (n = 115), P = 0.42). Stool samples tested for rotavirus (Rotazyme, Dakopotts AS, Denmark) and cultured for Salmonella and Shigella. Source of funding: Leiras, Turku, Finland and Valio, Helsinki, Finland
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Randomization schedule
Allocation concealment?	Yes	Randomization numbers sequentially assigned to participants as enrolled
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up > 90% in both groups

---

---

Methods	Randomized controlled trial; 2 centres Duration: 6 months, 22 January to 15 July 1996
Participants	Inclusion criteria: inpatients; infants and children with ≥ 3 watery stools in last 24 hours, diarrhoea for < 7 days; stools positive for rotavirus antigen (IDEIA Rotavirus, UK). Mean dehydration about 4% in both groups. Exclusion criteria: 20 participants who received exclusively or mainly IV fluids were excluded. 86/97 (89%) enrolled participants were positive for rotavirus. Number completing study: 21/21 (100%) in the probiotics and 25/25 (100%) in the placebo group. (20 allocated to a low dose probiotic group).
Interventions	Participants randomized to one of 3 groups: 20 in the probiotic small dose (10^7 CFU/day) group, 21 in the probiotic large dose group, 25 in the placebo group. Data from the large dose group were used in this review. Live L. reuteri (10^10-10^11 CFU/day for maximum 5 days) Live L. reuteri (10^7 CFU/day for maximum 5 days) Placebo Interventions started with ORF
Outcomes	Duration of diarrhoea (time to last watery stool in a 24 hour period with no watery stools) Stool frequency on day 2 of treatment Weight gain No comment regarding adverse events.
Notes	Study location: Finland (very low child and adult mortality). Cause of diarrhoea: all rotavirus. Nutritional status: no data presented. Hydration status: mean dehydration about4% in both groups. Data from high dose probiotic group used for continuous outcomes. Duration of diarrhoea before admission greater in probiotic group (4.2 (SD 1.4) days) than in the placebo group (2.9 (SD 1.2) days). Number with persistent diarrhoea on day 3 derived from graph. Source of funding: BioGaia Biologicals AB
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	No	Participants receiving IV fluids excluded

---

---

Methods	Randomized controlled trial; 1 centre Duration: 5 months, 29 January to 3 July , 1995
Participants	Inclusion criteria: inpatients; infants and children with ≥ 3 watery stools in last 24 hours; diarrhoea for < 7 days; ingested bovine dairy products. Exclusion criteria: immunosuppressive therapy or immune deficiency; allergy to bovine milk; serious underlying disorder; undergoing an investigational product during the preceding month. Number completing study: 41 participants initially enrolled; 19/19 (100%) in the probiotic group and 21/22 (95.5%) in the placebo group (1 participant in the placebo group removed because the probiotic agent (L. reuteri) was detected in stool; the probiotic was administered to his sibling).
Interventions	Live L. reuteri SD 2112 (10^10-10^11 CFU/day for a maximum of 5 days) Placebo Interventions started at recruitment.
Outcomes	Weight gain Duration of diarrhoea (last appearance of watery stools) Number of participants with watery diarrhoea according to day of treatment Stool frequency on days 2 and 3 Number of participants with vomiting according to day of treatment Less vomiting in the probiotic group. No adverse events attributed to probiotic.
Notes	Study location: Finland (very low child and adult mortality). Cause of diarrhoea: rotavirus identified in 18 (86%) in the probiotic group and 12 (63%) in the control group. Nutritional status: no data presented. Hydration status: mean dehydration at baseline greater in the probiotic (3.9% (SD 1.3)) than the control group (3.0 (SD 1.2)). Source of funding: BioGaia Biologicals, Inc., Raleigh,NC, USA.
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Randomiation schedule
Allocation concealment?	Yes	Randomization numbers sequentially assigned to participants as enrolled
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up > 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 1 year, September 1995 to August 1996
Participants	Inclusion criteria: inpatients; infants and children with acute, watery diarrhoea (stool frequency not stated) for ≤ 5 days. Exclusion criteria: mucous bloody stools or major systemic illness. Number completing study: 37/37 (100%) in the probiotic group and 36/36 (100%) in the placebo group.
Interventions	Heat-killed L. acidophilus LB (MA65/4E; Lacteol Fort sachets, Laboratoire du Lacteol du Docteur Boucard, Houdan, France; 2 x10^10 organisms and fermented culture medium 5 doses over 48 hours) Placebo Interventions mixed with 5 mL water and started with ORF.
Outcomes	Duration of diarrhoea (2 consecutive well formed stools or no stool passed for 12 hours) Recovery from diarrhoea by day of treatment Recovery from diarrhoea at 24 hours in rotavirus positive cases No comment regarding adverse events.
Notes	Study location: Thailand (low child and adult mortality). Cause of diarrhoea: bloody diarrhoea excluded. Nutritional status: severe malnutrition in 1 (2.7%) probiotic participant and 1 (2.8%) in the control group; mild/moderate malnutrition in 8 (21.6%) in the probiotic and 12 (33.3%) in the control group. Hydration status: no severe dehydration; all had mild/moderate dehydration. 40 children (17 probiotic and 23 placebo) had received antibiotics before admission. The effect of the probiotic in shortening the duration of diarrhoea more marked in children who had not received antibiotics before admission. Source of funding: Merck Ltd., Bangkok, Thailand. National Collection of Pasteur Institute provided the probiotic.
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Randomization code
Allocation concealment?	Yes	Numerically coded packages
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow-up >90% in both groups

---

---

Methods	Quasi-randomized controlled trial; 1 centre Duration: not stated
Participants	Inclusion criteria: inpatients; infants and children with acute rotavirus diarrhoea (stool characteristics described for each participant; stool frequency x 1-10/day; duration not stated); none had bloody stools. Exclusion criteria: none stated. Number completing study: 16/17 (94.1%) in the probiotic group and 11/15 (73.3%) in the control group.
Interventions	Live L. casei (1.5 g/day for up to 3 weeks) No additional treatment Not stated when interventions started. All participants received lactase (1.5 g/day in 3 doses) and albumin tannate (0.1/kg/day in 3 doses).
Outcomes	Efficacy, as judged by a clinician Time to first formed stool Average stool frequency before and after treatment Persistence of stool rotavirus antigen 1 week after intervention No adverse events attributed to probiotic.
Notes	Study location: Japan (very low child and adult mortality). Cause of diarrhoea: all rotavirus diarrhoea. Nutritional status: no data presented. Hydration status: no data presented. Results for time to first formed stool given for 16/17 (94.1%) participants in the probiotic group and 11/15 (73.3%) in the control group. Reasons for missing data not stated. Rotavirus antigen persisted in the stools of 1/9 (11.1%) children in the probiotic group and 2/8 (25%) in the control group. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	No	Allocation in order of hospitalization
Allocation concealment?	No	Allocation in order of hospitalization
Blinding? All outcomes	No	Open study
Incomplete outcome data addressed? All outcomes	No	Overall < 90% follow up in placebo group

---

---

Methods	Randomized controlled trial; 1 centre Duration: 10 months, September 2003 to June 2004
Participants	Inclusion criteria: inpatients and outpatients; aged 2 months to 6 years with acute diarrhoea (3 or more stools/day looser than normal that may contain blood or mucous for > 1 and < 5 days). Exclusion criteria: organic gut disease; underlying chronic illness; immunosuppression, exclusive breast-feeding Number completing study: 46/49 (93.9%) in probiotic group; 41/44 (93.2) controls. Withdrawals stated to be due to non-compliance or incomplete data.
Interventions	3 live strains of L. rhamnosus 573L/1, 573L/2, 573L/3 (2.4 x 10^10 CFU/day; Lakcid L, Biomed, Lublin, Poland) given orally in glucose Identical placebo Onset of intervention delayed >72 hours in many participants.
Outcomes	Duration of diarrhoea (either no abnormal movement for the last 12 hours or the time to the second normal stool) Weight gain after rehydration Number of stools/day Duration of IV fluids Number diarrhoea >7 days gGut colonization with probiotics No adverse events attributed to probiotic.
Notes	Study location: Poland (low child and adult mortality) Cause of diarrhoea: bloody diarrhoea included. Overall, 39/87 (45%) had rotavirus (22 probiotic and 17 control group), 5/87 (6%) had adenovirus, 9/87 (10%) had a bacterial pathogen and many children had norovirus infection. Nutritional status: no data presented. Hydration status: no severe dehydration. Mild/moderate dehydration in 34 (73.9%) in the probiotic and 31 (75.6%) in the control group. Source of funding: Wellcome Travel Award
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Computer-generated block randomization
Allocation concealment?	Yes	Sequential assignment of randomization numbers
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up > 90% in both groups

---

---

Methods	Randomized, single blind controlled trial; 1 centre Duration: 7 months; August 2007 to February 2008
Participants	Inclusion criteria: inpatients; infants and children with a history of acute watery diarrhoea (defined as ≥3 stools of liquid consistency/day < 72 hours duration) positive for rotavirus and with moderate to severe dehydration. Exclusion criteria: severe malnutrition; systemic infections requiring antibiotic therapy; severe chronic disease; identification of a second pathogen in the stool; ingestion of antibiotics; probiotics or nitazoxanide 3 weeks before admission; recurrence of diarrhoea after discharge. Patients in whom a cause of diarrhoea other than rotavirus were withdrawn (probiotic group: 3 with adenovirus, 2 with E. histolytica; control group: 3 with E. histolytica, 2 with Giardia, 1 with S. flexneri). Number completing study: 25/25 (100%) probiotic group; 25/25 (100%) control group.
Interventions	Participants were allocated to one of three groups: a nitazoxanide, a probiotic and a control group that received rehydration solutions only. Data from the probiotic group and controls used for this review. L. acidophilus, L. rhamnosus, B.longum, S. boulardii (total of 2.5 x 10^9 organisms/day administered for an average of 4.2 days). Unclear if they were live or killed organisms. Control (ORF only) Time when interventions started not described.
Outcomes	Duration of diarrhoea (time from admission until the presence of the first soft stool for at least 24 hours) Stool number and consistency Duration of fever Vomiting Duration of hospitalization No adverse events attributed to probiotic.
Notes	Study location: Bolivia (high child and high adult mortality) Cause of diarrhoea: all rotavirus diarrhoea. Nutritional status: severe malnutrition excluded; mild/moderate malnutrition in 5 (20%) in the probiotic and 15 (60%) in the control group. Hydration status: all had moderate to severe dehydration; no further data presented. Source of funding: the research was not sponsored by any pharmaceutical company
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Computerised admissions list
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	No	Single blind; only parents/caretakers not aware of group allocation. No placebo.
Incomplete outcome data addressed? All outcomes	Yes	Follow up > 90% in both groups

---

---

Methods	Randomized trial; 1 centre Duration: 1994-1995
Participants	Inclusion criteria: inpatients; infants and children Exclusion criteria: nosocomial rotaviral infection Number completing the study: 50/50 (100%) probiotic group; 50/50 (100%) control group.
Interventions	Live L. acidophilus ND (4 x 10^9 bacteria/day; duration not stated) Standard treatment Time when interventions started:
Outcomes	Average number of stools/day stool consistency at 5 days No adverse events attributed to the probiotic.
Notes	Study location: Czech Reublic (very low child and adult mortality) Cause of diarrhoea: nosocomial rotavirus diarrhoea excluded; 16 (32.0%) probiotic and 21 (42.0%) control group had viral diarrhoea. A total of 22 (44.0%) in the probiotic and 24 (48.0%) in the control group had bacterial diarrhoea. Nutritional status: no data presented. Hydration status: severe dehydration in 3 (6.0%) probiotic and 2 (4.0%) in the control group; all the rest had mild/moderate dehydration. No data presented that could be extracted for meta-analysis. Source of funding not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	No	No placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up > 90% in both groups

---

---

Methods	Randomized controlled trial; 1 centre Duration: 1 year, June 200 to May 2001
Participants	Inclusion criteria: consecutive inpatients aged 2 to 29 months with acute, non-bacterial diarrhoea (definition not stated) lasting >48 hours able to receive oral medication. Exclusion criteria: concomitant illness, antimicrobial, antidiarrhoeal or other drugs affecting gut motility, severe electrolyte disturbance or dehydration. Number completing study: 50 cases reported in both arms; number withdrawn because of the deterioration in diarrhoea, concomitant disease requiring other drugs unclear.
Interventions	Lyophilized Saccharomyces cerevisiae Hansen CBS 5926 (250 mg daily in 5mL cold liquid) 250 mg glucose daily in 5mL cold liquid Time of starting interventions and duration of administration not stated.
Outcomes	Stool frequency and consistency at 48 and 96 hours. No adverse events attributed to probiotic.
Notes	Study location: Turkey (low child and adult mortality) Cause of diarrhoea: non-bacterial diarrhoea Nutritional status: no data presented. Hydration status: none dehydrated. Lactose intolerance identified in 8% in the probiotic and 26% in the placebo group. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Unclear	Unclear if placebo identical or not
Incomplete outcome data addressed? All outcomes	Unclear	Number of children withdrawn not stated

---

---

Methods	Randomized controlled trial; 1 centre Duration: 1 year
Participants	Inclusion criteria: outpatients; infants and children aged 3 months to 2 years (urban population, middle social class); acute, mild to moderate diarrhoea. Exclusion criteria: use of probiotic in the preceding 7 days; chronic intestinal disease; short bowel syndrome; malabsorption; ≥ grade 2 malnutrition; severe disease (including dehydration requiring hospitalization at the time of the consultation); known immunodeficiency; immunosuppressant treatment (oral or IV corticoids in the preceding 7 days); oral nystatin; oral or parenteral imidazoles; other systemic antifungal agents; macrolides; drugs that alter intestinal motility (antispasmodics, cisapride, antiemetics and antidiarrhoeal drugs) in the preceding 7 days. Number completing study: 6/50 (12.0%) excluded from the probiotic group and 6/50 (12.0%) from the control group for lack of compliance with protocol medication.
Interventions	S. boulardii (250-500 mg twice daily. according to age for 6 days) Placebo ORF and antibiotics given when indicated.
Outcomes	Number of stools on day 4 and 7 Number participants with diarrhoea >7 days Number of participants with liquid stools on days 4 and 7 Duration of diarrhoea (time to stool frequency < 3/day or stool consistency improved for at least 24 hours) Effect when treatment was started within 48 hours after the onset of the diarrhoea No comment regarding adverse events.
Notes	Study location: Argentina (low child and adult mortality) Cause of diarrhoea: none had bloody diarrhoea; no other data presented. Nutritional status: ≥ grade 2 malnutrition excluded. Hydration status: dehydration requiring hospitalisation excluded; all had dehydration <7%. Stool frequency significantly lower in probiotic than placebo group on days 4 and 7. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Yes	Random computer-generated into blocks of 20
Allocation concealment?	Yes	Paediatricians recruiting patients received batches of coded containers
Blinding? All outcomes	Yes	Identical placebo
Incomplete outcome data addressed? All outcomes	No	Followup < 90% in both groups

---

---

Methods	Randomized open study; 1 centre Duration: March 2003 to January 2004; 11 months
Participants	Inclusion criteria: inpatients and outpatients; infants and children with watery diarrhoea (not defined) for < 5 days. Exclusion criteria: immunocompromised; suspected dysentery; diagnosed with persistent or chronic diarrhoea; chronic cardiac, pulmonary or haematological illness; undergoing antibiotic treatment in the last 2 weeks; severe dehydration or shock. 4/75 withdrawn (1 febrile seizure, 1 urinary tract infection, 2 with pneumonia). Two patients were withdrawn from each group. Number completing study: 36/38 (94.7%) in the probiotic group; 35/37 (94.6%) in the control group.
Interventions	Live L. acidophilus, Bifidobacterium infantis (Infloran; 6 x 10^9 CFU/day for 2 days) Control group did not receive probiotic Timing of administration not stated
Outcomes	Duration of diarrhoea Weight change/day Number bowel motions on day 2 Vomiting Duration of hospitalization Duration of diarrhoea reported for rotavirus diarrhoea. No adverse events attributed to probiotic.
Notes	Location: Thailand; low child and adult mortality Cause of diarrhoea: suspected dysentery excluded; overall, 34% had rotavirus and 12.1% Salmonella in stools. Nutritional status: no data presented. Hydration status: severe dehydration excluded; mild/moderate dehydration in 25 (69.4%) probiotic and 23 (65.7%) control group. Source of funding: AIS donation fund, Thai Red Cross Society
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	No	Open study; no placebo
Incomplete outcome data addressed? All outcomes	Yes	Follow up > 90% in both groups

---

---

Methods	Randomized controlled trial; 10 centres Duration: not stated
Participants	Inclusion criteria: adults with acute diarrhoea (characteristics and duration not stated). Exclusion criteria: not stated. 3 participants from each group withdrawn on day 4 or later (causes for dropouts stated to be unrelated to medication); 4 participants assigned to the probiotic group and 5 assigned to the placebo group did not complete the study (reasons not stated). Number completing study (for persisting diarrhoea outcomes): 40/47 (85.1%) in the probiotic group and 38/46 (82.6%) in the placebo group.
Interventions	Live Enterococcus SF 68 (Bioflorin; 225 x 10^6 bacteria/day for 7 days) Placebo Not stated when interventions started.
Outcomes	Number of cases cured by day of treatment (definition of cure not stated). No adverse events attributed to probiotic.
Notes	Study location: Switzerland and Lichtenstein (very low child and adult mortality). Cause of diarrhoea: no data presented. Nutritional status: no data presented. Hydration status: no data presented. Source of funding: not stated
Risk of bias
Item	Authors' judgement	Description
Adequate sequence generation?	Unclear	Not described
Allocation concealment?	Unclear	Not described
Blinding? All outcomes	Yes	identical placebo
Incomplete outcome data addressed? All outcomes	No	Follow up <90% in both groups

---

 

Characteristics of excluded studies [ordered by study ID]

---

Study	Reason for exclusion
Agarwal 2001	No non-probiotic group. Participants included in Agarwal 2002 study
Agarwal 2002	No non-probiotic group
Alexander 1971	Not a randomized controlled trial; no non-probiotic group
Alvisi 1982	Intervention groups not treated equally; antibiotics given to the non-probiotic group
Barone 2000	No non-probiotic group
Beck 1961	Not a randomized controlled trial
Bellomo 1979	Cause of diarrhoea unclear. Additional treatment given to children with persisting diarrhoea
Bellomo 1980	No non-probiotic group. Study included children with diarrhoea secondary to antibiotic treatment or associated with respiratory infection
Bellomo 1982	Cause of diarrhoea unclear
Bin Li Xie 1995	Intervention groups not treated equally; antibacterials given to the non-probiotic group
Brewster 2004	Secondary publication to Ritchie 2010
Camarri 1981	Intervention groups not treated equally; antibiotics given to the non-probiotic group
Cetina Sauri 1990	Secondary publication to Cetina-Sauri 1994
Chandra 2002	Prevention of rotavirus diarrhoea study
Chicoine 1973	Unclear if acute diarrhoea
Costa-Ribeiro 2000a	Unclear whether a randomized controlled trial
Costa-Ribeiro 2000b	Prevention of diarrhoea study
Cui 2004	No non-probiotic group
de dios Pozo-O 1978	Assessment of probiotic in the prevention of traveller's diarrhoea
Eren 2010	No non-probiotic group
Fang 2009	Study of effect of probiotic on rotavirus shedding in stools; no diarrhoea outcomes reported
Fourrier 1968	No non-probiotic group
Girola 1995	Children with gastroenteritis and antibiotic-associated diarrhoea studied together
Gracheva 1996	No non-probiotic group
Henker 2007b	Secondary reference to Henker 2007a
Heydarian 2010	No non-probiotic group
Isolauri 1991	No non-probiotic group
Kaila 1992	No non-probiotic group
Kaila 1995	No non-probiotic group
Korviakova 2000	Not a randomized controlled trial; probiotic versus antibiotic
Le Leyur 2010	Intervention group received an adapted lactose-free formula fortified with S. boulardii and control group received a standard formula; difference in diarrhoea outcomes between groups cannot be attributed to the probiotic
Lei 2006	Probiotic used was not specified
Lin 2009	Prevention study
Magreiter 2006	No non-probiotic control group
Majamaa 1995	No non-probiotic group
Mazo 2006	Prevention study
Michielutti 1995	Not a randomized controlled trial
Mitra 1990	No non-probiotic group
Moraes 2001	No non-probiotic group
Niv 1963	Not a randomized controlled trial; some children with diarrhoea thought to be caused by antibiotic treatment included
Ortlieb 1974	Participants with acute diarrhoea and antibiotic-associated diarrhoea combined
Pearce 1974	Intervention groups not treated equally; calcium carbonate given as the placebo and may have reduced diarrhoea in the non-probiotic group
Pedone 1999	Prevention of diarrhoea study
Pedone 2000	Prevention of diarrhoea study
Pene 1966	No non-probiotic group; participants with diarrhoea of various causes (infectious, post-antibiotics) grouped together
Rafeey 2008b	Secondary publication to Rafeey 2008a
Rautanen 1998	No data presented for placebo group
Saint-Marc 1991	Not a randomized controlled trial; no non-probiotic group
Salazar-Lindo 2004	Mean duration of diarrhoea reported from responders only; children with ongoing diarrhoea excluded from analysis
Salazar-Lindo 2007	Active placebo
Satoh 1984	Not a randomized controlled trial; no non-probiotic group
Savas-Erdeve 2009	Study of amoebiasis-associated diarrhoea and not acute infectious diarrhoea
Schrezenmeir 2004	Antibiotic-associated diarrhoea included in the study
Singh 1987	No probiotic specified
Sudarmo 2003	Other than the probiotic, unclear whether two intervention groups were treated the same. Probiotic group received high-lactose formula containing B. bifidum. Unclear whether control group received high-lactose or normal formula
Szymanski 2005	Preliminary publication of Szymanski 2006
Tojo 1987	Unclear whether diarrhoea acute and whether a randomized controlled trial

---

 

Characteristics of studies awaiting assessment [ordered by study ID]

---

Methods
Participants
Interventions
Outcomes
Notes	No details of study available

---

---

Methods
Participants
Interventions	Heat-killed L. acidophilus, Lacteol strain
Outcomes
Notes	No other details available

---

 

Comparison 1. Primary diarrhoea outcomes

---

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mean duration of diarrhoea	35	4555	Mean Difference (IV, Random, 95% CI)	-24.76 [-33.61, -15.91]
2 Diarrhoea lasting ≥4 days	29	2853	Risk Ratio (M-H, Random, 95% CI)	0.41 [0.32, 0.53]
3 Mean stool frequency on day 2	20	2751	Mean Difference (IV, Random, 95% CI)	-0.80 [-1.14, -0.45]

---

 

Comparison 2. Secondary diarrhoea outcomes

---

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Diarrhoea lasting ≥3 days	30	3022	Risk Ratio (M-H, Random, 95% CI)	0.62 [0.56, 0.70]
2 Mean stool frequency on day 3	14	2367	Mean Difference (IV, Random, 95% CI)	-0.63 [-1.18, -0.07]

---

 

Comparison 3. Strain of probiotic organisms

---

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

---

 1 Mean duration of diarrhoea

11

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

1.1 Live Lactobacillus casei strain GG

11

2072

Mean Difference (IV, Random, 95% CI)

-26.69 [-40.50, -12.88]

---

 2 Diarrhoea lasting ≥4 days

14

Risk Ratio (M-H, Random, 95% CI)

Subtotals only

---

2.1 Live Lactobacillus casei strain GG

4

572

Risk Ratio (M-H, Random, 95% CI)

0.59 [0.40, 0.87]

---

2.2 LIve Enterococcus LAB SF68

4

333

Risk Ratio (M-H, Random, 95% CI)

0.21 [0.08, 0.52]

---

2.3 Saccharomyces boulardii

6

606

Risk Ratio (M-H, Random, 95% CI)

0.37 [0.21, 0.65]

---

 3 Mean stool frequency on day 2

6

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

3.1 Live Lactobacillus casei strain GG

6

1335

Mean Difference (IV, Random, 95% CI)

-0.76 [-1.32, -0.20]

---

 

Comparison 4. Single organism versus combinations

---

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

---

 1 Mean duration of diarrhoea

35

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

1.1 Single organism

22

3196

Mean Difference (IV, Random, 95% CI)

-23.95 [-35.57, -12.32]

---

1.2 Two or more organisms

13

1375

Mean Difference (IV, Random, 95% CI)

-21.23 [-30.38, -12.09]

---

 2 Diarrhoea lasting ≥4 days

29

Risk Ratio (M-H, Random, 95% CI)

Subtotals only

---

2.1 Single organism

22

2136

Risk Ratio (M-H, Random, 95% CI)

0.45 [0.33, 0.60]

---

2.2 Two or more organisms

7

717

Risk Ratio (M-H, Random, 95% CI)

0.29 [0.12, 0.73]

---

 3 Mean stool frequency on day 2

20

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

3.1 Single organism

14

2040

Mean Difference (IV, Random, 95% CI)

-0.79 [-1.21, -0.38]

---

3.2 Two or more organisms

6

711

Mean Difference (IV, Random, 95% CI)

Not estimable

---

 

Comparison 5. Live versus killed organisms

---

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

---

 1 Mean duration of diarrhoea

32

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

1.1 Live organisms

29

3990

Mean Difference (IV, Random, 95% CI)

-26.55 [-36.95, -16.16]

---

1.2 Killed organisms

3

243

Mean Difference (IV, Random, 95% CI)

-10.39 [-30.75, 9.97]

---

 

Comparison 6. Dose of probiotic; live organisms

---

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

---

 1 Mean duration of diarrhoea

26

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

1.1 Low dose (≤10,000 million organisms/day)

16

2683

Mean Difference (IV, Random, 95% CI)

-25.88 [-39.04, -12.72]

---

1.2 High dose (>10,000 million organisms/day)

10

980

Mean Difference (IV, Random, 95% CI)

-27.02 [-38.64, -15.39]

---

 2 Diarrhoea lasting ≥4 days

17

Risk Ratio (M-H, Random, 95% CI)

Subtotals only

---

2.1 Low dose (≤10,000 million organisms/day)

13

1325

Risk Ratio (M-H, Random, 95% CI)

0.43 [0.29, 0.62]

---

2.2 High dose (>10,000 million organisms/day)

4

374

Risk Ratio (M-H, Random, 95% CI)

0.37 [0.12, 1.17]

---

 3 Mean stool frequency on day 2

15

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

3.1 Low dose (≤10,000 million organisms/day)

7

1455

Mean Difference (IV, Random, 95% CI)

-1.01 [-1.61, -0.41]

---

3.2 High dose (>10,000 million organisms/day)

8

861

Mean Difference (IV, Random, 95% CI)

-0.99 [-1.39, -0.60]

---

 

Comparison 7. Children with rotavirus diarrhoea

---

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mean duration of diarrhoea	12	701	Mean Difference (IV, Random, 95% CI)	-29.14 [-42.07, -16.20]
2 Mean stool frequency on day 2	3	164	Mean Difference (IV, Random, 95% CI)	-1.25 [-2.09, -0.41]

---

 

Comparison 8. Severity of diarrhoea; studies of outpatients

---

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

---

 1 Mean duration of diarrhoea

31

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

1.1 Studies of inpatients

26

3507

Mean Difference (IV, Random, 95% CI)

-20.90 [-31.44, -10.35]

---

1.2 Studies of outpatients

5

506

Mean Difference (IV, Random, 95% CI)

-42.81 [-55.07, -30.56]

---

 

Comparison 9. Mortality stratum for children and adults in the countries where trials were undertaken (children/adults)

---

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

---

 1 Mean duration of diarrhoea

32

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

1.1 Child and adult mortality low or very low

21

2075

Mean Difference (IV, Random, 95% CI)

-24.83 [-34.42, -15.23]

---

1.2 Either child or adult mortality high

11

2032

Mean Difference (IV, Random, 95% CI)

-24.75 [-43.40, -6.10]

---

 2 Diarrhoea lasting ≥4 days

26

Risk Ratio (M-H, Random, 95% CI)

Subtotals only

---

2.1 Child and adult mortality low or very low

19

1559

Risk Ratio (M-H, Random, 95% CI)

0.35 [0.23, 0.51]

---

2.2 Either child or adult mortality high

7

846

Risk Ratio (M-H, Random, 95% CI)

0.45 [0.26, 0.76]

---

 3 Mean stool frequency on day 2

19

Mean Difference (IV, Random, 95% CI)

Subtotals only

---

3.1 Child and adult mortality low or very low

14

1456

Mean Difference (IV, Random, 95% CI)

-0.99 [-1.35, -0.63]

---

3.2 Either child or adult mortality high

5

1231

Mean Difference (IV, Random, 95% CI)

0.00 [-0.78, 0.78]

---

 

Table 1. Heterogeneity in sensitivity analysis of primary outcomes¹

---

Sensitivity analysis	Outcome	Studies (no.)	χ^2	P value	I^2 (%)
Generation of allocation sequence	Mean duration diarrhoea Diarrhoea ≥4 days Stool frequency day 2	16 13 9	1077.2 46.2 26.9	< 0.00001 < 0.00001 0.0007	99 74 70
Concealment of allocation sequence	Mean duration diarrhoea Diarrhoea ≥4 days Stool frequency day 2	14 8 8	438.3 34.2 42.4	< 0.00001 < 0.0001 < 0.00001	97 8% 83
Blinding	Mean duration diarrhoea Diarrhoea ≥4 days Stool frequency day 2	26 16 14	1070.9 64.8 48.8	< 0.00001 < 0.00001 < 0.00001	98 77 73
Follow-up	Mean duration diarrhoea Diarrhoea ≥4 days Stool frequency day 2	25 19 15	672.3 52.3 54.5	< 0.00001 < 0.0001 < 0.00001	96 66 74

---

 1. Only trials considered adequate for quality assessment included; forest plots not shown