Probiotics are described as nonpathogenic living microorganisms that when ingested in sufficient large amounts exert positive health benefits on the host beyond any inherent general nutritional value (1). The probiotic intervention that has received the most attention in the published medical literature is in the treatment of acute diarrhoeal disease. The Lactobacillus species, in particular Lactobacillus GG (LGG), has been extensively studied in both adults and children (2). Saccharomyces boulardii has also been found to be an effective probiotic for the treatment of acute gastroenteritis and the prevention of antibiotic-associated diarrhoea in children (3,4). Probiotic treatment has shown to be moderately effective as adjunctive therapy in reducing the duration of proven or presumed infectious diarrhoea (5).
Aboriginal children living in remote and rural communities in Top End of the Northern Territory experience severe morbidity from the effects of acute diarrhoeal disease (6). The effectiveness of probiotic therapy in malnourished children with a high burden of infectious comorbidity and underlying tropical enteropathy remains unclear in this setting.
The aim of the present study was to formally evaluate the potential benefits of probiotics as an adjunctive therapy in this population of children. Although we included standardised diarrhoeal parameters such as duration, stool frequency, and stool output as important outcomes, our primary outcome measurement was short-term recovery from intestinal mucosal injury. This was measured objectively using the noninvasive sucrose breath test (SBT) as previously reported (7).
PATIENTS AND METHODS
This prospective, double-blind, placebo-controlled randomised clinical trial was approved by the Institutional Human Research Ethics Committee in Darwin. Written informed consent was obtained from the parent or guardian of all of the children by an Aboriginal health worker at the time of entry into the study.
Aboriginal children between the ages of 4 months and 2 years requiring admission to the Royal Darwin Hospital with a clinical diagnosis of acute diarrhoeal disease (defined as 3 or more loose stools during the 24 hours preceding presentation and diarrhoea duration of <7 days) were eligible for enrollment. Subjects were required to tolerate oral rehydration at the time of study commencement. Children were excluded if they had any of the following symptoms: oxygen requirement during the study period; chronic cardiac, renal or respiratory disease; previous gastrointestinal surgery; proven sucrose intolerance; suspected or known immunodeficiency; or received probiotic supplementation before enrollment. Infants younger than 4 months were excluded because of possible risk of inducing osmotic diarrhoea from oral sugar probe solutions and the lack of validation of the SBT in this age group.
Setting and Location
The present study was conducted at a single site on the Paediatric Infectious Diseases Ward of the Royal Darwin Hospital. This hospital is a tertiary referral centre serving a catchment population of approximately 30,000 Aboriginal children living in isolated rural communities across the Top End of the Northern Territory. Many of the children recruited into the present study had already experienced a high burden of infectious diseases related to poverty and social disadvantage. Study participants were cared for and closely monitored by paediatric nursing and medical hospital staff experienced in the management of gastrointestinal disease in indigenous children.
Patients with acute diarrhoeal disease were randomised to receive either Lactobacillus casei strain GG (Gelfilus, Valio Ltd, Helsinki, Finland) or placebo at a dose of 1 capsule 3 times per day for 3 days. L casei strain GG was supplied as hard gelatin capsule containing ≥5 × 109 colony-forming units in cellulose microcrystalline powder. Placebo capsule and contents were identical except the capsule contained no LGG. The powder from each capsule was reconstituted in 5 mL of sterile NaCl 0.9% and given via a nasogastric tube. Ward nurses prepared and administered the L casei strain GG and placebo.
Participants received L casei strain GG or placebo within 24 hours of admission after completion of baseline clinical assessment and study investigations according to the investigator's protocol. Oral antibiotics were not administered within 6 hours of receiving the study product during the course of the trial.
Treatment of Acute Diarrhoeal Disease
All of the children admitted to the ward with acute diarrhoeal disease were assessed for dehydration according to standard World Health Organization criteria and received rapid rehydration therapy including oral rehydration solution (Gastrolyte, Aventis Pharma Pty Ltd, Sydney, Australia) or intravenous solution with Ringer's lactate (Hartmann's) following the hospital protocol.
Baseline investigations included full blood count, urea, creatinine and serum electrolytes, venous blood gas analysis, a timed 90-minute blood lactulose-rhamnose ratio as described previously (8), stool microscopy and culture, including polymerase chain reaction for diarrhoeagenic Escherichia coli and urinanalysis.
Children received zinc 20 mg daily and single-dose oral vitamin A 50,000 to 100,000 IU as standard practice. Maintenance therapy was commenced after successful rapid rehydration and continued at the discretion of medical ward staff until diarrhoea had settled. Breast-feeding was continued throughout this period.
Specific Objectives and Hypothesis
The specific objective was to assess the efficacy of LGG in the management of proven or presumed acute infectious diarrhoeal diseases in Aboriginal children living in Top End of the Northern Territory. In the present study we tested the hypothesis that L casei strain GG would enhance short-term recovery of small intestinal injury and reduce diarrhoea.
We defined the primary outcome as the short-term change in small intestinal absorption capacity during the study intervention period. This was measured objectively using the noninvasive 13C-SBT. Subjects were first tested at baseline (day 1) and again on day 4 after completion of allocated randomised treatment. To further assess changes in gastrointestinal recovery we also included the following secondary endpoints consistent with published Cochrane review outcome measurements (5): diarrhoea duration, diarrhoea frequency, total stool output, and proportion of subjects with diarrhoea on days 3 and 4; change in body weight as measured on days 1 and 4; and total oral rehydration salts and intravenous rehydration therapy requirements during the intervention period. The safety and tolerability of LGG was also monitored.
13C-SBT (7) was performed following at least a 4-hour fast. Baseline (0 min) expiratory breath samples were collected from the nose and mouth using a nasal prong attached to a syringe. Subjects then received a sugar probe solution containing 2 g/kg of natural nonradioactive isotope 13C-sucrose (maximum 20 g) combined with 5 g of lactulose and 1 g of rhamnose (for the blood permeability lactulose-rhamnose ratio), with the volume adjusted to an osmolality of 400 mOsm/L. Breath samples were collected every 15 minutes for the first 90 minutes and then every 30 minutes until 150 minutes after ingestion; during that time subjects were fasting. A hand-held oxygen analyser was used to ensure that all of the samples were expiratory, defined as <19.5%. Breath samples were transferred to 10-mL glass tubes for storage and handling. The ratio of 13C to 12C in breath was measured using isotope ratio mass spectrometry (Europa Scientific, ABCA 20/20, Crewe, UK). Results were expressed as percentage of 13CO2 cumulative dose recovered at 90 minutes.
Diarrhoea duration was calculated in hours from the time the intervention was first given until the time of last loose stool in which fewer than 3 loose stools occurred within a 24-hour period. The proportion of patients with diarrhoea on days 3 and 4 was determined and the total number of diarrhoea stools with loose consistency during the admission recorded. Diarrhoea output was graded using a visual scoring system validated previously in this setting, with a score of 1 indicating a small amount of loose stool (≤10 cm2 of nappy surface area), 2 indicating a medium amount (10–20 cm2), and 3 indicating a large amount (>20 cm2). Diarrhoea severity assessments were meticulously recorded as part of ward routine practice by nursing staff.
Adverse events related to either the study products and/or protocol investigations were reported to the ethics committee approving the study.
Based on our previous study (7) of 13C-SBT in Aboriginal children with diarrhoeal disease (n = 18, 1.9 ± 1.0 mean and standard deviation) and without diarrhoeal diseases (n = 18, 4.1 ± 2.2), we determined that at least 33 children would need to be enrolled to provide the study with 80% power to detect a 40% increase in small-intestinal absorption capacity in the LGG group with a type I error of 0.05.
Children were allocated to receive either LGG or placebo using a computer-generated block randomisation sequence (StataCorp, College Station, TX) prepared from an independent research institute.
Allocation, Enrollment, and Assignment
The randomisation sequence series was contained in a set of sealed envelopes with a study number in order on the outside and code on the inside referring to the same numbered bottle stored in pharmacy. Patients were enrolled by the medical officer and Aboriginal health worker conducting the study. They were also responsible for opening the next sealed envelope, allocating the study product to patients, and assigning participants to the appropriate group. Both envelopes and coded bottles were prepared offsite and the randomisation and allocation information was concealed until recruitment, data collection, and analyses were completed. Study personnel, clinical ward staff, and patients remained blind to treatment assignment throughout the trial.
Data were entered and analysed according to the intention-to-treat principle using Stata 8 (StataCorp, College Station, TX). Normally distributed data are expressed as arithmetic means with 95% confidence intervals (CIs) and skewed data as geometric means with 95% CIs after logarithmic transformation. Continuous variables with homogeneity of variance were compared using the 2-sample t test and paired t test for repeated observations. Categorical variables were analysed using the χ2 test or the Fisher exact test. Scatterplots and Kaplan-Meier curve was constructed using Prism 4 (GraphPad Software Inc, San Diego, CA).
Safety and Quality Monitoring
External monitoring of the study and data review were conducted at 6 monthly intervals throughout the trial period for safety and quality evaluation.
Participant Flow Diagram
Five patients from the probiotics arm and 1 from the placebo arm were not available for evaluation after the randomisation process. The reasons for this included discharge before receiving intervention (n = 4) and parental withdrawal (n = 2) (Fig. 1). In the probiotic and placebo groups 33 of 38 and 31 of 32 were analysed, respectively.
Patients were recruited from June 2002 until March 2004.
There was no significant difference in the baseline demographics, clinical characteristics, or laboratory investigations between the 2 groups (Table 1). Aboriginal children with acute diarrhoeal disease frequently present with severe dehydration, acidosis, hypokalaemia, and abnormal intestinal permeability as a consequence of intestinal mucosal damage and lactose intolerance. Malnutrition, iron deficiency, and associated infective comorbidities are also common. Rotavirus accounts for less than one third of all of the enteric pathogens isolated.
Table 2 shows the results of percent cumulative 13CO2 recovery results obtained at 90 minutes on days 1 and 4 for the placebo and probiotics groups. Although we demonstrated a mean improvement in the percent cumulative 13CO2 recovery during 4 days from each group, this observed increase was not statistically significant for either the probiotics (P = 0.22) or placebo (P = 0.19) arm. Subgroup analysis of children (n = 17) in which rotavirus was subsequently identified produced similar results for each group (not shown).
In this clinical trial, probiotics showed no significant improvement on the functional absorptive capacity of the small intestine (as measured by the percent cumulative 13CO2 recovery at day 4) compared with placebo.
Probiotics did not reduce the duration of diarrhoea (Fig. 2), total number of loose stools, or total stool volume in this group of children (Table 3). On day 2 of the intervention, however, participants in the probiotic group had on average 1.41 (95% CI 1.00–1.99) fewer stools. Unfortunately, this effect was not sustained on days 3 and 4, and the proportion of children who presented without diarrhoea on these days was similar in both groups. Probiotic therapy did not appear to prevent or reverse the degree of dehydration required during the intervention period (Table 3). No significant differences were observed in any of the outcomes between the probiotic and placebo groups, when subgroup analysis of children with rotavirus only and/or those not exposed to antibiotics were examined (results not shown). We reported no adverse effect attributable to LGG in the present study protocol.
Probiotics had no significant impact on minimising any of the important diarrhoea symptoms in proven or presumed infectious acute gastroenteritis in this setting. Furthermore, we were unable to demonstrate any improvement in small intestinal recovery during the trial period using the methods we described. Our findings are consistent with other published trials that do not demonstrate a beneficial effect from use of probiotic therapy in diarrhoeal disease and will add further to the significance between study heterogeneity in outcomes already seen in a Cochrane review on the subject (5,9–12). However, the present study did show a more rapid healing in the LGG group, because there were significantly fewer patients with diarrhoea on day 2 and trending on days 3 and 4. There was also a trend for increased weight gain on day 4 in the LGG group compared with the placebo group. Importantly, our study emphasises the need for a cautious approach towards probiotic therapy in the acute management of diarrhoeal disease in children, especially where the clinical setting does not closely match published trial conditions.
In a recent meta-analysis of trials using only the probiotic strain L casei GG, Szajewska et al (13) concluded that the use of this probiotic strain is associated with “moderate clinical benefits and worthy of trial in healthy infants and children of developed countries.” LGG, however, appeared to be more beneficial in rotavirus disease, but was ineffective for diarrhoea caused by either invasive enteropathogens or in circumstances in which no pathogen was identified (13). The observed benefit with rotavirus was principally in reducing the risk of diarrhoea after 7 days and in shortening the duration of hospitalisation by −0.89 days (13). In our study we had insufficient participants with rotavirus diarrhoea alone to demonstrate any significant benefit from LGG. Somewhat surprising from this meta-analysis was the lack of effect from LGG on reducing stool volume and absence of any sustained reduction in the number of stools across the time intervals studied. These findings are consistent with our own results.
We are the first group to formally evaluate the effects of probiotic therapy on the absorptive capacity and integrity of the small intestine using the SBT. As previously reported, Aboriginal children with acute diarrhoeal disease in this setting have impaired absorptive capacity from both acute enteritis and underlying environmental enteropathy (14). The potential mechanisms of action of probiotic therapy for acute diarrhoeal disease and effects on intestinal epithelial function have been reviewed elsewhere (15–17). In the present study, however, L casei strain GG did not appear to enhance short-term recovery from small intestinal mucosal injury over placebo using the SBT. Lack of clinical improvement with all of the diarrhoeal parameters examined was also consistent with the primary outcome findings.
The strengths of the present study were to include children with nongastrointestinal infectious comorbidity and malnutrition, commonly seen in the developing world setting; not selectively excluding participants based on presence or absence or type of enteric pathogens identified; including children with recent antibiotic exposure; and using an objective measure of small intestinal function in addition to standard diarrhoeal parameters to evaluate probiotic efficacy.
A limitation of the present study was the smaller numbers of participants enrolled. The power of the present study was calculated based on changes in the SBT and may be underpowered to detect changes in clinical diarrhoeal parameters. We also acknowledge the limits of the SBT sensitivity in detecting small improvements in small-intestinal absorptive capacity during a 4-day period. Furthermore, the administration of probiotics may be too short to observe any beneficial effects in this population.
The lack of any sustained significant effect from LGG use in our study is likely to be multifactorial. The smaller numbers of rotavirus-only cases identified in the present study are a limitation. As shown in previous studies those with rotavirus demonstrate better outcomes from probiotic use (5). Furthermore, polymicrobial enteric infection in these children is far more common in these children when sophisticated diagnostic techniques are used as previously shown (18). Aboriginal children also frequently present with more moderate to severe diarrhoeal disease and experience a higher rate of complications (14). The presence of an underlying environmental enteropathy, together with malnutrition, may also have a significant impact on slowing the gut recovery process. Aboriginal mothers in the Northern Territory continue breast-feeding until about 18 months on average, and this may also have an impact on the efficacy of probiotics (19). Exposure to oral and or intravenous antibiotics could reduce the effectiveness of the probiotic therapy, although we did not find any difference from a subgroup analysis of those children who had received versus not received antibiotics in the present study. Furthermore, a recent study (20) has shown that higher dosage and longer duration of LGG administration compared with the present study was effective in reducing duration of diarrhoea and length of stay in hospital in Indian children with acute watery diarrhoea.
Although we confirmed the safety or tolerability of LGG in this trial, we cannot currently recommend this type of empiric therapy for Aboriginal children presenting to hospital from remote communities across the Northern Territory of Australia. Efforts would be better focused on interventions aimed at preventing infectious diarrhoea at a community level, including vaccination of infants for rotavirus and therapies with proven clinical effectiveness in this setting. Health providers in the developing world should carefully evaluate both the clinical and economic benefits of probiotic therapy before incorporating this form of treatment into routine clinical practice.
We acknowledge the support and cooperation of Josephine Brinjin, our Aboriginal health worker research assistant, the nursing staff of ward 7B at the Royal Darwin Hospital, and all of the families and children who participated in the present study.
1. O'Sullivan GC. Probiotics. Br J Surg 2001; 88:161–162.
2. Gorbach SL. Probiotics and gastrointestinal health. Am J Gastroenterol 2000; 95:S2–S4.
3. Vandenplas Y, Brunser O, Szajewska H. Saccharomyces boulardii in childhood. Eur J Pediatr 2009; 168:253–265.
4. Zanello G, Meurens F, Berri M, et al. Saccharomyces boulardii effects on gastrointestinal diseases. Curr Issues Mol Biol 2009; 11:47–58.
5. Allen SJ, Okoko B, Martinez E, et al. Probiotics for treating infectious diarrhoea. Cochrane Database Syst Rev 2003;CDC003048.
6. Kukuruzovic R, Haase A, Dunn K, et al. Intestinal permeability and diarrhoeal disease in Australians. Arch Dis Child 1999; 81:304–308.
7. Ritchie BK, Brewster DR, Davidson GP, et al. The 13C-sucrose breath test: novel use of a non-invasive biomarker of environmental gut health. Pediatrics 2009; 124:620–626.
8. Haase AM, Kukuruzovic RH, Dunn K, et al. Dual sugar permeability testing in diarrheal disease. J Pediatr 2000; 136:232–237.
9. Bhatanagar S, Sing K, Sazawal S, et al. Efficacy of milk vs yoghurt as part of a mixed diet in acute non-cholera diarrhoea among malnourished children. J Pediatr 1998; 132:999–1004.
10. Costa-Ribeiro H, Ribeiro T, Mattos A, et al. Limitations of probiotic therapy in acute severe dehydrating diarrhoea. J Pediatr Gastroenterol Nutr 2003; 36:112–115.
11. Salazar-Lindo E, Miranda-Langschwager P, Campos-Sanchez M, et al. Lactoabacillus casei strain GG in the treatment of infants with acute watery diarrhoea: a randomised double-blind, placebo controlled clinical trial. BMC Pediatr 2004; 4:18.
12. Basu S, Chatterjee M, Ganguly S, et al. Efficacy of Lactobacillus rhamnosus GG in acute watery diarrhoea of Indian children: a randomised controlled trial. J Pediatr Child Health 2007; 43:837–842.
13. Szajewska H, Skorka A, Ruszczynski M, et al. Meta-analysis: Lactobacillus GG for treating acute diarrohea in children. Aliment Pharmacol Ther 2007; 25:871–881.
14. Kukuruzovic R, Brewster D. Small bowel intestinal permeability in Australian Aboriginal children. J Pediatr Gastroenterol Nutr 2002; 35:206–212.
15. Vanderhoof J, Young R. Use of probiotics in childhood gastrointestinal disorders. J Pediatr Gastroenterol Nutr 1998; 27:323–332.
16. Isolauri E. Probiotics for infectious diarrhoea. Gut 2003; 52:436–437.
17. Boirivant M, Strober W. The mechanisms of action of probiotics. Curr Opin Gastroenterol 2007; 23:679–692.
18. Kukuruzovic R, Robins Brown R, Anstey N, et al. Enteric pathogens, intestinal permeability and nitric oxide production in childhood diarrhoeal disease. Pediatr Iinfect Dis J 2002; 21:730–739.
19. Salminen S, Isolauri E. Opportunities for improving the health and nutrition of the human infant by probiotics. Nestle Nutr Workshop Ser Pediatr Program 2008; 62:223–233.
20. Basu S, Paul DK, Ganguly S, et al. Efficacy of high-dose Lactobacillus rhamnosus GG in controlling acute watery diarrhea in Indian children: a randomized controlled trial. J Clin Gastroenterol 2009; 43:208–213.
© 2010 Lippincott Williams & Wilkins, Inc.