Article Text
Abstract
Objective We evaluated the role of diosmectite as an add-on treatment to the ‘recommended treatment’ of acute diarrhoea in children.
Methods We searched all published literature through the major databases: Medline via Ovid, PubMed, CENTRAL, Embase and Google Scholar till May 2014. Randomised clinical trials comparing diosmectite versus placebo were included (PROSPERO registration: CRD42014013783).
Main outcome measures The primary outcome measures were duration of acute diarrhoea (h), and day-to-day cure rates (%). The secondary outcome measures were stool output (volume), stool output (frequency) and adverse events.
Results Of 384 citations retrieved, a total of 13 randomised clinical trials (2164 children, 1–60 months old) were included in the meta-analysis. A dose of 3–6 grams per day of diosmectite was given for a duration from 3 days until recovery. Compared with placebo, diosmectite significantly decreased the duration of acute diarrhoea (mean difference, −23.39; 95% CI −28.77 to −18.01), and increased the cure rate (%) at day 5 (OR, 4.44; 95% CI 1.66 to 11.84), without any increases in the risk of adverse events. Diosmectite was effective in all types of acute childhood diarrhoea except dysentery. Because, most of the trials were open-label, and there was a high possibility of publication bias, the GRADE evidence generated was of ‘low quality’.
Conclusions Diosmectite may be a useful additive in the treatment of acute childhood diarrhoea. As the evidence generated was of ‘low quality’, future research is needed with higher quality designs before any firm recommendations can be made.
Trial registration number PROSPERO registration: CRD42014013783.
- Evidence Based Medicine
- Gastroenterology
Statistics from Altmetric.com
What is already known on this topic?
Diosmectite may be a useful additive in the treatment of acute childhood diarrhoea.
What this study adds?
There is ‘low quality’ evidence that diosmectite may be a useful additive in the treatment of acute childhood diarrhoea.
We need future research to have higher quality designs before any firm recommendations can be made.
Introduction
Acute diarrhoea kills more than 1.5 million children under 5 years of age every year globally, and is the second most common cause of death in this age group.1 The mainstay of treatment of an acute diarrhoeal episode includes oral rehydration therapy (ORT) and zinc.2 ORT aims to prevent or reverse dehydration, and has no effect either on the duration of diarrhoea or on the stool output. Zinc is not universally effective in the treatment of acute diarrhoea, and has been used mainly in developing country settings.3 For this reason, various medications and supplements (eg, loperamide, probiotics, prebiotics) have been studied in the treatment of acute childhood diarrhoea.4 ,5 However, there is lack of evidence of efficacy and some (eg, loperamide) may even have serious side effects.5
Diosmectite is a natural mineral clay having wide-ranging action on the gastrointestinal tract; they absorb bacteria, their toxins (endotoxin and exotoxin) and rotavirus; they increase water and electrolyte absorption and restore the barrier properties of intestinal epithelium after exposure to tumour necrosis factor α.6–11
A previous meta-analysis published in 2006 showed that diosmectite was associated with a moderate reduction in the duration of diarrhoea in children with acute infectious gastroenteritis.12 Many trials have been published after this. In order to provide the current best evidence and address some concerns (few trials contributing to the evidence, risk of publication bias and statistical conclusions based on heterogeneity) of the previous one, the present systematic review was conducted.
Methods
The review has been registered at the PROSPERO register: CRD42014013783.
Types of studies
Randomised controlled trials (RCTs) comparing diosmectite with placebo or no additional intervention with ≥80% follow-up (to reduce the risk of attrition bias in the included studies in case intention-to-treat (ITT) analysis has not been done).
Types of participants
Children of both sexes and up to 5 years of age with acute diarrhoea (defined below) and no/mild to some/moderate dehydration treated either in the hospital or in the outpatient/community setting were included. Exclusion criteria were, age <1 month (neonates), severe dehydration requiring intravenous rehydration, dysentery (gross blood in the stool), fever (≥39°C), history of persistent or chronic diarrhoea, malabsorption syndrome, antibiotic induced diarrhoea, severe malnutrition and inability to take oral medication.
Types of interventions
Trials, in which the intervention commenced within 7 days of onset of acute diarrhoea and consisted of administration of diosmectite orally at any dose and for any duration, were included. The intervention was administered as an adjunct to the standard treatment (ORT and zinc) of acute diarrhoea. Those in the control group received placebo or no additional intervention.
Types of outcome measures
Primary outcome measure
Duration of acute diarrhoea (h)
Cure rate (%) on day 5
Secondary outcome measures
Cure rate (%) at 48 h, 72 h and day 7
Stool output (volume, g/kg body weight)
Stool output (frequency)
Vomiting
Any other adverse events
Definitions
Acute diarrhoea was defined as defaecation frequency more than three times/day or more than the usual habit accompanied by changes in faeces consistency, without blood, lasting for less than 14 days. ‘Time to acute diarrhoea episode resolution’ was defined variably as the time (h) from the onset of diarrhoea to the first formed stool or to the first soft/formed stool followed by a non-watery stool or 24 h without stools or return to the average number of daily stools frequency. Recovery time was the time needed until the frequency of defaecation was less than or equal to three times per day with normal stool consistency without any complication. Cure rate was defined as the proportion of children whose diarrhoea has terminated by the given time.
Search methodology
The following major databases were searched systematically: Medline (1980 to April 2014) via Ovid, Pubmed (1980 to May 2014), Embase (1980 to April 2014), Cochrane CENTRAL (Issue 4, May 2014) and Google scholar (till May 2014). Besides these, searches were conducted on key conference proceedings, by contacting pharmaceutical companies, and by hand searching of the contents of key journals. No language restrictions were applied. Two reviewers reviewed the search results to identify relevant original human clinical trials. The details of search strategies have been attached as an online supplementary appendix S1.
Data extraction
Data extraction was done using a data extraction form that was designed and pilot tested a priori. Two authors independently extracted data including author, year, setting (country, inpatient/outpatient), participants (number, age and sex), inclusion and exclusion criteria, intervention (dose of diosmectite, schedule followed), results (outcome measures, effect, significance) and sources of funding/support. Any disagreement in the extracted data was resolved through discussion.
Assessment of risk of bias in the included studies
Two review authors independently assessed the methodological quality of the selected trials by using methodological quality assessment forms, and the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions.13 The handbook has seven criteria, which include—random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias) and any other bias. Any disagreements between the two review authors were resolved through discussion.
Data synthesis
The data from various studies were pooled and expressed as mean difference (MD) with 95% CI in case of continuous data (duration of acute diarrhoea and stool output volume), and OR with 95% CI in case of categorical data (cure rate at 48 h, 72 h, day 5, day 7 and stool output frequency). A p value <0.05 was considered statistically significant. We used the Q test (χ2 statistics) with an α of 0.1 to test heterogeneity among pooled estimates. For the primary outcomes when there was statistically significant heterogeneity in outcomes across studies, we tried to explore the cause by conducting subgroup analyses. A fixed-effects model was initially conducted. If significant heterogeneity existed between trials, potential sources of heterogeneity were considered and where appropriate a random-effects model was used. RevMan (Review Manager) V.5.0 was used for all the analyses.13
Grade of evidence
To assess the quality of evidence we used GRADE Profiler software (V.3.2).14 The software uses five parameters for rating the quality of evidence. The parameters used were—limitations to design of randomised controlled trials, inconsistency of results or unexplained heterogeneity, indirectness of evidence, imprecision of results and publication bias. The rating was done as—no, serious and very serious limitation.
Results
Description of studies
Of 384 citations retrieved, full texts of 30 articles were assessed for eligibility (figure 1). Out of these, a total of 13 RCTs were included.15–27 The detailed characteristics of trials have been described in table 1. A total of 2164 children ≤5 years old (excluding neonates <1 month) was included in the analysis.
Risk of bias in included studies
The details have been provided in online supplemental appendix S2.
Effect of interventions
Primary outcome measure
Time to acute diarrhoea episode resolution (h)
This was reported in 11 trials. But the result could be pooled from 10 trials (2002 participants), as one trial did not provide the measure of variance in either of the groups.16 The pooled result showed a significant reduction in the duration of acute diarrhoea (h) in the diosmectite group (MD, −23.39; 95% CI −28.77 to −18.01; p<0.00001) (figure 2). However, the included studies were significantly heterogeneous (χ2=112.19; p<0.00001). The following subgroup analyses were planned (some were planned a priori and some were planned post hoc) to find out the cause.
Definition of termination of diarrhoea: The post hoc subgroup analysis of trials using the definition of termination of diarrhoea as the duration from the beginning of therapy to passage of the last liquid stool (6 trials, 1305 participants) showed a significant reduction in the time to acute diarrhoea episode resolution (h) in the diosmectite group (MD, −20.57; 95% CI −23.64 to −17.5; p<0.00001). The heterogeneity was not significant (χ2=8.42; p=0.13) among the trials.
Age group: The pooled result from trials including children <24 months old (4 trials, 384 participants) showed a significant reduction in the duration of acute diarrhoea (h) in the diosmectite group (MD, −20.04; 95% CI −24.8 to −15.28; p<0.00001). The heterogeneity was not significant (χ2=5.11; p=0.16). The pooled result from trials including children <60 months old (3 trials, 975 participants) showed a significant reduction in the duration of acute diarrhoea (h) in the diosmectite group (MD, −22.07; 95% CI −24.8 to −19.34; p<0.00001). The heterogeneity was not significant (χ2=1.87; p=0.39). Only one trial each was found in the age group of <12 months and 24–60 months.25 ,26 Both the trials showed a significant reduction in the duration of acute diarrhoea (h) in the diosmectite group ((<12 months of age: 68 participants; MD, −31.6; 95% CI −33.04 to −30.16; p<0.00001); (24–60 months of age: 117 participants; MD, −18.03; 95% CI −24.7 to −11.36; p<0.00001)).
Blinding: We divided the trials into unblinded (open-label) and blinded subgroups. The results were significant in both the subgroups favouring the diosmectite ((open-label: 6 trials=1237 participants; MD, −22.49; 95% CI −27.3 to −17.67; p<0.00001); (double-blind: 5 trials=765 participants; MD, −26.12; 95% CI −37.8 to −14.43; p<0.0001)), but the included trials were significantly heterogeneous (open-label group: χ2=19.11, p=0.002; double-blind group: χ2=53.71, p<0.00001).
Adequacy of allocation concealment: For post hoc analysis, the trials were divided into adequate, no and unclear concealment subgroups. The results were significant in all the subgroups favouring the diosmectite, ((adequate concealment: 4 trials=703 participants; MD, −23.46; 95% CI −36.4 to −10.51; p=0.0004); (no concealment: 6 trials=1237 participants; MD, −22.49; 95% CI −27.3 to −17.67; p<0.00001); (unclear concealment: 1 trial=62 participants; MD, −41.4; 95% CI −60.81 to −21.99; p<0.0001)), but the included trials in two subgroups were significantly heterogeneous (allocation concealed group: χ2=52.43, p<0.00001; no allocation concealed group: χ2=19.11, p=0.002).
ITT: For post hoc analysis, the trials were divided into ITT and no ITT subgroups. The results were significant in both the subgroups favouring the diosmectite ((ITT: 8 trials=948 participants; MD, −24.81; 95% CI −32.03 to −17.59; p<0.00001); (no ITT: 3 trials=1054 participants; MD, −22.37; 95% CI −25.22 to −19.53; p<0.00001)). However, the included trials reporting ITT analysis were significantly heterogeneous (χ2=94.31; p<0.00001) compared with the no ITT analysis group (χ2=1.32; p=0.52).
Duration of treatment: The pooled result from trials in which the duration of therapy was 5 days (5 trials, 1269 participants) showed a significant reduction in the time to acute diarrhoea episode resolution (h) in the diosmectite group (MD, −20.97; 95% CI −25.08 to −16.86; p<0.00001), and the heterogeneity was not significant (χ2=7.45; p=0.11). The pooled result from trials in which the duration of therapy was until recovery (3 trials, 599 participants) showed a significant reduction in the time to acute diarrhoea episode resolution (h) in the diosmectite group (MD, −20.72; 95% CI −38.25 to −3.18; p=0.02), but the included trials were significantly heterogeneous (χ2=7.88; p=0.02).
When we compared different subgroups with the most common definition (ie, from the first drug administration to the passage of the last liquid stool prior to a formed stool) to the results from trials using other definitions to examine the influence of the choice of outcome definition on the results, the result was not vastly different from the overall result.
Cure rate (%) at day 5
Pooled result from four trials (254 participants) showed a statistically significantly higher cure rate in the diosmectite group (OR, 4.44; 95% CI 1.66 to 11.84; p=0.003). The heterogeneity was not significant (χ2=1.98; p=0.37).
Secondary outcome measures
Cure rate (%) at other time points
Cure rate (%) at 48 h: Pooled result from four trials (286 participants) showed a statistically significantly higher cure rate in the diosmectite group (OR, 13.08; 95% CI 7.16 to 23.88; p<0.00001). The heterogeneity was not significant (χ2=6.71; p=0.08).
Cure rate (%) at 72 h: Pooled result from four trials (254 participants) showed a statistically significantly higher cure rate in the diosmectite group (OR, 5.28; 95% CI 2.87 to 9.68; p<0.00001). The heterogeneity was not significant (χ2=1.21; p=0.75).
Cure rate (%) at day 7: Pooled result from two trials (1000 participants) showed a statistically significantly higher cure rate in the diosmectite group (OR, 1.74; 95% CI 1.19 to 2.54; p=0.004). The heterogeneity was not significant (χ2=0.45; p=0.5).
Stool output (volume, g/kg)
This was reported in four trials, but the result could be pooled from two trials (545 participants). The pooled result did not show a statistically significant reduction of the stool output in the diosmectite group by 72 h compared with the placebo group (MD, −10.13; 95% CI −24.19 to 3.93; p=0.16). The heterogeneity was not significant (χ2=0.95; p=0.33). Among the two other trials, one showed no effect of diosmectite on stool volume at various time points,19 while another showed a significant reduction in the stool outcome at various time points.17
Stool output (frequency)
This was reported in six trials, but the data was given in different formats (number of stools, time to passage of first formed stool and number of children with reduced stool frequency). Pooled result from three trials (196 participants) showed a statistically significant number of children with reduced stool frequency by 48 h in the diosmectite group (OR, 32.33; 95% CI 14.55 to 71.83; p<0.00001). The heterogeneity was not significant (χ2=0.12; p=0.94).
Out of the other three trials, pooled results from two trials (125 participants) showed a reduction in the frequency of defaecation in the diosmectite group at 72 h (MD, −0.58; 95% CI −0.9 to −0.3).16 ,18 Data from one trial that did not provide the measure of variance showed a reduction in the frequency of defaecation in the diosmectite group.22
Vomiting
Six trials provided data on vomiting. The pooled result from two trials showed no difference in the number of episodes of vomiting between the two groups (MD, 0.02; 95% CI −0.52 to 0.56; p=0.95).19 ,23 The heterogeneity was not significant (χ2=1.09; p=0.3). There was no difference in the number of vomiting episodes between the two groups in four other trials.17 ,18 ,22 ,24
Treatment related adverse events
Poor compliance with the treatment due to unpleasant taste: Three trials reported data on altered/bad taste. Pooled data from two trials showed no difference in the rate of this outcome between the two groups (OR, 0.91; 95% CI 0.78 to 1.07).15 ,16 The heterogeneity was not significant (χ2=0.73; p=0.51). Another trial reported the data for the diosmectite group (23%), but not for the placebo group.22
Constipation: Pooled data from two trials that reported data on constipation showed no difference between the two groups (OR, 5.8; 95% CI 0.7 to 47.1).18 ,20 The heterogeneity was not significant (χ2=0.03; p=0.87).
Any other adverse event: Six trials did not report any adverse events associated with diosmectite therapy.15 ,16 ,19 ,25–27
Publication bias
To assess whether there was bias in the published literature, a funnel plot was constructed using the MD and 1/SE values obtained from trials measuring one of the primary outcomes (duration of acute diarrhoea). In the absence of a publication bias, such a plot is expected to have a shape resembling an inverted funnel.28 From the funnel plot generated, the possibility of publication bias in the analysis is high (figure 3). This means, the trials that did not show diosmectite to be significantly better than placebo (negative) may not have been published, therefore, results of this meta-analysis may be biased in favour of diosmectite.
Grade of evidence
The evidence generated was of ‘very low quality’ for the 48 h cure rate; was of ‘low quality’ for time to acute diarrhoea episode resolution, 72 h and day 5 cure rates; and was of ‘moderate quality’ for stool output (volume), stool output (frequency) and day 7 cure rates (table 2).
Discussion
Summary of evidence
After an extensive search of the literature we could find 13 trials to be eligible for inclusion. Our results indicate that, compared with placebo, diosmectite significantly decreases the time to acute diarrhoea episode resolution (h), and increases the chance of cure rate (%) on day 5, without any increase in the risk of adverse events, when added to the standard treatment. The findings of the secondary outcomes were consistent with those of the primary outcomes. The funnel plot (constructed from trials measuring duration of acute diarrhoea) showed the possibility of publication bias to be high. When we constructed the GRADE of evidence from the available evidence, it was found to be of ‘low quality’ for primary outcomes, and of ‘moderate quality’ for most of the secondary outcomes.
The trials were variable with each other in some respects, and this intertrial variability contributed to the overall variability of the result. The trials that followed the uniform definition of termination of acute diarrhoea, reported no ITT analysis, reported outcomes for children <24 months old and <60 months old, and those treated with diosmectite for 5 days showed less variability in the result.
The present updated meta-analysis has some important differences from the previously conducted meta-analysis.12 First, 13 RCTs including 2164 participants were included in the present review, whereas 9 RCTs including 1238 participants were included in the previous one. Second, because of the small number of trials the previous review could not asses the publication bias, whereas the possibility of publication bias was noted to be high as the present review included more number of trials. Third, the previous review did not grade the quality of evidence, whereas the present review graded it (please see above).
Recently, the updated evidence-based guideline of the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and the European Society of Pediatric Infectious Diseases for management of acute gastroenteritis was published.29 The guideline recommends that diosmectite can be considered for the management of acute gastroenteritis (weak recommendation, moderate-quality evidence). But, in the present review more number of trials was included and the evidence was found to be of ‘low quality’.
Limitations
The variability of definitions of termination of diarrhoea is a limitation of the analysis. We could not take into account the type of aetiological agents (bacterial or viral) separately as none of the trials reported the outcome separately for both of them (only one trial reported the outcome based on rotavirus positive or negative status). We could not determine the effect of diosmectite administered in hospitalised (inpatient) children versus those taking outpatient treatment. As the dose range varied among the trials, we could not determine an optimal therapeutically effective dose of diosmectite. None of the trials included children above 5 years of age, so it is difficult to make any recommendation in this age group.
Further area of research
Future trials should have an uniform definition of termination of acute diarrhoea, and should report on the effect of diosmectite based on the aetiology. Trials should also report on the cost–benefit ratio. The effect of diosmectite should also be reported based on the severity, hospitalisation or outpatient status. The minimally effective dose and duration in different subgroups of children should also be determined.
Conclusion
Diosmectite may be a useful additive in the treatment of acute childhood diarrhoea. However, the evidence generated was of ‘low quality’ for the primary outcomes, and there was a high possibility of publication bias in the analysis. We need future research to have higher quality designs, so that it provides higher quality evidence than the current available evidence.
Acknowledgments
The authors thank Dr Nishant P Jaiswal, Scientist C, ICMR Advanced Centre for Evidence Based Child Health, PGIMER, Chandigarh, India for providing help in the database search.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online Appendix S1
- Data supplement 2 - Online Appendix S2
Footnotes
Contributors RRD, JS and SSN conducted the searches, collated and reviewed the studies, extracted the data, performed the analysis and wrote the draft. All authors approved the final version to be submitted. RRD will act as the guarantor.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.