Diarrhoea remains a leading cause of childhood death in middle and low income countries. The main complication is dehydration, which was treated with intravenous fluid infusion until the early 1960s. Oral rehydration solution (ORS) is now the mainstay of therapy and is particularly useful when intravenous fluids are in short supply, health services are basic, and there is a shortage of skilled personnel (Almroth 1995). The combination of salt and sugar enhances fluid absorption because sodium and glucose transport in the small intestine are coupled, and glucose promotes absorption of both sodium ions and water (Fordtran 1968). Diarrhoea is caused by derangement of fluid absorption and secretion from the gut, and coupling sodium and glucose allows absorption, even during active fluid secretion due to infection. Thus rehydration can take place even with large fluid losses, as seen in enterotoxic diarrhoea, such as that caused by cholera or infection with Escherichia coli (Guarino 2001).
ORS has proved both safe and effective worldwide in hospital settings, and is now widely used in the home to prevent dehydration (Mahalanabis 1973, Grant 1983).For more than two decades, the World Health Organization (WHO) has recommended the standard formulation of glucose-based ORS with 90 mmol/L of sodium and 111 mmol/L of glucose and a total osmolarity of 311 mmol/L. It remains unclear however, whether this is the optimum level of sodium. Laboratory work suggests that lower concentrations of sodium and glucose enhance solute induced water absorption (Farthing 1988, Hunt 1992). Papers report patients experiencing blood sodium levels above the normal of 150 mmol/L with standard solution (Finberg 1973).
The objective of this review is to critically appraise and evaluate all relevant randomized controlled trials addressing comparative effects of reduced osmolarity ORS with WHO standard ORS. One potential adverse effect of reduced osmolarity ORS is a deficiency of sodium is the blood (hyponatraemia), which can give rise to convulsions. We are also exploring the risk of this adverse outcome through trial and observational data.
We confined the review to children, as they are most vulnerable to dehydration and electrolyte imbalance from diarrhoea, and are the targets for large primary care child investments that include ORS sachet distribution. Severity, duration, and volume of diarrhoea are often primary outcomes in clinical ORS studies, but we sought a pragmatic outcome relevant to health providers. ORS aims to rehydrate children and avoid the need for intravenous fluid infusion. We therefore identified unscheduled intravenous fluid infusion as a primary outcome as this represents failed oral therapy.
To compare reduced osmolarity oral rehydration solution with the World Health Organization recommended strength for treating diarrhoea in children.
Criteria for considering studies for this review
Types of studies
Randomized controlled trials, defined as a trial in which the subjects followed were assigned prospectively to one of two or more interventions by random allocation. This excludes quasi-randomized designs.
Types of participants
Children with acute diarrhoea (history of less than 5 days).
Types of interventions
Experimental: Reduced osmolarity oral rehydration solution (total osmolarity 250 mmol/L or less with reduced sodium).
Control: World Health Organization standard oral rehydration solution (90 mmol/L sodium, 111mmol/L glucose, total osmolarity 311 mmol/L).
Types of outcome measures
Need for unscheduled intravenous fluid infusion during the course of treatment.
- Stool output.
- Children vomiting during rehydration.
- Asymptomatic hyponatraemia (defined as serum sodium less than 130mmol/L) during follow up.
Search methods for identification of studies
We used the following search terms to search all trial registers and electronic databases: child; diarrhoea; fluid therapy; oral rehydration; osmolar; and rehydration solutions.
We searched the following trial register: Cochrane Central Register of Controlled Trials (CENTRAL), published in The Cochrane Library (Issue 3, 2004).
We searched the following electronic databases: MEDLINE (1966 to July 2004); EMBASE (1988 to July 2004); and Current Contents (July 2004).
We also checked the citations of existing reviews and trial reports.
For unpublished data and ongoing trials, we contacted current researchers and key agencies, including the World Health Organization, the Centers for Disease Control and Prevention, Atlanta (USA), and the International Centre for Diarrhoeal Disease Research, Bangladesh.
Data collection and analysis
Selection of studies
SH and SK independently applied the inclusion criteria to all identified trials, and differences were resolved by discussion with the PG.
Data extraction and management
SH and SK extracted data on relevant outcome measures using a standardized data abstraction form.
Assessment of risk of bias in included studies
Each included trial was assessed in terms of adequacy of concealment of allocation, generation of allocation sequence, blinding, and follow up of patients, using the guidelines of the Cochrane Infectious Diseases Group. Studies excluded were detailed in the 'Characteristics of excluded studies'.
We used the Mantel-Haenszel odds ratio (OR) for binary outcomes. The odds ratios were not estimated when neither intervention group found any event, which are indicated in the MetaView figures. We used the Standardized Mean Difference (SMD) for continuous outcomes. We combined studies using a fixed effect method. For all estimates, we calculated 95% confidence intervals. We tested statistical heterogeneity using Chi-square statistic with a P-value less than 0.1 indicating statistical significance. We had prespecified potential sources of heterogeneity for analysis. We examined publication bias using a funnel plot, and a regression approach (Egger 1997) to assess funnel plot asymmetry. We conducted a sensitivity analysis in relation to adequate allocation concealment.
After presentation of the results, an expert consultation group from the World Health Organization recommended a stratified analysis by mmol sodium (less than 75 mmol and 75 to 85 mmol), and this analysis is now included (WHO 2001).
Description of studies
We identified 41 studies for inclusion, and 16 studies met the inclusion criteria. The progress through the stages of meta-analysis, using the process suggested in the QUOROM statement (QUOROM Group 1999), is shown below.
41 studies comparing oral rehydration solution (ORS) formulation for treating diarrhoea.
- 6 excluded as not randomized controlled trials (RCTs);
35 remaining RCTs of ORS comparing formulation for treating diarrhoea patients.
- 9 excluded as intervention was something other than reduced osmolarity ORS;
26 remaining RCTs reporting reduced osmolarity ORS in one treatment arm.
- 6 excluded if control group did not use World Health Organization (WHO) standard ORS;
20 remaining RCTs reporting comparison of reduced osmolarity ORS with WHO standard ORS.
- 2 excluded as not in children;
18 remaining RCTs reporting comparison of reduced osmolarity ORS with WHO standard ORS for treating children with diarrhoea.
- 2 excluded as no relevant outcomes reported;
16 remaining RCTs reporting comparison of reduced osmolarity ORS with WHO standard ORS in children with diarrhoea in relation to need of unscheduled intravenous fluid infusion therapy and some measures of clinical illness.
This leaves a total of 14 included studies. These were from Egypt (2), Bangladesh (3), Mexico (1), Colombia (1), India (3), Panama (1), and the USA (1). Two other studies were multicentre trials; one was conducted in Brazil, India, Mexico, and Peru, and the other in Bangladesh, Brazil, India, Peru, and Vietnam.
Participants were children with acute non-cholera diarrhoea in all trials except three, which included cholera patients (Bangladesh 1995b, CHOICE 2001, India 2000b). In all but one which included children up to 5 years old (India 2000a), the participants' ages ranged between 1 and 36 months. All children had some degree of clinical dehydration. One trial treated all children on day 1 with intravenous fluid infusion, and those still producing 80 ml/kg/24h were then randomized (Bangladesh 1995a). In five trials (CHOICE 2001, India 2000b, Panama 1982, WHO 1995, USA 1982) severely dehydrated children were included. Five trials included malnourished children (Bangladesh 1995b, Colombia 2000, Bangladesh 1995b, India 2000a, India 2000b, Mexico 1990a). The number of breastfed children was reported in eight trials (Bangladesh 1995a, Bangladesh 1995b, Bangladesh 1996a, CHOICE 2001, Colombia 2000, Egypt 1996b, India 2000b, WHO 1995). Fully weaned children were included in one trial (Egypt 1994).
We deviated slightly from the osmolarity definitions in our peer refereed protocol published in The Cochrane Library. For reduced osmolarity, we had defined this to be lower than 250 mmol/L, but some studies defined this as higher, and we therefore extended our limit to a total osmolarity of 270 mmol/L. For the WHO standard ORS, defined as a total osmolarity of 311 mmol/L, we also included two studies that used a slightly different WHO standard ORS with a total osmolarity of 331 mmol/L but with the same sodium and glucose combination (Panama 1982, USA 1982). All but two trials used a glucose based reduced osmolarity ORS; one used sucrose (Bangladesh 1996a), and one used L-alanine with glucose (Bangladesh 1995a).
Risk of bias in included studies
All of the studies were randomized controlled trials. Nine reported methods that assured adequate allocation concealment (WHO 1995, CHOICE 2001, Colombia 2000, Egypt 1996b, Bangladesh 1995a, Bangladesh 1995b, Bangladesh 1996a, India 2000a, India 2000b).
Six studies (CHOICE 2001, Egypt 1996b, Bangladesh 1995b, Bangladesh 1996a, India 2000a, India 2000b) were double blinded. One of the Mexico studies was described as single blinded (Mexico 1990b), but this study is currently excluded from the analysis as it is thought to be a subset of patients reported in another paper which is included, where no details of blinding are given (Mexico 1990a). Eight studies did not mention blinding.
Inclusion of all randomized participants
Included trials had losses to follow up of less than 10% of randomized participants in all cases.
Effects of interventions
Meta-analyses of the four outcomes are illustrated in the MetaView summary analysis.
Information for the primary outcome of the need for unscheduled intravenous fluid infusion was found in 11 trials (n = 1996). In the meta-analysis of 8 trials, a statistically significant reduction for unscheduled intravenous infusion for participants receiving reduced osmolarity oral rehydration solution (ORS) when compared with World Health Organization (WHO) standard ORS was demonstrated (odds ratio 0.59, 95% confidence interval 0.45 to 0.79). 3 of the 11 trials reported that none of their patients needed intravenous fluid infusion in either group, and the odds ratios were not calculated for these trials.
11 trials reported stool output during rehydration. These trials measured stool output in various ways using different units. We therefore used the standardized mean difference to analyse these data. Since the stool output in diarrhoeal disease showed a positive skewed distribution with clinical improvement, we used a log-normal approximation. The pooled standardized mean difference in the log scale is -0.23 (95% confidence interval -0.33 to -0.14), which suggests that the reduced osmolarity ORS resulted in significantly less stool output when compared with the WHO standard ORS. Data from one trial (India 2000a) were not combined with the others in the meta-analysis because this trial measured stool output for a much longer period beyond rehydration phase. The individual results of all 12 trials are summarized in Appendix 1.
For children vomiting during rehydration, six trials reported these data. The tendency was for fewer patients to vomit in the reduced osmolarity ORS group (Odds ratio 0.71, 95% confidence interval 0.55 to 0.92).
For presence of hyponatraemia, six trials reported this outcome. Three of these six trials did not observe hyponatraemia in any participants, irrespective of their allocated group. The meta-analysis of three trials, during which participants developed hyponatraemia, showed no significant difference between the groups (odds ratio 1.45, 95% confidence interval 0.93 to 2.26).
We tested for statistical heterogeneity of treatment effect across trials using the Chi-square statistic for all meta-analyses, and the statistic is presented in each meta-analysis diagram. Results suggest no evidence of statistical heterogeneity (P-value > 0.1) for any outcome considered.
A funnel plot was prepared with the primary outcome. The regression method used to assess funnel plot asymmetry yielded an intercept of -0.104 with a P-value of 0.12, indicating no significant evidence of publication bias.
Sensitivity analysis carried out included studies where allocation concealment was clearly described as adequate and suggested little difference from the original meta-analysis. For example, the pooled odds ratios of the seven trials for the primary outcome with adequate allocation concealment was 0.61 (95% confidence interval 0.46 to 0.82).
A stratified analysis by sodium content of the ORS is presented. Hyponatraemia was not detected in the three studies examining the very low sodium ORS.
We intended to examine treatment effects in cholera subgroup compared with non-cholera diarrhoea. A Cochrane Review of rice-based rehydration compared with glucose oral rehydration solution (ORS) showed that rice water was associated with lower stool volumes in cholera patients but not in diarrhoea from other causes (Fontaine 2000). The available data were insufficient however. Three studies (CHOICE 2001, Bangladesh 1995b, India 2000b) involved cholera patients, but a subgroup analysis for cholera patients was not available for meta-analysis. There were two studies (Farugue 1996, Alam 1999) in patients with cholera excluded from this review because they were in adults. Any recommendation for rehydration treatment for adults with cholera will need to take these and any other trials found through careful systematic searching into account.
This review examines trials of children admitted to hospital who were dehydrated because of diarrhoea. The trials do not provide any direct evidence for or against the use of ORS at home to prevent dehydration developing; nor do they provide any direct evidence that reduced osmolarity ORS is more effective in preventing dehydration in home-based care in this group.
We stand by our selection of unscheduled intravenous fluid infusion rather than volume of diarrhoea as the primary outcome, as specified in the original protocol. Some specialists consider that volume of diarrhoea is more appropriate, probably because it reflects the animal and human perfusion experiments that provide part of the rationale for a reduced osmolarity ORS. Unscheduled intravenous fluid infusion is pragmatic; it provides a measure of failed oral rehydration and has implications for the healthcare resources. For these reasons, we preserved this as the primary outcome.
When we reviewed the studies for inclusion, most trials reported unscheduled intravenous fluid infusion in the details of trial implementation, where exclusions and dropouts were described. As this was identified as our primary outcome at the protocol stage, we sought out these data and presented them as the primary analysis, and it is our opinion this shows a clear effect. This highlights the value for careful attention to the protocol for a systematic review before examining the trials, and provides an illustration of how non-specialist viewpoints can actually help in obtaining practical and useful answers from meta-analysis.
We found that reduced osmolarity ORS has beneficial effects over the WHO standard ORS in reducing needs for unscheduled intravenous fluid infusion, stool output during rehydration, and the number of patients with vomiting during oral rehydration treatment. Reduced osmolarity ORS has no further risk of developing hyponatraemia as compared to the WHO standard ORS. We are currently exploring the feasibility of obtaining data on convulsions (as evidence of symptomatic hyponatraemia) for the authors of the largest trial (CHOICE 2001).
The research evidence presented here relates to the ORS used for treating children with dehydration. ORS is used much more widely for preventing dehydration developing in children with diarrhoea. While this seems appropriate, the applicability to prevention is a judgement, and highlights the need for a systematic review to examine the policies of ORS provision and ORS formula in preventing dehydration in children with diarrhoea.
Findings from this review indicate reduced osmolarity ORS is more effective than WHO standard ORS in the first line treatment of dehydration in children with diarrhoea. It is not easy to be sure however, that this finding applies to a subgroup of patients with severe diarrhoea caused by cholera, where electrolyte loss is profound. This could increase the risk of hyponatraemia, result in adverse clinical events, and attenuate the advantages of reduced osmolarity ORS.
There is the possibility that policymakers and clinicians will judge that cholera reverses the balance of benefits and harms (that is, hyponatraemia will be more common, and outweigh the advantages of reduced osmolarity solution). If this is the case, then one option is to provide WHO standard ORS for people with suspected cholera, or in areas where cholera is prevalent. This is likely to be a logistical problem in areas where diarrhoea is common and coexists with cholera. The single formula sachet aids implementation of this lifesaving intervention. Providing different formulations complicates distribution. It means health workers have a more complicated task in providing the appropriate ORS. These factors may actually impair the effective delivery of any ORS to children.
Policymakers need to be careful if they decide against change a shift to reduced osmolarity solution in areas where cholera is common because of a putative risk around hyponatraemia. If they do this, then they are obliged to prove or disprove their belief in the superiority of WHO standard ORS through a randomized controlled trial in children with clinical cholera. The WHO has convened a expert working group to consider this review and related evidence. The group recommended that ORS for treating diarrhoea in children with non-cholera diarrhoea will be enhanced by shifting to a reduced osmolarity ORS, and propose a global shift to ORS with an osmolarity of 75 mEq/L of sodium (WHO 2001).
Implications for practice
Oral rehydration solution (ORS) has saved many children's lives in low and middle income countries, and the sachets are widely used in primary care, based on standards set by the World Health Organization (WHO). This review summarized data from existing studies, and provide some evidence that dehydrated children given a solution with a lower osmolarity were less likely to need an intravenous infusion than those given WHO standard ORS. These results have important implications for policy, and WHO and UNICEF, based on this review, related data, and expert discussions, are recommending reduced osmolarity ORS be accepted as standard (WHO 2001).
Implications for research
We found insufficient data on cholera in children to make recommendations for this condition. Since cholera is a secretory diarrhoea and electrolyte loss is profound, if reduced osmolarity ORS is to be used in cholera, more trials to investigate this should be undertaken.
There is a need for a good systematic review examining the influence of policies of ORS provision in preventing dehydration and hospital admissions in children with diarrhoea.
To: Christopher Duggan; Olivier Fontaine; Sheila Bird for comments. We note that the data presented and the views expressed are entirely the responsibility of the authors.
Data and analyses
- Top of page
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Index terms
Appendix 1. Stool output
Last assessed as up-to-date: 27 November 2001.
Protocol first published: Issue 4, 2000
Review first published: Issue 2, 2001
Contributions of authors
Seokyung Hahn and Yaejean Kim wrote the protocol, conducted the data extraction, data analysis and interpretation, and drafted the paper. Paul Garner advised on inclusion criteria and outcomes for the protocol, quality assessment and analysis, and helped write the review. Paul Garner is the guarantor.
Declarations of interest
We certify that we have no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter of the review (eg employment, consultancy, stock ownership, honoraria, expert testimony).
The World Health Organization provided funds for us to conduct this review.
Sources of support
- Liverpool School of Tropical Medicine, UK.
- Medical and Pharmaceutical Statistics Research Unit, University of Reading, UK.
- Seoul National University Hospital, Korea, South.
- World Health Organization, Switzerland.
- Department for International Development, UK.
Medical Subject Headings (MeSH)
Bicarbonates; Dehydration [etiology; *therapy]; Diarrhea [*complications]; Fluid Therapy [*methods]; Glucose; Osmolar Concentration; Potassium Chloride; Rehydration Solutions [*therapeutic use]; Sodium Chloride
MeSH check words
Child, Preschool; Humans; Infant