Summary of findings
Description of the condition
Cholera is one of the most serious types of infectious diarrhoeal disease, inflicting severe social and economic hardship in outbreak areas (WHO 2001a). In 2007, the World Health Organization (WHO) noted an increase in the number of reported cholera cases and deaths, with an estimated 177,963 cases resulting in 4031 deaths across 53 countries, representing a 46% increase on the mean number of cases reported between 2002 and 2005 (WHO 2008). Due to under-reporting (in light of travel and trade-related sanction concerns) and other surveillance system limitations, reported cases are thought to represent only a fraction of actual cases (WHO 2008). Caused by ingesting food or water containing the bacterium Vibrio cholerae, cholera can spread rapidly among populations lacking access to safe water and adequate sanitation facilities. Upon infecting the small intestine, the bacteria produce a protein enterotoxin that induces the hypersecretion of water and electrolytes by the small intestinal mucosa. Symptoms of cholera include acute watery diarrhoea, vomiting, and severe dehydration, which can lead to death within 24 hours if left untreated (Sack 2004).
Description of the intervention
Oral rehydration solution (ORS) was developed in the late 1960s. It is an important intervention for reducing the morbidity and mortality associated with diarrhoeal disease, regardless of etiology (WHO 2000). ORS has been highly effective in reducing the high mortality rates experienced during cholera outbreaks, which often reached 50 per cent before the introduction of this treatment (Quotah 1999). Utilizing a simple and inexpensive solution of sodium and glucose, ORS enhances the absorption of sodium and fluid in the small intestine, even in cases of enterotoxic diarrhoea, where fluid loss is often substantial.
The former standard formulation of ORS consisted of 90 mmol/L of sodium, 20 mmol/L of potassium, 80 mmol/L of chloride, 10 mmol/L of citrate, and 111 mmol/L of glucose, with a total osmolarity of 311 mmol/L (ORS ≥ 310). Initially intended to replace sodium losses in adults with cholera, this formulation was previously recommended by the WHO and the United Nations Children's Fund (UNICEF) for treating all types of diarrhoea in children and adults (Rabbani 2000; WHO 2001b; WHO 2002). Even though the expanded use of this solution has saved millions of lives, its optimal composition remains an issue of debate (Duggan 2004; Guarino 2000; Nalin 2004).
In 2001, a Cochrane review changed the worldwide ORS formula for treating diarrhoea of all causes, reducing the total osmolarity to 245 mmol/L (ORS ≤ 270; Hahn 2002); this is currently regarded as the standard global formula.
How the intervention might work
Potential problems with the ORS ≥ 310 formulation are that it may not lower stool output or duration of diarrhoea, which reduces its acceptance in many communities (Rabbani 2000). Alternative formulations, including those that use lower electrolyte concentrations or replace glucose with complex carbohydrates such as rice powder, or both, have been introduced with the aim of reducing osmolarity in order to promote greater salt and water absorption in the small intestine.
Why it is important to do this review
ORS ≤ 270 was found to be just as safe and more effective than ORS ≥ 310 for treating diarrhoea in children (Hahn 2002). Acknowledging the benefits of ORS ≤ 270 solutions, including reduced stool output and duration of diarrhoea, WHO and UNICEF now recommend that countries use and manufacture formulations with a total osmolarity of 245 mmol/L (WHO 2001b). However, there are concerns about potential adverse effects of using ORS ≤ 270 solutions to treat people with cholera (Hahn 2002; Nalin 2004; WHO 2001b). Because cholera is associated with significant electrolyte loss, especially among children, the use of ORS with reduced sodium levels may place patients at a greater risk of developing biochemical hyponatraemia (blood sodium levels < 130 mmol/L) (Fuchs 2001). This can result in severe illness, including seizures, respiratory arrest, coma (symptomatic hyponatraemia), and even death. Especially in areas where cholera is endemic, practitioners require the best available evidence about the balance between the benefits and risks of different ORS formulations.
To compare the safety and efficacy of ORS ≤ 270 and ORS ≥ 310 for treating dehydration due to cholera.
Criteria for considering studies for this review
Types of studies
Randomized controlled trials.
Types of participants
Adults and children with acute diarrhoea caused by V. cholerae, either confirmed (by stool microscopy or stool culture) or presumed.
Types of interventions
ORS formulations with an osmolarity of ≤ 270 mOsm/L (total osmolarity of 250 mmol/L with reduced sodium).
ORS formulations with an osmolarity of ≥ 310 mOsm/L (sodium 90 mmol/L, glucose 111 mmol/L, total osmolarity 311 mmol/L).
Types of outcome measures
Need for unscheduled intravenous infusion.
Symptomatic hyponatraemia as defined by trialists (symptoms include headache, lethargy, confusion, and seizures).
Biochemical hyponatraemia as defined by trialists.
Duration of diarrhoea.
Stool output in first 24 hours after admission or randomization.
Vomiting during rehydration.
Search methods for identification of studies
We attempted to identify all relevant trials regardless of language or publication status (published, unpublished, in press, and in progress).
We searched the following databases using the search terms and strategy described in Table 1.
- Cochrane Infectious Diseases Group's Specialized Register (April 2011).
- Cochrane Central Register of Controlled Trials (CENTRAL), published in The Cochrane Library (Issue 4, 2011).
- MEDLINE (1966 to April 2011).
- EMBASE (1974 to April 2011).
- LILACS (1982 to April 2011).
We provided individuals from the following key agencies and organizations with a list of the identified trials and asked for additional completed or ongoing trials: World Health Organization; Centre for Health and Population Research (ICDDR,B); Jawaharlal Nehru Medical College; National Institute of Cholera and Enteric Diseases, and the US Naval Medical Research Unit, Jakarta.
We also checked the reference lists of all studies identified by the above methods.
Data collection and analysis
Selection of studies
Two authors independently screened the results of the search to select potentially relevant studies and applied eligibility criteria using a pre-designed eligibility form based on the inclusion criteria. Corresponding full articles were retrieved and assessed using the eligibility criteria. Each of the articles was scrutinized to ensure that multiple publications from the same trial were included only once. Where there was ambiguity, we sought clarification from the trial authors and re-assessed the articles. We resolved any differences between the eligibility results through discussion. We excluded studies that did not meet the inclusion criteria and stated the reasons in the 'Characteristics of excluded studies'.
Data extraction and management
Using a specially designed data extraction form, two authors independently extracted information on methods, participants, interventions, and outcomes for each trial. One author entered the data into Review Manager 5 and this was independently checked by AM. We scrutinized data sources for multiple publications from the same data sets, referring to the original paper where there were any differences.
We extracted the number of participants randomized in each group and the numbers analyzed for each outcome. For dichotomous data, we extracted the number of events, and for continuous data we extracted the mean and standard deviation or information to estimate the standard deviation. We contacted the publication authors in the case of unclear or missing data. We resolved any differences through discussion.
Assessment of risk of bias in included studies
Two authors (JV and AM) independently assessed the risk of bias of the included studies using the latest Cochrane Collaboration tool for assessing the risk of bias. We followed the guidance to assess whether adequate steps were taken to reduce the risk of bias across six components: sequence generation; allocation concealment; blinding (of participants, personnel and outcome assessors); incomplete outcome data; selective outcome reporting; and other sources of bias. We categorized our judgements as 'yes', 'no' or 'unclear', indicating a low, high or unclear risk of bias respectively. The results were summarized using the 'risk of bias summary' and the 'risk of bias graph', in addition to the risk of bias tables. When necessary, we contacted trial authors for clarification. We resolved any disagreements through discussion.
Measures of treatment effect
We pooled estimates of effect using risk ratios (RR) for dichotomous data and mean differences (MD) for continuous data; and presented these results with 95% confidence intervals (CI). For continuous data that were expressed in different units, we calculated standardized mean difference (SMD).
Unit of analysis issues
Trials including more than two comparison groups were split and analyzed as individual pair-wise comparisons.
Dealing with missing data
If data from the trial reports were insufficient, unclear, or missing, we contacted the trial authors for additional information or clarification. We used the 'intention to treat' principle where there were no missing data. In the case of missing dichotomous data, we still used the 'intention to treat' principle but assumed that all the missing participants did not experience the event. For missing continuous data, we used the available case analysis.
Assessment of heterogeneity
We assessed heterogeneity by visually examining the forest plot to detect overlapping confidence intervals, and used the Chi
Assessment of reporting biases
We had planned to assess the likelihood of publication bias by examining the forest plot for asymmetry if we found sufficient trials (10 or more).
We analyzed data for glucose-based and rice-based ORS ≤ 270 separately using Review Manager 5. In the absence of homogeneity of treatment effects, we used a random-effects model of meta-analysis.
Subgroup analysis and investigation of heterogeneity
We investigated clinical heterogeneity based on the age of participants by comparing children (< 11 years) with adults (> 11 years), as children may be particularly at risk of developing hyponatraemia.
We had planned to carry out a sensitivity analysis for risk of bias if we found sufficient trials, in order to investigate the robustness of the results to the quality components.
Description of studies
We identified 12 studies that appeared to meet our inclusion criteria. However, we excluded five of these studies because one did not evaluate people with cholera, two did not administer ORS ≤ 270 or ORS ≥ 310, one did not employ randomization, and one reported cholera and non-cholera data in aggregate only (see 'Characteristics of excluded studies'). We had previously noted two ongoing trials (see Bangladesh; India); these two trials have now been completed but we could not obtain their full text articles.
The seven randomized controlled trials included in the analysis were either small in size or only included a small subset of participants with cholera, producing a combined sample size of 797 participants. All trials were published in English-language biomedical journals. We have provided details of these trials in the 'Characteristics of included studies' and have summarized them below.
Participants and location
All trials, including one multicenter study trial (Choice 2001), were conducted in low-income countries: Bangladesh (Alam 1999; Choice 2001; Faruque 1996), Brazil (Choice 2001), India (Alam 1999; Alam 2000; Bhattacharya 1998; Choice 2001; Dutta 2000), Indonesia (Choice 2001; Pulungsih 2006), Peru (Choice 2001), and Vietnam (Choice 2001).
The majority of participants were adults (> 11 years), but three trials (151 participants) assessed children (< 11 years) with cholera (Alam 2000; Choice 2001; Dutta 2000). Five trials included only,or predominantly, male participants. All trial participants were suffering from a severe degree of dehydration.
While all seven trials compared glucose-based ORS ≤ 270 and ORS ≥ 310, two trials also included an experimental rice-based ORS ≤ 270 trial arm (Bhattacharya 1998; Dutta 2000); the formulations are detailed in Table 2.
Before randomization, six trials administered intravenous rehydration solutions such as Ringer's lactate, Dhaka solution, or saline solution to correct severe dehydration (Alam 1999; Bhattacharya 1998; Choice 2001; Dutta 2000; Faruque 1996; Pulungsih 2006).
Six trials treated all participants with antibiotics (Alam 1999; Alam 2000; Bhattacharya 1998; Dutta 2000; Faruque 1996; Pulungsih 2006), while the seventh trial administered an antibiotic only when an intercurrent infection occurred (Choice 2001).
Four trials reported feeding (Alam 1999; Alam 2000; Choice 2001; Pulungsih 2006). Alam 1999 gave participants bread and bananas immediately after rehydration and standard meals three times daily thereafter. Children in Alam 2000 were fed curds and bread after they had begun rehydrating and breastfeeding was continued throughout. In Choice 2001, breastfeeding was also continued ad libitum and food appropriate to age was given to children during the maintenance phase. In Pulungsih 2006, noodles were offered immediately after rehydration and meals three times a day were given throughout. The three remaining trials did not report information on feeding.
The trials assessed the following pre-specified outcomes used in this review: need for unscheduled intravenous infusion; biochemical hyponatraemia; duration of diarrhoea; stool output in first 24 hours after admission or randomization; and vomiting during rehydration. Many of these indicators were measured at different time points (eg 24 hours after study inclusion, total study time). The trials also assessed other outcomes, which are described in the 'Characteristics of included studies'.
None of the trials assessed symptomatic hyponatraemia and death as pre-specified outcomes, although the incidence of clinical signs associated with hyponatraemia was either mentioned in the manuscript text or obtained through correspondence with the authors for five trials (Alam 1999; Choice 2001; Dutta 2000; Faruque 1996; Pulungsih 2006). Information on death was reported in the manuscript text of Bhattacharya 1998 and obtained by correspondence with the Choice 2001 trial authors.
Risk of bias in included studies
|Figure 1. Risk of bias graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.|
|Figure 2. Risk of bias summary: review authors' judgements about each methodological quality item for each included study.|
Generation of allocation sequence
All trials were reported as randomized. Six trials employed permuted block randomization for generating the allocation sequence, which we judged as having a low risk of bias (Alam 1999; Bhattacharya 1998; Choice 2001; Dutta 2000; Faruque 1996; Pulungsih 2006); however, none of the trials explicitly mentioned how the sequence was generated. The remaining trial (Alam 2000) provided insufficient information to enable us to judge whether there was a high or low risk of bias.
Three trials did not describe methods used to conceal allocation and we judged them as having an unclear risk of bias (Alam 1999; Bhattacharya 1998; Dutta 2000). Two trials were judged as having a low risk of bias: one (Alam 2000) used identical packets that were given a number by a faculty colleague not involved in the study, and the other (Pulungsih 2006) used sachets that were centrally prepared and sequentially numbered according to a randomization code. The remaining two trials (Choice 2001; Faruque 1996) provided insufficient information to be able to judge them as having a low risk of bias and we therefore judged them as having an unclear risk of bias.
Three trials reported that both participants and providers were blinded to treatment assignment (Alam 2000; Choice 2001; Pulungsih 2006) by stating that the trials were double-blind and that identical packets were used. These three were therefore judged as having a low risk of bias. Two trials (Alam 1999; Faruque 1996) reported that the trials were double blind but did not describe the blinding, and they were therefore judged as having an unclear risk of bias. Blinding methods were not described in the remaining two trials (Bhattacharya 1998;Dutta 2000 ), which were therefore judged as having an unclear risk of bias.
Incomplete outcome data
All seven trials were judged as having a low risk of bias since all randomized individuals were included in the final analysis of the seven trials. None of the seven trials had disproportionate numbers of losses to follow-up between the intervention and control arms. Numbers of losses to follow-up were very small in all seven trials.
No study protocol was found for any of the seven trials and all were therefore judged as having an unclear risk of bias. There were no indications in the trial reports to suspect selective outcome reporting.
Other potential sources of bias
All seven trials were judged as having a low risk of bias as there was no reason to suggest any other potential sources of bias.
Effects of interventions
See: Summary of findings for the main comparison ORS ≤ 270 mOsm/L (glucose-based) compared to ORS ≥ 310 mOsm/L (glucose-based) for treating cholera; Summary of findings 2 ORS ≤ 270 mOsm/L (rice-based) compared to ORS ≥ 310 mOsm/L (glucose-based) for treating cholera
Comparison 1: ORS ≤ 270 (glucose-based) versus ORS ≥ 310 (glucose-based)
Of the seven trials (718 participants) that evaluated ORS ≤ 270 (glucose-based), three assessed only children (< 11 years; n = 132) and four evaluated only adults (> 11 years; n = 586).
Need for unscheduled intravenous fluid infusion:
There was a non-significant tendency towards fewer unscheduled intravenous infusions for those administered ORS ≤ 270 (glucose-based) (RR 0.86, CI 0.66 to 1.12; n = 616, five trials; Analysis 1.1).
Those receiving ORS ≤ 270 were almost 70% more likely to develop biochemical hyponatraemia (blood sodium levels < 130 mmol/L) (RR 1.67, CI 1.09 to 2.57; n = 465, four trials; Analysis 1.2).
Severe biochemical hyponatraemia
Although the point estimate for severe biochemical hyponatraemia (blood sodium levels < 125 mmol/L) was in the same direction as biochemical hyponatraemia, the result was statistically inconclusive (RR 1.58, CI 0.62 to 4.04; n = 465, four trials; Analysis 1.3).
Duration of diarrhoea
For children, we found no statistically significant difference in the duration of diarrhoea between the two groups (MD -2.75 hours, CI -9.79 to 4.29 hours, random effects model, n = 97, two trials, Analysis 1.4).
For adults, we detected substantial heterogeneity between the trials (Chi
One study (Alam 2000) assessed children and reported the geometric means and standard deviations on a log scale, and so we could not pool it with other studies in a meta-analysis. The geometric mean, standard deviation and total for the ORS ≤ 270 group were 21.44, 1.32 and 19 respectively, and the corresponding values for the ORS ≥ 310 group were 19.97, 1.99 and 16 respectively. The mean difference was significant (MD 1.47, CI 0.33 to 2.61).
Stool output in first 24 hours after admission or randomization
We found no statistically significant difference in the stool output in the first 24 hours between the two formula groups (SMD -0.13, CI -0.43 to 0.17, random-effects model; n = 581, four trials; Analysis 1.5). Results from two studies (Faruque 1996; Pulungsih 2006) were skewed (mean/SD < 2) and therefore the results of the meta-analysis may not be reliable. We detected substantial statistical heterogeneity between the trials (Chi
Vomiting during rehydration
The proportion of people that vomited during rehydration was similar in the two groups (RR 1.14, CI 0.92 to 1.40; n = 363, two trials; Analysis 1.6).
Exploring heterogeneity: children (< 11 years) versus adults (> 11 years)
Subgroup analyses assessing children and adults separately appeared to show differences in the direction of treatment effect estimates for most outcomes. However, as the overall numbers of children were small and the confidence intervals tended to include both point estimates with some degree of overlap, it is difficult to determine whether these represent true differences. While the strength of treatment benefit for the outcome 'need for unscheduled intravenous infusion' appeared to be greater in children receiving ORS ≤ 270 (glucose-based) (RR 0.57, CI 0.29 to 1.11; n = 93, two trials; Analysis 1.1) than in adults (RR 0.93, CI 0.70 to 1.24; n = 523, three trials; Analysis 1.1) , the point estimate for children has a wider confidence interval and is not statistically significant at the predefined 5% level. Biochemical hyponatraemia may be more problematic for adults receiving the ORS ≤ 270 formula (RR 1.69, CI 1.06 to 2.69; n = 465, four trials; Analysis 1.2), yet this outcome for children was assessed in only one small trial with few events, resulting in a wider confidence interval and a lack of statistical significance (RR 1.58, CI 0.53 to 4.74; n = 39; Analysis 1.2). For the two diarrhoeal outcomes where statistically significant heterogeneity was evident, heterogeneity in adults persisted in the subgroup analysis.
Comparison 2: ORS ≤ 270 (rice-based) versus ORS ≥ 310 (glucose-based)
Two trials (102 participants) evaluated ORS ≤ 270 (rice-based).
Need for unscheduled intravenous fluid infusion
No information available.
No instances of symptomatic hyponatraemia were reported in the one trial that assessed this outcome (Dutta 2000).
While the point estimates suggest a reduced risk of biochemical hyponatraemia (blood sodium levels < 130 mmol/L) for those receiving ORS ≤ 270 (rice-based), these findings are not statistically significant (RR 0.66, CI 0.26 to 1.69; n = 102, two trials; Analysis 2.1).
Severe biochemical hyponatraemia (blood sodium levels < 125 mmol/L) was similar between the two groups; the confidence interval around the risk ratio is wide, reflecting the small number of events for this outcome (RR 0.35, CI 0.02 to 8.10; n = 102, two trials; Analysis 2.2).
Duration of diarrhoea
There was a statistically significant reduction in the duration of diarrhoea for those receiving the ORS ≤ 270 (rice-based) formula (MD -11.42 hours, CI -13.80 to -9.04; n = 102, two trials; Analysis 2.3).
None of the trials evaluated or reported on the other outcomes of interest.
Exploring heterogeneity: children (< 11 years) versus adults (> 11 years)
One small trial was available for each subgroup for three outcomes of interest. For biochemical hyponatraemia, there was no statistically significant heterogeneity between the two subgroups (Chi
Our review draws attention to the paucity of evidence on the effects of ORS ≤ 270 (glucose-based) compared with ORS ≥ 310 for treating people with cholera, with only seven trials evaluating 718 participants.
We intended to examine the safety of glucose-based ORS ≤ 270 for cholera by measuring the incidence of symptomatic hyponatraemia, as low blood sodium levels may be transient and therefore not necessarily result in serious illness. The 2001 WHO/UNICEF meeting of ORS formulation experts highlighted the importance of this outcome in people receiving treatment for cholera (WHO 2001b). An observational study found that the risk of symptoms associated with hyponatraemia in patients treated with ORS ≤ 270 was minimal and did not increase with the change in formulation (Alam 2006). As none of the trials found or explicitly evaluated symptomatic hyponatraemia, we could not assess this outcome. Instead, we measured the incidence of biochemical hyponatraemia. Asymptomatic hyponatraemia, while not providing a definitive marker for treatment failure, provides an important measure of potential risk for people with cholera.
We found that participants receiving ORS ≤ 270 were at greater risk of developing biochemical hyponatraemia (blood sodium levels < 130 mmol/L); however, the relatively few cases of severe biochemical hyponatraemia (blood sodium levels < 125 mmol/L) precludes firm conclusions regarding this outcome. These findings should, nevertheless, alert clinicians to the need for vigilance concerning the risk of hyponatraemia in non-trial settings.
For other outcomes, such as unscheduled intravenous infusion, stool output, vomiting, and duration of diarrhoea, there was little difference in effect between the two types of formulae. However, as most of the available trials are small with few events, they may have insufficient power to demonstrate important clinical differences even after pooling the results.
In separate analyses of two trials (102 participants) comparing ORS ≤ 270 (rice-based) with ORS ≥ 310 (glucose-based), we found no statistically significant differences except for the duration of diarrhoea, which was substantially shorter in the group receiving ORS ≤ 270 (rice-based). A similar finding was reported in a systematic review that compared rice-based ORS with glucose-based ORS ≥ 310 formulas in people with diarrhoea (Fontaine 1998).
Data available for assessing the safety and efficacy ORS ≤ 270 in children with cholera are also extremely limited, making it difficult to draw firm conclusions. Only one trial reported the risk of biochemical hyponatraemia in children (Dutta 2000) and it showed a trend favouring the ORS ≥ 310 formula, but this finding was not statistically significant.
WHO and UNICEF currently recommend formulations with a total osmolarity of 245 mmol/L for treating diarrhoea. It is not known, however, whether using ORS ≤ 270 is appropriate in cholera-endemic regions, where the balance between benefit and harm can be tenuous. Logistically, having one ORS formula is easier. However, the increased risk of biochemical hyponatraemia in those receiving ORS ≤ 270 solutions is of concern. Even though there were no instances of symptomatic hyponatraemia or death, the total patient experience in the trials is very small and these effects cannot be ruled out under wider practice conditions. Moreover, careful monitoring of blood sodium levels may be difficult in areas where healthcare resources are limited, especially during complex emergencies and large epidemics. Further trials in both adults and children with cholera should be undertaken to clarify these issues.
Quality of evidence
The quality of evidence was assessed using the GRADE methodology. Overall the quality is moderate, meaning that further research is likely to have an important impact on our confidence in the estimate of effect, and may change the estimate. See Summary of findings for the main comparison and Summary of findings 2.
Implications for practice
In people with cholera, ORS ≤ 270 results in more patients developing biochemical hyponatraemia, with no detectable benefits such as the need for unscheduled intravenous infusion, duration of diarrhoea, or stool volume, compared with the older ORS ≥ 310 formula. The increased risk of low blood sodium levels could have major implications in resource-constrained settings where clinicians may not have monitoring facilities and must rely on presumptive diagnosis. This review found no serious clinical consequences related to hyponatraemia in trial participants, but it is important to note that total patient experience in the existing trials is small.
WHO and UNICEF currently recommend an ORS ≤ 270 formulation for treating dehydration caused by all types of diarrhoea. While it may be easier to administer a single ORS formulation worldwide, the potential harms and limited evidence of improved efficacy for people with cholera should be kept in mind.
Implications for research
Further randomized controlled trials are needed to assess the balance between benefit and harm associated with the use of ORS ≤ 270 in people with cholera. These trials should be large enough to adequately assess important outcomes, including symptomatic hyponatraemia and death.
The protocol for this review was developed during the Mentorship Programme organized by the Cochrane Infectious Diseases Group in November 2001. The UK Department for International Development supports this Programme through the Effective Health Care Alliance Programme at the Liverpool School of Tropical Medicine.
We would like to thank Colleen Murphy who first initiated the review, and was the guarantor on the first published version. Additionally we would like to acknowledge Dr Seokyung Hahn, who contributed to the original review design and analysis.
We wish to thank Professor Paul Garner for his guidance in developing the protocol. We are also grateful to trial authors who provided us with clarifications and unpublished data.
Data and analyses
- Top of page
- Summary of findings [Explanations]
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Differences between protocol and review
- Index terms
Last assessed as up-to-date: 30 September 2011.
Protocol first published: Issue 3, 2002
Review first published: Issue 4, 2004
Contributions of authors
Colleen Murphy (CM) initiated the review and developed the eligibility and data extraction forms, with Seokyung Hahn (SH) and Jimmy Volmink (JV) providing input. CM and JV selected the trials for inclusion in the initial version of the review. CM, JV and Alfred Musekiwa (AM) extracted the data and assessed trial quality, and CM contacted authors for additional information for the first published version of the review. CM and AM entered the data and conducted the analysis. CM wrote the first draft of the review, with all reviewers contributing to the final text and analysis. AM responded to editor's comments and drafted this review update in line with RevMan 5.
Declarations of interest
Sources of support
- South African Medical Research Council, South Africa.
- Stellenbosch University, South Africa.
- Wits Reproductive Health and HIV Institute (WRHI), South Africa.
- Department for International Development, UK.
Differences between protocol and review
The title has been changed to 'Oral rehydration solution for treating cholera: ORS ≤ 270 mOsm/L solutions vs ORS ≥ 310 mOsm/L solutions'. The secondary outcome stool volume was changed to stool output to accommodate some stool weight measurements as reported by other trials.
We had planned the following analyses but they were not appropriate for the data available: (1) analysis of geometric means and standard deviation using log normal approximation; (2) analysis of time-to-event or censored data, when available, to estimate the log hazards ratio and its variance within each trial, using methods proposed by Parmar 1998; (3) examination of funnel plots for asymmetry indicative of publication bias; and (4) sensitivity analysis to determine the degree to which the results were influenced by the adequacy of allocation concealment.
Medical Subject Headings (MeSH)
Cholera [*complications]; Dehydration [*therapy]; Diarrhea [*complications]; Glucose; Hyponatremia [*etiology]; Osmolar Concentration; Randomized Controlled Trials as Topic; Rehydration Solutions [adverse effects; chemistry; *therapeutic use]
MeSH check words
Adult; Child; Humans
* Indicates the major publication for the study