PEDIATRICS Vol. 107 No. 4 April 2001, pp. 613-618

Multicenter, Randomized, Double-Blind Clinical Trial to Evaluate the Efficacy and Safety of a Reduced Osmolarity Oral Rehydration Salts Solution in Children With Acute Watery Diarrhea

CHOICE Study Group^*

From the Choice Study Group.

	ABSTRACT

Top Abstract Methods Results Discussion Conclusion References

Objective.  To compare the efficacy of a reduced osmolarity oral rehydration salts (ORS) solution (75 mmol/L of sodium [Na], 20 mmol/L of potassium [K], 65 mmol/L of chloride, 10 mmol/L of citrate, and 75 mmol/L of glucose; osmolarity, 245 mosm/L) with that of the standard World Health Organization (WHO) ORS solution.

Design.  A multicenter, double-blind, randomized, controlled clinical trial conducted in children with acute diarrhea in 5 developing countries to measure mean stool output in the 24 hours after randomization, proportion of children who required unscheduled intravenous therapy, proportion of children who vomited in the first 24 hours, and diarrhea duration after randomization.

Results.  A total of 675 children who ranged in age from 1 to 24 months and who had acute diarrhea and dehydration were enrolled in the trial; 341 were randomized to receive reduced osmolarity ORS solution, and 334 were randomized to receive the WHO ORS solution. The mean (SE) stool output (g/kg) in the first 24 hours (reduced osmolarity ORS solution vs WHO ORS solution = 114 [4] vs 125 [5]) and during the total study period (reduced osmolarity ORS solution vs WHO ORS solution = 320 [18] vs 331 [18]) were comparable. The proportion of children who vomited in the first 24 hours (reduced osmolarity ORS solution vs WHO ORS solution = 58% vs 62%) and the diarrhea duration in the 2 treatment groups, compared by log rank test, were similar. The proportion of children who required unscheduled intravenous therapy was significantly lower in children who received reduced osmolarity ORS solution (10%) as compared with those who received the WHO ORS solution (15%; odds ratio = 0.6, 95% confidence interval = 0.4-1.0). There was no significant difference in the incidence of hyponatremia (serum Na <130 mmol/L) at 24 hours between the 2 treatment groups (11% in reduced osmolarity ORS solution group vs 9% in the WHO ORS solution group; odds ratio = 1.3; 95% confidence interval = 0.8-2.2). The frequency of patients with serum Na <125 mmol/L at 24 hours was 13 of 341 (4%) in children who were treated with reduced osmolarity ORS solution versus 7 of 334 (2%) in children who received the WHO ORS solution.

Conclusions.  Treatment with reduced osmolarity ORS solution was associated with a 33% reduction in the need for unscheduled intravenous therapy and had no apparent effect on stool output and illness duration when compared with treatment with the standard WHO ORS solution. Children with acute diarrhea, therefore, may benefit from a reduced osmolarity ORS solution. The results of trials that examine the efficacy and safety of reduced osmolarity ORS solution in adult patients with cholera have to be taken into consideration before consensus on composition of oral rehydration formulation can be reached.  Key words:  acute diarrhea, oral rehydration solution, dehydration, low osmolarity solutions, parenteral therapy.

For 20 years, the World Health Organization (WHO) and the United Nations Children's Fund have recommended that all patients with dehydration as a result of acute diarrhea of any cause be treated with an oral rehydration salts (ORS) solution containing 90 mmol/L of sodium (Na) and 111 mmol/L of glucose (total osmolarity, 311 mosm/L) provided they are able to drink and do not have severe dehydration.¹ The extensive use of the WHO ORS solution in the Diarrheal Disease Control Program has been credited with saving millions of lives from diarrheal dehydration in developing countries.²

There is some disagreement, however, about the ideal composition of ORS solution, especially in regard to Na concentration and total osmolarity. On the basis of experiments in animals and humans, the European Society of Paediatric Gastroenterology and Nutrition has recommended the use of ORS solutions that contain 60 mmol/L of Na for children in developed countries because of concern about the risk of hypernatremia when solutions with higher Na concentrations are given to well-nourished children.³ It also recommended that osmolarity be maintained between 200 and 250 mosm/L, because water absorption from such solutions is better than from solutions that are isotonic with plasma.^4-7

Several clinical trials have been conducted recently in developing and developed countries using solutions that contained 60 to 75 mmol/L of Na and 75 to 90 mmol/L of glucose and the osmolarity of which ranged from 225 to 250 mosm/L in children with acute noncholera diarrhea.^8-14 Results from 5 controlled, double-blind trials that were conducted in developing countries^10-14 were reviewed in a Consultative Meeting on ORS Formulation, which was held in Dhaka, Bangladesh, in December 1994. In these 5 studies that involved 734 children who had acute watery diarrhea,^9-13 the use of a reduced osmolarity ORS solution resulted in clinically significant reductions in 24-hour stool output (25%) and in the need for supplemental intravenous therapy (33%) when compared with treatment with standard WHO ORS solution.¹⁴ The group of experts who gathered in Dhaka concluded that although the magnitude of reduction in the stool output and unscheduled intravenous therapy in the large multicenter trial (60 mmol/L of Na, 90 mmol/L of glucose; total osmolarity, 224 mosm/L) was clinically important,¹¹ the safety and efficacy of reduced osmolarity ORS solution in childhood diarrhea should be more precisely defined in a large trial using a single consensus formulation.

They recommended that a solution of reduced osmolarity containing 75 mmol/L of Na, 20 mmol/L of potassium (K), 65 mmol/L of Cl, 10 mmol/L of citrate, and 75 mmol/L of glucose and having an osmolarity of 245 mosm/L be evaluated. This consensus formulation was selected because its Na content is only modestly less than that in the standard WHO ORS solution, which may be important for treatment of cholera, and its glucose content is at least equal to that of Na. In addition, 1 study in children with acute diarrhea¹³ suggested that this solution was more efficacious than the standard WHO ORS solution.

Therefore, we designed a multicenter, randomized, double-blind clinical trial in children with acute watery diarrhea to evaluate the safety and efficacy of a reduced osmolarity ORS solution (as described above) versus the standard WHO ORS solution on stool output, duration of diarrhea, frequency of vomiting, and need for unscheduled intravenous therapy. The expert panel also recommended that the safety and efficacy of this reduced osmolarity ORS solution be tested in adult patients with cholera. Results of this study have been published recently.¹⁵

	METHODS

Top Abstract Methods Results Discussion Conclusion References

The study was conducted simultaneously at 5 research centers: 1) International Center for Diarrheal Diseases Research, Bangladesh, Dhaka, Bangladesh; 2) Centro Pediatrico Professor Hosannah de Oliveira-Universidade Federal da Bahia, Salvador, Brazil; 3) ICMR Advanced Center for Diarrhea Diseases Research, All India Institute of Medical Sciences and Kasturba Hospital, New Delhi, India; 4) Hospital Nacional Cayetano Heredia, Lima, Peru; and 5) Children's Hospital No. 1, Ho Chi Minh City, Vietnam. Study enrollment began in June 1995 and ended in February 1997. The ethics committees of each of the centers, as well as that of the WHO and the Harvard University School of Public Health, approved the protocol. Written informed consent was obtained from the parents or legal guardians of all participants.

Male children who met the following inclusion criteria were enrolled: age between 1 and 24 months, diarrhea for <72 hours (with passage of 3 or more watery stools in the 24 hours before admission), and signs of some or severe dehydration.¹⁶ Only male children were included because of the need to separate urine from stool using urine bags. Children were excluded if they had bloody diarrhea, clinical signs of systemic infection that required intravenous antibiotic therapy, or severe malnutrition defined as admission weight for height <65% of the National Center for Health Statistics standard (to account for rehydration), or the presence of obvious edema.

Randomization

Children were randomly assigned to receive either reduced osmolarity ORS solution or the WHO ORS solution. Randomization was stratified by site and age group (1-6 months; 7-24 months). The randomization list and numbered ORS packets (10 per child, each packet to be dissolved in 1 L of clean water) were prepared at the WHO (Geneva, Switzerland), using permuted blocks of variable length. Lengths 6, 8, and 10 were used to randomize the first 50 patients per stratum and per site, and lengths 4 or 6 were used for the remaining patients. The 2 ORS preparations were similar in appearance and packaged in identical polyethylene bags. Randomization took place immediately after the patient had been confirmed to be eligible to enter the study: for patients who were admitted with signs of severe dehydration, just after rapid intravenous fluid infusion had been completed, and for patients with some dehydration, on completion of the physical examination.

Case Management

The children were assessed and their cases were managed in accordance with WHO guidelines¹⁵ in which treatment is based on the assessment of dehydration status as none, some, or severe. Children who presented with severe dehydration were rehydrated with intravenous fluids at a rate of 40 mL/kg/hour for 2 hours and were randomized as soon as they were able to take rehydration fluids by mouth. Children who had some dehydration were offered 100 mL/kg of the assigned ORS solution over 6 hours with an interim evaluation at 3 hours. During rehydration, breastfeeding was continued ad libitum, but other foods and water were not offered. Ongoing losses of vomit and stool were replaced with the assigned ORS solution on a volume-to-weight basis. After rehydration was complete and clinical signs of dehydration were no longer present, the maintenance phase was initiated. During this phase, ongoing stool and vomit losses were replaced on a volume-to-weight basis with the assigned ORS solution until diarrhea ceased, and feeding appropriate to age was given. Cessation of diarrhea was defined as the time of passage of the last unformed stool before the passage of 2 consecutive formed stools, or no stool passed during a 12-hour period.

Exclusively breastfed infants continued breastfeeding. Partially breastfed infants (<5 breastfeedings per day) were offered cow's milk formula at 55 kcal/kg per day (younger than 4 months) or 75 kcal/kg per day (5-24 months) in addition to continued breastfeeding. Nonbreastfed infants who were younger than 4 months were offered at least 110 kcal/kg of cow's milk formula. In addition, cereal-based semisolid feeds made from locally available foods were offered to children who were older than 4 months and who were already taking such foods before entering the study, in amounts standardized for body weight. Plain water was offered ad libitum during the maintenance phase. Children who developed intercurrent infections after enrollment were treated with antibiotics and remained in the study.

Stool was collected in preweighed diapers or via metabolic beds and measured on scales that were accurate to 1 g. Urine was collected in urine bags, then measured in graduated cylinders. Vomitus was collected in preweighed pads. Nude body weight on admission, at the end of the rehydration phase, daily, and at discharge was measured on scales that were sensitive to 10 g. Body length was measured on admission with a length board to the nearest 0.1 cm. Intake and output data were recorded. Standard techniques were used for measurement of serum Na, serum K, hematocrit, and serum specific gravity at randomization and 24 hours later. The presence of rotavirus in stool was assessed using commercial enzyme-linked immunosorbent assay kits. In countries where cholera is endemic (all except Brazil and Vietnam), Vibrio cholerae was detected by stool culture.

Criteria for unscheduled intravenous therapy were the following: 1) during the rehydration phase, worsening or no improvement of hydration status at 6 hours despite appropriate ORS solution administration, or 2) during the maintenance phase: a) appearance of signs of severe dehydration despite appropriate ORS solution administration or b) if signs of some dehydration occurred and could not be corrected with appropriate ORS solution administration. When a child had 2 episodes of dehydration during the maintenance phase, ORS solution administration was discontinued and intravenous therapy was used for both maintenance and replacement of excess losses until cessation of diarrhea. Data were collected up until the time of the first intravenous therapy, at which point data were censored. The Na concentration of the intravenous solutions used ranged from 90 to 133 mmol/L.

Monitoring and Quality Assurance

All principal investigators attended a protocol development workshop in March 1995 where case report forms and a specific form for summarizing unscheduled intravenous administration were developed and finalized (see list of participants at the end of this article in the "Appendix"). The assessment of dehydration resulted from an observer variation study conducted during this workshop in which 10 experienced pediatricians independently assessed 12 patients. Only signs for which there was observer agreement were retained in final scoring. To ensure that the protocol was being followed similarly across sites, 3 monitors made visits to each center at initiation of patient enrollment, midway through recruitment, and before the data analysis workshop. For intermonitor calibration, a third of these visits were made by a pair of monitors; the remainder were made by a single monitor.

The Data Safety Monitoring Board (DSMB), composed of 1 statistician and 2 physicians who were not involved with the study, had a scheduled meeting midway through the trial and an earlier ad hoc meeting to address concerns raised by a higher-than-expected frequency of hyponatremia among the first few patients admitted to the study. The DSMB reviewed the treatment allocation for the relevant cases without the treatments being fully identified. After this review, the DSMB members concluded that they did not have sufficient grounds to unblind the study or to recommend modification in the conduct of the study.

Before enrollment, each principal investigator (PI) ensured that all study personnel were trained and procedures were standardized. Scale calibration was performed regularly. Observer variation exercises, such as those in which PIs had participated in Dhaka, were repeated at each site every 4 to 6 weeks to maintain uniform performance of study personnel within each center.

Sample Size

The primary outcomes were stool output in g/kg in the first 24 hours and duration of diarrhea. In previous studies, stool output and duration of diarrhea were observed to be reduced by 25% and 20%, respectively, in children who were treated with reduced osmolarity and alternative ORS formulations.^9-14 Standard deviation of stool output in the first 24 hours in patients who were treated with the WHO ORS solution is approximately equal to the mean. Therefore, 80% power to detect a plausible 25% reduction in mean stool output in the first 24 hours required randomization of 250 patients per treatment group. Mean and standard deviation for duration of diarrhea were assumed to be 40 and 25 hours, but because we planned to use life table analysis, a sample size calculation based on these data was only indirectly applicable and indicated a smaller sample size than was required to assess stool output. The percentage of patients who vomited in the first 24 hours was a secondary outcome. Review of studies on reduced osmolarity ORS solution suggested a one-third reduction (from 30% to 20%) in percentage of children who vomited in the first 24 hours. A minimum of 300 children per treatment group was required to detect this reduction with 80% power. Therefore, randomization of at least 600 children was the study's goal. However, the sample size calculated for this multicenter study did not allow precise determination of the impact of reduced osmolarity ORS solution on unscheduled intravenous infusion. Review of randomized studies suggested that reduced osmolarity ORS solution decreased by one third the need for unscheduled intravenous infusion. With 15% of the 1- to 24-month-old children requiring unscheduled intravenous infusion when treated with the WHO ORS solution, randomization of approximately 700 patients per treatment group would have been needed to detect a one-third reduction in the need for unscheduled intravenous infusion with 80% power. However, randomization of approximately 300 patients per treatment group, which is what was achieved with this trial, gave approximately 50% power with respect to a one-third reduction and over 80% power with respect to a 50% reduction in the need for unscheduled intravenous infusion.

Statistical Methods

Analysis of the data collected in this multicenter study was performed during a Data Analysis Workshop held in Ho Chi Minh City, Vietnam, in June 1997 (see list of participants in the "Appendix"). Comparison of stool output in the first 24 hours was made with and without adjustment for design effect (center and age stratum effects) and patient characteristics (cause, clinical characteristics on admission, nutritional status, and biochemistry). All analyses were done without transformation and after logarithmic and square root transformation, the last of which gave normally distributed residuals. Time to cessation of diarrhea was compared by log rank test, and adjustment for covariates was made via Cox proportional hazards regression.

	RESULTS

Top Abstract Methods Results Discussion Conclusion References

A total of 676 children were enrolled in the trial; 341 were randomized to receive the reduced osmolarity ORS solution, and 335 were randomized to receive the WHO ORS solution. One child was randomized for a second episode of diarrhea 7 months after his first recruitment; the data from his second episode were not included in the final analysis. A total of 125 children were withdrawn from the study before cessation of diarrhea; 63 received the reduced osmolarity ORS solution and 62 received the WHO ORS solution. Reasons for withdrawal were as follows: 1) 31 children had diarrhea for >7 days after randomization (18 in reduced osmolarity ORS solution group vs 13 in the standard WHO ORS solution group), 2) 20 consumed all 10 L of their assigned ORS solution before cessation of diarrhea (9 vs 11), 3) 17 required special care because of intercurrent illnesses and were transferred to the intensive care unit (7 vs 10), 4) 33 withdrew informed consent (20 vs 13), and 5) 24 children had 2 episodes of dehydration (9 vs 15) and therefore received intravenous therapy for maintenance rehydration as part of the standard local treatment policy. Data collected on all such children up to the time of withdrawal were included in the analyses. The admission characteristics of the children who were withdrawn were not different from the characteristics of those who continued in the study (data not shown).

Admission characteristics were well-matched between the 2 randomized groups (Table 1). Major center differences are evident: the children in India were less well-nourished, those in Bangladesh had more severe dehydration on presentation, those in Brazil and Vietnam were less frequently breastfed, and those in Peru had received antibiotics more frequently before enrollment. There was also no difference on admission between treatment groups in serum Na (137 mEq/L [7] in both treatment groups), in serum K (4.0 mEq/L [0.8] vs 3.9 mEq/L [0.8]) and in hematocrit (36% [5] in both treatment groups). When these and other patient characteristics were considered in regression analysis, center differences in outcome (eg, mean stool output in the first 24 hours and use of unscheduled intravenous fluids) were largely explained (data not shown).

Total and the initial 24-hour stool output (g/kg) were comparable between the two treatment groups (P = .65 and P = .08, respectively; Table 2). The proportions of children who vomited in the first 24 hours also were comparable (198 vs 807; P = .29). ORS solution intake in the initial 24 hours was lower in children who received the reduced osmolarity ORS solution than in those who were treated with the WHO ORS solution (P = .03); however, the total intake during the entire study was comparable between treatment groups (P = .63; Table 2). Diarrhea duration, compared by log rank test, was similar in the 2 groups (P = .56). Significantly fewer children who were treated with reduced osmolarity ORS solution required unscheduled intravenous therapy (34 of 341 [10%]) than those who were treated with the WHO ORS solution (50 of 334 [15%]; P = .052). The mean (SE) 24-hour stool output in patients who required unscheduled intravenous therapy was similar in both treatment groups (189 [13] g/kg/24 hours vs 205 [13] g/kg/24 hours) and approximately twice as large as the stool output of children who did not require unscheduled intravenous therapy (107 [4] g/kg/24 hours vs 111 [4] g/kg/24 hours). This clearly shows that patients who required unscheduled intravenous therapy had a significantly higher stool output than the children who did not require intravenous therapy.

The mean (SE) serum Na concentration at 24 hours was not different between groups (136 [0.4] mEq/L vs 137 [0.4] mEq/L). There was no significant difference in the incidence of hyponatremia (serum Na <130 mmol/L) at 24 hours between the 2 treatment groups (11% in reduced osmolarity ORS solution group vs 9% in the WHO ORS solution group; P = .33). The frequency of serum Na <125 mmol/L at 24 hours was 13 of 341 (4%) in children who were treated with reduced osmolarity ORS solution versus 7 of 334 (2%) in those who received the WHO ORS solution (P = .18). Only one hyponatremic child (reduced osmolarity ORS solution) had a brief generalized seizure, but all others were asymptomatic.

For the duration of the study, intake of milk, semisolid food, and water was comparable between treatment groups. Mean (SE) weight gain at 6 hours was 3.1% (0.1) for the reduced osmolarity ORS solution group and 2.7% (0.2) for the WHO ORS solution group (P = .03). At discharge, mean (SE) weight gains were 4.0% (0.2) and 4.2% (0.2), respectively (P = .64). Total urine output was also similar between the groups (P = .17).

Multifactorial analysis was used for adjusted comparison between ORS solutions for the following outcome variables: stool output at 24 hours and use of unscheduled intravenous therapy. There were no significant changes in treatment effects after adjusting for the covariates.

	DISCUSSION

Top Abstract Methods Results Discussion Conclusion References

In this large multicenter clinical trial of 675 children, we found no significant difference in stool output or duration of illness between patients who received reduced osmolarity ORS solution and patients who were treated with the standard WHO ORS solution. There was also no significant difference in the frequency of hyponatremia between treatment groups. The results of this study differ from previous trials that showed a significant reduction of stool output and duration of diarrhea with the use of reduced osmolarity ORS solution. Of the previously published trials in developing countries, at least two^11,13 had a sufficiently large trial size to obtain a precise estimate of impact. In these 2 earlier studies and in the current one, the osmolarity of the reduced osmolarity ORS formulations tested ranged between 224 and 245 mosm/L. In all of these trials, there was a substantial decrease in the need for unscheduled intravenous therapy with the use of reduced osmolarity ORS solution. The impact on stool output and illness duration was, however, variable, and it was maximal with the ORS formulation that had an osmolarity of 224 mosm/L. The compositions of the ORS formulation in Santosham et al's study¹³ and the current study were identical (75 mmol/L of Na and 75 mmol/L of glucose; osmolarity, 245 mosm/L), yet the 18% reduction in the initial stool output associated with reduced osmolarity ORS in the former study was not seen in this current multicenter trial. However, the reduction in stool output seen in Santosham et al's study probably was clinically not significant as it represented only a reduction of 1 to 2 excess bowel movements over the entire illness.

There was, however, a 33% reduction in the need for unscheduled intravenous therapy in the intervention group, which is consistent with previously published data.^8-13 The finding that children who were given standard WHO ORS solution had a significantly greater need for unscheduled intravenous therapy than those who were given reduced osmolarity ORS solution indicates a reduction in severity of the disease in children who were given reduced osmolarity ORS solution, as found in other studies.^9,17 Because the reduction in the need for intravenous therapy occurred primarily in children who had high stool output, transient glucose malabsorption among these children may be the reason behind the better outcomes. However, these children were relatively few in number, and this may explain the absence of a statistically significant difference in stool volume between the 2 treatment groups.

Some limitations of the study need to be noted. The present study was designed with 80% power to detect 25% differences in stool output between treatment groups. Smaller differences, if present, would not necessarily have been detected, but the clinical significance of these would be debatable. The study was not designed to detect differences in the incidence of hyponatremia. Therefore, it also should be noted that although the incidence of hyponatremia was not significantly different between groups (odds ratio = 1.3; 95% confidence interval = 0.8-2.2), a doubling of this incidence cannot be excluded by these data.

	CONCLUSION

Top Abstract Methods Results Discussion Conclusion References

For years, the WHO has encouraged the use of a single ORS formulation to be used for all ages and all causes of infectious diarrhea, and the effectiveness of this approach has been enormous. In this large study, we found that the use of a reduced osmolarity solution (with 75 mmol/L of Na and glucose and an osmolarity of 245 mosm/L) was associated with a 33% reduction in the need for unscheduled intravenous therapy and had no detectable effect on stool output and illness duration. In contrast, we recently reported that the same solution was equally as effective as the standard WHO ORS solution among adult cholera patients but that the incidence of asymptomatic hyponatremia was higher in those who received the reduced osmolarity ORS solution.¹⁴ If one single ORS formulation is to be used for both cholera and noncholera diarrhea, then the optimal composition of this formulation should be balanced against any added risks of hyponatremia.

	APPENDIX: CHOICE STUDY GROUP MEMBERS

N. H. Alam*, Bangladesh; S. Bhatnagar§, India; E. Chea-Woo*, Peru; C. Duggan*, United States; O. Fontaine*, Switzerland; G. J. Fuchs, Bangladesh; S. M. Gore*, Great Britain; G. Keusch* United States; D. T. Ly, Vietnam; D. Mahalanabis*, Bangladesh; R. N. Majumder*, Bangladesh; A. Mattos, Brazil; H. Ribeiro*, Brazil; M. Santosham*, United States; K. D. Singh*, India; P. N. Thanh*, Vietnam.

The asterisk symbol (*) indicates that this individual participated in the proposal development workshop; the double dagger symbol () indicates that this individual participated in the data analysis workshop and in the writing of the first draft of the papers and commented on subsequent drafts; the section symbol (§) indicates that this individual was responsible for writing the final draft.

Composition of the Teams That Participated in the Two Multicenter Trials:

International Centre for Diarrheal Diseases Research, Bangladesh (ICDDR,B), Dhaka (children trial): R. N. Majumder (PI), N. H. Alam, G. J. Fuchs.

Centro Pediatrico Professor Hosannah de Oliveira-Universidade Federal da Bahia, Salvador, Brazil: H. Ribeiro, Jr (PI), T. Ribeiro, A. Mattos, T. Guimaraes.

ICMR Advanced Center for Diarrhea Diseases Research, All India Institute of Medical Sciences and Kasturba Hospital, New Delhi, India: S. Bhatnagar (PI), M. K. Bhan, K. D. Singh, S. K. Saxena.

Hospital Nacional Cayetano Heredia, Lima, Peru: E. Chea-Woo (PI), E. Salazar-Lindo, Y. Alegre-Palomino.

Children's Hospital No. 1, Ho Chi Minh City, Vietnam: P. N. Thanh (PI), D. T. Ly, N. T. Ly.

Consultants: C. Duggan, ARCH Project, Boston (monitor); O. Fontaine, WHO, Geneva (monitor); S. M. Gore, MRC Biostatistics Unit, Cambridge (statistician); M. Santosham, Johns Hopkins University, Baltimore (monitor).

	ACKNOWLEDGMENTS

The two studies were funded by the Division of Child Health and Development (CHD) of the World Health Organization, Geneva; the Applied Research on Child Health (ARCH) Project from the Harvard Institute for International Development (HIID), Boston; and the United Nations Children's Fund (UNICEF), New York.

We thank the medical and nursing staffs of the five centers for their help, as well as the parents and patients. We thank the Data Safety and Monitoring Board (Dr R. B. Sack, Dr L. Moulton, and Dr R. Kleinman) for their assistance, as well as Dr N. F. Pierce for his advice in the design of the study and the writing of the results.

	FOOTNOTES

^* Members of the CHOICE Study Group and the composition of the teams that performed the studies are listed in the "Appendix."

Received for publication Mar 3, 2000; accepted Aug 25, 2000.

Reprint requests to Shinjini Bhatnagar, MD, Centre for Diarrheal Diseases and Nutrition Research, Department of Paediatrics, All India Institute of Medical Sciences (AIIMS), Ansari Nagar, New Delhi-110 029, India.

	ABBREVIATIONS

WHO, World Health Organization; ORS, oral rehydration salts; Na, sodium; K, potassium; DSMB, Data Safety Monitoring Board; PI, principal investigator.

	REFERENCES

Top Abstract Methods Results Discussion Conclusion References

1. Hirschhorn N The treatment of acute diarrhea in children: an historical and physiological perspective. Am J Clin Nutr 1980; 33:637-663 [Abstract/Free Full Text]
2. Rahaman MM, Aziz KMS, Patwari Y, Munshi MH Diarrhoeal mortality in two Bangladeshi villages with and without community-based oral rehydration therapy. Lancet 1979; 314:908-912
3. Report of an ESPGAN Working Group Recommendations for composition of oral rehydration solutions for the children of Europe. J Pediatr Gastroenterol Nutr 1992; 14:113-115 [Medline]
4. Lifshitz F, Wapnir RA Oral rehydration solutions: experimental optimization of water and sodium absorption. J Pediatr 1985; 106:383-389 [CrossRef][Medline]
5. Sandhu BK, Cristobal FL, Brueton MJ Optimizing oral rehydration composition in model systems: studies in normal mammalian small intestine. Acta Paediatr Scand 1989; 364:17-22
6. Elliot EJ The role of human perfusion techniques in the assessment of oral rehydration solutions. Acta Paediatr Scand 1989; 364:31-39
7. Hunt JB, Thillainayagam AV, Salim AFM, Carnaby S, Elliot EJ, Farthing MJG Water and solute absorption from a new hypotonic rehydration solution: evaluation in human and animal perfusion models. Gut 1992; 33:1652-1659 [Abstract/Free Full Text]
8. Rautanen T, El-Radhi S, Vesikari T Clinical experience with hypotonic oral rehydration solution in acute diarrhoea. Acta Paediatr 1993; 82:52-54 [Medline]
9. El-Mougi M, El-Akkad N, Hendawi A, Is a low osmolarity ORS solution more efficacious than standard WHO ORS solution? J Pediatr Gastroenterol Nutr 1994; 19:83-86 [Medline]
10. Mahalanabis D, Faruque ASG, Hoque SS, Faruque SM Hypotonic oral rehydration solution in acute diarrhoea: a controlled clinical trial. Acta Paediatr 1995; 84:289-293 [Medline]
11. International Study Group on Reduced Osmolarity ORS Solution Multicentre evaluation of reduced osmolarity oral rehydration salts solution. Lancet 1995; 345:282-285 [Medline]
12. Sarker SA, Majid N, Mahalanabis D Alanine and glucose-based hypo-osmolar oral rehydration solution in infants with persistent diarrhoea: a controlled trial. Acta Paediatr 1995; 84:775-780 [Medline]
13. Santosham M, Fayad I, Zikri MA, A double-blind clinical trial comparing World Health Organization oral rehydration solution with a reduced osmolarity solution containing equal amounts of sodium and glucose. J Pediatr 1996; 128:45-51 [CrossRef][Medline]
14. Division of Diarrhoeal and Acute Respiratory Disease Control. 25 Years of ORSJoint WHO/ICDDRB Consultative Meeting on ORS Formulation. Geneva, Switzerland: World Health Organization; 1995. CDR/CDD/95.2
15. Alam NH, Majumder RN, Fuchs GJ, CHOICE Study Group The efficacy and safety of a reduced osmolarity oral rehydration solution in adults with cholera. Lancet 1999; 354:296-299 [CrossRef][Medline]
16. Division of Diarrhoeal and Acute Respiratory Disease Control. The Treatment of Diarrhoea-A Manual for Physicians and Other Senior Health Workers. Geneva, Switzerland: World Health Organization; 1995. WHO/CDR/95.3
17. El-Mougi M, Hendawi A, Koura H, Hegazi E, Fontaine O, Pierce NF Efficacy of standard glucose-based and reduced-osmolarity maltodextrin-based oral rehydration solutions: effect of sugar malabsorption. Bull WHO 1996; 74:471-477 [Medline]