Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Clinical trial of a rice-powder oral rehydration beverage

Homero Martínez, Jean-Pierre Habicht, Cutberto Garza, and Felipe Mota



A clinical trial was conducted to assess the efficacy of a rice-based gruel as a rehydration solution, identified in a previous study to be widely used by mothers during diarrhoeal episodes in their children. Seventy children under five years of age, admitted to a paediatric hospital with clinical dehydration secondary to acute diarrhoea, were randomly assigned to receive rice-based gruel (n = 37) or oral rehydration salts (ORS) (n = 33). The hydration status was measured on admission and hourly until discharge, for a maximum of 8 hours, after which treatment was considered to have failed in children who had not been rehydrated Successful rehydration was achieved in 92% of the patients receiving rice-based gruel and 91% of those receiving ORS. Over 50% of all patients were rehydrated 4 to 5 hours after treatment was initiated; at these times the faecal output was statistically significantly lower in patients receiving rice-based gruel than in those receiving ORS. More patients were discharged from the study with hypernatraemia in the group receiving ORS than in the group receiving rice-based gruel, whereas the number of patients with hyponatraemia on discharge was similar in both groups. Faecal sodium concentrations were similar on admission in both groups but were statistically significantly lower at discharge in the group receiving rice-based gruel. We concluded that rice-based gruel could be safely used as an oral rehydration solution at the community level.


Diarrhoeal disease is one of the most common illnesses affecting children under five years of age in the developing world [1]. The loss of water and electrolytes from the body starts with the passing of the first watery stool, so it is present in all diarrhoeal episodes [2]; however, clinical signs of dehydration are not evident until there is an acute fluid loss of approximately 4% to 5% of body weight [3]. Dehydration takes its heaviest toll on infants' lives, as most deaths secondary to diarrhoeal disease are due to the excessive loss of water and electrolytes, reaching a level incompatible with human life [4]. These deaths could be largely averted if proper rehydration therapy were begun in a timely manner [5].

From the beginning of the first half of the nineteenth century, rehydration was achieved mainly by administering intravenous fluids [6, 7]. The discovery in the mid-1960s that oral glucose with sodium in an equimolar solution could significantly enhance the absorption of water and electrolytes has rightly been considered one of the most important discoveries of this century [8].

Oral rehydration therapy (ORT) is based on the fact that the bowel's ability to absorb fluids and salts is preserved during acute diarrhoea despite active secretion of water into the intestinal lumen [9, 10]. Glucose and other carriers are absorbed across the brush border of the epithelial intestinal cell by an active mechanism, in the course of which a molecule of sodium, linked to a molecule of glucose, is carried into the cell. The presence of sodium in the intestinal cell primes another active mechanism, the sodium pump [11]. This enzymatic system is in charge of maintaining a low level of sodium within the cell, moving it past the lateral end of the cell into the intercellular space, from which it passes to the intravascular space. Water passively follows the movement of sodium, due to solvent drag. Osmotic forces also play a role in moving more sodium, glucose, and water into the intestinal cell and the intravascular space [11].

The most appealing features of ORT are that it is noninvasive and a more efficient way of treating dehydration than intravenous fluid and electrolyte administration. Also, it can be taken past hospital boundaries to where it is most needed, that is, the community and the home [12, 13]. The last feature, however, is associated with major obstacles in the implementation of ORT. The problems are related to the poor acceptability of oral rehydration prepackaged salts (ORS) to mothers, who view this as a treatment that should stop diarrhoea, which it does not do [14], and not as a means to rehydrate a child, a concept that is often elusive in the home management of diarrhoea [15 18]. In addition, poor availability of ORS at the community level is translated into lack of packages at home when a child suffers an episode of diarrhoea [19, 20]. As an alternative to pre-packaged ORS, and in search of other effective means to provide early rehydration fluids to children with diarrhoea outside the hospital, home-made fluids have been evaluated [12, 21-25].

An ethnographic study of maternal practices in the management and treatment of diarrhoea in infants and young children in rural central Mexico found that more than 80% of mothers in the area used a rice-based gruel for children with diarrhoea, although it was not specifically intended as a rehydration beverage [26]. Although clinicians expressed some concern that the home preparation might not be ideal because of inadequate electrolyte content and hyperosmolality, it was preferable not to have to teach new practices (e.g., adding salt, avoiding sugar) when promoting this beverage at the community level.

Thus, it was important to do a trial of this rice gruel both for ethical reasons to ensure the beverage's efficacy before promoting it in the field, and to convince the medical establishment that home-made beverages are a suitable alternative to pre-packaged ORS. This paper reports on the second of three studies in a programme aimed at assessing the potential effectiveness of a rice-powder gruel in the home management of children with mild to moderate dehydration due to acute diarrhoea [27].


Study design

We enrolled 87 children of both sexes, less than five years of age, who were admitted to the Hospital Infantil de México in Mexico City with a history of acute diarrhoea of less than 15 days' duration, and with clinical dehydration that merited hospital treatment (mild, <5% acute weight loss; moderate, 5%-10% acute weight loss). Exclusion criteria were purging rates greater than 10 ml/kg body weight/h for more than two consecutive hours, and other concurrent infectious diseases or complications (other than dehydration) secondary to acute diarrhoea. Forty-five children were assigned randomly to an experimental group to receive the rice-powder gruel, and 42 to a control group to receive an oral rehydration solution. Other than treatment solutions, clinical management of both groups was similar and conformed to the norms established as standard treatment in the hospital's oral rehydration ward [28].

Despite experience showing that specific intestinal pathogens are not associated with dehydrating diarrhoea [29], faecal samples were obtained for bacterial cultures and to determine the presence of rotavirus and parasites; this was done to address specific concerns of the medical staff in relation to the efficacy of the test beverage. Serum, urinary, and faecal osmolality and sodium and potassium concentrations were measured; the specific density of urine was determined; and routine haematology evaluations were performed on admission. Blood and urine samples were obtained at discharge for similar determinations. Urine collection bags were used to separate urine from stools. The severity of dehydration was measured on admission and hourly until discharge, taking into consideration changes in body weight, pulse, respiration rate, irritability, drowsiness, sunken fontanelle, sunken eyes, mucosal hydration, skin turgor, presence of tears, purging rate, and urine output [30].

Rehydration was considered to be adequate when none of the signs of dehydration was present; no watery stools were passed in two consecutive hours; weight was stable for two consecutive hours; and two independent, concurrent clinical assessments of adequate hydration status were in agreement. All children were discharged from the hospital once their usual fluid diet had been reestablished. Treatment was considered failed in children whose hydration had not returned to normal within eight hours. Caretakers were asked to return with the children to the outpatient clinic within 24 hours of hospital discharge for follow-up evaluation.

The study was approved by the hospital's institutional review board, and signed informed consent from parents of all children was obtained before study enrollment.

Composition of rehydration solutions

The rice powder-based gruel (RBG) was prepared in a manner closely resembling the way mothers do it in their homes, as determined by chemical analyses during the background study [26]. Based on other published trials that successfully used rice-based beverages with added electrolytes [31, 32], the RBG contained 50 g rice powder/L water. It contained 60 g/L sucrose (table sugar), the maximum amount that clinicians in charge of the oral rehydration ward were willing to try, for fear of a possible high purging rate effect of RBG secondary to the osmolar load of sugar. The mean (± SD) sodium (Na) content was 2.5 (± 1.8) mmol/L, and that of potassium (K) was 2.0 (± 0.6) mmol/L. The total osmolality was 225.5 mOsm/L (± 6.3), due largely to the added sucrose [26]. The control ORS was prepared according to World Health Organization (WHO) recommendations [22], and contained Na 90 mmol/L, K 20 mmol/L, citrate 10 mmol/L, chloride 80 mmol/L, and glucose 20 g/L (111 mmol/L); total osmolality was 311 mOsm/L [33, 34].

Statistical analyses

Because of clinical concern about hyperosmolality, patients with high purging rates of diarrhoea were excluded from the study. All analyses were conducted with and without the inclusion of these patients. Comparisons between categorical data were made using an overall c 2 statistic or the Fisher-Irwin exact test when marginal frequencies of a 2 x 2 table were less than 5 [35]. Comparisons between mean values of selected parameters in the study and the control groups were done by the t test [36]. The 95% confidence interval around the difference between the proportion of successfully rehydrated patients in each group (mean ± 2 SE) was calculated [35-37]. Hydration outcomes were controlled by gender of patients.


Seventeen (20%) of the 87 patients originally recruited were excluded after entering the study because of purging rates greater than 10 ml/kg body weight/h. Eight were in the experimental group and nine in the control group. No significant differences in clinical status at admission were detected between subjects excluded from the study and those who completed the protocol.

Demographic and clinical characteristics on admission were similar between groups with and without the inclusion of the excluded subjects. The results with RBG and ORS were similar including and excluding high-purging patients. Consequently, only the analysis of patients who completed the protocol is presented, that is, 37 patients in the study group and 33 in the control group. Randomization was successful in producing comparable groups with no statistically significant differences in age, weight/age once rehydrated, duration of diarrhoea on admission, severity of dehydration, serum chemistries, and faecal output (table 1).

Seven patients in each group (19% experimental, 21% control) had stool cultures on admission that were positive for pathogens. The experimental group had three positive cultures for enteropathogenic Escherichia coli, two for Salmonella sp., one for Candida sp., one for Campylobacter sp., one for Aeromonas sp., and one with both Shigella sp. and Aeromonas sp. In the control group, one culture grew enteropathogenic E. coli, one Campylobacter sp., two Aeromonas sp., one Shigella sp., one both enterotoxigenic E. coli and Aeromonas sp., and one both enteropathogenic E coli and Shigella sp. No association was found between diarrhoeal episodes due to these pathogens and any of the clinical outcomes.

No significant differences were observed in the amounts of oral solution ingested by the subjects. Successful rehydration within the maximum eight-hour study period was achieved in 34 (92%) of 37 patients in the study group and 30 (91 %) of 33 in the control group. The 95% confidence interval around the differences between the two proportions was 0.14. Two of the three failures in the study group were patients who did not readily accept the RBG within the first two hours of treatment, and the medical resident in charge decided to switch them to ORS. The same resident was in charge of both patients. The other 11 medical residents did not observe any such difficulty associated with the RBG. The hydration status of the remaining four treatment failures in both groups did not return to normal within eight hours.

Other values used to measure the response to treatment were the average time between admission and return of hydration status to normal, mean weight gain, and percentage weight gain during hospital stay (table 2). Children in the study group gained slightly more weight than control children and took less time to rehydrate, but the differences are not statistically significant. The cumulative rate at which the hydration status became normal is shown in figure 1 (see FIG. 1. Cumulative percentage of patients successfully rehydrated during the study period (includes only patients without high purging rates). *p < .05). The faecal output at the mean time taken to rehydrate is shown in figure 2 (see FIG. 2. Faecal output at mean time taken to rehydration in RBG and WHO groups (includes only patients without high purging rates). p <.05; p < .01.). At four and five hours, when the hydration status of more than 50% of all patients was normal, the differences in faecal output between the study and control groups were statistically significant. These results were not affected by the patients' gender.

On admission, the serum osmolality was not significantly different between the study and control groups (285 and 293 mmol/L, respectively), and no patient had a serum sodium concentration below 130 mEq/L. The range of serum sodium and the percentage of children with low ( < 130 mEq/L), normal (130-150 mEq/L), and high (>150 mEq/L) serum sodium concentrations at discharge are summarized in table 2. The distribution of patients with low, normal, and high concentrations on admission and discharge in the two groups is summarized in sable 3. One patient in each group who had a low value on admission still had a low measurement at discharge (126 mEq/L study group, 128 mEq/L control group), although none developed clinical signs of hyponatraemia during observation. On the other hand, 10% of the patients in the control group who had normal values on admission were discharged with high levels, in contrast to none of the patients who received RBG. Forty-four percent of the patients in the control group with high serum sodium on admission remained hypernatraemic after receiving ORS, compared with only 29% of patients receiving RBG (p < .01).

TABLE 1. Selected clinical characteristics of the RBG and WHO groups on admission

Characteristics RBG group (n = 37) WHO group
(n = 33)
Age (mo)      
mean ± SD 8.1±5.2 8.4 ± 5.9 NSa
range (min, max) (1, 28) (1, 24)  
M 49% (18) 73% (24) .05b
F 51% (19) 27% (9)  
Weight (kg)      
mean ± SD 6.67 ± 1.6 6.75 ± 1.7 NSb
range (min, max) (3.3, 10.4) (3.3, 10.6)  
Nutrition status (wt/age)      
normal 43% (16) 33% (11) Nsb
I degree malnutrition 51% (19) 45% (15)  
II degree malnutrition 5% (2) 15% (5)  
III degree malnutrition 0 6% (2)  
Days with diarrhoea      
1-3 54% (20) 59% (19) NSb
4-7 30% (11) 25% (8)  
8-15 16% (6) 16% (5)  
<38°C 49% (18) 56% (18) NSb
>38°C 51% (19) 44% (14)  
present 62% (23) 53% (17) Nsb
absent 38% (14) 47% (15)  
Dehydration as estimated from weight gain at discharge      
no dehydration 5% (2) 3% (1) NSb
mild 81% (3) 79% (26)  
moderate 11% (4) 15% (5)  
severe 2% (1) 3% (1)  
Serum sodium (mEq/L)      
mean ± SD 143.6 ± 7.9 145 ± 8.2 NSa,b
range (min, max) (130, 162) (130, 160)  
low (<130 mEq/L) 0 0  
normal range (130-150 mEq/L) 81% (29) 69% (22)  
elevated ( > 150 mEq/L) 19% (7) 31 % (10)  
Serum osmolality (mmol/L)      
mean ± SD 285 ± 22 293 ± 22 NSa
range (min, max) (240, 341) (245, 349)  
Faecal output      
low ( < 10 ml/kg/h) 76% (28) 63% (20) NSb
high (10-20 ml/kg/h) 5% (2) 13% (4)  
very high (>20 ml/kg/h) 19% (7) 25% (8)  

NS = not significant
a. t test.
b. c 2 test.

In a subsample of 10 subjects for whom values were available, the faecal sodium concentrations were similar on admission in both groups but were significantly lower at discharge in the study group than in the control group (table 4).

TABLE 2. Response to treatment in the RBG and WHO groups

Responses RBG group WHO group Significance
(n = 37) (n = 33)
Time taken to rehydrate (h) (for patients successfully treated)
mean ± SD 3.67 + 1.55 4.2 ± 1.69 NSa
range (min, max) (1, 7) (1, 8)  
Weight gain during hospital stay (kg)
mean ± SD 0.233 ± 0.202 0.217 ± 0.18 NSa
range (min, max) (0, 0.82) ( - 0.15, 0.80)  
% weight gain during hospital stay
mean ± SD 3.1 ± 2.25 3.4 ± 2.9 NSa
range (min, max) (0, 10.2) ( - 2.1, 12.0)  
Serum sodium on discharge (mEq/L)
mean ± SD 140.3 ± 7.0 143.6 ± 7.4 0.05a,b
range (min, max) (126, 160) (128, 160)  
low ( < 130 mEq/L) 3% (1) 3% (1)  
normal range (130-150 mEq/L) 92% (33) 76% (22)  
elevated ( > 150 mEq/L) 5% (2) 21 % (6)  

NS = not significant. a. t test.

Fifty-four percent of all patients returned to their scheduled follow-up (57% study group, 52% control group). All were well hydrated and none experienced any clinical complications.

Contents - Previous - Next