Clinical trials of improved oral rehydration salt formulations: a review.

Reviews/Analyses Clinical trials of improved oral rehydration salt formulations: a review M.K. Bhan,1 D. Mahalanabis,2 0. Fontaine,3 & N.F. Pierce3 Reviewed are all the published clinical trials of glycine-based oral rehydration salts (ORS), L-alanine- based ORS, L-glutamine-based ORS, maltodextrin-based ORS, and rice-based ORS, as well as the results of several recently completed, but unpublished, studies of these formulations that were support- ed by WHO. All experimental ORS formulations contained the same concentrations of salts as citrate- based WHO-ORS; all trials were randomized comparisons with WHO-ORS, and all except those with rice-based ORS were double-blind studies. The rate of stool loss and, less frequently, the duration of diarrhoea were used as indicators of clinical performance to compare ORS formulations. The following conclusions were reached concerning the efficacy and use of modified ORS formula- tions. Rice-based ORS (50 g/l) is superior to WHO-ORS for patients with cholera, and for such patients it can be recommended in any situation where its preparation and use are practical. Rice-based (50 g/l) and WHO-ORS solutions are equally effective for treating children with acute non-cholera diarrhoea, when feeding is resumed promptly following initial rehydration, as has been consistently recommended by WHO. Since rice-based ORS is not superior to WHO-ORS for such children, there is no apparent reason to advise a change from glucose to pre-cooked rice in the recommended formulation for WHO-ORS. Maltodextrin-based ORS formulations (50g9l) and WHO-ORS appear to be equally effec- tive for treating children with acute non-cholera diarrhoea; there is no reason to advise a change from glucose to maltodextrin in the recommended formulation for WHO-ORS. Amino-acid-containing ORS formulations are not recommended for either non-cholera or cholera diarrhoea, since they are more costly and have no clinical advantage over WHO-ORS for children with acute non-cholera diarrhoea or over rice-based ORS for persons with cholera. Introduction The oral rehydration salts (ORS) solution recommen- ded by WHO is safe and effective for treating patients of all ages suffering from dehydration due to diarrhoea of any etiology, provided they are able to drink and that the dehydration is not severe (1). However, dehydration is an unusual event and 1 Additional Professor, Department of Pediatrics, All India Insti- tute for Medical Sciences, New Delhi, India. 2 Associate Director, Clinical Sciences Division, International Centre for Diarrhoeal Diseases Research, Bangladesh, Dhaka, Bangladesh. 3 Medical Officer, Diarrhoeal Diseases Control Programme, World Health Organization, 1211 Geneva 27, Switzerland. Requests for reprints should be sent to Dr 0. Fontaine at this address. Reprint No. 5553 parents may not understand the relation between diarrhoea and the need for hydration. In any case, the parents' primary concem, which is often shared by many health workers, is to see the diarrhoea stop. This may limit the acceptance of the WHO-ORS solution as a "treatment" for diarrhoea, since it neither reduces the rate of stool loss nor shortens the duration of diarrhoea (2-5). In part, this may ex- plain, the continuing widespread use of ineffective, and sometimes dangerous, "antidiarrhoeal" drugs and attibiotics. These considerations have prompted efforts to develop an ORS formulation that, like WHO-ORS, would be inexpensive, safe, effective and stable dur- ing prolonged storage, but would also reduce the rate of stool loss and the duration of diarrhoea. Such a product could be promoted as having antidiarrhoeal activity and should therefore have greater acceptabil- ity; it might also help to reduce demand for ineffec- Bulletin of the World Health Organization, 1994, 72 (6): 945-955 © World Health Organization 1994 945 M.K. Bhan et al. tive antidiarrhoeal drugs. Such an achievement would represent a significant advance in efforts to control diarrhoea morbidity and mortality through effective case management. Several approaches have been taken to develop such an improved ORS formulation. The present review describes the strategies used and summarizes the results of clinical trials of experimental formula- tions in adults and children suffering from dehydra- tion caused by cholera or acute non-cholera diar- rhoea. Approaches taken to develop an improved ORS formulation The effectiveness of the WHO-ORS solution depends on the coupled active transport of sodium ions and glucose across the brush border membranes of enterocytes (6), which results in passive absorp- tion of water and other electrolytes. This function remains largely intact during diarrhoea, irrespective of its cause. Two types of organic solutes are absorbed effi- ciently and relatively independently of each other in the small intestine, and each enhances the absorption of sodium ions and water: certain neutral amino acids-glycine, L-alanine, L-glutamine-and the dipeptides, tripeptides, proteins, and hydrolysed pro- teins from which they are derived; and glucose and the disaccharide (maltose), trisaccharide (malto- triose), oligosaccharide mixtures (maltodextrins of various grades), and polysaccharides (starches from rice or other cereals) from which glucose is derived. One approach to developing an improved ORS formulation was based on adding neutral amino acids or their dipeptides to WHO-ORS (7). Glycine was studied first because of its low cost and low toxicity (8). Subsequently, L-alanine was used when it was discovered that it enhanced the transport of sodium ions across the brush border membrane more effi- ciently than glycine or glycyl-glycine (9). Recently, L-glutamine has been used because it enhances the absorption of both sodium and chloride ions (10); also, it is the principal metabolic fuel of the small bowel mucosa and is involved in the repair of muco- sal damage (11, 12). A second approach was to replace the glucose in ORS with glucose polymers, such as maltodextrins or cooked cereal powders, thus increasing the amount of potentially available glucose in the ORS solution, while decreasing its osmolarity. The maximum con- centration of glucose that can be employed in solu- tions of ORS salts is about 20 g/l (110 mmol/l); higher concentrations make the solution hypertonic and may cause osmotic diarrhoea. This amount is only sufficient to promote the absorption of the so- dium (90 mmol/l) and water in the ORS solution. Water and salts secreted into the intestine are not reabsorbed and the rate of stool loss is essentially unchanged (13). It was hypothesized that ORS for- mulations in which glucose (20 g/l) was replaced by increased amounts of maltodextrin or cooked rice powder (50-80 g/l) would be hypotonic, thus avoid- ing osmotic diarrhoea, but would yield sufficient glucose to promote also the reabsorption of endoge- nous intestinal secretions and thus reduce the volume and duration of diarrhoea (7). A third approach combined the first two strate- gies: glucose polymers were used to provide a larger amount of glucose than in WHO-ORS, while reduc- ing osmolarity, and an amino acid (glycine) was added in an attempt to enhance further the absorption of sodium ions and water. Results of clinical trials of experimental ORS formulations Covered in the present review are data on all the published trials of glycine-based ORS, L-alanine- based ORS, L-glutamine-based ORS, maltodextrin- based ORS and rice-based ORS, as well as the results of several recently completed, but unpub- lished, studies of these formulations that were sup- ported by WHO. All the experimental ORS formula- tions contained the same concentrations of salts as the citrate-based WHO-ORS; all trials were random- ized comparisons with WHO-ORS; and all except those with rice-based ORS were double-blind. We used the stool volume and, less frequently, the duration of diarrhoea as indicators of clinical per- formance to compare ORS formulations. However, each of these parameters independently affects the total stool output during an episode of diarrhoea. When both are reduced, for example, the percentage reduction in total stool output is usually greater than the percentage reduction in either of the contributing measurements. Therefore, total stool output during a diarrhoeal episode was considered to be the single most useful outcome measure for comparing differ- ent ORS formulations.a ORS solutions with added amino acids Glycine-containing formulations. The first studies of glycine-containing ORS involved adults (14) and children (15) with acute diarrhoea caused mostly by a Recent advances in the development of improved ORS. Unpublished document WHO/CDD/CMT/91.1, 1991. 946 WHO Bulletin OMS. Vol 721994 Improved ORS formulation: a review Vibrio cholerae 01 or enterotoxigenic Escherichia coli (ETEC). These reported that the addition of gly- cine (110 mmol/l) to WHO-ORS solution caused significant reductions in total stool volume and dura- tion of diarrhoea (by 50% and 30%, respectively). However, the results of initial studies of a glycine- containing ORS solution (WHO-ORS solution containing an additional 110 mmol/1 of glycine) in children with acute non-cholera diarrhoea were inconclusive: two studies reported a clinical benefit (16, 17), whereas four others showed no clinical advantage for the formulation with added glycine (18-21). In an effort to clarify these findings, six glycine- containing ORS formulations were evaluated in nine clinical trials that were supported by WHO (22-27; and K.N. Jalan, H. Romer, E. Salazar-Lindo, person- al communications). Subjects were children with acute non-cholera diarrhoea of viral or bacterial eti- ology. Only one study admitted children up to 5 years of age; otherwise the upper age limit was 2 years or 3 years. The criteria for inclusion and exclu- sion in the studies were similar. Several strategies were used to keep the osmolarity of the experimental ORS formulations close to that of the WHO-ORS solution (311 mosmol/l). These included the follow- ing: reducing the glucose content of the solution; pro- viding part of the amino acid as a dipeptide (glycyl-glycine); and replacing glucose by a glucose polymer (maltodextrin). The mean theoretical osmo- larity of the experimental solutions was 343 mosmol/l (range, 326-387 mosmol/l) (Table 1). A meta-analysis of the results of seven of these trials, which included 643 subjects, has appeared (28). Two clinical trials that did not use the same standard recording forms were not included, although their results were similar to those of that meta-analysis (25; E. Salazar-Lindo, personal com- munication). The meta-analysis clearly showed that the ORS formulations containing glucose (or malto- dextrin) and glycine (or glycyl-glycine) were not clinically superior to the standard WHO-ORS solu- tion. Rather, the total stool output was increased by 4% (95% confidence interval (CI) = -6%, 14%) and the duration of diarrhoea was increased by 3% (95% CI = -7%, 13%) with glycine-containing ORS solu- tions. Neither of these differences, however, was sta- tistically significant. L-Alanine-containing formulations. Three ORS for- mulations containing L-alanine have been tested in five clinical trials (29, 30; and P. Santos-Ocampo, A. Madkour, F.C. Patra, personal communications). The key features and results of these trials are summar- ized in Table 2. One study involved adults with chol- era (29); another involved adults with diarrhoea of mixed etiology (cholera and non-cholera diarrhoea); and the other three were of under-5-year-olds with acute non-cholera diarrhoea caused mostly by rota- virus and ETEC. Two studies using the same experimental formula- tion, containing 111 mmol/l of glucose and 90 mmol/l of L-alanine (400 mosm/l), gave different clinical results in adults with cholera and children with acute non-cholera diarrhoea. This formulation was sub- stantially more effective than WHO-ORS solution in patients with cholera (29), the total stool output being reduced by 42%. The same formulation, how- ever, had no significant beneficial effect on stool output or duration of diarrhoea in children with acute non-cholera diarrhoea (30) (Table 2). The other three studies of children with non- cholera diarrhoea and of adults with diarrhoea of mixed etiology used experimental formulations with a lower concentration of L-alanine (50 mmol/l), while two also used a lower amount of glucose (12 g/l or 18 gIl). Although the differences observed between the two treatment groups in the study Table 1: Characteristics of the seven clinical trials with glycine-containing ORS solutions for non-cholera diarrhoea Solution Age of subjects Carrier substances in osmolarity No. of Author Year (years) formulation (mosmol/l) patients Santos Ocampo (ref. 22) 1988 <3 Glucose (20)a/glycine (4)/glycyl-glycine (4) 387 66 Jalan 1988 <5 Maltodextrin MD25 (20)/glycine (4)/glycyl-glycine (4) 326 33 Romer 1988 <3 Maltodextrin MD25 (20)/glycine (4)/glycyl-glycine (4) 326 150 Pizarro et al. (ref. 23) 1988 <2 Glucose (12)/glycine (4)/glycyl-glycine (4) 342 62 Kassem et al. (ref. 27) 1989 <3 Maltodextrin MD25 (20)/glycine (4)/glycyl-glycine (4) 356 150b Kassem et al. (ref. 27) 1989 <3 Maltodextrin MD25 (20)/glycine (4)/glycine (8) 356 150b Bhan et al. (ref. 24) 1990 <2 Maltodextrin MD25 (20)/glycine (4)/glycyl-glycine (4) 326 93 Grange (ref. 26) 1992 <3 Maltodextrin MD25 (20)/glycine (8) 356 89 a Figures in parentheses are the substance concentrations in g/l. b Three-cell study comparing two experimental formulations with WHO-ORS solution (50 patients per cell). WHO Bulletin OMS. Vol 72 1994 947 M.K. Bhan et al. involving adults were not statistically significant, the L-alanine-containing ORS solution reduced the total stool output by 22%. On the other hand, in the stud- ies involving children, although the osmolarities of these formulations (360-320 mosmol/l) were closer to that of the WHO-ORS solution (311 mosm/1), the experimental ORS formulations exhibited no clinical advantage over WHO-ORS (Table 2). L-Glutamine-containing formulation. Three clinical trials of an L-glutamine-containing ORS have recent- ly been completed (62 and R. Kumala, M.K. Bhan, personal communications). The first involved adults with cholera, while the other two involved children with acute non-cholera diarrhoea. The key features and results of these trials are given in Table 3. The study of the adults with chol- era used an experimental formulation containing 90 mmolIl of glucose and 90 mmol/l of L-glutamine (380 mosm/l). The two studies of children with non- cholera diarrhoea used an experimental formulation containing 50 mmol/l of glucose and 50 mmolI1 of L-glutamine (300 mosm/l). The clinical effectiveness of the L-glutamine- containing ORS solutions was similar to that of gly- cine- and L-alanine-containing ORS formulations: for the adults with cholera, there was a significant reduc- tion in the stool output and duration of diarrhoea but no clinical advantage over the WHO-ORS solution for the children with non-cholera diarrhoea, although the total osmolarity of the solution was slightly less than that of the WHO-ORS solution. Cereal-based ORS formulations Numerous trials comparing cereal-based ORS formu- lations with WHO-ORS have been reported. Most have involved rice-based formulations but a few have involved other cereals such as sorghum-wheat- or maize-based ORS. The results obtained with these other cereals generally resemble those obtained with rice-based ORS. Consequently, only the results of the clinical trials of rice-based ORS are reviewed here. These have been divided into the following categories: - thirteen trials summarized in a previously report- ed meta-analysis (31); - two recently completed trials (42, and A.M. Molla, personal communication); and - additional studies involving subjects of special concern: infants below 6 months of age (63); severely malnourished children (43); and chil- dren with a high prevalence of glucose malab- sorption (E. Chea-Woo, personal communica- tion). These are discussed in detail below. Meta-analysis of 13 clinical trials. Several early trials showed that ORS solutions containing 50-80 g/l of cooked rice powder in place of the 20 g/l of glucose in the WHO-ORS solution substantially reduced the rate of stool loss during episodes of acute diarrhoea (32-36). Other studies, however, reported no signifi- cant benefit for the rice-based ORS (37-40). Inter- pretation of these results is difficult because of the wide age range of the study subjects, differences in the practices (food was offered at various intervals after rehydration, often not until 24 or 48 hours later, or feeding practices were not described). Additional- ly, total stool output and duration of diarrhoea were reported in only one trial, although all studies report- ed the initial 24-hour stool output. Table 2: Characteristics and results of the clinical trials with alanine-containing ORS solutions Carrier % difference in:a. bsubstances (g/l) osmolarity No. of Mean total stool Mean diarrhoea Author Year Patients Glucose Alanine (mosmol/l) patients output duration Patra et al. (ref. 29) 1989 Cholera 20 8 400 94 -42 -15 (adults) (-62 to -21)C (-27 to 0) Patra 1993 Cholera and 18 4.5 350 154 -22 -9 non-cholera (-46 to 3) (-19 to 1) (adults) Sazawal et al. 1991 Non-cholera 20 8 400 129 -2 -5 (ref. 30) (<3 years of age) (-39 to 35) (-14 to 17) Santos Ocampo 1991 Non-cholera 20 4.5 360 140 23 8 (<3 years of age) (-21 to 69) (-15 to 31) Madkour 1991 Non-cholera 12 4.5 320 100 -18 -11 (<3 years of age) (-48 to 12) (-33 to 5) a [(Alanine-containing ORS - glucose ORS)/glucose ORS] x 100. b % difference in stool output and duration of diarrhoea refers to the reduction with the experimental ORS solution when the sign is negative. c Figures in parentheses are the 95% confidence interval. WHO Bulletin OMS. Vol 72 1994948 Improved ORS formulation: a review Table 3: Characteristics and results of clinical trials with glutamine ORS solutions Carrier % difference in: a, b osmolarity No. of Mean total stool Mean diarrhoea Author Year Patients Glucose Alanine (mosmol/l) patients output duration Kumala 1991 Cholera 16 13 355 181 -30 -13 (adults) (-34 to 28)C (-24 to 2) Bhan 1992 Non-cholera 9 7.3 300 125 -1 14 (<3 years of age) (-33 to 30) (-8 to 36) Ribeiro et al. 1992 Non-cholera 9 7.3 300 118 3 1 (ref. 62) (<3 years of age) (-26 to 32) (-16 to 19) a [(Glutamine-containing ORS-glucose ORS)/glucose ORS] x 100. b See footnote b, Table 2. c Figures in parentheses are the 95% confidence interval. To clarify these results, a meta-analysis was carried out in 1992 of all published or available randomized controlled trials comparing rice-based ORS (50-80 g/l) with WHO-ORS (31). A total of 13 studies were evaluated (32-41: A.M. Moechtar, E. Guiraldes, A.N. Alam, personal communications). These involved 531 adults and 424 children with chol- era or severe cholera-like diarrhoea, and 344 infants and young children with acute non-cholera diar- rhoea. The key features of the studies analysed are given in Table 4. In some studies, rice powder was cooked immediately before use and the salts then added; in others, packets containing pre-cooked rice powder and ORS were used. The meta-analysis showed that, irrespective of their age, cholera patients given rice-based ORS solution had substantially lower rates of stool loss than those given WHO-ORS solution. Stool output was reduced by a mean of 55 g per kg during the first 24 hours of treatment with rice-based ORS solu- tion, a reduction of 34% (95% CI = 25%, 43%). In contrast, the initial 24-hour stool output of children with acute non-cholera diarrhoea and who were given rice-based ORS solution was only reduced by a mean of 18 g per kg (18%); 95% CI = 6%, 30%), relative to children given WHO-ORS solution. Recently completed trials. Two recently completed studies were designed to address the weaknesses identified in the above trials, especially the impact of early feeding, as recommended by WHO, on the effectiveness of both the glucose-based ORS solution and the rice-based ORS solution (42; A.M. Molla, personal communication). The trials included 611 under- 18-month-olds with acute non-cholera diarrhoea and signs of mod- erate dehydration. Children who were predominantly breast-fed or with severe malnutrition were exclud- ed. The children were offered a standard rice and vegetable diet in six to eight helpings of equal vol- ume (150 g per kg per day) starting immediately after completion of rehydration (usually 4-6 hours). Data on total stool output and duration of diarrhoea were available for all but a few of the randomized patients. The key features and results of these trials are summarized in Table 5. In both studies, the 24-hour stool output, total stool output, and duration of diarrhoea were margi- nally reduced among patients given standard WHO- ORS solution compared with those given rice-based ORS solution. None of these differences was statisti- cally significant, however, except the duration of diarrhoea, which, in one study (42), was significantly shorter among patients given WHO-ORS solution. Studies involving subjects of special concern Infants below 6 months of age. Because pancreatic amylase is deficient at birth and does not reach adult levels until about 5 months of age, rice-based ORS solution may not be suitable for use with very young infants. This has been examined in a recent trial comparing rice-based ORS and standard WHO-ORS solutions in 100 infants below 6 months of age with acute diarrhoea (63). No significant differences were observed between treatment groups with regard to total stool output and duration of diarrhoea. The total stool output was reduced by 16% (95% CI = -13%, 55%) and the duration of diarrhoea by 8% (95% CI = -26%, 42%) with rice-based ORS solution. Since the confidence intervals for these comparisons include zero, they are not statistically significant. Infants with severe malnutrition. Children with severe malnutrition may also be deficient in pan- creatic amylase, which raises concem about the safe- ty and efficacy of treating them with rice-based ORS solution. In a recently published evaluation of rice- based ORS solution in 150 severely malnourished children with acute non-cholera diarrhoea, no signifi- cant differences were observed in total stool output WHO Bulletin OMS. Vol 72 1994 949 M.K. Bhan et al. Table 4: Characteristics of the clinical trials of rice-based ORS included in the initial meta- analysis Age of patients Amount of rice in Author Year (years)a ORS (g/l) No. of patients Cholera or cholera-like illness Moechtarb 1989 >12 50 164 Moechtarb 1989 >12 50 26 Alamc 1991 Adults 50 93 Alamc 1991 Adults 50 89 Molla et al. (ref. 34) 1985 >10 80 157 Molla et al. (ref. 34) 1985 <10 80 185 Molla et al. (ref. 36) 1989 <5 50 79 Molla et al. (ref. 33) 1983 <5 80 52 Alam et al. (ref. 35) 1985 <8 50 39 Alam et al. (ref. 35) 1985 <8 50 13 Patra et al. (ref. 32) 1982 <5 50 52 Acute non-cholera diarrhoea Guiraldes 1991 <2 50 97 Dutta et al. (ref. 38) 1988 <4 50 70 Bhan et al. (ref. 37) 1987 <5 50 64 El Mougi et al. (ref. 41) 1988 <1.5 50 51 Mohan et al. (ref. 39) 1985 <3 50 46 a None of the studies included infants aged <3 months. b Single studies in which results were stratified for analysis. c Clinical trial with a factorial design (4-cell trial). for patients treated with rice-based ORS or WHO- ORS solutions (43). Total stool output was reduced by 13% (95% CI = -15%, 41%) with rice-based ORS solution but this was not statistically signifi- cant. Children with an increased risk of glucose malab- sorption. A study conducted in Peru reported that glucose malabsorption, leading to an increased rate of treatment failure during oral rehydration therapy, occurred frequently (44). Based on previous studies, the rate of treatment failure due to clinically evident glucose intolerance in young children with acute non-cholera diarrhoea was expected to be no more than 2% (45-50). However, in the Peruvian study, 30% of the patients aged 3-24 months had detectable glucose malabsorption leading to treatment failure in 7% (the proportion of patients requiring intravenous therapy after rehydration was completed), well in excess of the expected 2% failure rate. The safety and efficacy of rice-based ORS solu- tion was evaluated in a study involving children from a similar population with a high prevalence of glucose malabsorption (E. Chea-Woo, personal com- munication). A total of 45% of patients admitted to the study had detectable glucose malabsorption, Table 5: Results of recently completed clinical trials of rice-based ORS in young children with non-cholera diarrhoea who were offered optimal feeding early in the illness % difference in:a, b No. of Age of patients Mean total stool Mean diarrhoea Author Year patients (months) output duration Fayad et al. (ref. 42) 1993 441 3-18 17 25 (-4 to 37)c (10 to 41) Molla 1993 170 3-12 19 2 (-20 to 57) (-22 to 25) a [(Rice ORS - glucose ORS)/glucose ORS] x 100. b See footnote b, Table 2. c Figures in parentheses are the 95% confidence interval. WHO Bulletin OMS. Vol 72 1994950 Improved ORS formulation: a review defined as the presence of >1% glucose in the stools and stool pH <5.5. However, no significant differ- ences were observed in the total stool output or in the duration of diarrhoea with the rice-based ORS solution. Treatment failure rates were high, but not significantly different between treatment groups (24% for patients treated with rice ORS solution ver- sus 27% for those treated with WHO-ORS solution). Revised meta-analysis of rice-based versus WHO-ORS solution. The results of these recently completed clinical trials, involving 828 non-severely malnourished infants and young children with acute non-cholera diarrhoea in which the WHO-ORS solu- tion was compared with the rice-based ORS solution, were combined with the results of the five studies conducted on 344 similar children and already pre- sented in the 13-study meta-analysis (31). The revised meta-analysis, the results of which are sum- marized in Fig. 1, therefore took into account nine studies and included a total of 1172 children. The results indicate that stool output was reduced by 7% (95% CI = -3%, 15%) for patients receiving rice- based ORS solution compared with those receiving standard glucose-containing ORS solution; this dif- ference was not statistically significant. Maltodextrin-based ORS formulation These studies were prompted by the initially encour- aging results reported with rice-based ORS. Malto- dextrins were considered to be an attractive altema- tive to cooked rice powder for the following rea- sons: they are glucose polymers with low osmotic ac- tivity; they dissolve easily in water without cooking; they are inexpensive; and they could easily replace glucose in packaged ORS (53-55). An ORS formulation containing 50 g/l of mini- mally hydrolysed maltodextrin (MD02) (227 mosmol/l) in place of glucose was evaluated in four clinical trials involving young children with acute non- cholera diarrhoea (51, 52 and K.N. Jalan, M. El- Mougi, personal communications). The key features and results of these trials are shown in Table 6. The efficacy of maltodextrin-based ORS was similar to that of WHO-ORS. There was no signifi- cant beneficial effect of maltodextrin-based ORS on either total stool output or duration of diarrhoea. Discussion Role of clinical trials in evaluation of ORS formulations The clinical efficacy of an ORS formulation is deter- mined by the extent to which the water and electro- lytes it contains are absorbed from the intestine. This Fig. 1. Mean percentage reduction in 24-hour stool out- put among individual studies of children with non-chol- era diarrhoea given rice-based ORS solution compared with children given WHO-ORS solution. ([Rice ORS-glu- cose ORS)/glucose ORS] x 100). The pooled (weighted) estimate of percentage reduction in mean stool output (95% confidence interval) is shown under "overview". 0 0- Chile Egypt India India [ India V u, n ( Egypt co 'a Mexico E Pakistan 00 § Peru a, z Overview -80 -60 -40 -20 0 20 40 60 % difference in mean rate of stool output can only be determined using controlled clinical trials involving patients with diarrhoea in which oral intake and faecal output of water and electrolytes are carefuly measured. Two processes within the intestine determine the efficacy of an ORS formulation. The first is rapid osmotic equilibration of the ingested solution with extracellular fluid, which occurs mostly in the duo- denum and proximal jejunum (56) and results from the flow of water across the semipermeable bowel mucosa. Water is absorbed from hypotonic ORS solutions and added to solutions that are hypertonic. The second is absorption of water and electrolytes that is mediated by actively absorbed organic molecules, such as glucose and certain amino acids. This occurs throughout the small bowel (2, 3). In addi- tion, when metabolized by colonic bacteria, undiges- ted starch entering the colon, produces short-chain fatty acids that stimulate electrolyte absorption in the colon (57). Although experimental studies of absorption from intestinal segments in humans or animals, or in- vitro systems, can help to guide the design of clinical WHO Bulletin OMS. Vol 72 1994 I II I I I T iT 11 ---dT IL - ----iT I T +- 4 . I 'L T F+ 951 M.K. Bhan et al. Table 6: Results of clinical trials of maltodextrin ORS with non-cholera patients % difference in:a. b No. of Age of patients Mean total stool Mean diarrhoea Author Year patients (months) output duration Akbar et al. (ref. 51) 1991 69 <36 -10 -8 (-73 to 52)c (-37 to 22) Santos Ocampo 1993 116 <36 35 -8 et al. (ref. 52) (-26 to 51) (-31 to 15) El Mougi 1990 89 <24 23 -9 (-32 to 46) (-42 to 60) Jalan 1990 32 <36 -25 -5 (-85 to 36) (-71 to 81) a [(Maltodextrin ORS - glucose ORS)/glucose ORS] x 100. b See footnote b, Table 2. c Figures in parentheses are the 95% confidence interval. trials and interpret their outcome, they cannot deter- mine the combined effect of the processes listed above on net absorption throughout the length of the gut and, thus, cannot replace clinical trials or neces- sarily predict their outcome. Moreover, most such studies are performed with healthy intestine and the few available models of infection-induced intestinal secretion do not parallel the full range of etiological agents or pathogenic processes that cause diarrhoea and for which ORS solutions are used. Efficacy of modified ORS formulations in cholera The clinical trials summarized in this review show that the efficacy of experimental ORS formulations depends on the etiology of diarrhoea but not on the age of the subjects. Both amino-acid-containing and rice-based formulations were superior to WHO-ORS for adults and children with cholera. The superior performance of these solutions was manifested by substantially reduced rates of stool output, duration of diarrhoea, and total diarrhoeal stool volume. For patients with cholera, similar clinical bene- fits occurred with solutions that were either hypoton- ic (rice-based ORS solution: ca. 200 mosmol/1) or hypertonic (amino-acid-containing ORS solutions: 360-421 mosmol/l) relative to WHO-ORS solution (311 mosmol/l). The improved performance of the amino-acid-containing solutions indicates that the absorption-promoting effect of added amino acids exceeds the potentially adverse effect of hypertonic- ity, which would initially cause a flow of water from the blood into the bowel lumen, increased luminal fluid volume and, if the ORS solution and added water were not eventually absorbed, increased stool volume. The net benefit of these solutions would not have been evident from animal studies that point to the dangers of hypertonic solutions and the benefits of hypotonic solutions (58, 59). In contrast, the clinical benefit of rice-based ORS was probably due to its hypo-osmolarity, which would enhance water absorption, and to its potential to release more glucose than WHO-ORS without incurring an "osmotic penalty" (7, 60). The possibil- ity that amino acids released from the digestion of proteins in rice powder also contributed to the effica- cy of rice-based ORS cannot be excluded. Efficacy of modified ORS formulations in acute non-cholera diarrhoea The results of studies of children with acute non- cholera diarrhoea differed markedly from those of adults or children with cholera. Amino-acid-supple- mented solutions had no beneficial effect, and rice- based ORS solution had only a modest advantage over WHO-ORS solution, which disappeared when feeding with a normal cereal-based diet was resumed promptly after initial rehydration was completed. For young infants, children with severe malnutrition, and children with a high prevalence of glucose malab- sorption, rice-based ORS solution had the same effi- cacy as WHO-ORS solution. Maltodextrin-based ORS had no beneficial effect over WHO-ORS solu- tion. Although the clinical trials summarized here were not designed to explore the mechanisms of action of different ORS solutions, comparison of these results with those for patients with cholera sug- gest that the pathogenic effects of agents causing non-cholera diarrhoea in young children prevented such solutions from having greater efficacy than WHO-ORS solutions. The most likely explanation for this is that the breakdown of polysaccharides or proteins, or the absorption of glucose or amino acids, WHO Bulletin OMS. Vol 72 1994952 Improved ORS formulation: a review or both, was partially impaired in these children. Such effects might be greater if diarrhoea is caused by certain etiological agents, but this could not be determined. It is known, however, that the causative agents of acute non-cholera diarrhoea include several that invade or damage bowel mucosa (e.g., rotavirus, Shigella spp., enteropathogenic E. coli) and might be expected to have multiple effects on intestinal func- tion. In contrast, V. cholerae 01 is not invasive and does not cause mucosal damage; cholera results almost entirely from the action of cholera toxin, which alters the normal processes for absorption of sodium and chloride ions by intestinal mucosa, but does not affect carrier-linked absorption of sodium ions (61). It is of interest that rapid intestinal transit did not appear to limit the effectiveness of ORS solu- tions, since transit is most rapid in patients with cholera. Proposed use of modified ORS formulations Below are summarized the conclusions and recom- mendations conceming the use of modified ORS for- mulations. * Rice-based ORS (50 g/1) is superior to WHO- ORS for patients with cholera. Its use for cholera patients can be recommended for any situation where its preparation and use are practical. * Rice-based (50 g/l) and WHO-ORS solutions are equally effective for treating children with acute non-cholera diarrhoea, when feeding is resumed promptly following initial rehydration, as has been consistently recommended by WHO. Rice-based and WHO-ORS solutions are also equally effective for young infants, children with severe malnutrition, and children with an increased risk of glucose malab- sorption. Since rice-based ORS is not superior to WHO-ORS for such children, there is no reason to advise a change from glucose to pre-cooked rice in the recommended formulation for WHO-ORS. Such a change would be very costly to implement and be without advantage, except for patients with cholera. * Maltodextrin-based ORS formulations (50 g/l) and WHO-ORS appear to be equally effective for treat- ing children with acute non-cholera diarrhoea. There is therefore no reason to advise a change from glucose to maltodextrin in the recommended for- mulation for WHO-ORS. * Amino-acid-containing ORS formulations are not recommended for either non-cholera or cholera diar- rhoea since they are more costly and have no clinical advantage over WHO-ORS for children with acute non-cholera diarrhoea or over rice-based ORS for persons with cholera. Acknowledgements We thank Dr J. Tulloch for valuable comments and sug- gestions in the preparation of this paper, and Ms M. King for secretarial help. Resume Formules am6liorees de sels de r6hydratation orale: le point Cet article pr6sente les resultats d'une revue de tous les essais cliniques publies, ainsi que les r6sultats de plusieurs etudes recemment ache- vees mais non encore publiees, sur les sels de rehydratation orale (SRO) a base de glycine, de L-alanine, de L-glutamine, de maltodextrine ou de riz et qui ont ete finances par le Programme de Lutte contre les Maladies diarrheiques (LMD) de l'Organisation mondiale de la Sant6 (OMS). Les SRO exp6rimentaux, dont l'efficacit6 etait 6valu6e dans ces 6tudes, contenaient les memes concen- trations de sels que les SRO a base de citrate recommand6s par l'OMS. Tous les essais cli- niques 6taient des comparaisons randomis6es de SRO exp6rimentaux avec les SRO de l'OMS, et tous, a l'exception des essais de SRO a base de riz, etaient des 6tudes en double-insu. Le volume des selles et, moins fr6quemment, la dur6e de la diarrhee, ont 6t6 utilis6s comme indicateurs pour comparer l'efficacite clinique des diff6rentes com- positions de SRO. Cette revue a permis de tirer un certain nombre de conclusions quant a l'effica- cit6 et l'utilisation de ces formules modifi6es de SRO: - Les SRO a base de riz (50 g/l) sont sup6rieurs aux SRO de l'OMS pour traiter les malades atteints de cholera; leur utilisation chez ces malades peut etre recommand6e dans toutes les situations ou leur preparation et leur utilisa- tion sont r6alisables en pratique. - Les SRO a base de riz (50 g/l) et les SRO de l'OMS ont la meme efficacite pour le traite- ment des enfants presentant une diarrh6e aigue d'origine non cholerique, quand l'alimen- tation est redemarr6e rapidement apres la rehydratation initiale, comme cela est recom- mand6 par l'OMS. L'efficacit6 des SRO a base de riz n'6tant pas superieure a celle des SRO de l'OMS, il n'y a aucune raison de recom- WHO Bulletin OMS. Vol 72 1994 953 M.K. Bhan et al. mander le remplacement du glucose par le riz dans la composition des SRO de l'OMS. Les SRO a base de maltodextrines (50 g/1) et les SRO de l'OMS ont la meme efficacit6 pour le traitement des enfants pr6sentant une diar- rhbe aigue d'origine non cholerique; il n'y a donc aucune raison de recommander le rem- placement du glucose par des maltodextrines dans la composition des SRO de l'OMS. Les SRO a base d'acides amines ne sont recommand6s ni pour le traitement des diar- rh6es d'origine non cholerique, ni pour le trai- tement des choleras, car ils sont plus chers et n'offrent aucun avantage clinique par rapport aux SRO de l'OMS pour les enfants pr6sen- tant une diarrhee aigue d'origine non chole- rique ou par rapport aux SRO a base de riz pour les personnes atteintes de cholera. References 1. Avery ME, Snyder JD. Oral therapy for acute diar- rhoea. The underused simple solution. New England journal of medicine, 1990, 323: 891-894. 2. Pierce NF et al. Effect of intragastric glucose electro- lyte infusion upon water and electrolyte balance in Asiatic cholera. Gastroenterology, 1968, 55: 333-343. 3. Hirschhorn N et al. Decrease in net stool output in cholera during intestinal perfusion with glucose-con- taining solutions. New England journal of medicine, 1968, 279: 176-180. 4. Mahalanabis D et al. Use of an oral glucose elec- trolyte solution in the treatment of pediatric cholera: a controlled study. Journal of tropical pediatrics and environmental child health, 1974, 20: 82-87. 5. Sack SA et al. Oral hydration in rotavirus diarrhoea: a double-blind comparison of sucrose with glucose electrolyte solution. Lancet, 1978, 2: 280-283. 6. Schultz SG. Sodium coupled solute transport by small intestine; a status report. American journal of physiology, 1977, 223: E249-E254. 7. Mahalanabis D, Patra FC. In search of a super oral rehydration solution: can optimum use of organic solute mediated sodium absorption lead to the development of an absorption-promoting drug? Jour- nal of diarrhoeal diseases research, 1983, 1: 76-81. 8. Milne MD. Pharmacology of amino acids. Clinical pharmacology and therapeutics, 1968, 9: 484-516. 9. Wapnir RA et al. Oral hydration solution (OHS): effectiveness of alanine in a model of osmotic diar- rhoea. Pediatric research, 1986, 20: 252A. 10. Souba WW, Smith RJ, Wilmore DW. Glutamine metabolism by the intestinal tract. Journal of paren- teral and enteral nutrition, 1985, 9: 608-616. 11. Askanazi J et al. Muscle and plasma amino acids following injury. Influence of intercurrent infection. Annals of surgery, 1980, 192: 78-85. 12. Cersosimo E et al. Role of glutamine in adaptation in nitrogen metabolism during fasting. American journal of physiology, 1986, 259: E622-E628. 13. Hirschhorn N. The treatment of acute diarrhoea in chil- dren: an historcal and physiological perspective. Ameri- can journal of clinical nutrition, 1980, 33: 637-663. 14. Nalin DR et al. Effect of glycine and glucose on sodium and water absorption in patients with chol- era. Gut, 1970, 11: 768-772. 15. Patra FC et al. In search of a super solution: con- trolled clinical trial of a glycine glucose oral rehydra- tion solution in infantile diarrhoea. Acta paediatrica Scandinavica, 1984, 73:18-21. 16. Bai NS, Arunambika, Philips E. A comparative study of Speedoral versus the WHO formula in acute diarrhoea in children. Medicine and surgery, 1988, 28:19-22. 17. Padmanabhan AS, Ganga N, Ashok TP. A con- trolled trial of glycine-supplemented oral rehydration solution (ORS) in childhood diarrhoea. Physicians update, 1988, 1: 37-39. 18. Vesikari T, Isolauri E. Glycine-supplemented oral rehydration solutions for diarrhoea. Archives of dis- ease in childhood, 1986, 61: 372-376. 19. Santosham M et al. Glycine-based oral rehydration solution: reassessment of safety and efficacy. Jour- nal of pediatrics, 1986, 109: 795-801. 20. Briones M et al. [A saline solution containing oligo- saccharides and glycine for oral hydration]. Boletin medico del Hospital Infantil de Mexico, 1988, 45: 437-441 (in Spanish). 21. Carmuega E et al. [Clinical trial of a super oral hydration serum]. Archives Argentinos de Pediatria, 1987, 85: 77-81 (in Spanish). 22. Santos Ocampo PD et al. Clinical trial on the effi- cacy and safety of added glycine and glycyl-glycine to oral rehydration solution. Philippine journal of pediatrics, 1988, 36: 99-120. 23. Pizarro D et al. Comparison of efficacy of a glucose/glycine/glycyl-glycine electrolyte solution versus the standard WHO/ORS in diarrhoeic dehy- drated children. Journal of pediatric gastroenterol- ogy and nutrition, 1988, 6: 882-888. 24. Bhan M. et al. Glycine, glycyl-glycine and maltodex- trin based oral rehydration solution. Acta paediatrica Scandinavica, 1990, 79: 518-526. 25. Khin Maung U et al. Comparison of glucose/elec- trolyte and maltodextrin/glycyl-glycine/electrolyte oral rehydration solution in acute diarrhoea in children. Journal of pediatric gastroenterology and nutrition, 1991, 13: 397-401. 26. Grange A. Evaluation of a maltodextrin/glycine oral rehydration solution. Annals of tropical paediatrics, 1992, 12: 353-358. 27. Kassem AS et al. Glycine and glycyl-glycine enriched oral rehydration solutions in the manage- ment of infantile diarrhoea. Alexandria journal of pediatrics, 1989, 3: 215-226. 28. International Study Group on Improved ORS. Impact of glycine-containing oral rehydration solu- tions on stool output and duration of diarrhoea: a meta-analysis of seven clinical trials. Bulletin of the World Health Organization, 1991, 69: 541-548. 29. Patra FC et al. Oral rehydration formula containing alanine and glucose for treatment of diarrhoea: a controlled trial. British medical journal, 1989, 298: 1353-1356. 954 WHO Bulletin OMS. Vol 72 1994 Improved ORS formulation: a review 30. Sazawal S et al. Alanine-based oral rehydration solu- tion: an assessment of efficacy in acute non-cholera diarrhoea among children. Journal of pediatric gas- troenterology and nutrition, 1991, 12: 461-468. 31. Gore SM, Fontaine 0, Pierce NF. Impact of rice- based oral rehydration solutions on stool output and duration of diarrhoea: a meta-analysis of 13 clinical trials. British medical journal, 1992, 304: 287-291. 32. Patra FC et al. Is oral rice electrolyte solution superior to glucose electrolyte solution in infantile diarrhoea? Archives of disease in childhood, 1982, 57: 910-912. 33. Molla AM et al. Rice-based oral rehydration therapy in acute diarrhoea: a superior therapy and a me- dium for calorie supplementation. In: Eeckles RE, Ransome-Kuti 0, Krooneenberg CC, eds. Child health in the tropics: Sixth Nutricia-Cow & Gates Symposium, Louvain, 18-21 October 1983. Dor- drecht, Nijhoff, 1985: 65-70. 34. Molla AM, Ahmed SM, Greenough III WB. Rice- based oral rehydration solution decreases the stool volume in acute diarrhoea. Bulletin of the World Health Organization, 1985, 63: 751-756. 35. Alam AN et al. Hydrolysed wheat based oral rehy- dration solution for acute diarrhoea. Archives of dis- ease in childhood, 1987, 62: 440-444. 36. Molla AM et al. Food-based oral rehydration salt solution for acute childhood diarrhoea. Lancet, 1989, 2: 429-431. 37. Bhan MK et al. Efficacy of mung bean and pop rice based oral rehydration solutions in comparisons with the standard glucose electrolyte solution. Journal of pediatric gastroenterology and nutrition, 1987, 6: 392-399. 38. Dutta P et al. Comparative efficacy of three differ- ent oral rehydration solutions for the treatment of dehydrating diarrhoea in children. Indian journal of medical research, 1988, 87: 229-233. 39. Man Mohan et al. Rice powder oral rehydration solu- tion as an alternative to glucose electrolyte solution. Indian journal of medical research, 1988, 87: 234-239. 40. Kenya PP et al. Cereal-based oral rehydration solu- tions. Archives of disease in childhood, 1989, 64: 1032-1035. 41. El Mougi M et al. Controlled clinical trial on the effi- cacy of rice powder based oral rehydration solution on the outcome of acute diarrhoea in infants. Jour- nal of pediatric gastroenterology and nutrition, 1988, 7: 572-576. 42. Fayad IM et al. Comparative efficacy of rice-based ORS plus early feeding versus glucose-based ORS plus early feeding. Lancet, 1993, 342: 772-775. 43. Razafindrakoto 0 et al. La solution de rehydratation orale (SRO) a base de riz, alternative de la SRO de l'OMS dans la diarrh6e aigue des malnutris. Archives fran9aises de Pediatrie, 1992, 50: 101-105. 44. Salazar-Lindo E et al. Bicarbonate versus citrate in oral rehydration therapy in infants with watery diar- rhoea: a controlled clinical trial. Journal of pediat- rics, 1986, 108: 55-60. 45. Nalin DR et al. Comparison of low and high sodium and potassium content in oral rehydration solutions. Journal of pediatrics, 1980, 97: 848-853. 46. Clements ML et al. Comparison of simple sugar/ salt versus glucose/electrolyte oral rehydration solu- tion in infant diarrhoea. Journal of tropical medicine and hygiene, 1981, 84: 189-194. 47. Sack DA et al. Oral rehydration in rotavirus diar- rhoea: a double-blind comparison of sucrose with glu- cose electrolyte solution. Lancet, 1978, 2: 280-283. 48. Black RE et al. Glucose versus sucrose in oral rehydra- tion solutions for infants and young children with rotavi- rus-associated diarrhoea. Pediatrics, 1981, 67: 79-83. 49. Santosham M et al. Oral rehydration therapy of infantile diarrhoea: a controlled study of well nour- ished children hospitalized in the United States and Panama. New England journal of medicine, 1982, 306: 1070-1076. 50. Sack RB et al. The use of oral replacement solu- tions in the treatment of cholera and other severe diarrhoeal disorders. Bulletin of the World Health Organization, 1970, 43: 351-360. 51. Akbar MS et al. A randomized double-blind clinical trial of a maltodextrin containing oral rehydration solution in acute infantile diarrhoea. Journal of diar- rhoeal disease research, 1991, 9: 33-37. 52. Santos Ocampo PD et al. A randomized double- blind clinical trial of a maltodextrin containing oral rehydration solution in acute infantile diarrhoea. Journal of pediatric gastroenterology and nutrition, 1993, 16: 23-28. 53. Daum F et al. Intestinal osmolality and carbohy- drate absorption in rats treated with polymerized glucose. Pediatric research, 1978, 12: 24-26. 54. Kerzner B et al. Jejunal absorption of short and long chain glucose oligomers in the absence of pancreatic amylase. Pediatric research, 1981, 15: 250-253. 55. Klish W et al. The effect of intestinal solute load on water secretion in infants with acquired monosac- charide intolerance. Pediatric research, 1980, 14: 1343-1346. 56. Fordtran JS, Rector FC, Carter NW. The mechanism of sodium absorption in the human small intestine. Journal of clinical investigation, 1968, 47: 884-890. 57. Bowling TE et al. Reversal by short-chain fatty acids of colonic fluid secretion induced by enteral feeding. Lancet, 1993, 342: 1266-1 268. 58. Lifschitz F, Wapnir RA. Oral hydration solutions: experimental optimization of sodium and water absorption. Journal of pediatrics, 1985, 106: 383-389. 59. Wapnir RA, Lifschitz F. Osmolality and solute con- centration - their relationship with oral hydration solution effectiveness: an experimental assessment. Pediatric research, 1985, 19: 894-898. 60. Molla AM et al. Rice powder electrolyte solution as oral therapy in diarrhoea due to Vibrio cholerae and Escherichia coli. Lancet, 1982, 1: 1317-1319. 61. Phillips AR. Water and electrolyte losses in chol- era. Federation proceedings, 1964, 32: 705-712. 62. Ribeiro HJ et al. Treatment of acute diarrhoea with oral rehydration solutions containing glutamine. Journal of the American College of Physicians, 1994, 13: 251-255. 63. Maulen-Radovan I et al. Safety and efficacy of a rice-based rehydration salt solution in the treatment of diarrhea in infants less than 6 months of age. Journal of pediatric gastroenterology and nutrition, 1994, 19: 78-82. WHO Bulletin OMS. Vol 72 1994 955