Study Patients The study was implemented in the Clinical Research and Service Centre of the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B), Centre for Health and Population Research, during June 1996 to January 1998. The children were selected from the outpatient department and admitted into a special study ward. Only boys between 5 to 12 months of age who presented with the following characteristics were admitted: (1) a history of frequent loose stools (.3 per day) for the last consecutive 14 days or more; (2) an absence of concurrent illnesses including severe infections (pneumonia, meningitis, sepsis, tuberculosis, and measles), and severe malnutrition (weight for age ,60% of National Centre for Health Statistics standard); (3) no history of treatment with antimicrobial or antidiarrheal agents within 7 days; and (4) ability to take oral feed. Informed, voluntary consents were obtained from parents before admitting their children into the study. This study was approved by the Ethical Review Committee of ICDDR,B.
Study Design This was a double-blind, controlled clinical trial of 62 children randomly assigned using a block randomization technique to 1 of 3 treatment groups: rice-based banana (n 5 22), rice-based pectin (n 5 19), or the rice-based diet alone (n 5 21). A computer-generated table of random numbers was used. The sample size estimation was based on an assumption of at least 40% improvement in the treatment groups at a 5% significant level with 80% statistical power. All enrolled patients successfully completed a 7-day treatment protocol in the hospital and were discharged after complete clinical recovery.
Clinical Management A complete history was recorded, and physical examination was performed by a physician. The child was placed on a “cholera cot” to facilitate collection and measurement of stool separate from urine. Urine was collected by applying polyvinyl urine collection bags. A physician evaluated the patient twice daily, specifically assessing the clinical severity of the illness. The patient’s nutritional status, body weight, weight of stool and urine, and their physical characteristics were recorded by specially trained nursing staff. Eight-hourly records of fluid intake (intravenous fluid and ORS) and output (stool, urine, and vomit) were maintained daily until discharge. After admission (1) children in the banana group were given a rice-based diet containing 250 g/L of cooked, green bananas; (2) children in the pectin group received the same rice-based diet containing pectin (Sigma, St. Louis, MO) at a dose of 4 g/kg body weight; and (3) children in the control group received the rice-based diet alone. The dose of pectin is based on a review of the literature, which indicated a wide range of dosages varying from 1 g/kg to 8 g/kg body weight.25 The diets were supplied in bottles to beds and were fed to the children by their mothers freely (ad libitum) under the direct observation of the nursing staff. All study diets were given to children for 7 days.
Preparation of Study Diets: Raw, green banana (Musa paradisiaca sapientum) was procured locally; the fruit was cooked in boiling water for 7–10 minutes with the skin intact. The skin was removed and the pulp blended to a thick homogenous consistency. One hundred grams (wet weight) of the pulp was mixed with 1.0 L of the cooked rice flour to prepare the rice-based diet, which was offered to children freely. The amount of rice flour added varied among the 3 diet groups to provide the same amount of calories per unit volume of each type of the study diet, i.e., 54 kcal/dL. The rice-diet was prepared by cooking locally available rice flour in 1.0 L of water and then adding other ingredients. Despite the differences in rice and banana contents, all final preparations looked like a white, homogenous suspension, without substantial variation in taste and density. However, to prevent against possible observers’ bias, we added 3 different coded flavors, i.e., vanilla, strawberry, and lemon, to each of the study diets. Moreover, at the bedside, stool measurements were made by the cleaning staff under the supervision of nurses, who had no knowledge of the type of treatment allocated to the study children by a different group of pharmacy staff. All staff were rotated at 8-hour shifts during the 24 hours. Breast-fed children continued to breast-feed during the study period. Vitamins and mineral supplements to meet 100% recommended daily allowance were provided daily.
Composition of Green Bananas The composition of green, unripe bananas, as described by Faisant et al.19 obtained by modern methods of chemical analysis is as follows (g/kg): protein 38, dietary fiber 92 (corresponding to cell materials), and total a-glucans 770, including oligosaccharides 60 (free glucose 47 plus oligosaccharides of dextrose polymer 2 to 12, 13) and insoluble starch 7–10. Thus, 30 g banana flour contained 23.1 (a-glucans), of which 1.8 g were oligosaccharides (including 1.4 g glucose) and 21.3 g insoluble starch.
Laboratory Investigations Stool or rectal swab specimens were taken on admission for isolation of enteropathogens (Shigella, Salmonella, V. cholerae, enterotoxigenic E. coli, rotavirus, ova, and parasites). Venous blood was obtained for determination of electrolytes (Na1, K1, Cl2, HCO32), glucose, protein, hematocrit, and total and differential white cell count.
Definitions of Outcome Variables “Liquid stool” was defined as feces taking the shape of the container and easily transferable without sticking to the container surface. The term “soft stool” was defined as being thicker, taking the shape of the container, but cannot be poured easily. “Formed stool” was defined when it retains its shape, does not stick to the container, and is characterized by the absence of mucus. Duration of diarrhea was defined as the number of hours elapsed from the start of the study diet until the passage of the last liquid stool, not followed by another abnormal stool within the next 24 hours.
Statistical Analysis Data in the 3 treatment groups were analyzed using analysis of variance in most instances, and x 2 significance test was used in comparing qualitative data. Kaplan-Meier survival analysis was used to compare treatment groups with regard to recovery from diarrhea; a P-value, 0.05 was considered statistically significant.