This open label randomized controlled trial was conducted from June, 2017 to June, 2018, in the department of pediatrics of a tertiary care teaching hospital in India. Children aged 9 months to 5 years, presenting within 48 hours of onset of acute watery diarrhea (AWD) with no dehydration (based on WHO criteria) [3] and whose parents were willing to admit their children in the hospital for three days were considered for the study. It was ensured that the parents did not have any financial burden due to the hospital stay. From among these, children who could not be fed orally, who were under-nourished (weight for age Z score <-2) [7], those having any concurrent or pre-existing severe illness, and those who had received any antibiotic or anti-motility agents in the last seven days were excluded. Children with blood in stool were also excluded. After getting written informed consent from the parents, one of the investigators randomized the eligible children into two groups based on block randomization with variable block size using computer-generated randomization sequence. Allocation concealment was done using sequentially numbered, opaque, and sealed envelopes. All the investigators enrolled and assigned the participants to interventions. Blinding could not be done in view of the logistics involved. Assuming the proportion of children with formed stools at 72 hours (primary outcome) to be 60% in the control arm [4] and 80% in the intervention arm (absolute difference of 20%), the sample size estimated with a power of 90% and alpha error of 5% was 105 in each group. Considering a dropout rate of 20%, it was decided to recruit 125 children in each group.

Children in the Control group, received standard care for acute diarrhea (zinc sulphate 20 mg orally once daily for 14 days, plus frequent breast milk, if on breast feeding, plenty of home available fluids, such as rice kanji, butter milk or tender coconut water, ORS 10 mL/kg/loose stool, and regular diet). Children in Green banana supplemented diet (GB) group received standard care, as above, and cooked green banana, in addition. The raw green banana was boiled for ten minutes following which the skin was peeled and the pulp was mashed to semi-solid consis-tency using a spoon. The cooked green banana was offered either alone (with little salt for taste) or mixed with any other food of their choice; the dose of green banana was 50 g twice daily for children younger than 1 year, 100 g twice daily for those aged 1-3 years, and 100g thrice daily for those aged 3-5 years [4]. The green banana supplemented diet was continued until diarrhea stopped or till the 14th day of illness, whichever was earlier. Children who developed dehydration were managed as per WHO protocol [3]; other complications were treated uniformly as per the institutional protocol. Children who persisted to have diarrhea beyond 72 hours from the beginning of intervention were advised to continue the standard treatment at home. The parents of the GB group were also advised to give cooked green banana daily until cessation of diarrhea – the cooking method was explained to the parents prior to discharge. Compliance for home available fluids, ORS and zinc for both the groups and in addition cooked green banana for the intervention group was ensured by daily telephonic contact until the 14th day of illness and the details were collected. The parents were asked to bring the child for review if there were any concerns on telephonic discussion. Children with diarrhea persisting for more than 14 days were also reviewed again in the hospital.

The primary outcome assessed was the proportion of children who improved at 72 hours of intervention (passing formed stools with the usual frequency), as observed by the mother and verified by one of the investigators. The incidence of dehydration, persistent diarrhea (diarrhea persisting beyond 14 days), lactose intolerance (as evidenced by perianal excoriation and positive stool reducing substances test) and any other complications were evaluated as the secondary outcomes. The study was approved by the Institutional ethics committee, and the trial was prospectively registered in Clinical Trials Registry of India.

Statistical analyses: Data collection was done with the help of a semi-structured pretested proforma and was transcribed into Microsoft Excel spread sheet. Data analysis was done using the software STATA version 12. The baseline socio-demographic variables were com-pared between groups using chisquare test. An intention to treat analysis was performed. The proportion of children recovering within 72 hours, and the development of complications in both the groups were compared using the chisquare test. In addition, time to event analysis was carried out for recovery using Kaplan Meier survival graphs and the recovery time between the groups was compared using log rank test. All statistical testing was carried out at 5% level of significance.

Of the 250 children recruited, 108 children in the green banana group and 116 children in the control group completed the study (Fig. 1). There was no significant difference between the groups with respect to the base-line demographic characteristics, socioeconomic status, duration of the illness (between the onset of illness and reporting for treatment), fever, vomiting and the pre-valence of breastfeeding (Table I). Children in the intervention group received GB diet for a median (IQR) duration of 3 (2) days (range, 2-14 day of illness). depending on the duration of diarrhea. The proportion of children recovering within 72 hours was significantly higher (62.4%) in the GB group as compared to the Control group (47.2%) [RR (95% CI) 1.3 (1.05-1.7), NNT=7; P=0.002] (Table II). Moreover, on comparing the time to recovery using Kaplan Meier graph, the pro-portion of children with diarrhea at 72 hours was significantly less in the green banana group (Fig. 2).

Fig. 1 Consort flow diagram depicting study methodology.

Table I Baseline Characteristics of Children With Acute Diarrhea Enrolled in the Study (N=250)

GB: green banana.

Fig. 2 Kaplan Meier graph showing recovery time in the two groups.

During the three days of stay in the hospital, none of the children in the two groups developed severe dehydration. Significantly higher number of children in the control group developed some dehydration (P=0.006) and persistent diarrhea (P=0.01) (Table II). None of the children in either group developed any other complications, including acute kidney injury.