Rehabilitation of Malnourished Infants With F-100 or Diluted F-100

Study design: This is a randomized controlled trial comparing three recovery diets and will be conducted at a Nutrition Rehabilitation Unit (NRU).

Duration of the study: 24 months. Study site: The trial will take place at the Dhaka Hospital of ICDDR, Bangladesh. This hospital has a high case load, peaking during April-October, and the WHO guidelines for the management of severely malnourished children with minor modifications are followed. Diarrhoea is the entry point for all patients attending the hospital including children.

Methods During the stabilization phase or the acute phase management, children will be fed every 2 hourly with 130ml/kg/day F-75 along with other necessary management according to WHO guidelines (8). Judgment regarding suitability for transfer to the rehabilitation phase will be made by the regular doctors at the longer-stay ward according to set criteria: i) return of appetite (i.e. hungry, easily finishes a feed), ii) naso-gastric feeding is not required, iii) absent or insignificant oedema, iv) cessation of vomiting, v) less than 3 watery stools/day, and vi) no clinical signs of infection.

When ready for the rehabilitation phase, eligible children will be transferred to the NRU and randomized to receive any of the three diets (Infant formula, F-100 or diluted F-100). For the first 2 days on the test diets, they will be fed 130ml/kg/day in 12 (2-hourly) feeds and then 10ml more at each feed until some is left. This is equivalent to the transition phase in the WHO guidelines. During these 2 days the infants' pulse and respiration rates will be monitored 3 hourly to find out the signs of possible fluid overload. Thereafter they will be fed at 3 hour intervals. Infants will be fed by trained research personnel and will be encouraged to eat until full. Being 'full' will be defined as two successive refusals when feed is re-offered. Water will be freely available between feeds. Breastfed children will continue to be breastfed. Pulse and respiration rates and temperature will be monitored every 8 hours. If, any of the infants shows the signs of fluid overload, they will be transferred to longer-stay or special care ward and will be managed according to the standardized management protocol practiced at Dhaka Hospital of ICDDR,B.

All the children who are accompanied by mothers will be counselled by professional breastfeeding counsellor for relactation if the child is already weaned from the breast.

Measurement of the diets: All feeding episodes will take place under the supervision of study personnel. Food portions will be weighed before and after serving using an electronic balance with one-gram precision, and the actual amount consumed will be calculated by subtracting leftover food from the amount offered. Pre-weighed napkins will be provided; any food that is regurgitated, vomited or spilled will be swabbed, weighed and subtracted from the amount offered.

Measurement of breast milk intake: If, the particular child is breastfed or the re-lactation has been established, then the mothers will be requested to breastfeed day and night, according to the usual feeding pattern at home prior to the study or if the child demands. At each nursing, breast milk will be offered until the child refuses further intake, and the amount consumed will be measured by test-weighing before and after nursing, while ensuring that the child wears the same clothing for each weighing. Each nursing episode initiated after taking a post-feeding weight will be considered a new episode, and the intake will be similarly measured. Breast milk intake will be measured by recording the difference in the child's weight before and after the feed, using an electronic balance with two-gram precision, and the weight change will be adjusted for child-specific insensible losses, which will be measured on the same balance for at least 30 minutes each day. The duration of each episode of breast-feeding will be measured using a stopwatch, and the duration and frequency of breastfeeding in 24 hours will also be recorded.

Discharge criteria: Body weight gain 15% of the admission bodyweight or as soon as the infant becomes oedema free or if the child has an oedema free weight-for-length Z-score ≥ 2.

Follow-up: At the time of discharge, the mother/caregiver will be asked to bring his/her child to the nutritional follow-up unit after 2 weeks, then monthly for 3 months. In every visit the child will be measured for weight, height and mid-upper arm circumference and will also be examined for any possible illness(s). If, any of the child fails to come for scheduled follow-up, study personnel will visit their respective homes to know the well-being and status of the subjects.

Deuterium Oxide Dilution Technique for the Assessment of Body composition:

The following procedure for assessment of body composition will be conducted on the day the infant will be transferred to the NRU and repeated at the time of discharge.

Deuterium oxide (2H2O) will be used for body composition assessment. Upon enrolment, baseline saliva samples will be collected. After collection of baseline saliva samples 5g of deuterium oxide labelled water will be fed to the study infant. Children should avoid drinking water during the equilibration period, i.e., 3-4 hours after ingestion of the stable isotope dose. According to the recommendation from International atomic Energy Agency (IAEA) field manual (9), two post-dose saliva samples will be collected at 3rd and 4th hours. The concentration of deuterium oxide in saliva samples will be measured with Fourier Transform Infrared Spectrometer (FTIR) and for total body water (TBW) estimation. Afterward from measured TBW, the amount of fat free mass (FFM) can be estimated. Body fat mass (FM) is the difference between body weight and fat free mass (10).

Therefore for each child, saliva samples will be collected 6 times (2 baseline and 4 post dose) during the course of NRU stay.

Calculation of body composition:

The dilution space of 2H (VD) is 4.1 % higher than TBW due to exchange of H with non-aqueous H in the body TBW (kg) = VD /1.041 Where, VD (kg) = Dose 2H2O (mg)/enrichment 2H in saliva (mg/kg) FFM (kg) = TBW (kg) / hydration coefficient*

• Hydration factor = Hydration of FFM (%FFM)/100 Fat mass (FM) is calculated by difference between body weight and FFM FM (kg) = body weight (kg) - FFM (kg) Results are often expressed as % body weight FM (%) = FM (kg) / body weight (kg) x 100
• Please see Table 2; Hydration factor for children (11, 12) Randomization After assessing suitable for initiating the rehabilitation phase, infants will be randomized to the test diets in blocks of 6 using a random numbers table (EPI-INFO 6.04, CDC Atlanta, GA). Allocation will be concealed in sealed envelopes. The codes will be kept secure and not broken until data analysis is complete. Medical and nursing staff and the children's carers will be blind as to the formula given. Infant formula, F-100 and diluted F-100 will be similar as much as possible in appearance and will be prepared by the dietician. The invitation to participate in the research will be given at admission before randomization.

Data collection

1. Rate of weight gain (g/kg/day): Body weight will be measured daily with a digital scale, precision 2g. Calculation of weight gain (13,14,15): Weight gain in g/kg.day = (W2 - W1) x 1000/(W1 x number of days from W1 to W2) Where:

• W1 = lowest weight during the NRW stay in kg (usually the 1st day of NRW-weight in case of marasmic children, but it may vary in case of oedematous children)
• W2 = weight in kg on the day of calculation

1. Intake (kcal/kg/day): During the stay at NRU, amounts of the test diets offered and eaten, and water consumed, will be recorded by the research assistants.
2. Renal solute load (mOsmol/l): Potential renal solute load: PRSL = N/28 + Na + C1 + K + Pa Where, N= Nitrogen content of the diet (mg/28=mosm) Na= Sodium content of the diet (mg/23=mosm) Cl= Sodium content of the diet (mg/35=mosm) K= Potassium content of the diet (mg/39=mosm) Pa= Available Phosphorous content of the diet (mg/31=mosm) 0.9 mOsm of PRSL are retained per gram of weight gain (4). RSL will, therefore, be estimated using the equation: RSL est = PRSL - (0.9 x gain) Where: RSL est and PRSL are expressed as mOsmoles per day and gain is weight gain in grams per day (7).
3. Serum sodium and plasma osmolality: Two ml blood will be taken from antecubital vein of the child on days 1, 3, and 7 for measuring serum sodium, potassium, chloride, total carbon dioxide, creatinine and osmolality. This will allow monitoring of osmolar state of the child i.e. hypernatremia. Spot urine samples on the above mentioned days will be checked for specific gravity and osmolality/creatinine ratio. Urine will be collected in paediatric urine collector bags. Standard methods will be used for measuring electrolytes, creatinine, osmolality and specific gravity in the Clinical Biochemistry Laboratory of ICDDR,B.
4. Extra-renal water losses: Loss of body fluids due to diarrhoea and vomiting will be recorded, as will fever, environmental temperature and humidity.

Outcome measures:

1. Rate of weight gain (g/kg/day)
2. Food Intake (kcal/kg/day)
3. Breast milk intake (if the child is breastfed) (g/kg/d) and (kcal/kg/d)
4. Renal solute load (mOsmol/l):

Sample size estimation:

In NRU of Dhaka Hospital of ICDDR,B, 124 children suffering from SAM who were managed with infant formula along with other recommended WHO-guidelines had a daily mean ± SD weight gain of 10.25 ± 6.7 g/kg/day. If, we assume that with F-100 or diluted F-100 the rate of weight gain will be 15 g/kg/d, then we require 40 children in each group with 80% power and 5% level of significance. If, the investigators further consider 10% attrition then each group will require 44 infants. So, the investigators will study 44 infant in each group giving a total of 132 children in three study groups.

data Analysis: All recording forms will be pre-coded and checked daily for completeness and consistency. Differences among the groups will be assessed using chi-squared tests for proportions and analysis of variance (ANOVA) for normally-distributed continuous variables. Paired t-test for paired comparison before and after the intervention will be performed. The analysis will be on an intention-to-treat basis. Statistical analysis will be with the Statistical Package for the Social Sciences, version 11.5 for Windows (SPSS Inc, Chicago, Ill). Subgroup analyses will be conducted to test if age has a differential effect on renal solute load among the groups.