Burden: According to Bangladesh Demographic Health Survey 2014, prevalence of stunting in under-five children is 36%. Severe acute malnutrition (SAM) is prevalent in about 3% of them. In absolute numbers, about 450,000 children suffer from SAM while several million children suffer from moderate acute malnutrition (MAM).

Knowledge gap: Investigators have already demonstrated that children with SAM have immature gut microbiota that is partially corrected with treatment. Through our earlier studies conducted in Bangladesh,investigators now also have idea about beneficial microbiota and food ingredients that support proliferation of these beneficial microbiota. Which composite food is better for children with SAM and MAM is not known.

Hypothesis (if any): The central hypothesis of our ongoing Breast Milk, Gut Microbiome and Immunity (BMMI) initiative proof of concept phase is that healthy growth in infancy and early childhood requires normal functional maturation of the gut microbiota. Related hypotheses are that (i) persistent defects in the development of this microbial 'organ' are causally related to stunting as well as the metabolic, immunologic and cognitive manifestations of undernutrition; and (ii) new approaches for durable repair of this microbiota immaturity including Microbiota-Directed Complementary Foods (MDCF) will provide a way to improve clinical outcomes.

Objectives:

1. Test 5 MDCF prototypes and nominate a lead MDCF formulation that has the greatest effect in promoting the representation of a broad range of age-discriminatory taxa and that has acceptable organoleptic properties. This nominated lead will be advanced to a fully powered clinical POC study in children with post-SAM MAM in 2018.
2. Determine whether once daily administration is as effective as twice daily administration of the lead MDCF in repairing microbiota immaturity.
3. Assess the durability of repair of immaturity by the lead MDCF by including a 4 week post-intervention phase in the final pre-POC study design.
4. Determine what effect enteropathogen burden (determined by PCR-based analysis of fecal samples) has on responses of age-discriminatory taxa to MDCFs, and reciprocally, the effect of MDCFs on enteropathogen burden.

Methods:

Investigators will conduct 3 Pre-POC studies in 12-18 month old children with MAM (Weight-for-Length Z-score, WLZ: between <-2 to -3) and stunting (Length-for-Age Z-score, LAZ: between <-2 to -3) over the course of approximately two years. The investigators will use a stratified randomization procedure based on age (i.e.2 levels: 12-15 months and >15-18 months) to prevent imbalance between the treatment arms. This study will be undertaken at Mirpur area of Dhaka city. The investigators will design and produce 5 MDCF prototypes at the icddr,b Food Processing Laboratory in sufficient quantities for clinical studies (3 MDCF prototypes for study 1 and 2 MDCF prototypes for study 2). These formulations will be matched in energy density and micronutrient content of ready-to-use supplementary foods (RUSFs) used for MAM in Bangladesh and other countries, and will meet all other requirements for a complementary/supplementary food for 12-18 month old children with MAM and stunting.

Investigators will conduct 3 Pre-POC studies among 12-18 month old children with MAM (WLZ <-2 to -3) and stunting (LAZ <-2 to -3) over the course of approximately two years.

Study design: Randomized controlled intervention trial.

Study site: This study will be conducted at Bauniabadh, Mirpur in Dhaka city, which is a peri-urban slum area.

Study participants: The study participants will be 12-18 months old children of either sex with MAM (WLZ: <-2 to -3) and stunting (LAZ: <-2 to -3).

Initial screening and enrolment: Children will be screened and enrolled by household survey in section 11 of Mirpur by Field Research Assistants (FRAs) following pre-specified inclusion criteria. Fulfilling the enrolment criteria and upon receiving the consent for study participation from the parents or legal guardians, the children with the respective mother/caregiver will be enrolled and randomly assigned to one of the four arms according to computer-generated random numbers using permuted blocks of varied lengths of four and twelve for study 1. Similarly, children and mother/caregiver will be enrolled and randomly assigned to one of the three arms according to computer-generated random numbers using permuted blocks of block lengths of three and nine for study 2 and 3 respectively. The code of assigned type of MDCFs will be kept in closed opaque envelopes until assigned to an enrolled child.

Feeding sessions: The children and mothers/caregivers will be requested to come to the nutrition centre established in the community at Mirpur preferably between 9 to 11 am and 3-5 pm on day 1. The mothers will be requested not to give/allow any food and breast milk in the preceding 2 hours of observed meal time on the day of observed feeding. On the same day at the nutrition centre, the child will be offered 25 grams (in two meals preferably between 9-11 am and 3-5 pm) of MDCFs as decided by random allocation. During the feeding time, the mothers will be asked to spoon feed the respective pre-weighed diets to their children until s/he refuses to eat further, as described below. After a two-minute pause, the same diet will be offered a second time until s/he refuses again. After a second two-minute pause, the diet will be offered a third time until refused again. After this third refusal, the feeding episode will be considered as 'terminated'.7 The duration of the feeding (excluding the intervening 'rest periods') will be recorded by stopwatch, and the total duration of the feeding will be noted. If the child demands more after finishing all the food, she/he will be offered the amount that was offered at the beginning. This feeding episode will last for maximum 60 minutes. Measured volumes of plain water will also be given and the amount of water taken during this 60 minutes meal period will be measured. The feeding episode will take place under the direct supervision of trained study personnel. Children will be considered as refusing further intake if they move their head away from the food, cry, clamp the mouth or clinch the teeth, or become agitated, spit out the food or refuse to swallow.8 The amount of MDCF actually ingested will be calculated by subtracting the left over from the offered amount. Pre-weighed napkins will be provided; any food that is regurgitated, vomited or spilled will be swabbed, weighed and subtracted from the amount offered.

Acceptability will be categorized as 'accepted eagerly' if they ate food readily, did not make any fuss, spit out, vomit or cry during the observed meal. Children will be categorized as 'accepted but not eagerly' if they ate the offered food but either made fuss, spit out, vomited or cried during the observed meal. The third category will be children who will not consume the offered food at all.9

The amount of consumed food (g), energy (kcal) and category of acceptability will be analyzed. The enrolled children will be monitored by daily home visits by trained Field Field Research Assistants for any possible side effects/adverse events (e.g. rash, urticaria due to food allergy or any significant changes in clinical status) for a week. If any side effects/adverse events are observed then they will be treated according to the standard management provided at Dhaka Hospital of icddr,b. Perception and assessment of organoleptic characteristics of the assigned MDCFs by the respective mother/caregiver will also be assessed by using the 7-point Hedonic scale. Higher value, that is 7, will be ranked as 'like extremely' and lowest value 1 will be ranked as 'dislike extremely'.10

At the beginning of the study, information will be sought on the demographic characteristics (families' wealth, standard of housing, family structure and parental characteristics etc.), and trained FRAs will record the children's nude weight using a digital scale with 2 g precision (Seca, model 728, Germany), length (using infantometer, Seca, model 416, Germany), and mid upper-arm circumference to the nearest mm (using a non-stretch insertion tape). Anthropometrics will be done according to the standard procedures and all measurements will be taken thrice and the middle one will be recorded. All interventions in each study will be administered to children at the Mirpur health clinic. Mothers/primary caregivers will be be advised to maintain their child's current dietary and breastfeeding practices.

Pre-POC Studies 1 and 2 will each be 8 weeks in duration and have 4 and 3 arms respectively, with 12-15 children enrolled per arm with a minimum of 10 children per arm completing the study. There will be a 2 week run-in period to define pre-treatment MAZ, and additionally, enteropathogen burden will be assessed by PCR-based measurement of a broad range of gastrointestinal pathogens in the same fecal specimens used to calculate MAZ scores (these will not be used as screening tools for subsequent treatment). Children will be randomly assigned to one of the three arms. Efforts will be made to balance arms with respect to gender and stratify the randomization on age (e.g. 2 levels; 12-15 months and >15-18 months). The treatment phase will be for 4-weeks, during which time each test product will be administered at the test center twice daily, separated by at least 6 hours. There will be a 2-week post intervention phase in which 2 additional fecal samples samples will be collected (at the end of weeks 7 and 8) to assess changes in the microbiota that occur after cessation of treatment.

Pre-POC Study 3 will have 3 arms. It will test the lead MDCF identified in study 2 and benchmark its efficacy at repairing microbiota immaturity compared to the current RUSF standard of care over a 4 week intervention period. We will examine the impact of 1x vs 2x per day dosing of the lead MDCF on efficacy in this study and also include a 4-week post-intervention period (with fecal sampling at the end of weeks 7, 8, 9 and 10) to assess the durability of changes in microbiota maturity after cessation of intervention and return to home diet.

Fecal sample collection Weekly fecal samples (1-2g) will be collected within 30 minutes of production (at home and transported back to the icddr,b) using the N2 dry shipper specimen collection SOP used in the recent study of local RUTFs for the treatment of SAM (see Appendix A). Specimens will be stored at -80◦C before being sent to the Gordon Lab at the end of each study for rapid analysis of microbiota maturity (measured before, during and after cessation of treatment with each MDCF prototype) and PCR-based assessment of enteropathogen burden in fecal samples (measured before and after MDCF treatment). The relative abundance of each of the bacterial strains that were specifically targeted by the next generation MDCF will be determined through analysis of the 16S rRNA datasets (performed in the Gordon Lab and funded by the primary BMMI Project POC Phase grant)

Proposed Pre-POC Studies

Study 1: Test 3 prototype MDCFs and the current rice-lentil RUSF standard of care for MAM to establish the effect size of each on MAZ repair in a 4 week 2x /day intervention, with a 2 week post-intervention phase to assess durability of MDCF-induced changes in the microbiota:

Arm 1 - Current RUSF used for MAM, administered 2x /day for 4 weeks Arm 2 - Prototype MDCF1, administered 2x /day for 4 weeks Arm 3 - Prototype MDCF2, administered 2x /day for 4 weeks Arm 4 - Prototype MDCF3, administered 2x /day for 4 weeks

Arm 1 - Rice-lentil based RUSF (rationale: reference standard of care for MAM; not designed based on knowledge of its effects on the gut microbiota or microbiota immaturity).

Arm 2 - MDCF prototype with four complementary food ingredients plus powdered milk (rationale: lead with evidence from preclinical gnotobiotic models of repair of microbiota immaturity and positive effects on growth).

Arm 3 - MDCF prototype with four complementary food ingredients without powdered milk (rationale: determine whether withdrawal of milk impacts the effects of the four complementary food ingredient combinations on the microbiota and host biomarkers).

Arm 4 - MDCF prototype with two complementary food ingredients (chickpea and soy flour) without powdered milk (rationale: compare the efficacy of this formulation to that of the MDCF formulation containing four complementary food ingredients which is administered to group 3).

Fecal sample collection:

Fecal samples will be collected from each child at enrollment, 1 week after enrollment, at randomization (2 weeks) and weekly during the intervention and post-intervention phases.

-Assuming 15 study participants / arm and 4 arms, a total of 540 fecal specimens will be collected in Study1.

Blood sample collection:

2 mL of blood samples each will be collected from every child prior to intervention and just after the intervention is complete.

-Assuming 15 study participants /arms, a total of 120 blood samples will be collected in Study1.

Study 2: Select most efficacious MDCF from study 1 and compare with 2 additional MDCF prototypes using the same design used in study 1:

Arm 1 - 'Winner' from study 2 administered 2x /day for 4 weeks Arm 2 - Prototype MDCF4 administered 2x /day for 4 weeks Arm 3 - Prototype MDCF5 administered 2x /day for 4 weeks

Fecal sample collection:

Fecal samples will be collected from each child at enrollment, 1 week after enrollment, at randomization (2 weeks) and weekly during the intervention and post-intervention phases.

-Assuming 15 study participants / arm and 3 arms, a total of 405 fecal specimens will be collected in Study 2.

Blood sample collection:

2 mL of blood samples each will be collected from every child prior to intervention and just after the intervention is complete.

- 15 study participants /arms, a total of 90 blood samples will be collected in Study 2.

Study 3:

Select lead MDCF from studies 1, 2 and conduct final 'bake-off' vs current RUSF and also examine the impact of 1x vs 2x per day administration with a 4 week post-intervention period to provide additional information on the durability of MDCF-sponsored changes in the microbiota:

Arm 1 - Current RUSF used for MAM 2x /day for 4 weeks followed by a 4 week post-interventional phase Arm 2 - Lead MDCF identified in study 2, 2x /day for 4 weeks followed by a 4 week post-interventional phase Arm 3 - Lead MDCF identified in study 2, 1x /day for 4 weeks followed by a 4 week post-interventional phase

Fecal sample collection:

Feces would be collected weekly from each study participant during the course of the 8-week study including the 2 weeks prior to treatment, the four weeks during treatment and the two weeks following cessation of treatment (11 samples/ child).

-Assuming 15 study participants / arm and 3 arms, a total of 495 fecal specimens will be collected in Study 3.

Assuming a 8-week study duration (studies 1 and 2) and 8 weeks for shipping of fecal specimens to Washington University, extraction of DNA, PCR amplification of the V4 region of the bacterial 16S rRNA genes, sequencing of these amplicons and data analysis, each study cycle from completion of enrollment to completion of data analysis is anticipated to be 16 weeks. With this cycle time, we anticipate being able to evaluate candidate MDCFs by the end of June 2017 and complete all 5 by the end of April 2018, at which point a lead MDCF will be nominated for the POC study.

Information gathered will be used to select human fecal samples for transplantation into germ-free animals; these animals, who will be fed the diets of their human microbiota donors, will be used to further characterize the mechanisms that link MDCF prototypes, the gut microbiota, and host physiology/metabolism.

Targeted Ultra Performance Liquid Chromatography-Mass Spectrometry (UPLC-MS) and Gas Chromatography-Mass Spectrometry (GC-MS) will be used to profile analytes of specific interest in plasma and/or fecal samples including bile acids and short chain fatty acids (SCFAs); markers of mitochondrial function (e.g., β-hydroxybutyrate, acylcarnitines/acylCoAs, TCA cycle intermediates); amino acids in fecal samples [branch chain amino acids, tryptophan and tryptophan metabolites related to growth and inflammatory status, including those produced by bacterial tryptophan metabolism (e.g. indole acetic acid derivatives)].

Blood sample collection:

2 mL of blood samples each will be collected from every child prior to intervention and just after the intervention is complete.15 study participants / arm and 3 arms, a total of 90 blood samples will be collected in Study 3.

In order to construct a library of gut microbiota of healthy growing children of the same community, one spot fecal sample (1-2 gm) and spot blood sample (2 mL) will be collected from 30 children each who would be aged 12-18 months of either sex, having WLZ and LAZ : >-1. This will be done only after obtaining consent from the study participants.

Through our initial lab work as well as earlier studies done on gnotobiotic animals in Washington University Centre for Genome Science has led us to suggest that a combination of food ingredients (chickpea, soy flour, peanut and green banana) with or without milk powder will be worth studying with respect to the diet's impact on stimulating proliferation of growth-discriminatory microbiota as well as cost and sustainability.

Study participants will be asked to come to the designated feeding center daily for nutritional therapy to avoid the issue of food sharing. They will be provided with a phone number to reach the clinic staff, and a member of the staff may visit a family's household for directly observed nutritional therapy with prior arrangement if needed.

Five day food frequency questionnaires will be completed at enrollment, prior to randomization and weekly during the entire study period.

Preparation of different MDCFs:

Preparation of complementary foods for Pre-POC clinical studies:

Based on compatible combinations of complementary food ingredients identified in the preclinical studies described above, investigators will design and produce 5 MDCF prototypes at the icddr,b Food Processing Laboratory in sufficient quantities for clinical studies (3 MDCF prototypes for Study 1, 2 MDCF prototypes for Study 2 and the qualified MDCF from Study 2 for Study 3). These formulations will be matched in energy density and micronutrient content with ready-to-use supplementary foods (RUSFs) used for MAM in Bangladesh and other countries, and will meet all other local requirements for a complementary/supplementary food for 12-18 month old children with MAM and stunting. The icddr,b Food Processing Laboratory will also produce the current rice-lentil RUSF standard of care used for treatment of these children for use in studies 1 and 3.

Development of the product

1. Identification of food ingredients: Locally available food ingredients including those identified in the preclinical diet oscillation studies (e.g. chickpeas, soy flour, peanuts, raw banana, soybean oil, sugar and with and without milk powder) will be used to manufacture prototype MDCFs.

   MDCFs will be fortified with a premix of vitamins and minerals to meet the micronutrient requirements for supplementary foods in Bangladesh (see below).

2. Experiments on recipe development: The theoretical formulation of MDCF components shall be made based on linear programming to identify the combinations of ingredients that would result in the most nutritious recipes. The energy density of MDCF is targeted at ~5 kcal/g, and caloric distribution is targeted to be 45-50 percent from fat and 8-12 percent from protein. Based on expert opinion and consensus within the research team, micronutrient content shall be set to cover 70 percent of the requirements of Non SAM children aged 12-18 months. Experiments for developing recipes and preparation of samples will be done at icddr,b Food Processing Laboratory following a standardized production procedure to control the quality of MDCF from each production batch and ensure that no unexpected contamination and nutrient loss occur during processing. All the ingredients will be cleaned and sorted. Potential recipes shall be produced in small batches after (cooking/ steaming/ roasting as appropriate for each of the ingredients) and by mixing all ingredients in an electric blender. When necessary, consistency of the recipes shall be adjusted by varying the amount of dry ingredients and soybean oil. Furthermore, the combination of minerals and vitamins shall be adjusted if there is any unpleasant taste for addition of high dose(s) of micronutrients. A small amount (1 percent) of soy lecithin shall be added to the recipe in order to improve the consistency and prevent oil separation.

3. Assessment of organoleptic properties of the recipes: The new recipes will be assessed for appearance, flavor, consistency of the diets and overall degree of liking by using the 7-point Hedonic scale. Higher value, that is 7, will be ranked as 'like extremely' and lowest value of 1 will be ranked as 'dislike most'.

4. Study of food safety and shelf life issues: Investigators will evaluate the stability of the product including microbiological quality to obtain key information needed for setting up initial 'best before' date and assess the stability of vitamin and mineral content and potential for oxidation. Standard microbiologic tests (total viable count, yeasts, moulds, coliforms, Escherichia coli, Bacillus cereus, Staphylococci, Listeria monocytogenes, Cronobacter sakazaki) will be performed at the icddr,b Food Safety Laboratory.

5. Finalization of 5 prototype MDCFs for the pre-POC clinical trials: Based on all of the above mentioned criteria, our team of researchers will select 5 MDCFs for acceptability trials. investigators will also estimate the cost/kg of the ingredients of the candidate MDCFs.

Training of field staff Training will be provided to all field team members on different procedures regarding collection of stool samples. The training will be facilitated by core group of trainers as well as investigators of the study.

Anthropometry:

The age of the child will be verified against documentation (birth certificate or immunization card, if available) or caregiver's report of the child's birth date. Length will be measured by a infantometer sensitive to 0.1 cm (SECA, model no. 4161721009, Hamburg, Germany). Body weight will be measured by a balance sensitive to 2 g (SECA 7281321009, Hamburg, Germany). Length-for-age (LAZ), weight-for-length (WLZ) and Weight-for-Age (WAZ) Z-scores will be calculated calculated following the Multicentre Growth Reference Study (MGRS) WHO growth standards 11. Edema will be examined by pressing the upper side of both feet for 3 seconds. Mid-upper arm circumference (MUAC) will be measured using TALC MUAC tape (UK). Standardization of the measuring equipments: Regular standardization of the baby weighing scale, and the length/length board with a known weight and length will be done.

Analyses of plasma and fecal samples

Plasma samples collected this exploratory pre-POC study will be sent to Jeffrey Gordon's lab located in the Center for Genome Sciences and Systems Biology at Washington University in St. Louis. Advanced mass spectroscopic- and immunoassay-based methods will be used to obtain new knowledge about the role of gut microbiota immaturity and the effects of attempting acute repair of this immaturity with lead microbiota-directed complementary food (MDCF) prototypes on biomarkers and mediators of healthy growth.

This initial test of the effects of MDCFs on microbiota immaturity and host responses to dietary intervention will provide key datasets that will enable a power calculation for our planned proof of concept study in a larger population. In the current exploratory pre-POC each child in each group of 12-15 study participants will serve as his/her own control. Additional comparisons will be made within and across the four treatment groups (i.e., those consuming reference RUSF standard, and one of three different MDCF formulations).

Feces would be collected weekly from each study participant during the course of the 8-week study including the 2 weeks prior to treatment, the four weeks during treatment and the two weeks following cessation of treatment. In addition, a 2 mL sample of blood would be obtained just prior to initiating the 4 week MDCF intervention and just as the 4-week long intervention is completed.

Targeted Ultra Performance Liquid Chromatography-Mass Spectrometry (UPLC-MS) and Gas Chromatography-Mass Spectrometry (GC-MS) will be used to profile analytes of specific interest in plasma and/or fecal samples including bile acids and short chain fatty acids (SCFAs); markers of mitochondrial function (e.g., β-hydroxybutyrate, acylcarnitines/acylCoAs, TCA cycle intermediates); amino acids in serum plus fecal samples [branch chain amino acids, tryptophan and tryptophan metabolites related to growth and inflammatory status, including those produced by bacterial tryptophan metabolism (e.g. indole acetic acid derivatives)].

In addition, key mediators/biomarkers of linear growth (e.g., growth hormone and IGF-1), energy utilization (insulin, leptin), and bone biology [IL-6, osteoprotegerin, the C-terminal peptide of type I collagen (CTX, a marker of osteoclast activity/bone resorption), and the amino-terminal propeptide of Type 1 procollagen (P1NP, a marker of osteoblast activity/bone formation)], and systemic inflammation (CRP, AGP) will be quantified using established ELISA/Luminex assays.

Information gathered will be used to select human fecal samples for transplantation into germ-free animals; these animals, who will be fed the diets of their human microbiota donors, will be used to further characterize the mechanisms that link MDCF prototypes, the gut microbiota, and host physiology/metabolism. In addition, these plasma and fecal biomarkers will be used to determine the features of our preclinical models are best correlated with responses in children; this knowledge will help inform the design of next generation MDCF prototypes with superior efficacy on growth outcomes.

Sample size calculation:

The primary outcome of this study will be the change in of Microbiota-for-Age Z (ΔMAZ) score after 4 weeks of intervention.. Our previous work5 showed that the mean MAZ of 18-month-old children with MAM was -1.60+0.36 (SEM) while for age-matched healthy children, the mean MAZ was -0.52+0.22, with a difference between means of 1.08+0.41. Using R's power.t.test function for a paired t-test (i.e. pairing samples from a child before and after intervention), with and assumed SD of 1 and a one-sided hypothesis, 15 children in each arm will enable us to capture an improvement in MAZ of >0.7 with 80% power and a significance threshold of 0.05. Investigators therefore anticipate being sufficiently powered to confidently detect a change in MAZ of the magnitude that distinguished the microbiota of healthy children from that in children with MAM in our prior study

Data analysis :

Data analysis will be done with the same approach used for calculating microbiota maturity and the application of these tools in a prior RUTF intervention study treating children with SAM at icddr,b5.

The primary outcome measure in each study is the mean treatment-induced change in MAZ score from baseline for each arm in the study. Differences (95% CI) between the means of treatment arms will be assessed using one-way ANOVA followed by Tukey's HSD or similar post hoc test (when appropriate). The MDCF prototype with the greatest mean effect size the in these pre-POC studies, and acceptable organoleptic properties will be selected for the follow-on POC clinical trial to examine clinical outcomes in children with post-SAM MAM.

Inclusion Criteria:

• All of the following criteria must be met for a study participant to be eligible to participate in the study -
• Parent(s) willing to sign consent form
• Child age 12-18 months and no longer exclusively breast fed
• WLZ between <-2 and -3; LAZ: between <-2 and -3
• Parent(s) willing to bring child to the feeding centre twice daily for 4 weeks for nutritional therapy and provide weekly fecal samples from their child for the duration of the study. The informed consent document will explicitly request permission to use the collected fecal samples for future studies, including but not limited to culturing component bacterial strains.

Exclusion Criteria:

• Medical conditions: Children with tuberculosis (diagnosis based on WHO 2014 guidelines which have been incorporated in the national TB control guidelines of Bangladesh)12 or any congenital/acquired disorder affecting growth i.e. known case of trisomy-21 or cerebral palsy; children on an exclusion diet for the treatment of persistent diarrhea; having known history of soy, peanut or milk protein allergy.
• Febrile illness or antibiotic use within the last 30 days
• Receiving concurrent treatment for another condition
• Any known history of or screened positive for clotting disorders or severe anaemia (<8mg/dl)
• Failure to obtain informed written consent from parents or caretakers