We implemented SEEM in the rural district of Pakistan; the primary objective was to test the biomarkers of EE in malnourished and well-nourished children. Matiari is a rural district that is 185km/115 miles north of the city Karachi, where the Aga Khan University (AKU) flagship campus is based. Its total population is 0.769 million with approximately 143,000 households in around 1600 villages. ⁴ The major source of economic growth in this area is from agriculture-related activities. Matiari is subdivided into 18 Union Councils (the equivalent of municipalities in the United States), which are its administrative subunits. The average population per union council is 30,000 people with an average household size of 6.7 persons. The Department of Pediatrics and Child Health of AKU has been conducting community-based research projects in this district since 2002 and our research team has successfully implemented several large community-based trials on maternal, neonatal, and child health. ^{5– 15} In our current study, SEEM, collection of field data and laboratory samples (blood, urine, and stool) involved traveling to the residence of the families versus the use of a field office as a central sample and data collection point. This decision was driven by feasibility given that our study population primarily lived in widely scattered villages from where access was limited due to poor road and public transportation infrastructure. It was logistically difficult to ask parents to bring their children to a field office or any fixed facility for sample collection.

The Matiari district shoulders an immense burden of maternal, neonatal, and childhood illnesses as evidenced by a stillbirth rate of 42·8 (40·5–45·2)/1000 total births, neonatal mortality of 46.9 (44.4–49.4)/1000 live births, and a maternal mortality ratio of 259 (198–319)/100,000 live births. ¹⁶ The Matiari site has a field laboratory with basic equipment needed for processing biological urine, blood, and stool samples and is staffed by technicians trained at AKU that receive periodical refresher training via regular visits to our central research coordinating facility - the AKU Infectious Diseases Pediatric Research Laboratory (IDRL) at Karachi. The site is also equipped with an electronic Data Management Unit (DMU) which enables both electronic transcriptions of the data collected in the field on hard/paper copies, and direct electronic data collection via portable tablets through customized applications. Furthermore, the DMU has data storage servers that enable data synchronization with central servers at AKU, Karachi. We have previously successfully conducted several studies on maternal and child health at this study site. ^{5– 15} However, we found SEEM to be unique in terms of the additional biological specimens and data collection processes we had to implement given that these samples would then be processed for high throughput sensitive ‘omics analyses such as microbial RNA and metabolomics.

The study protocol was approved by the Aga Khan University Ethics Review Committee (ERC, 3836-Ped-ERC-15) in December 2015, and the study is registered on clinicaltrials.gov (ID: NCT03588013). ¹⁷

This was a community intervention study.

The primary objective of the SEEM study was to understand the clinical phenotypes of children at risk for EE in order to identify predictive ‘omic profiles along with detailed microbiome and transcriptomics as potential therapeutic targets. A cohort of 416 children was selected for a longitudinal study; 365 of whom were diagnosed with acute malnutrition or wasting (defined as weight-for-height Z (WHZ) score < -2) while 51 healthy controls or well-nourished children (defined as WHZ score ≥ 0 and height-for-age Z (HAZ) score ≥ -1 on two consecutive follow-ups from three to six months) were recruited. Our detailed study methods and results have been published. ^{18, 19} In brief, birth surveillance was established in which parents registered their newborns within 30 days of birth followed by monthly anthropometric assessment until six months of age. Between the time of birth registration and six months of age, data regarding young infant feeding practices, morbidity, and health-seeking behavior in the case of illness was also collected. Anthropometry measurements including height, weight, and mid-upper arm circumference (MUAC) of both parents and children, with an additional measure of head circumference were recorded from our sample of children. Anthropometry data collected between three and six months was used to calculate WHZ scores based on World Health Organization (WHO) growth reference standards; ²⁰ children were enrolled as cases and controls based on criteria defined above. At the time of enrollment, we collected data of the Household Food Insecurity and Access Scale, ²¹ illnesses and treatment received, socioeconomic status, maternal and paternal education ( Table 1). Enrolled children were then followed weekly for morbidity data (diarrhea, respiratory tract infection, antibiotic use, and hospitalization) and monthly for anthropometric measurements. Blood, stool, and urine samples were collected at the time of enrollment and at nine months of age for biomarker assessment. Breast milk samples from mothers were also collected once at the time of enrollment for oligosaccharides assessment.

Between six and nine months of age, nutritional education counseling was provided monthly to cases i.e. acute malnourished children. At nine months, if the child had a persistent WHZ score < -2, a chick-pea-based and locally made Ready-to-use Supplementary Food known as AchaMum was initiated for two months as per WHO guidelines for the management of moderate acute malnutrition. ²² For severe acute malnutrition, however, while the WHO recommendation is for Ready-to-use Therapeutic Food (RUTF), we provided AchaMum according to a child’s weight (similar to RUTF) due to logistic and administrative issues in importing RUTF. If the child’s WHZ score consistently remained < -2 after the intervention, he/she was referred for further clinical assessment by a pediatric gastroenterologist for the screening of celiac disease and detailed luminal evaluation via an upper gastrointestinal (UGI) endoscopy.

Implementation challenges:

1. Collection of 24 hours food intake frequency data

A 24 hours food recall data was collected through a specialized structured instrument designed to collect all food items consumed by the child in the last 24 hours. This provided a quantitative estimate of the non-breast milk food intake. The form was administered between six and 24 months on every alternate month. This data was collected by an experienced dietitian or an individual with knowledge regarding food preparation techniques, portion sizes, recipes, and food intake. Six community health workers (CHWs) conducted the dietary intake form after getting training in the Naushahro Feroze district, an area where a similar food recall form was administered as part of the Malnutrition and Enteric Disease Study (MAL-ED) cohort. ²³

As the mothers in this area participate in fieldwork including crop harvesting, it was challenging to find optimal times to contact and interview them. We also found that some families temporarily migrated to different villages in response to employment opportunities. If a new recipe was found, which was not mentioned in the existing MAL-ED coding list, then CHWs cooked these recipes in the field office kitchen with the same method as cooked per the report by the mothers. The purpose of this activity was to identify the caloric value of each food item mentioned in the recipe. It was not only the data collection that was challenging but converting this data into a quantitative form for data analysis was a huge task. For this, the CHWs had to work for several hours on every single form to search for the codes already assigned to food items by the MAL-ED team and to assign new codes for new food items.

2. Management of acute childhood illnesses

All sick enrolled children were visited at their homes and assessed by study physicians as per Integrated Management of Childhood Illnesses guidelines. ²⁴ Medications were prescribed for mild-to-moderate and serious illnesses, the child was referred to the nearest healthcare facility with facilitated transport.

3. Anthropometry measurements and quality assurance

Weighing machines were monitored by daily standardization through calibrated weights of 0.5,1, 2, and 5 kilograms. Performance of CHWs was monitored in several ways: 1) standardization sessions were performed every three months to ensure reliability, precision & accuracy; 2) repeated measurements were taken from each child to increase the accuracy and precision, for 3-5% of the participants these repeat measurements were performed within 24-72 hours of the first measurement; 3) if measurements differed between two consecutive follow-ups, these were anthropometric measurements were recorded daily and outlier readings were then verified independently by the field team monitors, and 4) real-time feedback was provided to the field worker identified as responsible for the error. Furthermore, we employed the use of the Emergency Nutrition Assessment software ²⁵ for digit preference of weight, height and MUAC measurements, and these scores were reviewed weekly.

4. Upper GI endoscopy consent

After a nutritional intervention, children who did not adequately respond (WHZ remained < -2) were thus selected for further evaluation. This included a detailed clinical workup and, if indicated, an upper GI endoscopy with biopsies as per pediatric gastroenterology guidelines. ^{26, 27} Challenges encountered in the group of children selected for UGI endoscopies included: 1) 13 out of 89 biopsies had to be postponed as children were deemed unfit for anesthesia (seven children suffered from flu, fever and runny nose; three suffered from low hemoglobin levels; one had diarrhea; one had low platelets, and in one case the mother was too ill to accompany her child to Karachi). Children with low hemoglobin were prescribed oral iron until their hemoglobin levels reached an acceptable level (i.e. hemoglobin >=9mg/dl) for procedure eligibility; 2) 13 children (12.7%) were determined to be ineligible due to pre-existing diseases and disorders, which included tuberculosis, neurological disorders, thalassemia, congenital anomaly and, epilepsy; and 3) obtaining consent for the procedure from the parents required need of an additional explanation of risks and benefits of the procedure. As per protocol, the parents were re-consented by the pediatric gastroenterologist when they arrived at the AKU Hospital before the procedure.

5. Laboratory samples collection challenges

Breast milk sample collection

Sugars found in breast milk are known to serve as prebiotics and help in establishing the health-promoting gut microbes in infants. The presence of female study physicians led to a reduction of anxiety and addressed the fears the mothers were harboring. The sample was transported to and stored at Matiari field lab at 2-8°C.

Stool samples collection for microbiome testing

Fecal sample collection for the purpose of microbiome analyses was performed for the first time in our study in a rural setting in Pakistan. This required the development of new feasibility measures and modification of transport protocols. An important aspect of fecal sample collection in the microbiome collection protocol, given the rapid degradation of bacterial RNA species at room temperature, includes a critical time window i.e. stool samples had to be collected and cryopreserved ideally within 30 minutes. This was particularly challenging given that the households could be up to 90 minutes’ drive from the field office. To allow rapid freezing of samples, dry cryo-shippers were procured to facilitate transportation back to the central IDRL at AKU, Karachi, a vehicle was modified to enable transportation of the field-site collection staff and the dry shipper along with the ability to park as close to the collection site as possible for appropriate rapid storage and cryo-preservation. Further challenges faced included multiple visits, at times due to diarrhea, insufficient quantity of stool produced and the challenge of sitting in open courtyards by CHWs and children during extreme weather conditions for several hours at a time. We found that the provision of access to a cellphone helped coordinate the best time for stool collection.

Urine sample collection for Lactulose: Rhamnose test

The lactulose: rhamnose (L:R) is a dual sugar absorption assay used to assess increased intestinal permeability as a surrogate measure of EE. ²⁸ The L:R test is performed by preparing oral lactulose and rhamnose solutions for the child to drink. ²⁹ Urine samples are obtained while fasting (before intake of lactulose and rhamnose), at 30, 60, and finally at 90 minutes after intake. These urine samples are then tested for the excretion of sugars. ²⁹ The biggest challenge was the collection of a fasting urine sample as it was difficult for mothers to keep their infants in a fasting state. The CHWs counseled and coordinated appropriate times for the collection of fasting samples and this aided in the sample collection. Additionally, the collection of urine samples at multiple time points required close monitoring. This process took several hours to complete. Furthermore, Matiari has average summer temperatures ranging from 38°C to 41°C, ³⁰ which leads to decreased urine sample collection in summer due to dehydration. This environmental limitation was a significant barrier in conducting the L: R sample collection.