Developmental and Nutritional Changes in Children with Severe Acute Malnutrition Provided with an N-3 Fatty Acids Improved RUTF and Psychosocial Support: A Pilot Study in Tanzania

1. Introduction

Severe acute malnutrition (SAM) affects almost 15 million children under the age of five in Sub-Saharan Africa (SSA) [1]. While ready-to-use therapeutic food (RUTF) improves survival and nutritional recovery, cognitive development is known to often remain impaired after an episode of SAM [2,3].

During early life, brain development is influenced by both biological mechanisms and environmental factors such as nutritional intake, adequate stimulation and responsive caregiving [4]. These factors can interplay with one another and all play a role in long-term child developmental outcomes [5].

One of the nutritional factors that have been considered important for children’s cognitive development include an adequate intake of polyunsaturated fatty acids (PUFAs) [6]. Previous studies have shown that children with SAM who were provided with RUTF continued to have suboptimal, or even deteriorating, PUFA levels after weight recovery [7,8]. Docosahexaenoic (C22:6n-3; DHA) and arachidonic acid (C20:4n-6; AA) are n-3 and n-6 long chain PUFAs, respectively, that are found in abundance in the cells of the central nervous system and play important roles in brain development [6]. AA and DHA can either be supplied by dietary sources, including breast milk and seafood, or synthesized from their essential precursors, linoleic acid (C18:2n-6; LA) and alpha-linolenic acid (C18:3n-3; ALA), respectively [9]. However, large dietary amounts of LA may impair the conversion of ALA to DHA [10]. Consequently, the 2022 Codex Alimentarius guideline for RUTF now recommends lowering the upper limits for LA compared to previous recommendations [11,12].

In addition to adequate nutrition, psychosocial (PS) support is important to support the recovery of development in children treated for SAM [13,14,15]. Evidence from the 1970s showed that intense PS programmes, based on weekly home visits, benefited the development of children treated for SAM [16]. Thus, the World Health Organization (WHO) include PS in their ten steps-guideline for hospital management of SAM [17]. Studies evaluating the effects of PS interventions in children treated for SAM have demonstrated an improvement in gross motor development, clinic attendance, and nutritional recovery [18,19,20,21,22]. However, the PS interventions conducted in these studies are often resource-intensive and not feasible for implementation in the community and scale-up in health care systems [16]. In practice, support for stimulation and responsive caregiving is rarely offered during SAM management in hospitals, and it is still not part of the guidelines for management in the community, where most children are now treated [12].

This paper describes the results from a pilot study of two interventions provided to children during treatment of SAM to help improve cognitive and nutritional outcomes: RUTF with a PUFA composition modified to comply with the 2022 Codex guidelines and a context-specific PS intervention. We report changes in anthropometry, child development, whole blood PUFAs, caregivers’ stimulation and support including family care indicators, caregiver-child interaction, and maternal depression after 8 weeks’ combined intervention.

2. Materials and Methods

Study design and setting: The BrightSAM (Brain development, growth, and health in children with SAM) trial development study was conducted in Mwanza, Tanzania, June 2020 to February 2022 with an original target of including a total of 200 participants: 100 children with SAM and 100 participants without acute malnutrition. The main aim was to inform a future randomized 2-by-2 factorial trial, that will investigate the individual and combined effects of modified RUTF and/or a context-specific PS intervention in improving developmental and nutritional outcomes in children treated for SAM. During the trial development study, we developed and piloted the RUTF and PS interventions and assessed acceptability and implementation feasibility of these two interventions to be integrated in the existing SAM management in Tanzania and similar settings. The qualitative formative research that informed the design of the content and delivery mode of the PS intervention, and mixed methods evaluation of implementation fidelity and feasibility for future trial will be reported separately.

We recruited children with SAM from Bugando Medical Centre (BMC), a tertiary referral hospital for all districts and regional hospitals within Lake Zone, and 3 district hospitals: Misungwi, Nyamagana, and Buzuruga. We trained a network of community health workers in close contact with these hospitals to use mid-upper-arm circumference (MUAC) tapes for screening SAM in low-income communities within their catchment areas in Mwanza region. A MUAC of <115mm was regarded as SAM and any identified child was referred to the nearest district hospital for further clinical assessment and recruitment into the study if eligible.

Children with SAM who were hospitalized were included when transitioning to RUTF from F-75, while children with SAM without oedema, infections, a passed the appetite test were started on RUTF directly through outpatient management clinics of SAM. The study inclusion criteria were: residence within Mwanza, confirmed SAM diagnosis as either MUAC <115 mm, or weight-for-height (WHZ)-score <-3 as per WHO growth standards [23] or bilateral pitting edema, children’s age 6 to 36 months, and age of a caregiver giving consent ≥18 years. Children with SAM were excluded if they were allergic to peanuts or other RUTF ingredients, or had any overt disability limiting the feasibility of delivering interventions or conducting assessments. Non-malnourished reference children (MUAC >125 mm and WHZ >-2) were frequency-matched by age (+/- 2 months) and sex to children with SAM and recruited from the same neighborhoods in order to provide reference levels of whole blood PUFAs, child development and family environment indicators.

RUTF composition and administration: The RUTF used in the study was produced by Nutriset, France, in accordance with Codex 2022 [11]. The RUTF was formulated using high-oleic peanuts as a main ingredient and contained 3.7 g/100 g n-6 fatty acids and 1.02 g/100 g n-3 fatty acids (additional information about nutritional composition, Suppl Table S1). Each sachet contained 90 gm of RUTF paste. RUTF was provided during the rehabilitation phase of SAM management in amounts dependent on the child’s body weight: 2 sachets/day for a child weighing 5 to 6.9 kg, 3 sachets/day for a child weighing 7 to 9.9 kg, and 4 sachets/day for a child weighing 10 to 14.9 kg [24].

Psychosocial intervention: We developed a short, focused, and context-specific PS intervention by adapting the WHO/UNICEF (United Nations Children’s Fund) Care for Child development package [25] based on qualitative research involving key stakeholders involved in treating children with SAM (details will be published separately). The adaptation was based on input from professionals in organizations engaged in early childhood development such as Tanzania Home Economics Organization (TAHEA) located in Mwanza, healthcare professionals, and caregivers of children with SAM. We piloted the developed intervention with mothers of children with SAM and set up weekly group meetings at the research clinic of the National Institute for Medical Research in Mwanza. During group sessions, mothers were taught about nutrition, preparation of a balanced diet using locally available low-cost foods, water sanitation and hygiene (WASH), responsive parenting, and child stimulation. They were shown how to provide caregiver-led play and communication to their children and how to make toys from locally available materials.

Sociodemographic and clinical data: At enrolment of both SAM and non-malnourished children, a research clinician collected data on sociodemographic characteristics including age, sex, child’s birth weight, parental education, occupation, marital status, and information on household assets from the Demographic Health Survey Tanzania [26]. Information about household assets were utilized to compute socioeconomic status (SES) quintiles using principal component analysis [27]. Caregivers were also asked to report risk factors for developmental delay, such as hearing and visual deficits, and HIV exposure.

Early life stress measures: We assessed early life adversities using an adapted version of a tool created by researchers in India which has demonstrated high correlation with child development [28,29]. This tool was modified to remove items not relevant in the Tanzanian setting. The 16 adversities that we examined in the tool fell into three categories: 1) child stressors, 2) maternal stressors, and 3) socio-economic adversities. Child stressors included: premature birth; child hospitalisation; inadequate supervision; and separation of mother and child for more than a week. Maternal stressors included: death of one or more of mother’s close family members since becoming pregnant; mother seriously injured or sick since becoming pregnant; mother single, widowed, divorced, or separated; mother ill or seriously injured during pregnancy; mother screened positive for mild, moderate or severe depression on the Patient Health Questionnaire (PHQ-9) in the past two weeks; and problematic alcohol/drugs use and/or selling by a person staying in the household. Socio-economic adversities included: family being in the lowest SES quintile; very low mother’s or father’s education i.e. no education or only primary schooling; father’s or mother’s occupation irregular i.e. unemployed, seasonably employed, or casual labourer; and family debt or inability to buy food for the family during the past six months.

Anthropometry: Anthropometric assessment was conducted by a trained research assistant. Weight was measured to the nearest 0.1 kg using a digital scale (ADE model M112600), height/length was measured to the nearest 0.1 cm using a stadiometer/ wooden height measuring board [30]. The measuring boards and scales were checked and calibrated regularly. MUAC was measured to the nearest 0.1 cm at the midpoint between the olecranon and the acromion process of the left arm using a non-elastic measuring tape. Anthropometric measurements were taken in triplicate and the median used in analyses. The STATA package "zscore06" using the WHO standards was used to calculate WHZ and height-for-age (HAZ)-scores [31]. Stunting was defined as HAZ <−2.

Assessment of child development: Child development was assessed using the Malawi Development Assessment tool (MDAT) that was created and validated for use in African settings [32]. The assessments at baseline and 8 weeks follow-up were done in a quiet room by trained research assistants with the caregiver present during the complete assessment. MDAT examines gross motor, fine motor and language skills through direct observation and socio-emotional skills through caregivers’ report. The tool has demonstrated good construct validity and sensitivity in predicting moderate to severe developmental delay in children from birth to 6 years of age [32]. MDAT z-scores for this study were based on reference data from a Malawian non-malnourished population [32].

Caregivers’ stimulation and support assessment: The Observation of Mother Child Interaction measure (OMCI) [33], the Family Care Indicators (FCI) [34], and the PHQ-9 [35] were used in the study, as these measures were seen as potential mediators of intervention effects on developmental outcomes in children. OMCI is a 19-item observational tool which measures caregivers’ interaction behaviours (12 items) and child interaction behaviours (7 items). A caregiver was asked to play and talk with his/her child as he/she would normally, using a picture book provided, for a period of five minutes without interruption. An observer counted each behaviour and then coded it as either 0=Never, 1=very rarely, 2=rarely, 3=once in a while and 4=many times. All maternal items were summed to obtain maternal scores and child items summed to obtain child scores. To reflect a positive and responsive mother-child interaction the 4 negative items were reverse-coded to give a total score that ranged from 0 to 76 when maternal and child scores were combined. Prior to actual data collection research assistants were trained as directed by Rasheed et al [33] and attained sufficient inter-rater reliability (Kappa>0.8).

We used FCI to collect information on: a) availability and number of reading materials, b) availability and variety of play materials, and c) family interaction in the home [34]. A raw score for each component was obtained by adding up positive answers to obtain score ranges of 0-3 for source of playing materials, 0-7 for variety of playing materials, and 0-6 for family [36].

Depression symptoms among caregivers were assessed using the PHQ-9 [35], which is a 9-item depression-screening tool. Responses for all 9 items are based on the caregiver’s experience during the past two weeks. Each item is scored on a 4-point scale. An example question is “over the last two weeks how often have you been bothered by little interest or pleasure in doing things?”. A total PHQ-9 score is calculated by adding up all individual item scores. A total score of 1-4 indicates minimal/no depression, 5-9 mild depression, 10-14 moderate depression, 15-19 moderately severe depression, and 20-27 severe depression.

Blood sampling and analysis of PUFAs: A 1.5 mL sample of venous blood was collected in sodium citrate tubes from children with SAM at baseline and after 8 weeks and from non-malnourished children just once. Blood was transported to the laboratory at 2 to 8˚C. One mL of the venous sample was used to saturate 1 cm² of a chromatography paper strip treated with 50μg 2,6-di-tert-butyl-4-methylphenol (butylated hydroxytoluene) and 1,000-μg deferoxamine mesylate salt (both from Sigma-Aldrich, St. Louis, MI, USA) for analysis of whole-blood fatty acid composition. Data are given as percent by weight of individual fatty acids relative to the total fatty acid concentration in each sample (FA%). We regarded a high triene to tetraene ratio (defined as Mead acid (C20:3n-9):AA (arachidonic acid, C20:4n-6) as an indicator of low overall PUFA status [37]. Two indicators of low n-3 PUFA status were used: a high n-6 DPA (docosapentaenoic acid, C22:5n-6):DHA (docosahexaenoic acid, C22:6n-3) ratio and a high n-6: n-3 PUFA ratio [38,39].

Data management and statistical analysis: Data were captured electronically using CSPro and analysed in STATA 17 (StataCorp, College Station, TX, USA). Descriptive data are presented as mean ± standard deviation (SD), or number (%) as appropriate. Changes in anthropometry and child development at 8 weeks were assessed for both children with SAM and non-malnourished children using linear regression adjusted for sex, age and month of inclusion to account for possible seasonal effects. Difference in change between groups was also assessed. Mean levels of PUFAs, family care indicators, mother-child interaction, and maternal depression were compared between groups at baseline and changes in children with SAM and their families at 8 weeks was assessed using similar linear regression models.

3. Results

Recruitment occurred June 2020 to February 2021. Due to recruitment challenges during the COVID-19 pandemic, we did not reach the original recruitment target (200 participants), but managed to enrol 170 participants: 82 children with SAM and 88 non-malnourished children. Among 82 children with SAM enrolled: 28 children received inpatient management of SAM with F-75 before transitioning to RUTF and the remaining 54 children started on RUTF directly as cases of outpatient management of uncomplicated SAM. During follow-up, 5 (3.0%) children with SAM died, 7 (8.5%) were lost to follow-up. There were no deaths among non-malnourished children and 10 (11.4%) were lost to follow-up.

3.1. Background characteristics of children with SAM and non-malnourished reference children

Table 1 shows the baseline characteristics of children in the two groups. Children with SAM were more likely to be HIV-infected or exposed, to be from the lowest SES quintile, and to have mothers who never attended school. Children with SAM were 2 months younger (95% CI: -4.0, 0.1) than the non-malnourished children. More than half of caregivers of children with SAM indicated concern that their child had a developmental delay, while this was not a concern to any of the caregivers of non-malnourished children. Families of children with SAM experienced early life stress more than non-malnourished children. There was little difference between all children who enrolled at baseline and those who remained until the end of 8 weeks intervention (Table 1).

3.2. Anthropometry

At baseline, children with SAM had a lower length/height and higher prevalence of stunting than non-malnourished children (95.1% vs 56.8%). Among children with SAM having 8 weeks data, both HAZ and WHZ scores increased during intervention and 41 (55%) of those children with SAM who had a follow-up data attained nutritional recovery that was defined as no SAM/MAM diagnosis using all three criteria used by WHO (oedema, WHZ, or MUAC) at 8 week intervention. Non-malnourished children gained considerably in HAZ and had a decline of mean WHZ-score over the 8 weeks (Table 2).

3.3. Child developmental outcomes

Internal consistency for MDAT using Cronbach’s alpha was high for all domains in our setting (Standardized Cronbach’s alpha of ≥0.85). At baseline, the children with SAM had much lower MDAT z-scores in all domains compared to the non-malnourished reference children (total MDAT score -2.37 (-2.92; -1.82)). After 8 weeks, improvement in all domains of development were observed in children with SAM when compared to their baseline, with the largest increase in the fine motor z-score 0.75 (0.46; 1.04). The MDAT z-scores of non-malnourished children also increased at endline assessment, but less than the changes observed among children with SAM (Table 2).

3.4. PUFAs

Children with SAM had lower levels of total n-3 PUFA, DHA and EPA but similar levels of n-6 PUFAs compared to non-malnourished children at baseline. After 8 weeks of intervention, total n-3 PUFAs, and EPA increased, total n-6 PUFAs decreased, while DHA remained at the initial level (Table 3).

3.5. Caregivers’ stimulation and support

At baseline, children with SAM were less likely to have books or other play materials in their home and the FCI scores indicated that they interacted less with family members than non-malnourished children. The overall score for the observed interaction between mothers and children was also lower in the SAM group when compared to non-malnourished children. (Table 4 and Suppl. Table S2).

At 8^th weeks, the number and variety of playing materials had increased in the families of children with SAM and the observed interaction between mothers and children indicated some improvement.

Mothers of children with SAM had higher depression score than mothers of non-malnourished children at baseline, but at follow-up these had decreased notably.

4. Discussion

This trial development study enabled us to develop and pilot the processes for testing an RUTF that complies with 2022 Codex guidelines [11] alongside an integrated context-specific PS intervention for children with SAM. Children with SAM improved their development scores over the 8 weeks of intervention with larger changes in MDAT z-scores than their non-malnourished comparison group. It is likely that this was due both to the impact of the interventions and because the SAM group had more development to “catch up” on after an episode of severe malnutrition.

Similar benefit for child development was seen in a clinical trial conducted in Malawi testing high-oleic RUTF with additional preformed DHA in children with SAM [40]. This indicates that the improvements in child development scores seen in our study may be caused by the modified RUTF essential fatty acid profile. However, much of the improvement in development in these SAM children may just be due to a general improvement in wellbeing for those children who were provided with RUTF. When children feel better, they are also more likely to perform better and score higher on developmental tests [41]. The individual and combined effects of the two interventions therefore need to be tested in a randomised controlled trial to be able to conclude on individual and combined effects.

At baseline, the levels of n-3 PUFA, DHA and EPA were lower in children with SAM in our study sample (Table 3). However, these levels seemed quite relatively good when compared to PUFA levels among children with SAM/MAM studied elsewhere [38,42]. Looking at baseline levels of DHA for instance, the values were more than twice as that observed in children with SAM/MAM studied in other settings [38,42]. Our study population lives in a close proximity to Lake Victoria with easy access to lake foods including fish.

We found improved levels of total n-3 fatty acid post-intervention similar to a trial in Malawi [40] that tested three different formulations of RUTF: high oleic-peanut RUTF with preformed DHA, high oleic-peanut RUTF, and standard RUTF on child development. In that trial, high oleic-peanut RUTF with preformed DHA was the only formulation that improved the level of DHA post management. In our study, we observed the DHA levels of children remained unchanged after supplementation. Previous studies with earlier versions of RUTFs have demonstrated levels of long chain total n-3 fatty acids, EPA, and DHA that declined post-intervention [7,8,43]. Although one study suggested that lowering dietary levels of LA may improve synthesis of DHA [44], our findings suggest that additional preformed DHA may be needed to increase DHA levels. Our findings are in line with other studies that have shown that consuming RUTF with preformed DHA is the only way so far that increases the level of DHA post management in children with SAM [40,44,45].

The potential effects of PS programmes which promote responsive caregiving practices for children with SAM, is well described [16,46,47]. In practice, however, PS is often not provided during inpatient SAM treatment or within community-based management of SAM. We developed a programme for PS that aimed to be feasible and accessible to provide in hospital and clinic settings for children with SAM. This programme encouraged and supported responsive and nurturing caregiving with more caregiver-led interactive play and communication, and creation of play materials from locally available materials. In contrast to earlier studies that focussed on intensive family-based psychosocial intervention provided by trained personnel [16,19], our PS programme was designed to be self-sustainable by using existing outpatient management of SAM to provide PS alongside RUTF supplementation and thus be integrated with the existing health care services.

The children admitted with SAM in our study had less caregiver stimulation and support than did non-malnourished children – likely due to socioeconomic situations, low educational levels and stressors within these families. Financial constraints and lack of education may limit opportunities for positive parent-child interaction and may prevent children from having conducive environments to play. Many mothers of SAM children had increased signs of depression, which may be due to socio-economic constraints and other family issues, preventing them being able to support their children in the way they would like [48]. It is vital that these issues are considered when providing PS programmes for children with SAM as maternal depression and family stress may limit recovery or lead to re-admission unless tackled through social welfare and other support programmes.

This pilot study has a number of strengths. It included a population of children with SAM during their critical period of brain development (6 to 36 months) who were provided with a locally created feasible intervention linked with local organisations. We also conducted a detailed evaluation of anthropometry, child development and other factors (socioeconomic status including some measures of adversities, maternal child interaction and family care indicators) which may influence our intervention. Moreover, we included other factors (socioeconomic status, maternal child interaction and family care indicators) which may influence our intervention. Furthermore, since Tanzania household food security varies greatly across agricultural cycles of the year [49] we controlled for season during analyses by incorporating month of enrolment in the regression models. Despite these strengths, this was a trial development study and it was limited by having a small sample size, lack of randomization or blinding, and lack of follow-up blood samples from non-malnourished children. We do not regard 170 participants recruited instead of the planned 200 as a major limitation as the included sample still provided sufficient variation to answer our objectives. Lastly, community screening was based on MUAC only. If we had also used WHZ and systematically assessed for nutritional oedemas the studied sample would have represented SAM population more broadly.

5. Conclusions

The combined interventions of a PUFA-modified RUTF and a context-specific PS intervention benefitted child development, reduced maternal depression, and increased total n-3 PUFA, while DHA levels were maintained. High-oleic peanuts not likely to make a difference alone in improving DHA concentrations post management of SAM if DHA is not added directly in the RUTF.