Systematic review of the treatment of moderate acute malnutrition using food products.

Abstract There is currently a lack of international guidance on the most appropriate treatment for moderate acute malnutrition (MAM), and discrepancies in national treatment guidelines exist. We aimed to explore whether food interventions are effective for MAM children 6–59 months old and whether they result in better outcomes compared with no treatment or management with nutrition counselling. A systematic literature search was conducted in October 2018, identifying studies that compared treating MAM children with food products versus management with counselling or no intervention. A total of 673 abstracts were screened, 101 full texts were read, and one study was identified that met our inclusion criteria. After broadening the criteria to include micronutrients in the control group and enrolment based on out‐dated anthropometric criteria, 11 studies were identified for inclusion. Seven of these found food products to be superior for anthropometric outcomes compared with counselling and/or micronutrient supplementation; two of the studies found no significant benefit of a food product intervention; and two studies were inconclusive. Hence, the majority of studies in this review found that food products resulted in greater anthropometric gains than counselling or micronutrient interventions. This was especially true if the supplementary food provided was of suitable quality and provided for an adequate duration. Improving quality of and adherence to counselling may improve its effectiveness, particularly in food secure contexts. There is currently a paucity of comparable studies on this topic as well as a lack of studies that include important functional outcomes beyond anthropometric proxies.

stature and productivity (Bhutta et al., 2017;Black et al., 2013) However, there is currently no consensus on how best to treat moderate acute malnutrition (MAM). The WHO has recognized the present lack of global guidelines for the treatment of MAM and called for more evidence in this area to inform related policies (WHO, 2017). Additionally, at national level, there are discrepancies in treatment strategies for MAM, if any are present at all. We know that children defined as MAM (WHZ ≥ −3 and <−2 and/or MUAC ≥11.5 and <12.5 cm) are at higher risk of mortality, morbidity, and deteriorating into SAM; hence, finding an effective method of supporting this group is crucial for meeting the second Sustainable Development Goal of achieving zero hunger by 2030 and ending all forms of malnutrition (James et al., 2016;L. M. Lenters, Wazny, Webb, Ahmed, & Bhutta, 2013).
Certain national guidelines for MAM treatment suggest the provision of supplementary food products, whereas others state that caregivers of MAM children should be provided with nutrition counselling alone. There is some debate about the necessity of supplementary foods for MAM: Do they result in better outcomes than no treatment or management with nutrition counselling? With the rise of noncommunicable diseases in low-income settings and our current lack of understanding of the exact causes, we need to be confident that MAM interventions are optimizing immediate survival as well as long-term health (Fabiansen et al., 2017;Shrimpton & Rokx, 2012;WHO, 2014). Moreover, food product interventions can be costly and lack sustainability; hence, their effectiveness should be established with concrete evidence and the outcomes of children managed with alternative methods assessed.
This review therefore aimed to identify and to synthesize the current evidence on outcomes of MAM children treated with food interventions compared with no treatment or management with nutrition counselling. Through identifying the current state of knowledge and highlighting evidence gaps, we hope to inform future research and international guidelines for the treatment of MAM.

| Search strategy
The systematic literature search was conducted on October 22, 2018 in Pubmed, Cochrane, and ScienceDirect databases, as well as resources catalogued on the following websites: Emergency Nutrition Network, Valid International, Evidence Aid, and State of Acute Malnutrition. Reference lists of eligible publications were also searched for relevant titles. Both peer-reviewed publications and grey literature were eligible for inclusion. The search term strategy and eligibility criteria were guided by the "Population, Interventions, Control, and Outcome (PICO) framework" presented in Table 1.
The following search terms were used: "Moderate acute malnutrition" or "untreated MAM" or "moderately malnourished children" or "supplementary food" or "RUSF," including Medical Subject Headings and free text terms, restricted to "human studies" and "children: 0-18 years" where search function allowed. No restrictions on publication date and language were applied.

| Screening and study selection
Two researchers conducted the screening process independently (EEE and AB); titles and abstracts of all search results were screened, guided by the PICO framework. Potentially eligible full texts were independently assessed for inclusion and a discussion between reviewers then defined the final results. Data were extracted from the included texts using a standardized form. Risk of bias was assessed at study level using the quality appraisal checklist from the National Institute for Health and Clinical Excellence public health guidelines (Higgins & Green, 2011;National Institute of Health and Care Excellence, 2012). A meta-analysis of recovery rates was planned; however, heterogeneity of methods and outcomes as well as limited number of eligible studies prevented this.
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses checklist was followed for reporting results (Moher et al., 2015).

| RESULTS
We screened 673 abstracts, ultimately identifying only one study that strictly meets the predefined PICO framework (see Search result flow

Key messages
• There is currently a lack of analogous studies comparing the two different management approaches recommended for MAM in national guidelines: food supplementation or nutrition counselling. Studies that include functional outcomes are especially lacking.
• The evidence in this review suggests that supplementary foods do support recovery from MAM and are often more effective at promoting anthropometric recovery than nutrition counselling, with or without the addition of micronutrient supplements.
• The type and duration of supplementary food provided is important; supplementation with high-quality protein and adequate micronutrient content, for 3 months, is recommended.
• Improving quality and adherence to counselling interventions may improve effectiveness. diagram in Figure 1). Due to this very limited number of eligible studies, we discussed widening the inclusion criteria and identified two studies that provided micronutrient supplementation to the control group and eight studies that did not enrol children based on current, common definitions of MAM; however, MAM children were part of the sample, for example, enrolment based on low weight-for-age or MUAC < 12.9 cm.
Ten of the studies were randomized controlled trials; the other one was a prospective cohort study. A summary of the results is presented in Table 2. Seven of the studies found food products to be superior with regard to anthropometric outcomes (weight, height, or MUAC), compared with nutrition counselling and/or micronutrient supplementation; two of the studies found no significant benefit of a food product intervention compared with control; and two of the studies were not conclusive either way. A more detailed summary of the studies including sample size, setting, and a summary of the outcomes is presented in Table A1, and a summary of the internal and external risk of bias assessment based on National Institute for Health and Clinical Excellence quality appraisal checklist is in Table A2.
The one study that met the original PICO criteria was a clusterrandomized controlled trial in Burkina Faso with 18 clusters across 3 study arms (Nikièma et al., 2014). MAM was defined as WHZ < −2 and ≥−3 based on WHO, 2006, reference. One arm provided a soybased ready-to-use supplementary food (RUSF); another provided fortified corn-soy-blended flour (Supercereal Plus; UNICEF, 2016), and the control arm provided weekly personalized counselling to caregivers. The recovery rate (WHZ > −2) after 3 months in the control arm was significantly lower (57.8%) than the Supercereal Plus arm (74.5%) and the RUSF arm (74.2%; p < .0001). However, when a "per protocol" analysis was applied, there was no significant difference in recovery rate between the groups, suggesting that the beneficial effect of the food products was due to lower defaulting in those groups than in the control group.  (Fauveau et al., 1992) RCT Bangladesh N = 134 6-12 months, MUAC > 11.0 and <12.9 cm, and living in bamboo structure Supplementary food (rice, wheat, lentils, and oil; SF) Nutrition education (C) Maybe, food group have larger weight gain in first 3 months but not whole 6 months ++/+ (Continues) Food security status also differed across the studies and could affect the results, particularly the success of counselling as treatment.
The study by Nikièma et al. (2014) was conducted in a "relatively foodsecure" context. One other study states that it was conducted in a relatively food-secure setting, taking place in an urban area of Iran (Javan, Kooshki, Afzalaghaee, Aldaghi, & Yousefi, 2017). They found food supplementation with counselling to be superior to multivitamins and counselling alone; although there was some spontaneous recovery (WHZ > −2; 32%) in the counselling group, this was much lower than in the food supplementation group (80%). Three studies mention that their study populations were likely to be food insecure. Roy et al. (2005) suggest that, although food supplementation had the best weight gain, an "intensive counselling" group still had better weight gain than the "standard counselling" group, despite low food security, whereas Christian et al. (2015) conclude that counselling alone is not sufficient in areas of food insecurity.
Of the studies that used WHZ < −2 and ≥−3 as the definition of MAM, the recovery rate was between 75% and 68% when providing a supplementary food intervention compared with counselling and/or micronutrient supplements, where recovery rate was between 58% and 32% (Javan et al., 2017;Nikièma et al., 2014).
Across the papers, there was an inference that the type of food product provided is important. The study by Fauveau et al. suggested that the lack of iron and zinc in the supplementary food was the reason for the limited longer term effect of the supplement (Fauveau et al., 1992). Additionally, Christian et al. found consistent anthropometric benefits (WHZ and HAZ) of a soy-based RUSF as well as multiple benefits of the chickpea-based RUSF. However, far fewer beneficial outcomes were shown for the rice-lentil-based RUSF and no benefit at all of fortified wheat-soy-blended flour compared with nutrition counselling (Christian et al., 2015).
Of the two studies that saw no benefit from a food product, nei- in Guinea Bissau using "at risk" infants and children (Schlossman et al., 2017). This was defined as children aged 6-59 months with either WHZ < 2 or WAZ < 1 or HAZ < 2. Mean WHZ at enrolment was −1.0, suggesting that most children sampled were not MAM cases, and the results were not stratified by MAM status at enrolment. All groups saw significant improvements in weight and height. The authors suggest that improvements in the control group were due to increased focus by community health workers on nutrition recommendations as a result of the study. However, it is not possible to conclude whether or not there were anthropometric improvements in MAM children specifically, in either the control or intervention groups.

| DISCUSSION
One study was found that met the original PICO criteria, comparing MAM treatment with a food product to either counselling or no intervention. After broadening the criteria to include studies that provided micronutrient supplements to the control group and/or did not use Lack of adherence to counselling programmes may be one of the limitations influencing their effectiveness among control groups in these studies. The "per protocol" analysis by Nikièma et al. suggests that, if adhered to, the counselling programme may be as effective as the food intervention. Hence, finding ways to improve adherence to counselling interventions should be explored. However, it is also important to note that possible confounding factors may be influencing results, such as greater socioeconomic status, willingness to change behaviour, or greater care-seeking behaviour among adhering mothers.
Studies exploring the use of counselling would benefit from including direct measures of behaviour change, rather than just attendance at sessions.
Food security status may also be an important factor in determining the effectiveness of counselling interventions. The study by Nikièma et al. (2014) was conducted in a relatively food-secure context, which may have added to its success in the per protocol analysis. However, the only other study to mention that it took place in a relatively food-secure setting (urban Iran) still found food supplementation to be superior in recovering WHZ than micronutrients and counselling (Javan et al., 2017). It is generally felt that counselling alone is not sufficient in areas of food insecurity, as concluded by Christian et al. (2015).
The quality and content of nutrition counselling interventions also require consideration. A review of dietary counselling for the treatment of MAM was conducted by Ashworth andFerguson in 2009 (Ashworth &Ferguson, 2009); 10 studies were examined; the majority had quasi-experimental or observational study designs. The review concludes that counselling messages tended to be vague and were unlikely to be effective. They also found that it was difficult to achieve adequate zinc and vitamin E from home foods alone. In the studies they reviewed, they found that rates of weight gain tended to be slow, with estimates of 1 to 2 g/kg/day. This was likely due to repeated exposure to pathogens and high rates of stunting, as well as counselling being delivered by minimally trained personnel or volunteers. In successful programmes, frequent, regular exposure to a few simple, uniform, age-appropriate messages, together with an opportunity for interaction between caregiver and counsellor, was found to be important.
Not all studies in our review found food supplements to be superior to nutrition counselling. A key consideration is the type of supplementary food provided, as well as the dosage and length of treatment.
Studies in the review specifically highlighted the micronutrient content and protein quality of supplements as likely influential factors. The majority of studies provided supplements for at least 3 months; however, one study provided one sachet of RUTF for 14 days and was found to be ineffective at preventing SAM in MAM children recovering from illness. The health status of MAM children at admission should also be a consideration when selecting the most appropriate treatment; this may be at an individual level or at a more regional level based on morbidity rates. There are currently no international guidelines on the treatment of "complicated" MAM cases, another area which needs attention (Polnay, 2019). Bangladesh, four in West Africa, one in Jamaica, and one in Iran. However, emergency contexts are not well represented, with only one study being conducted during a nutrition emergency (Grellety et al., 2012). In addition, the level of food security, burden of infectious disease, and access to health services may also influence the generalizability of these results to all contexts.
In conclusion, there is currently a paucity of studies on this topic, which use standard definitions of MAM and recovery, as well as a lack of studies including important functional outcomes beyond anthropometric proxies. Future research is needed to address these evidence gaps. The majority of studies in this review found that food products resulted in greater anthropometric gains than counselling or micronutrient interventions. The type and duration of supplementary food provided is important, and these studies showed that supplementation with high-quality protein and adequate micronutrient content, for 3 months, was most commonly recommended. In addition, it was suggested that improving quality and adherence to counselling interventions in food secure settings may improve their effectiveness. Policymakers should consider these factors when improving the recommendations for MAM treatment, which are urgently needed.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.