The management of Moderate Acute Malnutrition (MAM) in children aged 6-59 months: A systematic review and meta-analysis

Background: Malnutrition is a leading cause of morbidity and mortality in children aged under five years, especially in low and middle income countries. Although severe acute malnutrition (SAM) is considered the most serious form of malnutrition, moderate acute malnutrition (MAM) affects greater numbers globally and, unlike SAM, guidelines lack a robust evidence base. This systematic review and meta analysis assessed the evidence for lipid based nutrient supplements (LNS) fortified blended flours (FBF) and nutrition counselling, in the treatment of MAM. Methods: Five databases were systematically searched for studies conducted in LMICs that compared the effectiveness of food based products versus any comparator group in promoting recovery from MAM in children aged 6-59 months. Where appropriate, pooled estimates of effect were estimated using random effects meta analyses. Results: A total of thirteen trials were identified for inclusion. All used active controls rather than 'standard care', which is often minimal in most settings. There was evidence of increased probability of recovery (as assessed by gaining normal weight for height and/or MUAC) among children treated with LNS compared to children treated with FBF (RR 1{middle dot}05, 95%CI 1{middle dot}01 ; 1{middle dot}09, p=0{middle dot}009). Treatment with an LNS was also associated with a lower risk of persistent MAM at the end of treatment compared with a FBF (RR 0{middle dot}82, 95%CI 0{middle dot} 71 ; 0{middle dot}95, p=0{middle dot}007). Conclusion: Based on a relatively small number of studies mainly from Africa, LNS are superior to FBF in improving anthropometric recovery from MAM. The true benefit of MAM treatment may be underestimated due to all studies using active controls rather than usual care which is minimal. More high quality evidence is needed to evaluate nutrition education/counselling alone as a MAM intervention. Studies should also assess a wider range of outcomes including body composition, morbidity and development, not weight gain alone.


Introduction
Malnutrition is a leading cause of morbidity and mortality in children aged under five, especially in low-and middle-income countries (LMIC) 1 . Acute malnutrition (AM) comprises wasting and/or nutritional oedema. Wasting is defined by low weight-for-length/height z-score (WLZ/WHZ) and/or low mid-upper arm circumference (MUAC). AM is often subdivided into severe acute malnutrition (SAM) and moderate acute malnutrition (MAM). In 2019, of the 47 million children under five years of age who were acutely malnourished, 32·7 million suffered from MAM 1 .
To date, much focus has been on treating children with SAM since individual case-fatality is higher.
However, MAM also matters: it affects greater numbers of children globally; children with MAM have a three-fold increased risk of mortality compared to those without MAM; children with MAM are at-risk of deteriorating to SAM and undergo poorer physical and cognitive development compared to their well-nourished counterparts 1,2,3 . MAM treatment is context-specific and commonly involves one of two options: 1) improving the adequacy of the home diet through nutrition education/counselling 4 , or 2) supplementary feeding with an energy-dense product. The latter may be necessary in settings with food insecurity or where dietary diversity is poor, and involve targeted supplementary feeding programmes (SFPs) 5 . In recent years, treatment of MAM has predominantly focused on the use of fortified blended flours (FBF) and lipid-based nutrient supplements (LNS) as standard care 5,6 . A key barrier to scale-up is the lack of World Health Organization (WHO) guidelines on MAM. This review aims to inform an upcoming 2021 WHO guideline review where MAM (moderate wasting) is one of four key topics being examined 7 .
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01. 16.21249861 doi: medRxiv preprint

Methods
We followed the PRISMA (Preferred Reporting-Items for Systematic Reviews and Meta-Analyses) reporting framework 8 .

Inclusion Criteria
Eligibility and inclusion into the review were based on the following PICO (Population, Intervention, Comparator, and Outcome) framework: -Children aged 6-59 months diagnosed with MAM and living in a LMIC. MAM was defined as having a WHZ <-2 and ≥-3 based on WHO 2006 Growth Standards 9 , and/or a MUAC <12·5cm and ≥11·5cm, without bipedal oedema.
-A supplementary food product used for the treatment of MAM. This includes: FBFs such as corn-soy blend (CSB), LNSs or ready-to-use therapeutic/supplementary food (RUTF/RUSF), or any complementary food supplement to be consumed in addition to the home diet.
-A comparative treatment group containing participants with MAM, who are receiving either: no specific intervention ('usual care' control, since MAM is not routinely identified or treated in all settings); active control, which includes an alternative food/supplement, nutrition education/counselling, any intervention/combination of interventions used to treat MAM.
-The primary outcome of interest was recovery from MAM, defined as having a WHZ >-2, and/or a MUAC >12·5cm, without, bipedal oedema, and based on WHO 2006 Growth Standards. Secondary outcomes of interest were; persistent MAM, progression to SAM, death, defaulting, and any adverse-effects to treatment including body composition.
Only published studies with a control/comparator group were included. For generalizability and is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint The following databases were searched by two authors independently, with final search completed on the 29 th October 2020: EMBASE (Ovid), MEDLINE (Ovid), CINAHL (EBSCO), Global Health and ClinicalTrials.gov.
Searches were limited to human-only studies, published between the years 2006 and 2020, and written in English language. A detailed search strategy is in Appendix 1.
The reference lists of identified studies were also checked for papers that met inclusion criteria.

Data Collection
Studies identified in the search were initially screened by title and abstract to determine if they met eligibility criteria. They were then screened from their full text. Reasons for exclusion were documented.
Data was collected using a data collection form ( Table 1) generated from the 'checklist of items to consider in data collection or data extraction,' as part of the Cochrane Methods handbook 10 .
Additional information was sought on the differences in characteristics of children who recovered from MAM, versus those who did not recover/progressed to SAM/defaulted/died.

Risk of Bias
Risk of bias was determined using the Cochrane Collaboration's Tool for Assessing Risk of Bias 11 , which considers the following: 1) random sequence generation, 2) allocation concealment, 3) blinding of participants and personnel, 4) blinding of outcome assessment, 5) incomplete outcome data, and 6) selective reporting.
For each criterion, a judgement was made as to whether the trial was at a 'low-risk,' a 'high-risk,' or an 'unclear-risk' of bias ( Table 2).

Data Analysis
For our primary and secondary outcomes we used STATA v·14 12 to perform random-effects metaanalyses to estimate pooled risk ratios (RR) for the outcomes. Random-effects meta-analyses were chosen to incorporate the expected random variation in the effect of each intervention across the studies into the pooled estimates. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01. 16.21249861 doi: medRxiv preprint We also used descriptive synthesis to summarise findings about characteristics of children who recovered from MAM versus those who did not.
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint

Results
Our search (Appendix 1) generated 1,968 references after removing 309 duplicates. Thirteen papers were eligible for inclusion in the systematic review and meta-analysis: Figure 1. Reasons for exclusion of papers included: unpublished results; using NCHS references for defining MAM; using an alternative measurement of MAM (other than WHZ or MUAC); using a different definition for recovery from MAM (i.e. WHZ > -1), and/or investigating MAM prevention rather than MAM treatment (Appendix 2).

Study Characteristics
Individual study characteristics are presented in Table 1.
Twelve studies were conducted in African countries, the other in Iran 13 . Two were in urban areas; ten in rural settings; one was set in both an urban and rural area.
Sample sizes ranged from 81 to 2,712. Enrollment ages differed ( Table 1) is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. Also investigated effect on anaemia: found no effect on mean Hb levels between groups. All products exceeded WHO energy density recommendations (>0·8kcal/g) 14 (Appendix 3).

RCT
The primary outcome of all except two studies was the proportion of children with MAM who recovered. The remaining, investigated increments in the fat-free mass index (FFMI) or fat-mass accretion (kg) of participants 15 .
One study did not define recovery as obtaining a WHZ > -2, but instead defined recovery as WHZ > -1 13 .
Nevertheless, data was available on the proportions of children who reached a WHZ > -2, which was used in this analysis. The four cluster-RCTs reported outcomes adjusted for the clustered design.
In all studies, once a participant was classified as recovered from MAM, they were discharged from the treatment programme. Two studies treated participants for a maximum of 8 weeks 16 ; the remaining trials treated for a maximum of 12 weeks. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Lipid-Based Nutrient Supplement (LNS) vs Fortified Blended Four (FBF)
Seven trials compared an LNS and a FBF. One study was removed from the final meta-analysis, as it used a type of FBF, corn-soy blend (CSB), which has since been replaced by improved versions (CSB+/++) 16 . The reason was to reduce heterogeneity. Random-effects meta-analysis of the remaining trials (n=7,667) showed that children treated with an LNS were 4% more likely to recover with CSB+/++ (RR 1·05, 95%CI 1·01-1·09, p=0·009) ( Figure 2).
There was no evidence of heterogeneity across studies (p=0·825).
Progression to SAM whilst receiving either an LNS or a FBF was reported in five trials (n=7,043). Random-effects meta-analysis showed no significant difference in the risk of developing SAM amongst children treated with an LNS compared to FBF (RR=0·87, 95%CI 0·74-1·00, p=0·066). There was no evidence of heterogeneity (p=0·879).
One study reported no evidence of a difference in the risk of diarrhoea (p=0·980) or vomiting (p=0·220) during treatment with an LNS versus a FBF 17 .
One study investigated fat-free mass accretion after treatment 15 . This showed that children treated with LNS had, on average, 0·083 kg/m² higher FFMI than those treated with CSB (95%CI 0·003-0·163). Furthermore, there was no evidence of effect modification for the differences in FFMI by season, admission criteria, baseline FFMI, stunting, inflammation, and breastfeeding.
Changes in body composition when treating MAM with two different LNS products was investigated in two studies 15 . One observed that products containing 20% milk protein were better at promoting fat-free mass accretion than products which did not contain milk protein (FFMI difference: 0·097 kg/m 2 , 95%CI -0·002; 0·196).
When considering protein quality, one trial noted no association between LNS protein quality and recovery from MAM (p = 0·61) 24 .
In one study, treating MAM using a locally produced LNS (A-RUTF), which substituted half the amount of peanut

Different Formulations of FBF
One study investigated treating of MAM with two different formulations of a FBF; CSB+ and a nutrient-dense is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint produced CSB+ with malted sorghum. Recovery from MAM did not differ significantly in children treated with MSBP versus CSB+ (n=94 (90%) vs n=81 (81%), p=0·055).

Food Supplement vs Nutrition Education / Counselling
Two studies compared treatment of MAM with a food supplement versus a non-food supplement 13,18 . One compared child-centered counselling (CCC) with a blended food (CSB++) and LNS (RUSF) in Burkina Faso; the other compared an LNS (RUSF) with usual diet and nutrition education in Iran.
In Azimi 2020 no children progressed to SAM or died and only one child defaulted.
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint

Discussion
This was a systematic review and meta-analysis investigating the effectiveness of treatments for moderate acute malnutrition (MAM) in children aged 6-59 months in LMICs. Ten of the thirteen included trials compared the effectiveness of different supplementary food products in promoting MAM recovery, reflecting the recent focus on the use of products in MAM management 6 . Children treated with an LNS, had a 5% increased probability of recovery compared to those treated with CSB++. Furthermore, treatment with an LNS was associated with a lower risk of persistent MAM, compared with a FBF. Treatment with a food product was associated with a higher probability of recovery, a lower risk of progressing to SAM, and a lower risk of defaulting, compared to children who did not receive supplementation.
In the past, others have also reviewed MAM management 27,28,29,30,31 . Results for our primary outcome -recovery The following may contribute to the lower recovery rates seen amongst MAM children treated with FBFs compared with LNSs: 1) FBFs resemble staple foods and thus may be more openly shared amongst other household members 33,34,35 . 2) Children must consume roughly eight-times the mass of FBF compared to LNS, thus potentially discouraging breastfeeding / eating other foods provided at home 33,35 . 3) FBFs are lower in fat content and energy density compared to LNSs (Appendix 2). This is supported by evidence showing that LNS supplementation does not replace other foods in the diet 36,37 , is highly acceptable in an African context 38,39 , and is shared amongst household members substantially less than FBFs 40,41 . This review has highlighted some key gaps in the current body of evidence, which can be grouped into population, intervention/comparator and outcome related.
Out of thirteen included studies, twelve were conducted in African populations, making it difficult to generalise conclusions to South Asian contexts. This is problematic given that over half of all wasted children globally live in South Asia 1 , and likely have different growth trajectories and energy needs compared to children . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint from Africa 20 . Though promising early evidence highlights the effectiveness and acceptability of LNSs in South Asia 42 , more research is needed to understand similarities and differences compared with other populations.
Another striking feature of studies in our review is that all used active controls. This risks underestimating the true benefit of MAM treatment since the field reality is that children with MAM are neither identified not get any treatment in many settings worldwide. Future studies should consider more 'usual care' controls (i.e. minimal/no specific care) to better understand true impact of MAM care. This would be ethically acceptable in some settings in light of a recent recommendation that children with MAM presenting to primary care should not be routinely treated 43 . Cost-effectiveness data would be another important part of such work.
In eleven included trials, anthropometry and in particular recovery of 'normal' weight was the main outcome measure. Whilst anthropometry is widely used as the key measure of nutritional status, it is malnutrition-associated risk of mortality and morbidity that really matters rather than body size alone 44 . Wasting, low WHZ, which defines MAM is strongly associated with high risk of mortality 45,46 . However, there is increasing evidence that weight recovery does not necessarily mean return to low clinical risk of those who were never malnourished: children who have had MAM remain at high risk of relapse post treatment 47 . Children with MAM, who recovered as part of the LaGrone 2012 study, were followed-up for a further 12-month period in order to assess the long-term effects of treatment with CSB++ and RUSF 48 . During this period, only 63% of children remained well-nourished, highlighting the vulnerability of children post MAM recovery according to anthropometric measures alone. Measuring the proportions of children who progress to SAM, require admission or die during treatment, are more meaningful measurements of short-term clinical effectiveness of interventions, as these highlights unwanted negative outcomes along the spectrum of anthropometric status. However, these outcomes are relatively rare, can be more complex to measure and certainly in the case of mortality require far sample sizes for robust analysis -so are not commonly included in research.
Other more clinically meaningful outcome measures include body composition and it is encouraging to see studies in our review starting to assess this 49 . Body composition is especially valuable since it also gives information about potential longer term future risk, including of NCD 50 . There is concern that routine supplementation of MAM children with energy-dense products will encourage unhealthy weight-gain that risks predisposition to NCDs 51 . Balancing this evidence from one of the studies in our review suggests the opposite; that the majority of weight gain during supplementation with an energy-dense product is fat-free mass 15 , and that there are no significant differences in changes in body composition depending on the energy density of products 25  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01. 16.21249861 doi: medRxiv preprint and South Sudan investigated a combined protocol for treating acute malnutrition compared to standard care 20 .
They noted no significant difference between recovery rates (risk difference 0·03, 95%CI -0·05-0·10, p=0·52), and cost-effectiveness of the combined protocol (US$123 less per child recovered) compared to standard care. These results are echoed in Maust 2015 52 . Overall, combined management may be simpler, more cost-effective and reach more children.
We acknowledge some limitations of our review and meta-analysis: 1) Relevant papers could have been missed by restricting the search to papers published after 2006 and limiting to English language. 2) In the risk of bias assessment, we did not search for original paper protocols when determining selective reporting, and therefore, it is unclear as to how much reporting bias contributed to concealment of undesirable results. This was not considered an issue for the primary outcome, recovery, as all papers provided relevant data. 3) We only found a small number of eligible papers, and thus generalisability may be limited, particularly as most papers were conducted in African LMICs. Nevertheless, the small quantity of papers identified highlights the need for more robust research in MAM management, specifically those that address points raised in this review.
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint

Conclusion
How best to treat children with MAM is a key question for global child health. We found most current evidence focuses on the use of food supplements and involves studies with active controls rather than no-treatment control as is often the case in everyday practice. This risks underestimating the true benefits of treatment. In studies we identified, supplementation with LNS improves anthropometric recovery and prevents progression to SAM compared to supplementation with FBF. The role of nutrition counselling/education alone, or in combination with a supplementary food-product, warrants further research, particularly in areas with good food security.
Outcomes in current research focus on weight-recovery: future trials should include more clinically-meaningful outcomes, such as progression to SAM, admission, death and/or changes in body composition.

Author Contributions
Gluning, I -designed the search strategy, conducted the search and meta-analysis, and wrote the initial manuscript. Kerac M -conceived the project, supervised and provided feedback on manuscript drafts. Bander A -conducted a second search and review of references Bailey J -provided feedback on manuscript drafts. Opondo C -conceived the project, supervised, assisted in meta-analysis and provided feedback on drafts of manuscript All authors: read and approve of final version of the manuscript

Declaration of Interest
The authors have declared that no competing interests exist.

Funding
There was no funding source.
. CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted January 20, 2021. ; https://doi.org/10.1101/2021.01.16.21249861 doi: medRxiv preprint