Extending Supplementary Feeding for Children Younger Than 5 Years With Moderate Acute Malnutrition Leads to Lower Relapse Rates : Journal of Pediatric Gastroenterology and Nutrition

Secondary Logo

Journal Logo

Original Articles: Nutrition

Extending Supplementary Feeding for Children Younger Than 5 Years With Moderate Acute Malnutrition Leads to Lower Relapse Rates

Trehan, Indi*; Banerjee, Somalee*; Murray, Ellen*; Ryan, Kelsey N.*; Thakwalakwa, Chrissie; Maleta, Kenneth M.; Manary, Mark J.*

Author Information
Journal of Pediatric Gastroenterology and Nutrition 60(4):p 544-549, April 2015. | DOI: 10.1097/MPG.0000000000000639


See “Extending Supplementary Nutrition in Malnutrition: More Is Better?” by Forbes and Brewster on page 424.

Approximately 5% of children <5 years of age worldwide experience moderate acute malnutrition (MAM), defined as having a weight-for-height z score (WHZ) between −2 and −3 (1). Compared with well-nourished children, the mortality rate for children with MAM is 3-fold higher (2), and survivors experience deficits in cognitive development (3,4). Supplementary feeding programs remain the standard of care for treating children with MAM as outpatients in rural settings (5), with the hope of reducing the rate of these adverse outcomes (6).

Unfortunately, evidence suggests that children successfully treated for MAM continue to be at an increased risk for malnutrition and death 12 months after initial recovery (7). Improved treatment protocols are thus needed to reduce the long-term risk of adverse outcomes, while continuing to achieve anthropometric recovery at an affordable cost. Operational practice varies when treating children with MAM, with some programs treating until children reach a fixed anthropometric target (most often to WHZ > −2), whereas others treat children for a fixed duration (8). As yet, there have been no direct comparisons to inform which protocol is more effective (9), and the optimal exit criteria from targeted supplementation programs remain a high-priority research topic (10). Similarly, the trend toward using a fixed mid-upper arm circumference (MUAC) level of 12.5 cm because diagnosis and discharge criteria (9,10) have not been explicitly compared with a fixed duration of treatment.

This pilot study tested the hypothesis that clinical outcomes 12 months after recovery from MAM would be better in children treated for a fixed duration of 12 weeks, rather than being treated only until they reach the recovery goal of WHZ > −2. By comparing these 2 treatment protocols, we also explored whether any information about anthropometric predictors of adverse outcomes during the follow-up period can be gleaned to inform optimal graduation criteria in operational practice for MAM supplementary feeding programs (9,11,12).


This was a quasiexperimental concurrently controlled prospective study comparing the 12-month clinical outcomes of children successfully treated for MAM with supplementary feeding for 12 weeks without respect to anthropometry, to those children fed to the anthropometric goal of WHZ > −2. The study was conducted among children who recovered from MAM after participating in a randomized controlled clinical trial comparing 3 supplementary foods at 18 different sites in southern Malawi in 2009–2011. Recovery from MAM was defined as achieving WHZ > −2; upon recovery supplementary feeding was stopped for children at 14 of these 18 sites (treat-to-goal group). Children at 4 randomly selected study sites were provided with the additional intervention of extended supplementary feeding to complete 12 total weeks of total therapy (treat-to-time group). The study was approved by the University of Malawi College of Medicine Research and Ethics Committee and the Washington University in St Louis Human Research Protection Office.


All of the children ages 6 to 59 months who recovered from MAM as part of a randomized controlled trial comparing the clinical efficacy of Corn-Soy Blend Plus-Plus (now called Supercereal Plus, produced locally by Rab Processors, Blantyre, Malawi), soy ready-to-use supplementary food (soy ready-to-use supplementary food [RUSF], produced locally by Project Peanut Butter, Blantyre, Malawi), and soy/whey ready-to-use supplementary food (soy/whey RUSF, Nutriset, Malaunay, France) in the treatment MAM were recruited for this study (13,14). Children who did not recover during the initial MAM treatment study were offered treatment for severe acute malnutrition (SAM) and excluded from this study. Children with chronic debilitating illnesses other than tuberculosis and human immunodeficiency virus (HIV) were also excluded.

Study Procedures

After providing verbal and written consent, caregivers provided demographic information and completed the validated Household Food Insecurity Access Scale (HFIAS) (15). Trained nutrition researchers and senior pediatric research nurses assessed each child for acute malnutrition using standard methodologies for anthropometric measurements: weight was measured using an electronic scale to the nearest 5 g; length was measured in triplicate to the nearest 0.2 cm using a rigid length board; MUAC was measured with a standard insertion tape to the nearest 0.2 cm; and each child was evaluated for the bilateral pitting edema diagnostic of kwashiorkor (16).

Anthropometry and the HFIAS were repeated at scheduled follow-up visits 3, 6, and 12 months after enrollment. Caregivers and village health workers were also encouraged to refer children back to the clinics any time there was a concern for nutritional deterioration. At each visit, caregivers were also asked about the child's appetite, infectious symptoms, health center visits, and medication use during the preceding 2 weeks.

Caregivers of children who relapsed to MAM at any visit were given focused nutritional counseling and supplemental soy RUSF rations and treated to recovery through biweekly clinic visits. Those diagnosed with SAM were treated with ready-to-use therapeutic food as outpatients or transferred to inpatient care as appropriate (16). If a child missed a scheduled follow-up visit, community health workers sought them out in their villages and encouraged them to return to the clinic as soon as possible.


The primary outcomes for this study were survival and nutritional status during the 12 months after the treatment intervention completed. Each child was classified as having “remained well nourished,” defined as WHZ > −2 or MUAC ≥ 12.5 cm at every follow-up visit for 12 months; “relapsed to MAM,” defined as −3 < WHZ ≤ −2 and MUAC < 12.5 cm at any point during the follow-up period; “developed severe acute malnutrition (SAM),” defined as WHZ ≤ −3 (marasmus) and/or bipedal edema (kwashiorkor) at any point during the follow-up period; “died”; or “defaulted,” defined as not completing the full 12 months of follow-up.

The criteria of both MUAC < 12.5 cm and WHZ < −2 to define relapses into MAM were used, whereas in operational clinical practice generally either criteria are used (16). These more strict criteria were intentionally chosen to help identify a true decrease in the child's nutritional health because the use of WHZ criteria alone is often complicated by short-term linear growth (7). Linear growth is commonly seen as a child recovers, which often makes it difficult to achieve recovery by WHZ criteria if real-time length measurements are used for the calculation; thus, recovery goals from MAM are defined on the basis of the initial length at the time of diagnosis (17). A child may grow in stature and body mass, both signs of recovery, yet appear to relapse when they return for follow-up visits because updated calculations of their WHZ using their new increased length makes them appear to have a low WHZ. The inclusion of MUAC as an additional and necessary relapse criterion is thus meant to avoid this conundrum.

Adverse outcomes during the follow-up period included the development of MAM or SAM, loss to follow-up (defaulting), or death. The first adverse outcome identified during the follow-up period was used to determine the final classification.

Data Analyses

Anthropometric z scores were calculated using Anthro or AnthroPlus (World Health Organization, Geneva), based on the 2006 World Health Organization Child Growth Standards (18). Comparisons of outcomes between the treat-to-time and treat-to-goal groups were made using either Fisher exact test or the χ2 test with Yates correction for dichotomous variables and Student t test for continuous variables. P values <0.05 were considered statistically significant. The intention-to-treat approach was used for all of the analyses.

To determine risk factors for poor outcomes while controlling for baseline differences in the enrollment characteristics of children in the 2 groups, logistic regression models for remaining well nourished and death during the follow-up period were created. The regression models were created using a stepwise backward method where the criteria for inclusion of a term in the final model was P < 0.10. Covariates initially included in the models were intervention group (treat-to-time vs treat-to-goal), age, sex, whether the child's mother was alive, whether the child's father was alive, whether the mother was the primary caregiver of the child, whether the father was present in the home, mother's HIV status, child's HIV status, number of children in the household <5 years of age, the month in which treatment was initiated, the child's initial MUAC, WHZ, HAZ, HFIAS score, and the caregiver's report of appetite at enrollment. Covariates with coefficients with a 95% confidence interval (CI) that did not include 1 were considered significant. Food insecure months were defined as January through April, as the annual harvest in southern Malawi generally occurs in April through May.

To assess the influence of a range of MUAC and WHZ measurements at the time of graduation from MAM therapy on the rate of remaining well nourished throughout the follow-up period, receiver operating characteristic curves were created. The number of additional children needed to treat to a range of higher MUAC and WHZ recovery thresholds to keep 1 additional child well nourished throughout the follow-up period was also calculated.


Between November 2009 and January 2011, 382 children with MAM were enrolled in the treat-to-time group (Fig. 1), whereas 1967 children were enrolled in the treat-to-goal group (7). Characteristics of these 2 groups of children at the time they each had reached WHZ > −2 were clinically generally similar (Table 1), except that the children in the treat-to-time group were more likely to be girls, to still be breast-feeding, and were significantly less stunted. The children in the treat-to-goal group had a higher MUAC and also a higher rate of maternal HIV.

Flow of patients in the treat-to-time group. MAM = moderate acute malnutrition; WHZ = weight-for-height z score; SAM = severe acute malnutrition.
Characteristics of study children with MAM upon successful completion of initial MAM therapy

During the extended feeding time between when children in the treat-to-time group achieved nutritional recovery (WHZ > −2) and the end of the 12-week time span, 8 (2%) of these 382 children developed kwashiorkor and 4 (1%) developed marasmus, despite continued supplementary feeding (Fig. 1). These 12 children thus did not complete their 12 weeks of MAM therapy, but instead were transferred into a treatment protocol for SAM, receiving 175 kcal · kg−1 · day−1 of ready-to-use therapeutic food, until they recovered (WHZ > −2 without edema) (19,20). These children proved to be a particularly high-risk group during the 12-month follow-up period, with only 5 of the 12 (42%) remaining well nourished during that time.

The remaining 370 children in the treat-to-time group showed evidence of continued growth during the extended 12-week feeding period, increasing their average WHZ from −1.61 ± 0.47 to −1.14 ± 0.67 and their average MUAC from 12.4 ± 0.7 to 13.0 ± 0.8 cm (P < 0.0001 for both comparisons). No decrease in stunting was seen, with 282 (76%) having HAZ < −2 and 143 (39%) still severely stunted with HAZ < −3, with an overall average HAZ of −2.75 ± 1.19.

Of the 382 children enrolled in the treat-to-time group, 337 (88.2%) remained well nourished during the first 3 months of follow-up, 301 (78.8%) remained well nourished throughout the first 6 months of follow-up, and 272 (71.2%) remained well nourished throughout the 12-month follow-up period (Fig. 1 and Table 2). These rates were all significantly higher than the rates among children in the treat-to-goal group (7), an improvement of 9.3% (95% CI 5.0–13.7, P < 0.0001) at 3 months, 10.0% (95% CI 5.0–15.0, P < 0.0001) at 6 months, and 8.7% (95% CI 3.4–13.9, P = 0.0015) at 12 months. These 272 children in the treat-to-time group who remained well nourished throughout the study had significantly higher WHZ and HAZ at the end of the follow-up period, compared with the 1230 children in the treat-to-goal group who remained well nourished (Table 2).

Outcomes of children who were successfully treated for moderate malnutrition 12 months later*

By 3 months, 6.5% (95% CI 2.5–10.5, P = 0.002) fewer children in the treat-to-time group had relapsed to MAM or SAM than in the treat-to-goal group; by 6 months, 5.0% (95% CI 0.35–9.6, P = 0.04) less had relapsed; and by 12 months, 3.6% (95% CI −1.2–8.4, P = 0.16) less had relapsed. A nonsignificant trend toward increased survival was also seen in the treat-to-time group; the relative risk for mortality in the treat-to-time group at 3 months was 0.18 (95% CI 0.025–1.35, P = 0.07) at 3 months, 0.19 (95% CI 0.48–0.79, P = 0.008) at 6 months, and 0.49 (95% CI 0.23–1.05, P = 0.08) at 12 months.

Two logistic regression models were developed to identify baseline characteristics that were associated with children in both cohorts who either remained well nourished or had died during 12 months of follow-up (Table 3). Children in both the treat-to-goal and treat-to-time groups were all included together in these models. Owing to the relatively small number of deaths, the model for risk of death showed poor predictive power. Although the treat-to-time group had a significantly higher rate of remaining well nourished than did the treat-to-goal group in univariate analysis (Table 2), the treatment protocol used proved not to be a significant predictor of remaining well nourished in these multivariate models. Instead, the characteristics with the strongest statistical association with remaining well nourished were greater WHZ and MUAC at the onset of the follow-up period, and the characteristics with the strongest association with death were HIV infection and low MUAC.

Variables associated with remaining well nourished or death during the follow-up period in logistic regression models*

MUAC at the end of initial MAM therapy proved to be slightly better than WHZ in predicting which children would remain well- nourished during the subsequent 12 months (supplementary Figures 1–3, https://links.lww.com/MPG/A405, https://links.lww.com/MPG/A406, https://links.lww.com/MPG/A407). Children who finished their initial MAM treatment with MUAC measurements of at least 12.9 cm or with WHZ of at least −1.75 were significantly more likely to remain well nourished during the course of the follow-up period (supplementary Table 1, https://links.lww.com/MPG/A408); these odds improved in a dose–response manner as their anthropometry increased. If the WHZ recovery threshold were increased from −2 to −1.5, nearly 16 children would need to be treated to keep 1 additional child well nourished; if the threshold were increased to −1, then the additional number needed to treat would be nearly 7 relative to the WHZ of −2 threshold.


Children treated for MAM for 12 weeks were more likely to remain well nourished during the subsequent 12 months than children treated simply to WHZ > −2. We also observed a trend toward lower rates of SAM and death among children in the treat-to-time group. Those who did remain well nourished had higher WHZ and HAZ parameters at the end of those 12 months. This likely reflects the better anthropometric indices at the start of the 12-month follow-up period and suggests that these indices achieved at the end of 12 weeks of MAM therapy can be sustained during the subsequent 12 months.

It may seem that treating children with MAM for a fixed duration of time was more effective in reducing relapse during the subsequent 12 months, but logistic regression modeling actually identified greater MUAC and WHZ as the most important variables that predicted which children would remain well nourished; whether they were in the treat-to-time or treat-to-goal group proved not to be a significant factor. This suggests that optimized anthropometric criteria are more important than a predetermined extended duration of feeding.

Given that so many children did not remain well nourished using the standard recovery threshold, we explored whether higher MUAC and/or WHZ thresholds could be used to determine a more optimal time to end therapy for MAM. This would inevitably involve a tradeoff in cost as this will extend the duration of feeding for many additional children, but may be a necessary measure in vulnerable populations (11). Continuing to feed a child who may no longer be strictly wasted, but nevertheless maintains a WHZ and/or MUAC on the cusp of MAM, could provide a safety net against subsequent acute infectious illnesses and food insecurity in this vulnerable population (11,21). Increasing the WHZ recovery threshold to −1.5 would greatly increase the specificity of predicting which children remained well nourished to 67% but would require treating 16 children to this level rather than to the WHZ of −2 level to keep 1 additional child well nourished. Similarly, increasing the recovery criteria to WHZ of −1.0 would be nearly 94% specific and would require treating only 7 additional children to keep 1 additional child well nourished compared with treating only to WHZ of −2. Presumably, similar significant benefits in mortality rates would also be realized by an approach that involved treating to a higher WHZ threshold, but the present study was limited in its sample size to make an accurate estimate of this benefit. This higher WHZ of −1.5 as a recovery threshold has been recommended in the past (22) but international recommendations (23), research studies (5,24), and operational programs (11,13) have generally used WHZ of −2 as the recovery threshold, although there is no clear consensus on this issue (8–10). Our results suggest that a return to this higher WHZ of −1.5 may help reduce the rate of long-term nutritional relapse. In addition, our data here suggest that MUAC of 12.5 cm may not be sufficient to prevent relapse in the subsequent year after recovery, which may be a particularly relevant finding given that more supplementary feeding programs are operationally relying more frequently on MUAC as entry and graduation criteria. In particular, a larger MUAC at the completion of supplementary feeding was significantly predictive of lower mortality rates (while WHZ was not), adding to the growing body of evidence that MUAC is a better predictor of mortality among malnourished children (25,26).

The results of this study may not be applicable to areas outside of rural sub-Saharan Africa, in areas with lower rates of falciparum malaria and HIV-1 (27), or where the incidence of acute malnutrition does not follow a similar seasonal pattern. In addition, the initial treatment for MAM was conducted in the context of a randomized clinical trial that provided specialized fortified foods with intensive nutritional counseling (13); outcomes would likely be worse in a programmatic setting with fewer resources.

Future interventions should attempt to reduce the relapse rate by studying and restoring physiologic and immunologic competence (28), in addition to anthropometry, as these may indeed be the key prognostic variables. Potential interventions could include provision of bed nets, antimalarial prophylaxis, micronutrient supplementation, deworming, closer integration with immunization programs and with HIV testing and treatment programs, and regular follow-up with nursing staff to help educate and advise caregivers about proper feeding techniques for their growing child.

The study results should be considered preliminary as children were not specifically randomized to different recovery thresholds, but instead treated to either a prespecified goal weight (WHZ of −2) or for a prespecified treatment duration (12 weeks). This design, however, did generate a range of anthropometry at graduation and allowed us a unique opportunity to conduct these analyses. A randomized prospective study testing the durability of optimizing MAM treatment in keeping children well nourished for a prolonged period after treatment ends is warranted. Given the enormous burden that acute malnutrition plays in global child mortality, this study adds to the larger context of work demonstrating that there is still much that can be done (1) to optimize the care of acutely malnourished children to improve both their short- (29) and long-term (7,30) outcomes.

In conclusion, we show that in this 12-month observational study following 12 weeks of supplementary feeding for MAM, 71% of children who recovered from MAM remained well nourished, whereas <2% died. This is the first report of a direct comparison of long-term outcomes for children treated to a target WHZ compared with children treated for a fixed duration of time. Higher anthropometry at the end of MAM therapy appears to be the most important predictor of long-term nutritional status, regardless of the duration of therapy provided, and thus more stringent recovery criteria such as WHZ of −1.5 and/or MUAC of 13 cm should be considered for moderately malnourished children to decrease the rate of relapse.


1. Black RE, Victora CG, Walker SP, et al. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet 2013; 382:427–451.
2. Olofin I, McDonald CM, Ezzati M, et al. Associations of suboptimal growth with all-cause and cause-specific mortality in children under five years: a pooled analysis of ten prospective studies. PLoS ONE 2013; 8:e64636.
3. Martorell R, Ho TJ. Malnutrition, morbidity, and mortality. Pop Dev Rev 1984; 10:49–68.
4. Nassar MF, Shaaban SY, Nassar JF, et al. Language skills and intelligence quotient of protein energy malnutrition survivors. J Trop Pediatr 2012; 58:226–230.
5. Nikiema L, Huybregts L, Kolsteren P, et al. Treating moderate acute malnutrition in first-line health services: an effectiveness cluster-randomized trial in Burkina Faso. Am J Clin Nutr 2014; 100:241–249.
6. World Food Programme. Nutrition at the World Food Programme: Programming for Nutrition-Specific Interventions. Rome: World Food Programme; 2012.
7. Chang CY, Trehan I, Wang RJ, et al. Children successfully treated for moderate acute malnutrition remain at risk for malnutrition and death in the subsequent year after recovery. J Nutr 2013; 143:215–220.
8. Ashworth A, Ferguson E. Dietary counseling in the management of moderate malnourishment in children. Food Nutr Bull 2009; 30:S405–S433.
9. World Health Organization, United Nations Children's Fund, World Food Programme, United Nations High Commissioner for Refugees. WHO, UNICEF, WFP and UNHCR Consultation on the Programmatic Aspects of the Management of Moderate Acute Malnutrition in Children Under Five Years of Age. Geneva: World Health Organization, United Nations Children's Fund, World Food Programme, United Nations High Commissioner for Refugees; 2010.
10. Annan RA, Webb P, Brown R. Management of Moderate Acute Malnutrition (MAM): Current Knowledge and Practice. CMAM Forum. http://www.cmamforum.org/Pool/Resources/MAM-management-CMAM-Forum-Technical-Brief-Sept-2014.pdf. Published September 2014. Accessed October 2014.
11. Briend A, Van den Broeck J, Fadnes LT. Target weight gain for moderately wasted children during supplementation interventions—a population-based approach. Public Health Nutr 2011; 14:2117–2123.
12. Lenters LM, Wazny K, Webb P, et al. Treatment of severe and moderate acute malnutrition in low- and middle-income settings: a systematic review, meta-analysis and Delphi process. BMC Public Health 2013; 13 (suppl 3):S23.
13. LaGrone LN, Trehan I, Meuli GJ, et al. A novel fortified blended flour, corn-soy blend “plus-plus,” is not inferior to lipid-based ready-to-use supplementary foods for the treatment of moderate acute malnutrition in Malawian children. Am J Clin Nutr 2012; 95:212–219.
14. Wang RJ, Trehan I, LaGrone LN, et al. Investigation of food acceptability and feeding practices for lipid nutrient supplements and blended flours used to treat moderate malnutrition. J Nutr Educ Behav 2013; 45:258–263.
15. Coates J, Swindale A, Bilinsky P. Household Food Insecurity Access Scale (HFIAS) for Measurement of Food Access: Indicator Guide. Washington DC: USAID; 2007.
16. World Health Organization, United Nations Children's Fund. Child Growth Standards and the Identification of Severe Acute Malnutrition in Infants and Children. Geneva: World Health Organization, United Nations Children's Fund; 2009.
17. Shoham J, Duffield A. Proceedings of the World Health Organization/UNICEF/World Food Programme/United Nations High Commissioner for Refugees Consultation on the management of moderate malnutrition in children under 5 years of age. Food Nutr Bull 2009; 30:S464–S474.
18. WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards: Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight-for-Height and Body Mass Index-for-Age: Methods and Development. Geneva: World Health Organization; 2006.
19. Manary MJ, Sandige HL. Management of acute moderate and severe childhood malnutrition. BMJ 2008; 337:a2180.
20. Trehan I, Manary MJ. Management of severe acute malnutrition in low-income and middle-income countries. Arch Dis Child 2015; 100:283–287.
21. Lazzerini M, Rubert L, Pani P. Specially formulated foods for treating children with moderate acute malnutrition in low- and middle-income countries. Cochrane Database Syst Rev 2013; 6:CD009584.
22. World Health Organization. The Management of Nutrition in Major Emergencies. Geneva: World Health Organization; 2000.
23. United Nations High Commissioner for Refugees, World Food Programme. Guidelines for Selective Feeding: The Management of Malnutrition in Emergencies. Geneva: United Nations High Commissioner for Refugees, World Food Programme; 2011.
24. Lagrone L, Cole S, Schondelmeyer A, et al. Locally produced ready-to-use supplementary food is an effective treatment of moderate acute malnutrition in an operational setting. Ann Trop Paediatr 2010; 30:103–108.
25. Berkley J, Mwangi I, Griffiths K, et al. Assessment of severe malnutrition among hospitalized children in rural Kenya: coomparison of weight for height and mid upper arm circumference. JAMA 2005; 294:591–597.
26. Briend A, Maire B, Fontaine O, et al. Mid-upper arm circumference and weight-for-height to identify high-risk malnourished under-five children. Matern Child Nutr 2012; 8:130–133.
27. Trehan I, O’Hare BA, Phiri A, et al. Challenges in the management of HIV-infected malnourished children in sub-Saharan Africa. AIDS Res Treat 2012; 2012:790786.
28. Rytter MJ, Kolte L, Briend A, et al. The immune system in children with malnutrition-a systematic review. PloS ONE 2014; 9:e105017.
29. Trehan I, Goldbach HS, LaGrone LN, et al. Antibiotics as part of the management of severe acute malnutrition. N Engl J Med 2013; 368:425–435.
30. Kerac M, Bunn J, Chagaluka G, et al. Follow-Up of Post-Discharge Growth and Mortality after Treatment for Severe Acute Malnutrition (FuSAM Study): a prospective cohort study. PloS ONE 2014; 9:e96030.

anthropometry; children; long-term outcomes; Malawi; mid-upper-arm circumference; moderate acute malnutrition; ready-to-use supplementary food; wasting; weight-for-height z score

Supplemental Digital Content

© 2015 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,