Ready-to-use therapeutic food for treating undernutrition in children from 6 months to 5 years of age

  • Protocol
  • Intervention



This is the protocol for a review and there is no abstract. The objectives are as follows:

To assess the effects of RUTF on health outcomes such as recovery rate, relapse during the intervention period, anthropometrical status, weight gain and mortality in children with moderate or severe undernutrition.


Description of the condition

Undernutrition occurs when the quantity of one or more macronutrients available to body tissues is inadequate to sustain optimal bodily functions (Manary 2008), and this is usually accompanied by numerous micronutrient deficiencies. Undernutrition is a broad concept that includes a variety of clinical conditions such as kwashiorkor, marasmus, marasmic kwashiorkor, underweight, wasting or stunting, and micronutrient deficiencies. For the purpose of this review, the terms undernutrition and malnutrition are used interchangeably. This review is focused on macronutrient undernutrition, so it includes all of the above conditions, which could also be accompanied by different degrees of micronutrient deficiencies.

Groups that are commonly affected by undernutrition are infants and young children, pregnant and lactating women, and the elderly. More than 77 million children are born every year in the 36 countries with the highest burden of malnutrition (21 of the countries are in Africa, 13 in Asia, and two in Latin America) (Bhutta 2008; Black 2008). Of these children, about 7.4 million die before the age of three years and a further 0.6 million die between the ages of three and five (Bhutta 2008). Short term consequences include mortality and morbidity, for example, pneumonia, diarrhoea, fatigue, impaired thermoregulation (Black 2008). In the long run, undernutrition in children may affect adult size, intellectual ability, economic productivity, reproductive performance, and increase the risk of metabolic disorders and cardiovascular disease (Black 2008).

In children under five years of age, undernutrition can be classified as moderate or severe. Moderate undernutrition - often referred to as moderate acute malnutrition (MAM) in the literature - is defined as a weight for height z score (WHZ) between two and three standard deviations (SDs) below the mean. Severe undernutrition - often referred to as severe acute malnutrition (SAM) in the literature - is defined as a WHZ of more than three SDs below the mean, or an mid-upper arm circumference (MUAC) of less than 115 mm, or the presence of nutritional edema (Collins 2003; Manary 2008; WHO and UNICEF 2009). Moderate or severe undernutrition without bilateral pitting edema is termed marasmus, whereas the presence of bilateral pitting edema is termed kwashiorkor (Manary 2008). A detailed classification system for MAM and SAM is presented in Table 1 and Table 2.

Table 1. Classification system for acute malnutrition in children under 60 months (Collins 2003)
  1. The WHO and UNICEF now recommend that the cut-off value for the mid-upper arm circumference for severe acute malnutrition should be increased to 115 mm (WHO and UNICEF 2009). The adoption of this higher cut-off value will sharply increase the case loads, which may influence the cost of nutrition programmes greatly (WHO and UNICEF 2009). However, detecting more children earlier as severely malnourished will lead to a shorter period needed to treat them, which may bring down the cost per child (WHO and UNICEF 2009).

Complicated malnutritionSevere uncomplicated malnutritionModerate uncomplicated malnutrition
<80% of median weight for height (< -2 SD score)<70% of median weight for height (< -3 SD score)

70 to 80% of median weight for height (≤ -3 SD score to > -2 SD score)


no edema


Bilateral pitting edema


Bilateral pitting edema



Mid-upper arm circumference < 110 mm


Mid-upper arm circumference < 110 mm


Mid-upper arm circumference 110 to 125 mm

AND one of the following:


Lower respiratory tract infection

High fever

Severe dehydration

Severe anaemia

Not alert



Clinically well




Clinically well


Inpatient stabilisation careOutpatient therapeutic careOutpatient supplementary feeding
Table 2. Classification of severe acute malnutrition (Collins 2006)
  1. IMCI = Integrated Management of Childhood Illness.

    *Grade 1 = mild edema on both feet or ankles.

    **Grade 2 = moderate edema on both feet, plus lower legs, hands or lower arms.

    ***Grade 3 = severe generalised edema affecting both feet, legs, hands, arms and face.

    The WHO and UNICEF now recommend that the cut-off value for the MUAC for severe acute malnutrition should be increased to 115 mm (WHO and UNICEF 2009). The adoption of this higher cut-off value will sharply increase the case loads, which may influence the cost of nutrition programmes greatly (WHO and UNICEF 2009). However, detecting more children earlier as severely malnourished will lead to a shorter period needed to treat them, which may bring down the cost per child (WHO and UNICEF 2009).

    ¤IMCI criteria: 60 respirations/min children age < 2 months; 50 respirations/min for age 2 to 12 months; 40 respirations/min for ages 1 to 5 years; 30 respirations/min for age > 5 years.

Severe acute malnutrition with complicationsSevere acute malnutrition without complications

Bilateral pitting edema grade 3*** (severe edema)


Mid-upper arm circumference < 110 mm


Mid-upper arm circumference <110 mm and bilateral pitting edema grades 1* or 2** (marasmic kwashiorkor)


Bilateral pitting edema grades 1* or 2** with mid-upper arm circumference ≥ 110 mm


  • Appetite

  • Clinically well

  • Alert

Mid-upper arm circumference <110 mm or bilateral pitting edema grades 1* or 2**


One of the following:

  • Anorexia

  • Lower-respiratory tract infection¤

  • Severe palmar pallor

  • High fever

  • Severe dehydration

  • Not alert

Inpatient care IMCI/WHO protocolOutpatient therapeutic care protocols

Although some diseases and conditions may contribute to the onset of undernutrition (for example, HIV/AIDS, TBC, kidney failure), the majority of cases are rooted in poverty and food insecurity. Furthermore, undernutrition and infection have a reciprocal effect since a lower host response to infection contributes to compromised nutritional status and vice versa (Kruger 2008; Naude 2008). Infections are associated with anorexia (loss of appetite) and decreased food intake; fever increases energy expenditure; and diarrhoea decreases nutrient absorption; finally resulting in wasting and higher mortality from infectious diseases (Kruger 2008; Naude 2008).

Description of the intervention

Ultimately, the only way to bring an end to undernutrition is to address economic deprivation and inequity. However, major reductions may also be facilitated through specific nutritional interventions (Black 2008). Hospitalised standard treatment for severely malnourished children first entails treatment with F75 (the starter milk-based therapeutic formula; also called phase 1 or stabilisation phase) (ACF International Network 2009). During this stabilisation treatment the edema (if present) starts to disappear, leading to weight loss (fluid loss). F75 aids in jump starting the metabolism and restoring hydroelectric equilibrium (ACF International Network 2009). Thereafer F100 (a milk-based therapeutic diet; also called phase 2) is given to initiate weight gain.

Ready-to-use foods (RUF) are energy-dense food with a low moisture content that can be eaten directly from the packaging. When such a product is used for nutritional rehabilitation of children with moderate or severe undernutrition, it is called ready-to-use therapeutic food (RUTF). RUTF was originally developed as a home-based alternative to F100. RUTF, either in the form of a solid or semi-solid feed, has a similar nutrient profile to F100 (other than containing iron), and it is more energy- and nutrient-dense than F100 (Collins 2006; WHO 2007). The nutritional contents of RUTF as recommended by the World Health Organization (WHO) are displayed in Table 3.

Table 3. Nutritional composition of RUTF, as recommended by the WHO (WHO 2007)
Moisture content2.5% maximum
Energy520 to 550 kCal/100g
Protein10% to12% total energy
Lipids45% to 60% total energy
Sodium290 mg/100g maximum
Potassium1110 to 1400 mg/100g
Calcium300 to 600 mg/100g
Phosphorus (excluding phytate)300 to 600 mg/100g
Magnesium80 to 140 mg/100g
Iron10 to 14 mg/100g
Zinc11 to 14 mg/100g
Copper1.4 to 1.8 mg/100g
Selenium20 to 40 μg
Iodine70 to 140 μg/100g
Vitamin A0.8 to 1.1 mg/100g
Vitamin D15 to 20 μg/100g
Vitamin E20 mg/100g minimum
Vitamin K15 to 30 μg/100g
Vitamin B10.5 mg/100g minimum
Vitamin B21.6 mg/100g minimum
Vitamin C50 mg/100g minimum
Vitamin B60.6 mg/100g minimum
Vitamin B121.6 μg/100g minimum
Folic acid200 μg/100g minimum
Niacin5 mg/100g minimum
Pantothenic acid3 mg/100g minimum
Biotin60 μg/100g minimum
n-6 fatty acids3% to 10% of total energy
n-3 fatty acids0.3% to 2.5% of total energy

RUTF can either be imported or produced locally. Two examples of commercially produced RUTF are a peanut-based paste called Plumpy'nut® (developed by Nutritset and the Institute for Research and Development, France) and a solid biscuit made from cooked wheat which is called BP100® (developed by Compact, Denmark) (Collins 2004; Navarro-Colorado 2005). Both are fortified with micronutrients and have very low water activity, which discourages microbial growth (Brewster 2006; WHO 2007; Kruger 2008). This is an important characteristic since clean safe water is not freely available in many poor communities. Children as young as six months can consume RUTF with a homogenous paste texture. Solid RUTF can be soaked in clean boiling water and eaten as porridge by such young children, or it can be consumed as a cookie by older children.

Communities can learn to produce their own RUTF, as is the case of Malawi where a peanut-based RUTF is produced (Sandige 2010). A typical recipe for a peanut-based RUTF is displayed in Table 4. Examples of other countries that manufacture RUTF are Ethiopia, Niger and the Democratic Republic of Congo in Africa as well as Sri Lanka, Indonesia and Pakistan in Asia (DFID 2009). The manufacturing equipment and technology needed to produce RUTF is simple and can be transferred to any country with minimal industrial infrastructure (WHO 2007). The scale of production will determine the methods of quality control and exact cost (Manary 2006), but on average a RUTF costs about US $3 per kilogram when locally (non-commercially) produced (WHO 2007). Children with SAM normally need 10 to 15 kg of RUTF given over a period of six to eight weeks for recovery (WHO 2007).

Table 4. A typical recipe for RUTF (Manary 2006)
  1. *Strict quality control is essential.

Ingredient% weight
Full-fat milk30
Vegetable oil15
Peanut butter*25
Mineral-vitamin mix1.6

Recipes for RUTF do not have to include peanut or milk powder although the WHO recommends that at least half of the proteins should come from a milk source (WHO 2007). Peanuts can cause allergic reactions and are also known to have a high risk for aflatoxin contamination. Milk powder is expensive and must often be imported (Collins 2004). The cost of milk powder in Malawi constitutes more than half of the cost of the final RUTF (Collins 2004). When considering developing a RUTF for non-commercial production, the following basic ingredients should be present (Collins 2004):

  • A staple food as the main ingredient (preferably a cereal);

  • A protein supplement from a plant or animal food (for example, beans, groundnuts, milk, meat, chicken, fish, egg). The RUTF must, however, be cost-effective and therefore legumes and oilseeds are mostly used;

  • A vitamin and mineral supplement (a vegetable and/or fruit);

  • An energy supplement (a fat, oil or sugar) to increase the energy density.

The food safety of the production process should be strictly monitored, with careful attention given to avoid contamination from microorganisms or other harmful substances (for example, heavy metals, pesticides, anti-nutritional factors such as phytate or protease inhibitors) (WHO 2007). Three recipes for locally produced RUTF are given in Table 5, Table 6, Table 7. The nutritional information and water activity of these recipes as well as for Plumpy'nut® are displayed in Table 8, Table 9, Table 10.

Table 5. Recipe of RUTF-1: Rice-Sesame (Collins 2004)
IngredientQuantities (%)
Roasted rice flour20
Soyamin 9029
Roasted sesame seeds paste8
Sunflower oil19.4
Icing sugar22
Vitamin and mineral premix (CMV therapeutique, Nutriset)1.6
Table 6. Recipe of RUTF-2: Barley-Sesame (Collins 2004)
IngredientQuantities (%)
Roasted pearl barley flour15
Soyamin 909
Roased sesame seeds paste27
Sunflower oil24
Icing sugar23.4
Vitamin and mineral premix (CMV therapeutique, Nutriset)1.6
Table 7. Recipe of RUTF-3: Maize-Sesame (Collins 2004)
IngredientQuantities (%)
Roasted maize flour33.4
Roasted sesame seeds paste27
Roasted chick peas flour25
Sunflower oil12
Icing sugar15
Vitamin and mineral premix (CMV therapeutique, Nutriset)1.6
Table 8. Nutritional information of various RUTF recipes (Collins 2004)
  1. *Protein and fat are reported to contribute 11% and 57% in energy input respectively. Total energy is reported to be 530 kcal/100g and moisture < 5%.

    **The energy has been calculated using Atwater factors.

    ***Carbohydrate is by difference assuming protein to be nitrogen multiplied by 6.25.

NutrientUnitRUTF-1 (100 g)Energy (%)RUTF-2 (100 g)Energy (%)RUTF-3 (100 g)Energy (%)Plumpy'nut®* (100 g)Energy (%)
Energy**kCal551 567 512 530 
EnergykJ2307 2373 2142 2218 
Ashg43 3.9 4.9 4 
Moistureg2.9 3.1 2.9 < 5 
Table 9. Mineral content of various RUTF recipes (Collins 2004)
MineralRUTF-1 (mg/kg)RUTF-2 (mg/kg)RUTF-3 (mg/kg)Plumpy'nut® (mg/kg)
Na256.5256.5189.9< 290
Table 10. Water activity in various RUTF recipes (Collins 2004)
RUTF recipeWater activity

How the intervention might work

Adequate energy, protein and micronutrient intake is vital for maintaining a functioning immune system or restoring a dysfunctional one (Naude 2008). Undernutrition is a condition where the body is in great need of nutrients. Infants and young children have a small body size which limits the amount of food that can be given in a feeding (Lin 2008). Lower energy-density foods together with a low frequency of feeding can result in an inadequate intake of energy.

The following characteristics of RUTF may contribute to the possible beneficial effect in the treatment of undernutrition:

  • Balanced, nutritious, home-based therapy;

  • Affordable;

  • Can be eaten safely at home, even where hygienic conditions are poor (WHO 2007);

  • Long shelf life;

  • No special storage (for example, refrigeration) or preparation required.

Some types of RUTF are more energy-dense than others. The amount of calories in the different RUTF products may be an important factor for successful treatment.

Why it is important to do this review

The vast majority of children with undernutrition live in low- and middle-income countries (LMIC). Many of these children are never brought to health care facilities (WHO 2007; Black 2008) due to reasons that may include a lack of money for transport to facilities and/or long distances; parents' lack of health status awareness, and a lack of health care resources to treat thousands of malnourished children in facilities (Kruger 2008). Hospital admission of patients with uncomplicated SAM may not always be desirable as this unnecessarily exposes them to additional risks of nosocomial infections and takes the mother or caretaker away from other children for prolonged periods, which may increase the risk for sibling undernutrition (Collins 2003). The alternative treatment for moderate and severe uncomplicated undernutrition may therefore be a community-focused nutritional intervention, such as RUTF, which does not require specialised health care personnel and expensive equipment (Kruger 2008).

Both the WHO and the United Nations Children's Fund (UNICEF) now recommend the use of RUTF in both health care facilities and in the community as therapeutic feeding for children with SAM (WHO and UNICEF 2009), see Table 11. Although recent clinical trials have evaluated RUTF for the treatment of MAM, there are no official guidelines for the treatment of MAM. It is thus essential that the effects of RUTF for the treatment of undernutrition be systematically evaluated, comparing children with uncomplicated MAM or SAM given RUTF versus F100 in health care facilities and in community settings. The findings of this systematic review will be of significant value to people in LMIC as well as to organisations involved in preparing clinical guidelines for practitioners and policy makers in LMIC (for example, WHO, UNICEF, government departments).

Table 11. SAM management as recommended by the WHO and UNICEF (WHO and UNICEF 2009)
  1. *Children who eat at least 75% of their calculated RUTF ration for the day.

Independent additional criteria
  • No appetite

  • Medical complications

  • Appetite

  • No medical complications

Type of therapeutic feedingFacility-basedCommunity-based
  • F75 (Phase 1)

  • F100/RUTF (Phase 2)

  • And 24 hour medical care

  • RUTF

  • And basic medical care

Discharge criteria (transition criteria from facility to community-based care)
  • Reduced edema

  • Good appetite (with acceptable* intake of RUTF)

  • 15 to 20% weight gain


To assess the effects of RUTF on health outcomes such as recovery rate, relapse during the intervention period, anthropometrical status, weight gain and mortality in children with moderate or severe undernutrition.


Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs), including those defined as quasi-randomised, i.e. trials that use a method of randomisation not strictly speaking random, such as alternation or date of birth. Trials that randomise clusters will also be included.

Types of participants

Children between six months and five years of age who are moderately or severely undernourished, regardless of country, setting or disease status and irrespective of the method of diagnosis.

Types of interventions


  • RUTF as defined by the study authors (either commercially- or non-commercially produced).


  • Alternative RUTF type (for example, corn/soy-based versus peanut-based);

  • Treatment as usual.

Any trial in which the effects of a RUTF are potentially confounded by another intervention will be excluded, i.e. where multiple interventions are involved, comparison groups should receive the same treatment apart from the experimental RUTF.

Types of outcome measures

Primary outcomes
  • Recovery rate as defined by the study authors;

  • Deterioration or relapse during the intervention period as defined by the study authors;

  • Mortality.

Secondary outcomes
  • Average weight gain per kilogram body weight per day during the intervention period;

  • Anthropometrical status at all reported points during and beyond the intervention period (for example, WHZ, weight for age z score (WAZ), height for age z score (HAZ), MUAC);

  • Cognitive function and development after the intervention period (for example, Denver Developmental Screening Test, Bayley Scales of Infant Development);

  • Adverse outcomes as reported by investigators (for example, allergic reactions, refusal of feeds due to poor palatability, diarrhoea).

Search methods for identification of studies

A comprehensive and exhaustive search strategy will be used in an attempt to identify all relevant studies regardless of language or publication status (published, unpublished, in press and in progress).

Electronic searches

A complete search of electronic databases will be conducted to assemble all the relevant studies for assessing the effects of RUTF on undernutrition. The following databases and trials registers will be searched:

  • The Cochrane Central Register of Clinical Trials (CENTRAL), part of The Cochrane Library



  • African Index Medicus 


  • ISI Science Citation Index


  • ZETOC (Conference search)

  • (

  • Current Controlled Trials (

  • WHO International Clinical Trials Registry Platform (ICTRP) (

A detailed search strategy has been developed by the Trial Search Coordinator of the Cochrane Developmental, Psychosocial and Learning Problems Group and will be modified according to the requirements of each database. The MEDLINE search strategy is as follows:

1 nutrition disorders/

2 malnutrition/

3  exp protein-energy malnutrition/

4  wasting syndrome/

5  Emaciation/

6  infant nutrition disorders/

7  child nutrition disorders/

8 deficiency diseases/

9 (undernutrition or under-nutrition).tw.

10 (undernourish$ or under-nourish$).tw.

11 (malnutrition or mal-nutrition).tw.

12 (malnourish$ or mal-nourish$).tw.

13 (nutrition$ adj defic$).tw.

14 marasmus$.tw.


16 emaciat$.tw.

17 (wasted or wasting).tw.

18 (stunted or stunting).tw.

19 or/1-18

20 Food, Fortified/

21 exp Dietary Supplements/

22 (therapeutic adj3 (food$ or diet$)).tw.

23 (enrich$ adj3 (food$ or diet$)).tw.

24 (fortifi$ adj3 (food$ or diet$)).tw.

25 (supplement$ adj3 (food$ or diet$)).tw.

26 (ready adj3 food$).tw.

27 (RUTF or RTUF).tw.

28 or/20-27

29 19 and 28

30 Infant/

31 exp Child/

32 (baby or babies or infant$ or child$ or boy$ or girl$ or toddler$ or preschool$ or pre-school$ or kindergarten$).tw.

33 30 or 31 or 32

34 29 and 33

35 randomized controlled

36 controlled clinical

37 randomi#ed.ab.

38 placebo$.ab.

39 drug therapy.fs.

40 randomly.ab.

41 trial.ab.

42 groups.ab.

43 or/35-42

44 exp animals/ not

45 43 not 44

46 34 and 45

Searching other resources

Targeted researchers, paediatricians and community dieticians will be contacted in order to obtain additional references. The reference lists of included studies and appropriate reviews will also be scrutinised in order to identify relevant studies. Lastly the contact author of each included study will be approached since they may be aware of additional trials (published, unpublished, or ongoing) in the field.

Data collection and analysis

Selection of studies

The authors AS and ML will independently screen the title and abstract of studies identified by the search and apply the pre-specified eligibility criteria in order to identify eligible studies. Where at least one author considers a study to be relevant we will obtain the full text and independently assess it for eligibility. Where there is missing information or clarity is needed we will contact the authors of the primary studies. We plan to resolve any remaining disagreement by consensus among the four authors. We will list studies that we first thought are relevant but which we later exclude in the table 'Characteristics of excluded studies' together with reasons for exclusion.

Data extraction and management

The authors AS and ML will independently extract data using a standardised, pre-piloted extraction form with disagreements being resolved by consensus among the authors. For each study we will collect the following: source (for example, contact details and citation); confirmation of eligibility; methods (for example, ethics approval and study design); participants (for example, age and co-morbidity); interventions (for example, description, dose and duration); outcomes (for example, description and time point(s) collected); results (for example, number of participants randomised per arm and numerical results for pre-specified outcomes); safety (for example, number and description of adverse effects or events per arm), and miscellaneous information (for example, funding source and references to other relevant studies).

Where reported information is unclear or contradictory, or where important data are missing, we will contact the study author(s).

The extracted data will be entered into the following three tables: (1) Characteristics of included studies, (2) Characteristics of studies awaiting classification, and (3) Characteristics of ongoing studies.

Assessment of risk of bias in included studies

The authors AS and ML will independently assess each included study for risk of bias using the guidelines provided in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). The specific criteria appear in Appendix 1. The domains that will be assessed are adequate sequence generation, allocation concealment, blinding, incomplete outcome data, selective reporting and other potential sources of bias. Each included study will be rated 'yes' (low risk of bias), 'no' (high risk of bias) or 'unclear' (uncertain risk of bias) according to each of the six domains. Any disagreement will be discussed with JV.

Cluster-randomised trials will be additionally evaluated according to the following criteria: recruitment bias, baseline imbalance and loss of clusters (Higgins 2008). The specific criteria appear in Appendix 2.

Measures of treatment effect

We will use Review Manager 5 (RevMan) to manage the data and to conduct the analysis (RevMan 2008). We will calculate relative risks (RR) for dichotomous data and mean differences (MD) for continuous data, except where continuous data are reported using different units, in which case we will calculate standardised mean differences (SMD). We will present all results with 95 percent confidence intervals (CI). We will present any highly skewed continuous data in a table.

Unit of analysis issues

Because of the nature of undernutrition we do not expect cross-over trials. For cluster-randomised trials (CRT) we will follow methods for adjusting for clustering as described in the Cochrane Handbook (Higgins 2008). If a CRT has properly accounted for the cluster design it can be included in a meta-analysis by simply using the effect estimate and its standard error (SE) and using the generic inverse variance method in RevMan. If an appropriate analysis has not been performed, the CRT will be incorporated into a meta-analysis (if relevant) by using an 'approximate method'. This entails calculation of an 'effective sample size' for the comparison groups by dividing the original sample size by the 'design effect' which is 1 + (c-1)ICC, where c is the average cluster size and ICC is the intra-cluster correlation coefficient. For dichotomous data both the number of participants and the number experiencing the event will be divided by the same design effect, while for continuous data only the sample size needs to be reduced (means and SDs should be left unchanged). If available we will extract the required information from the articles; otherwise attempt to contact the study authors. If we fail to obtain the required information we will perform sensitivity analyses using a high ICC of 0.1, a moderate ICC of 0.01 and a small ICC of 0.01. Although the values are relatively arbitrary we prefer to use them to adjust the effect estimates and its SEs due to the implausibility that the ICC is actually 0. We will then combine the estimates and their corrected SEs form the CRT with those from parallel group designs using the generic inverse variance method in RevMan.

In the case of multiple intervention groups we will first explore the possibility of creating a single pair-wise comparison (Higgins 2008). If that is not feasible we will select one pair of interventions (depending on the relevance to our systematic review) and exclude the others for analysis purposes (Higgins 2008).

We will report data of all collection time points during and after the intervention period (follow-up). Where data allow we plan to group time points as follows: ≤1 month of RUTF treatment, >1 to ≤ 2 months of RUTF treatment and >2 to 6 months of RUTF treatment.

Dealing with missing data

Important missing data will be classified and presented in a table as (1) pre-randomisation, (2) immediately post-randomisation or as (3) drop-outs during the intervention phase, alongside reasons for the absence. We will then investigate whether an intention-to-treat (ITT) analysis has been performed or not (for example, per-protocol analysis). We will not simply accept the description by the authors in the text, but look at the analysis itself to decide whether the correct procedure for an ITT was followed. We will attempt to obtain essential missing data along with the reason for absence by contacting the study author(s) via email.

Assessment of heterogeneity

The assessment of heterogeneity will be performed by both visual inspection of forest plots and statistically by means of the Chi2 test for heterogeneity (significance level P < 0.1). Heterogeneity will be quantified using the I2 test (Higgins 2002) where I2 values of 50% or more indicate a substantial level of heterogeneity (Higgins 2003).

Assessment of reporting biases

We will assess funnel plots to explore the possibility of small study bias when there are ten or more included studies. Different explanations for funnel plot asymmetry will be considered such as publication bias, the effect of different study sizes and poor study design.

Data synthesis

We anticipate a high degree of heterogeneity due to the inclusion of patients with a variety of conditions related to undernutrition (for example, marasmus, kwashiorkor, stunting, wasting) and various types of RUTF as the intervention (for example, corn/soy-based, peanut-based; and commercially- versus non-commercially produced RUTF). For this reason we will use a random-effects model to combine the results where appropriate. Where substantial statistical heterogeneity exists, study results will not be pooled in a meta-analysis but displayed in a table. Furthermore, studies in which the participants were pre-treated (for example, with F75) before receiving RUTF or the control treatment will not be pooled in a meta-analysis together with other studies.

Subgroup analysis and investigation of heterogeneity

Conducting a large number of subgroup analyses increases the likelihood of false positive results and therefore it is important to carefully select the relevant characteristics to be investigated in advance (Higgins 2008). Although seven characteristics are identified, we feel that all of them are of significant clinical importance:

  • Different types of undernutrition (for example, MAM versus SAM, stunted versus wasted);

  • Health care facility-based RUTF intervention versus community-based RUTF intervention;

  • Different types of RUTF products (for example, corn/soy-based versus peanut-based RUTF, commercially- versus non-commercially produced RUTF);

  • RUTF products of a different energy-density, expressed as kCal/100g (for example, < 520 kCal/100g, 520 to 550 kCal/100g, > 550 kCal/100g);

  • Total energy intake from RUTF per day, expressed in kCal/kg/day;

  • Age group of children: 6 to 12 months, as this is the ideal period to start weaning from a milk-based diet; 13 months to 5 years, as these children consume a mixed diet (mostly not breast milk although the child may still be taking some);

  • Children with or without disease (for example, HIV/AIDS, TB, malaria).

These analyses will be exploratory as they involve non-experimental (cross-over) comparisons and we will treat any conclusion with caution.

Sensitivity analysis

If possible we will perform sensitivity analyses in order to assess the influence of study quality (using adequacy of allocation concealment as a marker) and study design (for example, cluster-randomised controlled trials versus individually randomised controlled trials, trials using random method of allocation versus trials using quasi-random method) on the findings. Depending on the presence of substantial missing outcome data, we will also consider performing a sensitivity analysis to investigate the effect of the missing data on the pooled results.


Appendix 1. Assessment of risk of bias in included RCTs

Domain 1: Sequence generation (Higgins 2008)

Adequate: investigators described a random component in the sequence generation process such as the use of

  • A random number table;

  • Coin tossing;

  • Trowing dice;

  • Shuffling cards or envelopes.

Inadequate: investigators described a non-random component in the sequence generation process such as the use of

  • Odd or even date of birth;

  • The day or date of admission;

  • The hospital or clinic record number;

  • Preference of the participant;

  • The results of a laboratory test or series of tests.

Unclear: there is insufficient information to permit judgement of the way in which sequence generation was performed.

Domain 2: Allocation concealment (Higgins 2008)

Adequate: neither participants nor investigators enrolling participants could foresee assignment due to

  • Central allocation (e.g. via the telephone or pharmacy-controlled);

  • Sequentially numbered drug containers of a matching appearance;

  • Sequentially numbered, opaque and sealed envelopes.

Inadequate: both participants and investigators enrolling participants could foresee upcoming assignment based on, for example

  • Using an open random allocation schedule;

  • Assigned envelopes were unsealed, non-opaque or not numbered appropriately;

  • Date of birth;

  • Case record number.

Unclear: there is insufficient information to permit judgement to the sequence generation process.

Domain 3: Blinding (Higgins 2008)

Adequate: when anyone of the following are applicable

  • No blinding, but the review judge that the outcome would not be influenced by a lack of blinding;

  • Blinding of both the key study personnel and participants are ensured, and it is unlikely that blinding could have been broken;

  • Either participants or some key study personnel were not blinded, but the outcome measurement was blinded and the non-blinding of others are not likely to introduce bias.

Inadequate: when anyone of the following are applicable

  • No blinding or incomplete blinding;

  • Blinding of key study personnel and participants were attempted, but it is likely that the blinding could have been broken;

  • Either key study personnel or participants were not blinded which is likely to introduce bias.

Unclear: there is insufficient information to permit judgement, or the study did not address this outcome at all.

Domain 4: Incomplete outcome data (Higgins 2008)

Adequate: when anyone of the following are applicable

  • No missing outcome data;

  • The reasons for missing outcome data are unlikely to be related to the true outcome;

  • Missing outcome data are balanced in numbers across intervention groups;

  • Missing data have been imputed using appropriate methods;

  • For dichotomous data, the proportion of missing outcomes compared with the observed event risk is not enough to have a clinically relevant impact on the intervention effect estimate;

  • For continuous data, the plausible effect size among missing outcomes is not enough to have a clinically relevant impact on the observed effect size.

Inadequate: when anyone of the following are applicable

  • The reason for missing outcome data is likely to be related to true outcome;

  • The application of simple imputation is potentially inappropriate;

  • 'As-treated' analysis done with substantial departure of the intervention received from that assigned at randomisation;

  • For dichotomous data, the proportion of missing outcomes compared with the observed event risk is enough to introduce clinically relevant bias in the intervention effect estimate;

  • For dichotomous outcome data, the plausible effect size among missing outcomes is enough to induce clinically relevant bias in the observed effect size.

Unclear: there is insufficient reporting of exclusions to permit judgement, or the study did not address this outcome at all.

Domain 5: Selective outcome reporting (Higgins 2008)

Adequate: when anyone of the following are applicable

  • The study protocol is available and all of the pre-specified outcomes are addressed in the review in the pre-specified way;

  • The study protocol is not available, but it is clear that the published reports include all the pre-specified and expected outcomes.

Inadequate: when anyone of the following are applicable

  • Not all of the pre-specified primary outcomes have been reported;

  • One or more of the primary outcomes is reported using measurements of analysis methods that were not pre-specified;

  • One or more reported primary outcomes were not pre-specified;

  • One or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta-analysis;

  • The study report fails to include results for a key outcome that would be expected to have been reported for such a study.

Unclear: there is insufficient information to permit judgement of compliance.

Domain 6: Other potential threats to validity (Higgins 2008)

Adequate: when the study seems to be free of other sources of bias.

Inadequate: when there is the possibility of at least one important risk of bias such as

  • The quality of the specific study design is in question;

  • The study is stopped early due to some data-dependent process;

  • The study has been claimed to have been fraudulent.

Unclear: when there may be a risk of bias, but there is either

  • Insufficient information to assess whether an important risk of bias exists;

  • Insufficient rationale or evidence that an identified problem will introduce bias.

Appendix 2. Additional assessment of risk of bias in included cluster-randomised trials

Domain 1: Recruitment bias

Recruitment bias can occur when individuals are recruited to the trial after the clusters have been randomised (Higgins 2008). The types of participants recruited can be influenced by the knowledge of whether the specific cluster is an intervention or a control cluster.

  • Adequate: when no recruiting was done after randomisation.

  • Inadequate: when additional recruiting was done after randomisation.

  • Unclear: when no reporting was done regarding the timing of recruiting all participants.

Domain 2: Baseline imbalance

Cluster-randomised trials often randomise all clusters at once, therefore a lack of allocation concealment should not usually be a problem (Higgins 2008). However, when there are only a small number of clusters, there is a possibility of chance baseline imbalances between the randomised groups. This may effect either the clusters or the individuals.

  • Adequate: when the baseline comparability of clusters is sufficient, or when statistical adjustment for baseline characteristics occurred (Higgins 2008).

  • Inadequate: when there are significant differences between clusters and no statistical adjustments for baseline characteristics were made accordingly.

  • Unclear: when no reporting was done regarding baseline characteristics, or when it is not clear whether the differences between the clusters were significant.

Domain 3: Loss of clusters

It is possible that complete clusters may be lost from a trial, and have to be omitted from the analysis (Higgins 2008). In the same way as for missing outcome data in individually randomised trials, this may lead to bias in cluster-randomised trials. In addition, missing outcomes for individuals within clusters may also lead to a risk of bias in cluster-randomised trials.

  • Adequate: there were no missing data, or the missing data were addressed in the correct manner.

  • Inadequate: there were missing data and it was dealt with in a way that could have introduced bias.

  • Unclear: when no reporting was done regarding missing data (either complete clusters or individuals within clusters), or when it is unclear whether the authors of the primary study have dealt with the missing data adequately (e.g. acceptable statistical adjustments).


Protocol first published: Issue 2, 2011

Contributions of authors

AS initiated and developed the idea. AS and JV wrote the protocol, while ML and EN provided input.

Declarations of interest

Anel Schoonees - none known

Martani Lombard - none known

Etienne Nel - none known

Jimmy Volmink - none known

Sources of support

Internal sources

  • Stellenbosch University, South Africa.

External sources

  • No sources of support supplied