Preamble

Severe malnutrition in children who are 6–59 months of age was defined in previous publications (2) as weight-for-height (or length) less than −3 Z-score, or less than 70% of the median National Center for Health Statistics (NHCS)/WHO reference values, or the presence of oedematous malnutrition. The manual recommended admitting children with severe acute malnutrition to hospital for initial treatment and rehabilitation, and to continue treatment as outpatients (transfer to a nutritional rehabilitation centre) when children have completed the initial phase of the treatment, have no complications, and are eating satisfactorily and gaining weight (2). Since ready-to-use therapeutic food became available in the 1990s, reports have demonstrated that most children with severe acute malnutrition can be treated safely without admission to hospital (19). Consequently, in 2007 a joint United Nations statement endorsed outpatient care of children who are 6–59 months of age with severe acute malnutrition and who present with no medical complications and with good appetite (10). The same statement also endorsed the use of mid-upper arm circumference measurements as an independent criterion for screening.

The transition from the NCHS growth reference to the WHO growth standards in 2006 (4) prompted the revision of cut-off values for indicators of severe acute malnutrition. A 2009 joint United Nations statement endorsed a low mid-upper arm circumference less than 115 mm as a criterion for diagnosing severe acute malnutrition in children, in light of the high predictive value for mortality (9). Mid-upper arm circumference is measured using simple arm bands that are marked in millimetre graduations and are sometimes colour coded. These armbands can be used and interpreted by health-care workers who receive appropriate training. The statement also noted the programmatic advantage of using a single mid-upper arm circumference cut-off value to identify children with severe acute malnutrition in this age group (9). However, mid-upper arm circumference and weight-for-height indicators do not always correlate when used to identify children with severe acute malnutrition (20–22). About 40% of children who are classified as having severe acute malnutrition using one of these indicators are similarly classified using the other indicator (8). The agreement between these criteria varies considerably between settings and geographic location (9, 23). Comparison of the sensitivity–specificity curves (receiver operating characteristic curves) in community studies shows that mid-upper arm circumference is better at identifying children with a high risk of death (24).

The 1999 recommendations included discharging children from hospital care when they achieved a weight-for-height ≥−1 Z-score or ≥90% of the median NCHS/WHO reference values (2). The 2009 joint United Nations statement proposed using a single discharge criterion of 15% (or 20%, depending on the local context) weight gain over oedema-free weight on enrolment, for children admitted based on weight-for-height or mid-upper arm circumference, as well as absence of oedema for 2 weeks (9). Data from children treated for severe acute malnutrition in outpatient care in Malawi and Ethiopia have suggested that 15% weight gain would result in 50% of children with severe acute malnutrition meeting or exceeding 80% of the median weight-for-height of the NCHS reference (25). However, while some programmes adopted this approach, there was concern about its validity as an indicator of nutritional recovery, and in many settings only the mid-upper arm circumference cut-off value of ≥125 mm was applied (26). The discharge cut-off value for mid-upper arm circumference of ≥125 mm was based on historical cohort studies from Bangladesh, Malawi and Uganda that suggested that mortality risk at this cut-off value did not exceed 1/10 000 per day (25). The safety of using changes in mid-upper arm circumference as an indicator of progress of recovery during nutritional rehabilitation, and a single cut-off value to indicate “recovery”, has not been validated.

In light of these experiences, WHO, with support from the guideline development group aimed to provide guidance on the following:

• Admission and discharge criteria for children who are 6–59 months of age with severe acute malnutrition:

  —

  Admission cut-off values for the respective screening indicators

  —

  Discharge cut-off values for the different admission indicators

  —

  Admission criteria for inpatient care and outpatient care

  —

  Transition from inpatient care to outpatient care after stabilization

Summary of the evidence

A systematic review was conducted to examine admission and discharge criteria for severe acute malnutrition in children who are 6–59 months of age (27). The search identified 11 relevant epidemiological studies. Three of the 11 studies used cut-off values for admission that did not correspond to the WHO definition of severe acute malnutrition, namely mid-upper arm circumference <120 mm (28), mid-upper arm circumference <130 mm (29) and weight-for-height <−2 Z-score (30). However, these three studies reported stratified results from which outcomes of children with severe acute malnutrition could be extracted. All studies, except one, were conducted in African countries and five out of the 11 studies took place in the same setting in Malawi where there was a high proportion of children with oedema. The majority of studies involved uncomplicated cases of severe acute malnutrition that were managed in an outpatient programme. Five of the studies used mid-upper arm circumference as an inclusion criterion, and the other seven used weight-for-height for enrolling children, but reported on gain in mid-upper arm circumference during nutritional rehabilitation.

Low values of weight-for-height and mid-upper arm circumference both identify children with an increased risk of mortality, but the children identified by each indicator may be different. In one study including 34 937 children aged 6–59 months from Afghanistan, Angola, Burkina Faso, Burundi, Chad, Ethiopia, Malawi, Sierra Leone, Niger and India, the criterion of mid-upper arm circumference <115 mm did not detect severe acute malnutrition in up to 75% of children who were identified as having severe acute malnutrition defined by weight-for-height <−3 Z-score in the WHO growth standards (31). In two smaller studies, this proportion was around 40% (32, 33). The proportion of cases identified by a low weight-for-height that were undetected by a low mid-upper arm circumference was higher in males than in females and increased with age.

No randomized studies were identified that compared the outcomes of nutritional rehabilitation of children with severe acute malnutrition admitted on the basis of mid-upper arm circumference versus weight-for-height Z-score. Only one study compared the mortality risk of hospitalized children according to their mid-upper arm circumference versus their weight-for-height Z-score (32). The mortality risk was similar for children presenting with a weight-for-height ≤−3 Z-score or with a mid-upper arm circumference ≤115 mm, i.e. 10.1% and 10.9%, respectively. The highest risk for mortality (25.4%) was observed in children who had both weight-for-height ≤−3 Z-score or a mid-upper arm circumference ≤115 mm. Children with a mid-upper arm circumference ≤115 mm presented more frequently with signs of recent or current oedematous malnutrition, stunting and subcostal indrawing, and were more frequently girls and of younger age than those admitted with a weight-for-height ≤−3 Z-score (32). Bipedal oedema was present in 38.0% of children with a low mid-upper arm circumference versus 13.9% in those with a weight-for-height ≤−3 Z-score.

Four studies, including an unpublished report, reported on outcomes of children diagnosed on the basis of mid-upper arm circumference only and managed in outpatient care without a comparison group admitted on the basis of weight-for-height (25, 28, 29, 33). The mortality risk for children with severe acute malnutrition was reported in three of the studies and was overall relatively low (≤2.1 %). The median recovery times ranged from 44.4 ± 29.7 days (33) to 50.5 ± 25.8 days (28). The daily gain of mid-upper arm circumference ranged from 0.17 ± 0.16 mm in children admitted with a mid-upper arm circumference <110 mm and treated in a supplementary feeding programme (25) to 0.51 ± 0.3 mm in Burkina Faso in children with a mid-upper arm circumference ≤110 mm and receiving ready-to-use therapeutic food (28). Two of the studies stratified the results by level of mid-upper arm circumference at admission (28, 29). In both studies, children admitted with a lower mid-upper arm circumference displayed a greater daily gain in weight and mid-upper arm circumference. Furthermore, in Burkina Faso, mortality was greater among children with a lower mid-upper arm circumference at admission and referral for inpatient care was more common. This latter study also reported that the daily gain in mid-upper arm circumference was higher in older and taller children, and in boys (28).

Six additional studies also reported on gain in mid-upper arm circumference during nutritional rehabilitation, but admission was based on a weight-for-height Z-score of <−3. The daily gain in mid-upper arm circumference was in the same range as in studies that admitted children on the basis of a low mid-upper arm circumference only. Overall, mid-upper arm circumference increased by 0.2 to 0.4 mm/day during rehabilitation, with no obvious differences among studies admitting children on the basis of mid-upper arm circumference or weight-for-height. In all studies, the increase in mid-upper arm circumference paralleled the daily weight gain, which ranged between 3.0 and 6.5 g/kg.

Two observational studies reported outcomes when mid-upper arm circumference was used as a discharge criterion for malnourished children from nutritional rehabilitation programmes (28, 29). In Burkina Faso, the time to discharge of children with mid-upper arm circumference <120 mm from an outpatient programme (n = 5689) was reported for the period April to December 2008 when discharge was based on 15% weight gain, and then for the period April to December 2009 when discharge was based on achieving mid-upper arm circumference ≥124 mm (without consideration of weight gain). When 15% weight gain was used as the discharge criterion, the mean time to discharge of children was 53 ± 25 days; in the later period when discharge was based on achieving a mid-upper arm circumference ≥124 mm, children were discharged after an average duration of 36 ± 20 days. In children whose mid-upper arm circumference was initially ≤114 mm, the average duration of treatment to achieve at least 15% weight gain was 48 ± 23 days, whereas for children whose mid-upper arm circumference was initially between 115 mm and 119 mm, the average duration of treatment was 55 ± 26 days. In contrast, when mid-upper arm circumference ≥124 mm was used as the discharge criterion, the average duration of treatment for the more malnourished children (mid-upper arm circumference initially ≤114 mm) was 47 ± 25 days, which was longer than for the less malnourished children (initial mid-upper arm circumference between 115 mm and 119 mm) who were discharged on average after 33 ± 16 days. These observations provide indirect evidence that 15% weight gain is not an appropriate discharge criterion, because it results in the more severely malnourished children getting the shortest duration of treatment and being discharged when still malnourished. In this study, the mortality risks up until the end of treatment when using 15% weight gain versus mid-upper arm circumference ≥124 mm as a discharge criterion were comparable and were 1.5 %. The second study in Guinea-Bissau used a mid-upper arm circumference ≥130 mm as a discharge criterion but did not separate outcomes for children who were admitted with severe acute malnutrition from those with moderate acute malnutrition (29).

No published data were found that reported the final outcome status of children with severe acute malnutrition following discharge from treatment programmes, e.g. relapse rates; subsequent mortality when discharged was based on a mid-upper arm circumference measurement compared to other indicators such as weight-for-height Z-score, mid-upper arm circumference-for-age Z-score, or mid-upper arm circumference-for-height Z-score.

The overall quality of evidence was rated as low for the relationship between anthropometry and mortality risk on admission. There was no evidence available informing the relationship between anthropometry and the risk of mortality post discharge. The quality of evidence for other important outcomes, including time to recovery, growth and daily weight gain was rated very low or low, owing to methodological and reporting issues. A Grading of Recommendations Assessment, Development and Evaluation (GRADE) summary of the evidence was not developed, as the published data were epidemiological reports rather than comparative studies.

Only one of the studies reported outcomes in children identified, monitored or discharged based on mid-upper arm circumference and also by weight-for-height, and this was carried out among hospitalized children. Few of the studies addressed the precision and accuracy of anthropometric measurements, which may affect reliability. The fact that most of the studies were done in Malawi, where the majority of cases are oedematous and the prevalence of HIV infection is high, may affect the external validity (30, 34–38). No studies were retrieved reporting on a number of important outcomes, i.e. costs, adverse effects and population coverage.

Discussion

• The guideline group noted that the development of ready-to-use therapeutic food has made outpatient management of children with severe acute malnutrition more feasible and safer. Active community screening makes it possible to identify and treat these children early as outpatients. It also allows children to be appropriately managed while avoiding the risk and problems of inpatient care, such as nosocomial infections, patient transport and costs and disruptions to families. Screening in the community has been facilitated by the feasibility of using mid-upper arm circumference to identify severely malnourished children.
• Depending on the age group, and excluding consideration of oedematous malnutrition, anthropometric assessments based on weight-for-height using the WHO growth standards (8) are likely to identify a larger population with severe acute malnutrition than NCHS growth reference values. This has a major implication for countries in South and South-East Asia, where large numbers of infants and young children would be defined as having severe acute malnutrition, using the WHO growth standards.
• The guideline group noted that while weight-for-height and mid-upper arm circumference are good methods to assess wasting, both require training and methodical implementation to achieve reliable assessments. Even after training, significant variation has been noted in healthcare workers' measurements and classification during standardization tests.
• The guideline group agreed that children with severe acute malnutrition identified using weight-for-height do not consistently have a mid-upper arm circumference <115 mm, and vice versa. The group discussed the importance of each indicator for screening and admission of children for treatment in different settings. The evidence shows that the two indicators do not always identify the same populations.
• It is doubtful that health-care workers could detect the small incremental changes of between 1.4 mm and 2.8 mm per week of mid-upper arm circumference during the first month of treatment for monitoring recovery (39).
• There is limited evidence on the most appropriate cut-off value for mid-upper arm circumference to indicate nutritional recovery. The 2006 WHO growth standards (4) indicate that a mid-upper arm circumference of 125 mm corresponds to a mid-upper arm circumference-for-age Z-score <−3 for males older than 33 months and for females older than 44 months. So using this single cut-off value may result in children who still have severe acute malnutrition being discharged, especially older male children. At younger ages, this problem is less significant, though a proportion of children may be discharged while still being moderately malnourished. Adopting a higher cut-off value for mid-upper arm circumference would avoid this problem, though it would result in children being kept in treatment for significantly longer. Using mid-upper arm circumference-for-age or mid-upper arm circumference-for-height Z-scores ≥−2 would allow discharge of children who are no longer malnourished, and avoid inappropriate treatment of young children.
• The guideline group noted that simple and uniform cut-off values are programmatically helpful, but that the evidence suggests that the sensitivity and specificity of measurements of mid-upper arm circumference compared to weight-for-height differ significantly at both upper and lower age (and/or height) ranges and by sex. Examining the clinical relevance and programmatic feasibility of different cut-off values for mid-upper arm circumference for these age (and/or height) groups is an important area for future research, possibly by re-analysing existing data sets.
• The simplicity of mid-upper arm circumference and its superior correlation to risk of death makes it an attractive diagnostic tool for severe acute malnutrition. However, a single cut-off value for mid-upper arm circumference for admission and discharge of all children aged 6–59 months may not adequately reflect the variability and associated mortality risks of all children in this age range.
• The 1999 WHO manual does not refer to “visible severe wasting” as a diagnostic criterion for severe malnutrition, although visible severe wasting is included in the Integrated management of childhood illness (40, 41) and versions of the WHO Pocket handbook (42) and the 2007 joint United Nations Statement (10). Evidence does not support using visible severe wasting as a stand-alone criterion for children who are less than 5 years of age. However, the guideline group endorsed the principle that a trained clinician should always undress children identified with severe acute malnutrition and examine them naked to identify other medical complications.
• The guideline group noted that triaging of children with severe acute malnutrition to treatment as outpatients or inpatients should be guided primarily by the children's clinical condition, including appetite, and social circumstances, such as whether children are disabled or there are other mitigating circumstances, including significant social or access issues.
• The guideline group noted that the decision to transfer children from inpatient care to outpatient care (i.e. discharge from hospital) after the initial phase of stabilization should be guided primarily by the children's clinical condition, including appetite and response to treatment, and also social circumstances.
• The guideline group noted that percentage weight gain should no longer be used as a criterion for discharge from treatment.
• The guideline group noted that correlations of anthropometric indicators with risk of death and response to treatment are urgently needed to inform clinical recommendations and guidelines for managing severe acute malnutrition in infants less than 6 months old and in children 5 years and older, especially in the context of HIV.

Relevant standing recommendations

The recommendations below were confirmed as current in the context of existing WHO recommendations included in WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 (2); joint United Nations statement on Community-based management of severe acute malnutrition, 2007 (10); WHO child growth standards and the identification of severe acute malnutrition in infants and children: a joint statement by the World Health Organization and the United Nations Children's Fund, 2009 (9); and WHO Integrated management of childhood illness: caring for newborns and children in the community, 2011 (IMCI) (41), noting in particular that:

• in children who are 6–59 months of age, severe acute malnutrition is defined as:

  —

  weight-for-height ≤−3 Z-score, or

  —

  mid-upper-arm circumference <115 mm, or

  —

  presence of bilateral oedema;

• visible severe wasting is not included as a diagnostic criterion (2). However, all malnourished children should be clinically examined when undressed, as part of routine management;
• all anthropometric indicators are assumed to be derived from the WHO growth standards (8);
• children with severe acute malnutrition with medical complications or failed appetite test¹ should be admitted to hospital for inpatient care;
• admission may also be warranted if there are significant mitigating circumstances such as disability or social issues, or there are difficulties with access to care;
• children with severe acute malnutrition and without these signs or mitigating circumstances can be managed as outpatients by providing appropriate amounts of ready-to-use therapeutic food.

Recommendations

Criteria for identifying children with severe acute malnutrition for treatment

1.1.

In order to achieve early identification of children with severe acute malnutrition in the community, trained community health workers and community members should measure the mid-upper arm circumference of infants and children who are 6–59 months of age and examine them for bilateral pitting oedema. Infants and children who are 6–59 months of age and have a mid-upper arm circumference <115 mm, or who have any degree of bilateral oedema should be immediately referred for full assessment at a treatment centre for the management of severe acute malnutrition.

strong recommendation, low quality evidence

1.2.

In primary health-care facilities and hospitals, health-care workers should assess the mid-upper arm circumference or the weight-for-height/weight-for-length status of infants and children who are 6–59 months of age and also examine them for bilateral oedema. Infants and children who are 6–59 months of age and have a mid-upper arm circumference <115 mm or a weight-for-height/length <−3 Z-scores of the WHO growth standards (8), or have bilateral oedema, should be immediately admitted to a programme for the management of severe acute malnutrition.

strong recommendation, low quality evidence

Criteria for inpatient or outpatient care²

1.3.

Children who are identified as having severe acute malnutrition should first be assessed with a full clinical examination to confirm whether they have medical complications and whether they have an appetite. Children who have appetite (pass the appetite test) and are clinically well and alert should be treated as outpatients. Children who have medical complications, severe oedema (+++), or poor appetite (fail the appetite test) or present with one or more IMCI danger signs³ should be treated as inpatients.

strong recommendation, low quality evidence

Criteria for transferring children from inpatient to outpatient care²

1.4.

Children with severe acute malnutrition who are admitted to hospital can be transferred to outpatient care when their medical complications, including oedema, are resolving and they have good appetite, and are clinically well and alert. The decision to transfer children from inpatient to outpatient care should be determined by their clinical condition and not on the basis of specific anthropometric outcomes such as a specific mid-upper arm circumference or weight-for-height/length.

strong recommendation, low quality evidence

Criteria for discharging children from treatment

1.5.

1. Children with severe acute malnutrition should only be discharged from treatment when their:

   —

   weight-for-height/length is ≥−2 Z-score and they have had no oedema for at least 2 weeks, or

   —

   mid-upper-arm circumference is ≥125 mm and they have had no oedema for at least 2 weeks.

2. The anthropometric indicator that is used to confirm severe acute malnutrition should also be used to assess whether a child has reached nutritional recovery, i.e. if mid-upper arm circumference is used to identify that a child has severe acute malnutrition, then mid-upper arm circumference should be used to assess and confirm nutritional recovery. Similarly, if weight-for-height is used to identify that a child has severe acute malnutrition, then weight-for-height should be used to assess and confirm nutritional recovery.
3. Children admitted with only bilateral pitting oedema should be discharged from treatment based on whichever anthropometric indicator, mid-upper arm circumference or weight-for-height is routinely used in programmes.
4. Percentage weight gain should not be used as a discharge criterion.

strong recommendation, low quality evidence

Follow-up of infants and children after discharge from treatment for severe acute malnutrition

1.6.

Children with severe acute malnutrition who are discharged from treatment programmes should be periodically monitored to avoid a relapse.

strong recommendation, low quality evidence

Rationale

The guideline development group noted that the quality of evidence was low for the relationship between anthropometry and mortality risk on admission and very low/low for the relationship between anthropometry and time to recovery, growth and weight gain. However, the guideline group agreed that the recommendation should be strong. The group recognized that the associations between anthropometric outcomes and high mortality risks originated largely from observational or population-based studies rather randomized trials. Other evidence informing the recommendations was also based on programmatic experience and this characteristically downgrades the quality of evidence. However, it was felt that examining these issues through randomized trials would be very difficult and possibly unethical. The guideline group also considered that there is high biological plausibility for the associations between poor anthropometry and mortality. The benefits of identifying and treating children with severe acute malnutrition through screening with mid-upper arm circumference were determined to far outweigh the possible disadvantage of overdiagnosing malnutrition and causing children to be further assessed. The recommendations were considered to be feasible and sustainable in low-resource settings, even though there were no cost data available. As has been reported in a limited number of settings, community screening does increase the rate of identification of children with severe acute malnutrition. This will result in increased costs to the health system as treatment, either as inpatients or in the community. However, it is also likely that earlier identification will enable children to be successfully treated at home without the costs of hospitalization or complications associated with later presentation, or through nosocomial infection. Community-based management will also probably reduce transportation costs to the family and time away from home. The guideline group viewed the other recommendations in a similar way and that their overall benefits far outweighed any potential harm. However, there are several important research questions that need to be addressed to give greater certainty to these recommendations.

Implications for future research

Discussion with guideline development group members and stakeholders highlighted the limited evidence available in themes related to the priority areas listed next.

• To refine cut-off values of mid-upper arm circumference to identify severe acute malnutrition in children who are 6–11 months, 12–23 months and 24–59 months of age, through assessment of treatment outcomes.
• To test strategies to improve active community screening and routine health-facility screening, and investigate barriers to service access and uptake, to enhance treatment coverage.

Other issues (no specific order)

• To evaluate the validity of mid-upper-arm circumference values versus weight-for-height Z-score as discharge criteria for end of treatment (in relation to response to treatment, relapse and mortality) and determine the appropriate cut-off values.
• To assess the sensitivity and specificity of mid-upper-arm circumference measurements at the lower and higher age ranges of children who are 6–59 months of age, controlled for stunting and the presence of oedema.
• To establish mid-upper-arm circumference thresholds to identify severe acute malnutrition in infants who are less than 6 months of age and children who are 5 years of age and older.
• To assess the response to treatment according to initial anthropometric criteria and clinical and biochemical characteristics.
• To assess the correlation of anthropometric indicators with risk of death and response to treatment of severe acute malnutrition in infants who are less than 6 months of age and children who are 5 years of age and older, especially in the context of high and low HIV prevalence.

Preamble

The standard recommendation for the management of severe acute malnutrition has been inpatient care during the stabilization phase and the first part of the rehabilitation phase, and then transfer to a nutrition rehabilitation centre (when they exist) for the rest of the rehabilitation phase (42). However, inpatient treatment is resource intensive (43), disrupts the family and places children at risk of nosocomial infections. Reports of children with severe acute malnutrition being effectively treated at home (34, 35, 44) have prompted categorization of severe acute malnutrition as being with or without medical complications (43, 45). Cases are uncomplicated if there are no signs of severe clinical illness and the child has a good appetite. Using this classification, WHO, UNICEF and the World Food Programme recommend treating children with uncomplicated severe acute malnutrition as outpatients whenever possible (10). Children with severe acute malnutrition who have severe bilateral oedema (+++) have an increased risk of mortality compared to children with severe acute malnutrition but with lesser degrees of oedema. For this reason, it is unclear whether the severity of oedema, in the absence of other complications and with adequate appetite, should independently influence where to treat children with severe acute malnutrition.

The guideline development group aimed to review the evidence and provide guidance on the following question:

• Which children with severe acute malnutrition who also have oedema should be managed in hospital compared to at home?

Summary of the evidence

A systematic review examined the evidence on the effectiveness of managing children who are more than 6 months of age with uncomplicated oedematous malnutrition grade +/++,⁴ in ambulatory settings (46). No randomized controlled trials comparing the outcomes achieved in inpatient and outpatient care for this specific group were identified. However, eight reports were found that described outcomes of children with severe acute malnutrition who had bilateral oedema and who were treated as outpatients. The data were from case-series that did not have control groups and so comparisons were not therefore possible.

Among uncomplicated cases of oedematous malnourished children, recovery rates exceeded 88% and case-fatality rates remained below 4%. Oedema classification varied among the studies and in some studies it was not clear whether children with severe oedema were excluded from outpatient treatment. These outcomes are consistent with those recommended in the Sphere guidelines (47) and the Prudhon index for case-fatality rates (48). However, complicated cases were associated with high case-fatality rates during the first 4 days of inpatient stabilization, especially in the context of HIV infection (49). The weight gain of children with oedema who were treated as outpatients was reported in only two studies and the time to nutritional recovery was prolonged and did not meet the minimum Sphere standards (47). Authors suggested that this was due to sharing of ready-to-use therapeutic food at home or absence of food supplementation in the home-based treatment group (51). Methods to assess weight gain were not, however, uniform across the studies. No association was reported between the severity of oedema and the likelihood of nutritional recovery or not.

The quality of the evidence was rated as very low, as the absence of a control group in the reports of outpatient treatment of children with severe acute malnutrition and oedema make it difficult to make conclusions about the effectiveness and safety of outpatient treatment for children with oedema +/++. As a result, no GRADE table for this evidence base was developed.

Discussion

• The guideline development group considered that the systematic review supported the general principle that severely malnourished children with oedema, but no other medical complications, can be effectively and safely treated as outpatients.
• It was also noted that the severity of oedema reported in children is influenced by the familiarity of the staff with seeing oedematous cases. Inexperienced staff have been observed to overestimate the extent of oedema.
• While data were not available to comment on the severity of oedema and the safety and effectiveness of managing children with severe malnutrition and oedema at home, the association of increased mortality risk with more severe oedema (independent of site of care) was noted by the guideline group.
• The guideline group considered that evidence from inpatient settings supported managing children with severe malnutrition who have oedema +++ as inpatients and those with oedema +/++ as outpatients.
• The guideline group noted, however, that caution is required in interpretation of these results, as there are limitations in the quality of the design of the included studies.

Relevant standing recommendations

The recommendations were confirmed as current in the context of existing WHO recommendations included in the WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 (2) and Joint United Nations statement on Community-based management of severe acute malnutrition, 2007 (10), noting in particular that:

• infants and children who have only + or ++ bilateral pitting oedema but present with medical complications or have no appetite should be admitted for inpatient care;
• infants above 6 months of age and children who have + or ++ bilateral pitting oedema but who have no medical complications and have appetite should be managed as outpatients;
• children with bilateral pitting oedema who are admitted for inpatient care should transition from the stabilization to the transition phase when appetite returns and oedema is reducing.

Recommendations

2.1.

Children with severe acute malnutrition who have severe bilateral oedema +++,⁵ even if they present with no medical complications and have appetite, should be admitted for inpatient care.

strong recommendation, very low quality evidence

Rationale

The guideline group noted that the quality of evidence was very low but was in consensus that the recommendation should be strong. The recommendation essentially makes explicit what is presently implicit in existing WHO recommendations. The recommendation is therefore not new but clarifies where children with severe bilateral oedema should be treated, i.e. in hospital. The recommendation prioritizes the safety of children with this complication. The recommendation is feasible, as, in previous guidelines, all children with severe acute malnutrition were admitted. Hence, this recommendation places no additional cost burden on health systems than previously.

Implications for future research

Discussion with guideline development group members and stakeholders highlighted the limited evidence available in themes related to the priority areas listed next.

• What is the predictive value of different degrees of oedema (+, ++ or +++) on recovery of children with severe acute malnutrition managed as inpatients or outpatients?

Preamble

Epidemiological and clinical studies have documented the high prevalence of pneumonia, bacteraemia and urinary tract infections in children with severe acute malnutrition, which often may not be symptomatic.

The WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 recommended that all children with severe acute malnutrition should be admitted to hospital and treated with a course of antibiotics (2). If clinical complications were present, then parenteral antibiotics were recommended, depending on local resistance patterns and availability. For children without obvious clinical sepsis, oral antibiotics were recommended. Because severe acute malnutrition suppresses the immune response, it is hard to detect infection.

The availability of ready-to-use-therapeutic foods now allows large numbers of children who have severe acute malnutrition and are over 6 months of age, but who do not have medical complications, to be treated as outpatients. These children were previously treated in hospital because ready-to-use therapeutic foods that could be safely used in the community were not available. The 2007 joint United Nations statement on Community-based management of severe acute malnutrition (10) recommends that, in addition to the provision of ready-to-use therapeutic food, children with severe acute malnutrition receive a short course of basic oral medication to treat infections.

In 2007, however, the evidence to support the use of oral antibiotics in children without signs of clinical sepsis was limited. The emergence of successful outpatient management of children who have severe acute malnutrition but do not demonstrate clinical complications, has raised the question of whether these children would benefit from a course of oral antibiotics as part of routine management – in line with WHO's 1999 recommendation for treatment in hospitals (2).

The choice of antibiotic, for either inpatient or outpatient management, should consider patterns of antibiotic resistance, which vary between countries (52, 53); high rates of resistance to second-line antibiotics such as chloramphenicol, gentamicin and cephalosporins have been reported (52), while sensitivity to ciprofloxacin remains generally high. In vitro resistance, however, does not necessarily reflect clinical efficacy.

The guideline development group aimed to update guidance on the following question:

• Do children with uncomplicated severe acute malnutrition need to be treated with antibiotics and, if so, then which antibiotic should be used?

Summary of the evidence

A systematic review was conducted to examine the clinical safety and efficacy of antibiotic interventions among children with uncomplicated severe acute malnutrition (54).

One study reported outcomes in hospitalized children with severe acute malnutrition who all received antibiotics. In a randomized controlled trial (2002–2003), there were no significant differences in any of the efficacy outcomes among Sudanese children with severe acute malnutrition (complicated and uncomplicated) who were treated with either 5 days of oral amoxicillin or 2 days of intramuscular (IM) ceftriaxone (55). Given that the study population was restricted to hospitalized children and there was no control group, this study did not provide direct evidence on the safety and efficacy of antibiotics in children with uncomplicated severe acute malnutrition.

Two studies provided evidence on the efficacy of antibiotics in children managed as outpatients. In a retrospective analysis of children in Malawi with uncomplicated severe acute malnutrition managed as outpatients, outcomes in one cohort who received 7 days of oral amoxicillin in addition to ready-to-use therapeutic food were compared with those of a cohort who received ready-to-use therapeutic food only. At 12 weeks' follow-up, mortality rates were less than 5% in both cohorts and rates of nutritional recovery were similar (56). However, these two cohorts of children were quite different at baseline and therefore the results are difficult to interpret.

In a prospective, randomized, double-blind trial in southern Malawi, children with uncomplicated severe acute malnutrition managed as outpatients were randomized to receive either 1 week of amoxicillin 80–90 mg/kg/day, cefdinir 14 mg/kg/day or placebo (57). The rate of nutritional recovery of children who received antibiotics was significantly better than that of those who received placebo (amoxicillin 89.0%, cefdinir 91.4%, placebo 85.1%, P < 0.0002). This result was consistent among both children with bilateral oedema (amoxicillin 91.6%, cefdinir 92.9%, placebo 88.3%, P < 0.007) and those with severe wasting (amoxicillin78.9%, cefdinir 85.1%, placebo 73.7%, P < 0.03). Mortality rates were significantly lower in the groups that received antibiotics (amoxicillin 4.63%, cefdinir 3.86%, placebo 7.48%, P < 0.002). When analysed by type of malnutrition, mortality was reduced in children who had bilateral oedema (amoxicillin 3.49%, cefdinir 3.56%, placebo 7.48%, P < 0.0006). However, the study was not specifically powered to undertake the latter analysis.

One before-and-after study of ampicillin and gentamicin as second-line antibiotics was also identified. The design of the study was such that it was not possible to estimate the relative effect of the introduction of the antibiotic regimen independently of an algorithm for managing hypoglycaemia that was introduced in parallel (58).

No studies of ceftriaxone, ciprofloxacin or co-trimoxazole as a first-line treatment for children with severe acute malnutrition were identified.

The overall quality of this evidence was rated as low, owing to methodological and reporting bias. See Annex 1, Table 1 for details of the GRADE assessment of the evidence.

Discussion

• The guideline group considered that the use of broad-spectrum antibiotics such as amoxicillin for children with severe acute malnutrition is supported by epidemiological data demonstrating a high prevalence of infections in these children, including children with uncomplicated severe acute malnutrition receiving outpatient management.
• It is possible that providing antibiotics will result in some adverse events such as diarrhoea, skin reactions and hypersensitivities. However, it is unlikely that the frequency of such complications will be significantly more than in well-nourished children.
• However, it was noted that the one randomized trial that specifically examined outcomes in this population was in a setting where the pattern of severe acute malnutrition is predominantly oedematous and HIV is also prevalent.
• Increased antimicrobial resistance is a potential consequence of prescribing broad-spectrum antibiotics for the outpatient treatment of children with severe acute malnutrition.

  —

  However, children with severe acute malnutrition are at high risk of mortality – higher for example, than infants with fast breathing, where antibiotics are also currently recommended.

  —

  The reported reduction in mortality in children with uncomplicated severe acute malnutrition treated with antibiotics as outpatients (7.48% to 4.63%) is significant and the guideline group thought it was justified to recommend this intervention to this population, even at the risk of increasing antimicrobial resistance in the community.

  —

  Community-level antimicrobial resistance is not a certain consequence of providing antibiotics to children as part of outpatient care but will depend on factors that include the prevalence of uncomplicated severe acute malnutrition, the quality of health-care worker counselling and adherence to medications at home.

  —

  Antimicrobial resistance of community-acquired infections may be already present and the consequence of antibiotic use for other reasons.

  —

  Research is needed to examine these issues.

• It was noted that small bowel bacterial overgrowth is common in children with severe acute malnutrition and affects intestinal function.
• The mechanism by which oral antibiotics may benefit children with uncomplicated severe malnutrition is unclear but may include reducing excessive proliferation of small bowel flora, translocation of gut bacteria into the bloodstream, or modification of the microbiome.
• The cost of providing oral antibiotics to all children with uncomplicated severe acute malnutrition as outpatients would be about US$0.45 to US$0.60 per child for 5 days' treatment (59). WHO guidelines previously recommended that all children with severe acute malnutrition should be admitted and all children given appropriate antibiotics (2). Hence, while guidelines have changed to provide treatment for these children in outpatient care, treating all children with antibiotics has been the standard of care. While the budget to pay for antibiotics may be different between hospital and primary health-care facilities, the total cost to the national health authority would theoretically not be different. It is likely that all children in hospital are indeed given antibiotics and hence this cost is real. Reports from programmes indicate that the introduction of community-based screening has increased the identification of children with severe acute malnutrition – both those requiring hospital care and those that can be managed as outpatients. The funds required for antibiotics have therefore increased because of community-based screening improving coverage and service uptake. Managing children as outpatients is likely to be cheaper than hospital care, even if oral antibiotics are included as part of standard care. No cost analysis of these approaches was available.
• While providing antibiotics would increase the transaction time between health-care workers and mothers/caregivers of children, and highlights the importance of reliable supply-chain management, the potential reduction in mortality justifies these opportunity costs.
• WHO endorses the principle that antimicrobials should only be given to patients when infection is present. In the context of severe acute malnutrition, “uncomplicated” should not be understood to mean that infection is not present. Rather “uncomplicated” implies the absence of symptoms for infection and that the child can be otherwise managed as an outpatient.
• The guideline group noted that research is urgently needed to examine outcomes in populations where wasting predominates, such as in South-East Asia, and where HIV is less common.

Relevant standing recommendations

The recommendations were confirmed as current in the context of existing WHO recommendations included in the WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 (2), noting in particular that:

• children admitted with severe acute malnutrition and complications such as septic shock, hypoglycaemia, hypothermia, skin infections, or respiratory or urinary tract infections, or who appear lethargic or sickly, should be given parenteral (IM or intravenous [IV]) antibiotics;
• children admitted with severe acute malnutrition and with no apparent signs of infection and no complications should be given an oral antibiotic.

Recommendations

3.1.

Children with uncomplicated severe acute malnutrition, not requiring to be admitted and who are managed as outpatients, should be given a course of oral antibiotic such as amoxicillin.

conditional recommendation, low quality evidence

3.2.

Children who are undernourished but who do not have severe acute malnutrition should not routinely receive antibiotics unless they show signs of clinical infection.

strong recommendation, low quality evidence

Rationale

The guideline group noted that the quality of evidence regarding the efficacy of antibiotics in children with uncomplicated severe acute malnutrition was low and agreed that the first recommendation should therefore be conditional. However, the guideline group acknowledged that there is an extensive evidence base highlighting the adverse consequences of indiscriminate use of antibiotics, including antibiotic resistance and, though this evidence was not directly reviewed in the meeting, the group considered the evidence to be of high quality and that the second recommendation should be strong.

A conditional recommendation indicates that a decision by a national or local health authority to include, or not to include, an intervention into local protocols is context sensitive and should be based on consideration of local circumstances.

The guideline group considered the findings from the one randomized controlled trial to be informative and robust. The reduced mortality associated with the intervention was significant and such benefit should be offered to this group of very high-risk children.

The WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 (2) recommends oral antibiotics for children with severe acute malnutrition but without overt signs of sepsis. The classification “uncomplicated” that is now applied to children with severe acute malnutrition who can be managed at home, identifies the same children referred to in the 1999 manual, who previously received oral antibiotics in hospital.

It should be noted that the classification “uncomplicated” should not be interpreted to mean that infection is not necessarily present but that the overall condition of the child means that outpatient management can be considered.

Furthermore, providing antibiotics to the specific population of children with uncomplicated severe acute malnutrition does not mean that any child with some degree of undernutrition – e.g. low weight-for-age, moderate stunting or moderate wasting (weight-for-height <−2 Z-score but >−3 Z- score; mean upper-arm circumference <125 mm but >11.5 cm) – should also be routinely given antibiotics.

Providing antibiotics to children with uncomplicated severe acute malnutrition is unlikely to cause them harm, e.g. anaphylaxis or other hypersensitivities. It may, however, contribute to antimicrobial resistance in the community.

Monitoring of antimicrobial resistance in the population should be routine practice when broad-spectrum antibiotics are used for any community treatment of children with life-threatening conditions. Outpatient treatment with a broad-spectrum antibiotic is already recommended for specific conditions, e.g. treating children with fast breathing. Future adaptations of national policies and guidelines to give oral antibiotics to children with uncomplicated severe acute malnutrition intensifies the need to monitor patterns of antibiotic resistance at population level.

The strength of the first recommendation is conditional for two main reasons:

1. the magnitude of population-based antimicrobial resistance that may follow implementation of the recommendation among the target population is unknown. It is also not known whether implementation of the recommendation will result in significant reductions in the therapeutic efficacy of antibiotics, through antimicrobial resistance, when used for treating other patients with life-threatening conditions;
2. there were significant contextual factors regarding the evidence underpinning the recommendation, namely, most children in the study had bilateral oedema and a significant proportion were HIV exposed. These factors may limit the generalizability of the recommendation. Despite these concerns, the guideline group considered that it would be inappropriate and unethical to deny children with a high mortality risk access to an intervention that may significantly reduce mortality.

Further research is essential to:

• establish the efficacy of routine antibiotics in children with uncomplicated severe acute malnutrition living outside southern Africa, and therefore the generalizability of the recommendation;
• monitor the effect of this intervention on population-based antimicrobial resistance;
• establish the cost effectiveness of the intervention in different regions.

Implications for future research

Discussion with guideline development group members and stakeholders highlighted the limited evidence available in themes related to the priority areas listed next.

• What is the clinical effect and cost effectiveness of giving oral antibiotics to children and infants with severe acute malnutrition who do not require inpatient management in:

  —

  settings with predominantly wasting (e.g. West Africa, South Asia); and

  —

  non-HIV settings (randomized controlled trial with mortality as main outcome)?

• What is the effect of giving broad-spectrum antibiotics to infants and children with severe acute malnutrition without complications, who do not require inpatient management, on:

  —

  the prevalence of population-based antimicrobial resistance;

  —

  therapeutic efficacy.

Other issues (no specific order)

• What clinical algorithms or point-of-care technologies can identify the presence of significant bacterial infections in infants and children with uncomplicated severe acute malnutrition?
• What is the positive and negative predictive value of the appetite test for identifying children with severe acute malnutrition and clinically important infection?
• What are the most effective antibiotics for managing children with complicated severe acute malnutrition who are admitted for inpatient care:

  —

  stratified by HIV status, complications, type of severe acute malnutrition (oedematous versus wasting) and age;

  —

  taking account of in vivo and in vitro resistance versus effectiveness?

Preamble

Globally, between 100 and 140 million children are vitamin A deficient, 4.4 million of whom are estimated to have xerophthalmia (7, 60). Many children living in low-resource settings have a marginal deficiency because of poor intake of either preformed vitamin A, or its precursor, carotene, from the diet. Vitamin A is essential to maintain mucosal barriers and for normal humoral and cellular immune responses. In response to infections, inflammatory processes may disrupt vitamin A metabolism and the release of vitamin A from body stores.

In addition to impairing immune responses, vitamin A deficiency causes the epithelial lining to produce less mucus, which enables bacterial adherence and thereby the invasion of pathogenic microbes (61). Untreated vitamin A deficiency in all children, including severely malnourished children, leads to blindness and increased susceptibility to infection (2) and mortality. There is, however, evidence from randomized trials of vitamin A toxicity and adverse health outcomes in certain settings, such as in patients with pneumonia (62).

Several observational studies have reported the association between severe acute malnutrition and vitamin A deficiency. In Brazil, low serum retinol (less than 0.70 μmol/L) was more prevalent among hospitalized children with severe acute malnutrition than well-nourished hospitalized children, after controlling for several important factors (63). In a subgroup analysis, lower serum retinol concentrations in children with severe acute malnutrition were associated with increased diarrhoeal morbidities (P = 0.021). Among hospitalized Egyptian children with oedematous malnutrition and wasting, plasma vitamin A concentrations were significantly reduced (P < 0.05) (64). A study from Bangladesh that aimed to examine predictors and outcomes associated with shigellosis versus other forms of dysentery, also described the relationship between vitamin A and diarrhoea in children (65). Lower serum retinol concentrations (0.8 μmol/L) were reported in children with shigellosis, especially those with more severe disease and also those with low weight-for-age Z-score. Infection with the more virulent strain, Shigella dysenteriae type 1, showed the lowest serum retinol concentrations. A case-control study in Bangladesh showed that longer duration of diarrhoea (10–14 days and more than 14 days) and weight-for-age less than 60% of the reference median (adjusted odds ratio = 3.8; 95% confidence interval [CI] = 1.8 to 8.0) were independently associated with xerophthalmia, whereas weight-for-age of 60–70% of the expected median showed no association (66).

The WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 recommends giving a high dose of vitamin A (50 000 to 200 000 IU, according to age) to all children with severe acute malnutrition on admission, even if they have no signs of vitamin A deficiency. Children with eye signs of vitamin A deficiency or recent measles should receive a second high dose on the following day and a third high dose 2 weeks later (2).

Commercially available therapeutic formulas (F-75 and F-100) and ready-to-use therapeutic food that complies with the WHO specifications are fortified with vitamin A. This raised questions as to whether it is both safe and necessary to routinely give a high dose of vitamin A to children with severe acute malnutrition, especially those with oedema or signs of hepatic dysfunction, when they are fed with therapeutic foods that already contain vitamin A, and whether there is any risk associated with receiving a single high dose.

There is no question that all children with low, or potentially low, vitamin A status/intake need an increased intake of this nutrient. In resource-poor settings, this is most commonly achieved by supplementation, in particular among children with malnutrition, measles, severe or persistent diarrhoea and other forms of impaired absorption. The question being addressed concerns only the appropriate dose and timing of supplementation.

The guideline development group aimed to review and provide guidance on the following questions:

• What is the effectiveness and safety of giving high-dose vitamin A supplementation to children with severe acute malnutrition when they are receiving a WHO-recommended therapeutic diet containing vitamin A?
• How does the timing of high-dose vitamin A supplementation (i.e. at the beginning, after stabilization or after rehabilitation) affect the effectiveness and safety of the management of children with severe acute malnutrition?

Summary of the evidence

A systematic review examined the effectiveness and safety of giving single high-dose vitamin A supplementation as part of management of children with severe acute malnutrition who are receiving low doses of vitamin A (several times higher than the recommended nutrient intake) through the daily therapeutic diet (67). The review examined the effect of the timing of high-dose vitamin A supplementation (i.e. at the beginning of treatment or after rehabilitation) on treatment outcomes for children with severe acute malnutrition.

Three randomized controlled trials conducted in sub-Saharan Africa and Bangladesh compared daily low-dose (5000 IU) versus single high-dose vitamin A supplementation (100 000 IU for children who were less than 1 year of age, 200 000 IU for children who were aged 1 year or older) in children with severe acute malnutrition (68, 69). The first trial in the Democratic Republic of the Congo found no differences in mortality, acute lower respiratory tract infections, or duration of diarrhoea between malnourished children who received low-dose and high-dose vitamin A supplementation compared with placebo (69). However, children with bilateral pitting oedema who received low-dose vitamin A supplementation had a lower incidence of diarrhoea (relative risk = 0.21; 95% CI = 0.07 to 0.62) compared with placebo; no effect was shown in the high-dose group. Furthermore, severely malnourished children without oedema who received the high-dose supplementation had a two-fold increased risk of severe nosocomial diarrhoea compared with children who received placebo (69). The second trial, in Senegal, gave hospitalized children with severe acute malnutrition either high-dose vitamin A supplementation at admission, or the low-dose supplements daily until discharge (68). The incidence and duration of respiratory infection were lower in the low-dose group than in the high-dose group. In children with oedema on admission, mortality was significantly lower in the low-dose group (adjusted odds ratio = 0.21; 95% CI = 0.05 to 0.99). No differences were detected for diarrhoea morbidity or mortality. The third trial in Bangladesh compared the efficacy of a single high dose (200 000 IU) of vitamin A in addition to a daily low dose (5000 IU) in the management of children with severe acute malnutrition who also had diarrhoea with or without acute lower respiratory tract infections (70). The addition of high-dose vitamin A to daily low-dose supplements did not modify the time to resolution of diarrhoea, recovery from acute lower respiratory tract infections or mortality. No adverse effects were seen with high-dose supplementation of vitamin A.

No direct evidence was available on whether the timing of high-dose vitamin A supplementation, given either early or later in the treatment of children with severe acute malnutrition, influences nutritional recovery, including weight gain, mortality, signs of symptomatic vitamin A deficiency or biological indicators of vitamin A status. Subgroup analyses from the two studies comparing high-dose and daily low-dose vitamin A supplements suggested that, in severely malnourished children with oedema, low-dose vitamin A is more protective than high-dose vitamin A supplementation against diarrhoea incidence (69) and mortality (68). Other trials compared different high-dose vitamin A supplementation with placebo but did not select children on the basis of being malnourished. Several studies gave children vitamin A supplements once at enrolment (71–75), and two studies tested the effect of high-dose supplementation over multiple days (62, 76). Because these studies were heterogeneous and were not done exclusively in children with severe acute malnutrition, it was not possible to draw conclusions about the timing of high-dose supplementation. However, one study among hospitalized children who did not have severe acute malnutrition did observe adverse effects of high-dose vitamin A supplementation on children with respiratory tract infection (62, 76).

In summary:

• low-dose (5000 IU) vitamin A supplementation given daily to children with severe acute malnutrition, from the time of admission until discharge from treatment, is more effective in reducing the mortality of children with oedema, the incidence of severe diarrhoea, and the incidence and duration of respiratory infection than single high-dose vitamin A supplementation (100 000 IU for children who are less than 1 year of age; 200 000 IU for children aged 1 year or older) on day 1 of admission;
• high-dose vitamin A supplementation appears to confer some benefit (compared to receiving no supplementation at all) in children with severe acute malnutrition who present with severe diarrhoea or shigellosis or have clear signs of vitamin A deficiency;
• the evidence for the efficacy and safety of vitamin A supplementation in children with severe acute malnutrition with non-measles pneumonia and other acute lower respiratory tract infections is inconclusive;
• high-dose vitamin A supplementation reduces mortality in children with severe acute malnutrition complicated by measles-specific respiratory infections.

The overall quality of this evidence was rated as low, owing to methodological and reporting bias. See Annex 1, Table 2 for details of the GRADE assessment of the evidence.

Discussion

• The vitamin A intake of children who are fed therapeutic food (F-75, F-100 or ready-to use therapeutic foods) that complies with WHO specifications exceeds the recommended nutrient intake for well-nourished children and seems adequate for malnourished children; there is no clear rationale for giving a single high-dose vitamin A supplement, unless children have eye signs of vitamin A deficiency or have had measles or diarrhoea recently.
• Clarifying the recommendations to avoid double provision of vitamin A (through therapeutic foods and as supplements) unless absolutely required (e.g. eye signs of deficiency or recent measles) reduces opportunity costs and simplifies the work for health-care providers.
• Avoiding double dosing with vitamin A will be cost saving, though the savings were not estimated in any of the studies.
• There may be some programmatic advantages in giving a single high dose of vitamin A, for example in families that are unable to give their children regular therapeutic foods or multivitamin supplements, once the children are taken home.
• Even healthy children may experience adverse effects such as irritability, vomiting, blurred vision and raised intracranial pressure from high-dose vitamin A supplements; these effects may be more pronounced in children with severe acute malnutrition, especially those with bilateral pitting oedema or with compromised liver function. Clarifying and rationalizing the use of vitamin A in this population will help to avert such adverse events, while still identifying the children who will benefit most in terms of reducing mortality.
• It is not clear whether there is a risk of vitamin A toxicity in children who receive a high-dose vitamin A supplement before discharge from treatment for severe acute malnutrition and later receive another high dose during a national vitamin A campaign. The recommendation should follow the WHO guidance for all children, which is not to give the child their age-appropriate high dose more frequently than once a month.

Relevant standing recommendations

The recommendations were confirmed as current in the context of existing WHO recommendations included in the WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 (2), the joint United Nations statement on Community-based management of severe acute malnutrition, 2007 (10) and the WHO publication, The treatment of diarrhoea: manual for physicians and other senior health workers, 2005 (77), noting in particular that:

• a high dose (50 000 IU, 100 000 IU or 200 000 IU, depending on age) of vitamin A should be given to all children with severe acute malnutrition and eye signs of vitamin A deficiency on day 1, with a second and a third dose on day 2 and day 15 (or at discharge from the programme), irrespective of the type of therapeutic food they are receiving;
• a high dose (50 000 IU, 100 000 IU or 200 000 IU, depending on age) of vitamin A should be given to all children with severe acute malnutrition with recent measles on day 1, with a second and a third dose on day 2 and day 15 (or at discharge from the programme), irrespective of the type of therapeutic food they are receiving.

Recommendations

4.1.

Children with severe acute malnutrition should receive the daily recommended nutrient intake of vitamin A throughout the treatment period. Children with severe acute malnutrition should be provided with about 5000 IU vitamin A daily, either as an integral part of therapeutic foods or as part of a multi-micronutrient formulation.

strong recommendation, low quality evidence

4.2.

Children with severe acute malnutrition do not require a high dose of vitamin A as a supplement if they are receiving F-75, F-100⁶ or ready-to-use therapeutic food that comply with WHO specifications (and therefore already contain sufficient vitamin A), or vitamin A is part of other daily supplements.

strong recommendation, low quality evidence

4.3.

Children with severe acute malnutrition should be given a high dose of vitamin A (50 000 IU, 100 000 IU or 200 000 IU, depending on age) on admission, only if they are given therapeutic foods that are not fortified as recommended in WHO specifications and vitamin A is not part of other daily supplements.

strong recommendation, low quality evidence

Rationale

The guideline group noted that the quality of evidence was low because there were serious differences in the criteria used to define study populations, and that one study included, a priori, children with diarrhoea. Also, two of the three studies were published by the same principal investigator. However, the guideline group agreed that the recommendation should be strong. The recommendations clarify the role of vitamin A in the management of children with severe acute malnutrition and this was considered an important safety measure. They reduce the likelihood of double dosing with vitamin A and therefore should reduce opportunity costs and the likelihood of adverse events related to vitamin A. Vitamin A supplementation has been feasible in standard care of children with severe acute malnutrition. Vitamin A is widely available and studies in non-malnourished children have found vitamin A supplements to be a cost-effective child-survival intervention. Mortality reductions in this high-risk group of children are likely to be similarly cost effective. While additional randomized trials are warranted (see below), the guideline group agreed that it was justified to assign the recommendations as strong, in order to avoid the potential harm of children receiving too much vitamin A.

Implications for future research

Discussion with guideline development group members and stakeholders highlighted the limited evidence available in themes related to the priority areas listed next.

• What is the efficacy of daily low-dose vitamin A supplementation compared to single high-dose vitamin A in the treatment of children with severe acute malnutrition either with bilateral pitting oedema or presenting with severe diarrhoea or shigellosis?
• What is the most effective way to improve and sustain the vitamin A status of children with severe acute malnutrition after discharge from treatment?
• Are there regional differences in the response to, and safety of, vitamin A supplementation in children with severe acute malnutrition?

5.1. Therapeutic feeding of children with severe acute malnutrition and acute diarrhoea treated as outpatients

5.2. Therapeutic feeding of children with severe acute malnutrition and persistent diarrhoea

5.3. Therapeutic feeding of children during the transition period from initial stabilization to nutritional rehabilitation (in inpatient settings)

6.1. Fluid management in children with severe malnutrition and dehydration without shock

6.2. Fluid management of children with severe acute malnutrition and shock

Preamble

HIV-infected children commonly present with moderate or severe acute malnutrition (13). The exact cause of wasting in these children is probably multifactorial but includes altered glucose and lipid metabolism, raised basal metabolic rate, especially when opportunistic infections are present, multiple micronutrient deficiencies, higher rates of diarrhoea and malabsorption, frequent co-infections and higher rates of food insecurity and poverty. The health and survival of HIV-infected children also suffer when their mothers become ill or die and cannot provide care as usual. As a result of being immune compromised, HIV-infected children suffer high mortality rates, even without the added complication of severe acute malnutrition (140). However, HIV-infected children with severe acute malnutrition are nearly three times more likely to die while on nutritional rehabilitation therapy, compared to their HIV-negative counterparts (13).

HIV-infected children are susceptible to higher rates of persistent diarrhoea and to opportunistic infections such as tuberculosis (TB), cryptosporidiosis, disseminated candidiasis and cryptococcosis. Yet, despite these well-recognized clinical patterns, there has been little research on the management of HIV-infected children with severe acute malnutrition.

The effectiveness of antiretroviral drug treatment to significantly reduce the frequency of opportunistic infection and improve the survival of HIV-infected adults and children provides cause for hope. WHO recommends that all infants below 24 months of age should be started on antiretroviral drug treatment, irrespective of clinical condition, and that children who are 24 months of age or older should start if, among other criteria, there is unexplained malnutrition (141). For these reasons, WHO recommends screening of children to determine HIV status where there are any clinical signs consistent with HIV infection, and to assess eligibility for antiretroviral drug treatment. In the context of severe acute malnutrition, children who are known to be HIV exposed should be tested for HIV status, in order to determine their status and eligibility for antiretroviral drug treatment. In settings where HIV infection is common (HIV prevalence more than 1%), children with malnutrition should be tested for HIV, in order to establish their HIV status and to determine their need for antiretroviral drug treatment. WHO provides simplified dosing schedules for antiretroviral drugs based on weight bands.

In many settings, individuals and families who are infected or affected by HIV are the target of marketing of commercial products claiming to provide specific benefit to persons infected with HIV. These products, frequently nutritional products, have generally not been tested to demonstrate efficacy or not. In light of this influence, it is important to clarify which interventions provide clinical benefit to adults and children infected with HIV (142).

Four key issues related to the care of HIV-infected children with severe acute malnutrition, for which guidance should be developed were identified:

• What is the optimal timing for initiating and dosing of antiretroviral drug treatment?
• What are the optimal feeding regimens for HIV-infected children with severe acute malnutrition and do these differ from those for uninfected children with severe acute malnutrition?
• What is the value (effectiveness and safety) of vitamin A supplementation?
• What are the most effective therapeutic strategies for managing diarrhoea?

Summary of the evidence

A systematic review was conducted to examine these aspects of the treatment of HIV-infected children with severe acute malnutrition (143). No randomized controlled trials in HIV-infected children with severe acute malnutrition were identified that directly addressed any of the prioritized questions. Indirect evidence was available from other randomized trials and observational studies that reported outcomes of HIV-infected children who were undernourished, of whom some were severely malnourished. However, relevant anthropometric indices were not generally reported to determine whether any of the children fulfilled WHO definitions of severe acute malnutrition.

Nine observational studies reported outcomes of HIV-infected children started on lifelong antiretroviral drug treatment. Low weight-for-age was associated with increased mortality (144–146), and the development of immune reconstitution inflammatory syndrome (147). One retrospective study among malnourished children admitted for therapeutic feeding reported improved recovery when antiretroviral drug treatment was started within 21 days of admission compared to children who had delayed initiation of, or who were not considered eligible, for antiretroviral drug treatment (148). There were no data regarding the use of different antiretroviral regimens and their relative outcomes in HIV-infected children with severe acute malnutrition. Two studies reported pharmacokinetic data on HIV-infected children. Suboptimal dosing was not found when dosing was based on weight-band tables compared to requirements estimated using calculations of surface area (149). One study from India reported no difference in trough or 2-h nevirapine concentrations in children with weight-for-age Z-score <−2, compared to weight-for-age Z-score ≥−2, though stunted children had significantly lower median 2-h nevirapine concentrations compared to non-stunted children (150).

For HIV-infected children with severe acute malnutrition, there were no controlled comparative trials examining the effect of WHO-recommended therapeutic feeding approaches compared with other therapeutic feeds. Observational studies reported clinical recovery when ready-to-use therapeutic food was given (151, 152), but these did not compare outcomes with other therapeutic feeds such as F-100. One randomized trial examined the effect of a combination prebiotic and probiotic included as part of a therapeutic feeding approach. There was no benefit evident from inclusion of the combination product (89).

No studies directly tested the value of vitamin A in the care of HIV-infected children with severe acute malnutrition. However, indirect evidence from studies evaluating the effect of vitamin A supplements (200 000 IU every 4 months) in HIV-infected children who were 15–36 months of age without severe acute malnutrition reports improved growth and decreased all-cause mortality when children are regularly given supplements of vitamin A (153). There were no reports of adverse effects of vitamin A in HIV-infected children and there is currently no reason to believe that vitamin A might have an adverse effect in HIV-infected children with severe acute malnutrition.

Although diarrhoea is common in both HIV-infected children and children with severe acute malnutrition, there were no controlled studies identified that directly investigated diagnostic strategies or clinical interventions for HIV-infected children with severe acute malnutrition and diarrhoea. Indirect evidence suggests that Cryptosporidium is a common pathogen causing diarrhoea and that disaccharide intolerance is also common. While one study examined clinical outcomes following therapeutic feeding with either an elemental milk diet or a skimmed milk and soya diet, there is not even indirect evidence about the use of F-100 in children with HIV infection and diarrhoea.

There was similarly no direct evidence about the effect of zinc supplements in HIV-infected children with severe acute malnutrition who have diarrhoea. Zinc has been found to be safe in HIV-infected children without severe acute malnutrition (153). However, the effect of zinc on prevention or management of diarrhoea in HIV-infected children, with or without severe acute malnutrition, is yet to be established. Given that there are no reported adverse effects of giving zinc to children with HIV infection, there appears to be no basis to omit zinc from the routine management of HIV-infected children with severe acute malnutrition.

The overall quality of this evidence has been rated as very low, owing to indirectness and imprecision. A GRADE table was not developed as there were no studies that directly reported the outcomes of interest in HIV-infected children with severe acute malnutrition.

Discussion

• The former WHO recommendations on severe malnutrition (2) did not recommend HIV testing of children with severe acute malnutrition, even in settings of high HIV prevalence. At that time, there was little experience with the use of antiretroviral drugs in children, nor was there significant coverage of antiretroviral drugs as an intervention. Circumstances today are very different.
• There is high-quality evidence that antiretroviral drug treatment significantly improves the survival of children who are infected with HIV. WHO recommends that children are tested for HIV, especially in high-prevalence settings, when there are clinical signs consistent with HIV infection, or when there is known exposure (141). In the context of severe acute malnutrition, there is strong justification for routinely testing all children for HIV, in order to identify a population that would benefit from a highly effective intervention, namely antiretroviral drug treatment.
• There remains little direct evidence on the management of HIV-infected children with severe acute malnutrition. The questions identified by the guideline development group remain largely unanswered and recommendations are largely based on extrapolations of evidence in other populations and personal experience. There are several studies currently ongoing examining the pharmacokinetics of antiretroviral drugs in HIV-infected children with severe acute malnutrition. They are also designed to examine some of the mechanisms and risk factors for children developing immune reconstitution inflammatory syndrome. However, other important issues remain unexamined and research is urgently needed. Apart from prioritizing the initiation of antiretroviral drug treatment, the guideline group considered there was no basis, either from indirect evidence or on the grounds of biological plausibility, to recommend management approaches for HIV-infected children with severe acute malnutrition that differed significantly from what is recommended for uninfected children.

Relevant standing recommendations

The recommendations below were confirmed as current in the context of existing WHO recommendations included in the WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 (2), WHO Diarrhoea treatment guidelines for clinic-based healthcare workers, 2005 (132), WHO Guidelines for an integrated approach to the nutritional care of HIV-infected children, 2009 (142), WHO guidelines on Antiretroviral therapy for HIV infection in infants and children, 2010 (141), and WHO Guidelines for intensified tuberculosis case-finding and isoniazid preventive therapy for people living with HIV in resource-constrained settings, 2011 (154), noting in particular that:

• all HIV-exposed infants and children (including those with severe acute malnutrition) should be tested for HIV status;
• in settings where HIV infection is common (HIV prevalence more than 1%), children with severe acute malnutrition should be tested for HIV, in order to establish their HIV status and to determine their need for antiretroviral drug treatment;
• all HIV-infected infants and children who are less than 24 months of age should be initiated with lifelong antiretroviral drug treatment, irrespective of clinical staging (including severe acute malnutrition) and CD4 count. Co-trimoxazole prophylaxis should also be initiated;
• all HIV-infected children who are over 24 months and less than 5 years of age should be started on lifelong antiretroviral drug treatment based on their CD4 count (≤750 cells/mm³) or CD4 percentage (≤25%), or if they have WHO clinical staging 3 or 4 (including severe acute malnutrition). A new recommendation is due to be published in 2013;
• children living with HIV who have any one of the following symptoms – poor weight gain, fever, current cough or contact history with a TB case – may have TB and should be evaluated for TB and other conditions.

Recommendations

7.1.

Children with severe acute malnutrition who are HIV infected and who qualify for lifelong antiretroviral therapy should be started on antiretroviral drug treatment as soon as possible after stabilization of metabolic complications and sepsis. This would be indicated by return of appetite and resolution of severe oedema. HIV-infected children with severe acute malnutrition should be given the same antiretroviral drug treatment regimens, in the same doses, as children with HIV who do not have severe acute malnutrition. HIV-infected children with severe acute malnutrition who are started on antiretroviral drug treatment should be monitored closely (inpatient and outpatient) in the first 6–8 weeks following initiation of antiretroviral therapy, to identify early metabolic complications and opportunistic infections.

strong recommendation, very low quality evidence

7.2.

Children with severe acute malnutrition who are HIV infected should be managed with the same therapeutic feeding approaches as children with severe acute malnutrition who are not HIV infected.

strong recommendation, very low quality evidence

7.3.

HIV-infected children with severe acute malnutrition should receive a high dose of vitamin A on admission (50 000 IU to 200 000 IU depending on age) and zinc for management of diarrhoea as indicated for other children with severe acute malnutrition, unless they are already receiving F-75, F-100 or ready-to-use therapeutic food, which contain adequate vitamin A and zinc if they are fortified following the WHO specifications.

strong recommendation, very low quality evidence

7.4.

HIV-infected children with severe acute malnutrition in whom persistent diarrhoea does not resolve with standard management should be investigated to exclude carbohydrate intolerance and infective causes, which may require different management, such as modification of fluid and feed intake, or antibiotics.

strong recommendation, very low quality evidence

Rationale

The guideline group agreed that the recommendation should be strong, while noting that the quality of evidence was very low. No studies reported on the effect of the timing of starting antiretroviral drug treatment and mortality. Initiation of antiretroviral drug treatment and rates of recovery and weight gain were, however, noted. Initiation of antiretroviral drug treatment is already recommended for HIV-infected children who have severe acute malnutrition on clinical criteria. The first recommendation above aims to clarify how quickly treatment should be started, not whether it should be started. The guideline group considered that it is feasible to start antiretroviral drug treatment when appetite returns and, even though cost data were not presented or formally estimated, this timing would not have significant cost implications compared to starting 1–2 weeks later. The guideline group considered that the other recommendations serve to clarify relevant care and appropriately promote consistency of care for the three populations, namely HIV-infected children, children with severe acute malnutrition and HIV-infected children with severe acute malnutrition. The guideline group considered there to be major advantages, e.g. for training and supply-chain management, and little risk of harm in harmonizing recommendations.

Implications for future research

Discussion with guideline development group members and stakeholders highlighted the limited evidence available in themes related to the priority areas listed next.

• Does early initiation of antiretroviral drug treatment (as soon as metabolic complications are stabilized and sepsis is treated) in HIV-infected children with severe acute malnutrition improve outcomes and reduce adverse events such as immune reconstitution inflammatory syndrome or dyslipidaemia, compared with later initiation?
• Are the pharmacokinetic characteristics of HIV-infected children with severe acute malnutrition being started on antiretroviral drug treatment different from those of HIV-infected children who do not have severe acute malnutrition? Is there any need to modify the dose of antiretroviral drugs in HIV-infected children who are severely malnourished, to either avoid toxicity or ensure adequate therapeutic levels?
• Are the pharmacokinetic characteristics of HIV-infected children with severe acute malnutrition receiving other drugs, such as isoniazid, different from those of HIV-infected children who do not have severe acute malnutrition?
• What is the optimal dosing of antiretroviral drug treatment in HIV-infected children with severe acute malnutrition, to optimize clinical recovery and viral suppression and avoid the development of metabolic complications such as immune reconstitution inflammatory syndrome?
• What is the most effective therapeutic feeding approach for HIV-infected children with severe acute malnutrition who have persistent diarrhoea?
• Are the pathophysiological abnormalities of HIV-infected children with severe acute malnutrition, with or without oedema, the same as those of children with severe acute malnutrition but without HIV?

Preamble

Severe acute malnutrition is increasingly being recognized in infants who are less than 6 months of age (155). In addition to aetiologies such as low birth weight, persistent diarrhoea and recurring sepsis or chronic underlying diseases or disability, the development of severe acute malnutrition in this age group commonly reflects suboptimal feeding practices, especially breastfeeding practices. Infants who are less than 6 months of age should be exclusively breastfed to gain optimal nutrition and the greatest protection against infections. Yet, rates of exclusive breastfeeding worldwide remain disappointingly low, with only an estimated 25–31% of infants who are 2–5 months of age being exclusively breastfed (7).

Severe acute malnutrition is often associated with higher mortality in young infants than in older infants and children (24). While this is not unexpected, given the greater vulnerability of young infants and the range of possible underlying or contributing pathologies, the magnitude of this excess mortality risk and how much of it is avoidable are unknown. Few studies have examined the risk factors for young infants developing severe acute malnutrition; similarly, there are few reports describing approaches to managing severe acute malnutrition in this age group. Risk factors for increased mortality are likely to include recent weight loss, failure to gain weight, failure to feed effectively and the presence of bilateral oedema (24). To date, management of severe acute malnutrition in this age group has focused on establishing, or re-establishing exclusive breastfeeding, and, where this has not been possible, there are some reports of using formula feeds and early introduction of complementary foods to treat these infants (24).

However, there are important physiological differences between young infants and older children that justify separate consideration of the management of severe acute malnutrition in this age group. Before 6 months of age, physiological processes, including thermoregulation and renal and gastrointestinal functions, are relatively immature compared with those of older children and may require modified management approaches or clinical interventions. Clinical signs of infection and hydration status may also be more difficult to identify and interpret in the younger infant. As a result, criteria for admitting and discharging infants with severe acute malnutrition have not been adequately defined. Furthermore, it is unclear whether criteria and management approaches for older children with severe acute malnutrition, including outpatient care, can be appropriately extended to this population.

The guideline development group aimed to provide guidance on the following questions:

• What are the criteria for defining severe acute malnutrition in infants who are less than 6 months of age?
• What are the criteria for hospital admission of infants who are less than 6 months of age with severe acute malnutrition?
• What are the essential interventions, especially feeding approaches, for infants who are less than 6 months of age with severe acute malnutrition?
• What are the criteria for transferring infants who are less than 6 months of age and have been treated in hospital for severe acute malnutrition to outpatient care, or discharging them from treatment?

Summary of the evidence

Systematic reviews were conducted to review the admission and discharge criteria for the management of infants who are less than 6 months of age with severe acute malnutrition (156), and inpatient treatment of severe acute malnutrition in infants who are less than 6 months of age (157).

No studies were identified that directly examined admission and discharge criteria using the WHO growth standards (8). No studies were found in the peer-reviewed literature that reported outcomes when WHO therapeutic feeding recommendations for children who are over 6 months of age with severe acute malnutrition are applied to severely malnourished infants who are less than 6 months of age.

Infants who are less than 6 months of age with severe acute malnutrition were included in some studies (158–161) examining therapeutic feeding approaches. Subgroup analyses were not possible however, because of small sample sizes or inadequate detail of treatment for this subgroup. A randomized controlled trial in the Democratic Republic of the Congo (37, 162) recruited 161 infants who were less than 6 months of age with severe acute malnutrition defined as weight-for-length less than 70% of the median NCHS reference values, and who did not have oedema. However, infants who were not gaining weight at home or were too weak to breastfeed, or where the mother reported an inability to breastfeed, were also included. Infants were randomized to receive either diluted F-100 therapeutic milk (73.11 kcal/100 mL) or an isocaloric standard generic infant formula milk (76.7 kcal/100 mL). Infants were otherwise given the same treatment, which included support for continued breastfeeding, supplemental suckling when needed and vitamins, folic acid and antibiotics. The volumes of dilute F-100 or generic infant formula milk supplement were decreased by half when the infant was gaining at least 20 g/day for three consecutive days and were completely stopped when the infant maintained 10 g/day weight gain. No differences were found in weight gain, total duration of treatment, or treatment outcome (death, recovery or default).

A doctoral thesis, however, reported the use of WHO therapeutic feeding protocols for children who were 6–59 months of age when they were applied to infants who were less than 6 months of age. Infants from 13 countries in Africa were admitted for treatment of severe acute malnutrition, having been identified by weight-for-length <70% of the median reference value (NCHS). Infants who were less than 6 months of age had a higher mortality rate, particularly those who were not breastfed, than children with severe acute malnutrition who were older but treated with the same protocol. Attempts to put the infants to the breast up to 24 times a day were abandoned because of rapid weight loss (107). The supplemental suckling technique was subsequently successfully introduced to feed such infants (38). The increased mortality was considered to be due to the use of undiluted F-100, which resulted in a high renal solute load.

No studies examined feeding approaches in the transition and rehabilitation phases of treatment of severe acute malnutrition in infants who were less than 6 months of age. In some national guidelines that include supplemental suckling treatment as part of care for these infants, the volume of supplemental milk is gradually stepped down and ultimately stopped, as the infant improves and the supply of breast milk is re-established (24). However, this practice has not been referred to as “transitioning” to a rehabilitation phase. No references were found regarding the use of cup or spoon feeding in the context of young infants with severe acute malnutrition, although their effectiveness has been shown in young children without severe acute malnutrition in other settings.

The overall quality of this evidence has been rated as very low, owing to indirectness and imprecision. As no studies were found in the peer-reviewed literature to address the specific questions identified by the guideline group, it was not possible to develop GRADE tables for the assessment of evidence.

Discussion

• There is little empirical evidence available to inform recommendations on the management of severe acute malnutrition in infants who are less than 6 months of age. However, the guideline group acknowledged that programmatic guidance is important for health-care workers treating severe acute malnutrition in this age group. However, a summary of case-reports on how infants who are less than 6 months of age with severe acute malnutrition are being managed in some settings is available (24).
• The guideline development group placed value on highlighting general nutritional and infant feeding principles and their application to young infants with severe acute malnutrition.
• It was also agreed that general principles on the care of sick neonates and infants, such as identifying and referring infants who are seriously ill for other reasons, can reasonably be extended to infants with severe acute malnutrition.
• The group highlighted the relevance and value of other existing recommendations, even though they are not explicitly for infants with severe acute malnutrition. These include WHO Guiding principles for feeding non-breastfed children 6–24 months of age, 2005 (163), or the WHO Infant and young child feeding counselling: an integrated course, 2006 (164).
• It was considered that the use of the recently published 2011 WHO growth velocity standards (8) and the assessment of growth velocity (weight) was equally important as, or even more important than, a single weight-for-length value. This highlights the value of maintaining and examining growth-monitoring charts at primary health-care level for infants and young children.
• As with older children, severe acute malnutrition in infants who are less than 6 months of age can be “uncomplicated” or “complicated”. These infants should be managed in outpatient or inpatient settings respectively. The potential benefits of inpatient treatment should be carefully considered against the potential risks, especially nosocomial infections.
• The guideline development group acknowledged that generic infant formula milk can be provided safely in some inpatient settings. However, mothers or caregivers, who are expected to give formula milk to their infants after they are discharged from inpatient care, need clear guidance on safe preparation and use of such replacement feeds. This information is included in the WHO integrated course on infant and young child feeding (164).
• The guideline development group also emphasized that when programmes provide formula milk as part of case-management of infants with severe acute malnutrition, it should be implemented appropriately and not confuse or compromise the wider public health message concerning exclusive breastfeeding for infants who are less than 6 months of age.
• There is limited evaluated experience and evidence on managing infants who are less than 6 months of age with bilateral oedema, especially regarding the use of F-75 and diluted F-100.The question was considered by participants of a WHO meeting held in 2004 to review programmatic data and theoretical considerations on the management of severe acute malnutrition in young infants. The participants concluded that young infants with oedema should be fed with F-75 during the stabilization phase (164).
• In contrast, participants of the same meeting concluded that F-75 would not be appropriate for infants with severe acute malnutrition who are severely wasted and without oedema (165). Rather, generic infant formula or dilute F-100 would be appropriate and safer. However, there were no data presented at the meeting to substantiate these conclusions.
• The meeting report described the diversity of opinions on the feeding strategy for the rehabilitation phase. However, there was consensus that children who are less than 6 months of age with severe acute malnutrition should not be given undiluted F-100, owing to high renal solute load and risk of insufficient water intake and hypernatraemic dehydration (165).
• Infants who are less than 6 months of age have small daily weight gains that cannot easily be assessed with standard spring scales. Health-care workers taking care of infants with severe acute malnutrition should use scales with at least a 20 g precision.
• It was noted that, in some sites, ready-to-use therapeutic food is being given to infants who are 5–6 months of age, sometimes as a pragmatic response to situations where exclusive breastfeeding is either not possible or difficult for a variety of reasons. Participants recognized that there is no evaluated experience or evidence to indicate the effectiveness and safety of this approach.
• As with several other aspects of care, an urgent research agenda was identified.

Relevant standing recommendations

The recommendations were confirmed as current in the context of existing WHO recommendations included in the WHO Management of severe malnutrition: a manual for physicians and other senior health workers, 1999 (2) and Management of the child with a serious infection or severe malnutrition. Integrated Management of Childhood Illness (IMCI) 2000 (16), noting in particular that:

• severe acute malnutrition in infants who are 0–5 months of age is defined as:

  —

  weight-for-length less than−3 Z-score, or

  —

  presence of bilateral pitting oedema;

• infants who have been identified to have poor weight gain and who have not responded to nutrition counselling and support (IMCI) should be admitted for further investigation and treatment;
• any infant or child with a general danger sign as defined by the IMCI (16)¹¹ should be admitted for urgent treatment and care.

Recommendations

8.1.

Infants who are less than 6 months of age with severe acute malnutrition with any of the following complicating factors should be admitted for inpatient care:

1. any serious clinical condition or medical complication as outlined for infants 6 months of age or older with severe acute malnutrition;
2. recent weight loss or failure to gain weight;
3. ineffective feeding (attachment, positioning and suckling) directly observed for 15–20 min, ideally in a supervised separated area;
4. any pitting oedema;
5. any medical or social issue needing more detailed assessment or intensive support (e.g. disability, depression of the caregiver, or other adverse social circumstances).

strong recommendation, very low quality evidence

8.2.

Infants who are less than 6 months of age with severe acute malnutrition should receive the same general medical care¹² as infants with severe acute malnutrition who are 6 months of age or older:

1. infants with severe acute malnutrition who are admitted for inpatient care should be given parenteral antibiotics to treat possible sepsis and appropriate treatment for other medical complications such as TB, HIV, surgical conditions or disability;
2. infants with severe acute malnutrition who are not admitted should receive a course of broad-spectrum oral antibiotic, such as amoxicillin, in an appropriately weight-adjusted dose.

strong recommendation, very low quality evidence

8.3.

Feeding approaches for infants who are less than 6 months of age with severe acute malnutrition should prioritize establishing, or re-establishing, effective exclusive breastfeeding by the mother or other caregiver.

strong recommendation, very low quality evidence

8.4.

Infants who are less than 6 months of age with severe acute malnutrition who are admitted:

1. should be breastfed where possible and the mothers or female caregivers should be supported to breastfeed the infants. If an infant is not breastfed, support should be given to the mother or female caregiver to re-lactate. If this is not possible, wet nursing¹³ should be encouraged;
2. should also be provided a supplementary feed:

   —

   supplementary suckling approaches should, where feasible, be prioritized;

   —

   for infants with severe acute malnutrition but no oedema, expressed breast milk should be given, and, where this is not possible, commercial (generic) infant formula or F-75 or diluted F-100¹⁴ may be given, either alone or as the supplementary feed together with breast milk;

   —

   for infants with severe acute malnutrition and oedema, infant formula or F-75 should be given as a supplement to breast milk.

3. should not be given undiluted F-100 at any time (due to the high renal solute load and risk of hypernatraemic dehydration);
4. if there is no realistic prospect of being breastfed, should be given appropriate and adequate replacement feeds such as commercial (generic) infant formula, with relevant support to enable safe preparation and use, including at home when discharged.
   In addition:
5. assessment of the physical and mental health status of mothers or caregivers should be promoted and relevant treatment or support provided.
   strong recommendation, very low quality evidence

8.5.

Infants who are less than 6 months of age and have been admitted to inpatient care can be transferred to outpatient care when:

1. all clinical conditions or medical complications, including oedema, are resolved, and
2. the infant has good appetite, is clinically well and alert, and
3. weight gain on either exclusive breastfeeding or replacement feeding is satisfactory, e.g. above the median of the WHO growth velocity standards or more than 5 g/kg/day for at least 3 successive days, and
4. the infant has been checked for immunizations and other routine interventions, and
5. the mothers or caregivers are linked with needed community-based follow-up and support.

strong recommendation, very low quality evidence

8.6.

Infants who are less than 6 months of age can be discharged from all care when they:

1. are breastfeeding effectively or feeding well with replacement feeds, and
2. have adequate weight gain, and
3. have a weight-for-length ≥−2 Z-score.

strong recommendation, very low quality evidence

8.7.

For infants who are less than 6 months of age with severe acute malnutrition and who do not require inpatient care (Recommendation 8.1), or whose caregivers decline admission for assessment and treatment:

1. counselling and support for optimal infant and young child feeding should be provided, based on general recommendations for feeding infants and young children, including for low-birth-weight infants;
2. weight gain of the infant should be monitored weekly to observe changes;
3. if the infant does not gain weight, or loses weight while the mother or caregiver is receiving support for breastfeeding, then he or she should be referred to inpatient care;
4. assessment of the physical and mental health status of mothers or caregivers should be promoted and relevant treatment or support provided.

strong recommendation, very low quality evidence

Rationale

The guideline group noted that the quality of evidence was very low but the group was in agreement that the recommendations should be strong. The majority of recommendations outline care practices that are commonly regarded as standard of care and reflected in other WHO recommendations, especially the IMCI (16) and the WHO Pocket book of hospital care for children (42). Others are documented in reviews of clinical practice that are cited above. The guideline group also noted that guidelines have not previously been written for this specific age group of infants with severe acute malnutrition. It is therefore of value to affirm principles for this population that are already accepted for other groups, e.g. establishing effective exclusive breastfeeding by the mother or other caregiver, or assessing the physical and mental health status of mothers. The guideline group regarded the recommendations to be important principles of care that may not be easily subjected to randomized controlled trials. The group considered that recommending these practices, even in the absence of clinical or costing data, would benefit infants with severe acute malnutrition with no risk, or negligible risk, of harm. The guideline group hoped that outlining areas of good practice would highlight the need for research in this population of infants that would begin to build an evidence base for best clinical care, e.g. the most effective and safest therapeutic feeding approaches (see below).

Implications for future research

Discussion with guideline development group members and stakeholders highlighted the limited evidence available in themes related to the priority areas listed next.

• In infants who are less than 6 months of age, what is the predictive value of population-derived thresholds for weight-for-height, mid-upper arm circumference and reduced growth velocity (weight-for-age) with or without oedema to identify infants at high risk of mortality?

  —

  Consider analysis of published or other existing data.

  —

  Consider the feasibility of each assessment.

• What are the most effective and safest therapeutic feeding approaches, including different food “recipes” in addition to breastfeeding, for infants who are less than 6 months of age with severe acute malnutrition?
• How is breastfeeding most effectively re-established among infants with a low weight-for-height/poor weight gain?

Other issues (no specific order)

• Do infants who are less than 6 months of age with weight-for-length less than −3 Z-score also have reduced growth velocity and is there any difference according to epidemiological setting such as African, South Asian and South-East Asian sites?
• In infants who are less than 6 months of age, what is the feasibility and accuracy of using weight-for-height, mid-upper arm circumference and reduced growth velocity (weight-for-age) with or without oedema to identify infants in need of therapeutic management?
• What criteria most effectively identify infants who are less than 6 months of age with the metabolic abnormalities/adaptations associated with severe acute malnutrition in older children?
• What is the effectiveness, tolerance and safety of ready-to-use therapeutic food as an adjuvant to breastfeeding in infants who are less than 6 months of age with severe acute malnutrition?

  —

  Consider specific subgroups, e.g. with oedema, 4–6months old, those with underlying diseases such as HIV.

• What is the recommended folic acid supplementation for infants who are less than 6 months of age with severe acute malnutrition?
• Are there adjustments in the drug dosage or selection of drugs required when managing infants who are less than 6 months of age with severe acute malnutrition?
• Is there any reason to expect that the efficacy of vitamin A supplementation for infants who are less than 6 month of age with severe acute malnutrition is different from that observed in infants who are 1 to 5 months of age without severe acute malnutrition?