Weight-for-height and mid-upper-arm circumference should be used independently to diagnose acute malnutrition: policy implications

Summary of research¹

Location: Global

What we know: Overlap between mid-upper-arm circumference (MUAC) and weight-for-height Z-scores (WHZ) when assessing acute malnutrition (AM) prevalence varies by country. This has significant implications for programming.

What this article adds: A recent study examined the direction and degree of discrepancy between MUAC and WHZ of children aged 6-59 months in 1,832 anthropometric surveys from 47 countries, mainly in Africa. Overall, using MUAC or WHZ, 16.3% of children were identified with global acute malnutrition (GAM) and 3.5% with severe acute malnutrition (SAM). The proportion of overlap between the two indicators was 28.2% for GAM (15-38.5%) and 16.5 % for SAM (6.1-29.8%). Overlap for individual countries was especially low for SAM. The numbers of children diagnosed by either criterion varied dramatically by country: GAM varied from minus 57% to plus 72%. For SAM, in four of the 38 countries, less than 25% of severely malnourished children would be identified and admitted for treatment if a MUAC-only admission policy were being used. Overall, 41% of children were younger than 26.5 months and 61% were shorter than expected. For all countries examined, the discrepancies were not adequately explained by any single hypothesis. The authors argue that MUAC-only criteria may not be appropriate where WHZ deficits predominate, and in such contexts recommend that both indicators are used as admission criteria.

Background

In 2009, WHO estimated about a 40% overlap between mid-upper-arm circumference (MUAC) and weight-for-height Z-scores (WHZ) when assessing acute malnutrition (AM) prevalence; this is observed to vary by country. To test this, a recent study examined the relationship between MUAC and WHZ for admission to treatment programmes, since this has implications for programming cost, workload, case detection, coverage and treatment.

Methods

Anonymous data were collected from 1,832 anthropometric surveys with over 75 malnourished children from 47 countries in Africa (1,619), Asia (166), Central America (two) and the Caribbean (45) between 1986 and 2014 with children measured aged from six to 59 months. Eleven additional surveys from eight countries where fewer malnourished children were identified were also analysed. Most of the surveys used two-stage, cluster sampling. All surveys followed standard WHO methods for measuring weight, height and MUAC. Indices were calculated using Emergency Nutrition Assessment (ENA) software for Standardised Monitoring and Assessment of Relief and Transitions (SMART).

The prevalence of global acute malnutrition (GAM) and severe acute malnutrition (SAM) was calculated using either absolute MUAC or WHZ (WHO 2006 standards). For each country, the total number of children diagnosed as acutely malnourished by either criterion alone or by both criteria was summed from all the surveys conducted in that country.

Results

Of the original 1,404,396 children with plausible data in the 1,832 surveys, 0.49% were excluded for oedema and 1.4% were then excluded using SMART flags, leaving a total of 1,384,068 children. Most of the children (88.1%) were from an African country. Key findings were as follows.

Overall, 16.3% of children were identified with GAM by either WHZ<-2SD or MUAC<125 mm and 3.5% were identified as having SAM by either WHZ<-3SD or MUAC<115 mm. The proportion of overlap between the two indicators was 28.2% for GAM and 16.5% for SAM, with analysis of all the children from surveys with more than 75 malnourished children. The degree of overlap ranged from 15.0% in Sri Lanka to 38.5% in Sierra Leone for GAM and 6.1% in Sri Lanka to 29.8% in Mozambique for SAM. For the 47 individual countries, the degree of overlap was consistently low (GAM: 29.9±15.3%, SAM 16.0±5.4%, mean±SD). The overlap was much smaller for SAM than for GAM.

There were slightly fewer countries in this analysis that had a higher proportion of children malnourished by MUAC-only than by WHZ-only (GAM 19 vs 28 countries; SAM 18 vs 20). The numbers of children diagnosed by one criteria or the other varied dramatically from one country to another.

• For GAM, the difference ranged from minus 57% to plus 72%; thus, in 11 countries, more than 75% of malnourished children would be identified using MUAC-only criteria, whereas in nine countries, including the Philippines, Sri Lanka and Senegal, less than 25% of malnourished children would be selected if only MUAC were used as the admission criterion.
• For SAM, the difference is even more dramatic. MUAC would not identify more than 75% of severely malnourished children in any country in which more than 75 SAM children were identified. In four of the 38 countries, less than 25% of severely malnourished children would be identified and admitted for treatment if a MUAC-only admission policy were being used (13% would be identified in Sri Lanka and 14% in Senegal).

Overall, only 41% of the children were younger than 26.5 months (the proportion never reached 50%); however, 61% of the children were shorter than would be expected for a child of 26.5 months growing normally. There is a tendency for there to be fewer children diagnosed as GAM by WHZ when there are more short children. The regression is significant (r2 =0.19, P<0.01, y=67.5– 0.14×). There is no relationship between the age distribution of the children and the relative importance of WHZ or MUAC for diagnosis of GAM (r2 =0.00).

Discussion and recommendations

The authors discuss a number of potential hypotheses and outstanding questions regarding the results.

Firstly, shorter or younger children are more likely to fall below the absolute cut-off point for MUAC. As the age categories did not differ significantly from one country to another, this does not adequately explain the different directions of the discrepancy observed.

Second, in countries where the children are more stunted, a higher proportion of children will have a MUAC below the cut-off point at any particular WHZ prevalence, simply because they are smaller. While there is a tendency for countries with more stunted children to have more diagnosed as AM by MUAC alone, the association is very weak, with only about 19% of the variation explained on this basis.

Third, absolute MUAC is less dependent on body proportions than WHZ, which may overestimate AM in children with a low sitting-to-standing height ratio (SSR) and underestimate those with relatively short limbs (legs weigh less than the torso). The study data does not support the explanation of variations in limb length accounting for the discrepancies between predominantly WFH or MUAC criteria.

Fourth, many studies have documented ethnic differences in fat distribution or ‘patterning’ in normally nourished populations. The effect of malnutrition on the relative loss of fat from the limbs and trunk and proportional loss of muscle from various body muscles is unknown. Thus, muscle and fat losses may affect MUAC and WHZ differentially.

Fifth, different population body shapes (endomorphic, mesomorphic and ectomorphic) might explain only some of the discrepancy observed.

The relationship between MUAC and WFH is complex. It is probable that the factors outlined affect some of the populations but not others; in combination they generate the discrepancy. More understanding of the factors at play is needed before a decision is made to abandon WHZ as an independent criterion for the diagnosis of acute malnutrition.

The authors argue that the superior power of MUAC to predict mortality risk in children is a strong argument for MUAC-only admission criteria if it predicts the death of the same children that WHZ would identify. However, since the two variables appear to identify different children, this will not be the case and it might be more helpful to consider the two indicators as complementary and additional, a hypothesis supported by the higher risk of death of those children with both MUAC and WFH deficits (Isanaka et al, 2015). Furthermore, addition of other deficits such as a low height-for-age or weight-for age progressively increase the risk of death.

The authors ask whether recent longitudinal studies of mortality risk conducted mostly in Bangladesh and Malawi are applicable globally, given the variations observed in this study’s analyses; the move towards using MUAC-only criteria may be appropriate for some countries but not for others, such as in Asia, where WHZ deficits predominate. WHZ as an independent admission criterion should be maintained until mortality risks are adequately assessed. The authors also propose that all future anthropometric surveys, including national DHS surveys, should include measurement of both MUAC and WHZ (and oedema) and report the prevalence of GAM and SAM using both MUAC and WHZ.

Conclusion

For all countries examined, the discrepancies observed between the indicators were large and not adequately explained by any single hypothesis. The perceived need for humanitarian intervention can be affected by the measurement chosen to assess the prevalence of malnutrition, which will vary from region to region. The dramatic difference in prevalence between countries using the two diagnostic criteria will influence decision-making and the distribution of resources. The authors conclude that WHZ and MUAC are complementary indicators that should both be used independently to guide admission for treatment of malnourished children. Using WHZ-only or MUAC-only estimates of prevalence will underestimate the burden of acute malnutrition.

References

Isanaka S, Guesdon B, Labar AS, Hanson K, Langendorf C, Grais RF, 2015. Comparison of Clinical Characteristics and Treatment Outcomes of Children Selected for Treatment of Severe Acute Malnutrition Using Mid Upper Arm Circumference and/or Weight-for-Height Z-Score. PLoS One. 2015;10(9):e0137606.