This is a major debate in our field. MUAC and W/H measure different things. MUAC is a criteria-referenced indicator. This means that we pick MUAC thresholds that identify children with a high risk of an outcome such as near term death if left untreated (and a good chance of survival if left untreated). W/H is a norm-referenced indicator. This means that we pick thresholds that identify children that look different from the reference population without any reference to risk of a negative outcome. The problem with W/H is that the functional significance for a given value varies with body shape and body shape varies from place to place. This means that the functional significance of a W/H of a given value varies from place to place. Sahel pastoralists (e.g.) can have low W/H and be healthy (long legs lower W/H but are also a sign of prolonged good health and good nutrition). Andean agriculturalist (e.g.) can have high W/H and still be healthy (big chests are needed to capture enough oxygen but no enough to support long legs). In these examples both the NCHS and WHO references will classify "wasted" or "overweight" but neither classification is appropriate in terms of fitness or outcome. This renders W/H next to useless. W/H is not (as you have above) the "better descriptor of nutritional status". Where is your evidence for this statement? All the evidence that I have seen is that it is the worst "descriptor of nutritional status". Worse in terms of functional outcome / need than uncorrected MUAC, MUAC/A, MUAC/H, H/A, and W/A. In terms of functional outcome / need W/H performs little better than tossing a coin. The Sudanese data are likely a case of long limbs and short bodies in healthy children. W/H is selecting healthy children because healthy children in Sudan look wasted compared to the reference population. Also, less healthy children in Sudan look like the reference population (because they have short limbs due to stunting). So we have W/H picking the healthy and rejecting the sick. Where low W/H does not reflect body shape is when we have low W/H and low MUAC so we pick these kids up. Since the most at-risk children will also tend to be stunted with a "normal" W/H we would not pick them up with W/H. We could use H/A but that can be difficult to do accurately in many settings. As for MUAC "preferentially" identifying younger children ... this is judgemental. MUAC selects younger children than W/H but not "preferentially" or in a "biased" manner. It selects the children most at need of intervention. These happen to be younger children because these younger children are most at risk of SAM through a combination of poor IYCN and infection. Perhaps "preferentially" is correct because I definitely prefer an indicator that targets intervention at the most needy rather than fills up programs with the least needy. Any other behaviour is irrational. BTW (1) ... We can integrate CMAM so a prevalence threshold for intervention need not used. BTW (2) ... MUAC is also affected by body shape but to a very much lesser extent than is W/H.

As Nikki Blackwell pointed out ... This means that we pick MUAC thresholds that identify children with a high risk of an outcome such as near term death if left untreated (and a good chance of survival if left untreated) Should have read: This means that we pick MUAC thresholds that identify children with a high risk of an outcome such as near term death if left untreated (and a good chance of survival if TREATED) Oops.