From Emmanuel Forsén: I haven't been able to find anything published on this issue. It takes a deliberate study design collecting height measures on kids in a program that otherwise use the combination MUAC for admission and 15% GoW for discharge, i.e deliberately decided against measuring height. Current research is mostly focusing on establishing a suitable discharge criteria connected to MUAC for admission, and WHZ is more or less out of this picture. WHZ has been dropped because MUAC is easier and a better predictor of mortality. 15% GoW has it's evident weaknesses, and I think we will soon se published material underlining this, that is in the coming months.

The advantage of % weight gain is that in can be used in programmes admitting children on MUAC and not measuring height. The problem with % weight gain, however, is that it leads to longer duration of treatment for the less severely malnourished children. This is the opposite of what we want (ie longer duration for the most malnourished). Arguably, this effect was predictable on theoretical grounds (less weight gain needed for the most malnourished to get 15% WG, faster weight gain among them) but has been confirmed by at least one study not yet published but presented at a WHO meeting in 2008. Another study now submitted for publication shows that you don’t have this undesirable effect if you use MUAC as discharge criteria for children admitted on MUAC. So this is the way to go for programmes admitting children on MUAC.

In response to a comment by Tamsin that MUAC is a better indicator of mortality, in our CMAM target area in Bangladesh, we see that SAM identification by W/H Z-score finds more than twice the number than by MUAC. Even though taking height/length takes more time and less mobile, majority of SAM cases will be missed in the community if we only stick with MUAC. Regarding discharge, we are seeing WHZ recovers faster than MUAC. Most of the time, gaining 15% weight takes the most time even though they have come out of the SAM criteria by MUAC and WHZ.

Dear Ehsanul, By using also WFH to select children, you will find more children, but these will not be necessarily in the high risk category. If you want to be on the safe side, better use a higher MUAC cut-off than introduce WFH. All this is based on anthropometry mortality studies using ROC curves, done more than 20 y ago… in particular in Bangladesh and more recently repeated with data from Senegal using WHO growth standards. Also note that using MUAC will facilitate high coverage, key to programme success. See: Briend A, Maire B, Fontaine O, Garenne M. Mid-upper arm circumference and weight-for-height to identify high-risk malnourished under-five children. Matern Child Nutr. 2012 Jan;8(1):130-3.

Beware of this sort of thinking. MUAC, W/H, visible severe wasting, and so-on are all surrogates or markers for complex and hidden processes. None alone or together really get to the underlying processes. Just because we define SAM as (e.g.) WHZ < -3 does not mean that a child with WHZ < -3 is severely malnourished and at greatly elevated risk of near-term mortality. W/H is biased by body shape. This means that in (e.g.) pastoralists living at low altitude and in warm climates W/H tends to select a lot more kids than alternative measures but in (e.g.) high latitude and cold climates W/H tends to select very few kids compared to alternative measures (it tends to call normal kids obese). It is, therefore, incorrect to say things like the "majority of SAM cases will be missed in the community if we only stick with MUAC" if you do not take into account context and, in particular, body shape. Currently available data indicates that MUAC is the better measure than W/H in almost all regards (see [url=http://www.brixtonhealth.com/FNB_27_3_2006_B.pdf]here[/url] for a review of many of the issues). It is very likely to be the case that a child with low W/H but high MUAC is at low risk of near-term mortality and in need of nutritional rehabilitation.

Just to note that I did not make the comment about MUAC being a better predictor of mortality, but was posting a response as the moderator on behalf of Emmanuel Forsen. ENN is currently engaged with Save the Children UK on a practitioner and expert consultation on anthropometric indicators of SAM, which has a particular focus on the current issues and dilemmas faced by organisations in the field. This consultation aims to gather practitioner data and experiences as well as to understand where there are gaps or inconsistencies in the published evidence-base. The consultation concludes in December, at which time we will be in a position to share a position paper and the outcomes more widely. We welcome inputs to this in the form of unpublished data or experience from the field. Please contact post@ennonline.net Thank you.

Tamsin, You will find that MUAC is a better predictor of mortality than W/H. You will also find the W/H is worse at this than MUAC, MUAC/A, MUAC/H, W/A, and H/A. That makes W/H the worst predictor of mortality. The literature is clear and consistent about this particular finding. Nothing is served (and maybe beneficiaries are harmed) by pretending that this is not the case. Mark

Thanks Mark. It was just a clarification, rather than an entry into the discussion. Many organisations seem to be grappling with discharge criteria at present, so in terms of the consultation, further information on what programmes are doing in the field and what their experiences/data are showing, is of interest. Best wishes, Tamsin