The issue is more complicated than "short posture" in terms of height achieved. In Somali people (my research experience) it is low sitting to standing height ratio (i.e. long legs / short trunk) that reduces W/H and leads to misdiagnosis (i.e. normal child as wasted). It can go the other way ... in populations with large chest circumference and short limbs, typical of high altitude populations, there is a tendency for normal children to be diagnoses as overweight. This is part of the "obesity epidemic" in the US Latino population. There is suggestive evidence that narrow trunks my be responsible for misdiagnosis in some SE Asian populations. From a biology / nutrition / evolutionary perspective, ... pastoralists living in warm climates at low altitude, herding on foot in arid / semi arid regions, and practicing polygyny can be expected to have low sitting to standing height ratios. MUAC may also be associated with limb length but this does not appear to affect MUAC so strongly as to lead to much misdiagnosis. See: Myatt M, Duffield A, Seal A, Pasteur F, The effect of body shape on weight-for-height and mid-upper arm circumference based case definitions of acute malnutrition in Ethiopian children, Annals of Human Biology, 2009, 36(1):5-20 You can get this [url=http://www.brixtonhealth.com/MyattBodyShape.pdf]here[/url]. It is not a great deal of work to confirm or deny the body-shape hypothesis. Let me know if you would like to pursue it.

From Jen Peterson: Dear Mija – I have used W/H and MUAC in Guinea, Sierra Leone and Niger with Fulani children, and believe that they provide an adequate indication of nutritional status. In general, my understanding is that children under five have the capacity to grow at the same rate, regardless of their ethnicity. I was told those genetic factors kick in mostly at adolescence (ou bien? What do I know – I am an agronomist!!). What I have focused on in the field (where balances are inaccurate/unavailable and height boards impractical) is not so much the raw weight, but constant INCREASES in weight (no matter how badly balanced the scale is), or height (using home measurements on the wall of the hut). Three months without weight gain and/or height increases => health center. Oddly enough, the Fulani children I worked with in Niger were SHORT (50% stunting; although the general Fulani population is taller than other ethnic groups in Niger), whereas in Guinea they were really tall (and generally thin). In general I found Fulani malnutrition levels in Guinea, Sierra Leone and Niger were lower than other groups I worked with in all three countries, partially because they drink more milk products. We have the same issue here in Madagascar – many people believe that children are stunted (up to 70% in the central highlands) because of genetics. However, when we loads those kids up on diversified diets and load their pregnant moms up with improved diets, they grow like weeds! Some of them are as tall as their parents by age 10 – no kidding. It is a wonder to see. Could be scary by the time they are teenagers :>) Hope this helps! Cheers – Jen

Thanks very much both Mark and Jen. Basically, in the Peulh/Flani children I refer to we see pastoralists that we can therefore compare with the Somali children Mark is refering to, i.e. potentially overdiagnosed as wasted. Mark, do you have data on your Somalia experience published somehow so i could refer to that? Now to all of you and also for you Jen: What if we make these children settle (which happens in refugee situations sometimes, livestock will keep on moving though) and the children are separated for longer periods of time to these good protein and calcium resources, i.e. milk. In the literature I see these children having therefore more access to starchy food, fat, sugar. Will it make them as a group relatively shorter ? I.e. the long legs will be less long, and therefore proprtionally they would 'normalise' meaning trunk and legs (or sitting/standing ratio) become more comparable with other ethnic groups. This would mean one would find less overdiagnosis of wasting, correct?

Mija, There is a link on my previous post. Here it is again: [url]http://www.brixtonhealth.com/MyattBodyShape.pdf[/url] If you want the data then I can let you have that. I also have unpublished data from Somalia. You question is a difficult one. Here is some "hand waving" ... There may be some genetic effect since having long legs may be advantageous to pastoralists. Physical models of walking and running predict that longer legs allow faster natural walking and running speeds (i.e. the walking and running speeds requiring the minimum expenditure of energy) than shorter legs (see below for the physics). Faster natural walking and running speeds allow larger ranges with similar energy expenditures. Afar and Somali pastoralists herd on foot without dogs or horses. The ability to walk long distances efficiently may facilitate access to wider grazing ranges and allow larger herds to be maintained. This may lead to the development of a virtuous circle (see Figure 7 in the linked paper). In such a virtuous circle, ‘momentum’ could be checked by exogenous factors such as drought, livestock disease, and competition for grazing land or endogenous factors limiting limb length. Both Afar and Somali pastoralists practice polygyny and the association between wealth and child survival predicted by such a virtuous circle may bestow a considerable reproductive advantage upon long-legged individuals. Long-legged individuals may have long-legged progeny. This effect may take some time (i.e. generations) to disappear. The magical milk factors may be lacking but the genes and, physical environment (heat, altitude), and cultural environment (long legs desirable) will remain. We really need a human biologist to review this. Short note on physics of walking ... in case anyone misses high school physics and algebra ... a simple physical model of walking represent legs as a uniform cylinders of length L with each leg acting as a pendulum swinging from the hip. The walking speed requiring the minimum expenditure of energy (the natural walking speed) is related to the natural period (T) of the pendulum. The natural period (T) of a cylindrical pendulum of length L is: T = 2 * pi * sqrt((2 * L) / (3 * g)) where g is the acceleration due to gravity (9.80665 m/s2 at sea level). This is the time required for a complete swing of the pendulum which is twice the time needed to make a single step. The model assumes that step length (SL) is proportional to leg length (L): SL ~ alpha * L where alpha is a constant. The natural walking speed is thus: v = SL / (T / 2) Substituting for T and SL yields: v = (alpha / pi) * sqrt((3 / 2) * g * L) In this model: v ~ L So longer legs means higher walking speeds with less energy expenditure (important when herding on foot) Similar models for natural running speeds yield the same result. What a cleaver bloke that Isaac Newton was! Perhaps we need a physicist too.

Mark, ok I think that is interesting but......despite the formulas and the vey interesting physics here, my question still stands I think: Will the risk of overdiagnosis of acute malnutrition (in this case exclusively wasting) in Fulani children become reduced when they become more sedentary (as in some refugee camps)? Or did I miss the point of your explanation?

I think it is a very difficult question. We have to judge the relative contribution of genes, diet, and environment on body shape. All are clearly important. What you asking is the effect of altering diet alone. Who knows? If there is an effect it will probably take some time (years at least, maybe generations) to be important. I think that we need to be careful of thinking of a pastoralist diet as a feast of meat and milk. Here is a link: [url]http://www.brixtonhealth.com/SCS.pdf[/url] to a report that I wrote some time ago which contains some data on diet (and body shape) in a pastoralist group. It is unclear whether (a) sedentary pastoralists lose all meat and milk from their diet, and (b) whether the nutritional composition of a GFD in a refugee camp is very different from the usual diet. WRT (a), it seems to me that the Warusha people of Tanzania (many of whom are settled Maasai following a rinderpest panzootic in the 1890's - c. 90% of livestock lost - and loss of extensive grazing lands in Kenya to British colonials at about the same time) have not lost their long leg / short trunk body shape (or the Maasai language and mores). They do keep some cattle and so have more milk and meat in their diet than "pure agrarians". Personally ... I would avoid the most biased measure (W/H) and stick with the least biassed and more useful measure (MUAC). This effectively solves the misdiagnosis problem.

Hello Mija, I'm not an expert or anything, but perhaps you could solve this by getting a sampling of children and measuring their sitting height/standing height ratio and using a Cormic Index to help you to know if using W/H is appropriate or needs to be adjusted. Although MUAC does seem to be less sensitive to body shape issues, you may have programmatic reasons for staying with W/H rather than using MUAC. For example, if other agencies in the region are using W/H and you want to compare in the two areas, or you want to build a trend against data that used W/H. It appears the general approach right now is to do them both if you have the time and resources. Then if there is a question during the analysis, you will be sure to have the data you need. It will also build the body of knowledge on how MUAC relates to W/H. Hope that helps.

hi to add to the point about about diet, see the the report on Nigeron the FEWSNET website eg : http://www.fews.net/pages/livelihoods-country.aspx?gb=ne&l=en page 17 or 18 of the document titled 'Niger livelihood descriptions' explains that the diet of the Peuhls is in fact primarily cereal based. Regards Sonya,

Anonymous 618, Yes. The approach taken in my linked article could be used. It is just an elaboration on what you have above. A single survey of SMART type with a typical SMART sample size (maybe a little larger) would probably do. I think that doing both MUAc and W/H (for surveys) it fine, even neccessary. I worry about using W/H programmatically since it could make for a very big program filled with older and healthier kids squeezing the time we have available for monitoring and for the younger and sicker kids. I have seen the use of W/H damage coverage and cure-rates. Sonya, Thank you for the link. It is very useful. I have done some work with pastoralists (I have also lived in a pastoralist community in mid-Wales for about 20 years) and the material you link to tallies with my experience. For the bulk of the population, meat and milk are present but in small quantities with milk consumed more frequently than meat in all but the extreme dry periods. I think that it is useful to see livestock as a "cash crop" for livestock sales, leather sales, milk / milk-product sales, and residual carcass products (e.g. glues, bone meal, musical instruments, bone jewelry) rather than as the main food source.

Hello Mija, to all, I am not an expert, but I been work 4 years in Mali and in Niger and during a discussion with Pr Golden I pointed out that during SMART survey (excluding here the fact that probably the SMART is not suited for the pastoral populations and that a new more appropriate methodology is in court of test I think) during the hunger gap season we could find a very high rate of MAS, but two months later we could'nt catch them at all. This observation could point the fact of when the have again access to the food they seemed to have a fast recovery without treatment. It do not answer your question but it seems to me that besides the morphology of these populations we can also face a point with a physiological adaptation .... and who could have an influence also on the question of Mija? Thank you C.

We need to be careful about "fast recovery without treatment" since mortality associated with SAM (i.e. real SAM by MUAC) is high. A conservative estimate is that about 20% for children with SAM (i.e. real SAM by MUAC) will die. Mortality associated with low MUAC is consistent between cohorts from Africa and Asia suggesting that we can discount the "physiological adaptation" argument (if that means that SAM is not a particularly serious disease in particular groups). The "physiological adaptation" argument seems to me to be little more than an attempt to explain away the low predictive value of the W/H indicator in some populations. This is not my area of expertise ... has anyone ever demonstrated "physiological adaptation" other than with the observation that low W/H kids do not dying at an expected rate in some populations. The alternative epidemiological / ethnomethodological explanation (i.e. W/H has a low and variable predictive value) seems more compelling. It might be the case that you don't find the SAM cases two months later is because many have died. Death is nature's way of lowering prevalence. Note : I am not necessarily contradicting Professor Golden. He is a proponent of W/H. Using this indicator it is likely that many more children recover spontaneously as the positive predictive value (i.e. of death) associated with W/H is worse then any other anthropometric indicator in common field use.

Hi Mark, I miss express myself and be sure that I'm not promote a "fast recovery without treatment" and I understand well the relation between mortality and SAM. I just mentionned a exemple regarding the fact that those children was diagnoses acute malnourish two month before by the W/H or by MUAC (important to know that in the same month we put in place two surveys one was vulnarability survey (food security) using MUAC and after we confirm our result with the SMART) but two month later when we pass again for the blanket -the blanket was to late I know- and looking for the same child we foud this child but the were not malnourish anymore (or at risk)... But your right we need too push more the investigation about that and this small sample size are not representative of all the pastoral population. Thank's C.

Chantal, Thanks for the clarification. Very interesting observation. How did you manage follow-up with transhumant pastoralists?

Good question.... And I think I have to foward back to the monitoring of this blanket (2010) to give you more information. I remember that we found almost 70% of the child where more the origin of "sedentary" pastorals the other 30% was really related to transhumant pastoralist people.... This "ethnie" are really interesting for me and I think they will really develop some psyshiological "capactity" over the year, wich are not in contradiction with acute malnutrition at certain moment or high rate of mortality.... C.

From Mike Golden: Mark states "I am not necessarily contradicting Professor Golden. He is a proponent of W/H." This misrepresents my position. Since 1994 we have advocated to use both W/H and absolute-MUAC as admission criteria. Certainly for screening there is no reasonable alternative. Advocacy for the use of MUAC as an independent criterion for admission is no longer necessary - the arguements are long finished. The old NCHS w/h standards discriminated against short (young) children - much more when using Z-scores than %median. The new standards are very much better in this respect - but not perfect when compared to the Prudhon index of risk of death. There is an exponential reduction in risk of death with age from birth: younger children have a much higher risk than older children nomatter their anthropometry. The use of absolute MUAC is biased (appropriately) towards youger children and against older children. W/h does not have this effect - and is appropriate to use as the admission critierion for older/taller children. The age profile of malnourished children is mainly the less than 24 month old children - for these MUAC is appropriate and will pick up the bulk of children. When survey data are used to examine mortality risk the analysis should be by height/age category or weighted to match the age distribution children admitted for malnutrition otherwise the analysis is flawed. In conclusion - we should retain w/h as an independent category for admission - there are problems with both indicies. W/H has the body shape problem - absolute MUAC has problems with short children (<60cm) and tall children - and although it is a simple measure - it is not easy to take accurately and precisely -there needs to be research on this aspect of MUAC measurement

Regarding the question posed by Mike Golden on the reliability and precision, our group in Kilifi, Kenya recently published a report where we found MUAC to be reliably and accurately measured by Community Health Workers compared to WFLz among infants. You can read more on http://onlinelibrary.wiley.com/doi/10.1111/j.1365-3156.2012.02959.x/abstract Martha

The issue of error in MUAC measurement is often trotted out as an argument against the use of MUAC. This was, I think, first asserted (as a self-evident truth and without supporting evidence) by Waterlow (1972). Evidence supporting this assertion has remained elusive. It is interesting to review the evidence since it tells a very different story. W/H has been consistently found to be more prone to error (when we include measurement, recording, and calculation error) than MUAC. The paper from the Kilifi team (see Martha's post) is recent evidence of this. It is worth adding that error with MUAC is greatly reduced when using colour-banded MUAC straps. We do this everywhere now.

Hi everyone in this discussion, it brings to mind something to share; currently we have switched to using only MUAC as the admission criteria (in addition to discharges from OTP) into the TSFP program currently running in Warrap state south sudan instead of both MUAC and W/H Z-scores. There is an impression that in the geographical context, W/H Z-score criterion captures a lot of children even those who are not moderately malnourished because the children here are quite tall (the population is Dinka). Though we have just shifted to MUAC admission criterion only, I think it is a good move because using Z-scores may in a way lead to very large numbers of children being admitted where some of these may not necessary require the current expensive preventive measures like TSFP but other less-expensive preventive livelihood improvement measures. Can anyone share their experiences to using only MUAC only criterion in TSFP?

hi martha, here at ALIMA we would be very interested in knowing your MUAC admission and discharge criteria for your TSFP also in your program - the average length of stay? and, even though you are using MUAC in and MUAC out if you have an idea of the average wt gain / kg / day in your program thank you nikki

Hello Nikki, I am not in position togive you the information you seek as we have just started using the MUAC-only criteria.