Menu ENN Search
Language: English Français

WFH versus MUAC

This question was posted the Assessment forum area and has 21 replies. You can also reply via email – be sure to leave the subject unchanged.

This question has been closed, so you cannot submit a new reply. Recommended answers have been marked with a star.

Anonymous 81

Public Health Nutritionist

Normal user

4 Sep 2015, 12:44

I would like experts input in this regard. I wish Mark Myatt to be one of the respondent of my question. Much has been said about the discrepancy of MUAC and WFH in some population particularly in South Sudan (Dinka and Nuer), Ethiopia (Gambella and Somali) and Somalia. I don't have problem with this findings of discrepancy between MUAC and WFH. As a result of this findings, there is strong push to use MUAC alone in these population. I want quote the following recommendation given to south Sudan, "MUAC alone is appropriate for admission in CCM/OTP in south Sudan". So my question is, is there evidence that shows those children <-3 SD WFH but normal MUAC have no risk? Are you still confidence those excluded children are not malnourished? if these excluded children (<-3SD) are grow in normal environment, do you think they will be still <-3SD?

Andrew Seal

UCL and NIE Regional Training Initiative

Frequent user

5 Sep 2015, 11:22

A very well focussed question and I am afraid the answer is that the nutrition community does not have the data to fully answer it.

What we do know is that:
Low MUAC is a good predictor of risk of death
MUAC is a very useful screening and admission tool and should be used in SAM treatment programmes
Weight for Height (WH) will identify some children who are malnourished (by definition) but who do not have a low MUAC
if you use WH admission criteria (either instead of or in addition to MUAC criteria) you will admit more children into your programme
It is not a good idea to have separate community screening and programme admission criteria (e.g. 2 different MUAC cut-offs)
Keeping programme protocols as simple as possible is good

What we do not know is:
What is the risk level for children with low WH but not low MUAC?

An interagency consultation meeting was organised by ENN to address the question you raise in 2012.
"A consultation of operational agencies and academic specialists on MUAC and WHZ as indicators of SAM"

The meeting Recommendation 1 was:

"At community level, there should be active case finding using MUAC to identify children requiring management of SAM

At health facility level (fixed or mobile), there should be systematic case finding using MUAC to identify children requiring management of SAM. If a child is not identified by MUAC, WHZ should be measured where it is feasible (capacity in terms of materials, time and trained human resources) without jeopardizing other essential health services; WHZ should be measured in particular where there are relevant clinical conditions, visible severe wasting, maternal concern and/or contextual factors (e.g. acute or prolonged emergency where more older children are affected)."

It is worth noting that dual anthropometric admission criteria have been advocated for for a long time:
"MUAC is a suitable tool for initial screening but admission to the feeding programmes should
be based on W/H. However, because presence of bilateral oedema or a MUAC < 110mm
indicate an increased risk of death and acute malnutrition, these attributes also justify
admission to a TFP, regardless of W/H." MSF Nutrition Guidelines from (1995)

I hope the above thoughts are useful. The full consultation report is available on the ENN site.
Good luck with the decision making process.

Sameh Al-Awlaqi

Public Health and Nutrition Consultant

Normal user

5 Sep 2015, 17:52

Hi there,

Yes,it is an interesting question,we don't have a clear answer so far. Based on WHO updates on the management of severe acute malnutrition in infants and children 2013(can be found on this link;it is similar to my colleague Andrew mentioned earlier( community level using MUAC/Bilateral Oedema, and facility level using MUAC OR WHZ OR Bilateral Odema).

Currently I have observed a regional trend in some African/Asian countries and NGOs to use MUAC as a sole admission criteria. MSF has published an article on Field exchange 50th edition recommends MUAC-based (and oedema) programming for most of its emergency responses, with ongoing review. But it also emphasizing the importance of conducting researches to set up appropriate MUAC cut-off points and how to reduce the risk of missing SAM children based on WHZ if we consider MUAC alone.

In Yemen the three admission criteria are used to admit a SAM child,while in Sudan there is almost a complete shift to use MUAC as the only admission and discharge criteria. As mentioned earlier by Andrew;MUAC is a predictor of mortality and it has a good sensitivity during emergency,cheap and easy to use.However,it is not considered as a measurement of wasting,we use W/H as a wasting indicator in 6-59 months child and the link between MUAC and WHZ is not fully understood besides the possibility of missing WHZ<-3 by using MUAC only.

The question comes again here: What is the risk of missing WHZ<-3 SAM children if we use MUAC only?

I think we need to study and research this question in our programming ,I have observed some of SAM children were missed depending on MUAC only,you can easily note some children with MAM cut-off points if their weight is measured against height;we can discover SAM<-3.

In 2014 a field study in Cambodia has been conducted and published on PLOS one nutrition and ENN 2014 regarding the optimal screening of children with acute malnutrition ( can be found here ; recommends that both of MUAC and WHZ should be regarded as an independent from each other for admission. Also,it indicates the use of MUAC -only cut-off points < 115 would have missed 90% of severely acute malnourished children based on WHZ <-3;reversely,WHZ-only will miss 80% <115 mm MUAC! They are also recommend the use of 133 mm MUAC cut-off points screening at community level to refer children for a second stage WHZ at facility level.

I am sure we will find a very useful inputs to this topic from our experienced colleagues around the world.


Mark Myatt

Consultant Epideomiologist

Frequent user

6 Sep 2015, 11:53

Since you ask ...

My response will be from the perspective of an epidemiologist.

A clinician may focus on treatment efficacy. The efficacy of the CMAM protocol can be defined as how well the protocol works in ideal and controlled settings. It is measured by the cure rate:

    Cure rate (%) = (Number cures / Number treated) * 100

This is usually estimated in a clinical trial.

An epidemiologist will focus on effectiveness at both the individual and at the program level.

For the CMAM protocol, the cure rate (treatment efficacy) is close to 100% in uncomplicated incident cases (i.e. in cases with MUAC at or just below the admission criteria and cases with mild oedema). There is, therefore, little room for large improvements in the efficacy of the CMAM protocol.

We cannot significantly change the efficacy of the CMAM protocol. We can, however, change the effectiveness of the CMAM protocol.

The effectiveness of the CMAM protocol can be defined as "the cure rate in a beneficiary cohort under program conditions". Effectiveness depends, to a large extent, on:

[b]Severity of disease :[b] Early treatment seeking and timely case-finding and recruitment of severe acute malnutrition (SAM) cases will result in a beneficiary cohort in which the majority of cases are uncomplicated incident cases. The cure rate of the CMAM protocol in such a cohort is close to 100%. Late treatment seeking and weak case-finding and recruitment will result in a cohort of more severe and more complicated cases. The cure rate in such a cohort may be much lower than 100%.

Compliance : Programs in which the beneficiary and the provider adhere strictly to the CMAM protocol have a better cure rate than programs in which adherence to the CMAM protocol is compromised. Poor compliance can be a problem with the beneficiary (e.g. sharing of ready-to-use therapeutic food [RUTF] within the household) or a problem with the provider (e.g. RUTF and drug stock-outs), and both have a negative impact on effectiveness.

Defaulting : This is the ultimate in poor compliance.

An effective program must, therefore, have:

(1) Thorough case-finding and early treatment seeking. This ensures that the beneficiary cohort consists mainly of uncomplicated incident cases that can be cured quickly and cheaply.

(2) A high level of compliance. This ensures that the beneficiary receives a treatment of proven efficacy.

(3) Good retention from admission to cure (i.e. little or no defaulting). This also ensures that the beneficiary receives a treatment of proven efficacy.

Coverage is one factor (the other being effectiveness) in the capacity of a program to meet need. It can be expressed as:

    Coverage (%) = (Number in the program / Number in need) * 100

Coverage depends on:

(A) Thorough case-finding and early treatment seeking. This ensures that the majority of admissions are uncomplicated incident cases, which leads to good outcomes.

(B) A high level of compliance. This ensures that the beneficiary receives a treatment of proven efficacy.

(C) Good retention from admission to cure. This is the absence of defaulting.

Note that that the "123" of achieving high individual effectiveness and the "ABC" of achieving high coverage in CMAM programs are the same.

Program effectiveness (met need) is:

    Met need = effectiveness * coverage

Coverage and effectiveness depend on the same things and are intimately linked to each other. Good coverage supports good effectiveness. Good effectiveness supports good coverage. Maximizing coverage maximises effectiveness and met need. An effective program is built upon individual effectiveness and coverage.

Let us see how this plays out. We can look back at the CMAM vs. TFC debate of the previous decade. At the time this was a heated debate with many harsh words spoken. I think we can be more measured ...

In terms of clinical effectiveness the high dependency inpatient therapeutic feeding (TFC) model does very well. Cure-rates of 90% were achievable even with very severely malnourished caseloads. Weight velocities were also high (e.g. > 10 g/kg/day). This good performance was achievable without high coverage because care was delivered in high dependency units (HDU) using appropriate treatment products and protocols. It seems likely that the low coverage of TFC programs meant that they tended to treat mostly very severe cases and that TFC proponents believed that the case they saw were typical SAM cases and that all SAM cases needed treatment in HDUs. This is an example of drawing wrong inferences from a self-selected sample.

In comparison, CMAM programs do not look so good. Cure-rates may be lower (e.g. 80%) and weight velocities are typically half that achieved in TFCs even when the severity of SAM in most cases is not as sever as those seen in TFCs.

From a purely clinical perspective the TFC would be the program of choice. This conclusion changes when we look at the problem from an epidemiological perspective and factor in coverage.

TFC coverage was usually low (i.e. typically between 5% and 15%. The early CTC (CMAM) programs achieved coverage of 70% to 80%. If we take the best coverage for TFCs and the worst coverage for CMAM programs we get:

    Met need (TFC) = 15% * 90% = 13.5%
    Met need (CMAM) = 70% * 80% = 56.0%

The clinically inferior program turns out to be four times more effective that then the clinically superior program. Even "bad" CMAM programs with (e.g.) 75% cure rates and 20% coverage (met need = 15%) outperform the best TFCs. All but the very worst CMAM programs outperform the best TFC programs.

I have gone into this detail to make the point that to make the right choice in terms of delivering effective programming requires us to consider more than one factor. In the TCF vs. CMAM debate (e.g.) narrow clinical considerations compounded by inferences drawn from self-selected samples would have led to us making the wrong decision. We would still be treating and curing tens of thousands of children rather that hundreds of thousands of children.

I also wanted to make the point (a golden rule in public health medicine) that coverage is the most effective route (the "Royal Road") to program effectiveness. Any discussion of CMAM admission criteria needs, therefore, to consider coverage very carefully.

There is considerable evidence favouring MUAC as the best criteria for admission into CMAM programs. This has been covered on this forum and elsewhere. In brief, MUAC is the better predictor of mortality, low MUAC children respond well to treatment with the CMAM protocol, and MUAC is cheap and practicable. The data are reviewed here.

I will suspend my disbelief for a moment and treat MUAC and WHZ as equally valid case-definitions for SAM.

The practicability of MUAC means that case-finding with high spatial and temporal coverage of screening is possible. This is something that is very definitely not feasible with WHZ. The ability to achieve high case-finding coverage is, IMO, a key reason to favour MUAC as a primary admission criteria. Many proponents of "strong" mixed MUAC / WHZ programming ignore this. They often take SMART survey data and apply case-definitions and tabulate MUAC cases against WHZ cases and find a number of "discordant" cases (i.e. cases with WHZ < -3 and MUAC > 114 mm). They then point to the discordant cell and say "We must use WHZ on an equal or superior footing to MUAC". Such an analysis leads us into error as it assumes that case-finding and recruitment is 100% exhaustive (i.e. has 100% coverage) or is equal for both MUAC and WHZ (equality or exhaustiveness may be present in very bad programs where it approaches zero - I have seen programs like this!). 100% or equal coverage of screening is extremely unlikely to be the case. Looking at data that I have to hand regarding a bottleneck analysis of MoH-run CMAM programming in a west African state I find that coverage of screening by MUAC of 50% per quarter (i.e. 50% of children have their MUAC taken in their homes in any three month period) is routinely achievable. Such a figure is not achievable for WHZ outside of special settings (e.g. well ordered camps). It is, in fact, unlikely to exceed a very small percentage. If we take coverage into account then the number "discordant" cases (we might call this "WHZ cases missed by MUAC") is greatly diminished. That is, if you properly account for the practicability of what is being proposed then the problem can bee seen in proper proportion.

Also, we must not ignore the impact of the considerable additional cost and staff required to achieve anything but very low screening coverage for WHZ. What tends to happen is that the extent of screening by both MUAC and WHZ contracts as WHZ eats the resources available for case-finding. Coverage of screening in "strong" mixed programming suffers and the programs ability to meet need also suffers. Again, we must consider the practicability and practical consequences of what is being proposed.

I think a "strong" mixed MUAC / WHZ program cannot be effective and that introducing WHZ into MUAC programming can be counterproductive.

We might want to consider using a "weak" mixed MUAC / WHZ programming model. Such a model would concentrate on MUAC-based community-based case-finding. It would also use MUAC in all health service contacts (e.g. at clinics and during EPI extension activities, IYCF activities, &c.). WHZ could be applied in clinical facility settings. If clinical staff saw MUAC below 135 mm they could measure weight and height and calculate WHZ. Any child with MUAC < 135 mm and WHZ < -3 would be referred to / admitting into CMAM services. In my experience care would need to be taken to avoid WHZ taking resources away from community-based screening, sensitisation, and mobilisation activities. You need to note that height boards are not usually available at many health facilities and are not part of essential clinic supply packs. Also, height / length measurement is not part of the IMCI syllabus. We must be careful that we do not allow the (vestigial) WHZ tail to wag the MUAC dog.

All these words and still no direct answer to your question!

Let me try ...

We know that MUAC is a universally good predictor of near term mortality.

We know that low MUAC kids respond well to treatment with the CMAM protocol.

We know that WHZ is a poor predictor of mortality. The clear evidence is that it performs worse than MUAC, MUAC/A, MUAC/H, H/A, and W/A. That is, it performs worse than any other commonly used anthropometric indicator. We also know that the predictive value of WHZ ranges between worse than tossing a coin to only a little better than tossing a coin.

We know that WHZ is strongly influenced by body shape. In some settings, low WHZ is associated with good H/A which suggests that low WHZ may be predictive of survival, good educational performance, high work productivity, &c. The body shape problem means that WHZ is sensitive but not-specific in warm climate / low altitude / pastoralist populations (i.e. we may treat many children who do not need it). It also means that in cold climate / high altitude settings WHZ has low sensitivity by high specificity (i.e. we fail to treat many children who need it).

We do know that MUAC and WHZ are quite strongly correlated with each other.

When we consider this in the round it is reasonable to surmise that much of the (weak) association between WHZ and mortality may be due to low MUAC. This would means that many of the WHZ+ / MUAC- children will be at lower risk of near term mortality than the MUAC+ children. I think there is some reason to believe that the discordant cases [b]taken as a group[b] are not at elevated mortality risk (a couple of studies do shows this). I would hesitate to conclude that no member of the discordant group is at elevated mortality risk. I think that many of the discordant cases at elevated risk may be picked up by WHZ at health facility level in the "weak" mixed MUAC / WHZ programming model. This will be, in large part, due to them being sick enough to prompt attendance at a clinical facility.

This article reviews the option of adding WHZ to MUAC only programs and concludes that to identify high-risk malnourished children, there is no benefit in using both WHZ less than –3 and/or MUAC less than 115 mm, and that using MUAC alone is preferable.

If I were very concerned about missing children I would increase the MUAC admission threshold rather than use WHZ.

At the risk of boring you ... I have started to notice articles of this type being touted as evidence for the need for a "strong" mixed MUAC / WHZ program. This is very weak evidence as it attempts to draw inferences from a self-selected sample (see the TFC vs. CMAM discussion above). This is something that should only be done with much caution and, when done, all efforts should be made to judge the direction and magnitude of the biases inherent in the sample. To give an example, in the "weak" mixed MUAC / WHZ model we would expect the low WHZ cases found at clinical facilities to be sicker than other cases because they were recruited as sick children attending at clinical facilities. If we looked at our patient cohort we would not be surprised to see high mortality in these children. This is an effect of sampling bias. When we see this it would not be legitimate to conclude that we must use a "strong" mixed MUAC / WHZ model.

I hope this is some help.

Sameh Al-Awlaqi

Public Health and Nutrition Consultant

Normal user

6 Sep 2015, 13:36

Hi Mark,

Thank you for this feedback,actually this is an article per se, I need to read it again to capture all information and reference you mentioned! (Thanks for your amazing efforts).

Please allow me to express my initial comments/questions to your answer:

1. I do agree with you that MUAC is important and more practical in community settings.
2. MUAC is very good for increasing coverage,which will be reflected on program effectiveness and the "met need".

I think,in my humble opinion, as a nutritionists we need to think in both directions: Coverage, and detection of as many as we can SAM children based on MUAC/WHZ. Using the weak model you have suggested can improve case finding in community very much,but again, we need to consider individual setting and community,qualified staff and availability of health facilities.The setting of 135 mm or less cut-off points for referring children for second stage screening could minimize the missing of SAM children based on WHZ.

Let me ask a question here:

When MUAC is used as a sole admission and discharge criteria,even in a facility-based nutrition program, what is the significance of missing SAM WHZ? To what extent this will affect their health and nutrition status? Do we need to wait until they have complications and have (acceptable MUAC cut-off point)to be admitted into the program? This is my concern; can MUAC replace WHZ permanently without an effect on SAM children at community/facility level?

I am asking this because I am observing some calls to consider MUAC as a sole admission and discharge criteria in various settings(including health facilities).

I have observed that some SAM children can acquire weight and achieve target WHZ faster than target MUAC, and vice versa, this could have some linkage to certain geographic localities(Asia,Africa), but I think this subject needs more field researches.What do you think?

Thanks for your inputs,




Frequent user

6 Sep 2015, 15:09

Dear Sameh,

A few comments about this recurring discussion about MUAC and WFH.

There is general recognition now that MUAC is more closely associated with the risk of dying than WFH.

Also, it has been shown repeatedly for at least 30 y that MUAC and WFH do not select the same children.

It is unfortunate that many people overlook that these two points are closely linked. If MUAC identified the same children as WFH, it would not be superior in assessing the risk of death.

There are several explanations about the relevance of MUAC to assess the risk of death. The explanations are compatible and can be true together. Compared to WFH, MUAC selects preferentially young children with some degree of stunting who are more at risk than those with only wasting as risk factor. Also, MUAC is directly related to muscle and fat mass, body compartments which can be used as fuel and source of nutrients when energy and nutrient intakes are insufficient. This contrasts with indices based on weight, as about 75% of body weight in children is related to organs which cannot be tapped as nutrient and energy stores.

This link with the risk of death made MUAC more and more popular for CMAM programmes in areas where preventing malnutrition related deaths is the priority. The possibility to use it in the community is an added advantage.

Now, are all children with an increased risk of death captured by MUAC < 115 mm ? The answer is clearly NO, as the relationship between MUAC and risk of death is continuous and regular, and in particular there is no discontinuity around MUAC 115 mm. A child with a MUAC of 116 mm is hardly better than a child with 115mm.

In this regard, have a look at the Myatt el al paper, especially figure 3:

A review of methods to detect cases of severely malnourished children in the community for their admission into community-based therapeutic care programs. Food Nutr Bull. 2006 Sep;27(3 Suppl):S7-23

So to be on the safer side, you could increase your MUAC cut off, as was done by MSF in several projects described in the Field Exchange issue 50 you refer to. Now if you do that, you will also include in your programme some children with MUAC > 115 mm and WHZ < -3.

How far should we go in increasing MUAC cut off ? This is a difficult question, as the number of children to include increases dramatically when you increase the MUAC cut off. And this can have negative effects on quality of care and programme coverage. Also it considerably increases programme costs. So any change of MUAC cut-off should be considered only when you already have a high treatment coverage of children with MUAC < 115 mm.

A “false good idea” is to increase MUAC to such level as to include all children with WHZ < -3. This in practice means increasing MUAC cut off to unrealistically high levels. The paper you refer to from Cambodia suggests that you need to go as high as 133 mm to reach this objective, which would put more that 50% of children in most populations in therapeutic feeding. The idea of having MUAC measured in the community and having all those below a high cut off (say 130 mm) referred to health centres and checked in terms of WFH has been proposed and tested many times in early CMAM days and we now know it does not work and that its main effect is to decrease treatment coverage. And it does not make sense either when the objective is to reduce malnutrition related mortality, as low MUAC children should get the priority and WFH is largely irrelevant.

I hope this helps

Mark Myatt

Consultant Epideomiologist

Frequent user

6 Sep 2015, 16:28

WRT "The setting of 135 mm or less cut-off points for referring children for second stage screening could minimize the missing of SAM children based on WHZ". This has been tried ans the effect has been to damage coverage. This was such a common problem in teh early days of CMAM that is was a name problem (i.e. "the problem of rejected referrals"). It was estimated (see here) that almost 40% of coverage failure was due to this problem alone. My advise is to avoid this course of action.

If I had serious concerns about missing cases I would ...

(1) Put a great deal of effort into achieving high spatial and temporal coverage or screening by MUAC by having MUAC taken at every possible contact (in extension services such as EPI, IYCF, GMP, &c. and at all clinical contacts) and by community mobilisation and sensitisation with MUAC taken by community-based volunteers and my mothers using ALIMA's "Mothers Understande and Can use it" method.

(2) Place CMAM services as close as possible to the population using health posts, CHWs, and mobile teams.

(3) I would avoid using WHZ as it is very difficult to avoid it damaging coverage.

(4) Look to increasing the MUAC threshold to 125 mm and having a lower intensity intervention (less RUTF, fortnightly contact, &c.) for the 115 - 124 mm cases. I'd pay a lot of attention to picking up non-response in this group.

I fear that by concentrating on discordant cases which, as far as we know, are not at very elevated mortality risk will direct efforts away from treating MUAC < 115 mm cases which we know are at high mortality risk.

Sameh Al-Awlaqi

Public Health and Nutrition Consultant

Normal user

6 Sep 2015, 19:01

Thanks André and Marc for your input.

I do agree with what have been mentioned by André and the question of "How far should we go in increasing MUAC cut off ?" Taking into consideration its impacts and the consequences of having a high standard level MUAC cut-off points.

It is true that the rejected referrals is one of the major reasons behind low coverage and non-attendance in CMAM sites, but it is not only WHZ rejected referrals, it can be MUAC also,especially when the MUAC reading is near 115mm, as health workers have to repeat it to confirm the reading before admissions.
So we are expecting referral-rejection in general,and referring to the article you shared "Determinants of coverage in Community based Therapeutic Programme" the way in which the rejections are handled is one of three common factors to non-attendance.

In addition to the rejection at health facility level,it is found during barrier analysis surveys in some areas in Yemen that prolonged waiting in OTP center for already admitted child is one of the major factors of defaulting,which was also applicable on prolonged waiting on screening areas that discourage the mother to bring their children for screening.

Working closely with the community and CHVs, crowd control at facility level and waiting area, and proper handling of rejection by health workers and clear messages provided to caregivers by CHVs before referral can reduce this rejection cases.I do agree with you regarding the importance of expansion of programme to cover as many areas as possible and to take the advantage of every possible contact with children to get MUAC reading can increase coverage and include more children at risk of dying.

Including mothers in screening is very good ideas especially areas with no CHVs(Community Health Volunteers), in the past I was remembering giving a MUAC tape for few number of mothers of discharged children to follow her children's nutrition status , given the fact the 2-3 months stay in programme allows the mother to observe many anthropoemetric techniques and to apply MUAC screening for her child, it works well and some children was referred as a relapse during conflict period,some of them referred when the child reached to MAM cut-off points ( Yellow zone), so the child was brought for confirmation of this finding and weight measurement.Understanding of caregivers of what these anthropoementry means is one of the causes toward a good acceptance of nutrition services by the communities which will be reflected positively on programme coverage.

I don't think that concentrating on " discordant cases" will direct efforts away from treating <115mm rather than letting us to treat SAM children and to reduce the risk of death and mortality taking into consideration a good coverage programme.

In this regard, I would like to consult you( André and Mark and other colleagues) in one idea: Do you think if we can analyse a current MUAC-based program from 2012- 2015 in certain locality could help us to understand the magnitude of these discordant cases and to understand more WHZ trend in these children ( luckily the data on WHZ is available in OTP cards).

Thanks for your valuable inputs,appreciated very much.

Kind regards,


Anonymous 81

Public Health Nutritionist

Normal user

7 Sep 2015, 04:14

Dear Andrew Seal, Sameh, Mark, and André
Thank you very much for your valuable comments. My question is very specific to certain population. However, the response seems general. So, I am requesting Mark and André to reply only specific for long leg population. In this discussion, rejection of admission was repeatedly mentioned. So, don’t you think the same case of rejection could be there for majority of children with physical wasted and <-3 WFH but normal MUAC.
My last question is, are you rejecting the six recommendation that were decided in 2012 meeting In London. One of the recommendations shown below.
Recommendation 1:
At community level, there should be active case finding using MUAC to identify children requiring management of SAM
At health facility level (fixed or mobile), there should be systematic case finding using MUAC to identify children requiring management of SAM. If a child is not identified by MUAC, WHZ should be measured where it is feasible (capacity in terms of materials, time and trained human resources) without jeopardizing other essential health services; WHZ should be measured in particular where there are relevant clinical conditions, visible severe wasting, maternal concern and/or contextual factors (e.g. acute or prolonged emergency where more older children are affected).


Frequent user

7 Sep 2015, 08:08

Dear Sameh,

I am surprised by your comment about rejected cases with MUAC due to discordant measures between CHW (or mothers) and health workers at the treatment facility. This should not happen if CHW workers (or mothers) are adequately trained with the same MUAC tapes. And if good MUAC tapes are used. Also, you need to have some tolerance in measure discrepancies. It would be stupid if a child measured at 115 mm by a mother is rejected if the supervisor finds 116 mm. Please have a look at the figure 2 of the MUAC paper from Nikki Blackwell

You can see that discrepancies concentrate near 115 mm. Define in advance an acceptable range and you should avoid these rejected referrals which are so damaging for the programme (and also for families).

The issue of discrepancy between MUAC and WFH has been discussed in nearly all meetings on SAM and CMAM for the last 10 years or more and this debate is incredibly repetitive. I have nothing against an analysis of your OTP data to look at what happens in your settings, but I wonder what new information you will get from it. You most probably will find that MUAC preferentially selects young children usually wasted (not always below -3 z WFH) and with some degree of stunting and a slight excess of girls. But we have known that for years already.


Frequent user

7 Sep 2015, 08:14

Dear anonymous 81 (no other name ??)

I have no problem with the statement you quote from the 2012 meeting (which I attended). Please kindly note that the emphasis nowadays should be on community screening, and there is no mention of WFH measure at the community level in this statement. Please also note the careful wording regarding the use of WFH in health facilities. In particular, the additional workload should not be at the expense of other essential health activities.

I hope this helps

Mark Myatt

Consultant Epideomiologist

Frequent user

7 Sep 2015, 09:04

First on Sameh Al-Awlaqi's comments / suggestions ...

WRT barriers / bottlenecks ...

When we started CMAM the use of MUAC other than as the first stage of a two-stage screening process (i.e. MUAC then WHZ with only WHZ cases being admitted) was the subject of much debate. The two-stage screen with a referral to centre for the second stage meant very large numbers of rejected referrals. This was a coverage killer. Note that applying the two stage screen in the community does not have this problem but this proved difficult to sustain with high levels spatial and temporal coverage. Moving to MUAC admission criteria reduces the problem considerably. When you do this it is advisable to have a policy of admitting on referral (i.e. all referrals are admitted) and using a referral monitoring system to identify referrers making inappropriate referrals and needing further training / supervision. The case (e.g.) referred by a traditional healer that is rejected can cause coverage problems that are very difficult to overcome. It is sometimes difficult for providers to allow a "direct admission from the community" policy. Many agencies seem to need to take a "gatekeeper" role. One of the appeals of WHZ is that it keeps gatekeeping functions firmly in the hands of the provider rather than devolving this to the community.

Waiting times ... yes. Centre organisation is something you do have control over. It is often worth doing old-style "time and motion" studies to identify patient flow bottlenecks and work around them. I see a lot of programs that do not implement good queue management processes.

WRT mothers ...

The recruitment of "positive deviant" mothers (i.e. mothers who clearly understand and follow the CMAM protocol and help new mothers at centres) was a key part of the original CTC / CMAM. The ALIMA "MUAC" model takes this further.

WRT concentrating on discordant cases ...

Finding these cases requires you to use WHZ with the same spatial and temporal coverage as MUAC. Since WHZ is very resource intensive and needs to be done using program staff (e.g. mothers cannot do this) the cost of this will probably be at least an order of magnitude higher (i.e. > 10 times) then the cost of your MUAC screening program. Unless you can free resources, something will likely break. We tried this in the "CTC Research Program" (the operational research program that developed and tested CMAM programming models) and we found that it was expensive and could not be done with any useful level of coverage.

There can be some harm in even a "weak" mixed MUAC / WHZ program since moist facilities do not have the staff or equipment for height measurements and W/H calculations. Putting this in place can be expensive and time-consuming and, while you concentrate on doing that, other screening activities are likely to suffer.

The practicalities mean that this type of attempt to improve coverage usually damages coverage.

WRT a study in MUAC only programs ...

I think you need to be very careful with trying to draw inferences with self-selected samples. You clinic population will be such a self-selected sample. In the program you describe you will have MUAC+ / WHZ- and MUAC+ / WHZ+ but not have MUAC- / WHZ+ children. The use of survey data may be useful but you will need to weight numbers by some reasonable expectation of you WHZ screening coverage.

I am concerned about the use of terms such as "magnitude" without definition. If this is just a number of such cases then it is of little value. What is needed is the number of cases at substantially elevated mortality risk above baseline mortality risk. There is no compelling evidence to suggest that these discordant cases are at substantially elevated risk.

What might help us here is to get access to the original (1980s / 1990s) cohort data and to estimate mortality in MUAC+ / WHZ+, MUAC+ / WHZ-, MUAC-, WHZ+, and MUAC- / WHZ-. This will give us the evidence wee need to decide whether we need to concentrate resources on the discordant cases and to argue for the funds needed to do this in anything but a piecemeal and partial basis.

Now to Anonymous 81's post ...

Sorry to have no addressed the specific context sufficiently.

General (1) : There is no compelling evidence to suggest that the discordant (MUAC- / WHZ+) children are consistently at elevated mortality risk. We can investigate this using existing datasets. If you are very concerned about these cases then most are eligible for admission TSFP. NOTE: I have used "consistently" here because, unlike MUAC, the meaning of a given WHZ threshold in terms of mortality risk in not very consistent.

General (2) : CMAM is, I think, a child survival program. This means we want to use admission criteria that best identify children at near term mortality risk who can be successfully treated with the CMAM protocol. I am at a loss to understand why any rational person would pick the worst performing (i.e. WHZ) indicator for use in CMAM programs. It would be better using MUAC, HAZ, and WAZ before even thinking about using WHZ. If, however, the point of CMAM is to attempt to correct a type of thinness associated with the surface area to mass ratio then WHZ would be pretty good (and MUAC not so good). For child survival programming (and the raising of mortality risk in this thread suggests that this is what we all think we are doing) WHZ is a damaging and expensive distraction.

BTW: One of the nice things about MUAC is that it tends to pick up low HAZ and low WAZ children. This is one reason why it works so well fror our purposes.

Specific (1) : W/H and related measures such as BMI are known to be strongly affected by body-shape. "Body shape" includes, but is not restricted to sitting to standing height ratio (SSR) which is lower in long-limbed populations. The effect of low SSR is to yield lower WHZ. This is particularly true in the absence of stunting. This means that, in some children at least, we will see reduced WHZ in the healthiest and wealthiest. What this means is that we can expect, for some children - particularly the older children, lower mortality risk with lower WHZ. In your specific setting you can expect MUAC- / WHZ+ older children not to have elevated mortality risk. WHZ is generally useless in CMAM programming and is especially useless in this specific setting as it will fill your program with older and healthy children who are not at greatly elevated (above baseline) mortality risk.

General (3) : We can expect a similar effect in South Asian and SE Asian populations but here "body shape" might be narrow trunk widths / shoulders.

You ask:

    My last question is, are you rejecting the six recommendation
    that were decided in 2012 meeting In London. One of the
    recommendations shown below. 

    Recommendation 1:
    At community level, there should be active case finding using
    MUAC to identify children requiring management of SAM.

I wholeheartedly support this.

    At health facility level (fixed or mobile), there should
    be systematic case finding using MUAC to identify children
    requiring management of SAM. If a child is not identified
    by MUAC, WHZ should be measured where it is feasible (capacity
    in terms of materials, time and trained human resources)
    without jeopardizing other essential health services; WHZ
    should be measured in particular where there are relevant
    clinical conditions, visible severe wasting, maternal
    concern and/or contextual factors (e.g. acute or prolonged
    emergency where more older children are affected).

I support this. I do not think that WHZ has much value in CMAM programming. I would, if resources allowed, prefer to admit cases on clinical sign such as visible severe wasting and per-oedematous signs of kwashiorkor rather than waste time and money on WHZ. These signs do not need special equipment or three staff to measure. I wholeheartedly agree with the caveat "where it is feasible (capacity in terms of materials, time and trained human resources) without jeopardizing other essential health services". Since we usually operate in resource scarce settings this caveat will usually apply and we will not use WHZ.

General (4) : While I am agreeing ... André is correct ... we can have a slightly elastic boundary. It makes little sense, and may harm coverage, to send away a child with a MUAC of 115mm or 116mm.

I hope this is of some use.

Sameh Al-Awlaqi

Public Health and Nutrition Consultant

Normal user

7 Sep 2015, 14:37

Hi Anonymous 81,

Excuse me for taking the space allocated for answering your question by having a general discussion as you mentioned,but this question came on a right time for me. Thank you for raising this question.

Dear Mark and André,

Thank you for this very invaluable information and sharing your technical expertise;it helps a lot. I will keep observing the field work and will take your recommendations into consideration, I will keep in touch through this forum. Thank you again.

Best regards,


Mark Myatt

Consultant Epideomiologist

Frequent user

17 Sep 2015, 09:07

just FYI ... A new related thread has started here.

Benjamin Guesdon

Research coordinator / ACF-France

Normal user

2 Oct 2015, 09:55

Dear all,

At this stage, I think it would be helpful to review the hypotheses of André and Mark and the extent to which they are infirmed or confirmed by recently published evidence. I agree that this recent evidence is far from perfect, yet it is good sometimes to see how observations match theory.

So, on the one hand, we have the following hypotheses which have been systematically put forward on this forum:
- Children with a MUAC<115mm have a higher risk of death in the short term than children with WHZ<-3 (which is in fact quite different from the statement that MUAC is a better predictor of mortality than WHZ according to ROC curves see here for an objective explanation)
- Children with a MUAC<115mm are more in need for SAM treatment (short-term rehabilitation of acute nutritional deficits and appropriate management of clinical complications) than children with WHZ<-3
- Children with WHZ<-3 and MUAC=115mm (WHZonly) are in fact healthy older children with long legs

As expressed in the former consultation of experts mediated by EN-net it is important to highlight that the evidence supporting these hypotheses is far from being conclusive for everybody.

On the other hand, we have the following recently published evidence:

- Midupper arm circumference and weight-for-length z scores have different associations with body composition: evidence from a cohort of Ethiopian infants.
Grijalva-Eternod CS, Wells JC, Girma T, Kæstel P, Admassu B, Friis H, Andersen GS.
Am J Clin Nutr. 2015 Sep;102(3):593-9. doi: 10.3945/ajcn.114.106419. Epub 2015 Jul 29.

- Inconsistent diagnosis of acute malnutrition by weight-for-height and mid-upper arm circumference: contributors in 16 cross-sectional surveys from South Sudan, the Philippines, Chad, and Bangladesh.
Roberfroid D, Huybregts L, Lachat C, Vrijens F, Kolsteren P, Guesdon B.
Nutr J. 2015 Aug 25;14(1):86. doi: 10.1186/s12937-015-0074-4.
This one had been wrongly quoted by Mark as being based on self-selected samples of children: on the contrary, it is based on cross sectional data gathered in representative samples of children

- Comparison of weight-for-height and mid-upper arm circumference (MUAC) in a therapeutic feeding programme in South Sudan: is MUAC alone a sufficient criterion for admission of children at high risk of mortality?
Grellety E, Krause LK, Shams Eldin M, Porten K, Isanaka S.
Public Health Nutr. 2015 Oct;18(14):2575-81. doi: 10.1017/S1368980015000737. Epub 2015 Mar 25.

- 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.
Isanaka S, Guesdon B, Labar AS, Hanson K, Langendorf C, Grais RF.
PLoS One. 2015 Sep 16;10(9):e0137606. doi: 10.1371/journal.pone.0137606. eCollection 2015.

All these works challenge the aforementioned hypotheses and rather support the alternative hypotheses:
- MUAC<115mm is primarily an indicator of growth retardation in young children, rather than an indicator of acute nutritional needs (contrarily to WHZ)
- MUAC<115mm overdiagnose SAM in young stunted children and underdiagnose SAM in older children
- Children with WHZ<-3 and MUAC=115mm (WHZonly) are at similar high risk of death compared to children with MUAC<115mm, and they respond at least as well to treatment
- Among children with MUAC<115mm, those presenting with WHZ<-3 are the most at risk of immediate death, while those with WHZ>=-3 have more chronic vulnerability profiles.

Further investigations are required and are coming. In the meanwhile, it would be good to stay a bit cautious and to keep in mind both types of hypotheses.

Finally, I would like to say that, although I fully agree with the importance of the feasibility argument for SAM management programmes (especially the difficulty to detect and refer low WHZ children), feasibility is a context and time-specific concept which should not interfere with the need for a deep understanding of the needs for this type of programme.


Mark Myatt

Consultant Epideomiologist

Frequent user

3 Oct 2015, 09:10

We keep coming back to tangentially relevant or misleading evidence:

Givjalva-Eternod et al. (2015) - Interesting but not, I think, directly relevant as the "outcome" measure is of body composition (not mortality). The finding of this paper are at odds with other papers that address this issue. Further work is probably needed although I am not sure if it will be relevant to CMAM programming.

Roberfroid et al. (2015) - Showing that MUAC and WHZ select different children is exactly the point. WHZ perform badly at what we need it to do. MUAC performs better that WHZ at what we need it to do. If MUAC and WHZ selected the same children then MUAC and WHZ would perform equally well as what we need then to do. I avocate for the use of MUAC exactly because because it is different from WHZ (which does very badly at what we need it to do). [/b]Please point me to the post where I misrepresent the Roberfroid et al. (2015) paper. I will check and, if needed, write a correcting note.[/b]

Grellety et al. (2015) - This has been discussed at length above. This is not appropriate or useful evidence.

Isanaka et al. (2015) - Here we have evidence that children with low MUAC respond to treatment. This is supports using MUAC for admission. We already knew this. The article does not address the issue of mortality risk in untreated cohorts which is what we need to decide whether the discordant MUAC-/WHZ+ children need the full CMAM protocol or could be treated in (e.g.) TSFP type programs. Clinical studies miss the point somewhat (see above re Grelley et al. (2015)).

It does seem that we need more evidence (although I fear that it will be ignored by those with a "dogmatic attitude in favour of low WFH children"). What you have presented above is not the evidence we need. At best these articles present tangentially relevant evidence. Grellety el al. (2015) is so inappropriate and biased as to be misleading. Your "Further investigations ... " suggest more of the same pseudo-evidence.

I am not sure where the evidence for some of the statements about MUAC and WHZ comes from. We know from the CTC research program that (e.g.) low MUAC kids in SFP die at five or six times the the rate as the same kids in OTP. This was a key piece of evidence for allowing MUAC to be used in OTP programs. I cannot understand the logic behind classifying the identification and admission of children who need and can benefit from the program as over-diagnosis. This is the sort of "logic" that gives rise to the sort of programming described in Grellety et al. (2015).

Please ... rather than keep producing and flaunting tangentially relevant or misleading evidence why not start collecting so useful and appropriate evidence?

Benjamin Guesdon

Research coordinator / ACF-France

Normal user

3 Oct 2015, 15:29

Dear Mark,

You wrote: "I am not sure where the evidence for some of the statements about MUAC and WHZ comes from."

These hypotheses are in my opinion quite well supported by the papers I quoted, as well as by the evidence reviewed here, and by Roberfroid et al. for WHO. They also cope quite well with the evidence (not talking here about the comments) you usually present. The reverse is not true: your usual hypotheses are challenge by the recent evidence.

Indeed :
- Knowing that mortality risk decreases with age, a parameter like MUAC, which is strongly related to age, and which is also strongly related to stunting, can be expected to be a good predictor of mortality. It does not need to be an indicator of acute nutritional needs to display this characteristic.
- Young and stunted children without acute nutritional needs, as would be children misdiagnosed as SAM by MUAC, can be expected to display some weight or MUAC gains, especially when treated with high caloric foods. You can even call this recovery if weight or MUAC gain is the only recovery criteria you are interested in measuring.
- However, young and stunted children without acute nutritional needs can hardly achieve the short term rates of weight gain demonstrated by children with acute nutritional needs.

Mark, with the data you have in hands, can you tell us how weight gains of MUAC only children in the short term, let’s say at 2 weeks, compare with weight gains in the two other SAM categories (i.e children with both MUAC<115mm and WHZ<-3; and children with MUACabove115mm and WHZ<-3)?
I am sure you will be interested in the observation made on this point by Isanaka et al.

It is true however that the hypothesis that MUAC-only children are rather chronic than acute could be challenged by the observation of a high immediate mortality risk in these children. Published evidence of this is however not there. I only retrieved a citation of a “natural experiment” observing mortality in children with MUAC<110mm and height below 75cm and WHZ above 70%of the mean (NCHS) in one of your publications, Mark (Myatt 2006). Can you have a look at your files and tell us how many of these children where MUAC<115mm and WHZabove-3, how many were MUAC<115mm and WHZ<-3, and what was their mortality rate?

That would be definitely an interesting piece of information.


Benjamin Guesdon

Research coordinator / ACF-France

Normal user

3 Oct 2015, 15:35

sorry there is a problem with the link I wanted to add above: I was referring to the great EN-net mediated review on this issue.


Frequent user

5 Oct 2015, 04:42

You keep referring to the 2010 WHO working document of which you are a co-author. I posted a comment on the weaknesses of of this document on the thread on MUAC for admission and discharge. Please reply to these comments before considering this can be used against my current position:

MUAC is better predictor of death than WFH. The Grellety et al paper cannot be used in any way to refute this.

Children with low MUAC respond to treatment and have a rapid weight gain when getting SAM treatment.

Now you say:

“Young and stunted children without acute nutritional needs, as would be children misdiagnosed as SAM by MUAC, can be expected to display some weight or MUAC gains, especially when treated with high caloric foods. You can even call this recovery if weight or MUAC gain is the only recovery criteria you are interested in measuring.”

We are several on this forum to plea WFH supporters to provide evidence in favour of their position or at least to tell us which kind of study will support their attitude. A claim which cannot be tested and for which there is no sound scientific rationale should be regarded as non scientific.

In the statement above you claim that the weight gain of some low MUAC children is not a criteria of improvement. How can you say this ? And test this ? Do you think any committee will allow a trial in which young stunted low MUAC children will remain un-supplemented to test your hypothesis ? And what is the rationale supporting the idea that giving a high energy high nutrient diet promoting weight gain will not be beneficial for them ?

This kind of statement is just creating confusion, and reflects a dogmatic attitude. I suggest you come back on this discussion when you have something clear to propose to support your ideas.

Mark Myatt

Consultant Epideomiologist

Frequent user

5 Oct 2015, 12:33

I have started a thread here.

I hope that we can come together there to work out what we need to do to move forward with an informed debate on issues around MUAC-only programming (a wider topic than has been debated here) and the appropriate treatment of children with MUAC > 115 mm and WHZ < -3.

See you there!

Benjamin Guesdon

Research coordinator / ACF-France

Normal user

6 Oct 2015, 09:00

Dear all,

When I said that it was good sometimes to see how observations match theory, I never meant that it would be a pleasant exercise for everybody. There is no point however in leaving the ground, losing tempers and insulting people engaged in providing more evidence, asking questions, putting forward existing and alternative hypotheses, or simply referring to balanced and collective examination of the scientific literature.

I will now go back to the scientific discussion.

André, you wrote: “MUAC is better predictor of death than WFH”
First, according to the past cohorts and their ROC curves, even age alone would be performing better than WHZ: this does not mean that all children under 2 have acute nutritional needs and should be treated for SAM. Second, this statement does not mean that children with MUAC<115mm have a higher mortality risk than children with WHZ<-3, does it? In which of your publications (or comments) have you shown this? Here I would have missed something.

Mark and André, you wrote: children with MUAC<115mm “exhibit rapid gains in both weight and MUAC usually following a typical growth curve (i.e. very rapid gains in the first few weeks slowing and approaching an asymptote)” and “children with low MUAC respond to treatment and have rapid weight gain”
In fact, it seems that not all children with MUAC<115mm exhibit this type of growth curve, only the ones presenting with both a MUAC<115mm and WHZ<-3 (and the WHZonly category as well). See the paper by Isanaka et al 2015: initial response to treatment in terms of weight gain is much lower in MUAConly SAM children than in the two other SAM categories (children with both criteria, and WHZonly children). This was also highlighted in the review by Roberfroid et al for WHO in 2013. This is also confirmed by many field observations.

Mark, you wrote: “When we don’t recruit them into OTP and put them into SFP they do not do well”, thereby referring to the “natural experiment” you mention in the 2006 paper.
Again, among these children with excess of mortality risk in SFP, what was the proportion of children with both MUAC<115mm and WHZ<-3, and what was the proportion of children with only MUAC<115mm? That would be interesting and useful information for the current discussion.

To conclude, I will not adopt a dogmatic attitude and extrapolate beyond the observations to draw high-impact recommendations restricting the target of SAM management programmes to one or the other of the criteria. I will rather go back to work, and try with my colleagues to provide the evidence we are lacking now.


Tamsin Walters

en-net moderator

Forum moderator

6 Oct 2015, 13:17

Dear all,

This discussion is now running on two threads, so I propose to close this one and any further comments can be posted here in the Prevention and treatment of MAM forum area under the "Only MUAC for admission and discharge?" discussion.

Alternatively, a new thread has been opened to discuss ideas for research/collection of evidence to help us move the debate forward. It can be found here .

Best wishes,

Back to top