Dear ENN, I was wondering if there was any information or experience available in the use or adjustment of BMIs in adult populations in populations that are skewed from the norm (such as the Nuer and Dinka in South Sudan). We are having some difficulties with interpreting BMI data for within these populations and would like to have a method of cross comparison to other areas of the country too. I have heard and read of cormic index adjustment (and read some on the forums and the Somali Micronutrient study). I have a few questions regarding this. If we were to do this, would we need to do sitting height/standing height ratio for the whole population (noting their ethnicity) in the area of intervention or would a sample be sufficient? If we are to do a sample in order to get the cormic index adjustment ratio, how would we calculate the sample? And would we need a separate one for the Nuer and the Dinka (with the assumption the SH/S ratio is different?) and other groups? What experience exists in this? The Somali Micronutrient study noted that it was not used to adjust for the BMI - why is this? For a cross-comparison within areas and ethnicities, would the adjusted BMI be sufficient and reliable? Alternatively, is there any more information on the use of MUAC in adults - all the literature note the poor connection with mortality in adults and indeed many cut-offs have been suggested. Furthermore, what are the cut-offs for emergency intervention in this group of population? We have the similar problem in children, where emergency thresholds exist in weight-for-height but are lacking in MUAC..... I have heard of the use of anthropologists to adjust the BMI in South Sudan and work that has been done before as this is a recurrent problem - is there anyone who can guide me on this? Thanks a lot for your help.
Dear Anonymous 502: Nick Norgan at Loughborough University has done the most work on adjusting BMI for different body habitus. The best description of the Norgan correction for adjusting BMI for sitting height / standing height ratio (also called the Cormic Index) is in the paper: Norgan NG. Relative sitting height and the interpretation of the body mass index. Ann Hum Biol 1994;21:79-82. The application of this correction and demonstration of the difference it makes in the estimated prevalence of malnutrition among adults in an emergency-affected population is given in the paper: Salama, et al. Malnutrition, measles, mortality, and the humanitarian response during a famine in Ehiopia. JAMA 2001;286:563-571. Norgan's correction is done on each individual, so you would need sitting height and standing height on each survey subject.
Anonymous

Answered:

9 years ago
The issues that you raise are not just for adults. It is now clear that weight-for-height type measures (e.g. WHZ, BMI) are strongly affected by body shape in both children and adults. This is one reason why MUAC is now being used to assess children and PLWs (with pregnant women the BMI is also biased upwards - this has been covered elsewhere on this site). The Cormic Index adjustment corrects BMI for body shape using sitting to standing height ratio (SSR). This is usually done for individuals. It may be possible to arrive at an average correction but I am not aware if this has been done (i can see that ethnic group and age might complicate this). If you are using survey data then you need only correct the BMI for individuals in the survey. This will need you to collect sitting and standing height for all in the survey). This will need you to take a sample that properly represents your population of interest. Correction will allow between country comparisons but the correction is not often done so the number of comparisons you will be able to make will likely be few. I am not sure that the available data does show "poor connection with mortality in adults". An RNIS review in 2000 indicates the opposite and recommended the use of MUAC for acute (primary and secondary) undernutrition. As with children, the cut-offs used have varied over time. Currently we have: MUAC < 230 in PLWs for GAM MUAC < 210 in other adults (FOR GAM) MUAC < 185 in other adults (for SAM) These cut-offs are based on considerable evidence. The lower cut-offs mentioned in the RNIS review were derived from extreme famine conditions (Somalia 1992) and have been superseded. There is no real issue with prevalence thresholds with MUAC. These are the same as for W/H. Why would they not be? The thresholds were designed when we thought that WHZ measured something useful. I hope this is of some use.
Mark Myatt
Technical Expert

Answered:

9 years ago
Dear Mark, Thank you very much as always for the insightful comments. It looks like in order to come to a correction of BMI we would therefore have to do SSR for all the sample we have. From this, if we take ethnicity as well, would it be possible to arrive at an average correction rate? What would be the sample needed to come up with an average correction rate? There was a micronutrient study in Somalia by Andrew Seal et al, in which an average was arrived at but was not used (despite the positive correlation). I quote from the National Micronutrient Study, 2009 by FSNAU and UCL: "The mean sitting height/standing height ratio (Cormic Index) was 50.5% and there was a significant positive correlation between the BMI and the index (P=0.01). Correction of BMI for using the Cormic Index is required for before comparison of prevalence data between populations. However, different equations for this correction exist in the literature and no consensus currently exists on the best method for use with Somali populations. For these reasons the data in this report is presented as unadjusted BMI." Has there been any further consensus on this issue? is there an equation we can use? We are also using MUAC but find it difficult in these populations. We have been using Ferro-Luzzi cut-offs of <200mm for SAM males and <190mm for SAM females and <230mm GAM males and <220mm GAM females. The discrepancy between BMI (unadjusted) and MUAC has been very confusing thus far and therefore it is difficult to obtain the idea of the severity of the situation. In terms of thresholds, 15% as an "emergency" threshold seems applicable often in terms of WFH, but the MUAC rates rarely seem to be at this "emergency" threshold. Have there been studies of this that have correlated the MUAC thresholds with this emergency threshold and when to "trigger" interventions as is often used with WFH GAM rates of 15%? Thanks a lot for the input, as always. Regards
Anonymous

Answered:

9 years ago
Thank you for your kind comments. Note that if all you want to do is make between survey comparisons in the same population then you do not need the correction. Yes. You would have to do SSR for the sample. You might want to do this in a small survey (a mean can usually be estimated with a small sample size - n = 60 in each group might be enough) and estimate the means SSR for each ethnic group and for males and females separately. See [url-http://www.unsystem.org/scn/archives/adults/ch05.htm]here[/url] for details of the correction and how to apply individual and average corrections. The Ferro-Luzzi cut-offs were, I think, derived from BMI cut-offs. They are not in current use. The current cut-offs for prevalence are 210 mm (GAM) and 185 mm SAM. PLS with MUAC < 230 mm need nutrition support. WRT : "In terms of thresholds, 15% as an "emergency" threshold seems applicable often in terms of WFH, but the MUAC rates rarely seem to be at this "emergency" threshold" ... it will depend on where you are. In South Sudan you will find that GAM is low by MUAC but high by WHZ. This is (in large part) due to WHZ being affected (biased downwards) by body shape. In high altitude and cold climate countries you may, in emergencies, see the opposite. This is because WHZ is biased upwards by body-shape. In South Sudan you may early see the 15% threshold breached when using MUAC. Note : MUAC is also affected by body shape but not to the extent that is creates a significant bias in prevalence. The 15% threshold is a "back of the envelope" figure for GAM and is not specifically linked to WHZ. I hope this helps.
Mark Myatt
Technical Expert

Answered:

9 years ago
Dear Brad and Mark, Thank you both for your replies - I got delayed in various other things but back to following this up actively. Thanks for the clarification on the Ferro-Luzzi cut offs. Do you have papers that indicate the evidence of the adult MUAC cut offs that you are proposing Mark? Brad - I have been trying to get my hand on the two papers that you suggested (Norgan, 1994 and Salama et al,2001) but have been unable to find them on the web and have no access to the journals themselves. Would anyone have them? Or be able to better point me in this direction? Thank you both for your help, looking forward to hearing more. Regards
Anonymous

Answered:

9 years ago
Just to correct my previous message ... See [url=http://www.unsystem.org/scn/archives/adults/ch05.htm]here[/url] for details of the correction and how to apply individual and average corrections. There are many papers and reports with the MUAC cut-offs. They are in common use. Try a PubMed search. Also, check out publications from UNOs (e.g. WFP) that use these thresholds.
Mark Myatt
Technical Expert

Answered:

9 years ago
From Brad Woodruff: Dear Anonymous 502: The Salama article can be found at [url]http://jama.jamanetwork.com/article.aspx?articleid=194065[/url] Sorry, I've forgotten which Norgan paper I recommended. All Norgan's work was published in the early to mid-90s, so it's relatively less accessible than more recent publications. One of Norgan's papers which talks about adjustment of the BMI for Cormic index can be found here: [url]http://apjcn.nhri.org.tw/server/APJCN/4/1/137.htm[/url]
Tamsin Walters
Forum Moderator

Answered:

9 years ago

Dear All again,

I am sorry I keep on coming back to this issue of body shape again even after a hiatus.

I am now trying to mind-map the issue of MUAC in adults while still exploring the cormic index adjustment. If one would start a treatment programme for adults (note, not elderly only) using MUAC only (due to the BMI problem), what would the discharge criteria be? I know the advice has fluctuated in children from MUAC to percentage weight gain etc - how would this be in adults? Would we expect them to increase their MUAC within the given time they are in a programme or would we be getting a lot of non-respondents in this way? A suggestion I have heard is that we limit the length of stay to 2 months, and thereafter they are discharged regardless of MUAC, which goes against the grain (to me) of a nutrition programme. Basically, what would your advice be in an adult treatment programme if admission is by MUAC?

As always thanks a lot for your input on this forum and I am happy to be directed to more reading. I have looked through a lot of the studies from epicentre and other organisations on this issue but still find a dearth of information regarding adults! Especially non-pregnant females and adult men!!

Anonymous

Answered:

8 years ago

A fixed time period is commonly used to provide a discharge criteria of "non-responder". This should not be used as a simple catch-all. Response should be monitored so that non-response can be identified early and investigated and treatment changed as appropriate (e.g. you may want to check for and treat TB).

Some programs have additional non-anthropometric admission criteria. You might (e.g.) admit a lactating women with a MUAC less than 230 mm. It seems reasonable to discharge this women when she ceases breastfeeding provided that her MUAC was >= 210 mm.

For simple anthropometry only programs ... something like:
Admit to SFP : MUAC < 210 mm Discharge : MUAC >= 210 mm for two visits Clinically well Admit to TFP : MUAC < 185 mm Transfer to SFP : MUAC >= 185 mm for two visits If no SFP, discharge : MUAC >= 210 mm for two visits Clinically well
Seems reasonable. You may want to increase the discharge criteria to 230 mm.

It is important o not that cases may have an underlying disease such as TB or HIV (often together) so it is important to undertake clinical assessment and test / refer as appropriate.

I hope this is of some use.

Mark Myatt
Technical Expert

Answered:

8 years ago
Please login to post an answer:
Login