Nutrition assessment of adolescents
Nutrition assessment of adolescents Brad Woodruff 05.02.99
Nutrition assessment of adolescents Peter Salama 08.02.99
Re: Nutrition assessment of adolescents Eric Benefice 11.02.99
Re: Nutrition assessment of adolescents Tony Nelson 13.02.99
Nutrition assessment of adolescents David Hipgrave 17.02.99
Nutrition assessment of adolescents Andy Seal 18.02.99
Nutr assmt of adolescents and infants Claudio Schuftan 23.02.99
Nutrition assessment of adolescents Abe Parvanta 23.02.99
Indicators for adolescent malnutrition Saskia van der Kam 21.06.99
Indicators for adolescent malnutrition Michael Golden 22.06.99

Date: Thu, 11 Feb 1999 08:40:16 +0000

From: "Woodruff, Brad" <>

Subject: Ngonut: Nutrition assessment of adolescents


Dear NGO Nutters:

In collaboration with UNHCR, CDC conducted nutrition survey of adolescents age 10-19 years in the four camps in Kenya housing Sudanese and Somali refugees. Height, weight, age, and sex were recorded for a random sample of adolescents age 10-19 years. The body mass index (BMI) of each study subject was compared to the WHO BMI-for-age reference population. Subjects with a BMI less than the 5th percentile of their age group in the reference population were defined as having acute protein-energy malnutrition. Using this definition, it was found that in Kakuma Camp, 223 (57%) of the 391 survey subjects were acutely malnourished. In the three camps in Dadaab District in eastern Kenya, 237 (61%) of 391 survey subjects were acutely malnourished. However, ancillary data, such as morbidity and mortality rates, did not support these high estimates of the prevalence of malnutrition among adolescents in both populations. The WHO BMI-for-age for adolescents is based on measurements of American adolescents in the first National Health and Nutrition Examination Survey (NHANES I). We strongly suspect that this reference may not be an appropriate comparison for all adolescents worldwide.

1). We would be interested in hearing of others' experience in using this reference to measure acute malnutrition among adolescents.

2). We would also very much like to hear suggestions for other, perhaps more valid, methods of assessing the nutritional status of adolescents.

Rita Bhatia (UNHCR) and Brad Woodruff (CDC)

Date: 8 Feb 1999 14:31:04 +0100

From: "Peter Salama" <>

Subject: FW: Ngonut: Nutrition assessment of adolescents


Reply to the questions posed by Rita Bhatia and Brad Woodruff regarding adolescent criteria.

Concern has not used BMI/Age as the criteria for defining malnutrition in the adolescent population in the past. In fact we have been struggling with the issue of finding valid criteria for adolescents in southern sudan over the past year.

In programmes in parts of Central Africa we have used extended weight for height charts to define malnutrition for the purpose of admission and discharge into supplementary and therapeutic feeding centres. Feedback from the field suggests that these worked well in Burundi. In our programmes in Bahr el Ghazal with Dinka populations however, these charts (which I believe are extrapolations from standard weight for height charts) did seem to overestimate the prevalence of malnutrition. This became such an issue on discharge (from supplementary and therapeutic feeding programmes) with very healthy looking adolescents failing to attain the 85% weight for height (%M) after 5-6 weeks periods that we began to use a combination of MUAC and clinical criteria as discharge criteria for the older ie.-post-pubertal adolescents.

(NB. In our adult programmes in Sudan we use a combination of MUAC and clinical signs/symptoms such as inability to stand, dehydration, anorexia for admission and discharge).

One of the problems in comparing the Dinka population to any reference population is the relatively long leg length in comparison to body length so that the sitting height/standing height ratio is lower (significantly) than the reference population. A 30 cluster survey of adolescents and adults that we completed in January 1999 in Aweil West in Bahr el Ghazal in the Dinka population confirms prior work done in this field in sudan. A large proportion of the population in Kakuma refugee camp are Dinka so this would certainly be one factor to explain the high prevalence quoted. I imagine that the other Sudanese groups in Kakuma and the Somalis in Eastern Kenya would also have lower sitting height/standing height ratios but perhaps not as low as the Dinka. So as you point out, BMIs calculated in this population are not equivalent to the reference population and need to be corrected prior to comparison.

The other perennial problem that we've found is using stated age in any criteria. Even with good local events calendars we've had a good deal of difficulty getting accurate age estimations in Sudan this year. Clearly this is particularly crucial in the adolescent growth period where 1 year either way can make such a difference. And yet even with accurate age determinations, using a reference population one would also have to be reasonably confident that the age of onset and duration of puberty were comparable in the study and the reference population.

In sum what Concern is moving towards in these populations is extended weight for heights criteria in the pre-pubertal population and MUAC/clinical criteria in the post-pubertal. But we've still a lot of work to do!



Peter Salama, Medical Co-ordinator Emergencies

Concern World-wide

52/55 Lwr Camden St, Dublin 2, Ireland

telephone 00 353 1 475 4162, email

Date: Mon, 08 Feb 1999 17:33:35 +0100

From: Eric BENEFICE <>

Subject: Re: Ngonut: Nutrition assessment of adolescents


reply to Rita Bhatia (UNHCR) and Brad Woodruff (CDC)

A colleague sent me your message about BMI use in adolescents. In Senegal where we are studying puberty, we found that compared with european or american population, there exits an important delay in growth spurt and maturation among adolescents. Median age of menarche is over 16 years. As puberty go along with increase in fat mass (girls), FFM (boys) and body mass, you must take into account the maturational status when using BMI. Note that BMI are calculated from American population with prevalence of obesity > 25% (Hanes III survey conclusions). Also using 5em P of BMI for age during adolescence as threshold of maturation is not a recommendation of WHO but a suggestion (see: de Onis et al, Am J Clin Nutr 64: 650-658, p 653, last paragraph).

You may also consider 2 other issues 1) a late maturation is merely a consequence of chronic undernutrition during infancy and childhood; 2) there is a possibility for some catch-up in African population with prolonged growth during adolescence.

In conclusion: 1) BMI is not an appropriate indicator of malnutrition during adolescence unless you valid it for your specific population; 2) instead, you could use other simple indices such as arm circumference.


Dr. Eric Benefice


911 av Agropolis, BP 5045, 34032 Montpellier cedex1 (France)

tel: 04 67 41 61 48, fax: 04 67 54 78 00


Date: Sat, 13 Feb 1999 09:28:12

From: Tony Nelson

Subject: Re: Ngonut: Nutrition assessment of adolescents


There seems to be two schools of thought on the universal applicability of growth standards:

1. That there are universial growth standards applicable to all children (and by inference if significant children fall above or below the standard then they are either malnourished, stunted or overweight).

2. That there are definite population and genetic differences and growth standards should be developed for the population being assessed.

As I understand it, data showing high rates of stunting in south Asia are based on WHO (american) reference standards and do not consider the possiblity of genetic differences to partly explain the high rates of stunting.

In past years the same arguement was used to infer that Hong Kong children were smaller than their Caucasian counterparts because of suboptimal nutrition. Although there has been secular changes in the growth of Hong Kong children over the past 30-40 years, it is clear that local growth standards differ from those from other populations (Leung SS, Cole TJ, Tse LY, Lau JT. Body mass index reference curves for Chinese children. Annals of Human Biology 1998; 25: 169-174.)


Tony Nelson

Date: Wed, 17 Feb 1999 17:30:31 +0000

From: David Hipgrave <> (

Subject: Re: Ngonut: Nutrition assessment of adolescents


Dr Nelson's comments open a whole area of discussion on which I suspect many ngonutters would appreciate some expert debate.

Although the original question referred primarily to adolescents, might it not equally be asked about infants and children? Without going to extremes, how important might it be for standards to be developed for individual ethnic groups, and how would that change the way the world perceives nutritional standards in many countries?


Dr David Hipgrave, Project Manager

Strengthening Immunisation and Malaria Control, Viet Nam Health Alliance

Box 252, IPO, Hanoi, Viet Nam

Ph. 84 4 766 0484, Fax 84 4 766 0486

Date: Thu, 18 Feb 1999 12:08:22 +0000

From: Andy Seal <>

Subject: Ngonut: Nutrition assessment of adolescents


A few comments/ideas on the issues of anthropometric standards raised by Tony Nelson and David Hipgrave.

International nutritional reference data allows for international comparisons which can be so important for epidemiology, advocacy etc.

References are not the same as standards although, unfortunately, in reality they are often treated as the same thing.

The functional significance of any chosen cut-off may vary according to the particular population and local circumstances. Therefore, in certain situations it may be appropriate to set local "standards".

This dosen't necessarily mean that the expense and difficulties of constructing a locally valid and accepted data set have to be undertaken. (Indeed, the validity of local data sets may be time-limited due to secular changes.) It may just mean adopting different cut-offs using the same international reference data. Such a decision would need to be informed by epidemiology and operational realitites, and, just as importantly, by the ethical/political dimensions.


Andrew Seal PhD, Researcher in Public Health Nutrition

Centre for International Child Health, Institute of Child Health

30 Guilford Street, London WC1N 1EH

tel. +44 (0)171 242 9789 ext. 2468, fax +44 (0)171 404 2062


Date: Tue, 23 Feb 1999 11:59:20 +0000

From:Claudio Schuftan

Subject: Nutr assmt of adolescents and infants


Please keep the issues of anthropometric standards for adolescents separate from those for infants.

References are not the same as standards, granted.

But I think that what ultimately counts is the public health significance of the currently used (international) cut-off points for infants. Even IF current growth charts would not fully apply to particular populations and local circumstances, would the expenses and difficulties of constructing locally valid and accepted data sets be worth that expense? Given the magnitude of the problem of malnutrition in most developing countries, moving the cut-off point 2-3 percentiles down (as compared with the intl) would still leave the same hundreds of thousands of infants malnourished. Using the "old" (intl) cut-off points we still have a monumental problem that we have been unable to solve. Reducing the number of officially malnourished infants by a few thousand will still not let us off the hook of our responsibilities to combat malnutrition more decisively by addressing its immediate, underlying AND basic causes simultaneously.

Let's leave the fine-tuning for when we can say that we have overcome the worst.

Claudio Schuftan

Tue, 23 Feb 1999 23:47:21 +0100 (MET)

From: "Parvanta, Ibrahim" <>

Subject: RE: Nutrition assessment of adolescents


I concur with Dr. Seal's comments. As an example, it may of interest to colleagues that in the United States, the cutoff for classifying nutritional risk based on anthropometric indicators is the 5th percentile which is equivalent to about a Z-score of -1.65. If the internationally used Z-score of -2.00 were to be used in the U.S., then very few children would be classified as "at risk" and most of them would be related to signifcant medical reasons rather than inadequate or poor nutrition. Thus, using a higher Z-score (-1.65) allows public health and primary care programs in this country to identify "at risk" children who could benefit from nutritional interventions. In contrast, specialized medical programs in the U.S. which serve children with severe medical needs and "failure to thrive" use the 3rd pecentile (equivalent to -1.88 Z-score) to assess and monitor growth of children under their care.

Although genetic potential is important in assessing anthropometric status, in young children this is more appropriate at the individual rather than the population level. A recent publication (Mei, et al, 1998) supports the use of the CDC/WHO international growth reference. The authors present data that show that over the past 2 decades the prevalence of low height-for-age among S.E. Asian immigrants and refugees to the United States has dramatically decreased and that the entire height-for-age distribution curve for these children is now almost similar to that of the other ethnic groups in this country. Thus, the S.E. Asian immigrants are not shown to have a different genetic growth potential as a population. Rather, improvements in their living conditions, nutritional and health status, have resulted in improved linear growth.


Mei Z., et al. Improving trend of growth of Asian refugee children in the USA: Evidience to support the importance of environmental factors on growth. Asia Pacific J Clin Nutr. 7: 111-116, 1998.

Best regards,


Abe Parvanta

Division of Nutrition & Physical Activity

Centers for Disease Control and Prevention, Atlanta, GA 30341


tel: 770-488-5865, fax: 770-488-5369


Date: Mon, 21 Jun 1999 18:29:26 +0100

From:,Saskia van der Kam

Subject: indicators for adolescent malnutrition


Dear colleagues,

Some time ago on NGONUT the question was brought up about what indicators and what reference is the best for assessing malnutrition in adolescents. Apparently BMI and BMI-for-AGE is not very satisfactory.

I did not see a satisfactory answer on NGONUT. So I wonder, is there no answer, or did some individuals directly answer the authors of the question?

Anyway, I would like to bring up the issue again: what indicators, references and cut off points are used? in what population groups?

Are these somehow validated?


Saskia van der Kam, Médecins Sans Frontières Holland

Date: Tue, 22 Jun 1999 12:05:52 +0100

From: Michael Golden

Subject: indicators for adolescent malnutrition


Dear Saskia,

When the question of assessment of individuals taller than the weight-for-height standards came up in ACF, I generated an extended weight-for-height chart derived from the NCHS standards of height-for-age and weight-for-age to obtain the median weight-for-height. We have been using this since 1995; I believe that it has now spread to other NGOs at field level and is being quite widely used. Certainly, ACF field staff have found this to be easy to use and very convenient..

The chart/table is easy to use and to teach because it is exactly the same as the one used for the younger children - this is a tremendous advantage.

I have given the extended weight/height table below, using 1 cm increments in height and 100g in weight. if anyone wants this with finer gradations or different %median values, I will happily send it - but it is a bit large for this message.

It is not sensible to use the Z-score system with this chart. This is because one SD that I calculate is appreciably wider than it would be with a chart constructed from the raw data (the spread of weight for people all of the same height -but different ages - is much less than the spread of weight of people all of the same age who have the same median height as in the first instance but a spread of height in practice). For this reason we have used 70% of the median as the cut-off point as the admission criterion - and 85% as the criterion for transfer to supplemental feeding program - exactly the same as for younger children.

BMI itself does not work because of the large changes in absolute value of the normal pre-adult.

It is very unfortunate that BMI-for-age was inserted into the WHO manual at a late stage. This measure has much the same problems in assessing wasting as weight-for-age does and is correspondingly inaccurate in stunted individuals or those that do not know their age. It is conceptually very difficult to explain to field staff and it attempts to introduce the third type of cut-off (centile - when staff who are used to % of median are already struggling with Z-score).

BMI itself was generated to be a height independent measure of "fatness" so that tall and short adults could be compared using the same absolute numbers. Thus, although BMI-for-height would get over some of the problems of BMI-for-age, it is theoretically unsound as well as being conceptually difficult to teach.

None of these measures (or MUAC) have been properly evaluated to examine their sensitivity and specificity in determining outcome in the malnourished adolescent, so we do not yet have the data to make any statement (that is not just a guess) about which measure is the most appropriate for deciding upon admission and discharge of an adolescent. In the mean time we should use the easiest and most familiar for the staff.

There are some data but we, as yet, have insufficient numbers to do a meaningful analysis. Even when we do examine the outcome of admitted patients against the admission criteria, this may not relate linearly to the outcome that the person would have had had they not been admitted and received treatment - it is the latter data that would be best.

I recommend that, for the time being, for adolescents we use % median weight-for-height, as we do with younger children. And that we actively gather data, to examine the specificities and sensitivities of the different anthropometric assessments of severely malnourished adolescents in order to refine this recommendation.

(see weight/height table)


Prof. Michael H.N.Golden