You are on page 1of 5

Journal of Tropical Pediatrics Advance Access published September 17, 2010


Evaluation of Mid-upper Arm Circumference in Pre-school

Children: Comparison Between NCHS/CDC-2000 and
WHO-2006 References
by Viviane G. Nascimento,1 Thais Costa Machado,1 Ciro João Bertoli,2 Luiz Carlos de Abreu,1,3 Vitor E. Valenti,3,4 and
Claudio Leone1
Departamento de Saúde Materno-Infantil, Universidade de São Paulo (USP), São Paulo, SP, Brazil
Departamento de Medicina, Universidade de Taubate´, Taubate´, SP, Brazil

Downloaded from at Universidade Federal de S?o Paulo on October 12, 2010
Departamento de Morfologia e Fisiologia, Faculdade de Medicina do ABC, Santo Andre´, SP, Brazil
Departamento de Medicina, Universidade Federal de São Paulo (UNIFESP), São Paulo, SP, Brazil

Correspondence: Claudio Leone, Departamento de Saúde Materno-Infantil, Faculdade de Saúde Pública, Universidade de São
Paulo, Avenida Dr Arnaldo 715, Cerqueira Ce´sar 01246-904 São Paulo, SP, Brasil. Tel.: þ55-11-30850240,
E-mail: <>

We aimed to evaluate the classification of arm circumference (AC) in pre-school children by using
National Center for Health Statistics (NCHS/CDC-2000) and World Health Organization
(WHO-2006) references. We evaluated 205 children: weight, height and AC were assessed and the
body mass index (BMI) was calculated. The BMI values were classified into Z-scores by the WHO
referential. The AC was classified into Z-cores by two references, comparing the whole-sample value
and among groups (tercis) of BMI Z-score. The correlation was also evaluated between differences
of AC with BMI Z-score. The WHO referential classified the AC in Z-scores greater than the
NCHS/CDC, which is more specific and less sensitive than the NCHS/CDC for lean children and at
the same time more sensitive and less specific for children with overweight. In conclusion, a significant
difference in the AC classification occurs according to the referential used.

Key words: general paediatrics, international child health, orthopaedics, nutrition, statistics.

Introduction appropriate for growth evaluation. This reference

The arm circumference (AC), also known as brachial also includes new values for brachial perimeter
perimeter, has been proposed as an alternative to classification [8].
select children 5-years old in critical nutritional In order to assess nutritional risk in individuals
state, aiming to identify the risk degree, regardless and populations, researchers compare their data
of age, gender, weight and height [1, 2]. Its measure- with measured values of growth reference [9–12].
ment is an important indicator of mortality prognosis Besides the reference of WHO-2006 for AC, there is
[3–6], especially in cases of severe malnutrition [7]. also the National Center for Health Statistics
Recently, in 2006, the World Health Organization (NCHS) USA [7], restated in 2000 by the Centers
(WHO-2006) estimated new anthropometric param- for Disease Control (CDC) and available on
5 NutStat Epi-Info software.
eters for children aged between 0 and 60 months,
In a single set of individuals, the use of different
proposing a reference considered as the most
benchmarks may produce different risk classifica-
tions, and therefore, create values of nutritional
disorders prevalence in communities, which would
Funding derail any evaluations and comparisons over time
and/or between population [13, 14]. At the moment
Authorship has been granted with the financial that we choose to use a particular anthropometry
support from FAPESP (Fundação de Amparo à reference, it is important to know the differences
Pesquisa do Estado de São Paulo, Brazil), # 08/ in risk classification in which it will result when
53142-9. compared to other references. Therefore, this

ß The Author [2010]. Published by Oxford University Press. All rights reserved. For Permissions, please email: 1 of 5

investigation was undertaken to evaluate the classifi- 2000 by the CDC [13] and available in the NutStat
cation of AC in pre-school children by using two Epi-Info software.
references: NCHS/CDC-2000 and WHO-2006.
Statistical analysis
Methods The comparison of mean AC Z-scores was made by
using paired Student t-test, with a significance level
Study population  ¼ 0.05%. In the analysis of results, we also used the
This analysis is part of a larger cross-sectional study, Pearson correlation coefficient in order to estimate
which aim was to evaluate, by anthropometric correlation between the difference between AC
analysis, growth achieved by a sample of children Z-scores (determined by the two references) and the
aged between 2 and 4 years, who attended at public difference between BMI Z-scores. To calculate
nursery schools in Taubaté city, São Paulo, in 2008. the sample, we used the Power and Sample

Downloaded from at Universidade Federal de S?o Paulo on October 12, 2010
The variables investigated were: weight, height, body Calculations software (v.3.0.1) and statistical analysis
mass index (BMI) and AC. Data collection was was performed by Graph Pad Prism (v. 5.0.2).
performed at the beginning of the school year. We
excluded children who presented history of chronic Results
diseases that could influence their growth. The Mean Z-scores of AC presented a difference of
consent term was sent to all parents or responsible 0.49 Z-scores, the one estimated by the WHO was
by the daycare center, which was duly completed and the highest (p < 0.0001; Table 1).
signed before the start of collection data. This study Figure 1 shows that the classification by the refer-
was approved by the Ethical Committee in Research ence proposed by WHO in 2006 offsets the estimated
of our University. distribution curve to the right, toward higher values
In order to calculate the sample size, we used the of Z-scores than the observed curve of the NCHS/
following parameters: level of the test significance CDC-2000 reference, which in turn is much closer to
 ¼ 5%; test power (1–) ¼ 90% and a minimum the theoretically expected normal distribution.
difference between the mean Z-scores estimated by
the two benchmarks of a quarter standard deviation
(0.25), which resulted in a sample size of 170 children.
Taking into account the possibility of any loss of TABLE 1
data, we added 20% to the total number of children Comparison of arm circumference (AC) Z-scores in
calculated that resulted in a sample of 204 children. 205 pre-school children according to the reference
Children were selected from the list of daycare used: NCHS/CDC – 2000 and WHO – 2006
centers provided by the Department of Education
and Culture of Taubaté in a probabilistic way, Reference Mean Standard Minimum Maximum
which resulted in a total of 205 children (87 girls
and 118 boys). NCHS/CDC – 2000 0.08 1.16 2.26 4.22
WHO 0.57 1.12 2.00 4.01
Variables evaluation
After obtaining consent from children parents or
responsible, anthropometric data were collected in
the daycare center, according to Lohman et al. [15]
method. The child’s weight was measured by using
the Seca electronic scale with a split of 0.1 g, and
height was measured by stadiometer wall (Seca)
with two meters, centimeters and millimeters subdiv-
ision. The measure of AC was performed with tape
fiberglass extendable using the non-dominant arm of
the child, placing the tape horizontally at the
mid-point between acromion and olecranon. The
measurements were performed by trained profes-
sionals, according to Hauspie technique (2004) [8].
The values of BMI [BMI ¼ weight (kg)/height2 (m)]
in Z-scores were calculated from the reference FIG. 1. Comparison of arm circumference Z-scores
values proposed by the WHO (2006) [8]. AC was distribution for age in pre-school children according
classified in Z scores by two benchmarks: the WHO to the reference used: NCHS/CDC – 2000 and WHO
in 2006 and the reference from the National Center – 2006. Paired Student t-test; p ¼ 0.0001; AC, arm
for Health Statistics (NCHS), USA [7] restated in circumference.

2 of 5 Journal of Tropical Pediatrics


The differences of mean Z-score in the first, second Discussion

and third tertile of BMI were 0.478, 0.539 and 0.453, The results of this study indicate that when analyzing
respectively, all statistically significant (p < 0.05; AC in pre-school children, depending on the bench-
Fig. 2). mark that is used, it may direct to different interpret-
Figure 3 displays that there was an inverse ations regarding the diagnostic of nutritional status.
correlation between the differences of Z-scores of We suggest that anthropometric assessment alone is
AC and the differences of Z-scores of BMI not sufficient for preparation of definitive diagnosis
(r ¼ 0.2945; p < 0.0001). The regression slope was nutrition; it is just one tool to be considered,
0.0301  0.0068 (p < 0.0001). however, in selection situations it may influence
interventions which were adopted as result of the
epidemiological situation.
AC has been considered an important indicator of

Downloaded from at Universidade Federal de S?o Paulo on October 12, 2010
malnutrition, especially during population screening
because it presents good correlation with BMI and it
is an easy measurement to be performed. Although it
was originally proposed primarily for malnutrition
screening [16, 17], AC was also used as an indicator
of nutritional status in general, including in Z-scores
form, in order to control the influence of gender and
age in its interpretation [7, 18, 19].
In order to assess whether AC could be used
for screening of pre-school obese children a previous
investigation [20] studied 1090 children aged between
12 and 59.9 months and proposed cut-off points
for AC for age because they found no advantage
in using the AC/height ratio. These authors also
suggest its use as a more practical method for
screening of pre-school children who are obese due
to the facility and speed of its attainment in that
FIG. 2. Comparison of Z-score distribution of arm Besides its use in prevalence surveys of overweight
circumference in pre-school children by tertile of and obesity risk and for screening populations, the
BMI according to the reference used: NCHS/CDC arm circumference may also be useful for clinical
– 2000 and WHO – 2006. ANOVA followed by assessment of nutritional status of critically ill chil-
Tukey post-test (*p < 0.05, p < 0.05). dren, since it has good correlation with arm muscle
mass and, thus, with the base nutritional condition of
the children [21]. This may be particularly interesting
when it comes to children admitted to intensive care
units, where it is difficult to precisely evaluate other
anthropometric variables due to their critical condi-
tion, or in cases of children with large tumors or with
large edema, which surely contribute to alter their
real weight [21].
Currently, two references that allow the classifi-
cation of arm circumference in Z-scores are available
(including software to make it more accurately), the
NCHS/CDC-2000 and WHO-2006, which were
drawn from different population samples. The
NCHS/CDC-2000 reference corresponds to the
NCHS, 1977, which were developed in the first half
of the 70s, from a study of children fed with formula
and breast milk differently from the standard recom-
mended by WHO. In 2000, the CDC conducted a
FIG. 3. Correlation of arm circumference Z-scores review of anthropometric references, including a
estimated from NCHS/CDC – 2000 and WHO–2006 new sample of children younger than 3-years old
references in pre-school children according to and modified the mathematical model used for
their BMI Z-scores. Pearson: r ¼ 0.2945 (95% curves generating and smoothing [13]. On the other
CI ¼ 0.4146 to 0.1644); p < 0.0001. R2 ¼ 0.08675. hand, for arm circumference the EpiNut of the

Journal of Tropical Pediatrics 3 of 5


Epi-Info 3 software continued to be based on data children. J Trop Pediatr 2010 [Epub ahead of print; 8
from NCHS 1977. January 2010].
In 2006, the WHO published the results of their 5. Santos VF, Morais TB. Nutritional quality and osmo-
multicenter study conducted between 1997 and 2003, lality of home-made enteral diets, and follow-up of
growth of severely disabled children receiving
in order to generate new growth curves for evaluation home enteral nutrition therapy. J Trop Pediatr
of children aged between 0 and 60 months; they col- 2010;56:127–8.
lected information from approximately 8500 children 6. Buffa R, Baali A, Lahmam A, et al. Assessment of nu-
from different cultures and ethnicities. The youngest tritional status in the Amazigh children of Amizmiz
children (<18-months old) were exclusively breastfed (Azgour Valley, High Atlas and Morocco). J Trop
during the first 4 months of life, recommended by the Pediatr 2009;55:406–8.
WHO. Based on those findings, the WHO sought to 7. Mei Z, Strawn-Grummer LM, de Onis M, et al. The
establish what would be a normative pattern of development of MUAC-for height reference, incluing a
growth and development of children up to 5-years comparison to other nutritional status screening indica-

Downloaded from at Universidade Federal de S?o Paulo on October 12, 2010
tors. Bull. World Health Organ 1997;75:333–41.
old [8].
8. WHO Multicentre Growth Reference Study Group.
It is also important to be considered that the WHO Child Growth Standards: Head Circumference-
differences between the two references regarding AC for-age, Arm circumference-for-age, Triceps Skinfold-
are higher in absolute values of Z-score, when for-age and Subscapular Skinfold-for-age: Methods
evaluating children which tend to thinness, a fact and Development. Geneva: World Health
that recommends higher caution in the interpretation Organization, 2007.
of this indicator in epidemiological situations of high 9. Magezi SR, Kikafunda J, Whitehead R. Feeding and
prevalence of malnutrition. At the same time, this nutritional characteristics of infants on PMTCT
apparent disadvantage may become an advantage in programs. J Trop Pediatr 2009;55:32–5.
epidemiological nutritional transition periods [22, 23] 10. de Fátima Costa Caminha M, da Silva Diniz A,
Falbo AR, et al. Serum retinol concentrations in
when a significant increase in prevalence of over- hospitalized severe protein-energy malnourished
weight and obesity go through, such as emerging children. J Trop Pediatr 2008;54:248–52.
countries. 11. Semba RD, de Pee S, Sun K, Bloem MW, et al.
We reported that depending on the reference Coverage of the national vitamin A supplementation
(NCHS/CDC-2000 and WHO-2006), it may be program in Ethiopia. J Trop Pediatr 2008;54:141–4.
directed to distinct interpretations with respect to 12. Chandra AK, Debnath A, Tripathy S. Iodine nutrition-
nutritional status; both the references are used in al status among school children in selected areas of
Brazil [20, 24]. We suggest clinical professionals to Howrah District in West Bengal, India. J Trop
consider our data before evaluate clinical status of Pediatr 2008;54:54–7.
13. Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC
the children. Our findings are based on a population
growths charts for the United States: methods and
situated in the southeast of the country; it is possible development. Vital Health Stat 2002;246:1–190.
that different results are found in the north of Brazil. 14. Leone C, Bertoli CJ, Schoeps DO. Novas curvas de
Thus, future investigations in different areas of Brazil crescimento da Organização Mundial da Saúde:
are necessary to verify if our data are applicable in comparação com valores de crescimento de crianças
different regions of the country. pré-escolares das cidades de Taubaté e Santo André,
In conclusion, when analyzing AC in pre-school São Paulo. Rev Paul Pediatr 2009;27:40–7.
children, depending on the reference (NCHS/ 15. Lohman TG, Roche AF, Matorell R. Antropometric
CDC-2000 or WHO-2006) it may be directed to Standardization Reference Manual. Illinois: Human
distinct interpretations with respect to nutritional Kinetics Books, 1988.
16. Hauspie RC, Cameron N, Molinari L. Methods in
status. human growth research. In: Cameron N (ed.).
Measuring Growth. England: Cambridge University
References Press, 2004; pp. 68–107.
1. Briend A, Wojtyniak B, Rowlands MG. Arm circum- 17. Jelliffe DB, Jelliffe EF. The arm circunference as a
ference and other factors in children at high risk of public health index of protein-calorie malnutrition of
death in rural Bangladesh. Lancet 1987;2:725–28. early childhood: current conclusion. J Trop Pediatr
2. Grover Z, Ee LC. Protein energy malnutrition. Pediatr 1969;15:253–60.
Clin N Am 2009;56:1055–68. 18. Frisancho AR. New norms of upper limb fat and
3. Nackers F, Broillet F, Oumarou D, et al. Effectiveness muscle areas for assessment of nutritional status. Am
of ready-to-use therapeutic food compared to a corn/ J Clin Nutr 1981;34:2540–5.
soy-blend-based pre-mix for the treatment of childhood 19. de Onis M, Yip R, Mei Z. The development of
moderate acute malnutrition in Niger. J Trop Pediatr MUAC-for-age reference data recommended by a
2010 [Epub ahead of print]. WHO Expert Committee. Bull World Health Organ
4. Boeke CE, Mora-Plazas M, Forero Y, et al. Intestinal 1997;75:11–8.
protozoan infections in relation to nutritional status 20. de Almeida CA, Del Ciampo LA, Ricco RG, et al.
and gastrointestinal morbidity in Colombian school Assessment of mid-upper arm circumference as a

4 of 5 Journal of Tropical Pediatrics


method for obesity screening in preschool children. 23. Popkin BM. The nutrition transition and its health
J Pediatr 2003;79:455–60. implications in lower-income countries. Public Health
21. Feferbaum R, Delgado AF, Zamberlan P, et al. Nutr 1998;1:5–22.
Challenges of nutritional assessment in pediatric ICU. 24. De Almeida CA, Ricco RG, Del Ciampo LA, et al.
Curr Opin Clin Nutr Metab Care 2009;12:245–50. Comparison of four anthropometric methods of nutri-
22. Popkin BM, Keyou G, Zhai F, et al. The nutrition tional assessment and evaluation of the agreement be-
transition in China: a cross-sectional analysis. Eur J tween two reference populations. J Trop Pediatr
Clin Nutr 1993;47:333–46. 1999;45:345–50.

Downloaded from at Universidade Federal de S?o Paulo on October 12, 2010

Journal of Tropical Pediatrics 5 of 5