Adc 96 11 1008
Adc 96 11 1008
Adc 96 11 1008
prevalence will change when WHO standards, rather than a humanitarian emergency; over 30% a famine/humanitarian
NCHS references, are used to de ne cases. catastrophe. We emphasise that the IPC cut-offs for acute mal-
To contextualise our ndings, we compared infant nutrition are not normally applied to single age groups, and that
under-6-month wasting with that in older children (from 6 to anthropometric indicators, on their own, are not normally used
under 60 months) from the same populations. to classify emergency situations. They serve here to demon-
strate the extent of differences between infants under 6 months
METHODS and children, and the NCHS and WHO growth norms.
Study design, setting and population Second, we differentiated between severe and moderate
We performed secondary analysis of 21 demographic and wasting, predicting WHO-based prevalence from NCHS-
health survey (DHS) datasets. DHS are large national sur- based prevalence using univariable linear regression.
veys, standardised across and within countries ( http://www. Third, to illustrate the implications for treatment pro-
measuredhs.com/ ). grammes, we estimated the numbers of individuals affected.
We selected 21 countries from a reference population of Population statistics were from the 2004 United Nations
36 that account for the majority of the global malnutrition population database.9,15 We assumed that infants under
disease burden 9 and that had available DHS anthropometry 6 months were 1/10th of the total 0 to under 60 months
data collected in the past 10 years. We registered our pro- population.16 We accounted for differences in population
ject via http://www.measuredhs.com/accesssurveys/access_ size by calculating a population weighting for each coun-
instructions.cfm. try. Assuming that our 21-country sample was represen-
tative, we extrapolated the pooled, weighted prevalence
estimate to the population in all developing countries. The
Variables and data handling objective was to illustrate the magnitude of NCHS/WHO
Current de nitions of wasting10 are summarised in table 1. changes rather than to derive de nitive statistics. We lacked
We calculated NCHS z-scores from weight, height/length, the information to calculate useful con dence intervals.
age and sex variables using Emergency Nutrition Assessment To assess our estimate validity, we compared our gures
for software for standardised monitoring and assessment of against other published data.
relief and transitions (SMART).11 Extreme values are more Finally, to improve understanding of why wasting preva-
likely to represent measurement or database errors than an lence changes, we used published NCHS17 and WHO18 tables
individual who is truly very small or very large. Following to plot WHZ 3 and 2 cut-off curves.
commonly used nutrition survey criteria,12 we thus excluded
individuals with: weight-for-height z-score (WHZ) (NCHS)
<4 or >+ 6; or weight-for-age z-score (WAZ) (NCHS) < 6 or RESULTS
>+ 6; or height-for-age z-score (HAZ) (NCHS) < 6 or >+ 6; In our 21-country sample, 15 534 infants under 6 months and
or incompatible combinations of HAZ and WHZ: (HAZ >3.09 147 694 children aged from 6 to under 60 months had a valid
and WHZ <3.09) or (HAZ <3.09 and WHZ >3.09). We calcu- WHZ (NCHS). Survey details are shown in supplementary
lated WHZ (WHO) for these same individuals. appendix 1 (available online only).
Figure 1 shows wasting prevalence by country and by age
group. The prevalence of wasting in infants under 6 months
Sample size is related to the prevalence of child wasting: r2 =0.66 (using
The DHS survey size is large enough for robust national NCHS), r2 =0.84 (using WHO). Prevalence is lowest using
prevalence estimates.13 To determine whether our sample NCHS-based case de nitions: 1.115%, (median 3.7%, IQR
of 21 countries was reective of all developing countries, we 1.86.5%). Seven of the 21 countries have acceptably low
compared our ndings against other published data. (<3%) wasting. WHO-based prevalence is higher: 2.034%
(median 15%, IQR 6.217%). Only one country remains in the
Data analysis acceptable category. Among children (gure 1b) NCHS/WHO
Using SPSS version 16 and Excel 2003 we performed three differences are minimal.
analyses. Figure 2 separates severe and moderate wasting. Highlighted
First we looked at country-level wasting prevalence using by the steeper slope of the regression line, WHO standards
the international integrated food security phase classica- result in more diagnoses of severe wasting, particularly
tion (IPC).14 This is used to determine the severity of an emer- among infants under 6 months. Moderate infant wasting also
gency and guide the need for interventions: more than 3% to increases when using WHO standards. In contrast, moderate
under 10% wasting prevalence reects moderately food inse- child wasting decreases.
curity; 1015% an acute food and livelihood crisis; over 15% Table 2 shows regression equations for gure 2. Regression
slopes indicate the magnitude of change in wasting prevalence
Table 1 Case definitions of wasting using NCHS growth references when case de nitions change from NCHS to WHO: severe
and WHO growth standards6 wasting in infants under 6 months is 3.5 times greater and
severe child wasting 1.7 times greater. Moderate wasting in
Weight-for-height z-score*
infants under 6 months is also greater with WHO standards.
NCHS WHO In contrast, moderate child wasting decreases with WHO.
Wasting <2 <2 Table 3 presents wasting in terms of the numbers affected.
Moderate 3 to <2 3 to <2 Rounded gures emphasise that these are estimates and assume
Severe <3 <3 that our sample is representative of all developing countries.
(eg, z-score 1 = 1 SD below mean)
Finally, gure 3 shows WHO and NCHS WHZ 3 and WHZ
*z-scores represent SD below the National Center for Health Statistics (NCHS) or 2 cut-off curves for boys. Girls curves are similar and are
WHO population mean (eg, z-score -1 = 1 standard deviation below mean). not shown. The gap between WHZ (WHO) and WHZ (NCHS)
Figure 1 Country prevalence of wasting (<2 weight-for-height z-score) as defined by National Center for Health Statistics (NCHS) growth
references (striped) and WHO growth standards (shaded). Countries are ordered by increasing infant under-6-month wasting prevalence (NCHS).
Boxed comments (ie, Famine, Humanitarian emergency) refer to the integrated food security phase classification, IPC see Methods section.
(A) Wasting prevalence among infants aged from 0 to under 6 months. (B) Wasting prevalence among children aged from 6 to under 60 months.
Figure 2 Scatter plot of country prevalence of (A) severe and (B) moderate wasting (weight-for-height z-score <3 and 3 to <2
respectively), as diagnosed using either National Center for Health Statistics (NCHS) or WHO growth norms. Regression and identity lines are
shown. Each country survey is represented by one filled and one unfilled circle.
Table 2 Univariable linear regression models predicting percentage wasting prevalence (WHO) from wasting prevalence (NCHS)
Regression slope 95% CI (slope) Constant (%) 95% CI (constant) Pearsons r Residual SD (%)
Severe infant wasting 3.54 (2.6 to 4.4) 2.3 (0.9 to 3.7) 0.88 2.1
Severe child wasting 1.68 (1.5 to 1.8) 0.1 (0.2 to 0.5) 0.98 0.4
Moderate infant wasting 1.43 (1.1 to 1.8) 2.0 (0.4 to 3.6) 0.89 1.9
Moderate child wasting 0.86 (0.8 to 0.9) 0.0 (0.4 to 0.4) 0.99 0.5
Table 3 Approximate numbers of infants and children in all our sample being representative of other developing countries.
developing countries (millions) affected by severe and moderate A 2006 review quoted 13.1 million and 47.1 million as severely
wasting, as diagnosed using NCHS and WHO weight-for-height and moderately wasted, respectively, using NCHS. 22 These
z-score gures are based on a mix of surveys, some including others
excluding infants under 6 months. Our NCHS-based estimates
Infants Children Total
(infants under 6 months and children combined) are 9.3 million
0<6 months 6<60 months 0<60 months
severely and 40.7 million moderately wasted. A 2008 review
n=55.5 million n=500 million n=555.5 million using WHO standards quoted 19.3 million severely wasted.9
Severe wasting NCHS 0.8 8.5 9.3 Our gure is 19.8 million.
weight-for-height WHO 3.8 16 20
<3 z Implications for policy and practice
Moderate wasting NCHS 2.2 38 41 First, we recommend that nutrition surveys more routinely
weight-for-height WHO 4.7 34 38
3 to <2z include infants under 6 months. Our data help estimate infant
wasting prevalence, but specic settings are likely to have spe-
NCHS, National Center for Health Statistics. cic epidemiological patterns.
Second, we suggest that programmes should consider their
same raw anthropometric measurements (whether accurate or capacity to treat infants who are identied as wasted. This is
not) are used to calculate an individuals WHZ (NCHS) and particularly important before adopting WHO-based case de-
WHZ (WHO), the effect on NCHS/WHO changes is unlikely nitions. Many programmes already struggle to deal with the
to be marked. smaller number of NCHS-diagnosed wasted infants.4
Finally, we call for a review of the effects of diagnosing
Strengths and weaknesses in relation to other studies greater numbers of infants under 6 months as wasted. Current
The results extrapolated from our 21-country dataset are com- treatment guidelines focus on the anthropometry for diag-
parable to gures cited elsewhere.9, 22 This is consistent with nosing infant wasting.4 An increased survey prevalence thus
Figure 3 Difference in WHO and National Center for Health Statistics (NCHS) 2 and 3 z-score cut-offs. Arrows on the figure show median
length/height at different ages for boys (using WHO growth standards).
equates to greater numbers eligible for treatment. This has of both severe and moderate infant under-6-month wasting.
possible risks: the evidence base underlying current treat- Policy makers and programme managers should consider the
ments for infants under 6 months is weak;4, 23 if clinically implications of this change. An international policy state-
well, exclusively breastfed infants under 6 months are labelled ment on infant under-6-month wasting would ll an impor-
as small (ie, below 3 or 2 WHZ), mothers might become tant gap because neither the 2009 statement on WHO growth
concerned 24, 25 and inappropriately introduce top-up foods standards6 nor the 2007 statement on the management of
or breastmilk substitutes. 26 This would have adverse con- wasting3 address this age group.
sequences given the well-documented protective effects of
exclusive breastfeeding.9, 27 Contributors MK, MM and AS conceptualised the study. MK drafted the initial
manuscript. HB and MK performed the main data analysis. CGE, JS, TJC and AS
contributed further analyses. All authors contributed to the development of the
Unanswered questions final manuscript.
The generalisability of our results could be con rmed by Acknowledgements The authors thank MEASURE DHS (Macro International
examining other datasets. Work is also needed to explore risk Inc, Calverton, USA) and all countries surveyed for DHS datasets. The authors also
factors for wasting in infants under 6 months and to deter- thank Melody Tondeur for helpful comments on an earlier draft of the paper. Full
mine which infants benet most from which treatments. DHS datasets used in this paper are available from ORS Macro, USA.
We were unable to explore why age-related differences in Funding This paper was written as part of the MAMI (Management of Acute
NCHS and WHO growth curves are so marked. We recogn- Malnutrition in Infants) project, funded by the UNICEF led Inter Agency Standing
Committee Nutrition Cluster (www.humanitarianreform.org/humanitarianreform/
ise that WHO standards represent an important advance on Default.aspx?tabid=74). MK, CG, MM and AS were all part-funded by MAMI. TJC
NCHS references. 28 We note their technical superiority and was funded by the Medical Research Council (grant number G0700961).
that they are based on a highly selected population of healthy, Competing interests None.
optimally fed infants with relatively low statistical variance. 29
Provenance and peer review Not commissioned; externally peer reviewed.
However, this gold standard of growth could be difcult to
achieve for many infants in developing countries.
For clarity, we examined only z-score case de nitions, REFERENCES
1. Manary MJ, Sandige HL. Management of acute moderate and severe childhood
which are preferred for nutrition reporting. For admissions, malnutrition. BMJ 2008;337:a2180.
many feeding programmes also use weight-for-height percent- 2. Victora CG, Adair L, Fall C, et al. Maternal and child undernutrition:
age of median.7 Moving from weight-for-height percentage consequences for adult health and human capital. Lancet 2008;371:340 57.
of median (NCHS) to WHZ (WHO) may result in different 3. World Health Organization (WHO). Community-based management of severe
acute malnutrition. A joint statement by the World Health Organization, the World
changes to those described. Food Programme, the United Nations System Standing Committee on Nutrition
Finally, we suggest that alternative diagnostic criteria for and the United Nations Childrens Fund, 2007. http://www.who.int/nutrition/
infants under 6 months be considered alongside anthropom- topics/statement_commbased_malnutrition/en/index.html. (accessed 19
etry: for example, different z-score cut-offs; mid-upper arm September 2010).
4. Emergency Nutrition Network, UCL Centre for International Health and
circumference, 30 body mass index31, 32 and clinical criteria.
Development, Action Contre la Faim. Management of Acute Malnutrition in
Infants (MAMI) project, 2009. http://www.ucl.ac.uk/cihd/research/nutrition/
CONCLUSIONS mami. (accessed 19 Sept 2010).
5. de Onis M, Garza C, Onyango AW, et al. WHO child growth standards.
Wasting among infants under 6 months is prevalent in many of Acta Paediatr 2006;Issue supplement S450;7.
the developing countries examined in this study. Using WHO 6. World Health Organization (WHO). WHO child growth standards and the
standards to de ne wasting results in a greater prevalence identification of severe acute malnutrition in infants and children. A joint
statement by the World Health Organization and the United Nations Childrens 19. Prost MA, Jahn A, Floyd S, et al. Implication of new WHO growth standards
Fund, 2009. http://www.who.int/nutrition/publications/severemalnutrition/97892 on identification of risk factors and estimated prevalence of malnutrition in rural
41598163/en/index.html. (accessed 19 Sept 2010). Malawian infants. PLoS ONE 2008;3:e2684.
7. Kerac M, Egan R, Mayer S, et al. New WHO growth standards: roll-out needs 20. de Onis M, Onyango AW, Borghi E, et al. Comparison of the World Health
more resources. Lancet 2009;374:100 2. Organization (WHO) child growth standards and the National Center for Health
8. Inter-Agency Standing Committee (IASC). Nutrition cluster informal consultation. Statistics/WHO international growth reference: implications for child health
Geneva: IASC, 2527 June 2008. programmes. Public Health Nutr 2006;9:9427.
9. Black RE, Allen LH, Bhutta ZA, et al. Maternal and child undernutrition: 21. Myatt M, Khara T, Collins S. A review of methods to detect cases of severely
global and regional exposures and health consequences. Lancet malnourished children in the community for their admission into community-
2008;371:243 60. based therapeutic care programs. Food Nutr Bull 2006;27:S723.
10. World Health Organization (WHO). Management of severe malnutrition: a 22. Collins S, Dent N, Binns P, et al. Management of severe acute malnutrition in
manual for physicians and other senior health workers, 1999. http://www.who. children. Lancet 2006;368:19922000.
int/nutrition/publications/en/manage_severe_malnutrition_eng.pdf. (accessed 23. World Health Organization (WHO). Severe malnutrition: report of a consultation to
19 September 2010). review current literature. Geneva, Switzerland: WHO, 2004.
11. Emergency Nutrition Assessment (ENA). Software for standardized monitoring 24. Sachs M, Dykes F, Carter B. Feeding by numbers: an ethnographic study of how
and assessment of relief and transitions (SMART), version October 2007. http:// breastfeeding women understand their babies weight charts. Int Breastfeed J
www.nutrisurvey.de/ena/ena.html. (accessed 19 Sept 2010). 2006;1:29.
12. Dean A, Dean J, Coulombier D, et al. Epi Info, version 6: a word processing, 25. Laraway KA, Birch LL, Shaffer ML, et al. Parent perception of healthy infant and
database, and statistics program for public health on IBM-compatible toddler growth. Clin Pediatr (Phila) 2010;49:343 9.
microcomputers (user manual). Atlanta, GA: Centers for Disease Control and 26. Binns C, Lee M. Will the new WHO growth references do more harm than good?
Prevention, 1996. http://www.cdc.gov/epiinfo/Epi6/ei6manl.htm. (accessed Lancet 2006;368:1868 9.
19 September 2010). 27. Victora CG, Smith PG, Vaughan JP, et al. Evidence for protection by breast-
13. Aliaga A, Ren R. Optimal sample sizes for two-stage cluster sampling feeding against infant deaths from infectious diseases in Brazil. Lancet
in demographic and health surveys. Working paper 30. Calverton, MA: 1987;2:319 22.
ORC Macro, 2006. http://www.measuredhs.com/pubs/pub_details. 28. Garza C, de Onis M. Rationale for developing a new international growth
cfm?ID=589&srchTp=type. (accessed 19 September 2010). reference. Food Nutr Bull 2004;25(Suppl 1):S5 14.
14. IPC Global Partners. Integrated food security phase classification technical 29. WHO Multicentre Growth Reference Study Group. Enrolment and baseline
manual, version 1.1. Rome: FAO, 2008. ISBN: 978-92-5-106027-8. Reprint 2009. characteristics in the WHO Multicentre Growth Reference Study.
15. UN Department of Economic and Social Affairs, Population Division. Acta Paediatr Suppl 2006;450:715.
World population prospects, 2004 revision. New York: United Nations, 2005. 30. Mathenge A, Mwangome M, Fegan G, et al. Assessment of severe malnutrition
16. United Nations Statistics Division. Demographic yearbook, 2007. http://unstats. among hospitalized Kenyan infants under 6 months old. Abstract presented at
un.org/unsd/demographic/products/dyb/dyb2007.htm. (accessed 18 September CAPGAN meeting, Commonwealth Association of Paediatric Gastroenterology
2010). and Nutrition). Malawi Med J 2009;21:25.
17. World Health Organization (WHO). Measuring change in nutritional status. 31. Cole TJ, Flegal KM, Nicholls D, et al. Body mass index cut offs to define thinness
Geneva: WHO, 1983. in children and adolescents: international survey. BMJ 2007;335:194.
18. World Health Organization (WHO). Child growth standards, 2006. http://www. 32. Cole TJ. A critique of the NCHS weight for height standard. Hum Biol
who.int/childgrowth/en/. (accessed 18 September 2010). 1985;57:183 96.