3582277563

Research

Development of a WHO growth reference for school-aged children and adolescents

Mercedes de Onis/ Adelheid W Onyango," Elaine Borghi,^J Amani Siyam,¹ Chizuru Nishida^: & Jonathan Siekmann^J

Objective To construct growth curves for school-aged children and adolescents that accord with the WHO Child Growth Standards for preschool children and the body mass index (BMI) cut-offs for adults.

Methods Data from the 1977 National Center for Health Statistics (NCHS)ZWHO growth reference (1-24 years) were merged with data from the under-fives growth standards' cross-sectional sample (18-71 months) to smooth the transition between the two samples. State-of-the-art statistical methods used to construct the WHO Child Growth Standards (0-5 years), i.e. the Box-Cox power exponential (BCPE) method with appropriate diagnostic tools for the selection of best models, were applied to this combined sample.

Findings The merged data sets resulted in a smooth transition at 5 years for height-for-age, weight-for-age and BMI-for-age. For BMI-for-age across all centiles the magnitude of the difference between the two curves at age 5 years is mostfy 0.0 kg/m² to 0.1 kg/m². At 19 years, the new BMI values at +1 standard deviation (SD) are 25.4 kg/m² for boys and 25.0 kg/m² for girls. These values are equivalent to the overweight cut-off for adults (s= 25.0 kg/m²). Similarly, the +2 SD value (29.7 kg/m² for both sexes) compares dosely with the cut-off for obesity 30.0 kg/m²}.

Conclusion The new curves are dosely aligned with the WHO Child Growth Standards at 5 years, and the recommended adult cut-offs for overweight and obesity at 19 years. They fili the gap in growth curves and provide an appropriate reference for the 5 to 19 years age group.

Bulletin of the World Health Organization 2007,85:660-667.

Uoe uaduam en hanf¹s dece /dswrop figuie a la fx>de farflcte. Al hnalcW amato se /aaWa una traduiodn ai espaAol. .łxiy ¹1^    ^ i^MiJi ¹;>¹!

Introduction

The nccd to dcvclop an appropriate single growth reference for the scrccning, surveillancc and monitoring ot school-aged children and adolescents has becn stirred by two contcmporary cvcnts: the inereasing public health conccrn ovcr childhood obesity¹ and the April 2006 rclcasc of the WHO Child Growth Standards for preschool children based on a prcscriptivc approach.¹ As countrics procccd with the implcmcntation of growth standards for children under 5 years of age. the gap across all centiles between these standards and cxisting growth rcfcrcnccs for older children has bccomc a matter of great conccrn. It is now widcly acccptcd that using descrip-tivc samples of populations that rcflcct a secular trend towards ovcrwcight and obesity to construct growth rcfcrcnccs rcsults inadvcrtently in an undesirable upward skewness leading to an underes-timation of ovcrwcight and obesity, and an ovcrcstimation of undcrnutritionT The reference prcviously recommended by WHO for children abovc 5 years of age. i.e. the National Center for I Tealth Statistics (NCTTSJ/WHO inter-national growth reference.¹ has scvcral drawbacks.' In partieubr. the body mass indcx-for-agc reference, dcvcloped in 1991." only starts at 9 years of age, groups data annually and covcrs a lim-ited pcrccntilc rangę. Many countrics pointed to the nccd to havc body mass indcx (BMI) curvcs that start at 5 years and permit unrcstrictcd calculation of pcrccntilc and z-scorc curvcs on a con-tinuous age scalę from 5 to 19 years.

'lhc nccd to harmonize growth as-sessment tools conccptually and prag-matically prompted an cxpcrt group mecting in January 2006 to cvaluatc the fcasibility of dcvcloping a single inter-national growth reference for school-aged children and adolescents. ^a The cxpcrts agreed that appropriate growth rcfcrcnccs for these age groups should be dcvclopcd for clinical and public health applications. Ihcy also agreed that a multiccntrc study, similar to the one that led to the dcvclopment of the WłTO Child Growth Standards for 0 to 5 years. would not be fcasiblc for older children. as it would not lic possiblc to control the dynamics of their cnviron-ment. Thcrcforc, as an alternativc, the cx-perts suggested that a growth reference be constructcd for this age group using cxisting historical data and discusscd the critcria for selccting the data sets.

WHO subscquently initiated a pro-ccss to identify cxisting data sets from various countrics. This proccss resulted in an initial identification of 115 candi-datc data sets from 45 countrics, which were narrowed down to 34 data sets

660

1

Department of Nutribon, World Health Organization, 20 Avenue Appia, 1211 6eneva 27. Switzerland. Corresponóence to Mercedes de Onis (e-mall: deonism®«vho.mt). doi: 10.2421JBLI.07.043497

(Suómitred: 25 April 2007 - Finał temed veision receu ed. )2 Tuty 2007 - Acceęied: 15 Suty 2007)

Bulletin of the World Health Orgamzation | September 2007,85 (9)