[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Research Letter
February 2013

Multiethnic WHO Growth Charts May Not Be Optimal in the Screening of Disorders Affecting Height: Turner Syndrome as a Model

JAMA Pediatr. 2013;167(2):194-195. doi:10.1001/jamapediatrics.2013.436

In 2006, the World Health Organization (WHO) published multiethnic growth charts for children younger than 5 years compiled of growth data of 6669 infants from 6 different countries (the United States, Norway, Oman, Brazil, India, and Ghana).1 Originally, the rationale for the construction of the multiethnic WHO standards stemmed from an earlier study2 that reported nearly similar growth patterns in infants and children from diverse ethnic backgrounds when their physiological needs are met and the environment supports healthy development.

Adult height varies significantly among populations and is strongly genetically determined, with heritability estimates around 80%.3 Therefore, the childhood linear growth patterns are probably not as strongly determined by the optimal nutritional, environmental, and psychosocial factors as the weight patterns. In this respect, the original assumption of uniform growth among populations may be incorrect and adversely affect the screening of growth disorders.

At least 125 countries have adopted the WHO growth charts for nationwide use in childhood growth monitoring (Mercedes de Onis, WHO, written communication, November 7, 2011). Nevertheless, the WHO charts have not been sufficiently tested in the context of the screening of abnormal linear growth in the general population in developed countries. We tested the performance of the WHO charts against a population-specific growth reference in the screening of Turner syndrome (TS). This condition is often regarded as a model condition to justify screening of height disorders because short stature or poor growth may be the only presenting sign, and an early diagnosis is important because of treatable comorbidities and timely introduction of height-promoting therapies to improve adult height.4


Longitudinal height data of 124 Finnish girls with TS (64 with karyotype 45,XO) containing 2020 measurements were assessed either by using the WHO charts or a recently renewed population-specific growth reference based on growth data of 26 636 individuals.1,5 The accuracy of both growth references for detecting growth patterns associated with TS was studied by calculating the cumulative percentage of girls with TS with at least 1 measurement either less than the third or the first height-for-age percentile.


The cumulative percentage (ie, sensitivity) for detecting TS was significantly higher when the population-specific growth reference was used (Figure). By the age of 2 years, 72% and 55% of all girls with TS (81% and 67% of girls with karyotype 45,XO) had at least 1 height measurement less than the third and first percentiles, respectively, in population-specific growth reference. In contrast, only 36% and 20% of all girls with TS (41% and 23% of girls with karyotype 45,XO) had a height measurement less than these percentiles when the WHO standard was used (P ≤ .001, McNemar test). From age 2 to 5 years, the gap in the cumulative sensitivity between the 2 growth references grew even wider, especially if the first percentile was used as a cutoff (Figure).

Figure. Percentage of girls with Turner syndrome (TS) with height for age less than the third (A) or the first (B) percentile of either the Finnish (upper lines) or the World Health Organization (WHO) (lower lines) growth reference for all girls with TS (n = 124, solid lines) and those with an XO karyotype (n = 60, dashed lines).

Figure. Percentage of girls with Turner syndrome (TS) with height for age less than the third (A) or the first (B) percentile of either the Finnish (upper lines) or the World Health Organization (WHO) (lower lines) growth reference for all girls with TS (n = 124, solid lines) and those with an XO karyotype (n = 60, dashed lines).


Height assessment in children growing at the outer percentiles with potential morbidity may be flawed if we use WHO standards. In a recent study from Norway and Belgium, proportions of the healthy children outside the ±2 SDs of WHO standards were different than expected.6 Furthermore, Finnish children are 0.2 to 0.8 SD taller on average than the WHO standards.1,5 Thus, the concept of “similarity” of heights among children from different ethnic backgrounds is questionable because genetic differences have an impact on linear growth.

Our study is, to our knowledge, the first to report the suboptimal accuracy of screening of a height disorder with the WHO standard in comparison with population-specific reference. Further studies in other populations and with other growth disorders are warranted. Before implementing the WHO growth charts in height screening, their performance should ideally be tested in the population they are intended for.

Back to top
Article Information

Correspondence: Dr Dunkel, Centre for Endocrinology, William Harvey Research Institute, Barts and the London, Queen Mary University of London, Charterhouse Square London, London EC1M 6BQ, England (l.dunkel@qmul.ac.uk).

Published Online: December 3, 2012. doi:10.1001/jamapediatrics.2013.436

Author Contributions:Study concept and design: Saari, Sankilampi, and Dunkel. Acquisition of data: Saari and Dunkel. Analysis and interpretation of data: Saari, Sankilampi, and Dunkel. Drafting of the manuscript: Saari, Sankilampi, and Dunkel. Critical revision of the manuscript for important intellectual content: Saari, Sankilampi, and Dunkel. Statistical analysis: Saari. Obtained funding: Dunkel. Administrative, technical, and material support: Sankilampi and Dunkel. Study supervision: Sankilampi and Dunkel.

Conflict of Interest Disclosures: None reported.

World Health Organization Multicentre Growth Reference Study Group.  WHO Child Growth Standards: Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight-for-Height and Body Mass Index-for-Age: Methods and Development. Geneva, Switzerland: World Health Organization; 2006
Graitcer PL, Gentry EM. Measuring children: one reference for all.  Lancet. 1981;2(8241):297-2996114335PubMedGoogle ScholarCrossref
Silventoinen K, Sammalisto S, Perola M,  et al.  Heritability of adult body height: a comparative study of twin cohorts in eight countries.  Twin Res. 2003;6(5):399-40814624724PubMedGoogle Scholar
Stochholm K, Juul S, Juel K, Naeraa RW, Gravholt CH. Prevalence, incidence, diagnostic delay, and mortality in Turner syndrome.  J Clin Endocrinol Metab. 2006;91(10):3897-390216849410PubMedGoogle ScholarCrossref
Saari A, Sankilampi U, Hannila ML, Kiviniemi V, Kesseli K, Dunkel L. New Finnish growth references for children and adolescents aged 0 to 20 years: length/height-for-age, weight-for-length/height, and body mass index-for-age.  Ann Med. 2011;43(3):235-24820854213PubMedGoogle ScholarCrossref
Júlíusson PB, Roelants M, Hoppenbrouwers K, Hauspie R, Bjerknes R. Growth of Belgian and Norwegian children compared to the WHO growth standards: prevalence below -2 and above +2 SD and the effect of breastfeeding.  Arch Dis Child. 2011;96(10):916-92119948662PubMedGoogle ScholarCrossref