Article Text

Download PDFPDF
Are universal standards for optimal infant growth appropriate? Evidence from a Hong Kong Chinese birth cohort
  1. L L Hui1,
  2. C M Schooling1,
  3. B J Cowling1,
  4. S S L Leung2,
  5. T H Lam1,
  6. G M Leung1
  1. 1
    Department of Community Medicine and School of Public Health, The University of Hong Kong, Hong Kong, People’s Republic of China
  2. 2
    Department of Health, The Government of Hong Kong Special Administrative Region, Hong Kong, People’s Republic of China
  1. C Mary Schooling, The University of Hong Kong, Department of Community Medicine, 5/F William MW Mong Block, Room 5-21, 21 Sassoon Road, Pokfulam, Hong Kong, People’s Republic of China; cms1{at}


Objective: In 2006 the World Health Organization (WHO) published new optimal growth standards for all healthy infants worldwide. To assess their general applicability to a recently transitioned Chinese population, we compared them with infant growth patterns in a representative sample of Hong Kong infants.

Design and settings: Weight at birth and at 1, 3, 9, 12, 18 and 36 months, length at 3 and 9 months and height at 36 months were obtained for over 80% of all infants born in April and May 1997 (3880 boys and 3536 girls). Age and sex specific z scores were calculated relative to the WHO growth standards for term singletons.

Results: Weight for age was close to the 50th percentile of the WHO growth standards for both boys (mean z score: 0.00) and girls (0.04) at most time points before 3 years of age. However, our participants were shorter at 3 years, where the z scores in height were −0.34 and −0.38 for boys and girls, respectively. Restricting the analysis to a subset matching the WHO criteria for healthy infants without restrictions on growth gave similar results.

Conclusions: Although the WHO study group concluded there was a striking similarity in length/height among different populations, Hong Kong Chinese toddlers are, on average, shorter. Epigenetic constraints on growth coupled with the rapid epidemiological transition in Hong Kong may not have allowed sufficient generations for infants and children to reach their full genetic height potential, and with it the WHO standards. A universal infant growth standard may not be appropriate across all populations.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Growth charts or standards provide a normative reference to monitor child health. Growth charts are in common clinical use to check individual growth trajectories so that deviations indicating potential health problems can be detected. At a population level growth standards, particularly for height, are also used to monitor whether child health and care needs are being met effectively such that children are achieving their full growth potential.1

Growth charts have usually been specifically developed for particular populations at certain points in time. For instance, locally developed growth charts from a cross-sectional survey in 1993 are currently used in Hong Kong.2 In contrast, in 2006, the World Health Organization (WHO) issued growth charts which provide universal standards for optimal growth of all healthy infants worldwide, with general applicability for current and future generations.1 3 However, these growth standards could be invalid or even misleading for the one fifth of the global population in East Asia. Firstly, the WHO sample for the growth charts did not include infant populations from China, who were initially included but later dropped,4 or elsewhere in East Asia.5 Secondly, the WHO assumes and states that under optimal environmental conditions infants and children can achieve their full genetic height potential within one generation,3 without regard to their parents’ size. However, there is a limit to inter-generational height increases such that incremental increases in height take place over many generations.6

We took advantage of a recent, large, well-characterised and population-representative Hong Kong birth cohort from 1997 to test the hypothesis that the WHO growth standards have universal applicability. We examined whether the growth of this Hong Kong cohort met the new WHO weight, length/height and body mass index (BMI) standards7 both for the entire sample and for a subset who met the WHO criteria of living in “optimal” growth conditions.


Subject recruitment and data collection

The 1997 Hong Kong birth cohort is a population based study of 8327 infants born in Hong Kong in April and May 1997, designed primarily to investigate the health impact of second-hand smoke exposure.8 Recruitment took place at the Maternal and Child Health Clinics (MCHCs), which provide postnatal care including immunisations, physical examinations and growth monitoring at no cost to infants and children from birth to 5 years. Almost all (92%) infants born in Hong Kong attend an MCHC at least once.9 Recruitment was aimed at ethnic Chinese, comprising >95% of the population, and there was a 95% response rate.8 Thus, this cohort covers 88% of all relevant births, but over 90% of the ethnic Chinese.

The participants’ main care providers (mostly mothers) were interviewed shortly after birth, and at 3, 9 and 18 months. Information on socio-economic status (education and type of housing), birth characteristics, feeding, smoking and proxy-reported health service use was collected. Weight and recumbent length in infants or height in toddlers and children were measured, to 0.1 kg and 0.1 cm, respectively, by nursing staff and recorded on medical charts whenever the participants were brought to the centres (eg, for vaccinations or well-baby checks). In 2005, all anthropometric measurements were retrieved from the original medical records. The study was approved by the University of Hong Kong-Hospital Authority Hong Kong West Cluster Joint Institutional Review Board and the Ethics Committee of the Department of Health, The Government of Hong Kong Special Administrative Region, People’s Republic of China.

Inclusion and exclusion criteria

Consistent with the scope of the WHO standards, we excluded pre-term (<37 weeks’ gestation) infants and multiple births and included term low-birth-weight infants. We also excluded measurements more than three standard deviations from the sample median.7

Outcome measure

Our outcome was infant weight, length/height and BMI for age from birth to 3 years of age.

Statistical analysis

We calculated z scores relative to the WHO standards with 95% confidence intervals, to assess whether age and sex-specific weight, length/height or BMI differed significantly from the WHO standards. The z scores for weight were regressed on age in months to test for a trend over time. All z scores were reported to two decimal places.

MCHCs are primarily intended for well-baby checks. However, low-birth-weight or sick infants may make more MCHC visits, thus contributing disproportionately to the measurements. We used only the closest measurement per participant within 1 month of the scheduled well-baby checks at 1, 3, 9, 12, 18 and 36 months for weight, and at 3, 9 and 36 months for length/height, except at 1 month and 36 months where we used measurements within 15 days and 4 months, respectively. We checked measurements which changed by >1.5 in z score between consecutive measurements. Length/height measurements at the three time points were not always available; we checked if those with and without measurements differed significantly in terms of socio-economic status and stature at 7 years.

In addition to comparing the entire birth cohort with the WHO standards, we repeated the analysis for a sub-sample selected to match the WHO criteria for living in an “optimal” environment. The WHO criteria are: exclusive or predominant breastfeeding for at least 4 months, introduction of complimentary foods by 6 months, no maternal smoking, no socio-economic constraints on growth and an absence of significant morbidity that might affect growth (such as repeated episodes of diarrhoea).1 Similarly, we selected our “optimal” subset as infants of non-smoking mothers who were still breastfed at 3 months. Hong Kong infants in the 1990s were more than adequately fed,10 so we did not exclude any infants on the basis of late introduction of complementary feeding. We did not exclude any infants due to socio-economic constraints on growth as fewer than 1% of children from the two countries most economically comparable (in terms of gross domestic product (GDP) per capita) to Hong Kong (ie, Norway and USA) were excluded on this basis.5 However, we excluded infants who had been hospitalised for diarrhoea or intestinal related diseases during the first 18 months of life. To ensure we had comprehensively excluded infants who might have any morbidity affecting growth, we also carried out a sensitivity analysis further excluding any infants with any hospitalisation in the first 18 months of life.

Statistical analyses were performed using R version 2.3.1 (R Development Core Team, Vienna, Austria). We used the Akima package in R to interpolate the WHO standards onto a daily scale, so that we could calculate accurate z scores at exact daily ages.


Sample characteristics

After excluding twins (n = 123), non-ethnic Chinese (n = 20), pre-term infants (n = 426) and 342 with missing gender or gestational age, there were 7416 infants, including 64 boys and 97 girls with low birth weight (<2.5 kg). There were 3880 boys and 3536 girls with 20 898 and 19 119 weight measurements and 7956 and 7230 length/height measurements before 40 months of age, respectively, after excluding 0.4% of the weight and 0.2% of the length/height measurements as being more than three standard deviations from the sample median and allowing only one measurement per participant at the appropriate ages. There were no significant differences in parental education or participants’ height at 7 years between those with and without length/height measurements at routine follow-up time points (data not shown). Only 369 males and 358 females (∼10%) met the criteria for inclusion in the “optimal” subset; most exclusions were due to formula feeding.

Comparison with WHO growth standards

Mean birth weight in both boys and girls were lower than the WHO standards (mean z scores −0.16 and −0.14, respectively). However, subsequent weights slightly exceeded the standards until 36 months, when the mean weight in boys was comparable to the WHO growth standards, while girls were lighter (mean z score −0.14) (table 1). There was no evidence of an age dependent trend in weight z scores for boys or girls (data not shown).

Table 1 Mean and 95% confidence intervals (CI) for z scores of weight, length/height and BMI, with reference to the WHO growth standards

Length at 3 and 9 months was comparable to the WHO standards in girls (mean z scores −0.05 and −0.02, respectively), while boys were slightly short (mean z scores −0.18 and −0.19, respectively). At 36 months, both boys and girls were shorter (z scores −0.34 and −0.38, respectively). Participants with height measurement at 36 months had similar lengths at 3 and 9 months (in boys and girls at 3 and 9 months, respectively: −0.18, −0.06, −0.19 and −0.04) to those without height at 36 months (−0.18, −0.03, −0.18 and 0.00).

Mean BMI was higher in both sexes at 3, 9 and 36 months. The z score standard deviations were consistently below 1, because Hong Kong Chinese are more homogeneous than the six different ethnic groups used to construct the WHO standards. Weight, length/height and BMI were similar for the optimal subset, except that optimal boys were heavier at birth and less short at 3 months, while BMI was greater for both optimal boys and girls (table 2). A sensitivity analysis using a more restricted optimal subset of 259 boys and 327 girls with no hospital admission in the first 18 months of life gave very similar results to the original optimal subsets (data not shown).

Table 2 Mean and 95% confidence intervals (CI) for z scores of weight, length/height and BMI, with reference to the WHO growth standards, in the optimal group*


The underlying assumption for establishing an international growth standard for infants and children is that “children in all countries can achieve their full growth potential when their nurturing follows health recommendations and care practices associated with healthy outcomes”.3 Field tests in four less developed countries showed that the WHO growth standards were effective at detecting stunting and underweight, thus demonstrating that “the standards were technically and clinically sound”.11 On the other hand, our population based sample of full-term, singleton infants in the affluent society of Hong Kong met the standards for weight from birth to 3 years, with no evidence of weight faltering, but were shorter, particularly at 3 years, with linear growth below the WHO trajectory. Although Hong Kong 3-year-old children were on average shorter by only about 1.3 cm, this represents a difference in the entire height distribution, which in keeping with Rose’s prevention paradox,12 may represent large impacts on population health across the life course.

Length/height was not always measured on schedule at the MCHC. However, the availability of length/height measurements appears to be due to individual practice at the MCHCs rather than any attributes of the infants or their families. As measurements were retrieved from medical records, they could not be explicitly checked, however we eliminated inconsistent measurements and excluded length/height measurements taken between scheduled growth checks, to avoid over-representation of participants whose poor growth might be a cause of concern. Moreover, length and height in this study were similar to a smaller cross-sectional study from Hong Kong in 1993.13 Finally, we were constrained by the MCHC policy on growth monitoring to length at 3 and 9 months and height at 36 months; however, major and highly implausible catch-up growth after 3 years would be necessary for these infants to achieve the WHO standards for height thereafter.

Growth faltering in weight and length at around 6–12 months compared to the NCHS growth charts14 was reported in a previous generation of Hong Kong infants (born in the late 1960s and early 1970s)15 when economic development was less advanced.16 However, weight faltering was not evident currently, but there was still catch-down linear growth compared to the WHO standards. Hong Kong is now an affluent society, where, as in China, children are treasured, so it is unlikely that shortness is due to a less than optimal nurturing environment that restricts infant growth. The breastfeeding rate is low (20.4% at 1 month old) in Hong Kong.17 However, formula-fed infants usually grow faster in the entire infancy period.1820 Growth of the optimal subset, breastfed until 3 months, resembled the whole sample, so formula feeding does not explain the shorter height of Hong Kong toddlers.

We cannot rule out the possibility that southern Chinese are genetically shorter than northern Chinese21 22 and other ethnic groups.23 However, Chinese born and raised aboard are also taller than those living in China.24 Secular increases in early linear growth usually occur with improved living conditions, and make a major contribution to the secular trend in adult height across generations,6 which takes about 150 years or six generations to reach genetic height potential.6 Hong Kong Chinese are mainly first or second generation migrants from southern China, who moved from an environment with living conditions and heights similar to pre-industrial Europe25 26 to a rapidly developing economy with a similar GDP per head to western Europe by the 1990s.27 Hong Kong Chinese people are small (contemporaries of the birth cohorts’ parents were 7 cm shorter than the parents of the WHO study infants3 28) and have a short history of economic development, with potentially insufficient generations to realise their full genetic height potential, possibly because of epigenetic constraints on inter-generational increases in size. Unexplained “height growth limitation” has also been observed in long-term industrialised population,29 where obese children from low socio-economic status families were shorter than other obese children, despite more than adequate nutrition. One potential explanation is that the children’s families underwent different pathways over generations of economic transition, leaving varying epigenetic imprints of growth restraint. Moreover, we have shown that the effect of height on cardiovascular risk may be epidemiological stage specific and related to the population history,30 31 suggesting that height is not a marker of childhood environment alone.

While this study provides some support for our hypothesis that the population history of Hong Kong Chinese may have precluded the current generation of infants from reaching the WHO height standards, shortness before 1 year was less notable. There are two possible explanations. Hong Kong infants may usually have a lower length/height trajectory, which is currently obscured by the growth promoting effects of formula feeding.18 19 However, breastfed infants in the optimal subset had a similar trajectory. Alternatively, the physiological mechanisms controlling growth in infancy, childhood and puberty may be different.32 Factors controlling infant growth are not fully understood. Growth hormone may play a role,33 34 as may insulin growth factor I35 independently of growth hormone,36 although the growth hormone axis may act differently during the first year of life than during childhood.37 Epigenetic controls on the expression of growth hormone38 39 could result in epigenetic imprinting of growth restraint mainly after infancy during the childhood growth phase and might explain why our children were obviously short at 3 years of age.


Although the WHO study group concluded there was a striking similarity in length/height among children of different populations,3 Hong Kong Chinese toddlers are, on average, shorter compared with the WHO growth standards. We speculate this is the result of epigenetic constraints on growth rather than failure to thrive or stunting due to less optimal living conditions. Inappropriate expectations concerning length/height for age at an individual level could result in unnecessary worry and over-feeding. At a population level, use of inappropriately long length/height standards could result in “normalising” high BMI and overweight. Although the WHO growth standards are intended to be applicable to any population and indeed could be used to observe growth trajectories for Hong Kong infants, our findings indicate that the standards are probably not suitable for Hong Kong Chinese, and by extension for any other infants from East Asia with a similar socio-economic developmental history and macro environment. We suggest that the history of economic development may impact on the growth in length/height in children. It is unlikely that a single universal growth standard for infants and children could be valid.

What is already known on this topic

  • The new WHO growth standards for optimal infant and childhood growth are intended to be applicable to any population, although the reference populations did not include East Asians.

  • The WHO study group concluded there was a striking similarity in length/height among children of different populations.

What this study adds

  • Hong Kong Chinese toddlers at 3 years of age are, on average, shorter compared with the WHO growth standards. We speculate this is the result of epigenetic constraints on growth rather than failure to thrive or stunting due to less optimal living conditions.

  • Our study challenges the universal applicability of the growth standards and the underlying assumption that children in all countries can achieve their full growth potential when nurtured under optimal conditions.


We thank the Family Health Service, Department of Health, The Government of the Hong Kong Special Administrative Region for collaborating in the study and facilitating the recruitment and follow-up of subjects, and Dr LM Ho, Keith Tin and Eileen Yeung for providing assistance in data extraction and collation.


View Abstract


  • Funding: This study was supported by the Hong Kong Health Care and Promotion Fund (grant no. 216106) and the Health and Health Services Research Fund (grant no. 03040771), Hong Kong, China.

  • Competing interests: None.

Linked Articles

  • Atoms
    Howard Bauchner
  • Perspectives
    Stef van Buuren Jacobus P van Wouwe