Background New national growth references have been published in Belgium and Norway. The WHO recommends universal use of their 2006 Child Growth Standards based on data from breastfed children.
Objective To compare the growth of Belgian and Norwegian children with the WHO standards.
Participants 6985 children 0–5 years of age from Belgium and Norway.
Design Proportion of children below −2 SD and above +2 SD of the WHO standards was calculated for length/height, weight, body mass index and head circumference. Average SD scores of exclusively breastfed children of non-smoking mothers were compared with national reference data and with the WHO standards.
Results Generally, the number of Belgian and Norwegian children below −2 SD lines of the WHO standards was lower and above +2 SD higher than expected. The largest differences were for head circumference (0.97% Belgian and 0.18% Norwegian children below −2 SD, 6.55% Belgian and 6.40% Norwegian children above +2 SD) and the smallest for length/height (1.25% Belgian and 1.43% Norwegian children below −2 SD, 3.47% Belgian and 2.81% Norwegian children above +2 SD). The growth pattern of breastfed children of non-smoking mothers was in both countries more alike the local national growth references than the WHO standards.
Conclusions There are significant deviations in the proportion of children outside normal limits (±2 SD) of the WHO standards. This was true for all children, including those who were exclusively breastfed. Hence, adoption of the WHO growth charts could have consequences for clinical decision-making. These findings advocate the use of national references in Belgium and Norway, also for breastfed children.
Statistics from Altmetric.com
Because of the variations in growth between populations, there has been a long tradition of constructing growth references aimed at specific target populations. Recently, new national growth references from birth to adulthood were published for Belgian1 and Norwegian children.2 In 2006, the WHO also issued universal growth standards that intend to show how children grow under optimal conditions (hence, “standards”), irrespective from which world region or population they originate (hence, “universal”).3 To achieve this goal, the WHO set up a multi-centre growth reference study (MGRS) in six countries (Brazil, Ghana, India, Norway, Oman, USA) and only included children of high social class (except in Norway and the USA where social class was not an inclusion criterion) of non-smoking mothers that were breastfed according to recommendations.4 The results for length demonstrated that differences between the six countries were relatively small (although the medians differed up to 0.5 SD at certain ages), and universal use of these charts for all children worldwide was, therefore, recommended.5 To this date, site differences for weight, body mass index (BMI) or head circumference have not been reported.
What is already known on this topic
▶ In 2006, the WHO issued growth standards showing children's growth under optimal conditions.
▶ The WHO advises universal use of these standards.
What this study adds
▶ The growth of Belgian and Norwegian children deviates significantly from the WHO growth standards, even when they are breastfed according to recommendations.
▶ In both countries head circumference is systematically 0.5 SD or more above the WHO growth standard.
The WHO growth standards represent an improvement over the previous US National Centre for Health Statistics (NCHS)-based growth reference charts, but the implementation is nevertheless not straightforward in developed countries where updated references are available. Children in Hong Kong have shown to be shorter when compared with the WHO standard6 and children in The Netherlands longer.7 In the UK, birth weight and birth length have been found to be higher,8 and consequently, the WHO standards were only adopted from 2 weeks of age and onwards. When comparing SD scores for length/height, weight, BMI and head circumference, three of us (MR, KH and RH) have recently shown that Belgian breastfed children grow similar in length when compared with the WHO standards, but not for other traits like weight, BMI and head circumference.9 Collectively, these observations challenge the validity of the WHO standards as a universal tool for growth monitoring that should replace local references.7 This is an important issue because the choice of a local reference or the WHO international standards may affect thresholds used in clinical guidelines and can, therefore, have considerable consequences for both patients, parents and the healthcare system.
The aim of the present study was to compare the proportion of Belgian and Norwegian children outside the normal limits (defined by ±2 SD) of the WHO growth standards for length/height, weight, BMI and head circumference, with the expected number of 2.3%. A separate analysis was performed for children of a non-smoking mother breastfed according to recommendations.
Participants and methods
The Flanders Growth Study (FGS) is a representative cross-sectional survey of 18 051 participants 0–21 years of age, recruited in Flanders (northern part of Belgium), in 2001–2004.1 The Bergen Growth Study (BGS) is a cross-sectional survey of 8299 participants 0–19 years of age, recruited in the city of Bergen, Norway in 2003–2006.2 The present analysis was limited to children of Belgian (n=4754) and Norwegian (n=2231) origin between 0 and 5 years of age. All children were healthy and born at term (37–42 weeks of gestational age), but low birth weight children born at term were not excluded (table 1).
The FGS and BGS are highly comparable in design and modus operandi, and both surveys included biometry and an assessment of social status and infant feeding practice. Length (before 2 years of age), height (from 2 years of age), weight and head circumference were measured by trained staff members, using a standardised technique and equipment. Weight, length and head circumference at birth, feeding practice and smoking in pregnancy were recorded through a parental questionnaire in both studies. Children were considered as exclusively breastfed when they were breastfed for at least 6 months, and no formula milk was introduced before that age. This includes also children <6 months who did not yet receive any formula milk at the time of measurement. Solids were not accounted for, but these are typically introduced between 4 and 6 months of age. A total of 683 Belgian and 527 Norwegian children were assigned to the exclusively breastfed subgroup (excluding 69 Belgian and 43 Norwegian children because of smoking during pregnancy). In addition, 304 Norwegian children were exclusively breastfed for at least 6 months and received continuous breastfeeding (exclusive or not) up to at least 12 months of age (excluding 16 children because of smoking during pregnancy). Only 26 children in the Belgian sample met this criterion.
Participants were grouped into 4 months interval between birth and 1 year of age and yearly intervals thereafter. The prevalence of children below and above certain thresholds for length, height, weight, BMI and head circumference is reported by age group or adjusted for age (per year) otherwise. For the analysis of breastfed children of non-smoking mothers, all traits were converted to SD scores according to the WHO growth standards.3 The reference group consists of all children on which the Belgian and Norwegian growth charts are based, including the group of exclusively breastfed children by definition.
Means and proportions are always presented with 95% CIs. CIs for simple proportions are the exact Clopper–Pearson 95% confidence limits.10 Age-adjusted rates and confidence limits are based on the γ distribution,11 assuming an equal number of children in each age group. Size at birth was based on recall data from all participants in the sample and constitutes a separate age class. Global prevalence estimates are only based on postnatal growth measurements.
The Flanders Growth Study was approved by the ethics committee of the Vrije Universiteit Brussel and the Bergen Growth Study by the Regional Committee for Medical Research Ethics and the Norwegian Data Inspectorate.
Percentage of all children outside −2 SD and 2 SD
Compared to the WHO growth standards, we generally saw fewer children below −2 SD and more children above +2 SD than expected (fig 1). Remarkably, similar trends were observed in Belgian and Norwegian children.
At birth, the percentage of Belgian children below −2 SD was as expected for length, weight and head circumference, but too high for the BMI. The percentage of children above +2 SD was highly elevated for length (8%) and head circumference and comparable for weight and BMI. The percentage of Norwegian children below −2 SD at birth was as expected for length and BMI, but too low for weight and head circumference. The percentage of children above +2 SD was highly elevated for all variables, and up to 13% for length and head circumference.
The age-adjusted prevalence of all children 0–5 years of age with a length below −2 SD was 1.25% (0.92–1.65) in Belgium and 1.43% (0.90–2.17) in Norway. The corresponding percentage above +2 SD was 3.47% (2.90–4.14) and 2.81% (2.11–3.69), respectively. In both countries, there was a tendency to converge with the WHO standard for length towards the older age groups, and the proportion of children outside normal limits became reasonably close to the expected amount of 2.3% by the age of three (fig 1). For weight, the overall percentage below −2 SD was 0.65% (0.45–0.93) for Belgian and 0.56% (0.29–1.04) for Norwegian children. The percentage above +2 SD was, respectively, 3.08% (2.54–3.71) and 3.39% (2.56–4.44). Unlike for length, there was no apparent trend with age. A similar observation was made for the BMI, where the overall percentage below −2 SD was 0.84% (0.59–1.18) for Belgian and 0.54% (0.29–0.99) for Norwegian children. The corresponding percentage above +2 SD was 3.01% (2.45–3.67) and 4.29% (3.30–5.50). For head circumference, the overall percentage below −2 SD was 0.97% (0.70–1.33) for Belgian and only 0.18% (0.05–0.53) for Norwegian children. The prevalence of children above +2 SD was relatively high with, respectively, 6.55% (5.76–7.42) and 6.40% (5.19–7.83). Apart from birth, the largest discrepancy was observed between 1 and 3 years of age.
The effect of breastfeeding
To document the effect of breastfeeding, we compared the growth of a subgroup of 1210 breastfed children born to non-smoking mothers with the WHO growth standards in both countries. This was done by comparing mean SD scores rather than the prevalence outside the normal range, because the number of children below −2 SD or above +2 SD would become too small. The prevalence of exclusive breastfeeding for at least 6 months was 26.5% in the Belgian sample9 and 40.6% in the Norwegian sample (based on recall data from children 6 months to 5 years of age). This analysis demonstrated that the growth of exclusively breastfed children was more similar to that of all children in the local population (including those that were not breastfed) than to the WHO growth standards (fig 2). Between 6 months and 2 years of age, Belgian breastfed children had generally a smaller length and less weight and are, therefore, somewhat closer to the WHO growth standards, although the overall pattern was still more like that of the reference population. For Norwegian children, this was true for weight and BMI. To document the possible additional effect of continuous breastfeeding up to at least 12 months of age, the growth of a subgroup of 304 Norwegian children that comply with these criteria was compared to those that were exclusively breastfed for 6 months (fig 3). The length and weight were highly similar in the two groups, and no additional effect could be demonstrated in this study. Similar results were obtained for BMI and head circumference (data not shown).
In this paper, we have demonstrated significant deviations in the proportion of Belgian and Norwegian children that are located outside the normal limits of the WHO growth standards, generally defined as an SDS below −2 SD (2.3 percentile) or above +2 SD (97.7 percentile). Although these cut-offs do not warrant immediate action or referral to specialised care, they are often included in clinical guidelines, and adoption of the WHO growth charts could, therefore, affect the clinical decision making process.8 Compared to the WHO standards, the number of Belgian and Norwegian children below −2 SD is generally lower, and the number above +2 SD higher than expected, for length/height, weight, BMI and head circumference, with the largest discrepancy observed for head circumference. As the WHO standards are based on a sample of breastfed children born to a non-smoking mother, one could speculate that these differences mirror different inclusion criteria. However, our observation, that the growth of exclusively breastfed children in both Belgium and Norway was generally closer to that of the local reference population than to that of the WHO standards, refutes this hypothesis and probably reflects that genetic and environmental background factors have stronger influence on growth than feeding practices. This perception is strengthened by the similarity of the growth patterns observed in both countries as compared to the WHO growth standards.
In the present study, criteria for exclusive breastfeeding were chosen to mimic those of the MGRS. These criteria differ slightly from current WHO recommendations, which advise the introduction of solids at the age of 6 months.12 In the MGRS, the previous recommendation of introducing solids between 4 and 6 months of age was used. Solids were not accounted for in the present study, but it was previously shown that only a minority of breastfed children received solids before the age of 4 months.13 14 In children above +2 years, the MGRS inclusion criteria were less strict (at least 3 months of predominant breastfeeding, with no additional requirements about complimentary foods), while we applied the same definition throughout the whole age range. Due to the cross-sectional design of this study, the feeding practice of children younger then 6 months was only documented up to the time of measurement. This is in contrast with the MGRS, where the feeding practice of all children <2 years of age was documented over the whole trajectory. While this could affect the composition of the target group, any consequences on the growth of breastfed children between birth and 6 months cannot be due to breastfeeding per se.
The average length is larger for Norwegian children but relatively comparable for Belgian children when compared with the WHO growth standards, although these differences tend to disappear by 3 years of age. The proportion of children below −2 SD was nevertheless considerably smaller than expected. A comparable pattern of average growth in length was also observed in the Norwegian sample of the WHO standards,5 7 which suggests a possible population effect.
In the case of BMI, adoption of the WHO growth standards would lead to a lower prevalence of underweight and an increased prevalence of overweight in Belgium and Norway. A comparable observation was made in the UK by Wright et al,8 who saw this as a positive feature, partly because of the current increase in childhood overweight. This is, however, contradicted by the observation in our study that, compared to the WHO growth standards, weight and BMI in breastfed children is broadly comparable to that of all children. Any comparisons with the WHO standards for weight and BMI are nevertheless limited by the fact that the WHO did not publish site differences for these traits.15 Comparable results were obtained when plotting SD scores of weight for length/height (data not shown).
The most striking differences were observed for head circumference. When applied in Belgium or Norway, the proportion below −2 SD of the WHO growth standards was less than half the expected amount (only 8% of the expected number in Norway); approximately 70% of all children are above the median (data not shown); and the proportion above +2 SD would increase up to threefold. In Norway, in each annual birth cohort of approximately 60 000 children about 4000 would be above the 97th percentile (expected 1500) and only about 100 would be located below the 3rd percentile (expected 1500). Although clinical guidelines for head circumference are seldom based on a single measurement, clinicians and medical staff of well baby clinics and school health centres would find a very high proportion of their population with large heads, while small heads would be virtually non-existing. This might lead to a large uncertainty concerning referral criteria and unwarranted examinations of many children and introduce unnecessary worries in many parents.
One of the main reasons for constructing the WHO growth standards was the observation that breastfed children grow differently from the previous NCHS/WHO international growth references, with faster growth in length and weight during the first 6 months and slower growth thereafter.16 A tendency towards this pattern was observed in the present study, although the general trend was more like that of the local reference population. In our opinion, the present observations do not favour the use of the WHO growth standards for monitoring growth of children of Belgian or Norwegian origin, a result that may also be generaliseable to other North/West European populations. The WHO growth standards are nevertheless a valuable instrument, (a) as an independent platform for international comparisons and (b) as a tool for growth monitoring when a suitable local reference is unavailable, although differences in growth should be accounted for. Moreover, the concept of the WHO growth study has put breastfed children forward as the “norm” rather than the other way around, a view that we fully support.
In conclusion, we have demonstrated a significant deviation in the proportion of Belgian and Norwegian children that are located outside the normal limits for length/height, weight, BMI and head circumference of the WHO growth standards. Consequently, the implementation of the WHO growth standards would have unwanted clinical implications that are not supported by this study. This suggests that at least in Belgium and Norway, the use of a local reference based on a representative sample of healthy children is more appropriate for all children, bottle-fed or breastfed.
The authors are most grateful to the children and their families for taking part in the Flanders Growth Study and Bergen Growth Study.
PBJ AND MR contributed equally to this work.
Funding The Flanders growth study was supported by a grant from the Minister of Public Health of the Flemish Government. The Bergen Growth Study was supported by The Western Norway Regional Health Authority, University of Bergen, Novo Nordisk AS and Merck Serono SA.
Competing interests None.
Ethics approval The Flanders Growth Study was approved by the ethics committee of the Vrije Universiteit Brussel, and the Bergen Growth Study by the Regional Committee for Medical Research Ethics and the Norwegian Data Inspectorate.
Provenance and peer review Not commissioned; externally peer reviewed.
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.