Professional Documents
Culture Documents
12377 1
Abstract
Background: To evaluate if there are departures from the WHO Child Growth Standards (WHO-CGS) in postnatal
growth of healthy ‘Canadian’ children in Ontario up to age 2 years, including by infant feeding and ethnicity.
Methods: We included data on 9964 healthy, singleton children born in Ontario, Canada. Smoothed weight, length
and body mass index (BMI) percentile curves were generated using quantile regression for the Canadian cohort
from birth to age 2 years. Differences in percentile values were calculated comparing Canadian children vs. the
WHO-CGS.
Results: Canadian children under age 2 years were longer than the WHO-CGS at the 10th (0.8 cm), 50th (1.3 cm)
and 90th (1.9 cm) percentiles. Canadian children incrementally surpassed the WHO-CGS in weight after age
6 months, and in BMI after 9 months. By age 2 years, the 50th percentile weight of Canadian males was 823 g
(95% confidence interval (CI) 680, 965) higher than the WHO-CGS 50th percentile. Weight differences were seen
regardless of feeding practice, and were greatest among children of mothers born in Canada and Europe/Western
nations, and least for those of East Asian/Pacific or South Asian heritage. Among Canadian breastfed males, 18%
(95% CI 16, 19) of newborns and 26% (95% CI 20, 33) toddlers aged 2 years were classified by WHO-CGS as
weighing >90th percentile – much higher than the expected rate of 10%. Similarities were seen for differences in
BMI.
Conclusions: Healthy Canadian infants/toddlers are longer and heavier than the WHO-CGS norms. Explanations
for these discrepancies require further elucidation.
Keywords: infant and child growth, growth chart, growth standard, WHO, growth standards, length, weight, percentiles.
An important aspect of routine care for young chil- health or environmental constraints on growth. They
dren is serial measurement of weight and length.1 used a longitudinal design between birth and
Monitoring early growth helps practitioners assess 24 months, and a cross-sectional design between 18
the health of children, including feeding adequacy.2,3 and 71 months. Of the 1743 infants enrolled at birth,
In 2006, the World Health Organization (WHO) 882 (50%) complied with the criteria for the longitudi-
released universal child growth standards (CGS), nal group, including exclusive or predominant breast
intended to describe the optimal growth of children.4 feeding until at least 4 months.4
The WHO Multicentre Growth Reference Study Canadian authorities endorsed using the WHO-
(MGRS) took place between 1997 and 2003, and CGS to monitor infant growth from 0 to 2 years,5
included children from Brazil, Ghana, India, Norway, replacing the Centres for Disease Control and Preven-
Oman and the US, who were deemed to be free of tion 2002 growth charts that had predominantly
included formula-fed infants, who are prone to exces-
Correspondence:
sive growth.6 Although the WHO-CGS are thought to
Joel G. Ray, Departments of Medicine and Obstetrics and
Gynaecology, St. Michael’s Hospital, University of Toronto, describe optimal growth, regardless of ethnicity, a
Toronto, Ontario, Canada. number of international studies have found marked
E-mail: rayj@smh.ca departures from the WHO-CGS, not entirely
†
Members of the Canadian Curves Consortium are in explained by feeding practice.7–11 Despite widespread
Appendix 1.
Pacific, (vii) South Asia, and (viii) Other. Maternal We calculated differences and 95% confidence inter-
records were linked to the Ontario portion of the fed- vals (CI) in the 10th, 50th and 90th percentile WFA and
eral Immigration, Refugees and Citizenship Canada LFA; and in the 3rd, 50th and 97th percentile BMI-FA,
(IRCC) Permanent Resident Database (PRD), also between Canadian children and those of the WHO-
housed at ICES.15,17 This database contains information CGS. Estimated 95% CIs were calculated using the sex-
about original country of citizenship, for immigrants to and age-specific sample-based errors from our cohort
Canada from 1985 onward. Women not linked to the and the coefficients of variation from the WHO-CGS.23
PRD were classified as non-immigrant women. Over Differences were calculated by sex and infant feeding
90% of the latter comprise Canadian-born women, and type, as well as maternal world region of origin.
henceforth, are referred to as ‘Canadian-born’, but their Next, we estimated the proportion of exclusively
ethnic background is not documented within our breastfed Canadian children that would be considered
administrative datasets. an unhealthy size according to the WHO-CGS thresh-
Residential postal code at the time of birth was used olds. For weight and length, we determined the pro-
to determine the neighbourhood income quintile and portion (95% CI) in each age group who were <10th
rural residence, derived from Statistics Canada census percentile or >90th percentile according to the WHO-
data. CGS. For BMI, we calculated the proportion in each
age group who would be classified as wasted (<3rd
percentile) or overweight (>97th percentile) using the
Statistical analyses
WHO-CGS. All weights, lengths and BMI values were
We used nonparametric quantile regression methods converted to WHO-CGS percentiles. CIs for binomial
to generate smoothed weight-for-age (WFA), length- proportions are exact Clopper-Pearson 95% confi-
for-age (LFA), and BMI-for-age (BMI-FA) percentile dence limits.
curves for all children in our final cohort. Quantile Statistical analyses were performed using SAS for
regression produces virtually similar results to the Linux, Version 9.4 (SAS Institute, Cary, NC) and R for
Lamba-Mu-Sigma (LMS) method when the distribu- Linux, Version 3.1.2 (R Foundation for Statistical
tion of the response variable is roughly normal.21 Computing, Vienna, Austria). The study was
Curves were fit employing a cubic spline and the use approved by the institutional review board at Sunny-
of a smoothing algorithm, with knot locations deter- brook Health Sciences Centre, Toronto, Ontario.
mined by backward stepwise regression. Separate
curves were generated for males and females, and
Results
also by type of infant feeding and by maternal world
region of origin. BMI – the ratio of weight (in kg) to Formation of the final cohort of 9964 healthy, single-
recumbent length (in m2) – was calculated for a subset ton term-born infants in EMRALD is shown in Data
of infants who had both a weight and a length S1. There were 48 556 WFA measurements, 32 681
recorded at the same visit. BMI is highly correlated LFA measurements, and 31 037 BMI-FA measure-
with weight-for-length from 0 to 2 years of age.22 ments. The average child had 5 weight and 4 length
Percentiles were based on growth measurements at measurements (data not shown). Compared to all hos-
birth (weight only), and subsequent primary care vis- pital births in Ontario, those that linked to the
its centered around key time points of 7 days (range EMRALD database were somewhat more likely to
1–11), 28 days (range 17–34), 2 months (range 50– have a Canadian-born mother and reside in a rural
69 days), 4 months (range 100–129 days), 6 months area, but were otherwise comparable (Data S2). We
(range 160–189 days), 9 months (range 250–279 days), were unable to determine infant feeding practices for
12 months (range 350–379 days), 15 months (range 2188 (22%) infants.
350–379 days), 18 months (range 500–559 days) and
24 months (range 680–749 days). If a child had more
Weight-for-age
than one weight (or length) measurement within the
aforementioned age groups, then we selected that At birth, the 50th percentile weight of Canadian males
closest to the target age. Age groups were chosen to (3530 g vs. 3346 g) and females (3380 g vs. 3232 g)
correspond with those defined in the WHO-CGS was higher than that of the WHO-CGS (Data S3). By
methodology.23 6 months of age, there was consistent departure in
weight differences at the 10th, 50th and 90th per- estimate, except for breastfed and mixed-fed females
centile (Figure 1). For example, by age 2 years, the dif- at 24 months of age (Data S5).
ference at the 50th percentile was 823 g (95% CI 680,
965) among males, and 776 g (95% CI 635, 917) among
BMI-for-age
females (Data S3). These differences were even seen
for exclusively breastfed infants (Figure 1). The 50th percentile BMI was significantly lower in
The 50th percentile weight difference between the Canadian children than the WHO-CGS up to
Canadian cohort and the WHO-CGS varied signifi- 6 months for males and 9 months for females; after
cantly at birth by maternal world region of origin, age 12 months, the 50th percentile BMI was signifi-
mostly among infants of mothers born in Canada, Eur- cantly higher in the Canadian cohort vs. the WHO-
ope and Western nations (Data S4a and b). Thereafter, CGS, regardless of infant feeding (Data S6). The
differences were not appreciable up to age 6 months, threshold for infant wasting was generally higher in
but were once again evident after 9 months of age. the WHO-CGS than among Canadian males up to
12 months and females up to age 9 months (Figure 3).
Thereafter, there was little difference in 3rd percentile
Length-for-age
BMI among males, and females, until age 18 months,
From birth to age 2 years, compared to their WHO- when positive differences emerged (Figure 3). The
CGS counterparts, Canadian children were on aver- pattern of <3rd percentile BMI differences varied
age 0.8 cm longer at the 10th percentile, 1.3 cm longer slightly for breastfed males and breastfed females.
at the 50th percentile, and 1.9 cm longer at the 90th The threshold for an infant classified as overweight
percentile (Figure 2). The 50th percentile length was was generally higher in the WHO-CGS than that
higher among Canadian children than the WHO-CGS among Canadian infants up to 4 months of age
2.00 2.00
All Breas ed Formula-fed Mixed-fed
1.75 1.75
Difference in percen le weight, (kg)
1.50 1.50
1.25 1.25
1.00 1.00
0.75 0.75
0.50 0.50
0.25 0.25
0.00 0.00
– 0.25 – 0.25
– 0.50 – 0.50
0 1week 1 2 4 6 9 12 15 18 24 0 1week 1 2 4 6 9 12 15 18 24 0 1week 1 2 4 6 9 12 15 18 24 0 1week 1 2 4 6 9 12 15 18 24
1.50 1.50
1.25 1.25
1.00 1.00
0.75 0.75
0.50 0.50
0.25 0.25
0.00 0.00
– 0.25 – 0.25
– 0.50 – 0.50
0 1week 1 2 4 6 9 12 15 18 24 0 1week 1 2 4 6 9 12 15 18 24 0 1week 1 2 4 6 9 12 15 18 24 0 1week 1 2 4 6 9 12 15 18 24
Figure 1. Difference in weight-for-age between Canadian minus WHO-CGS males (top panel) and females (bottom panel), at the 10th
(red circles), 50th (black squares) and 90th (blue triangles) percentiles. The dashed line at 0 indicates no difference.
3.0 3.0
All Breas ed Formula-fed Mixed-fed
2.5 2.5
Difference in percen le length, (cm)
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
–0.5 – 0.5
–1.0 – 1.0
1 week 1 2 4 6 9 12 15 18 24 1 week 1 2 4 6 9 12 15 18 24 1 week 1 2 4 6 9 12 15 18 24 1 week 1 2 4 6 9 12 15 18 24
Age (months) Age (months) Age (months) Age (months)
Number:
Canada 2249 1769 2374 1975 1636 1073 1875 1441 1472 892 1098 806 1159 971 791 524 877 658 625 375 280 237 309 274 220 135 263 184 194 88 766 612 767 615 521 322 544 409 382 214
WHO 391 427 424 416 419 398 417 421 444 593 391 427 424 416 419 398 417 421 444 593 391 427 424 416 419 398 417 421 444 593 391 427 424 416 419 398 417 421 444 593
3.0 3.0
All Breas ed Formula-fed Mixed-fed
Difference in percen le length, (cm)
2.5 2.5
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
– 0.5 – 0.5
– 1.0 – 1.0
1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24
Age (months) Age (months) Age (months) Age (months)
Number:
Canada 2143 1700 2275 1874 1544 1048 1760 1348 1414 819 1012 780 1083 877 728 506 773 606 599 282 279 233 294 243 205 148 247 193 210 111 721 577 736 604 486 296 525 368 341 209
WHO 396 450 445 447 444 442 453 448 472 598 396 450 445 447 444 442 453 448 472 598 396 450 445 447 444 442 453 448 472 598 396 450 445 447 444 442 453 448 472 598
Figure 2. Difference in length-for-age between Canadian minus WHO-CGS males (top panel) and females (bottom panel), at the 10th
(red circles), 50th (black squares) and 90th (blue triangles) percentiles. The dashed line at 0 indicates no difference.
(Figure 3). Thereafter, the opposite was observed, a much higher than expected proportion of exclu-
such that by 24 months, the 97th percentile BMI was sively breastfed Canadian children were >90th per-
2.0 kg/m2 higher for Canadian males and 1.1 kg/m2 centile LFA on the WHO-CGS – as much as 30% at
for females, than their WHO-CGS counterparts (Fig- age 6 months (Figure 5). Correspondingly, the pro-
ure 3). Such departures were more pronounced for portion classified <10th percentile LFA on the
formula-fed than breastfed children (Figure 3). WHO-CGS was consistently lower than 10% during
the same age period (Figure 5).
3.0 3.0
All Breas ed Formula-fed Mixed-fed
Difference in percen le BMI (kg/m2)
2.5 2.5
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
– 0.5 – 0.5
– 1.0 – 1.0
– 1.5 – 1.5
1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24
Age (months) Age (months) Age (months) Age (months)
Number:
Canada 2162 1739 2309 1933 1464 1044 1655 1378 1415 824 1047 792 1121 950 704 512 774 629 599 359 283 242 315 276 196 135 231 184 193 84 735 595 739 597 473 306 499 386 362 200
WHO 387 426 423 408 411 392 411 416 442 587 387 426 423 408 411 392 411 416 442 587 387 426 423 408 411 392 411 416 442 587 387 426 423 408 411 392 411 416 442 587
3.0 3.0
All Breas ed Formula-fed Mixed-fed
2.5 2.5
Difference in percen le BMI (kg/m2)
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
– 0.5 – 0.5
– 1.0 – 1.0
– 1.5 – 1.5
1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24 1week 1 2 4 6 9 12 15 18 24
Age (months) Age (months) Age (months) Age (months)
Number:
Canada 2049 1673 2209 1829 1398 1017 1550 1279 1363 747 954 766 1040 855 673 487 686 573 578 270 274 235 295 245 190 151 231 192 204 102 700 566 718 587 431 283 463 345 325 195
WHO 392 449 445 444 441 439 451 442 469 591 392 449 445 444 441 439 451 442 469 591 392 449 445 444 441 439 451 442 469 591 392 449 445 444 441 439 451 442 469 591
Figure 3. Difference in BMI-for-age between Canadian minus WHO-CGS males (top panel) and females (bottom panel), at the 3rd
(orange circles), 50th (black squares) and 97th (green squares) percentiles. The dashed line at 0 indicates no difference.
In a post hoc analysis, we compared the 10th, 50th were most pronounced among children of mothers
and 90th percentile weights and lengths for the Cana- born in Canada, or Europe and Western nations, Mid-
dian cohort, as generated by the quantile regression dle East and North Africa, and Sub-Saharan Africa
method used herein vs. that by the LMS method used and the Caribbean. Infants of mothers from East Asia
by the WHO-MGRS Group.23 No important variations and the Pacific, South Asia, and Latin America
were observed (Data S8a and b). The same was seen appeared to grow more similar to WHO-CGS. By age
comparing the methods for the 3rd, 50th and 97th per- 2 years, there was a 10% excess in the proportion of
centiles (Data S8c). male infants classified as overweight.
30
25
15
10
0
0 1 week 1 2 4 6 9 12 14 16 18 24
Age (months)
30
25
Propor on (%, 95% CI)
20
15
region of origin (Data S4a and b). Otherwise, the In WHO MGRS, 1743 infants were enrolled at birth,
newborns in EMRALD were similar to all those in but only 882 (50%) complied with the criteria for lon-
Ontario. The ethnic ancestry of Canadian-born gitudinal follow-up to age 2 years, including exclu-
women was not available; nonetheless, the majority sive or predominant breast feeding for at least
were born 25–35 years earlier to parents of British 4 months of life.4,23 We attempted to follow the
and European ancestry.24 EMRALD had EMR data WHO-CGS criteria, and presented separate analyses
from 321 physicians across Ontario, and anthropo- by feeding practice. Even then, we could not deter-
metric data routinely collected at well-baby/child mine feeding for 22% of infants. We also lacked data
visits. Child weight and length were not obtained on maternal smoking, but the rate in Ontario is under
using the strict measurement criteria applied in the 10%.25
WHO MGRS. However, the potentially greater vari-
ability in growth measurements in EMRALD, par-
Implications
ticularly for length, which is more prone to
measurement error in young children, would not We found important differences in percentiles of
account the consistently greater lengths in Canadian weight, length and BMI of young Canadian children
children compared to the WHO-CGS. compared to the WHO-CGS, regardless of infant
40
35
30
Propor on (%, 95% CI)
25
20
15
10
0
1week 1 2 4 6 9 12 14 16 18 24
Age (months)
40
35
30
Propor on (%, 95% CI)
25
20
feeding practice. Since the WHO-CGS is currently contributing to the WHO MGRS, the tallest average
used across Canada, these differences could have a maternal height was among Norwegians, and the
substantial impact on how we interpret the ‘normal’ shortest was among Indian mothers, as were the cor-
growth of Canadian children. responding birth lengths of their infants.23 While only
Our healthy Canadian cohort was markedly longer about 20% of size at birth is determined by additive
than the WHO-CGS before 18 months, regardless of genetic variation, genetics come to explain nearly 80%
feeding practice. In a prior Canadian study of a of variation in weight, and 40% of variation in height,
selected cohort of 73 infants who were breastfed from 4 months onward.27,28 Thus, it is plausible that
according to the WHO guidelines, median newborn postnatal growth, especially after 6 months of age,
length was markedly greater at birth, but approxi- may be partly expressed by parental ethnicity, which
mated the WHO thereafter.14 This pattern was also may explain why certain groups of healthy children
seen in UK infants.9 In Norway and Belgium, greater deviate from the standards expressed in the WHO-
child length in the first 2 years of life was reported, CGS.
including among exclusively breastfed Norwegian In contrast to length, the direction and magnitude
children.7,8 of difference in weight and BMI percentiles between
Differences in maternal stature are hypothesised to Canadian children and the WHO-CGS depended on
play a role in birth length.9,26 Among the countries age. Compared to their WHO-CGS counterparts,
Canadian children had a higher 50th percentile discrepant when using the WHO-CGS. Whether such
weight at birth, and again, after 6 months, while the departures are due to obesogenic feeding practices or
50th percentile BMI was lower before 12 months normal healthy growth, needs to be determined.32
and higher thereafter. Correspondingly, a higher
than expected proportion of Canadian breastfed chil-
Conclusion
dren were wasted in the first 9 months and over-
weight the second year of life, according to the Our findings suggest that many healthy Canadian
WHO-CGS. Age-related differences in weight per- infants and toddlers weigh more and are longer than
centiles have been observed in a number of interna- the norms laid out in the WHO-CGS. Specifically, the
tional studies comparing national growth references degree to which these differences matter for short-
to the WHO-CGS.7–11 Rolland-Cachera et al.10 com- term and long-term health trajectories requires
pared national weight references from the US, UK, longitudinal examination before reconsideration of
Netherlands, Belgium and France to the WHO-CGS, current use of the WHO-CGS in Canada can be
and revealed a strikingly similar pattern of lower recommended.
median weights after birth and up to 6 months of
life, with typically higher weights thereafter.
Acknowledgements
Although the trough between birth and 6 months
observed herein, and in other Canadian studies,12–14 The authors thank Dr. Daniel Roth of The Hospital for
was less pronounced, the consistency of this pattern Sick Children for his constructive comments over the
of divergence from the WHO-CGS across different course of this project. This work was supported by a
populations, and independent of feeding practice, Healthcare Renewal Policy Analysis grant from the
raises questions about using the WHO-CGS to moni- Canadian Institutes of Health Research (CIHR). JGR
tor the growth of Canadian children, particularly holds a Canadian Institutes for Health Research Chair
before 6 months of age. in Reproductive and Child Health Services and Policy
Some have suggested that the unexpectedly higher Research, co-funded by the SickKids Foundation and
weights of WHO-CGS children aged 0–6 months may CIHR. MU holds a CIHR New Investigator Award.
be from selective dropout of smaller infants.29 Around The funders had no role in study design, data collec-
50% of infants enrolled in the WHO MGRS were non- tion and analysis, decision to publish, or preparation
compliant with breast feeding, and their growth mea- of the manuscript. This study was supported by the
sures were excluded. When an infant is smaller than Institute for Clinical Evaluative Sciences (ICES), which
expected, exclusive breast feeding is more likely to be is funded by an annual grant from the Ontario Min-
replaced with mixed for formula feeding.29,30 In istry of Health and Long-Term Care (MOHLTC). The
Malawi, among exclusively breastfed infants under opinions, results and conclusions reported in this
age 6 months, more infants fell below the threshold paper are those of the authors and are independent
<3rd percentile WFA on the WHO-CGS than on from the funding sources. No endorsement by ICES
another growth chart.31 These children were at higher or the Ontario MOHLTC is intended or should be
risk of unnecessary referral to a specialist, with inter- inferred. Parts of this material are based on data and
ruption of exclusive breast feeding, even when there information compiled and provided by the Canadian
was no evidence of faltering growth.31 Hence, while Institute for Health Information (CIHI). However, the
largely adopted for use in many settings, there is some analyses, conclusions, opinions and statements
uncertainty around the use of the WHO-CGS as a expressed herein are those of the author, and not nec-
guide to healthy child growth in the first 2 years of essarily those of CIHI. Conflict of Interest: The authors
life. Certainly, a similar assessment among Canadian report no conflict of interest.
infants and toddlers is required, including by parental
ethnicity.
References
We calculated BMI measures for Canadian infants
and toddlers in the first 2 years of life, offering 1 Spencer NJ. Failure to think about failure to thrive. Archives
another useful indicator of early childhood growth.22 of Disease in Childhood 2007; 92:95–96.
2 Kramer MS, Barros FC, Demissie K, Liu S, Kiely J, Joseph KS.
The BMI cut-points for classifying a Canadian breast-
Does reducing infant mortality depend on preventing low
fed infant as wasted, or a toddler as overweight, were
growth charts for assessing infant malnutrition: a (green circles) and WHO (blue triangles). The right
randomised controlled trial. Journal of Paediatrics and Child axis shows the difference (95% CI) in 50th percentile
Health 2014; 50:32–39.
weight (g) (black squares). The dashed line at 0 indi-
32 Salsberry PJ, Reagan PB. Dynamics of early childhood
overweight. Pediatrics 2005; 116:1329–1338. cates no difference.
Data S4. Difference (95% CI) in 50th percentile
weight-for-age between Canadian minus WHO-CGS
Appendix 1 (a) males (b) females, by maternal world region of ori-
gin. The dashed line at 0 indicates no difference.
Data S5. Difference (95% CI) in the 50th percentile
Members of the Canadian Curves Consortium length-for-age between Canadian minus WHO-CGS
males (top panel) and females (bottom panel). The left
Kathlyn Babaran-Henfrey, Emily Bartsch, Howard
axis shows the 50th percentile length (cm) in Canada
Berger, Douglas Campbell, Maria Chiu, Lisa Colizza,
(green circles) and WHO (blue triangles). The right
Leanne R. De Souza, Astrid Guttmann, Manavi
axis shows the difference (95% CI) in 50th percentile
Handa, Christopher Longo, Jonathon Maguire, James
length (cm) (black squares). The dashed line at 0 indi-
Meloche, Nir Melamed, Patricia Mousmanis, Alison
cates no difference.
Park, Joel Ray, Henry Roukema, Jennifer Roy, Michael
Data S6. Difference (95% CI) in the 50th percentile
Sgro, Graeme N. Smith, Karen Tu, Marcelo Urquia,
BMI-for-age between Canadian minus WHO-CGS
Pat Vanderkooy, William Watson, Nancy Watts,
males (top panel) and females (bottom panel). The left
Jacqueline Young.
axis shows the 50th percentile BMI (kg/m2) in Canada
(green circles) and WHO (blue triangles). The right
Supporting Information axis shows the difference (95% CI) in 50th percentile
BMI (black squares). The dashed line at 0 indicates no
Additional Supporting Information may be found in
difference.
the online version of this article at the publisher’s
Data S7. Proportion of exclusively breastfed Canadian
web-site:
male (top) and female (bottom) children that would
Data S1. Flow chart of formation of the cross-sectional be classified as wasted <3rd percentile BMI-for-age
cohorts with measurements of weight (Cohort 1), (orange circles) or overweight >97th percentile BMI-
length (Cohort 2), or both weight and length concur- for-age (green squares) according to the WHO-CGS.
rently (Cohort 3), between birth and age 750 days. The dashed line indicates the expected proportion of
Data S2. Characteristics of the 9964 singleton term 3%.
livebirths comprising the current study cohort from Data S8. (a) Difference in 10th, 50th and 90th per-
the EMRALD database, compared to all singleton centile weights of Canadian children, comparing
term hospital livebirths in Ontario between April 1, Quantile Regression (QR) vs. Lamba-Mu-Sigma
2002 and March 31, 2013. Standardised differences (LMS) methods. (b) Difference in 10th, 50th and 90th
were used to compare differences in means or percentile lengths of Canadian children, comparing
proportions. Quantile Regression (QR) vs. Lamba-Mu-Sigma
Data S3. Difference (95% CI) in the 50th percentile (LMS) methods. (c) Difference in 3rd, 50th and 97th
weight-for-age between Canadian minus WHO-CGS percentile body mass index (BMI) of Canadian chil-
males (top panel) and females (bottom panel). The left dren, comparing Quantile Regression (QR) vs.
axis shows the 50th percentile weight (kg) in Canada Lamba-Mu-Sigma (LMS) methods.