Content uploaded by Philippe G Steenhout
Author content
All content in this area was uploaded by Philippe G Steenhout on Sep 10, 2016
Content may be subject to copyright.
Growth of infants consuming whey-predominant term infant formulas
with a protein content of 1.8 g/100 kcal: a multicenter pooled analysis
of individual participant data
1,2
Dominik D Alexander,
3,4
Jian Yan,
5
* Lauren C Bylsma,
3
Robert S Northington,
5
Dominik Grathwohl,
6
Philippe Steenhout,
7
Peter Erdmann,
8
Evelyn Spivey-Krobath,
8
and Ferdinand Haschke
9
3
EpidStat Institute, Ann Arbor, MI;
4
EpidStat Institute, Seattle, WA;
5
Research and Development, Nestle
´Nutrition, King of Prussia, PA;
6
Nestle
´Research
Center, Lausanne, Switzerland;
7
Nestle
´Health Science and
8
Nestle
´Nutrition, Vevey, Switzerland; and
9
Paracelsus Medical University, Salzburg, Austria
ABSTRACT
Background: High protein intake during infancy may contribute to
obesity later in life in infants who are not exclusively breastfed.
Lowering the protein content of infant formula so it is closer to that
of mature breast milk may reduce long-term risk of overweight or
obesity in formula-fed infants.
Objective: We assessed the effects of whey-predominant formu-
las with a protein content of 1.8 g/100 kcal (lower than that in
most current formulas and closer to breast milk) on infant growth
by comparing against WHO growth standards and breastfed
infants.
Design: A multicenter pooled analysis was conducted with the use
of individual participant data (n= 1882) from 11 randomized con-
trolled trials of healthy term infants. Mixed-effects models that used
ANCOVA were generated to estimate weight-for-age zscore (WAZ),
as well as length-for-age, BMI-for-age, and head circumference–for-
age zscores at age 4 mo in infants fed a lower-protein infant
formula (LPF) or a lower-protein infant formula with additional
active ingredients (probiotics, prebiotics, or both) (LPFA) and
breastfed infants. Estimates, including 95% CIs, were compared
with a 60.5 SD of WHO growth standards, a benchmark for clin-
ically significant differences.
Results: The 95% CIs for pooled estimates of WAZ were within
60.5 SD of WHO growth standards for the LPF [0.07 (20.16,
0.29)] and LPFA [0.22 (0.01, 0.43)] groups. WAZ was higher in
the LPF (P,0.001) and LPFA (P= 0.003) groups than in the
breastfed infants, likely because breastfed infants had a relatively
low WAZ [20.23 (20.51, 0.05)] compared with WHO growth stan-
dards. The 95% CIs for all other zscores in the LPF and LPFA
groups were within 60.5 SD of WHO growth standards, except for
head circumference, for which the upper limit of the 95% CI
slightly exceeded 0.5 SD. No difference was observed in any
zscores between the LPF and LPFA groups.
Conclusion: Whey-predominant infant formula with a lower protein
content that more closely resembles that of breast milk supports healthy
growth comparable to the WHO growth standards and close to
breastfed infants. Am J Clin Nutr doi: 10.3945/ajcn.116.130633.
Keywords: multicenter, individual participant data, pooled anal-
ysis, low-protein infant formula, breast milk, infant growth, pro-
biotics, prebiotics
INTRODUCTION
Accumulating scientific evidence has shown that the kinetics
of early growth, such as rapid weight gain after birth, may be
associated with later risk of obesity and possible chronic disease
outcomes (1–4). Breastfeeding, compared with feeding tradi-
tional (high-protein) formulas, has been identified as a protective
factor against obesity later in life (5–8). Although the underlying
biological mechanisms are not entirely clear, one theory, known
as the “early protein hypothesis,” attributes this possible protective
relation to the protein content of feedings (5, 6). Specifically,
formula-fed infants may be exposed to a high amount of protein
that may increase their risk of later undesirable health outcomes.
Therefore, a better understanding of the potential relation be-
tween lower protein intake in early infancy and growth may
have important implications for obesity prevention.
Because of the difference in protein quality between breast milk
and infant formula, a higher protein content in many infant for-
mulas traditionally has been required to ensure that infants receive
adequate amounts of amino acids for growth and development (9).
However, advances in protein technology have led to the de-
velopment of a higher-quality whey-predominant protein (10) that
is used to manufacture lower-protein infant formula (LPF)
10
and
1
FundedbyNestle
´Nutrition, Vevey, Switzerland. Nestec provided study prod-
ucts and funding support for all 11 studies included in the analysis. This is a free
access article, distributed under terms (http://www.nutrition.org/publications/
guidelinesand-policies/license/) that permit unrestricted noncommercial use, dis-
tribution, and reproduction in any medium, provided the original work is properly
cited.
2
Supplemental Figures 1 and 2 and Supplemental Table 1 are available from
the “Online Supporting Material” link in the online posting of the article and
from the same link in the online table of contents at http://ajcn.nutrition.org.
*To whom correspondence should be addressed. E-mail: jian.yan@rd.
nestle.com.
Received January 19, 2016. Accepted for publication July 28, 2016.
doi: 10.3945/ajcn.116.130633.
10
Abbreviations used: BMIAZ, BMI-for-age zscore; HCAZ, head
circumference–for-age zscore; IPD, individual participant data; LAZ, length-
for-age zscore; LPF, lower-protein infant formula; LPFA, lower-protein
infant formula with additional active ingredients (probiotics, prebiotics, or
both); RCT, randomized controlled trial; WAZ, weight-for-age zscore.
Am J Clin Nutr doi: 10.3945/ajcn.116.130633. Printed in USA. Ó2016 American Society for Nutrition 1of10
AJCN. First published ahead of print September 7, 2016 as doi: 10.3945/ajcn.116.130633.
Copyright (C) 2016 by the American Society for Nutrition
lower-protein infant formula with additional active ingredients
(probiotics, prebiotics, or both) (LPFA). At 1.8 g protein/100 kcal,
the protein-to-energy ratio of LPF and LPFA is closer to that of
breast milk and represents the lowest regulatory permissible limit
for protein in infant formula in the United States and the Euro-
pean Union (11–13).
The protein concentration of LPF and LPFA has been dem-
onstrated to be safe while supporting early growth patterns and
metabolic outcomes closer to those of breastfed infants (14).
Although individual randomized controlled trials (RCTs) have
demonstrated that LPFs and LPFAs support adequate infant
growth, to our knowledge, anthropometric outcomes from these
randomized trials have not been synthesized with the use of
systematic methodology. Furthermore, as noted in the comments
on infant formula supplemented with probiotics and/or prebiotics
in 2010 by the European Society for Pediatric Gastroenterology,
Hepatology, and Nutrition (15), better understanding of the health
effects of such formulas compared with formula without probiotics
or prebiotics is warranted. Although preliminary pooled analyses of
weight-for-age zscore (WAZ) from some studies have been re-
ported in 2 recent reviews (16, 17), to our knowledge, no study to
date has systematically pooled all of the individual-level growth
data from different trials for infants receiving LPF, including those
containing active ingredients, and compared them against the
WHO growth standards and with breastfed infants. Pooled anal-
ysis of individual data across multiple trials can enhance the sta-
tistical precision to estimate growth parameters in infants fed LPF
or LPFA while accounting for geographic and cultural variability.
Therefore, we conducted a pooled analysis with the use of
individual participant data (IPD) across 11 randomized trials of
LPF and LPFA. In a traditional meta-analysis, summary statistics
are combined across studies, which may be subject to design
variation, use of differing analytic metrics, and variable defini-
tions of exposures and outcomes. Our approach allowed us to
perform a pooled analysis of IPD across 11 primary RCTs, thus
making it possible to create unified variable definitions and adjust
for influencing covariates such as birth characteristics (18).
Specifically, we 1) evaluated the growth of infants fed LPF or
LPFA and breastfed infants by comparing the anthropometric
zscores against the 2006 WHO growth standards (19), with
WAZ as the primary endpoint, and 2) compared anthropometric
zscores in LPF-fed, LPFA-fed, and breastfed infants.
METHODS
Study design and inclusion of individual studies
Currently, the most commonly used infant formula globally
with 1.8 g protein/100 kcal is a whey-predominant infant formula
(Supplemental Table 1) from Nestle
´Nutrition. This formula
has been tested in multiple clinical trials that have included
infant growth as an outcome, thereby providing a rich
source of data on infant formula with a specific and homoge-
neous protein content, in terms of both protein quantity and
quality. We had access to the participant-level data for all in-
cluded trials, which allowed the use of a more rigorous IPD
pooled analysis design than would a traditional meta-analysis
design based on summary statistics reported by the individual
studies. Eligibility criteria for inclusion of trials in the pooled
analysis included the following: 1) double-blind, randomized
controlled design; 2) evaluation of healthy term infants; 3) infants in
$1 study arm fed whey-predominant infant formula with a protein
content of 1.8 g/100 kcal; 4) results that included WAZ, length-for-
age zscore (LAZ), BMI-for-age zscore (BMIAZ), and head
circumference–for-age zscore (HCAZ); and 5) infants either ex-
clusively formula-fed from #4wkofageto$4moofage
(formula-fed arms) or exclusively breastfed from birth to $4mo
of age (breastfed arms). All included trials had parallel infant
formula group designs, with some having a nonrandomized
breastfed reference group as well. Research staff and infants’
parents were blinded to the type of formula during study follow-
up. All trials were conducted in accordance with the Declaration
of Helsinki and Good Clinical Practices and were approved by
respective institutional ethics committees.
Comparison with 2006 WHO Child Growth Standards
In the current analysis, zscores were calculated to compare
infants’ growth with the 2006 WHO standards. Zscores represent
the number of SD units above or below the median of the WHO
standard curves for healthy child growth, which were developed
based on the WHO Multicenter Growth Reference Study (19).
The WHO growth standards have been widely accepted as the
standard on how infants and young children should grow (20).
Data synthesis and statistical analyses
We pooled individual-level data across trials while following
a modified version of the Preferred Reporting Items for a Sys-
tematic Review and Meta-Analysis of Individual Participant Data
guidelines to ensure the validity and accurate reporting of data(21).
IPD from all trials were merged into a single database. To be
included in the pooled analysis, a subject needed to have data at
birth and month 4 and also have data for the 2 additional covariates
included in the statistical model, infant sex and delivery type. This
allowed us to take advantage of the poolingstructure of the analysis
by harmonizing covariates across different centers. All data were
carefully reviewed for missing values or data entry errors according
to standard quality control procedures. This was done for both
baseline information and outcome measurements. Among the few
database errors identified were negative ages, no birth weight
data, and inconsistent head circumference data. These accounted
for ,10 infants throughout all study centers, and data for these
participants were omitted from the analyses. The data manage-
ment flow diagram is presented in Supplemental Figure 1. Baseline
categorical data were analyzed with the use of a chi-square test, and
baseline continuous data were analyzed with the use of ANOVA.
Overall comparison of the 3 groups was conducted first, and if this
was significant at P#0.05, pairwise comparisons were done.
The primary objective was to assess the growth of LPF-fed,
LPFA-fed, and breastfed infants by comparing anthropometric
zscores with WHO growth standards. The primary endpoint was
WAZ at 4 mo of age. Four months was selected as a clinically
relevant time point that would avoid significant confounding
from the introduction of complementary feeding between 4 and
6 mo of age. To achieve the primary objective, WAZ, LAZ,
BMIAZ, and HCAZ were calculated with the use of publicly
available macros from the WHO website. The zscore estimates
at 4 mo from the individual studies were derived by ANCOVA
while adjusting for corresponding birth zscores, infant sex, and
2of10 ALEXANDER ET AL.
type of delivery. A 2-step procedure that used mixed-effects
(fixed and random) models was used for all primary analyses
(18). I
2
was calculated for the primary endpoint (WAZ) as an
indicator of the proportion of heterogeneity in the meta-analysis
model. In the first step, fixed-effects models with the use of IPD
were generated with ANCOVA that corrected for corresponding
birth zscores, infant sex, delivery type, and study. In the second
step, random-effects models were created to estimate the overall
weighted group mean effects while accounting for between-study
variance, with the use of the study-specific group mean estimates
and variance data. This allowed us to use and control for the in-
dividual data in all infants, estimate the within-group variance for
each study, and produce overall summary effects while accounting
for both within- and between-study variability. Finally, the zscore
estimates and 95% CIs were compared with 60.5 SD of the WHO
growth standards. Specifically, when the lower bound of the 95%
CI was .20.5 SD and the higher bound of the 95% CI was ,0.5
SD, anthropometric parameters were considered to be not statis-
tically or clinically different from the WHO growth standards. The
selection of 60.5 SD as the clinically significant benchmark for
WAZ is consistent with previous studies (20, 22, 23) and with the
recommendation from the American Academy of Pediatrics to use
3 g/d as a clinically relevant difference in weight gain in infant
feeding clinical trials (24). A growth difference of 3 g/d will result
in a 366 g difference in weight gain after 4 mo. When compared
against the WHO growth standard at 4 mo, 366 g translates to
a 0.52 SD for girls and a 0.50 SD for boys. Thus, as a measure of
clinical relevance, a 3-g/d difference in weight gain translates into
a60.5 SD for WAZ at 4 mo. The same benchmark of 60.5 SD
was also used as an indication of clinical significance for other
anthropometric parameters in the study, and this is consistent with
previous studies that applied 0.5 SD as a benchmark for assessing
infant LAZ and HCAZ (20, 22, 23).
The secondary objective of our analysis was to compare the
growth of infants fed LPF or LPFA with that of breastfed infants.
To achieve this objective, zscores were compared between the
3 feeding groups with the use of ANCOVA with feeding group
as the factor in the model while adjusting for corresponding birth
zscores, infant sex, delivery type, and study. As an exploratory
analysis, we evaluated the rate of weight gain based on change in
WAZ from birth to 4 mo of age (WAZ at 4 mo minus WAZ at
birth) between the 3 feeding groups. The change in WAZ
was classified as “slow” (,20.67), “gradual” (20.67 to 0.67), or
“rapid” (.0.67). Such classification based on increments of 0.67
SD has been used in previous studies (25, 26), and the 3 cate-
gories for change in WAZ are equivalent to downward, aver-
age, and upward crossings of major weight percentiles, with
a 0.67 SD change corresponding to a change in one major
percentile band (e.g., 50th to 75th) on a growth chart (25).
Rapid early weight gain (change of .0.67) has been linked to
an increased risk of overweight in later life (25). The distri-
bution of infants in these 3 categories was compared between
the LPF, LPFA, and breastfed groups with the use of a chi-
square test. A Fisher’s exact test subsequently was used for
pairwise comparisons. In previous studies (25, 26), the clas-
sification was applied to the change in WAZ from birth to
age 6 mo. However, to avoid potential confounding effects of
complementary feeding on the rate of weight gain, we focused
on the change from birth to 4 mo, when infants still were fed
exclusively with either formula or breast milk.
All statistical analyses were conducted with the use of SAS
Statistical Software, version 9.1. Data reported are estimated
means (95% CIs) unless otherwise noted.
RESULTS
Characteristics of infants included in the pooled analysis
Individual-level data from 11 RCTs were included in the
multicenter pooled analysis (Table 1). Studies were conducted
between 1998 and 2008 and were performed in 6 different
countries on 4 different continents. Ten studies used LPF as an
infant feeding arm; 9 studies used LPFA that contained prebi-
otics, probiotics, or both; and 5 studies used breastfed infants as
a reference arm. The overall dropout rate among all study par-
ticipants was w24%, with slightly higher rates of dropout in the
breastfed groups than in the formula-fed groups (30% compared
with 22%). We pooled data from a total of 1882 healthy term
infants; 737 in the LPF group, 965 in the LPFA group, and 180
in the breastfed group met all predefined requirements for the
WAZ endpoint analysis. The number of infants included in the
other growth endpoint analyses was somewhat lower because of
missing baseline values (Table 2). The proportion of male and
female infants was balanced (w1:1) in each of the 3 feeding
groups. There was a significantly higher rate of cesarean section
delivery in the LPF and LPFA groups than in the breastfed group,
and the 2 groups of formula-fed infants had significantly lower
WAZ and BMIAZ at birth than did the breastfed group (Table 2).
WAZ at 4 mo of age
The pooled WAZ estimate for the LPF group was 0.07 (20.16,
0.29) based on analysis of 737 infants from 10 studies (Figure 1).
The 95% CI range for the pooled estimate was well within
60.5 SD of the WHO growth standards; this was also true for
7 of the 10 studies with LPF arms. Results from individual studies
were relatively homogeneous, with the exception of the study
conducted in China (31), in which the mean estimate for WAZ
was 0.92 and the lower bound of the 95% CI was 0.79 (which
is .0.5 SD). A sensitivity analysis conducted without the data
from this study resulted in a pooled WAZ estimate of 20.03
(20.12, 0.05) based on analysis of 554 infants. Furthermore, the
I
2
heterogeneity test was 94% with the China study included, but
the model became homogeneous after removal of this study (I
2
=8%).
The pooled WAZ estimate for the LPFA group was 0.22 (0.01,
0.43) based on analysis of 965 infants from 9 studies (Figure 1).
Similar to the LPF analysis results, the 95% CI for the pooled
estimate was within 60.5 SD; this was also true for 6 of 9 studies
with LPFA arms. The China study (31) again appeared to be an
outlier, with a mean WAZ estimate of 0.81 and a lower bound of
the 95% CI of 0.68. When this study was removed in a sensitivity
analysis (n= 776), the pooled WAZ estimate was 0.14 (0.02, 0.26)
and the proportion of variance due to heterogeneity (I
2
)changed
from 92% to 42%.
The pooled WAZ estimate for the breastfed group was 20.23
(20.51, 0.05) based on analysis of 180 infants from 5 studies
(Figure 1), with an I
2
of 64%. The pooled mean WAZ estimate
was skewed slightly lower than the WHO growth standards by
data from 2 French studies (32, 33), both of which had mean
WAZ estimates and corresponding lower bounds of the 95% CIs
that were ,20.5 SD.
FORMULA PROTEIN CONTENT AND INFANT GROWTH 3of10
When compared between the 3 feeding groups (Table 3), the
WAZ estimates were significantly higher in the LPF (P,0.001)
and LPFA (P= 0.003) groups than they were in the breastfed
group. The mean WAZ differences were 0.30 (0.14, 0.46) for
the LPF group compared with breastfed infants and 0.24
(0.08, 0.39) for the LPFA group compared with breastfed
infants; both differences were ,0.5 SD. No significant dif-
ference was detected between the LPF and LPFA groups on
WAZ estimates (P=0.17).
LAZ at 4 mo of age
The pooled LAZ estimates for the LPF, LPFA, and breastfed
groups (Figure 2) were 20.02 (20.20, 0.17), 0.14 (20.06, 0.34)
and 0.19 (0.04, 0.34) based on analysis of 699, 921, and 180
infants from 10, 9, and 5 studies, respectively. The 95% CIs of
the pooled estimates were all within 60.5 SD. No statistically
significant differences in LAZ estimates were detected between
the LPF, LPFA, and breastfed groups (P.0.16, Table 3).
BMIAZ at 4 mo of age
The pooled BMIAZ estimates for the LPF and LPFA groups
(Figure 3) were 0.07 (20.19, 0.33) and 0.11 (20.10, 0.33)
based on analysis of 699 and 921 infants from 10 and 9 studies,
respectively. The 95% CIs for both groups were within 60.5 SD.
Similar to with the WAZ results, the China study (31) appeared
to be an outlier, with mean BMIAZ estimates of 0.89 and 0.69
with lower bounds of the 95% CIs of 0.70 and 0.52 (both of
which were .0.5 SD).
The pooled BMIAZ estimate for the breastfed group (Figure 3)
was 20.53 (20.79, 20.28) based on analysis of 180 infants
from 5 studies. Results from individual studies were homoge-
neous, with mean BMIAZ estimates for all individual studies
close to or below 20.5 SD. When compared with the LPF and
LPFA groups (Table 3), the breastfed group had significantly
lower BMIAZ (P,0.001 for breastfed infants compared with
the LPF group and P= 0.001 for BF infants compared with the
LPFA group). The difference was driven by the low BMIAZ
for the breastfed group, with BMIAZ estimates for the LPF
TABLE 1
Studies included in the multicenter IPD pooled analysis
1
Study, year Study arms included in the pooled analysis Reference
Italy, 1998 Formula-fed: 1.8 g protein/100 kcal Ra
¨iha
¨et al., 2002 (14)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs
Breastfed
Italy, 1999 Formula-fed: 1.8 g protein/100 kcal Barclay et al., 2003 (27)
Formula-fed: 1.8 g protein/100 kcal + prebiotics (Raftilose)
Formula-fed: 1.8 g protein/100 kcal + probiotics (BB12)
Formula-fed: 1.8 g protein/100 kcal + prebiotics (Raftilose) + probiotics (BB12)
Australia, 2002 Formula-fed: 1.8 g protein/100 kcal Gibson et al., 2009 (28)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + probiotics (B. lactis CNCM
I-3446)
Italy, 2003 Formula-fed: 1.8 g protein/100 kcal + LCPUFAs Puccio et al., 2007 (29)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + prebiotics (BMOSs) + probiotics
(BL999)
France, 2003 Formula-fed: 1.8 g protein/100 kcal + LCPUFAs Chouraqui et al., 2008 (30)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + probiotics (BL999 + LPR)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + prebiotics (BMOSs) + probiotics
(BL999 + LPR)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + prebiotics (BMOSs) + probiotics
(BL999 + ST11)
China, 2003 Formula-fed: 1.8 g protein/100 kcal + LCPUFAs Wu et al., 2016 (31)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + probiotics (BL999)
France, 2005a Formula-fed: 1.8 g protein/100 kcal Putet et al., 2016 (32)
Breastfed
France, 2005b Formula-fed: 1.8 g protein/100 kcal Hascoe
¨t et al., 2011 (33)
Formula-fed: 1.8 g protein/100 kcal + probiotics (BL999)
Breastfed
Italy, 2005 Formula-fed: 1.8 g protein/100 kcal + LCPUFAs Meli et al., 2014 (34)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + prebiotics (BMOSs)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + prebiotics + probiotics
(BMOSs + BL999 + LPR)
Breastfed
South Africa, 2007 Formula-fed: 1.8 g protein/100 kcal + LCPUFAs Cooper et al., 2015 (35)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + prebiotics (BMOSs) + probiotics
(B. lactis CNCM I-3446)
Greece, 2008 Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + probiotics (B. lactis CNCM I-3446) Baglatzi et al., 2016 (36)
Formula-fed: 1.8 g protein/100 kcal + LCPUFAs + probiotics (B. lactis CNCM
I-3446, higher level)
Breastfed
1
BB12, Bifidobacterium lactis BB12; BL999, Bifidobacterium longum (ATCC BAA999); BMOS, bovine milk–derived oligosaccharides; IPD, individual
participant data; LCPUFA, long-chain PUFA; LPR, Lactobacillus rhamnosus CGMCC 1.3724; ST11, Lactobacillus paracasei CNCM I-2116.
4of10 ALEXANDER ET AL.
and LPFA groups being similar to the WHO growth standards
(Figure 3).
HCAZ at 4 mo of age
The pooled HCAZ estimates for the LPF and LPFA groups
(Figure 4) were 0.37 (0.19, 0.55) and 0.45 (0.28, 0.62) based on
analysis of 736 and 959 infants from 10 and 9 studies, respectively.
Results from individual studies of these 2 groups were homoge-
neous, with mean HCAZ estimates close to 0.5 SD, except for
those for the South Africa study (35), in which mothers of infants
were HIV-positive. Interestingly, infants fed LPF and LPFA in the
South Africa study exhibited mean HCAZ estimates close to 1 SD.
The pooled HCAZ estimate for the breastfed group (Figure 4) was
0.27 (0.13, 0.4) based on analysis of 178 infants from 5 studies. No
significant differences between the LPF, LPFA, and breastfed
groups were detected on HCAZ estimates (P.0.4, Table 3).
Rate of weight gain based on change in WAZ from birth to
age 4 mo
The proportion of infants in weight gain categories based on
change in WAZ differed (P,0.001) between the LPF, LPFA,
and breastfed groups (Figure 5A). Specifically, compared with
TABLE 2
Baseline characteristics of infants included in the IPD pooled analysis according to feeding groups
1
LPF LPFA BF
Studies included, n10 9 5
Female 50 49 49
Cesarean section delivery 43 52 27*
Birth WAZ (n)20.10 60.84 (737) 20.12 60.90 (965) 0.14 60.77 (180)
#
Birth LAZ (n) 0.03 61.07 (700) 0.09 61.08 (921) 0.21 60.91 (180)
Birth BMIAZ (n)20.17 61.08 (700) 20.27 61.13 (921) 0.05 60.99 (180)
#
Birth HCAZ
2
(n) 0.14 61.07 (554) 0.22 61.06 (775) 0.30 60.93 (180)
1
Values are means 6SDs or percentages, unless otherwise indicated. Categorical data were analyzed with the use of
a chi-square test, and continuous data were analyzed with the use of ANOVA. An overall comparison between 3 groups was
conducted first, and if significant at P#0.05, pairwise comparisons between 2 groups were obtained. *LPF and LPFA are
significantly different from BF and are also different from each other at P#0.05.
#
LPF and LPFA are significantly different
from BF but not different from each other. BF, breastfed; BMIAZ, BMI-for-age zscore; HCAZ, head circumference–for-
age zscore; IPD, individual participant data; LAZ, length-for-age zscore; LPF, lower-protein infant formula; LPFA, lower-
protein infant formula with additional active ingredients (probiotics, prebiotics, or both); WAZ, weight-for-age zscore.
2
Birth HCAZ of infants from the study conducted in China (31) (which included LPF and LPFA arms) was not available.
FIGURE 1 IPD pooled analysis of WAZ in infants at 4 mo of age for the LPF, LPFA, and BF groups. Values are estimated means calculated from
ANCOVA with the use of random-effects models adjusted for birth WAZ, infant sex, delivery type, and study. The solid circles represent the estimated mean
from individual studies, and the horizontal lines represent the 95% CIs for the mean. The diamonds represent the pooled mean estimate, with the horizontal
tips of the diamond representing the lower and upper limits of the 95% CIs. BF, breastfed; IPD, individual participant data; LPF, lower-protein infant formula;
LPFA, lower-protein infant formula with additional active ingredients (probiotics, prebiotics, or both); WAZ, weight-for-age zscore.
FORMULA PROTEIN CONTENT AND INFANT GROWTH 5of10
the breastfed group, the LPF (P,0.001) and LPFA (P,0.001)
groups had a lower proportion of infants in the slow category
(22% and 18%, respectively, compared with 42%), a similar
proportion of infants in the gradual category (48% and 50%,
respectively, compared with 49%) and a higher proportion of
infants in the rapid category (30% and 32%, respectively,
compared with 9%). No difference was detected between the
LPF and LPFA groups (P= 0.13).
The China study (31) again appeared to be an outlier (Sup-
plemental Figure 2), with 57% and 54% of infants in the rapid
category for the LPF and LPFA groups, respectively. In addition,
the China study did not include a breastfed reference group;
therefore, no Chinese infants were included in the breastfed group
in the current analysis. Further analysis that excluded data from
the China study showed that the proportions of infants in weight
gain categories in the LPF (P,0.001) and LPFA (P,0.001)
groups still differed from those of the breastfed group, but were
numerically closer (Figure 5B), i.e., 21% and 26% instead of 30%
and 32% in the rapid category for the LPF and LPFA groups, re-
spectively. After the China study was excluded, the LPF group
differed from the LPFA group (P= 0.015), with a higher proportion
of infants in the slow category (27% compared with 22%), no
difference in the gradual category (52% compared with 52%), and
a lower proportion in the rapid category (21% compared with 26%).
DISCUSSION
We conducted a comprehensive pooled analysis on individual
data from 1882 healthy term infants in 11 RCTs to evaluate the
effects of a whey-predominant infant formula with lower protein
content with and without added active ingredients (prebiotics,
probiotics, or both) on growth parameters at 4 mo of age. Using
TABLE 3
Anthropometric zscore differences between feeding groups at age 4 mo
1
LPF vs. BF LPFA vs. BF LPFA vs. LPF
WAZ difference 0.30 (0.14, 0.46)* 0.24 (0.08, 0.39)* 20.06 (20.15, 0.03)
LAZ difference 0.04 (20.15, 0.23) 20.03 (20.21, 0.15) 20.07 (20.17, 0.03)
BMIAZ difference 0.37 (20.19, 0.33)* 0.11 (20.10, 0.33)* 20.53 (20.79, 20.28)
HCAZ difference 0.02 (20.16, 0.19) 20.01 (20.18, 0.15) 20.03 (20.12, 0.06)
1
Values are estimated zscore mean differences (95% CIs) between feeding groups (LPF 2BF, LPFA 2BF, and LPFA 2
LPF) calculated from ANCOVA while adjusting for corresponding birth zscores, infant sex, delivery type, and study.
*P#0.003. BF, breastfed; BMIAZ, BMI-for-age zscore; HCAZ, head circumference–for-age zscore; LAZ, length-for-age
zscore; LPF, lower-protein infant formula; LPFA, lower-protein infant formula with additional active ingredients (pro-
biotics, prebiotics, or both); WAZ, weight-for-age zscore.
FIGURE 2 IPD pooled analysis of LAZ in infants at 4 mo of age for the LPF, LPFA, and BF groups. Values are estimated means calculated from
ANCOVA with the use of random-effects models adjusted for birth LAZ, infant sex, delivery type, and study. The solid circles represent the estimated mean
from individual studies, and the horizontal lines represent the 95% CIs for the mean. The diamonds represent the pooled mean estimate, with the horizontal
tips of the diamond representing the lower and upper limits of the 95% CIs. BF, breastfed; IPD, individual participant data; LAZ, length-for-age zscore; LPF,
lower-protein infant formula; LPFA, lower-protein infant formula with additional active ingredients (probiotics, prebiotics, or both).
6of10 ALEXANDER ET AL.
pooled analysis methodology of IPD, we were able to further
analyze study data by harmonizing the covariates under study
with uniform analytic metrics (e.g., correcting for baseline
characteristics). Our pooled analyses generated summary associ-
ations with greater precision (i.e., enhanced statistical power) than
any of the individual studies. By virtue of combining individual
data, we created a single, larger analysis with greater analytic
control. This methodology is widely used (e.g., in the Harvard
Pooling Project of Prospective Studies of Diet and Cancer; https://
www.hsph.harvard.edu/pooling-project/about-the-study/).
The results of our analyses showed that a whey-predominant
infant formula with a protein content of 1.8 g/100 kcal supports
healthy growth that is comparable to the WHO growth standards.
Specifically, the WAZ, LAZ, and BMIAZ pooled estimates and
95% CIs at 4 mo in LPF- or LPFA-fed infants were well within
60.5 SD of the WHO growth standards. However, there was
some degree of data inflection within some of the models. This
mainly was due to outlier results from one study conducted in
Shanghai, a major urbanized city in China. The estimates for
WAZ and BMIAZ at 4 mo of age for LPF- and LPFA-fed infants
from the Chinese study (31) were considerably higher than those
from the other studies included in the analysis. This observation
is consistent with a recent publication (37) based on the Chinese
fourth National Survey on the Physical Growth and Development
of Children, which reported that urban Chinese infants were
heavier than those included in the WHO Multicenter Growth
Reference Study. Regardless, the exclusion of data from this
Chinese study did not modify the pooled results significantly. The
results of our analyses also showed that including specific active
ingredients (i.e., prebiotics, probiotics, or both) in LPF did not
significantly affect growth parameters at 4 mo of age. This is con-
sistent with a recent systematic review by Szajewska et al. (38),
which showed that infants fed Bifidobacterium lactis–supplemented
formula grew at a rate that was similar to that of infants fed
unsupplemented formula. One major difference between our
analysis of formula supplemented with active ingredients and
that of Szajewska et al. (38) is that our LPFA group included
several different types of active ingredients: different strains of
probiotics (e.g., B. lactis CNCM I-3446, BL999) or prebiotics
(bovine milk–derived oligosaccharides), or both. An interesting
finding is that the change in WAZ from birth to 4 mo of age
differed between the LPF and LPFA groups (Figure 5) after the
China study was excluded; however, interpretation of this finding
should be made in the context of the exploratory and descriptive
nature of this particular analysis.
It is noteworthy that HCAZ estimates of LPF- and LPFA-fed
infants were consistently higher than the WHO growth standards,
with infants in the South Africa study of HIV-positive mothers
exhibiting the greatest deviation. HCAZ estimates in breastfed
infants included in the analysis were also similarly higher than
the WHO standard. These HCAZ results mirrored the findings
from a recent systematic review (20), which showed that the
WHO head circumference data are at the lower end of head
circumference measurements from large studies of economically
advantaged children in 30 countries. The observed difference
between the WHO head circumference standard and recent studies
(including ours) may be the result of differences in head cir-
cumference measurement techniques; however, as noted in Natale
et al. (20), there is still a sizable difference between the WHO head
circumference data and a large European study that used a strict
FIGURE 3 IPD pooled analysis of BMIAZ in infants at 4 mo of age for the LPF, LPFA, and BF groups. Values are estimated means calculated from
ANCOVA with the use of random-effects models adjusted for birth BMIAZ, infant sex, delivery type, and study. The solid circles represent the estimated mean
from individual studies, and the horizontal lines represent the 95% CIs for the mean. The diamonds represent the pooled mean estimate, with the horizontal
tips of the diamond representing the lower and upper limits of the 95% CIs. BF, breastfed; BMIAZ, BMI-for-age zscore; IPD, individual participant data; LPF,
lower-protein infant formula; LPFA, lower-protein infant formula with additional active ingredients (probiotics, prebiotics, or both).
FORMULA PROTEIN CONTENT AND INFANT GROWTH 7of10
standardized measurement technique that mirrored the WHO study
methodology (39). Collectively, our IPD pooled analysis findings
support the conclusion of Natale et al. (20), and suggest that
additional research may be needed to justify using a single in-
ternational head circumference standard.
An interesting finding from our IPD pooled analysis was that
breastfed infants appeared to deviate from the WHO growth
standards for weight, length, and BMI to a larger extent than did
LPF- or LPFA-fed infants. Specifically, breastfed infants in our
study tended to be lighter and longer, and therefore manifested
FIGURE 4 IPD pooled analysis of HCAZ in infants at 4 mo of age for the LPF, LPFA, and BF groups. Values are estimated means calculated from
ANCOVA with the use of random-effects models adjusted for birth HCAZ (birth HCAZ for the China study (31) was not available, and birth WAZ was used
instead for this study), infant sex, delivery type, and study. The solid circles represent the estimated mean from individual studies, and the horizontal lines
represent the 95% CIs for the mean. The diamonds represent the pooled mean estimate, with the horizontal tips of the diamond representing the lower and
upper limits of the 95% CIs. BF, breastfed; HCAZ, head circumference–for-age zscore; IPD, individual participant data; LPF, lower-protein infant formula;
LPFA, lower-protein infant formula with additional active ingredients (probiotics, prebiotics, or both); WAZ, weight-for-age zscore.
FIGURE 5 Rate of weight gain between birth and 4 mo of age for the LPF, LPFA, and BF groups. IPD from all studies (A). The LPF (P,0.001) and
LPFA (P,0.001) groups were significantly different from the BF group, whereas no difference between the LPF and LPFA groups was detected (P= 0.13).
IPD from all studies excluding the China study (31) (B). The LPF (P,0.001) and LPFA (P,0.001) groups were significantly different from the BF group,
and the LPF group also was significantly different (P= 0.015) from the LPFA group. The WAZ change was calculated as WAZ at 4 mo of age minus WAZ at
birth. Values are percentage of infants in weight-gain categories based on WAZ change as slow (,20.67), gradual (20.67 to 0.67), or rapid (.0.67). A chi-
square test was used to compare the LPF, LPFA, and BF groups, and a Fisher’s exact test subsequently was used for pairwise comparisons. BF, breastfed; IPD,
individual participant data; LPF, lower-protein infant formula; LPFA, lower-protein infant formula with additional active ingredients (probiotics, prebiotics, or
both); WAZ, weight-for-age zscore.
8of10 ALEXANDER ET AL.
a significantly lower BMI than the WHO growth standards.
Multiple reasons may underlie this finding. For example, the
breastfed infants in our analysis were from France, Italy, and
Greece, all of which are countries not included in the WHO study.
In addition, the sample size (n= 180) of the breastfed group was
relatively small. Nonetheless, the relatively lower weight and
BMI of the breastfed infants included in our analysis was ac-
ceptable, as it was within 22 SD of the WHO standards. The
considerably homogeneous results of low BMIAZ in breastfed
infants in the 5 European studies included in our analysis also
are in agreement with the BMI zscores of the breastfed group
in the European Childhood Obesity Trial (6), which included
w500 breastfed infants from 5 European countries (Germany,
Belgium, Italy, Poland, and Spain). Furthermore, the breastfed
group from the Euro-Growth study also manifested relatively
low WAZ (compared to WHO Growth Standards) at 4 mo of age
(40). Additional research is warranted to explore any potential
long-term implications of the relatively lower weight and BMI
in early infancy in European infants.
Because early rapid weight gain has been shown to be asso-
ciated with obesity risk in later life (25, 26, 41, 42), the rate of
weight gain based on the change in WAZ was explored. In the
LPF, LPFA, and breastfed groups, w50% of the infants were in
the gradual weight-gain category. However, a relatively greater
proportion of breastfed infants (42% compared with #27% of
formula-fed infants) were in the slow category, and a relatively
greater proportion of formula-fed infants ($21% compared with
9% of breastfed infants) were in the rapid category, despite
consumption of formulas with a lower protein content closer to
that of breast milk. The observed difference in the rate of weight
gain may arise in part from a higher milk intake in formula-fed
infants than in breastfed infants (43). It also may be a statistical
caveat, because there were fewer infants in the breastfed group
(5 studies; n= 180). In addition, our ability to interpret these
findings is somewhat restricted because it is uncertain whether
the rate of weight gain in the breastfed group was skewed or
truly reflective of how infants should grow. For example, it is
theoretically desirable to have all infants in the gradual weight
gain category; however, only w50% of the breastfed infants in
the current analysis were in the gradual category, whereas 42%
were in the slow category.
One of the major strengths of our analysis was the ability to
pool individual-level data across 11 RCTs, with greater analytic
control being facilitated by the ability to harmonize covariates,
definitions, and analytic metrics. Another key strength was the
randomized controlled design of the included studies. This type
of design allows for greater control of influential confounding
factors, and is less susceptible to residual confounding than
observational studies. Moreover, there is control over the allo-
cation and, in theory, the compliance with the type of infant
feeding, given the experimental components of RCTs. Finally,
the analysis of data over a decade of research, at different study
centers, and across numerous geographic regions may have
enhanced the generalizability of our research findings.
Some limitations of our analysis also warrant mention. We
focused on infant formula with a lower-protein content (1.8 g
protein/100 kcal) and a specific protein quality (whey-predominant);
thus, findings may not be applicable to LPFs with a different protein
quality (e.g., casein-predominant). Maternal height and BMI (im-
portant factors influencing child growth, which were not controlled
in the analysis) may differ between the infant population evaluated
in the WHO growth reference study and the studies included in this
analysis. In addition, the RCTs used in the current analysis differed
somewhat in terms of operational methodology; not all trials in-
cluded both LPF and LPFA arms, and fewer trials included arms
with breastfed infants, thus resulting in a variable proportion of
infants representing the 3 study groups. However, given the fact that
we were able to conduct the analysis on individual-level data, we
had analytic latitude such that we could synchronize similar ex-
posure groups and outcome classifications.
In conclusion, it is well established that obesity is a major
public health burden in most developed countries, with a fore-
telling of risk in developing countries worldwide. Accumulating
scientific evidence suggests that excessive protein intake during
infancy may increase the subsequent risk of obesity. The results
of our pooled analysis of IPD from 11 RCTs have indicated that
feeding with a whey-predominant formula with a lower protein
content (1.8 g/100 kcal, lower than most currently available infant
formulas) that is closer to that of human milk, with or without pre-
or probiotics, supports healthy early growth comparable to the
WHO growth standards and close to that of breastfed infants. The
different rate of weight gain based on the change in WAZ between
breastfed infants and those fed LPF or LPFA warrants further
investigation.
We thank the lead investigators of the 11 studies: Niels Ra
¨iha
¨(Italy, 1998;
Italy, 1999), Maria Makrides (Australia, 2002), Giuseppe Puccio (Italy,
2003; Italy, 2005), Jean-Pierre Chouraqui (France, 2003), Weiping Wang
(China, 2003), Guy Putet and Jean-Charles Picaud (France, 2005a), Jean-
Michel Hascoe
¨t (France, 2005b), Peter A Cooper (South Africa, 2007), and
Christos Costalos (Greece, 2008).
The authors’ responsibilities were as follows—JY, RSN, DG, PS, PE, ES-K,
and FH: conceptualized and designed the study; RSN and DG: conducted the
analysis; DDA, JY, and LCB: wrote the paper with input from all authors; and
all authors: read and approved the final manuscript. DDA and LCB are em-
ployees of EpidStat Institute; JY, RSN, DG, PS, PE, and ES-K are employees
of Nestle
´; and FH was the chairman of the Nestle
´Nutrition Institute at the time
of initiation of the study.
REFERENCES
1. Singhal A, Lanigan J. Breastfeeding, early growth and later obesity.
Obes Rev 2007;8(Suppl 1):51–4.
2. Toschke AM, Grote V, Koletzko B, von Kries R. Identifying children at
high risk for overweight at school entry by weight gain during the first
2 years. Arch Pediatr Adolesc Med 2004;158:449–52.
3. Stettler N. Nature and strength of epidemiological evidence for origins
of childhood and adulthood obesity in the first year of life. Int J Obes
(Lond) 2007;31:1035–43.
4. Wells JC. The programming effects of early growth. Early Hum Dev
2007;83:743–8.
5. Weber M, Grote V, Closa-Monasterolo R, Escribano J, Langhendries
JP, Dain E, Giovannini M, Verduci E, Gruszfeld D, Socha P, et al.
Lower protein content in infant formula reduces BMI and obesity risk
at school age: follow-up of a randomized trial. Am J Clin Nutr 2014;
99:1041–51.
6. Koletzko B, von Kreis R, Closa R, Escribano J, Scaglioni S, Giovannini
M, Beyer J, Demmelmair H, Anton B, Gruszfeld D, et al. Can infant
feeding choices modulate later obesity risk? Am J Clin Nutr 2009;89:
1502S–8S. Corrected and republished from: Am J Clin Nutr 2009;
90(1):248.
7. Monasta L, Batty GD, Cattaneo A, Lutje V, Ronfani L, Van Lenthe FJ,
Brug J. Early-life determinants of overweight and obesity: a review of
systematic reviews. Obes Rev 2010;11:695–708.
8. Inostroza J, Haschke F, Steenhout P, Grathwohl D, Nelson SE, Ziegler
EE. Low-protein formula slows weight gain in infants of overweight
mothers. J Pediatr Gastroenterol Nutr 2014;59:70–7.
FORMULA PROTEIN CONTENT AND INFANT GROWTH 9of10
9. Lo
¨nnerdal B. Infant formula and infant nutrition: bioactive proteins of
human milk and implications for composition of infant formulas. Am J
Clin Nutr 2014;99:712S–7S.
10. Erdmann P, Neumann F. Process for treatment of a lactic raw materials.
Treatment of lactic materials. US patent 6787158 B1. 2014 Sep 7.
11. FDA. Title 21: food and drugs. Washington (DC): Office of the Federal
Government; 2015.
12. World Health Organization. CODEX alimentarius: standard for infant
formula and formulas for special medical purposes intended for infants.
Revised standard for infant formula. Rome (Italy): Food and Agriculture
Organization of the United Nations; 1981.
13. European Union. Commission Delegated Regulation (EU) 2016/127.
Official Journal of the European Union 2016;59.
14. Ra
¨iha
¨NC, Fazzolari-Nesci A, Cajozzo C, Puccio G, Monestier A,
Moro G, Minoli I, Haschke-Becher E, Bachmann C, Van’t Hof M, et al.
Whey predominant, whey modified infant formula with protein/energy
ratio of 1.8 g/100 kcal: adequate and safe for term infants from birth to
four months. J Pediatr Gastroenterol Nutr 2002;35:275–81.
15. Braegger C, Chmielewska AF, Decsi TF, Kolacek S, Mihatsch W,
Moreno LF, Piescik MF, Puntis J, Shamir R, Szajewska H, et al.
Supplementation of infant formula with probiotics and/or prebiotics:
a systematic review and comment by the ESPGHAN committee on
nutrition. J Pediatr Gastroenterol Nutr 2011;52:238–50.
16. Haschke F, Steenhout P, Grathwohl D, Haschke-Becher E. Evaluation
of growth and early infant feeding: a challenge for scientists, industry
and regulatory bodies. World Rev Nutr Diet 2013;106:33–8.
17. Haschke F, Grathwohl D, Detzel P, Steenhout P, Wagemans N. Post-
natal high protein intake can contribute to accelerated weight gain of
infants and increased obesity risk. Edition edited. In: Fewtrell MS,
Haschke F, Prescott SL, editors. 85th Nestle Nutrition Institute work-
shop, London, November 2014: Preventive aspects of early nutrition.
London, 2016.
18. Blettner M, Sauerbrei W, Schlehofer B, Scheuchenpflug T, Friedenreich
C. Traditional reviews, meta-analyses and pooled analyses in epidemi-
ology. Int J Epidemiol 1999;28:1–9.
19. WHO. 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.
20. Natale V, Rajagopalan A. Worldwide variation in human growth and
the World Health Organization growth standards: a systematic review.
BMJ Open 2014;4(1):e003735.
21. Stewart LA, Clarke M, Rovers M, Riley RD, Simmonds M, Stewart G,
Tierney JF. Preferred reporting items for systematic review and meta-
analyses of individual participant data: the PRISMA-IPD statement.
JAMA 2015;313:1657–65.
22. WHO Multicentre Growth Reference Study Group. Assessment of
differences in linear growth among populations in the WHO Multi-
centre Growth Reference Study. Acta Paediatr Suppl 2006;450(Suppl):
56–65.
23. WHO. An evaluation of infant growth: the use and interpretation of
anthropometry in infants. Bull World Health Organ 1995;73:165–74.
24. American Academy of Pediatrics. Clinical testing of infant for-
mulas with respect to nutritonal suitability for term infants. Elk
Grove (IL): American Academy of Pediatrics Committee on Nu-
trition; 1988.
25. Ekelund U, Ong K, Linne Y, Neovius M, Brage S, Dunger DB,
Wareham NJ, Rossner S. Upward weight percentile crossing in infancy
and early childhood independently predicts fat mass in young adults:
the Stockholm Weight Development Study (SWEDES). Am J Clin
Nutr 2006;83:324–30.
26. Ong KK, Loos RJ. Rapid infancy weight gain and subsequent obe-
sity: systematic reviews and hopeful suggestions. Acta Paediatr
2006;95:904–8.
27. Barclay D, Puccio G, Fazzolari-Nesci A, Giammanco A, Ra
¨iha
¨N,
Carrie-Fa
¨ssier A, Brown C, Chauffard F, Grathwohl D, Hager C, et al.
Growth and tolerance of a whey-based starter infant formula with
enhanced protein efficiency and containing Pro-, Pre-, or synbiotics. A
randomized controlled trial in term infants. J Pediatr Gastroenterol
Nutr 2003;37:388.
28. Gibson RA, Barclay D, Marshall H, Moulin J, Maire JC, Makrides M.
Safety of supplementing infant formula with long-chain polyunsaturated
fatty acids and Bifidobacterium lactis in term infants: a randomised
controlled trial. Br J Nutr 2009;101:1706–13.
29. Puccio G, Cajozzo C, Meli F, Rochat F, Grathwohl D, Steenhout P.
Clinical evaluation of a new starter formula for infants containing live
Bifidobacterium longum BL999 and prebiotics. Nutrition 2007;23:1–8.
30. Chouraqui JP, Grathwohl D, Labaune JM, Hascoet JM, de Montgolfier
I, Leclaire M, Giarre M, Steenhout P. Assessment of the safety, tol-
erance, and protective effect against diarrhea of infant formulas
containing mixtures of probiotics or probiotics and prebiotics in
a randomized controlled trial. Am J Clin Nutr 2008;87:1365–73.
31. Wu BB, Yang Y, Xu X, Wang WP. Effects of Bifidobacterium sup-
plementation on intestinal microbiota composition and the immune
response in healthy infants. World J Pediatr 2016;12:177–82.
32. Putet G, Labaune JM, Mace K, Steenhout P, Grathwohl D, Raverot V,
Morel Y, Picaud JC. Effect of dietary protein on plasma insulin-like
growth factor-1, growth, and body composition in healthy term infants:
a randomised, double-blind, controlled trial (Early Protein and Obesity
in Childhood (EPOCH) study). Br J Nutr 2016;115:271–84.
33. Hascoe
¨t JM, Hubert C, Rochat F, Legagneur H, Gaga S, Emady-Azar
S, Steenhout PG. Effect of formula composition on the development of
infant gut microbiota. J Pediatr Gastroenterol Nutr 2011;52:756–62.
34. Meli F, Puccio G, Cajozzo C, Ricottone GL, Pecquet S, Sprenger N,
Steenhout P. Growth and safety evaluation of infant formulae con-
taining oligosaccharides derived from bovine milk: a randomized,
double-blind, noninferiority trial. BMC Pediatr 2014;14:306.
35. Cooper PA, Bolton KD, Velaphi SC, Pecquet SS, Steenhout PG.
Normal growth of infants born from HIV + mothers, after normal or
caesarean delivery, fed a pre and probiotics supplemented starter for-
mula. The European Society for Paediatric Gastroenterology Hepatology
and Nutrition Amsterdam, 2015.
36. Baglatzi L, Gavrili S, Stamouli K, Zachaki S, Favre L, Pecquet S,
Benyacoub J, Costalos C. Effect of infant formula containing a low
dose of the probiotic Bifidobacterium lactis CNCM I-3446 on immune
and gut functions in C-section delivered babies: a pilot study. Clin Med
Insights Pediatr 2016;10:11–9.
37. Zong XN, Li H. Construction of a new growth references for China
based on urban Chinese children: comparison with the WHO growth
standards. PLoS One 2013;8:e59569.
38. Szajewska H, Chmielewska A. Growth of infants fed formula supple-
mented with Bifidobacterium lactis Bb12 or Lactobacillus GG: a system-
atic review of randomized controlled trials. BMC Pediatr 2013;13:185.
39. van’t Hof MA, Haschke F. The Euro-Growth Study: why, who, and
how. J Pediatr Gastroenterol Nutr 2000;31(Suppl 1):S3–13.
40. Ziegler EE. Obesity prevention: is the protein intake in infants too
high? The Nestle Nutrition Institute Newsletter. South East Asia and
Pacific Rim. 2013;(7):3.
41. Nettleton JA, Jebb S, Riserus U, Koletzko B, Fleming J. Role of dietary
fats in the prevention and treatment of the metabolic syndrome. Ann
Nutr Metab 2014;64:167–78.
42. Druet C, Stettler N, Sharp S, Simmons RK, Cooper C, Smith GD,
Ekelund U, Levy-Marchal C, Jarvelin MR, Kuh D, et al. Prediction of
childhood obesity by infancy weight gain: an individual-level meta-
analysis. Paediatr Perinat Epidemiol 2012;26:19–26.
43. Kramer MS, Guo T, Platt RW, Vanilovich I, Sevkovskaya Z, Dzikovich
I, Michaelsen KF, Dewey K. Feeding effects on growth during infancy.
J Pediatr 2004;145:600–5.
10 of 10 ALEXANDER ET AL.
