 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PEDIATRICS Vol. 106 No. 6 December 2000, pp. 1355-1366
From the Department of Pediatrics, USDA/ARS Children's Nutrition Research Center, Baylor College of Medicine, Houston, Texas.
 |
ABSTRACT |
|---|
Top
Abstract
MaterialsMethods
Results
Discussion
References
|
|---|
Background. Differences in the growth pattern of breastfed (BF) and formula-fed (FF) infants are well-recognized and have been attributed to differences in nutrient intake. However, the impact of qualitative and quantitative differences in nutrient intake on the body composition of BF and FF infants has been unclear. Furthermore, it is unknown whether putative differences in body composition persist beyond weaning.
Design. Prospective cohort study.
Methods. Repeated anthropometric and body composition measurements were performed on 40 BF and 36 FF infants at 0.5, 3, 6, 9, 12, 18, and 24 months of age. A multicomponent body composition model based on total body water by deuterium dilution, total body potassium by whole body counting, and bone mineral content by dual-energy x-ray absorptiometry was used to estimate fat-free mass (FFM) and fat mass (FM). Independent measurements of FFM and FM were made using total body electrical conductivity and dual-energy x-ray absorptiometry. By design, infants were either exclusively BF or FF from birth to 4 months of age; thereafter, the feeding mode was at the discretion of the parents. Infant food intake was measured at 3, 6, 12, and 24 months of age using 3-day weighed-intake records. Data were analyzed by repeated measures analysis of variance.
Results. Weight velocity was higher in FF than BF infants age 3 to 6 months, and higher in FF than BF girls 6 to 9 months of age. Adjusted for gender and baseline values, BF infants had lower total body water at 3 months, lower total body potassium at 3 to 24 months, and lower bone mineral content at 12 months. The multicomponent model indicated that FFM was lower in BF than FF infants at 3 months, and FM and %FM were higher in BF than FF infants at 3 and 6 months (boys only). Total body electric conductivity confirmed lower FFM in BF than FF infants at 3 months, as well as at 6 and 9 months; FM and %FM were higher in BF than FF at 3 and 6 months, and 9 months (boys only). Intakes of energy, protein, fat, and carbohydrate were lower in BF than FF infants at 3 and 6 months, and were positively correlated with weight gain and FFM gain, but not FM gain. No differences in nutrient intakes were observed at 12 or 24 months.
Conclusion. Infant feeding mode is associated with differences in body composition in early infancy which do not persist into the second year of life. Key words: breastfeeding, formula-feeding, body composition, infants, toddlers, total body water, total body potassium, total body electrical conductivity, dual-energy x-ray absorptiometry, fat mass, fat-free mass.
Differences in the growth pattern of breastfed (BF) and
formula-fed (FF) infants are well recognized and have been attributed to differences in nutrient intake.1 In studies with
clearly-defined feeding groups and sufficient sample sizes, rates of
weight gain are lower in BF than FF infants. BF infants consume not
only less energy than FF infants, but also disproportionately less
protein and fewer micronutrients.2-5 The higher nutrient
density of formulas, in particular the higher protein:energy ratio, may
promote greater accretion of fat-free mass (FFM). Studies in
pigs,6-9 rats,10 and mice11
indicate that diets higher in protein than the milks produced by the
respective species are associated with higher rates of weight, protein,
and potassium gain, and a decrease in the ratio of extracellular to
intracellular water. Balance studies in infants have shown that
increased nutrient intakes result in increased retention of nitrogen
and minerals.12-14 Newborn FF infants were shown to
retain more nitrogen, potassium, phosphorus, magnesium, and calcium
than BF infants, suggesting an acceleration of the chemical maturation
of lean tissue. Higher nitrogen intakes in FF infants resulted in
higher retention of nitrogen for a given weight gain.14
The effects of different nutrient retention rates on infant body composition have not been determined.
The impact of qualitative and quantitative differences in nutrient
intake on the body composition of BF and FF infants has been unclear.
Findings based on anthropometry and the few in vivo body composition
studies have been inconsistent.4,15-20 Further, it is
unknown whether putative differences in body composition persist beyond
weaning. Protective effects of breastfeeding on later adiposity have
been found in some studies,21,22 but refuted in other
studies.23-27 We, therefore, undertook the following
study to assess the impact of infant feeding mode on growth and body
composition during the first 2 years of life.
Study Design and Participants
Repeated anthropometric and body composition measurements were
performed on 76 healthy, term infants at 0.5, 3, 6, 9, 12, 18 and 24 months of age at the Children's Nutrition Research Center (CNRC).
Sample size was 76, 76, 75, 74, 74, 71, and 72 at 0.5, 3, 6, 9, 12, 18, and 24 months, respectively. Attrition was due to relocation outside
the Houston area. Food intake was measured at 3, 6, 12, and 24 months
of age using 3-day weighed intake records. By study design, the infants
were either exclusively BF (n = 40) or FF
(n = 36) from birth to 4 months of age; thereafter, the feeding mode was at the discretion of the parents. This study was
approved by the Baylor Affiliates Review Boards for Human Subject
Research, and informed written consent was obtained from each child's
mother.
Pregnant women were recruited from the Houston area through the CNRC
community-based referral system. The following screening criteria were
used: healthy, nonsmoking, ages 18 to 35 years, parity not >4, and no
chronic medications or alcohol/drug abuse. Unremarkable health history
and pregnancy course were confirmed by the volunteer's obstetrician.
All women agreed to either exclusively breastfeed or formula-feed their
infants for the first 4 months of life. Infants were required to be
healthy and full-term.
Infants were admitted to the CNRC Metabolic Research Unit from ~1000
to 1700 hours for a series of anthropometric and body composition
measurements. The relative timing of these measurements was as follows.
Anthropometric measurements were performed Infant Food Intake
At each 3-month or 6-month study interval, an infant feeding
history was taken. Caretakers were asked about breastfeeding frequency,
and the use of formula, milk, juice, solids, water, and vitamin-mineral
supplements. In addition, infant food intake was quantitated at 3, 6, 12, and 24 months with a 3-day weighed intake record.
In the case of BF infants, the intake of human milk was assessed by
test-weighing.28 Infant weights were measured before and
after each feeding on electronic, integrating scales with a precision
of ± 1.0 g (model 3862 MP, Sartorious, Göttingen,
Federal Republic of Germany). For the determination of milk composition
at 3 and 6 months, mothers expressed all their milk during a 24-hour
period with an electric pump (Egnell, Inc, Cary, IL) while in the
metabolic research unit for other studies. The 24-hour expressed milk
volumes were greater than the human milk intake by 20%. After each
pumping session, the milk was weighed and a 10% aliquot was
refrigerated and later pooled for the 24-hour analysis. The energy
content of human milk was determined by adiabatic bomb calorimetry
(Parr Instruments, Moline, IL). Nitrogen was analyzed by
the Kjeldahl method before and after trichloroacetic acid precipitation
of protein (Kjeltec Auto Analyzer 1030, Tecator, Hoganas, Sweden).
Nonprotein nitrogen was determined on the supernatant, and protein
nitrogen was estimated from the difference between total nitrogen and
nonprotein nitrogen. Nitrogen was assumed to be 15.7% of total protein
(conversion factor 6.38). Lactose was determined using an automatic
analyzer (YSI, model 127, Yellow Springs, OH). Fat was
determined by the Jeejeebhoy method.29 To convert gross
energy to metabolizable energy, a factor of 0.95 was
used.30
The intake of formula and other foods and beverages was determined for
3 days using a preweighing and postweighing method. The mother received
instruction on use of a digital food scale (Ohaus model LS2000,
Florham, NJ), and forms with which to weigh and record all food
consumed. Preweighed towels were provided to recover any losses. The
nutrient content of the 3-day diet was analyzed by the investigators
using the Nutrition Data System (version 2.3/5A/20, Nutrition
Coordinating Center, University of Minnesota, Minneapolis,
MN).31
Anthropometry
Infants were weighed naked on an electronic integrating scale
(Sartorius MC1, LC34, Gottingen, Federal
Republic of Germany; precision ± 1.0 g). Crown-to-heel
length was measured on a recumbent infant board to the nearest 1 mm
(Holtain Limited, Crymych, United Kingdom). Weight and length
velocities were calculated from the difference in weight or length
between adjacent study points and divided by the exact number of days
between measurements. Weights and lengths were converted to
weight-for-age, length-for-age, and weight-for-length z
scores using the National Center for Health Statistics (NCHS)
Growth Curves for Children.32
Circumferences of the head, chest, arm, and thigh were measured to the
nearest 1 mm using a nonextensible measuring tape. Skinfold thicknesses
at the triceps, subscapular, flank, and quadriceps were measured on the
left side of the body to the nearest 0.5 mm using Lange calipers
(Cambridge Scientific Industries, Cambridge, MD), according to the
method outlined by McGowan et al.33 All anthropometric
measurements were performed by the same trained individual with an
assistant to position the infant.
TBW
TBW was determined by dilution of an orally administered dose of
deuterium oxide (50 or 100 mg of
2H2O/kg body weight). Urine
samples were collected 3 to 5 hours after dosing or daily for 10 days
with cotton balls placed in the infant's diaper.34 Before
analysis, hydrogen gas was generated from undistilled urine samples by
zinc reduction in quartz vessels.35,36 2H
abundance in the urine samples was measured by gas-isotope-ratio mass
spectrometry (Delta-E, Finnigan MAT, San Jose, CA).
Deuterium dilution space was calculated from the average of 2 postdose
urine samples by the plateau method at 0.5 months of age, and from 10 daily urine samples by the extrapolation method at all other ages. The
deuterium dilution space (NH) was calculated
using the following equation:
MATERIALS AND METHODS
Top
Abstract
MaterialsMethods
Results
Discussion
References
1 h after feeding,
followed by the total body electric conductivity (TOBEC) measurement.
The 2H2O dose for the total
body water (TBW) measurement was administered 30 minutes after eating
to avoid regurgitation. The 15-minute whole body counting of
40K and the dual-energy x-ray absortiometry (DXA)
measurement were performed next. After discharge, a 3-day infant food
intake record was completed at home by the infant's caretaker (usually
the mother).
where D is the dose of
2H2O in g; A is the amount
of laboratory water, in g, used in the dose dilution;
is the amount
of 2H2O, in g, added to the
laboratory water in the dose dilution; E
is
the rise in 2H abundance, per mil, in the
laboratory water after the addition of the isotopic water;
Ed is the rise in 2H
abundance, per mil, in the postdose urine sample. Deuterium dilution
space was converted to TBW by dividing by 1.04.
Total Body Potassium (TBK)
TBK was estimated from the 40K naturally
present in the child's body using the CNRC whole body
counter.37 One gram of K emits 
-rays (1.46 MeV) at the
constant rate of 200.4 photons/min, which are detected by 12 NaI(Tl)
detectors arranged in two arrays above and below the child's body. The
detectors are inside a shielded room to reduce background interference. During the 15-minute count, the younger children were swaddled inside a
plastic bassinet for safety. The older children were secured on a
hammock with straps. A custom set of four phantoms was used for routine
quality control of the instrument. The in vivo precision for TBK
measurements of infants and toddlers is <2.5%, which equates to an
error of ~0.04 kg FM (fat mass) at 0.5 months and 0.27 kg FM at 24 months of age.37
TOBEC
TOBEC was used to measure FFM and FM (model HP-2, EM-SCAN, Inc., Springfield, IL).38 The child was undressed and swaddled in a sheet such that the arms and legs were fully extended and parallel to the main axis of the body and to restrict movement. The child was then placed supine on the TOBEC instrument sled, which was inserted into the TOBEC opening, and the maximum (peak) TOBEC number was recorded. FM of the child was calculated as follows:
![<UP>FM </UP>(<UP>kg</UP>)<UP> = WT − </UP>[<UP>0.0265&cjs3484;</UP>(<UP>TOBEC# × L</UP><SUB><UP>c</UP></SUB>)<UP> − 0.0313</UP>]](/content/vol106/issue6/fulltext/1355/img002.gif)
|
DXA
DXA was used to estimate bone mineral content (BMC), fat, and lean mass at three ages (0.5, 12 and 24 months) only. The whole body was scanned in the single-beam mode with a Hologic QDR-2000 instrument using Infant Whole Body Analysis, software version 5.56-5.71P (Hologic, Inc, Waltham, MA). For the DXA scans, the child was lightly dressed, diapered, and placed supine on the bed with an interposing paper sheet; no sedation was used. DXA measurements were usually performed on the younger infants while they were sleeping. At 12 to 24 months, measurements could be done while the child was awake and distracted by watching a video on a recorder nearby the DXA scanner. If the infants moved, the scan was stopped and restarted. A step phantom was placed beside the child for soft tissue calibration. The precision of the DXA measurement was determined for bone mineral (1.2%-4.1%), fat (3.0%-4.2%) and lean mass (1.0%-1.5%) based on repeated measurements in pigs weighing 4.6 to 15.7 kg.39 In terms of FM, an average precision of 3.6% equates to an error of 0.02 kg FM at 0.5 months and 0.11 kg at 24 months of age.
Multicomponent Body Composition Model
Body composition was estimated using a modified version of the multicomponent model published by Fomon et al40:
|
|
This model was based on the following assumptions. Intracellular and extracellular water can be estimated using TBW and TBK, and the known concentrations of potassium in each compartment (4 and 150 mEq/kg water, respectively). Body protein mass was based on the ratio of nitrogen to potassium (461 mg/mEq), and the fact that total body protein contains 16% nitrogen. To predict BMC when measurements were not available, a prediction equation was developed using linear regression of BMC on TBK, heel-crown length, age and gender at 0.5, 12, and 24 months. Heel-crown length and gender were not significant, and therefore, were eliminated from the model. The final prediction equation was as follows:
|
Statistics
Minitab (release 12, Minitab Inc, College Station, PA, 1998) was used for data description and statistical analyses, including Pearson correlation and linear regression. Repeated measures analysis of variance with fixed and time-varying covariates (5V; BMDP Statistical Software, Inc, Los Angeles) was used to test the effects of feeding mode on growth and body composition. The value of the dependent variable at 0.5 months was used as a covariate to adjust for initial anthropometric and body composition values. Current weight and length were entered as covariates in the body composition analyses to control for body size. The basic model included grouping factors for initial feeding mode (BF or FF) and gender, a time factor (3, 6, 9, 12, 18, and 24 months of age), and interactions between feeding mode, gender, and age. Significant interactions were further investigated by making comparisons between BF and FF infants with one-way analysis of variance. After evaluation of several growth models programmed into TableCurve 2D (version 4, Jandel Scientific, San Rafael, CA), the Jenss model, a four-parameter nonlinear function with exponential and linear components, was chosen to fit the individual weight and length data41,42 based on the highest adjusted r2. The equation used was:
|
| |
RESULTS |
|---|
|
Top
Abstract
MaterialsMethods
Results
Discussion
References
|
|---|
Maternal and infant characteristics are described in Table 1. Family income was distributed similarly in both feeding groups: 8% under $20 000, 24% between $20 000 and 34 999, 17% between $35 000 and 49 999, and 51% above $50 000. Maternal characteristics did not differ by infant feeding mode, except attained level of education which was higher in the BF than the FF group (P = .001).
|
There were no statistically significant differences in weight, length, or gestational age between feeding groups at birth (Table 1). The average duration of breastfeeding was 11.4 ± 5.8 months (mean ± standard deviation). Infants were exclusively breastfed for at least 4 months, with 3 exceptions: one was weaned at 109 days, another received formula at 102 days, and another was given cereal at 106 days of age. The percentage of BF children still breastfed at 3, 6, 9, 12, 18 and 24 months was 100%, 80%, 58%, 38%, 15%, and 5%. The average duration of formula feeding was 4.4 ± 4.5 months among the BF group and 11.9 ± 3.8 months among the FF group. The percentage of BF children given formula at 3, 6, 9, 12, 18 and 24 months was 0%, 40%, 48%, 30%, 10%, and 2%. For the FF children, the corresponding percentages were 100%, 100%, 94%, 47%, 6%, and 0%. Seventeen (42%) of the BF infants were not given formula. Among the FF infants, 9 were given small amounts of cereal and/or fruit before 4 months of age. The use of cow's milk averaged 8%, 65%, 82%, and 88% among BF infants and 28%, 67%, 89%, and 92% among FF infants at 9, 12, 18 and 24 months. Infant vitamin-mineral supplements were given to 38%, 38%, 40%, 48%, 55%, and 60% of BF infants and 11%, 14%, 17%, 22%, 22%, and 36% of FF infants at 3, 6, 9, 12, 18, and 24 months.
Mean intakes of human milk and formula are summarized in Table 2. Macronutrient intakes derived from the 3-day weighed-intake records are summarized in Table 3. Metabolizable energy intake and intakes of protein, fat, and carbohydrate were significantly lower in BF than FF infants at 3 and 6 months (P = .001). Protein (%ME) was lower at 3 and 6 months in BF than FF infants (P = .001). Fat (%ME) did not differ significantly between feeding groups. No significant differences in nutrient intake were observed between feeding groups at 12 and 24 months. Infants still breastfed received 99 ± 4%, 76 ± 29% and 38 ± 20% of energy from human milk at 3, 6, and 12 months, respectively. Infants still formula-fed received 99 ± 9%, 80 ± 16% and 31 ± 19% of energy from formula at 3, 6, and 12 months, respectively.
|
|
Weight and length measurements are summarized in Table 4. A three-way interaction among feeding mode, gender, and age was detected for weight, and adjusted for initial weight (P = .04). Weight was higher in the FF than BF girls at 9 and 12 months. Length tended to be lower in BF than FF infants (P = .07). A three-way interaction also was detected for weight velocity (P = .04). Weight velocity (g/d) was higher among FF than BF infants between 3 to 6 months, and higher in FF than BF girls between 6 to 9 months.
|
The Jenss model applied to weight revealed the following: 1) no differences between feeding groups for the intercept or birth weight (parameter A); 2) an interaction between feeding mode and gender (P = .02) for the infancy increment (parameter B), with lower values in BF than FF girls; 3) a significant difference (P = .05) for the rate of exponential decay (parameter C), with higher values in BF than FF infants; and 4) an interaction between feeding mode and gender (P = .04) for weight velocity in later infancy (parameter D), with higher values in BF than FF girls. The Jenss model applied to length revealed no significant differences between feeding groups. The coefficients for the Jenss model for weight and length are given in Table 5. The adjusted r2 indicated that the Jenss model fit the individuals' data satisfactorily.
|
No statistically significant differences were seen between feeding groups when compared with NCHS weight-for-age, length-for-age or weight-for-length z scores (Fig 1).
|
Head, chest, arm, and thigh circumferences also did not differ between feeding groups (data not shown). No differences in skinfold thicknesses at the triceps, subscapular, flank, and quadriceps were observed between feeding groups (Fig 2). Sums of skinfold thicknesses were not different between feeding groups.
|
TBW was significantly lower in BF than FF infants only at 3 months (P = .02) (Fig 3), yet TBK (g) was significantly lower in the BF than FF group at all time points (P = .04). Significant gender (M > F; P = .02) and time effects (P = .001) were also observed for TBK. The lack of a significant feeding mode*time interaction (P = .37) indicated that TBK was lower overall in the BF group. TBK of the BF infants was 101%, 96%, 92%, 94%, 94%, 98%, and 96% of the values observed for the FF infants at 0.5, 3, 6, 9, 12, 18, and 24 months of age, respectively. Mean TBK was 15.7 ± 1.6 g versus 16.7 ± 1.9 g at 12 months, 19.0 ± 2.0 g versus 19.5 ± 2.1 g at 18 months, and 21.7 ± 2.7 g versus 22.6 ± 2.3 g at 24 months in BF versus FF infants. BMC, adjusted for weight, length, bone area, and BMC at 0.5 months, was significantly lower in BF than FF infants at 12 months (P = .04), but not at 24 months.
|
FFM, FM, and %FM estimated by the multicomponent model are shown in Figure 4. Body composition differences between feeding groups were age-dependent and gender-dependent (P = .001). FFM was lower in BF than FF infants at 3 months (P = .01). FM and %FM were significantly higher in BF than FF infants at 3 months (P = .02), and in BF than FF boys at 6 months (P = .05). No significant differences were seen at 9, 12, 18, or 24 months.
|
TOBEC-derived FFM, FM and %FM differed between feeding groups at
specific ages (P = .001) (Fig
5). FFM was lower in BF than FF infants
at 3, 6 and 9 months (P 
.02). FM was higher in BF
than FF infants at 3 and 6 months (P .01). %FM was
higher in BF than FF infants at 3 and 6 months (P .03), and higher in BF than FF boys at 9 months (P = .05). No significant differences were seen at 12, 18, or 24 months.
|
DXA-derived estimates of FFM, FM, and %FM, adjusted for values at 0.5 months, did not differ between feeding groups at 12 or 24 months of age (Fig 6).
|
The changes in FM and FFM between study time points were calculated for the multicomponent model and TOBEC. Feeding mode, gender, and their interactions were not significantly different.
Among the FF infants, neither the duration of formula-feeding, nor the
age of introduction of solids or other milks, had an effect on body
size and composition, adjusted for gender and baseline values. Among
the BF infants, the age of introduction of formula was negatively
associated with weight at 6 to 24 months (P .03) and
with FFM at 6 to 9 months (P = .02), and the duration
of formula feeding was positively correlated with weight at 12 and 18 months (P .06), after adjustment for gender, initial
weight and duration of breastfeeding.
Duration of breastfeeding was not correlated significantly with weight, length, FFM, or FM. Adjusted for gender and baseline values, weight, length, TBK, FFM, FM, and %FM of BF infants who were breastfed for >12 months (n = 18) did not differ at any time point from infants breastfed for <12 months (n = 22). Adjusted for baseline BMC, bone area, weight and length, BMC at 24 months in infants who were breastfed for >12 months was lower than infants breastfed <12 months (289 vs 310 g; P = .05). Duration of breastfeeding was positively correlated to the exponential decay factor in Eq. 6 for length (r = 0.426; P = .007).
Adjusted for gender and baseline values, weight, FFM, FM, and %FM of
BF infants who were breastfed for >12 months (n = 18) did not
differ from the formula-fed infants at 12, 18, or 24 months. Length,
TBK, and BMC were significantly lower among the infants breastfed for
>12 months than FF infants at 12 months (P .02). BMC was also lower in these BF infants than the FF infants at 24 months
(P = .01).
Daily intakes of energy, protein, fat, and carbohydrate were positively correlated with weight (r = 0.23-0.48; P = .05-0.001) and FFM (r = 0.41-0.56; P = .001), but not with FM or %FM, at 3, 6, and 12 months. Daily intakes of energy, protein, fat, and carbohydrate were positively correlated with weight gain (r = 0.23-0.40; P = .05-0.001) and FFM gain (r = 0.28-0.64; P = .04-0.001) between 0.5 to 3 months, 3 to 6 months, and 9 to12 months, but not 18 to 24 months. Nutrient intakes were not associated with changes in FM or %FM. Protein (%ME) was positively correlated with weight gain at 3 to 6 months (r = 0.38; P = .001) and FFM gain at 3 to 6 months (r = 0.58; P = .001), but not FM gain.
| |
DISCUSSION |
|---|
|
Top
Abstract
MaterialsMethods
Results
Discussion
References
|
|---|
In this cohort of BF and FF infants, differences in weight gain and body composition coincident with differences in nutrient intake were seen during the suckling period of early infancy. FFM was lower, and FM and %FM were higher in BF than FF infants between 3 to 9 months of age. With the introduction of other milks and supplementary foods, differences in body composition between feeding groups diminished and were no longer apparent in the second year of life.
The growth pattern of BF infants is known to deviate from the
NCHS32 growth reference, which was derived primarily from
FF infants. A pooled World Health Organization (WHO) analysis of growth
data on BF infants from the United States, Canada, and Europe showed consistent downward trends after 2 to 3 months in NCHS weight-for-age, length-for-age, and weight-for-length z scores to means of
0.5, 0.29 and 0.32 at 12 months of age,
respectively.43 The weight curve of the BF infants in the
first 12 months also displayed more curvature compared with the NCHS
reference, although in the second year of life the BF infants grew more
rapidly, as was shown in our data by the Jenss model. The Jenss growth
model analysis confirmed lower weight gain in early infancy (parameter B) among BF girls compared with FF girls, and also indicated greater weight velocity in later infancy (parameter D). Furthermore, growth rates declined more rapidly in BF than FF infants (parameter C). Several studies, but not all,44-49 have confirmed that BF
infants grow more slowly in terms of weight than FF
infants.3,4,16,19,50-54
The growth of our infants is most appropriately compared with
contemporary studies of modern infant formulas and weaning practices. In the WHO pooled data sets from 1984 to 1994, the duration of breastfeeding was related to the magnitude of the decline in
weight-for-age and weight-for-length z scores; infants
breastfed for more than 12 months exhibited a continuing decline in
z scores.1 Infants who were heavier initially
were less likely to be given formula or solids. In our study, the
duration of breastfeeding did not impact infant growth or body
composition, although our statistical power was limited by our sample
size of 40. There was no evidence of differences in body size or
composition in infants breastfed for 12 months, with the exception of
BMC at 24 months. Further, the introduction of solids did not influence the growth of BF or FF infants, in agreement with Mehta et
al,55 who reported that early introduction of solids does
not alter growth or body composition during the first year of life, but results in a displacement of calories from formula. However, we did
find that the age of introduction and duration of formula use affected
the weight of the BF infants.
In the DARLING study, the mean weight of 46 BF infants was lower than that of 41 FF infants between 6 to 18 months51,56 As in our study, differences in weight were most pronounced in the second 6 months of age, especially among the girls, consistent with lower energy intakes of BF infants at 6 and 9 months of age. In contrast to our results, weight-for-length z score was significantly lower between 4 to 18 months in the DARLING study, and the sum of skinfold thicknesses of BF was less than that of FF infants 9 to 17 months of age. While solids were not given before 4 months, as in our study, the DARLING infants were either breastfed or formula-fed for the first 12 months of life. This represents an important difference in the weaning practices between the two studies. Introduction of formula in the first year of life influenced growth and body composition of our BF infants. In the WHO pooled analysis, exclusivity of breastfeeding was associated with lower weight-for-age and weight-for-length z scores at 12 months, as illustrated by the DARLING study, which had the lowest weight-for-length z scores.
Skinfold thicknesses have been used as an indirect indicator of body fatness in a number of studies comparing BF and FF infants. Greater increases in skinfolds in BF than FF infants were reported in early48,57 and later infancy,53 in contrast to other studies which detected greater increases in FF infants in early4,19,58 and later infancy.56 Several studies failed to detect differences between feeding groups.3,44,45,47,49,54 Although the pattern of skinfolds paralleled body fat changes, we did not detect any significant differences in skinfolds between feeding groups.
A limited number of studies comparing BF and FF infants used in vivo measurements of body composition. Shepherd19 made serial measurements of TBK for the first 3 months of life and Rutledge20 for the first 14 months of life. Shepherd found that at similar rates of weight gain, formula feeding resulted in a higher TBK/weight ratio at 3 months in girls. Rutledge did not observe differences in TBK accretion in BF and FF infants at similar rates of weight gain. The higher content of TBK in FF than BF infants throughout the entire 2 years in our study was surprising. According to the multicomponent model, the higher TBK was not associated with higher FFM throughout, primarily because TBW did not differ beyond 3 months. TBK was augmented in FF infants during the exclusive feeding period, and the differential between BF and FF infants remained fairly constant thereafter. During the exclusive feeding period, FF infants consumed approximately twice the quantity of potassium of BF infants. McCance and Widdowson59 observed that newborn piglets given sow's milk with twice as much potassium retained twice as much potassium as those given unsupplemented sow milk. The additional potassium made no difference in the percentage of water in tissues. The additional potassium retention was accounted for primarily by an increase in the concentration of potassium in skeletal muscle, with a minor increase in serum potassium.
Previous results from a cross-sectional study in which we used 18O dilution and TOBEC to estimate the body composition of BF and FF infants at 1 and 4 months of age18 confirm our present findings of lower FFM and higher FM in BF infants in early infancy. Expressed as a percentage of weight, TBW and FFM were lower, and FM was higher in the 4-month BF than FF infants. These results are in contrast to nonsignificant findings from another study, based on serial measurements of 18O in a small sample of BF and FF infants.17 Bellu16 also used TOBEC to measure the body composition of BF and FF at 12 months of age and reported lower body FM among the BF infants; however, these infants were significantly smaller at birth, and had lower lengths, weights and head circumferences at 12 months. In determining the effect of early infant nutrition on growth and body composition, it is necessary (but seldom done) to adjust for gender and initial body size and composition in the analysis. Lastly, de Bruin et al4 used TOBEC in a 12-month longitudinal study comparing 23 BF and 23 FF infants. Higher weight gains between 1 to 4 months in FF girls resulted in significantly higher FFM and FM, but not %FM, at 4 and 8 months of age.
Although lower nutrient intakes in BF than FF infants are clearly evident during the period of exclusive feeding,60 few comparative intake studies are available through the transitional and weaning periods. In a Dutch study, FF infants were shown to have higher macronutrient and gross energy intakes than BF infants during the 1 to 4 months period of exclusive feeding, but not at 8 and 12 months of age.4 In the DARLING study, FF infants had higher energy intakes than BF infants at 3, 6, 9 (girls only), and 12 months (girls only) and higher protein intakes at 3, 6, and 9 months (girls only).5 Energy intakes of FF infants were 15% to 20% higher than BF infants and persisted throughout the first year, primarily because of differences in energy from human milk or formula. In both studies, the higher nutrient intakes of FF infants promoted greater weight gains. In the DARLING study, energy and protein intakes at 3 months were significantly correlated with weight gain and FFM gain at 3 to 6 months (r = 0.30-0.35; P = .05). In the Dutch study, higher weight gains in the FF infants were associated with proportionately higher FFM and FM.
Ziegler has argued that the accretion of FFM of BF infants may be limited by their lower protein intake.61 In a small sample of 8 BF and 10 FF infants, weight gain and accretion of FFM based on deuterium dilution were significantly and proportionately higher in FF than BF infants between 42 to 84 days of age.62 In our study, differences in nutrient intake and weight gain were associated with disproportionate accretion of FFM and FM, resulting in differences in the %FM between feeding groups in early infancy. The multicomponent model indicated lower FFM, and higher FM and %FM in BF than FF infants at 3 months and 6 months (boys only). TOBEC confirmed and extended these findings to 9 months. FFM was lower in BF than FF infants at 3, 6, and 9 months. FM and %FM were higher in BF than FF infants at 3, 6, and 9 months (boys only). Our differences in body composition were observed during early infancy when human milk or formula was the primary source of nutrients, and when the discrepancy in nutrient intakes was observed between feeding groups. Although it seems plausible that the differences in body composition are a result of quantitative differences in energy and protein intake, high intercorrelations observed between nutrient intakes make it impossible to identify the causal factor (s). Furthermore, we cannot rule out the myriad of other qualitative and quantitative differences between human milk and formula, including insulinogenic amino acids, fatty acids, minerals, and growth factors, which may underlie differences in body composition between feeding groups.
Perhaps more importantly, the differences in body composition between BF and FF infants did not persist into the second year of life. Retrospective studies are conflicting with respect to the effect of breastfeeding on later adiposity. Protective, dose-responsive effects of breastfeeding on being overweight or obese were observed at 5 to 6 years of age in German children21 and at 12 to 18 years of age in Canadian children.22 In contrast, indicators of adiposity of formerly BF and FF infants were not significantly different later in childhood in a number of other studies.23-27 Follow-up anthropometric and body composition measurements are being made on our cohort at 5 years of age. Whether longer-term metabolic programming because of infant feeding mode occurs later in life remains unknown.
| |
ACKNOWLEDGMENTS |
|---|
This project was funded in part with federal funds from the USDA/ARS under Cooperative Agreement 58-6250-6001.
We wish to thank the women who participated in this study, and to acknowledge the contributions of Carolyn Heinz for participant coordination; Marilyn Navarrete for participant recruitment; Sopar Seributra and Sandra Kattner for nursing and dietary support; Maurice Puyau, Firoz Vohra, Judy Joo Posada, JoAnn Pratt, Nitesh Mehta, Zahira Colon, Kiyoko Usuki, Shide Zhang, and Deborah Roose for technical assistance; Anne Adolph for data management; Leslie Loddeke for editorial review; and Idelle Tapper for secretarial assistance.
We also are grateful to Ross Laboratories for their formula donations.
| |
FOOTNOTES |
|---|
This work is a publication of the US Department of Agriculture (USDA)/Agricultural Research Service (ARS) Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children's Hospital, Houston, Texas.
The contents of this publication do not necessarily reflect the views or policies of the USDA, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.
Received for publication Sept 30, 1999; accepted Mar 2, 2000.
Reprint requests to (N.F.B.) Children's Nutrition Research Center, 1100 Bates, Houston, TX 77030. E-mail: nbutte{at}bcm.tmc.edu
| |
ABBREVIATIONS |
|---|
BF, breastfed; FF, formula-fed; TBW, total body water; TBK, total body potassium; BMC, bone mineral content; DXA, dual energy x-ray absorptiometry; FFM, fat-free mass; FM, fat mass; TOBEC, total body electrical conductivity; CNRC, Children's Nutrition Research Center; NCHS/CDC, National Center for Health Statistics/Centers for Disease Control and Prevention; WHO, World Health Organization.
| |
REFERENCES |
|---|
|
Top
Abstract
MaterialsMethods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. K Saha, E. A Frongillo, D. S Alam, S. E Arifeen, L. A. Persson, and K. M Rasmussen Appropriate infant feeding practices result in better growth of infants and young children in rural Bangladesh Am. J. Clinical Nutrition, June 1, 2008; 87(6): 1852 - 1859. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. N Jordan and K. D Hall Dynamic coordination of macronutrient balance during infant growth: insights from a mathematical model Am. J. Clinical Nutrition, March 1, 2008; 87(3): 692 - 703. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Arsenault, D. Lopez de Romana, M. E. Penny, M. D. Van Loan, and K. H. Brown Additional Zinc Delivered in a Liquid Supplement, but Not in a Fortified Porridge, Increased Fat-Free Mass Accrual among Young Peruvian Children with Mild-to-Moderate Stunting J. Nutr., January 1, 2008; 138(1): 108 - 114. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Gunther, A. E Buyken, and A. Kroke Protein intake during the period of complementary feeding and early childhood and the association with body mass index and percentage body fat at 7 y of age Am. J. Clinical Nutrition, June 1, 2007; 85(6): 1626 - 1633. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Scholtens, U. Gehring, B. Brunekreef, H. A. Smit, J. C. de Jongste, M. Kerkhof, J. Gerritsen, and A. H. Wijga Breastfeeding, Weight Gain in Infancy, and Overweight at Seven Years of Age: The Prevention and Incidence of Asthma and Mite Allergy Birth Cohort Study Am. J. Epidemiol., April 15, 2007; 165(8): 919 - 926. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. N. Davis, M. J. Weigensberg, G. Q. Shaibi, N. C. Crespo, L. A. Kelly, C. J. Lane, and M. I. Goran Influence of Breastfeeding on Obesity and Type 2 Diabetes Risk Factors in Latino Youth With a Family History of Type 2 Diabetes Diabetes Care, April 1, 2007; 30(4): 784 - 789. [Abstract] [Full Text] [PDF]
|
|
