Human Nutrition and Metabolism
Complementary Feeding with Cow’s Milk Alters Sleeping Metabolic Rate in
Breast-Fed Infants1,2
Hinke Haisma,*†**3 Jonathan C. K. Wells,‡ W. Andrew Coward,†† Danton Duro Filho,*
Cesar G. Victora,* Roel J. Vonk,** Antony Wright,†† and G. Henk Visser†‡‡
ABSTRACT Although it is widely accepted that energy expenditure in infants is a function of feeding pattern, the
mechanism behind this is not well understood. The objectives of this observational study were as follows: 1) to
compare minimal observable energy expenditure (MOEE) between 2 subgroups of breast-fed infants, a BM group
in which breast milk was the only source of milk and a BCM group given cow’s milk in addition to breast milk; and
2) to identify potential mediators of a feeding pattern effect. For this purpose, infants were classified by feeding
group on the basis of a mother’s recall. Respiration calorimetry was used to measure MOEE in 62 infants (n ⫽ 35
BM, n ⫽ 27 BCM) aged 8.7 mo in Pelotas, southern Brazil. Breast-milk intake was measured using deuterium oxide,
complementary food intake by 1-d food weighing, total energy expenditure and total body water using doubly
labeled water; anthropometric indices were calculated. MOEE was 1672 ⫾ 175 kJ/d in BM compared with 1858
⫾ 210 kJ/d in BCM infants (P ⬍ 0.001). Mass-specific MOEE was 201 ⫾ 24.6 and 216 ⫾ 31.9 kJ/(kg 䡠 d) in BM and
BCM infants, respectively (P ⫽ 0.041). MOEE (kJ/d) was mediated by protein intake and fat-free mass (R2
⫽ 41.4%). We conclude that complementary feeding with cow’s milk alters the sleeping metabolic rate in
breast-fed infants. These findings deserve attention in relation to “metabolic programming” and the development
of obesity later in life. J. Nutr. 135: 1889 –1895, 2005.
KEY WORDS: ● minimal observable energy expenditure
● cow’s milk ● infants
Sleeping metabolic rate (SMR)4 and minimal observable
energy expenditure (MOEE) have been used as approximations of basal metabolic rate in infants. SMR refers to the
●
sleeping metabolic rate
●
breast milk
mean energy expenditure during a certain period of sleep, with
the length of this period varying among studies. MOEE is
defined as the mean energy expenditure during the lowest 5
consecutive minutes of a sleeping metabolic rate measurement. The latter is the more standardized entity, and has been
used as the basis of this work. In contrast to basal metabolic
rate (BMR), SMR and MOEE include the thermal effect of
food (TEF), but do not include the energy expended to stay
awake (the cost of arousal). In infants, SMR is ⬃60% of TEE,
and the metabolic intensity of the brain contributes to ⬃70%
of SMR (1). Schofield (2) compiled measurements of BMR
and developed prediction equations based on sex, age, and
weight; recently, a new set of prediction equations were developed for infants (3). The SMR of infants appears to be
dependent on gender, age, and weight, but ideally, prediction
equations should also include feeding mode. Several studies
found higher SMR in formula-fed compared with breast-fed
infants (4 – 6). A possible mechanism was suggested by Butte
et al. (7), who studied sleep organization in breast- and formula-fed infants and found a shorter duration of rapid eye
movement (REM) sleep in breast-fed infants, and a longer
percentage of time spent in non-REM sleep. Another possible
mechanism could be through the TEF. Infant formulas pre-
1
Presented in abstract form at the ESPGHAN precongress, Early Nutrition
and Its Late Consequences, 2–3 July 2004, Paris, France and at the 12th International Conference of the International Society for Research in Human Milk and
Lactation (ISRHML), September 10 –14, 2004, Queens’ College, Cambridge, UK
[Haisma, H., Wells, J.C.K., Coward, W. A., Duro Filho, D., Victora, C. G., Vonk,
R. J., Wright, A. & Visser, H. (2005) Effect of complementary feeding with
cows’ milk on sleeping metabolic rate in breast-fed infants. J. Paed. Gastroenterol. Nutr. (in press); and in Breastfeeding and health: early influences on later
health. In: Proceedings of the 12th International Conference of the Society for
Research on Human Milk and Lactation (Goldberg, G., Prentice, A. M., Prentice,
A., Filteau, S. & Simondon, K. B. (eds.). Springer Press, London, UK (in press)].
2
Supported by the International Atomic Energy Agency (RC10981/R2); Directorate General for International Cooperation, Netherlands Ministry of Foreign
Affairs.
3
To whom correspondence and reprint requests should be addressed.
E-mail: [email protected].
4
Abbreviations used: AEE, activity energy expenditure; BCM, infants receiving both breast and cow’s milk; BM, infants receiving breast milk as the only
source of milk; BMR, basal metabolic rate; FFM, fat-free mass; FFMI, fat-free
mass index; FM, fat mass; FMI, fat mass index; IGF, insulin-like growth factor;
MOEE, minimal observable energy expenditure; REM, rapid eye movement; SES,
socioeconomic status; SMR, sleeping metabolic rate; TBW, total body water;
TEE, total energy expenditure; TEF, thermal effect of food.
0022-3166/05 $8.00 © 2005 American Society for Nutritional Sciences.
Manuscript received 30 January 2005. Initial review completed 1 March 2005. Revision accepted 4 May 2005.
1889
Downloaded from jn.nutrition.org at MRC NUTRITION LIBRARY, CAMBRIDGE on September 5, 2006
*Universidade Federal de Pelotas, Departamento de Medicina Social, Pelotas, RS, Brazil; †Groningen
University, Zoological Laboratory, Haren, the Netherlands; **Groningen University, Laboratory of Paediatrics,
Groningen, the Netherlands; ‡Institute of Child Health, MRC Childhood Nutrition Research Group, London,
UK; ††MRC Human Nutrition Research, Elsie Widdowson Laboratory, Cambridge, CB1 9NL, UK; and
‡‡
Centre for Isotope Research, Groningen, the Netherlands
HAISMA ET AL.
1890
SUBJECTS AND METHODS
Study design. The study was conducted in Pelotas, a city in the
extreme south of Brazil (32° S and 52° W) with ⬃330,000 habitants
and 6000 births/y. Field work was conducted from October 2001 to
May 2002. Mean temperature ranged from 15.5 to 24.0°C, with
maximum temperatures ranging from 28.2 to 36.2°C. Relative humidity is high in Pelotas, and ranged from 79.9 to 89.6% (Centro de
Meteorologia, Universidade Federal de Pelotas). The study was designed as a cross-sectional study to assess food intake, growth, total
energy expenditure (TEE), SMR, and MOEE in 8-mo-old breast-fed
infants; 8 mo was chosen as the latest age at which the required
sample size of breast-fed infants was expected to be recruited within
the time frame of the study. All measurements were done at the
homes of the participating infants. The total length of the measurement period for each infant was 3 wk. Breast-milk intake measurements were conducted during the first 2 wk of the study, and intake
of complementary foods was measured in wk 2. TEE and SMR were
measured during wk 3. Details of the measurements are described
below.
Subjects and classification by feeding group. The study was
embedded in a larger study on the influence of SES on energy
requirements (13). Maternal education served as a proxy for SES. An
electronic database (SINASC) including all birth registrations in
Pelotas was used for the selection of mother-infant pairs. From the
database, a selection was made using the following criteria: 1) mother
living in urban Pelotas; 2) infant birth weight ⱖ 2500 g; 3) infant
gestational age between 37 and 41 wk; 4) single birth; 5) infant
healthy at birth; 6) maternal schooling ⱕ 3 y (low SES) or schooling
ⱖ 8 y (high SES). Mothers whose infants would be 8 mo old on a
week day (to allow data collection) were visited in their homes when
the infant was ⬃7 mo old. During this visit the data obtained from
the database were verified, and an additional inclusion criterion was
added, i.e., only infants who were still breast-fed were included.
Infants were subsequently classified into 2 different feeding categories (irrespective of whether they were from the low or high SES
groups): 1) breast-milk or BM group, i.e., infants given breast milk as
the only source of milk (with or without solids); 2) breast- and
cows’-milk or BCM group, i.e., infants given cow’s milk in addition to
breast milk (with solids). Classification was on the basis of a measure
of usual food intake using a food questionnaire applied on d 0 of the
study. This allowed a characterization of an infant’s feeding pattern
for a longer period of time. The questionnaire included questions on
the age at which certain foods had been introduced (for example,
cow’s milk, formula milk, fruits, meat, black beans).
SMR and MOEE. SMR was measured by respirometry using a
DeltatracTM MBM-100. The head and part of the body of the infant
were covered with a transparent plastic canopy; the adult mixing
chamber with an airflow of 40 L/min was used to avoid accumulation
of carbon dioxide in the canopy during the time of the measurement.
The use of this adult setup of the Deltatrac for infants was validated
by Wells (14). Oxygen consumption, carbon dioxide production, and
respiratory quotient were calculated by the Deltatrac software from
the constant air flow and the downstream gas concentrations; the
data were printed every minute. These data were subsequently entered into a computer and energy expenditure (J/min) was calculated
using Weir’s formula (15): (1.106 ⫻ VCO2 ⫹ 3.941 ⫻ VO2)
⫻ 4.186, where VCO2 is carbon dioxide produced (mL/min), and
VO2 is oxygen consumed (mL/min). SMR (kJ/min) was defined as the
mean of energy expenditure during the whole measurement period
(40 min up to 1 h), and MOEE (kJ/min) was assessed as the mean of
the 5 consecutive lowest 1-min values for energy expenditure. Measurements were made at a time the infant would usually sleep. This
could be any time of the day or night. Measurements made from 2200
to 0800 h were classified as night measurements. It was common for
the infant to be fed before the measurements, and then soothed to
sleep. The length of a sleep cycle was assessed for the first 10 infants
from the graph plotting energy expenditure against time. A cycle was
defined as the time between 2 peaks or 2 dips in energy expenditure,
whichever was first completed within a measurement period. This was
found to be ⬃40 min. Subsequent measurements were therefore
conducted for at least 40 min; if possible, however, measurements
were continued for 1 h. MOEE was considered to be the best standardized approximation of BMR, and analysis was based primarily on
MOEE.
Breast-milk intake and total daily energy expenditure. The
dose-to-the-mother 2H2O turnover method was used to measure
breast-milk intake but this was combined with the subsequent measurement of TEE using 2H218O. Details of the basic breast-milk
measurements were described elsewhere (16); briefly, the method
involves the administration of 0.5 mol/L (10 g) of 99.8% deuterium
to the mother, and collection of saliva samples from the mother
immediately before dose administration on d 0 (predose), and subsequently on d 1, 3, 13, and 14. Urine samples were collected from the
infant on d 0 (predose) and on d 1, 2, 3, 13, and 14. For the
measurement of TEE, an oral dose of 0.18 g/kg H218O and 0.10 g/kg
2
H2O was administered to the infant on d 14 shortly after the
collection of the d-14 sample for the breast-milk estimates. The dose
was slowly fed into the infant’s mouth using a nasogastric tube
attached to a syringe. Any spillage was collected using preweighed
tissues. The exact dose administered was calculated from the difference in weight of the dosing vial, syringe, nasogastric tube, and tissues
pre- and postdosing, and averaged 84% of the dose prepared. Subsequently urine samples were collected from the infant on d 15, 16, 17,
20, and 21. During the field work, samples were stored on ice, and
thereafter at ⫺20°C. Samples were shipped unfrozen to the laboratory
in Cambridge, UK for analysis. Equations used for the calculations of
breast-milk intake, total body water (TBW), and TEE are described
elsewhere (13).
Body composition. Fat-free mass (FFM) was calculated using a
hydration coefficient of 79.7% (17), and fat mass (FM) as the difference between body weight and FFM: FFM (kg) ⫽ TBW/0.797. FM
(kg) ⫽ body weight at d 14 –FFM. A fat-free mass index (FFMI), and
fat mass index (FMI) were calculated from FFM (kg)/height (m)2, and
FM (kg)/ height (m)2, respectively (18). Maternal body composition
was calculated from deuterium data only [Em(0)], as described elsewhere (16), using the same procedures except that the hydration of
FFM was assumed to be 73%.
Intake of complementary foods. Intake of complementary foods
was measured by food weighing using a mechanical scale that was
calibrated against standard weights. Field workers were trained and
monitored throughout the study. Weighing was started at the time of
the first meal of the day (other than breast milk) and continued until
after supper. A return visit was made the next day to measure any
leftovers from food eaten overnight. A Brazilian food composition
Downloaded from jn.nutrition.org at MRC NUTRITION LIBRARY, CAMBRIDGE on September 5, 2006
pared using cow’s milk often have a higher protein content
than breast milk, and protein contributes more to TEF than
lipids and carbohydrates. However, Butte et al. (6) measured
TEF as part of SMR measurements in breast- and formula-fed
infants, and found no difference. In line with the differences in
energy metabolism between formula- and breast-fed infants are
findings that growth patterns also differ between the 2 groups.
After an initially similar growth pattern, breast-fed infants
grow more slowly during infancy, and at 1 y of age formula-fed
infants weigh more (8,9). Evidence is increasing that the early
life period is important in relation to obesity and associated
diseases later in life; breast-feeding has been associated with a
reduced prevalence of obesity by some authors (10,11), although not by others (12).
The study described here was designed primarily to examine
the influence of socioeconomic status (SES) on the components of energy expenditure including MOEE in breast-fed
infants (13). The group of breast-fed infants was not homogenous in that some received breast and cow’s milk (BCM
group), whereas others received only breast milk (BM group).
This raised the question whether energy metabolism would be
influenced by this difference in feeding pattern. In line with
differences in SMR and MOEE between breast- and formulafed infants found by others, we studied the hypothesis that
MOEE would by higher in BCM than in BM infants.
SLEEPING METABOLIC RATE IN BREAST-FED INFANTS
variances were different for the groups studied). Factors considered as
possible confounders were maternal age, maternal nutritional status,
parity, SES, family income, years of schooling of both father and
mother, crowding, mother working, mother being away from the
child during the day, smoking habits, presence of tap water and
flushing toilet, sex of the child, ethnicity of the child, birth weight,
length at birth, health status, time of the SMR measurement (night,
day), duration of the measurement, and time between last feed and
start of the measurement. The criteria for confounders were those
described by Rothman and Greenland (24). None of the proposed
factors fulfilled these criteria. Special attention was given to SES
because the study had initially been designed to study differences in
components of energy expenditure between high- and low-SES
groups. However, because SES was not associated with either outcome (MOEE) or exposing variable (feeding group), and inclusion of
SES into a multivariate model including both feeding group and SES
did not reduce the difference between feeding groups by ⬎ 10%, SES
was not a confounder, and there was no need for adjustment of the
data by SES. Pearson’s correlation coefficient was used to study
univariate correlations between the major variables of interest and
MOEE (kJ/d). Those variables with a correlation with MOEE significant at the P ⫽ 0.10 level [means (95% CI)] were subsequently
entered into a covariance model, to study the extent to which the
effect of feeding pattern was mediated by these factors.
Ethics. The study was approved by the ethical committee of the
Universidade Federal de Pelotas, affiliated with the National Commission on Research Ethics of the Brazilian Ministry of Health, and
an informed consent was signed by the mother.
RESULTS
Subjects
In examining the characteristics of mother and infants
(Table 1), 35 infants were classified as BM and 27 infants as
BCM. The sex ratio and ethnicity of the infants did not differ
between the groups. Time since the last feeding and infant
nutritional status also did not differ between feeding groups.
However, in the larger sample (n ⫽ 77), which included those
infants in whom SMR measurements could not be made, FMI
differed between the feeding groups (BM, 4.5 ⫾ 1.5; BCM, 5.5
⫾ 1.7; P ⫽ 0.013).
TABLE 1
Characteristics of BM and BCM infants at 8.7 mo of age1,2
BM (n ⫽ 35)
Birth weight, kg
Length at birth, cm
Weight at 8.7 mo, kg
Length at 8 mo, cm
Weight gained from birth, kg
BMI, kg/m2
Head circumference at 8 mo, cm
FM, kg
FFM, kg
Fat at 8.7 mo, %
FMI, kg/m2
FFMI, kg/m2
Time since last feed, min
Sex ratio, male/female
Ethnicity, white/nonwhite
Health status, healthy/ill
SES, high/low
1
2
3.3 ⫾ 0.4
48.6 ⫾ 2.1
8.4 ⫾ 0.9
69.7 ⫾ 2.4
4.9 ⫾ 0.8
16.8 ⫾ 1.2
44.2 ⫾ 1.0
2.3 ⫾ 0.7
6.0 ⫾ 0.8
28.1 ⫾ 7.5
4.8 ⫾ 1.4
12.3 ⫾ 1.5
78.1 ⫾ 115
13/22
26/9
21/14
17/18
BCM (n ⫽ 27)
3.3 ⫾ 0.4
49.2 ⫾ 2.0
8.7 ⫾ 1.2
70.1 ⫾ 3.2
5.1 ⫾ 1.2
17.2 ⫾ 1.7
44.6 ⫾ 1.4
2.7 ⫾ 0.8
6.1 ⫾ 0.8
30.0 ⫾ 6.9
5.4 ⫾ 1.6
12.3 ⫾ 0.9
54.9 ⫾ 34.8
15/27
18/9
14/13
16/11
Values are means ⫾ SD or ratios.
Means did not differ between BM and BCM infants.
Downloaded from jn.nutrition.org at MRC NUTRITION LIBRARY, CAMBRIDGE on September 5, 2006
table (19) was used for calculation the energy, protein, lipid, and
carbohydrate content of the diet. Data on breast-milk composition
during the second half of infancy were obtained from the National
Academy of Sciences (20) as follows: energy content, 2721 kJ/L;
protein 12.1, g/L; fat, 40 g/L; and carbohydrate, 74 g/L.
The respiratory quotient, defined as the ratio of the production of
carbon dioxide and the consumption of oxygen, was estimated from
the calculation of the food quotient as described by Black et al. (21).
Thermal effect of food (TEF). Because SMR measurements were
performed in the postprandial state, energy expended as a result of the
TEF was included in the MOEE. In an attempt to disentangle the
contribution of TEF to MOEE, we estimated TEF on the basis of
macronutrient intake. The largest contribution to TEF was from
protein intake (22). The contribution of protein intake to the TEF
was calculated by dividing protein intake (g/d) by the factor 6.25 to
give nitrogen intake (g N/d). This value was then multiplied by 35.2
kJ/g N to provide an estimate of the energy released (kJ/d) (22).
Anthropometry. The infants were weighed without clothes using
a portable electronic UNICEF scale accurate to 0.1 kg. Length was
measured using a standardized anthropometer (AHRTAG babylength measures). Mothers were weighed without shoes but with
clothes using the same UNICEF scale, and maternal weight was
calculated as the difference between the weight with clothes and the
weight of clothes. Maternal height was measured to the nearest mm
using a locally developed portable stadiometer. Maternal BMI was
calculated from weight (kg)/height (m)2.
Morbidity. Twice-weekly morbidity questionnaires were applied.
The number of days the child presented with diarrhea, cough, runny
nose, or fever was registered during each of those visits to constitute
the total number of days on which these symptoms were present over
the total 3 wk of the study. Diarrhea was defined as ⱖ5 liquid stools/d.
A child was classified as being healthy if during the week of SMR
measurements ⱕ1 of the above-mentioned symptoms was present for
ⱕ1 d. Measurements were postponed if acute illness (fever) was
present on the day the measurement had been scheduled. In this case,
measurements took place within 1 wk of the day scheduled.
Sample size. Studies on the SMR of 9-mo-old infants (breastand bottle-fed combined) showed values of 218 kJ/(kg 䡠 d) (23) and
239 ⫾ 9.5 kJ/(kg 䡠 d) (3). At the time this study was planned, no data
existed on the SMR of 8-mo-old breast-fed infants. The only data
available on breast-fed infants were from Butte et al. (6) in 4-mo-old
infants. The SMR was 199 ⫾ 18.4 kJ/(kg 䡠 d), mean ⫾ SD. This SD
was used for sample size calculations. For an 8% difference in SMR
between infants from the high and low SES groups to be significant,
the study required a minimum of 20 infants in each group. These
calculations assumed a Type I error (␣) of 5%, two-tailed, and a Type
II error (␤) of 10%, that is, a statistical power of 90%.
Because the study was embedded in a larger study on energy
requirements in infants from high- and low-SES groups (with a
sample size of 77 infants), and there was some uncertainty on the SD
of SMR measurements in infants, all 77 participating infants were
included. SMR measurements were successful in 62 infants (35 BM,
27 BCM). In 5 cases, the indirect calorimeter was not available at the
time scheduled for the study. In another 4 cases, a maximum of 5
attempts was reached without a successful measurement (typical
reasons were that the child had no regular sleeping hours or slept only
while suckling at the mother’s breast). For a further 6 subjects, the
mother did not collaborate in scheduling a visit to her home to do the
measurement. Food and macronutrient intake was assessed in 58 of 62
infants. From 2 infants, breast-milk intake data were not available, 1
mother refused to take part in this specific component of the study,
and a 4th infant was ill during the time of the food assessments and
did not recover within the time frame required (i.e., within 1 wk from
the end of the study period). TEE and TBW was available from 55 of
62 infants. To maximize the power of the comparison of interest, i.e.,
SMR in 2 different feeding groups, we chose to include all 62 infants
in the comparisons of SMR. Univariate correlations of food intake or
body composition and SMR were analyzed using n ⫽ 58 and n ⫽ 55,
respectively. A complete dataset was available from 52 infants, and
this number was used in the regression analysis.
Statistical methods. Differences between feeding groups were
investigated using Student’s t tests or a Mann-Whitney Test (if
1891
HAISMA ET AL.
1892
Food and nutrient intake
TABLE 3
Sleeping metabolic rate
Differences by feeding group. In summarizing MOEE and
SMR by feeding group (Table 3), MOEE (kJ/d) was 1672
⫾ 175 kJ/d in the BM group vs. 1858 ⫾ 210 kJ/d in the BCM
group (P ⬍ 0.001). The regression coefficient of log MOEE
(kJ/d) to log weight (kg) was 0.347 (SE 0.111, P ⫽ 0.001).
This indicates that the effect of weight on MOEE was removed
(normalized) by dividing MOEE by weight raised to the power
0.347. Mass-independent MOEE also differed between groups
(P ⬍ 0.001). Equally, the regression on log MOEE (kJ/d) to log
FFM (kg) resulted in a coefficient of 0.380 (SE 0.115, P
⫽ 0.002), and MOEE normalized for FFM0.380 was 850 ⫾ 83.7
kJ/(kg FFM0.380 䡠 d) in BM infants compared with 942 ⫾ 96.3
kJ/(kg FFM0.380 䡠 d) in BCM infants (P ⫽ 0.001). The regression lines of log MOEE on log FFM by feeding group are shown
in Figure 1.
Analyses were repeated to include only those infants for
whom the duration of a measurement was ⬎1 h. This reduced
the sample size to 9 BM and 9 BCM infants; as a result, the
power was reduced. Nevertheless, the differences remained
borderline significant (P ⬍ 0.1) for MOEE (kJ/d) and MOEE
[kJ/(kg 䡠 d)], and significant (P ⬍ 0.05) for MOEE [kJ/(kg
FFM 䡠 d)], and MOEE [kJ/(kg0.347 䡠 d)].
Further analysis was done to study the effect of the criteria
Components of energy expenditure of BM and BCM infants1
Metabolic component
MOEE
kJ/d
kJ/(kg 䡠 d)
kJ/(kg FFM 䡠 d)
kJ/(kg0.347 䡠 d)
SMR
kJ/d
kJ/(kg 䡠 d)
kJ/(kg FFM 䡠 d)
kJ/(kg0.313 䡠 d)
TEE2
kJ/d
kJ/(kg 䡠 d)
kJ/(kg FFM 䡠 d)
kJ/(kg0.192 䡠 d)
AEE2
kJ/d
kJ/(kg 䡠 d)
kJ/(kg FFM 䡠 d)
1
2
BM (n ⫽ 35)
BCM (n ⫽ 27)
P-value
1672 (1607–1736) 1858 (1785–1932)
201 (192–211)
216 (205–227)
284 (270–298)
309 (293–324)
801 (772–829)
876 (846–911)
0.001
0.041
0.020
0.001
1908 (1846–1971) 2045 (1974–2116)
230 (220–239)
238 (227–249)
325 (311–340)
339 (323–355)
982 (953–1011) 1041 (1007–1074)
0.005
0.263
0.198
0.001
2315 (2094–2536) 2537 (2290–2784)
285 (257–313)
294 (262–325)
390 (350–430)
421 (377–466)
1547 (1402–1692) 1674 (1512–1836)
0.186
0.688
0.290
0.247
634 (401–867)
82.6 (54.5–110)
106 (66.9–146)
0.795
0.875
0.827
679 (419–940)
79.3 (47.8–111)
113 (69.4–156)
Values are means (95% CI).
Values are adjusted for ethnicity.
by which the infants were classified. As described in the
Subjects and Methods section, infants were classified on d 0 of
the study period on the basis of a mother’s recall. If infants
were classified on the basis of their actual intake of cow’s milk
on the day of food weighing (BM: n ⫽ 21; BCM: n ⫽ 34) the
difference remained significant (P ⬍ 0.01) for MOEE expressed as kJ/d, kJ/(kg 䡠 d), kJ/(kg FFM 䡠 d), and kJ/(kg0.347 䡠 d).
Similarly, if infants were classified on the basis of the recall on
d 0 of the study, and subsequently those infants in the BM
TABLE 2
Food and nutrient intake of BM and BCM infants1
Breast milk volume, mL/d
Cow’s milk volume, mL/d
Intake of yoghurt, g/d
Age until which EBF,2 mo
Energy intake, kJ/d
Energy from breast milk, %
Energy from cow’s milk, %
Energy from solids, %
Protein intake, g/d
Fat intake, g/d
Carbohydrate intake, g/d
Food quotient
BM (n ⫽ 32)
BCM (n ⫽ 26)
P-value
761 ⫾ 231
25 ⫾ 60
24 ⫾ 49
3.0 ⫾ 1.9
3160 ⫾ 800
69 ⫾ 22
1.6 ⫾ 4.0
29 ⫾ 21
17.1 ⫾ 5.7
34.7 ⫾ 9.6
104.9 ⫾ 32.6
0.86 ⫾ 0.02
464 ⫾ 352
232 ⫾ 227
33 ⫾ 58
2.3 ⫾ 2.0
3127 ⫾ 741
42 ⫾ 32
18 ⫾ 20
38 ⫾ 22
23.7 ⫾ 9.5
29.8 ⫾ 10.3
104.8 ⫾ 25.3
0.87 ⫾ 0.02
0.001
0.001
0.490
0.169
0.878
0.001
0.001
0.133
0.002
0.065
0.986
0.077
Values are means ⫾ SD.
EBF, exclusively breast-fed, i.e., receiving nothing but breast milk,
not even water.
1
2
FIGURE 1 Association of log FFM and log MOEE by feeding
pattern in infants fed breast milk as the exclusive source (BM) and in
those also given cow’s milk (BCM). For BM infants: log MOEE (kJ/d)
⫽ 2.952 ⫹ 0.350 ⫻ log FFM (kg); for BCM infants: log MOEE (kJ/d)
⫽ 3.005 ⫹ 0.336 ⫻ log FFM (kg).
Downloaded from jn.nutrition.org at MRC NUTRITION LIBRARY, CAMBRIDGE on September 5, 2006
In examining food and nutrient intake by feeding group
(Table 2), breast-milk intake was higher in the BM than in
the BCM group (761 ⫾ 231 vs. 464 ⫾ 352 mL/d, respectively,
P ⫽ 0.001). The infants classified as BM consumed 25 ⫾ 60
mL/d of cow’s milk on the day of food weighing. In the BCM
group, 5 infants had an intake of cow’s milk of 0 mL/d on the
day that food weighing took place. In the BM group, 3 of 32
infants were reported to be exclusively breast-fed (i.e., receiving nothing but breast milk) at the time the infant was 8 mo
old. Of the 26 infants in the BCM group, 18 were reported to
be receiving cow’s milk, 3 were receiving formula, and 5 were
receiving both cow’s milk and formula. On the day of food
weighing, only 1 infant was receiving formula. All BCM
infants were receiving solids. Energy (kJ/d), fat (g/d), and
carbohydrate intake (g/d) did not differ between the groups,
but protein intake (g/d) differed. Energy and macronutrient
intakes were not related to weight, length, or FFM.
SLEEPING METABOLIC RATE IN BREAST-FED INFANTS
1893
TABLE 4
Univariate correlation coefficients between MOEE and its potential determinants
Protein intake, g/d
FFM, kg
Weight, kg
Percentage cow’s milk
Cow’s milk intake, mL/d
Breast milk intake, mL/d
Length, cm
1
Protein
kJ/d
g/d
0.4872
0.4042
0.3912
0.3792
0.3312
⫺0.2971
0.2881
—
0.216
0.155
0.6052
0.6422
⫺0.3542
0.162
FFM
Weight
kg
Cow’s milk
Cow’s milk
%
Breast milk
mL/d
—
—
—
—
—
0.7022
⫺0.031
⫺0.048
0.069
0.6492
—
⫺0.028
⫺0.008
0.050
0.7352
—
—
0.9332
⫺0.8062
⫺0.011
—
—
—
⫺0.7092
0.035
—
—
—
—
⫺0.057
P ⬍ 0.05.
P ⬍ 0.01.
group who were receiving some cow’s milk on the day of the
food weighing and those infants in the BCM group who were
not receiving any cow’s milk on that day were excluded
(included BM: n ⫽ 23; BCM: n ⫽ 22), the difference in
MOEE remained significant (P ⬍ 0.005) when expressed as
kJ/d, kJ/(kg FFM 䡠 d), and kJ/(kg0.347 䡠 d). The difference became borderline significant (P ⫽ 0.070) for MOEE [kJ/
(kg 䡠 d)].
MOEE was 0.88 ⫻ SMR in the BM group compared with
0.91 ⫻ SMR in the BCM group (P ⫽ 0.006). Comparisons
between feeding groups based on SMR gave similar findings.
SMR was normalized for weight to the power 0.313. Differences in SMR (kJ/d, and kJ/kg0.313/d) between BM and BCM
infants were significant. However, SMR [kJ/(kg 䡠 d)] did not
differ between feeding groups (see Table 3). In addition, TEE
and AEE (Table 3) did not differ between feeding groups.
Univariate correlations. Associations of MOEE (kJ/d)
and major variables of interest (discussed below) were approximately linear, and univariate correlation coefficients are presented in Table 4. Energy (kJ/d), fat (g/d), and carbohydrate
intake (g/d) were not correlated with MOEE at a level of
significance ⬍ 0.10, and are not included in the table. In order
of significance, MOEE (kJ/d) was positively correlated with
protein intake (g/d), FFM (kg), weight (kg), percentage of
cow’s milk intake of total milk, intake of cow’s milk (mL/d),
and length (cm). MOEE (kJ/d) was negatively correlated with
breast-milk intake (mL/d). MOEE [kJ/d, kJ/(kg 䡠 d) and kJ/
(kg0.347 䡠 d)] was not associated with sex or ethnicity. Equally,
health status was not associated with MOEE [kJ/d, kJ/
(kg0.347 䡠 d)]. But MOEE [kJ/(kg 䡠 d)] was higher in ill compared with healthy children [ill (n ⫽ 27), 217 ⫾ 28.4 kJ/
(kg 䡠 d); healthy (n ⫽ 35) 201 ⫾ 27.4 kJ/(kg 䡠 d), P ⫽ 0.031].
BM and BCM infants did not differ in the number of days on
which diarrhea (P ⫽ 0.984), runny nose (P ⫽ 0.290), cough (P
⫽ 0.669), or fever (P ⫽ 0.879) were present. MOEE [kJ/d (P
⫽ 0.873), kJ/(kg 䡠 d) (P ⫽ 0.098), kJ/(kg0.347 䡠 d) (P ⫽ 0.407)]
did not differ between night- (n ⫽ 13) and daytime (n ⫽ 49)
measurements.
Analysis of covariance. Analysis of covariance was performed with feeding group as a fixed factor and the variables
mentioned in Table 4 entered as covariates (Table 5). MOEE
(kJ/d) was the independent variable in the model. Variables
with a Pearson correlation coefficient significant at the 0.10
level were entered as dependent variables. These included
feeding group (BM ⫽ 0 or BCM ⫽ 1), protein intake (g/d),
FFM (kg), weight (kg), percentage cow’s milk of total milk
intake, intake of nonbreast milk (kg), breast-milk intake (kg),
and infant’s length (cm).
The effect of feeding mode on MOEE was mediated
through both protein and FFM. Weight had a similar but
smaller effect as FFM. Adjustment of MOEE for protein intake
decreased the difference between feeding groups from ⫺186 to
⫺116 (95% CI: ⫺219, ⫺13.2) kJ/d (BM, 1681 ⫾ 169 kJ/d;
BCM, 1859 ⫾ 214 kJ/d; P ⫽ 0.028). Adjusting MOEE for FFM
decreased the difference from ⫺186 (95% CI: ⫺284, ⫺88.8)
kJ/d to ⫺177 (95% CI: ⫺277, ⫺78,7) kJ/d (BM, 1675 ⫾ 182
kJ/d; BCM, 1867 ⫾ 214 kJ/d; P ⫽ 0.001). MOEE (kJ/d) was
best explained if both protein and FFM were entered into the
model (see Table 5), with R2 ⫽ 41.4%. The difference between feeding groups remained significant (P ⫽ 0.021). When
SMR instead of MOEE was used as the outcome variable, the
results were very similar, with the difference in SMR again
best explained by protein intake and FFM (R2 ⫽ 32.1%).
Contribution of protein intake to TEF
The difference in protein intake between BM and BCM
infants was 6.6 g/d (higher values for the BCM infants). The
contribution of this difference in protein intake to the TEF
was 37.1 kJ/d, or 20% of the crude difference (186 kJ/d) in
MOEE (kJ/d).
DISCUSSION
The most striking finding was that within a group of breastfed infants, energy metabolism was increased if cow’s milk was
also given. Butte et al. (6) found higher values for formula-fed
infants compared with breast-fed infants, but no other study
has investigated SMR and MOEE within subgroups of breastfed infants.
The study was part of an observational study on energy
requirements of infants from groups with high and low SES,
TABLE 5
The effect of protein intake and FFM on the difference in
MOEE between BM and BCM infants
Model
Constant
Feeding group
Constant
Feeding group
Protein, g/d
FFM, kg
BM-BCM
95% CI
P-value
R2
1858
⫺186
1136
⫺124
8.4
86.6
1784, 1932
⫺284, ⫺88.8
747, 1525
⫺229, ⫺19.2
1.0, 14.9
23.6, 149.6
0.001
0.001
0.001
0.021
0.012
0.008
0.195
0.414
Downloaded from jn.nutrition.org at MRC NUTRITION LIBRARY, CAMBRIDGE on September 5, 2006
2
MOEE
1894
HAISMA ET AL.
a signaling function for these compounds as part of the parentoffspring conflict.
Comparison with prediction equations based on UK infants
(3) showed that our values were on average 93% of those
predicted. The difference was higher for BM compared with
BCM infants (P ⫽ 0.024), possibly reflecting the feeding
pattern of the UK infants. Furthermore, it was demonstrated in
birds (34) and small mammals (35) that SMR tends to be
lower at lower latitudes. Such a trend cannot be excluded in
metabolic patterns of infants, and the observation that MOEE
in Brazilian infants is lower than in the UK would agree with
this finding. Recently, a 5% reduction in energy expenditure
(TEF and SMR) was described as a result of higher temperature (22 vs. 16°C) (36), which is of the same order of magnitude as the difference in SMR between Brazilian and UK
infants (7%).
In conclusion, complementary feeding with cow’s milk appears to increase MOEE and SMR in breast-fed infants. The
effect occurs in part through a higher protein intake in BCM
infants, but even after adjusting for protein intake, the effect
of feeding group remained significant; either a bioactive factor
in cow’s milk is responsible for the higher SMR in BCM
infants, or alternatively, some factor associated with breast
milk or breast-feeding keeps SMR low in BM infants. The
higher FMI in BCM infants in the larger sample of 77 infants
suggests that the consumption of cow’s milk not only has an
effect on SMR, but also on body composition. This may be a
matter of some concern in relation to the “metabolic programming” theory (37) and the development of obesity later in life.
ACKNOWLEDGMENTS
We thank Luis Fernando Barros for his help with the custom
procedures involved in transport of the Deltatrac from the UK to
Brazil and Mara Santos for her help in administering the project.
LITERATURE CITED
1. Holliday, M. A. (1971) Metabolic rate and organ size during growth
from infancy to maturity and during late gestation and early infancy. Pediatrics 47
(suppl. 2):169 –179.
2. Schofield, W. N. (1985) Predicting basal metabolic rate, new standards and review of previous work. Hum. Nutr. Clin. Nutr. 39 (suppl. 1): 5– 41.
3. Wells, J. C., Joughin, C., Crisp, J. A., Cole, T. J. & Davies, P. S. (1996)
Comparison of measured sleeping metabolic rate and predicted basal metabolic
rate in the first year of life. Acta Paediatr. 85: 1013–1018.
4. Wells, J. C. & Davies, P. S. (1995) Sleeping metabolic rate and body
size in 12-week-old infants. Eur. J. Clin. Nutr. 49: 323–328.
5. Butte, N. F., Wong, W. W., Hopkinson, J. M., Heinz, C. J., Mehta, N. R. &
Smith, E. O. (2000) Energy requirements derived from total energy expenditure
and energy deposition during the first 2 y of life. Am. J. Clin. Nutr. 72: 1558 –1569.
6. Butte, N. F., Wong, W. W., Ferlic, L., Smith, E. O., Klein, P. D. & Garza, C.
(1990) Energy expenditure and deposition of breast-fed and formula-fed infants
during early infancy. Pediatr. Res. 28: 631– 640.
7. Butte, N. F., Jensen, C. L., Moon, J. K., Glaze, D. G. & Frost, J. D., Jr.
(1992) Sleep organization and energy expenditure of breast-fed and formula-fed
infants. Pediatr. Res. 32: 514 –519.
8. Dewey, K. G., Peerson, J. M., Brown, K. H., Krebs, N. F., Michaelsen,
K. F., Persson, L. A., Salmenpera, L., Whitehead, R. G. & Yeung, D. L. (1995)
Growth of breast-fed infants deviates from current reference data: a pooled
analysis of US, Canadian, and European data sets. World Health Organization
Working Group on Infant Growth. Pediatrics 96: 495–503.
9. World Health Organization (1994) Working Group on Infant Growth. An
Evaluation of Infant Growth. WHO/NUT/94.8. WHO, Geneva, Switzerland.
10. Arenz, S., Ruckerl, R., Koletzko, B. & von Kries, R. (2004) Breastfeeding and childhood obesity—a systematic review. Int. J. Obes. Relat. Metab.
Disord. 28: 1247–1256.
11. Grummer-Strawn, L. M. & Mei, Z. (2004) Does breastfeeding protect
against pediatric overweight? Analysis of longitudinal data from the Centers for
Disease Control and Prevention Pediatric Nutrition Surveillance System. Pediatrics 113: e81– e86.
12. Victora, C. G., Barros, F., Lima, R. C., Horta, B. L. & Wells, J. (2003)
Anthropometry and body composition of 18 year old men according to duration
of breast feeding: birth cohort study from Brazil. Br. Med. J. 327: 901.
13. Haisma, H. (2004) Energy Utilisation of Infants in Southern Brazil.
Downloaded from jn.nutrition.org at MRC NUTRITION LIBRARY, CAMBRIDGE on September 5, 2006
and the classification into BM or BCM infants was made post
hoc. Nevertheless, SES and the percentage of mothers working
outside the home did not differ between the 2 feeding groups,
and there were no other confounding factors that we are aware
of that could possibly have biased the results.
Analysis was done on the basis of MOEE for 2 reasons: 1)
the lower value of MOEE was considered a better estimate of
BMR as infants were fed and measurements include TEF; 2)
the external validity of our MOEE measurements was better
than the SMR measurements because of the shorter protocol
used for measurement of SMR (40 min vs. a minimum of 1 h
used by others). However, analysis of SMR by feeding group
gave similar results. The MOEE:SMR ratio was similar to the
0.89 observed by others (4,6).
A covariance model showed that the effect of feeding group
on MOEE was mediated by protein intake and body weight.
Explained variance of the model was 43.1%, suggesting that
either additional mediators were involved in the association
between MOEE and feeding group, or that the mediators
included were not measured properly, and that this explained
why all of the feeding group effect could not be eliminated.
Measurements of body weight were accurate, but it is more
difficult to obtain accurate food intake data. Food intake was
not measured on the same day as the SMR measurements, and
day-to-day variation of food intake in infants already receiving
solids was found to be ⬃12% in British infants (25), and 25%
in Asian infants (26). Residual mediation from protein should
therefore be considered.
Protein intake is the most important determinant of TEF
(22); theoretically, at least part of the protein effect on MOEE
could be explained by its contribution to TEF. However, Butte
et al. (6) actually measured TEF in breast- and formula-fed
infants, and found no difference. We calculated that 20% of
the difference in MOEE between feeding groups could be
attributed to TEF. High protein intake in early life was suggested to influence later health outcomes. Rolland-Cachera et
al. (27) suggested that the increase in insulin-like growth
factor (IGF)-I would influence maturation and trigger adipocyte multiplication, and they observed an association between
high protein intake at the age of 2 y and a higher incidence of
obesity at the age of 8 y. Similar results were obtained by
Scaglioni et al. (28). Nielsen et al. (29) found a more rapid
weight gain between 5 and 10 mo among the 10% of infants
with the highest protein content in their diet (ⱖ16 protein
energy %). The role of IGF-I was further studied by Hoppe et
al. They found that cow’s milk, but not meat, had a stimulating effect on serum IGF-I (30,31). This suggests that some
factor in cow’s milk other than protein could have been
responsible for the metabolic differences between BM and
BCM infants.
An alternative explanation for the effect of feeding group
may be through breast milk or the process of breast-feeding.
Because all infants in our study were breast-fed, any influence
of the feeding process itself would have to be related to time
spent at the breast, or the volume of breast-milk intake.
Breast-milk intake was negatively correlated with MOEE, but
did not mediate the feeding group effect. Time spent at the
breast was not measured in this study, but one would expect it
to be related to the volume of breast-milk intake. Nevertheless, some mothers allowed their infants to suckle for a long
time after the real feeding process had stopped, and we cannot
exclude the possibility of MOEE being reduced by time suckling. Apart from breast milk volume, the composition of breast
milk compared with cow’s milk could also have contributed to
the lower MOEE in BM infants. Breast milk contains benzodiazepine-like sedative compounds (32). Wells (33) suggested
SLEEPING METABOLIC RATE IN BREAST-FED INFANTS
26. Harbottle, L. & Duggan, M. B. (1994) Daily variation in food and
nutrient intakes of Asian children in Sheffield. Eur. J. Clin. Nutr. 48: 66 –70.
27. Rolland-Cachera, M. F., Deheeger, M., Akrout, M. & Bellisle, F. (1995)
Influence of macronutrients on adiposity development at follow up study of
nutrition and growth from 10 months to 8 years of age. Int. J. Obes. 19: 573–578.
28. Scaglioni, S., Agostoni, C., De Notaris, R., Radaelli, G., Radice, N.,
Valenti, M., Giovannini, M. & Riva, E. (2000) Early macronutrient intake and
overweight at five years of age. Int. J. Obes. 24: 777–781.
29. Nielsen, G. A., Thomsen, B. L. & Michaelsen, K. F. (1998) Influence of
breastfeeding and complementary food on growth between 5 and 10 months.
Acta Paediatr. 87: 911–917.
30. Hoppe, C., Udam, T. R., Lauritzen, L., Molgaard, C., Juul, A. &
Michaelsen, K. F. (2004) Animal protein intake, serum insulin-like growth factor
I, and growth in 2.5-y-old Danish children. Am. J. Clin. Nutr. 80 (2): 447– 452.
31. Hoppe, C., Molgaard, C., Vaag, A., Barkholt, V. & Michaelsen, K. F.
(2005) High intakes of milk, but not meat, increase s-insulin and insulin resistance in 8-year old boys. Eur. J. Clin. Nutr. 59 (3): 393–398.
32. Dencker, S. J., Johansson, G. & Milsom, I. (1992) Quantification of
naturally occurring benzodiazepine-like substances in human breast milk. Physchopharmacology 107: 69 –72.
33. Wells, J.C.K. (2003) Parent-offspring conflict theory, costly begging
behavior, signaling of nutritional need, and differential weight gain in early life. Q.
Rev. Biol. 78: 169 –198.
34. Weathers, W. W. (1979) Climatic adaptation in avian standard metabolic rate. Oecologia 42: 81– 89.
35. Lovegrove, B. G. (2003) The influence of climate on the basal metabolic rate of small mammals: a slow-fast metabolic continuum. J. Comp. Physiol.
B 173: 87–112.
36. Westerterp-Plantenga, M. S., van Marken Lichtenbelt, W. D., Strobbe, H.
& Schrauwen, P. (2002) Energy metabolism in humans at a lowered ambient
temperature. Eur. J. Clin. Nutr. 56: 288 –296.
37. Barker, D.J.P., ed. (1992) Fetal and Infant Origins of Adult Disease.
BMJ Publishing, London, UK.
Downloaded from jn.nutrition.org at MRC NUTRITION LIBRARY, CAMBRIDGE on September 5, 2006
Breast-Feeding Pattern and Socio-Economic Status in Relation to Obesity. Paediatric Laboratory, Section of Nutrition and Metabolism. University of Groningen,
Groningen, the Netherlands. http://www.ub.rug.nl/eldoc/dis/medicine/h.h.haisma/ [last accessed April 6, 2004].
14. Wells, J.C.K. (1994) Energy Metabolism in Breast-Fed and FormulaFed Infants. Cambridge University, Cambridge, UK.
15. Weir, J.B.d.V. (1949) New methods for calculating metabolic rate with
special reference to protein metabolism. J. Physiol. 109: 1–9.
16. Haisma, H., Coward, W. A., Albernaz, E., Visser, G. H., Wells, J.C.K.,
Wright, A. & Victora, C. G. (2003) Breast milk and energy intake in exclusively,
predominantly and partially breast-fed infants. Eur. J. Clin. Nutr. 57: 1633–1642.
17. Butte, N. F., Wong, W. W., Hopkinson, J. M., Smith, E. O. & Ellis, K. J.
(2000) Infant feeding mode affects early growth and body composition. Pediatrics 106: 1355–1366.
18. Van Itallie, T. B., Yang, M. U., Heymsfield, S. B., Funk, R. C. & Boileau,
R. A. (1990) Height-normalized indices of the body’s fat-free mass and fat
mass: Potentially useful indicators of nutritional status. Am. J. Clin. Nutr. 52:
953–959.
19. IBGE (1981) Estudo Nacional da Despesa Familiar-ENDEF-Tabelas de
Composição de Alimentos. Rio de Janeiro, Brazil.
20. Food and Nutrition Board (FNB), Institute of Medicine (2002) Dietary
Reference Intakes for Energy, Carbohydrates, Fiber, Fat, Protein and Amino Acids
(Macronutrients). National Academies Press, Washington, D.C.
21. Black, A. E., Prentice, A. M. & Coward, W. A. (1986) Use of food
quotients to predict respiratory quotients for the doubly-labelled water method of
measuring energy expenditure. Hum. Nutr. Clin. Nutr. 40: 381–391.
22. Kleiber, M. (1961) Calorigenic effect of food. In: The Fire of Life. John
Wiley & Sons, New York, NY.
23. Wells, J. C. & Davies, P. S. (1998) Estimation of the energy cost of
physical activity in infancy. Arch. Dis. Child. 78: 131–136.
24. Rothman, K. J. & Greenland, S. (1998) Modern Epidemiology. Lippincott Williams & Wilkins, Philadelphia, PA.
25. Black, A. E., Cole, T. J., Wiles, S. J. & White, F. (1983) Daily variation
in food intake of infants from 2 to 18 months. Hum. Nutr. Appl. Nutr. 37: 448 – 458.
1895