Randomized controlled trial of 4 compared with 6 mo of
exclusive breastfeeding in Iceland: differences in breast-milk
intake by stable-isotope probe1–3
ABSTRACT
Background: The WHO recommends exclusive breastfeeding
(EBF) for 6 mo after birth. However, the time at which breast milk
ceases to provide adequate energy and nutrition, requiring the introduction of complementary foods, remains unclear. Most studies
that investigated this issue were observational and potentially confounded by variability in social circumstances or infant growth.
Objective: We hypothesized that EBF infants would consume more
breast milk at age 6 mo than infants receiving breast milk and complementary foods.
Design: We measured anthropometric outcomes, body composition,
and breast-milk intake at age 6 mo in infants who were randomly
assigned at age 4 mo either to 6-mo EBF or to the introduction of
complementary foods with continued breastfeeding. We recruited
119 infants from health centers in Reykjavik and neighboring municipalities in Iceland. In 100 infants who completed the protocol
(50/group), breast-milk intake was measured by using stable isotopes, and complementary food intakes were weighed over 3 d in
the complementary feeding (CF) group.
Results: Breast-milk intake was 83 g/d (95% CI: 19, 148 g/d)
greater in EBF (mean 6 SD: 901 6 158 g/d) than in CF (818 6
166 g/d) infants and was equivalent to 56 kcal/d; CF infants obtained 63 6 52 kcal/d from complementary foods. Estimated total
energy intakes were similar (EBF: 560 6 98 kcal/d; CF: 571 6 97
kcal/d). Secondary outcomes (anthropometric outcomes, body composition) did not differ significantly between groups.
Conclusions: On a group basis, EBF to age 6 mo did not compromise infant growth or body composition, and energy intake at age
6 mo was comparable to that in CF infants whose energy intake was
not constrained by maternal breast-milk output.
Am J Clin Nutr
2012;96:73–9.
INTRODUCTION
Since 2001 the WHO has recommended exclusive breastfeeding (EBF) for the first 6 mo of life in all populations (1).
These recommendations were officially adopted by the governments of some but not all industrialized populations, including
the United Kingdom in 2003 (2, 3). In the developing world,
where the benefits of EBF to reduce diarrheal disease are paramount, the recommendations have been widely adopted.
Despite widespread official endorsement, the scientific basis of
the recommendations has been questioned, attributable in part to
limited evidence on when EBF ceases to provide adequate energy
and nutrition, and hence when complementary foods should be
introduced alongside breast milk. According to published data on
breast milk (intake, energy content) and infant energy requirements,
average intakes would fail to meet energy requirements in the
average 6-mo-old infant (4), although this approach relied on various
assumptions about breast-milk energy content. However, insufficient
relevant evidence is available, particularly in EBF infants. The
appropriate time for introducing complementary foods therefore
remains uncertain.
A major limitation is the scarcity of data from randomized
trials (5, 6). Mothers who breastfeed their infants at all differ from
those who do not breastfeed, and those who undertake 6 mo EBF
differ from those who breastfeed for shorter periods (7–11).
Fewer than 1% of UK mothers were practicing EBF up to age 6
mo in 2005 (12), and those who did were highly educated and
well financially supported during this period (13). Reported
phenotypic differences between EBF and non-EBF infants might
arise either from different nutritional intakes or from differences
in parental or socioeconomic circumstances. Two randomized
trials of 4 compared with 6 mo of EBF were conducted in Honduras, where low birth weight is common (14, 15). In each trial,
breast-milk intake was greater in the 6-mo EBF group, but anthropometric outcomes did not differ between groups. Equivalent
data from populations with larger body size during infancy are
lacking.
The effect of introducing complementary foods is also controversial, as such foods, which are variable, may add to or replace
1
From the Childhood Nutrition Research Centre (JCKW, MSF, and AL)
and the Department of Surgery (SE), UCL Institute of Child Health, London,
United Kingdom; the Unit for Nutrition Research, Landspitali National University Hospital of Iceland, and Faculty of Food Science and Human Nutrition, University of Iceland, Reykjavik, Iceland (OHJ and IT); the Division of
Global Health (PLH) and the Department of Pediatrics (REK), Massachusetts General Hospital for Children, Harvard Medical School, Boston, MA;
and the Directorate of Health and Reykjavı́k University, Reykjavik, Iceland
(GG).
2
Supported by Mead Johnson and the Eimskip Fund of the University of
Iceland.
3
Address correspondence and requests for reprints to JCK Wells, Childhood Nutrition Research Centre, UCL Institute of Child Health, 30 Guilford
Street, London WC1N 1EH, United Kingdom. E-mail: jonathan.wells@ucl.ac.uk.
Received November 8, 2011. Accepted for publication April 2, 2012.
First published online May 16, 2012; doi: 10.3945/ajcn.111.030403.
Am J Clin Nutr 2012;96:73–9. Printed in USA. Ó 2012 American Society for Nutrition
73
Downloaded from www.ajcn.org at LANDSPITALI UNIVERSITY HOSPITAL LIBRARY on June 21, 2012
Jonathan CK Wells, Olof H Jonsdottir, Patricia L Hibberd, Mary S Fewtrell, Inga Thorsdottir, Simon Eaton, Alan Lucas,
Geir Gunnlaugsson, and Ronald E Kleinman
74
WELLS ET AL
SUBJECTS AND METHODS
Recruitment for the study was undertaken from November
2007 to November 2009, with ethical permission granted by the
National Bioethical Committee in Iceland and the Massachusetts
General Hospital Institutional Review Board. Data collection was
approved by the Data Protection Authority in Iceland. Mothers
provided written informed consent to participate.
Recruitment
Mothers were recruited from 7 health centers in the Reykjavik
Capital Area and neighboring municipalities in Iceland. Mothers
were informed about the study when their infant was ;2 mo old
and were given study handouts if the infant was still exclusively
breastfeeding at 3 mo of age. If the infant was still exclusively
breastfeeding at 4 mo and the mother was interested in the study,
they were invited to participate. Mothers and infants attended
a screening visit and, after providing informed consent, were evaluated for the following eligibility criteria:
Infants: singleton birth, gestational age 37 wk, healthy (ie,
absence of congenital abnormalities or chronic health issues likely to affect growth, development, or iron status),
and exclusively breastfeeding at the time of assessment
Mothers: willing to be randomly assigned either to continue to
breastfeed exclusively until infant age of 6 mo or to introduce complementary foods alongside breast milk at infant
age of 4 mo
The operational definition of EBF was breastfeeding with no
additional liquid or solid foods other than vitamins and medications (18, 19), although up to a maximum of 10 feedings of
formula or water during the first 6 mo were allowed because of
practicalities of EBF. Newborn infants are at times given formula or sugar water at birth and may be given extra fluids after
returning home. Eligible infant-mother pairs were randomly
assigned to the 2 feeding groups. Vitamin D supplementation was
provided to both groups. Mothers in both groups received
counseling from nurses who were lactation specialists (International Board Certified Lactation Consultants). Specifically,
EBF mothers were advised to exclusively breastfeed for 6 mo and
CF mothers were advised to introduce complementary foods
within 7 d of being randomly assigned. During the 2-mo study
period, mothers were encouraged to contact the research staff at
the Unit for Nutrition Research (Reykjavik, Iceland) and a specialist at the health care centers if they had any questions or
concerns.
Background information
Baseline data on maternal age and education (categorized here
dichotomously as with/without university education), parity, and
mode of delivery were collected on eligible mother-infant pairs.
Anthropometric measurements
Data on body weight and length and head circumference were
abstracted from the infants’ charts at birth, and the infants were
measured again at 4 and 6 mo by using Seca 757 scales and a Seca
model 207 infantometer (Vogel & Halke) and flexible nonstretching tape. BMI was calculated as weight divided by length
squared. All data were converted to z-scores by using current
WHO reference data (20). Maternal weight was also measured
at 6 mo (Seca 761 scales; Vogel & Halke).
Dietary assessment
A 3-d weighed food record was obtained when the infant was
aged 157 6 7 d to estimate the intake of complementary foods in
the CF group. Parents were advised to keep records for 3 consecutive days and to weigh and record all complementary foods
given over this period by using electronic scales accurate to 1 g
(Philips model HR 2385). Data on diet were collected during the
4–6-mo study period to determine the date of introduction of the
first new food item in both groups. Detailed weighed measurements were not made in the EBF group, but diary records were
obtained in all subjects in case complementary foods were initiated in the EBF group before measuring breast-milk intake.
Energy intakes were calculated from complementary foods by
using nutrient calculation software (ICEFOOD, version 2002;
Public Health Institute of Iceland) (21), with added information
about infant foods.
Isotope measurements
Breast-milk intake was determined by the deuterium dose-tothe-mother method (22, 23). Briefly, each mother received orally
;10 g deuterium oxide (2H2O) diluted in ;50 g drinking water.
The dose consumed was determined accurately to 0.01 g. Predose urine samples were obtained from mothers and infants.
Additional urine samples were collected from the mothers on
days 1, 4, and 14 and from the infants on days 1, 3, 4, 13, and 14.
A second isotope dose (0.05 g deuterium/kg body weight) was
given orally to the infants on day 14 to calculate total body
water; further infant urine samples were collected 5 h after the
dose and on day 16.
Infant urine samples were obtained by placing an absorbent
cotton pad in the diaper, which was checked every 30 min for
wetness. The time of urination was taken as the midpoint between
the last time it was dry and the time it was wet. Urine was
transferred to 2-mL cryogenic tubes and stored at 220°C. Urine
and dose samples were transported frozen to the United Kingdom for isotope-ratio mass spectrometry (Delta XP; Thermo
Downloaded from www.ajcn.org at LANDSPITALI UNIVERSITY HOSPITAL LIBRARY on June 21, 2012
breast milk in the infant diet (16). In observational studies, infants
receiving complementary foods earlier in infancy may also have
systematically different initial growth patterns or other characteristics than those receiving complementary foods later in infancy
(11). We conducted a randomized trial to evaluate the effect of
timing of introduction of complementary foods on infants whose
mothers practiced EBF for the first 4 mo. The study was conducted
in Iceland, where EBF rates at age 6 mo were 12% (17). Mothers
who agreed to participate were randomly assigned either to
continue EBF for 6 mo (EBF group) or to introduce complementary foods alongside breast milk at age 4 mo [complementary
feeding (CF) group]. The primary hypothesis was that breast-milk
intake at 6 mo would be greater in EBF than in CF infants.
Secondary hypotheses were that complementary food intake would
differ between groups, whereas anthropometric outcomes and
body composition would not.
RANDOMIZED TRIAL OF EXCLUSIVE BREASTFEEDING
75
of breast-milk measurement. The index of potential harm assessed
was growth status.
Modeling
Randomization
Total body water was calculated by using the back-extrapolation
method in both mothers and infants (24). Dilution spaces were
assumed to overestimate body water by a factor of 1.044 (25).
Breast-milk intake was calculated by fitting the isotopic (tracer)
data to a model for water (tracee) turnover in the mothers and
infants, and the transfer of milk from mother to infant, on the basis
of equations and assumptions described previously (22, 23). A
“solver” function in Excel (Microsoft) was used to minimize the
sum of squares of differences between observed and fitted values
for maternal and infant data combined.
With the use of specially designed software (www.randomization.
com), permuted blocks of 2 and 4 with the sequence presented in
random order were used to generate assignments, which were
accessed by using a password-protected Web-based application after
eligibility criteria were confirmed. Assignments were generated by
one person who was not involved in any other aspect of the study.
After randomization, one mother who was randomly assigned to
the CF group was incorrectly instructed to the EBF group. We
performed the primary analyses with this mother included in the
EBF group (n = 50 EBF, 50 CF) but have reported outcomes for
the baseline analyses with this subject in the CF group (n = 49
EBF, 51 CF). Nurses who collected data on complementary food
intakes and anthropometric outcomes were not blinded to participant group status, but all mass spectrometric analyses and
isotopic modeling were blinded.
Statistical analyses
The sample size calculations were based on the primary hypothesis that EBF infants would have higher breast-milk intakes
than CF infants. Fifty mother-infant pairs per group allowed us to
detect differences of 75 g/d with 80% power and 5% significance, which is equivalent to ;0.6 SD (23). Recruitment continued until 100 mother-infant pairs completed the study.
Independent t tests and chi-square tests were used to compare
results between groups. Regression analysis, with a dummy variable for feeding group, was used to test group differences in
breast milk with adjustment for confounding variables. Exploratory analyses were used to identify factors associated with EBF
mothers introducing any non-breast-milk foods during the period
RESULTS
A total of 119 mother-infant pairs were recruited, of whom 100
completed the trial protocol (Figure 1). Among the 10 CF
mothers who did not complete the study, either their infants
were not given, or did not accept, complementary foods or the
mothers stopped breastfeeding or did not complete the study.
Seven EBF infants were given complementary foods after 4 mo,
whereas 2 EBF mothers did not complete the study. The primary
FIGURE 1. Profile of the recruitment process. CF, complementary feeding; EBF, exclusive breastfeeding.
Downloaded from www.ajcn.org at LANDSPITALI UNIVERSITY HOSPITAL LIBRARY on June 21, 2012
Fisher Scientific). Samples were analyzed in duplicate (precision
1.4 delta units) by using the equilibration method.
76
WELLS ET AL
TABLE 1
Baseline characteristics of subjects who completed the study compared with dropouts and of the 2 groups of mothers
and infants1
Maternal data
University education (n)
Primiparous (n)
Multiparous (n)
Vaginal delivery (n)
Weight (kg)
Age (y)
Infant data
Male (n)
Age at enrollment (d)
Gestational age (d)
Birth weight (kg)
Birth weight z-score4
Weight at age 4 mo (kg)
Weight z-score4
Length at age 4 mo (cm)
Length z-score4
Completing
(n = 100)
Noncompleting
(n = 19)
54
27
73
83
—
30.2 6 4.83
11
10
9
89
—
28.1 6 6.9
46
123.0 6
280.9 6
3.71 6
0.83 6
7.0 6
0.33 6
64.4 6
0.65 6
3.3
8.3
0.48
0.94
0.9
0.95
2.1
0.86
122.0
278.6
3.67
0.77
7.0
0.37
65.1
1.07
9
6
6
6
6
6
6
6
6
3.9
11.5
0.37
0.65
0.6
0.57
1.8
0.59
P
0.762
0.0272
0.482
—
0.22
0.92
0.77
0.29
0.75
0.77
1.00
0.89
0.13
0.04
CF group
(n = 51)
EBF group
(n = 49)
31
18
33
46
73.0 6 15.5
29.7 6 4.4
24
9
40
36
71.3 6 12.4
30.7 6 5.1
25
123.0 6
279.6 6
3.70 6
0.825 6
7.05 6
0.345 6
64.3 6
0.585 6
3.4
8.8
0.44
0.89
1.00
1.05
2.3
0.91
21
123.0 6
277.9 6
3.71 6
0.835 6
6.99 6
0.335 6
64.5 6
0.725 6
3.2
8.1
0.52
1.00
0.81
0.83
1.9
0.80
P
0.232
0.0572
0.022
0.5
0.2
0.52
0.95
0.3
0.9
0.9
0.9
0.9
0.7
0.4
1
All comparisons were made by independent t test except where indicated otherwise. CF, complementary feeding;
EBF, exclusive breastfeeding.
2
Chi-square test.
3
Mean 6 SD (all such values).
4
Calculated by using WHO reference data (18).
5
Significantly different from WHO mean values, P , 0.05.
Downloaded from www.ajcn.org at LANDSPITALI UNIVERSITY HOSPITAL LIBRARY on June 21, 2012
Mean (6SD) breast-milk intake in the EBF group was 901 6
158 g/d, which was higher than the WHO reference values of
854 6 24 g/d (D = 47 g/d; 95% CI: 2, 92 g/d; P = 0.04). The CF
group consumed 818 6 166 g breast milk/d, which was not
significantly different from the WHO reference values (D =
236; 95% CI: 283, 11 g/d). The EBF group therefore consumed 83 g/d (95% CI: 19, 148 g/d) more breast milk than did
the CF group (P = 0.012). Regression analysis showed that,
holding constant for primiparity and mode of delivery, the
EBF group still consumed significantly more breast milk (D =
83 g/d; 95% CI: 18, 154 g/d). None of the findings for growth,
body composition, or breast-milk intake (D = 76; 95% CI: 11,
141 g/d) changed if the analysis was conducted for the 2
groups as originally randomized (ie, 51 CF compared with 49
EBF infants).
The average daily intake of complementary foods among CF
infants was 92.2 g and included 43% water given as a drink or as
part of foods prepared by the parents, such as porridge. The
remaining foods included 17% infant formula, 16% fruit purée,
12% fruit/vegetables, 9% infant cereal (dry mass), and small
amounts of other foods. Mean (6SD) daily energy intake was
63.4 6 52.3 kcal, which is equivalent to 265.4 6 219 kJ, of
which 49% was attributable to infant cereals, 12% to fruit/
vegetables, 18% to infant formula, 13% to fruit purée, 3% to oils
or butter, 1% to infant vitamin or cod liver oil, and 4% to other
foods.
During the 2-wk period of isotope measurement, a small
proportion of mothers in the EBF group gave their infants either
water (n = 4; 8% of the group), sugar water (n = 1; 2%), formula
milk (n = 3; 6%), or rice porridge (n = 2; 4%). In total, 6 mothers
(12%) gave their infants some energy-containing fluids or foods
other than breast milk. In most cases, however, the amounts were
outcome (milk intake) was not measured in the infants whose
mothers did not comply with the protocol.
Data on maternal and infant characteristics at birth or at the
time of randomization are provided in Table 1. There were few
differences between completers and noncompleters; however,
noncompleting infants were 0.9 cm (95% CI: 20.1, 1.9 cm)
longer at 4 mo, which was significant when expressed in zscores (0.42; 95% CI: 0.01, 0.82; P = 0.04). Noncompleters
were also more likely to be primiparous (dropouts: 10 of 19,
52.6%; completers: 27 of 100, 27%; P = 0.027). Within the EBF
group, noncompleting infants did not differ in 4-mo anthropometric outcomes compared with completing infants. Within the
EBF group, noncompleting mothers were younger than completing mothers (mean 6 SD: 25.0 6 3.4 compared with 30.9 6
5.1 y; NS) and were significantly more likely to be primiparous
(n = 6 of 8 compared with n = 10 of 50, P , 0.001).
The mother-infant pairs in the 2 groups who completed the trial
are also compared in Table 1. There were no differences in
maternal characteristics or in infant anthropometric outcomes,
except that a smaller proportion of the EBF infants were delivered
vaginally (46% compared with 36%, P = 0.03), and there was
a borderline-significant trend for fewer primiparous mothers in
the EBF group (P = 0.057).
Infant anthropometric outcomes and breast-milk intake at age
6 mo are shown in Table 2. Although one mother was instructed
to join the wrong group, all mothers in the final analysis successfully followed the isotope protocol for breast-milk intake.
Relative to WHO reference values, the sample had significantly
greater anthropometric outcome z-scores in both groups from
birth to age 6 mo. However, there were no significant differences
between groups in anthropometric outcomes (Figure 2) or in
lean or fat mass by deuterium dilution.
77
RANDOMIZED TRIAL OF EXCLUSIVE BREASTFEEDING
TABLE 2
Age, size, body composition, and breast-milk intake of the infants at age 6 mo1
CF group
(n = 50)
Age (d)
Weight (kg)
Weight z-score
Length z-score
BMI z-score
Head circumference z-score
Lean mass (kg)6,7
Fat mass (kg)6,7
Breast-milk intake6 (g/d)
1
3
4
5
6
7
6
6
6
6
6
6
6
6
6
3
7.1
1.06
1.08
0.92
1.14
0.77
0.92
0.96
166
187.3
8.01
0.365
0.775
20.10
1.025
4.96
3.04
901
6 6.6
6 1.04
6 0.99
6 0.84
6 1.04
6 0.89
6 1.18
6 1.12
6158
Difference between
groups2
4
0.6 (23.3, 2.2)
0.05 (20.47, 0.37)
20.08 (20.49, 0.33)
0.17 (20.52, 0.18)
20.03 (20.41, 0.46)
0.07 (20.26, 0.40)
20.17 (20.28, 0.62)
0.33 (20.77, 0.11)
83 (19, 148)
P
0.6
0.8
0.7
0.3
0.9
0.6
0.4
0.14
0.012
CF, complementary feeding; EBF, exclusive breastfeeding.
Differences between groups tested by independent t test
Mean 6 SD (all such values).
Mean; 95% CI in parentheses (all such values).
Significantly different from WHO mean values by paired t test, P , 0.05.
Determined by using isotope measurements.
n = 43 and 46 for CF and EBF groups, respectively.
very small, occurred on one day only, and contained very little
energy. Because the isotope method estimates deuterium transfer
from the mother, these extra fluids do not compromise the accuracy of breast-milk intake values. Only 2 mothers gave their
infants any significant quantity of energy other than breast milk
during the isotope study period: one fed her infant 200 mL of
infant formula on one day only, whereas one mother provided
4–5 spoons of rice porridge almost daily. Breast-milk intake of
the 6 infants who received any extra calories was marginally (26
g/d; 95% CI: 2166, 113 g/d) higher than that in the remaining
44 EBF infants (mean 6 SD intakes: 924 6 192 compared with
898 6 155 g/d), suggesting they were hungrier, but the difference was not significant. If these 6 infants were excluded, the
remaining 44 EBF infants still consumed significantly more
breast milk than did the 50 CF group infants (D = 80 g/d; 95%
CI: 13, 146 g/d; P = 0.018).
DISCUSSION
Our trial, which, to our knowledge, is the first of its kind in
a population with average growth rates consistent with current
WHO reference data throughout infancy, showed that infants of
mothers randomly assigned to 6 mo of EBF showed satisfactory
growth and had average breast-milk intakes slightly above WHO
reference values (26). Infants randomly assigned to complementary foods at age 4 mo consumed significantly less breast
milk at age 6 mo. Our approach allowed experimental disentangling of the effect of introducing complementary foods on
breast-milk intake from the effects of baseline maternal or infant
factors that may affect both infant diet and growth. Our findings
were similar to those of 2 trials in Honduras, a population with
a high frequency of low-birth-weight infants, where breast-milk
intake at age 6 mo was greater in EBF than in CF infants, although
growth rates were similar (14, 15).
The comparison of intakes of breast milk and complementary
foods between groups of infants, whose behavior or social circumstances may drive the pattern of weaning, has long been
a challenge. Measurement of breast-milk intake is also challenging, because intakes are “invisible” and cannot be readily
studied using approaches common in other age groups. Techniques adapted for infants, such as test-weighing, may intrude on
the normal behavior of mothers and/or infants and hence bias the
results (27). Despite the enormous emphasis placed on the value
of breastfeeding for short- and long-term health, objective data on
breast-milk intakes remain sparse, although a review of objective
isotope data from 12 countries was recently published (28). A
third challenge is that rates of EBF at age 6 mo are typically
<15% in industrialized populations [Canada, 13.8%; United
States, 11.3%; Sweden, 10.1%; Norway, 7% (29–32)] and only
;1% in the United Kingdom (12).
We addressed these challenges in a randomized trial in Iceland
by using an established isotope method to obtain objective data
on breast-milk intake. In a national cohort in 1995–1997, rates of
EBF in Iceland were 88% at 1 mo, 69% at 3 mo, and 46% at 4 mo
(33), whereas in a 2005–2007 cohort, rates were higher (34).
Previous data on our research question are sparse. In a review
of observational data from 8 developed country populations, the
mean (6SD) breast-milk intake in 93 EBF infants at age 6 mo
was 854 6 118 g/d, although individual population averages
ranged from 800 6 120 to 925 6 112 g/d (26). There is concern
FIGURE 2. Changes in anthropometric z-scores between ages 4 and 6 mo
in EBF and CF groups. The average changes did not differ between groups
(all P values were .0.4). Error bars represent SD. CF, complementary
feeding; EBF, exclusive breastfeeding.
Downloaded from www.ajcn.org at LANDSPITALI UNIVERSITY HOSPITAL LIBRARY on June 21, 2012
2
186.7
7.96
0.28
0.605
20.08
0.945
5.13
2.71
818
EBF group
(n = 50)
78
WELLS ET AL
cite concern that their breast milk has become insufficient as
a source of nutrition for their infant as a reason for introducing
formula-milk or complementary foods (39, 40); hence, evidence
on the breast-milk intakes of EBF infants will help encourage
mothers to breastfeed. In the present study, we have shown that
the large majority (77%) of a sample of mothers practicing EBF
at 4 mo successfully continued EBF to 6 mo, with their infants
showing adequate growth. Another longitudinal study, in Glasgow, showed that EBF mothers recruited from breastfeeding
support groups increased breast-milk output between 4 and 6 mo
(41). Thus, both of these studies rebut our challenge and suggest
that EBF successfully meets energy requirements, but further
work will be required to determine whether all mothers in this and
other populations could achieve this outcome.
The strengths of our study include its randomized design and
the use of an objective method for measuring breast-milk intake.
Conducting the study in Iceland represents both a strength, because EBF is sufficiently common there for a randomized trial to
be feasible, and a limitation, because the cold environment and
high birth weight are not typical of most European populations,
which reduces generalizability. The 2 feeding groups differed at
baseline in the mode of delivery and maternal parity; however,
these differences were relatively minor and did not account for
our main findings regarding breast-milk or complementary food
intakes. We were not able to test whether those who dropped out
of the study were those with lower breast milk intakes—ie, potentially a self-selected group. Finally, our study was designed to
evaluate growth and energy intake and not other issues such as
development of dietary preferences, mineral status, or effects on
health such as diarrhea and allergy.
We are grateful to those who participated and to colleagues for the following contributions. The recruitment, data collection on breastfeeding, and provision of support for mothers were undertaken by the nurses: Alma Maria
Rognvaldsdóttir, Audur Egilsdottir, Dagny Gudmundsdottir, Sesselja Gudmundsdóttir, Elin Sigurbjornsdottir, Gudrun Gudmundsdottir, Ingibjorg Eiriksdottir, Jona Margret Jonsdottir, and Kristin Vigfusdottir. Dietician Margret
Thora Jonsdottir and PhD student Tinna Eysteinsdottir are also gratefully acknowledged for their contributions to data collection. The study was also supported by the Primary Health Care organizations in the Reykjavik Capital
Area, Akranes, and Sudurnes, and by the directors of the participating health
centers. Anna Fiorino prepared the randomization method using Jerry Dallal’s
Tufts-based software, developed the Web-based data entry system, and synthesized the study database. James Grinham performed mass spectrometric
analyses at UCL Institute of Child Health. The protocol can be obtained from
Ronald E Kleinman.
The responsibilities of the authors were as follows:—GG, AL, MSF, REK,
PLH, JCKW, and IT: designed the study; OHJ: was responsible for the nutritional and isotope data collection; SE: was responsible for the mass spectrometric analyses; and JCKW: undertook the isotope calculations and, after
extensive discussions among the study team, conducted the statistical analysis
and wrote the first draft of the manuscript. All authors critically revised the
manuscript and had full access to all of the data in this study and take complete
responsibility for the integrity of the data and the accuracy of the data analysis.
The study sponsors had no role in study design; in the collection, analysis,
and interpretation of data; in the writing of the report; or in the decision
to submit the manuscript for publication. None of the authors declared a conflict of interest.
REFERENCES
1. World Health Organization. Infant and young child nutrition. 55th
World Health Assembly. Geneva, Switzerland: World Health Organization, 2002.
Downloaded from www.ajcn.org at LANDSPITALI UNIVERSITY HOSPITAL LIBRARY on June 21, 2012
regarding the utility of these test-weighing values from highly
selected populations as a reference. Randomized trials enable
these concerns to be addressed. Data from 2 randomized trials
from Honduras indicated no adverse effect of 6 mo of EBF on
growth (14, 15), but whether this applies to populations with
larger infant sizes requires confirmation.
The mothers in our trial showed a high degree of compliance
with the protocol. The majority (88%) of completing EBF
mothers reported giving no energy-containing food to their infants up to the end of the isotope measurement, and of those who
provided such foods, only 2 did so in any significant quantity. The
exclusion of these infants did not affect our findings. However,
another 7 EBF infants were dropped from the study because they
were given complementary foods; hence, the total EBF compliance
rate was 77%. For CF infants, the compliance rate was 82%. Because
the primary outcomes were not measured in those not complying
with the full protocol, our analysis was restricted to the 100 infants
with isotopic data.
We calculated energy consumption with the assumption that
the metabolizable energy content of breast milk at age 6 mo is
62.1 kcal/100 g (4). Current WHO guidelines for energy requirements in breastfed infants at age 6 mo are 78 kcal kg21 d21
(35), which in this study would equate to 625 kcal/d for the average EBF infant in this study. The EBF infants received 560 6
98.2 kcal/d, which is equivalent to 90.1% of the recommendation.
The CF group received 508 6 103 kcal/d from breast milk plus
63 6 53.2 kcal/d from complementary foods, totaling 571 6 97
kcal/d, which is equivalent to 92.7% of the recommendation.
Our data therefore show an apparent inconsistency, in that the
infants appear to meet WHO recommendations for breast-milk
intake but not energy intake. This may be because the WHO
recommendations were calculated from total energy expenditure
data in predominantly, not exclusively, breastfed infants. Introducing formula milk or cow milk into the diet increases infant
energy expenditure (35, 36); hence, the value of 78 kcal kg21 d21
may be an overestimate of actual energy requirements in EBF infants. It is also possible that the energy content of breast milk in our
population is .62.1 kcal/100g, and this merits further research.
The similar estimated energy intakes of the 2 groups indicate
that complementary foods acted primarily to replace breast milk.
Given the similarity between groups in body weight throughout
the trial and in body composition and estimated energy intake
at age 6 mo, we assume that infant appetite was the key factor
driving energy intake.
There was no indication of energy inadequacy in the EBF
infants. Anthropometric outcome z-scores were consistently above
zero; although this partly reflects the infants’ large size at birth,
which is typical in the Icelandic population (37), z-scores at ages 4
and 6 mo in the 2 groups were very similar, as was lean mass at
age 6 mo, giving no indication of faltering in the EBF group.
However, each group also showed variability, and we were unable
to determine whether those infants within each group who grew
less between birth and 6 mo had insufficient breast-milk intakes
earlier in infancy.
We have highlighted the fact that previously published evidence on breast-milk intakes did not support the proposition that
the average infant could satisfy his or her energy requirements at
age 6 mo from breast milk alone (4, 38). This suggestion was not
intended to undermine breastfeeding but rather to promote further
research into this important issue. Mothers themselves commonly
RANDOMIZED TRIAL OF EXCLUSIVE BREASTFEEDING
23. Haisma H, Coward WA, Albernaz E, Visser GH, Wells JC, Wright A,
Victora CG. Breast milk and energy intake in exclusively, predominantly, and partially breast-fed infants. Eur J Clin Nutr 2003;57:
1633–42.
24. Davies PS, Wells JC. Calculation of total body water in infancy. Eur J
Clin Nutr 1994;48:490–5.
25. Racette SB, Schoeller DA, Luke AH, Shay K, Hnilicka J, Kushner RF.
Relative dilution spaces of 2H- and 18O-labeled water in humans. Am
J Physiol 1994;267:E585–90.
26. Butte NF, Lopez-Alarcon MG, Garza C. Nutrient adequacy of exclusive breastfeeding for the term infant during the first six months of life.
Geneva, Switzerland: WHO, 2002.
27. Lucas A, Ewing G, Roberts SB, Coward WA. How much energy does
the breast fed infant consume and expend? Br Med J (Clin Res Ed)
1987;295:75–7.
28. da Costa TH, Haisma H, Wells JC, Mander AP, Whitehead RG, Bluck
LJ. How much human milk do infants consume? Data from 12 countries using a standardized stable isotope methodology. J Nutr 2010;140:
2227–32.
29. Al-Sahab B, Lanes A, Feldman M, Tamim H. Prevalence and predictors
of 6-month exclusive breastfeeding among Canadian women: a national
survey. BMC Pediatr 2010;10:20.
30. Lande B, Andersen LF, Baerug A, Trygg KU, Lund-Larsen K, Veierod
MB, Bjørneboe GE. Infant feeding practices and associated factors in
the first six months of life: the Norwegian infant nutrition survey. Acta
Paediatr 2003;92:152–61.
31. Brekke HK, Ludvigsson JF, van Odijk J, Ludvigsson J. Breastfeeding
and introduction of solid foods in Swedish infants: the All Babies in
Southeast Sweden study. Br J Nutr 2005;94:377–82.
32. Centers for Disease Control and Prevention. Breastfeeding trends and
updated national health objectives for exclusive breastfeeding—United
States, birth years 2000-2004. MMWR Morb Mortal Wkly Rep 2007;
56:760–3.
33. Atladottir H, Thorsdottir I. Energy intake and growth of infants in
Iceland-a population with high frequency of breast-feeding and high
birth weight. Eur J Clin Nutr 2000;54:695–701.
34. Thorisdottir AV, Thorsdottir I, Palsson GI. Nutrition and iron status of
1-year-olds following a revision in infant dietary recommendations.
Anemia 2011;2011:986303.
35. Food and Agriculture Organization. Human energy requirements. Report
of a Joint FAO/WHO/UNU Expert Consultation. Rome, 17–24 October
2001. Rome, Italy: United Nations University, WHO/FAO, 2004.
36. Haisma H, Wells JC, Coward WA, Filho DD, Victora CG, Vonk RJ,
Wright A, Visser GH. Complementary feeding with cow’s milk alters
sleeping metabolic rate in breast-fed infants. J Nutr 2005;135:1889–95.
37. Johannsson E, Arngrimsson SA, Thorsdottir I, Sveinsson T. Tracking
of overweight from early childhood to adolescence in cohorts born
1988 and 1994: overweight in a high birth weight population. Int J
Obes (Lond) 2006;30:1265–71.
38. Reilly JJ, Wells JC. Duration of exclusive breast-feeding: introduction
of complementary feeding may be necessary before 6 months of age.
Br J Nutr 2005;94:869–72.
39. Schluter PJ, Carter S, Percival T. Exclusive and any breast-feeding
rates of Pacific infants in Auckland: data from the Pacific Islands
Families First Two Years of Life Study. Public Health Nutr 2006;9:
692–9.
40. Li R, Fein SB, Chen J, Grummer-Strawn LM. Why mothers stop
breastfeeding: mothers’ self-reported reasons for stopping during the
first year. Pediatrics 2008;122(suppl 2):S69–76.
41. Nielsen SB, Reilly JJ, Fewtrell MS, Eaton S, Grinham J, Wells JC.
Adequacy of milk intake during exclusive breastfeeding: a longitudinal
study. Pediatrics 2011;128:e907–14.
Downloaded from www.ajcn.org at LANDSPITALI UNIVERSITY HOSPITAL LIBRARY on June 21, 2012
2. Scientific Advisory Committee on Nutrition (SACN). Paper for discussion:
introduction of solid foods, agenda item: 3. Subgroup on Maternal and
Child Nutrition. London, United Kingdom: The Stationery Office, 2003.
3. American Academy of Pediatrics. Breastfeeding and the use of human
milk. Pediatrics 2005;115:496–506.
4. Reilly JJ, Ashworth S, Wells JC. Metabolisable energy consumption in
the exclusively breast-fed infant aged 3–6 months from the developed
world: a systematic review. Br J Nutr 2005;94:56–63.
5. Kramer MS, Kakuma R. The optimal duration of exclusive breastfeeding: a systematic review. Adv Exp Med Biol 2004;554:63–77.
6. Fewtrell M, Wilson DC, Booth I, Lucas A. Six months of exclusive
breast feeding: how good is the evidence? BMJ 2011;342:c5955.
7. Motil KJ, Kertz B, Thotathuchery M. Lactational performance of adolescent mothers shows preliminary differences from that of adult
women. J Adolesc Health 1997;20:442–9.
8. Fewtrell MS, Lucas A, Morgan JB. Factors associated with weaning in
full term and preterm infants. Arch Dis Child Fetal Neonatal Ed 2003;
88:F296–301.
9. Griffiths LJ, Tate AR, Dezateux C. Do early infant feeding practices
vary by maternal ethnic group? Public Health Nutr 2007;10:957–64.
10. Hawkins SS, Griffiths LJ, Dezateux C, Law C. The impact of maternal
employment on breast-feeding duration in the UK Millennium Cohort
Study. Public Health Nutr 2007;10:891–6.
11. Wright CM, Parkinson K, Scott J. Breast-feeding in a UK urban context: who breast-feeds, for how long and does it matter? Public Health
Nutr 2006;9:686–91.
12. Bolling K, Grant C, Hamlyn C, Thornton A. Infant feeding survey
2005. Leeds, United Kingdom: The Information Centre, 2007.
13. Kelly YJ, Watt RG. Breast-feeding initiation and exclusive duration at
6 months by social class–results from the Millennium Cohort Study.
Public Health Nutr 2005;8:417–21.
14. Cohen RJ, Brown KH, Canahuati J, Rivera LL, Dewey KG. Effects of
age of introduction of complementary foods on infant breast milk intake, total energy intake, and growth: a randomised intervention study
in Honduras. Lancet 1994;344:288–93.
15. Dewey KG, Cohen RJ, Brown KH, Rivera LL. Age of introduction of
complementary foods and growth of term, low-birth-weight, breast-fed
infants: a randomized intervention study in Honduras. Am J Clin Nutr
1999;69:679–86.
16. Quandt SA. The effect of beikost on the diet of breast-fed infants. J Am
Diet Assoc 1984;84:47–51.
17. Gunnlaugsson G. Brjóstagjöf á Íslandi. Börn fædd árið 1999-2002.
[Breastfeeding in Iceland. Children born 1999-2002.] Reykjavik, Iceland: Miðstöð heilsuverndar barna (Centre for Child Health Services),
2005 (in Icelandic).
18. WHO. Indicators for assessing breastfeeding practices. Report of an
informal meeting. Division of Child and Adolescent Health. Geneva,
Switzerland: WHO, 1991. (Publication WHO/CDD/SER/91.14.)
19. WHO, UNICEF, USAID, Academy for Educational Development,
University of California Davis, International Food Policy Research
Institute. Indicators for assessing infant and young child feeding
practices—part I: Definitions. Conclusions of a consensus meeting held
6–8 November 2007 in Washington, DC, 2008. Geneva, Switzerland:
WHO, 2008.
20. WHO. Software for assessing growth and development of the world’s
children, version 3.1. Geneva, Switzerland: WHO, 2010.
21. Steingrı́msdottir L, Þorgeirsdottir H, Olafsdottir AS. The Icelandic
National Nutrition Survey 2002. Reykjavik, Iceland: Public Health
Institute of Iceland, 2003.
22. Coward WA, Cole TJ, Sawyer MB, Prentice AM. Breast-milk intake
measurement in mixed-fed infants by administration of deuterium
oxide to their mothers. Hum Nutr Clin Nutr 1982;36:141–8.
79