Content uploaded by Roberta Cohen
Author content
All content in this area was uploaded by Roberta Cohen on Jan 13, 2014
Content may be subject to copyright.
Community and International Nutrition
Effects of Exclusive Breastfeeding for Four versus Six Months on Maternal
Nutritional Status and Infant Motor Development: Results of Two
Randomized Trials in Honduras
1
Kathryn G. Dewey,*
2
Roberta J. Cohen,* Kenneth H. Brown* and Leonardo Landa Rivera
†
*Department of Nutrition and Program in International Nutrition, University of California, Davis, California
95616-8669 and
†
Medicina Infantil, San Pedro Sula, Honduras
ABSTRACT To examine whether the duration of exclusive breastfeeding affects maternal nutrition or infant motor
development, we examined data from two studies in Honduras: the first with 141 infants of low-income primiparous
women and the second with 119 term, low birth weight infants. In both studies, infants were exclusively breastfed
for 4 mo and then randomly assigned to continue exclusive breastfeeding (EBF) until 6 mo or to receive
high-quality, hygienic solid foods (SF) in addition to breast milk between 4 and 6 mo. Maternal weight loss between
4 and 6 mo was significantly greater in the exclusive breastfeeding group (EBF) group than in the group(s) given
solid foods (SF) in study 1 (⫺0.7 ⫾ 1.5 versus ⫺0.1 ⫾ 1.7 kg, P ⬍ 0.05) but not in study 2. The estimated average
additional nutritional burden of continuing to exclusively breastfeed until 6 mo was small, representing only
0.1– 6.0% of the recommended dietary allowance for energy, vitamin A, calcium and iron. Women in the EBF group
were more likely to be amenorrheic at 6 mo than women in the SF group, which conserves nutrients such as iron.
In both studies, few women (10–11%) were thin (body mass index ⬍19 kg/m
2
), so the additional weight loss in the
EBF group in study 1 was unlikely to have been detrimental. Infants in the EBF group crawled sooner (both studies)
and were more likely to be walking by 12 mo (study 1) than infants in the SF group. Taken together with our previous
findings, these results indicate that the advantages of exclusive breastfeeding during this interval appear to
outweigh any potential disadvantages in this setting. J. Nutr. 131: 262–267, 2001.
KEY WORDS:
●
breastfeeding
●
complementary feeding
●
maternal nutrition
●
lactation
●
amenorrhea
●
motor development
There currently is an active debate about the recommended
age of introduction of complementary foods to breastfed in-
fants. The World Health Organization stipulates an age inter-
val of 4 – 6 mo (World Health Organization 1995), whereas
UNICEF and the American Academy of Pediatrics have used
the wording “at about 6 mo”(UNICEF 1999, American Acad-
emy of Pediatrics 1997). Most of the evidence used to evaluate
the optimal duration of exclusive breastfeeding (EBF)
3
has
focused on infant intake, growth and morbidity, with little
attention devoted to effects on the mother or on other func-
tional outcomes for the infant (Brown et al. 1998). It has been
argued that there may be tradeoffs between maternal and
infant needs and that a comprehensive assessment of risks and
benefits for both mother and infant is needed to formulate
appropriate feeding recommendations (Frongillo and Habicht
1997, McDade and Worthman 1998).
Only two randomized intervention trials have been con-
ducted to examine the effects of introducing complementary
foods at 4 versus 6 mo of age, both in Honduras (Cohen et al.
1994, Dewey et al. 1999). The first study included 141 infants
of low-income, primiparous women and the second included
119 term, low birth weight (LBW) (i.e., small-for-gestational
age) infants. In both studies, there was significant displace-
ment of breast milk intake when hygienic, nutrient-rich solid
foods were introduced and no significant impact on infant
growth (short or long term) or food acceptance to 12 mo of age
(Cohen et al. 1995a and 1995b). The social and cultural
feasibility of EBF for 6 mo was evaluated in both populations
(Cohen et al. 1995c and 1999), and although there were
several obstacles to achieving this goal, women who perse-
vered became enthusiastic proponents of this practice. The
present article describes findings from these two trials regard-
ing other important outcomes: maternal nutritional status,
lactational amenorrhea and infant motor development.
1
Supported by the Thrasher Research Fund; the World Health Organization;
UNICEF; the Institute for Reproductive Health, Georgetown University, under a
Cooperative Agreement with the U.S. Agency for International Development
(DPE-3040-A-00-5064-00 and DPE-3061-A-00-1029-00); and Wellstart Interna-
tional through the Health Services Division, Office of Health, Bureau of Global
Programs, Field Support and Research Development, U.S. Agency for Interna-
tional Development (Cooperative Agreement No. DPE-5966-A-00-1-45-00). The
views expressed by the authors do not necessarily reflect the views of these
institutions.
2
To whom correspondence should be addressed at Department of Nutrition,
University of California, One Shields Avenue, Davis, CA 95616-8669. E-mail:
kgdewey@ucdavis.edu
3
Abbreviations used: BMI, body mass index; EBF, exclusive breastfeeding
group; LBW, low birth weight; RDA, Recommended Dietary Allowance; SF,
group(s) given solid foods; SF-M, group given solid foods with maintenance of
preintervention breastfeeding frequency.
0022-3166/01 $3.00 © 2001 American Society for Nutritional Sciences.
Manuscript received 14 August 2000. Initial review completed 5 September 2000. Revision accepted 4 November 2000.
262
METHODS
Study design and selection criteria. Each study was designed as
a prospective observational study from birth to 4 mo, followed by a
randomized intervention trial to determine the impact of comple-
mentary foods from 4 to 6 mo and a follow-up period from 6 to 12 mo.
Subjects were recruited from the two main maternity hospitals in San
Pedro Sula, Honduras. Selection criteria for study 1 were that the
mother is primiparous, willing to exclusively breastfeed for 6 mo, not
employed outside the home before 6 mo postpartum, of low income
(⬍$150/mo), 16 y or older and healthy (not taking medication on a
regular basis) and that the infant is healthy, term and weighs ⱖ2000
g at birth. Selection criteria for study 2 (LBW) were similar except
that the infant birth weight was 1500–2500 g, the maternal age was
ⱖ15 y and there were no limitations on income or parity. Twins and
infants with severe medical conditions that might interfere with food
intake or growth were excluded from both studies.
At 16 wk, infants who were still exclusively breastfed were ran-
domly assigned to intervention groups. Subjects in study 1 were
assigned to one of three groups: 1) EBF to 26 wk, with no other
liquids (water, milk or formula) or solids (EBF), 2) introduction of
solid foods at 16 wk, with ad libitum breastfeeding (SF) or 3)
introduction of solid foods at 16 wk, with maintenance of preinter-
vention breastfeeding frequency (SF-M). Subjects in study 2 were
assigned to one of two groups: EBF or SF-M, as described earlier.
Complementary foods of high nutritional quality were provided in
jars and fed twice daily to infants in the SF and SF-M groups, as
described elsewhere (Cohen et al. 1994, Dewey et al. 1999).
Randomization was done by week of birth to facilitate provision of
feeding instructions to each group during their visits to the research
center. Subjects were not informed of their assignment until they had
completed the first 16 wk of the study. Measurements of infant breast
milk intake, milk composition and solid food intake were made at the
research center for all subjects in study 1 at 16, 21 and 26 wk
postpartum and for a subsample of 50% of subjects in study 2 at 16
and 26 wk postpartum (Cohen et al. 1994, Dewey et al. 1999). Home
visits were conducted weekly from 1 to 26 wk postpartum and
monthly thereafter (after the intervention phase) until 12 mo to
record maternal return of menses and infant growth, morbidity, motor
development and feeding practices. Infant blood samples were col-
lected at 6 mo in study 1 (Dewey et al. 1998a) and at 2, 4 and 6 mo
in study 2 (Dewey et al. 1998b). The study protocols were approved
by the Human Subjects Review Committee of the University of
California, Davis.
Maternal anthropometry. Maternal weight was measured shortly
after delivery and monthly thereafter using a digital scale accurate to
the nearest 0.2 kg. Accuracy of scales was checked weekly using
standard weights. Maternal height was measured using a metal tape
and headboard against a wall. Body mass index (BMI) was calculated
as weight (in kg)/height (in m
2
). The prediction equation of Pollock
et al. (1975) was used to estimate maternal percent body fat.
Duration of lactational amenorrhea. At each home visit,
women were asked if they experienced any menstrual bleeding since
the previous visit, and if so, the dates and duration of each episode
were recorded. Information on the use of hormonal contraceptives
was also collected. The definition of the first menstrual period was
based on the following criteria: a) it lasted ⬎1 day, b) it occurred after
56 d postpartum and c) it was followed by an interval of at least 21 d
but not ⬎70 d before the next bleed, as previously described (Dewey
et al. 1997). Data from study 1 were previously reported (Dewey et al.
1997) but are included here along with the new data from study 2 for
completeness.
Infant motor development. At each home visit, mothers were
also asked to report whether their infants could perform any of the
following 10 motor milestones and, if so, when it first occurred. Field
workers were trained to probe for the specific criteria listed for each
milestone. The milestones included were: 1) while lying face down,
the infant can raise the head and look forward; 2) while lying face
down, the infant can raise the head and chest, supporting the body
with the arms; 3) the infant can regularly roll over (from back to
front); 4) the infant can crawl (sustained movement); 5) from a lying
down position, the infant can get into a sitting position; 6) the infant
can stand while holding on to furniture; 7) the infant can pull to a
standing position; 8) the infant can walk while holding on to furni-
ture (“cruising”); 9) the infant can stand alone (for ⱖ30 s); 10) the
infant can walk unaided.
Data analysis. Data were analyzed by using SAS-PC software
(SAS Institute 1987). Group comparisons of maternal weight and
BMI were performed using Student’s t test and analysis of variance
(ANOVA), and the percentage who were amenorrheic at 6 mo was
compared using
2
tests. The duration of lactational amenorrhea was
analyzed using survival analysis (Kaplan-Meier, PROC: LIFEREG).
For motor development, ANOVA was used in initial analyses to
compare the average age at which each milestone was achieved across
groups. However, some infants had not achieved all of the motor
milestones by 12 mo. When this was the case (milestones 4 –10),
survival analysis (PROC: LIFEREG) was used to compare groups,
including censored values for infants who had not achieved that
particular milestone and for those who dropped out after 6 mo. Some
infants never exhibited crawling (five in each of the three interven-
tion groups in study 1; one in each of the two intervention groups in
study 2) or sitting from a lying position (only in study 1: one in EBF,
two in SF and one in SF-M), even though they achieved subsequent
milestones; in these cases, the values were considered missing rather
than censored. Survival analysis was not informative for the last
milestone (walking) because fewer than half of the infants were
walking by 12 mo; therefore,
2
analysis of the percentage who were
walking by 12 mo was used instead. In the latter analysis, subjects
who dropped out of the studies before 12 mo were excluded. In the
ANOVAs, exclusion of dropouts did not change the results, so values
for the total sample are presented.
RESULTS
Maternal nutritional status and lactational amenorrhea.
Table 1 shows the changes in maternal weight and BMI
between 4 and 6 mo for each of the two studies. Data for the
two solid food groups in study 1 were pooled because there was
no significant difference in maternal weight change between
TABLE 1
Maternal weight and body mass index 4 – 6 mo postpartum in
lactating Honduran women
EBF SF (pooled)
1
Study 1 (primiparous mothers)
n 50 91
Weight, kg
4 mo 53.4 ⫾ 7.6
2
53.4 ⫾ 9.6
6 mo 52.7 ⫾ 7.8 53.2 ⫾ 10.1
⌬⫺0.7 ⫾ 1.5 ⫺0.1 ⫾ 1.7
3
BMI, kg/m
2
4 mo 22.5 ⫾ 3.3 22.7 ⫾ 3.4
6 mo 22.0 ⫾ 3.1 22.6 ⫾ 3.6
⌬⫺0.5 ⫾ 1.6 ⫺0.1 ⫾ 0.8
3
Study 2 (mothers of LBW infants)
n 59 60
Weight, kg
4 mo 51.8 ⫾ 8.3 52.0 ⫾ 9.0
6 mo 51.5 ⫾ 8.4 51.9 ⫾ 9.2
⌬⫺0.3 ⫾ 1.6 ⫺0.1 ⫾ 1.7
BMI, kg/m
2
4 mo 22.9 ⫾ 3.2 23.0 ⫾ 3.5
6 mo 22.8 ⫾ 3.3 22.9 ⫾ 3.6
⌬⫺0.1 ⫾ 0.7 ⫺0.1 ⫾ 0.8
1
SF and SF-M groups combined in study 1.
2
Values are means ⫾ SD.
3
P ⬍ 0.05, EBF vs. SF groups, Student’s t test.
Abbreviations: BMI ⫽ body mass index; EBF ⫽ exclusively breast-
feeding; LBW ⫽ low birth weight; SF ⫽ solid foods given to infants and
breastfeeding ad libitum; SF-M ⫽ solid foods given to infants and
maintenance of baseline nursing frequency.
EFFECTS OF EXCLUSIVE BREASTFEEDING FOR 4 OR 6 MO 263
the SF and SF-M groups (⫺0.2 ⫾ 1.6 versus ⫺0.1 ⫾ 1.8 kg,
respectively). In both studies, the EBF and SF groups were very
similar in weight and BMI at 4 mo, before the intervention.
Average BMI at 4 mo was ⬃23 kg/m
2
in both studies. In study
1, the EBF group lost significantly more weight (difference of
0.6 kg) and BMI (difference of 0.4 kg/m
2
) during the 2-mo
intervention than the SF group, but there was no significant
difference between intervention groups in study 2. There was
no significant interaction between initial maternal BMI (⬍22
or ⱖ22) and intervention group in either study.
The theoretical maternal nutritional burden of continued
EBF between 4 and 6 mo, based on observed differences in
mean milk volume and energy output between intervention
groups, is illustrated in Table 2 for energy, vitamin A, calcium
and iron. These nutrients were chosen because there are body
reserves of each that may become depleted during lactation
(Institute of Medicine 1991). The differences in the output of
each nutrient are shown for the comparison of the EBF and SF
groups in study 1 (the maximum difference observed in breast
milk output) and for the EBF and SF-M groups in study 2 (the
minimum difference observed in breast milk output). Also
shown is the corresponding percentage of the RDA or esti-
mated body stores (for a well-nourished woman) that the
average difference represents. In study 1, the difference in the
change in milk volume from 4 to 6 mo between the EBF and
SF groups was ⬃110 mL/d, and the difference in milk energy
output was ⬃92 kcal/d (385 kJ/d). This represents ⬃3% of the
RDA of 2700 kcal/d. Over the 2-mo period, the total energy
difference would be ⬃5520 kcal (23 MJ) or ⬃4% of body fat
reserves (assuming an initial percent body fat of 25–30%; the
average for women in study 1 was 30%; Perez-Escamilla et al.
1995). In study 2, the difference in the change in milk volume
from 4 to 6 mo between the EBF and SF-M groups was ⬃67
mL/d, and the corresponding difference in milk energy output
was ⬃45 kcal/d (188 kJ/d). As a result, the additional energy
burden of EBF from 4 to 6 mo in study 2 was only 2% of the
RDA and 2% of estimated body fat stores (again, assuming
25–30% body fat; the average for study 2 was 29%). Interest-
ingly, the estimated total energy burdens (5520 and 2700 kcal
in studies 1 and 2, respectively) are in close agreement with
the between-group weight differences of 0.6 and 0.2 kg shown
in Table 1, assuming that the weight lost was nearly all fat (9
kcal/g; 38 kJ/g).
For the other nutrients in Table 2, the additional burden of
EBF from 4 to 6 mo represents 3– 6% of the RDA for vitamin
A, 2–3% of the RDA for calcium and 0.1– 0.2% of the RDA
for iron (the lower end of the range is for study 2, and the
upper end is for study 1). The corresponding percentages of
estimated body stores are 1–2% for vitamin A, 0.1– 0.2% for
calcium and 0.4– 0.7% for iron.
Although the average values for the nutritional burden of
continued EBF are all ⱕ6% of the RDA, there is considerable
variability in breast milk output during this interval and thus
in the nutritional burden for individual mothers. For example,
in study 1 the standard deviation for the change in breast milk
output in the SF group was 124 g/d, or 120% of the mean
change. Using this coefficient of variation, the 95th percentile
for the daily burden of continued EBF for these four nutrients
can be estimated as 266 kcal (1113 kJ), 240
g vitamin A, 101
mg calcium and 0.04 mg iron, which represents 11, 18, 10 and
0.3% of the RDA, respectively.
The nutritional burden of continued EBF may be modified
by differences in the duration of postpartum amenorrhea,
especially for nutrients such as iron. Table 3 shows the per-
centage of women in each study who remained amenorrheic at
6 mo postpartum, after excluding users of oral contraceptives
and those whose menses returned before 18 wk postpartum
(which could not have been influenced by the intervention).
The differences among the three intervention groups in study
1 were not statistically significant (P ⫽ 0.11), although the SF
group tended to be less likely to be amenorrheic. In study 2,
the percentage amenorrheic at 6 mo was significantly lower in
the SF-M group than in the EBF group (even though these two
groups had quite similar rates of amenorrhea in study 1). This
difference decreased slightly over time, with the respective
percentages being 55.6 versus 75.0% at 8 mo (P ⫽ 0.07) and
38.2 versus 53.5% at 10 mo (P ⫽ 0.18). Survival analysis
TABLE 2
Theoretical maternal nutritional burden of exclusive breast
feeding versus breastfeeding plus solid foods given to infants
of Honduran women at 4 – 6 mo postpartum
Energy Vitamin A
1
Calcium
2
Iron
3
kcal
5
gmg
Study 1 (primiparous
mothers)
4
Additional daily output 92 74 31 0.033
Percent RDA 3 6 3 0.2
Total 4–6 mo 5520 4440 1860 2.0
Percent body stores ⬃4
6
2 0.2 0.7
Study 2 (mothers of
LBW infants)
7
Additional daily output 45 45 19 0.02
Percent RDA 2 3 2 0.1
Total 4–6 mo 2700 2700 1140 1.2
Percent body stores ⬃2 1 0.1 0.4
1
Assumes milk concentration of 670
g/L; RDA ⫽ 1300
g/d; body
stores 209 mg.
2
Assumes milk concentration of 280 mg/L; RDA ⫽ 1000 mg/d;
body stores 1035 g.
3
Assumes milk concentration of 0.3 mg/L; RDA ⫽ 15 mg/d; body
stores 300 mg.
4
Difference in milk volume between EBF and SF, 110 mL/d.
5
Multiply by 4.184 to convert to kJ.
6
Assuming an initial percent body fat of 25–30%.
7
Difference in milk volume between EBF and SF, 67 mL/d.
Abbreviations: LBW ⫽ low birth weight; RDA ⫽ Recommended
Dietary Allowances.
TABLE 3
Maternal amenorrhea at 6 mo postpartum in lactating
Honduran women
1
EBF SF-M SF P-value
2
Study 1
3
(primiparous mothers)
n 40 35 31 —
Percent amenorrheic 80.0 85.7 64.5 0.11
Study 2 (mothers of LBW
4
infants)
n 45 38 — —
Percent amenorrheic 88.9 68.4 — 0.02
1
Excluding oral contraceptive users and those whose menses re-
turned before 18 wk postpartum.
2
2
test.
3
Study 1 data previously published (Dewey et al. 1997).
4
EBF ⫽ exclusively breastfeeding; LBW ⫽ low birth weight; SF
⫽ solid foods given to infants and breastfeeding ad libitum; SF-M
⫽ solid foods given to infants and maintenance of baseline nursing
frequency.
DEWEY ET AL.264
indicated a significant difference in the duration of amenor-
rhea between the EBF and SF groups in study 2 (median
duration 331 versus 255 d, P ⫽ 0.04) but not in study 1. The
percentage of women still breastfeeding at 12 mo was ⱖ90%
across intervention groups in both studies, and mean breast-
feeding frequency from 6 to 12 mo did not differ significantly
among intervention groups in study 1 (Cohen et al. 1995b) or
in study 2 (14 times per day in both groups).
Infant motor development. Motor development of infants
in the two studies is compared in Table 4. For seven of the
milestones (all except the first two milestones and crawling),
the LBW infants (study 2) were significantly delayed compared
with the infants in study 1. For example, half as many were
walking by 12 mo (22 versus 46%). Table 5 shows the mean
or median age of achievement of each milestone by interven-
tion group within each study. In both studies, there were no
significant differences among intervention groups for the mile-
stones that occurred before the intervention (on average),
indicating that the groups were similar at baseline. In both
studies, infants in the EBF group crawled sooner than infants
in the SF groups, although the difference was only marginally
significant in study 2; in the survival analyses with data from
both studies included (combining the SF and SF-M groups in
study 1), there was a significant difference (P ⫽ 0.007) be-
tween the EBF and SF groups. Crawling occurred, on average,
at ⬃7 mo, 1 mo after the 2-mo intervention period. In study
2, there was also a marginally significant difference between
groups (P ⫽ 0.09) in the age at which the infants were able to
sit, which occurred earlier in the EBF group. In study 1, infants
in the EBF group were more likely to have walked by 12 mo
than infants in the SF groups (P ⫽ 0.07 with three groups; P
⫽ 0.02 with the SF and SF-M groups combined: 60 versus
39%).
Although both studies were randomized trials and there
were no statistically significant differences in initial character-
istics across intervention groups (Cohen et al. 1994, Dewey et
al. 1999), there were slight differences in birth weight, infant
sex and maternal education that may have confounded these
results, particularly in study 1. Within study 1, birth weight
was not significantly correlated with the first four milestones
(including crawling), but the remaining milestones were
achieved earlier in infants with higher birth weights (P
⬍ 0.05). Infant sex was not significantly associated with any of
the milestones. Greater maternal education was associated
TABLE 4
Motor development milestones in a general sample of term
infants (study 1) or a sample of low birth weight, term infants
(study 2) in Honduras
Study 1
Study 2
(LBW)
1
P-
value
2
n 140 108
Raise head 1.2 ⫾ 0.7
3
1.0 ⫾ 0.8 0.02
Raise head and chest 1.9 ⫾ 0.9 1.9 ⫾ 1.5 0.79
Roll over 3.1 ⫾ 1.2 3.8 ⫾ 1.9 0.0003
Crawl 7.0 ⫾ 1.7 (4)
4
7.2 ⫾ 1.8 0.20
Sit, from lying position 6.9 ⫾ 1.3 7.8 ⫾ 1.6 (5) ⬍0.0001
Stand with assistance 7.5
5
(4) 8.5 (10) ⬍0.0001
Pull to stand 8.0 (6) 9.0 (15) ⬍0.0001
Walk with assistance 9.0 (8) 9.5 (16) ⬍0.0001
Stand alone 10.5 (28) 11.0 (38) ⬍0.001
Walk by 12 mo
6
46% (62/134) 22% (22/99) ⬍0.001
1
LBW ⫽ low birth weight.
2
ANOVA for the first three milestones; survival analysis for the
remainder (except walking, tested by
2
).
3
Mean age achieved (mo) ⫾ SD.
4
Number of censored values in survival analyses.
5
Median age achieved (mo), from survival analysis including cen-
sored values.
6
Percentage who walked by 12 mo, excluding dropouts.
TABLE 5
Motor development milestones of Honduran infants by feeding mode at 4–6 mo
Study 1 Study 2 (LBW)
1
Pooled
survival
analysis
2
EBF SF SF-M P-value
3
EBF SF P-value
3
P-value
(EBF vs. SF)
n 49 47 44 — 56 52 — 248
Raise head 1.1 ⫾ 0.7
4
1.3 ⫾ 0.7 1.2 ⫾ 0.6 0.24 1.0 ⫾ 1.0 1.0 ⫾ 0.6 0.74 —
Raise head and chest 1.8 ⫾ 0.8 2.0 ⫾ 1.0 1.8 ⫾ 0.9 0.50 1.9 ⫾ 1.6 1.8 ⫾ 1.3 0.67 —
Roll over 2.9 ⫾ 1.1 3.4 ⫾ 1.4 2.9 ⫾ 1.1 0.15 3.8 ⫾ 2.0 3.8 ⫾ 1.8 0.93 —
Crawl 6.3 ⫾ 1.8 (2)
5
7.3 ⫾ 1.7 (1) 7.2 ⫾ 1.4 (1) 0.02 6.8 ⫾ 1.7 (2) 7.4 ⫾ 1.9 (2) 0.08 0.007
Sit, from lying position 7.0 ⫾ 1.5 (3) 7.0 ⫾ 1.2 (1) 6.8 ⫾ 1.1 (1) 0.68 7.4 ⫾ 1.6 (3) 8.0 ⫾ 1.6 (2) 0.09 0.63
Stand with assistance 7.5
6
(2) 7.5 (1) 7.0 (1) — 8.5 (6) 9.0 (4) — 0.83
Pull to stand 8.0 (3) 8.0 (1) 8.0 (2) — 9.0 (8) 9.0 (7) — 0.90
Walk with assistance 8.5 (4) 9.5 (2) 9.0 (2) — 9.5 (10) 9.5 (6) — 0.82
Stand alone 10.0 (8) 10.5 (7) 10.5 (13) — 11.0 (24) 11.0 (14) — 0.61
Walk by 12 mo
7
60% (28/47) 41% (19/46) 37% (15/41) 0.07 18% (9/50) 27% (13/49) 0.31 —
1
EBF, exclusively breastfeeding; LBW, low birth weight; SF, solid foods given to infants and breastfeeding ad libitum; SF-M, solid foods given
to infants and maintenance of baseline nursing frequency.
2
Studies 1 and 2 combined, SF and SF-M groups combined, controlling for study (see text for description of studies 1 and 2).
3
Analysis of variance for first five milestones;
2
test for walk by 12 mo.
4
Mean age achieved (mo) ⫾ SD.
5
Number of censored values in survival analyses.
6
Median age achieved (mo), from survival analysis, including censored values.
7
Percentage who walked by 12 mo, excluding dropouts.
EFFECTS OF EXCLUSIVE BREASTFEEDING FOR 4 OR 6 MO 265
with the earlier achievement of raising the head and chest but
the later achievement of crawling; it was not a significant
predictor of the other milestones. Controlling for maternal
education in the pooled survival analysis for crawling did not
change the results. To control for birth weight in the analyses
of walking by 12 mo (for study 1 only), we performed a logistic
regression analysis with the independent variables being in-
tervention group (EBF versus combined SF groups), birth
weight (as a continuous variable) and infant sex (because its
inclusion improved the P-values for both group and birth
weight). Intervention group remained significant (P ⫽ 0.05)
in this model.
DISCUSSION
Continued EBF between 4 and 6 mo postpartum led to
significantly greater maternal weight loss in study 1 but not in
study 2. The difference between studies is probably related to
the fact that the net difference between intervention groups in
milk volume, and thus in energy demand, was larger in study
1 than in study 2. Is greater maternal weight loss good or bad?
The difference in the amount of weight lost in study 1 (0.6 kg)
is not large, but for thin women it may be of concern. There
was no interaction between initial maternal BMI and feeding
mode; in other words, weight loss was greater in the EBF group
than the SF group even in the low BMI subgroup in study 1.
On the other hand, the average BMI (in both studies) was not
low, and very few women had a BMI below 19 kg/m
2
(10% in
study 1 and 11% in study 2). It can be argued that the effect
of continued EBF on maternal weight is protective against
maternal obesity in affluent populations and those in transi-
tion but could contribute to maternal depletion in undernour-
ished populations (Adair and Popkin 1992, Winkvist and
Rasmussen 1999). From a public health perspective (consid-
ering both the mother and the infant), in the latter situation
it is probably safer to supplement lactating women than to
advocate a shorter duration of EBF.
The calculated impact of continued EBF from 4 to 6 mo
(compared with breastfeeding plus complementary feeding) on
maternal losses of nutrients other than energy is quite low.
The average daily additional burden is only 3–6% of the RDA
for vitamin A, 2–3% of the RDA for calcium and a minute
fraction of the RDA for iron (because there is very little iron
secreted in breast milk). Of course, many women do not
consume the RDA for certain micronutrients, so these per-
centages would be somewhat higher if based on actual nutrient
intake. These estimates do not imply that deficiencies of such
nutrients are unlikely among lactating women, only that the
risk of deficiency is not appreciably greater in exclusively
breastfeeding women than in those who introduce solid foods
before 6 mo. However, we did not collect data on maternal
micronutrient status, so further research is needed to verify
this conclusion.
Although maternal nutrient losses may be somewhat
greater with exclusive rather than partial breastfeeding, there
is an important tradeoff with regard to the duration of amen-
orrhea. The difference in amenorrhea among intervention
groups in study 1 was not large, but in study 2, the median
duration of amenorrhea was 1.3 mo longer in the EBF group
than in the SF group. This would translate into a “savings” of
15.6 mg of Fe, assuming menstrual losses of 0.4 mg/d (INACG
1981). After subtracting the additional iron losses in milk
attributable to EBF during the 4- to 6-mo interval (Table 2),
this represents a savings of ⬃5% of estimated body stores.
Increased duration of amenorrhea may also result in a longer
interval before the next pregnancy, which allows the mother
to be more fully repleted and more time to care for the infant
before another child is born.
In both studies, infants in the EBF group were reportedly
able to crawl earlier, and in study 1 they were more likely to be
walking by 12 mo (60 versus 39%) than infants in the SF
groups. There are several limitations to this component of the
studies. First, neither the mothers nor the field workers were
blind to group assignment. However, they had no reason to
suspect that there would be differences between intervention
groups (there were no a priori hypotheses regarding these
outcomes), so this should not have biased the results. Second,
no data were collected to validate the mothers’ reports of their
infants’ motor skills. This is a standard practice, but it is
difficult to compare data across studies because the definitions
of the milestones vary considerably. Nevertheless, the average
ages at which infants in study 1 achieved pull to stand, walk
with assistance and walk alone were similar to the 50th per-
centiles of the Denver (Frankenburg and Dodds 1967) and
Bayley (The Psychological Corporation 1969) scales, the val-
ues for crawling and sitting were similar to those reported for
Indonesian (Pollitt et al. 1994) and Pakistani (Yaqoob et al.
1993) infants and the values for walking were similar to those
reported for Pakistani (Yaqoob et al. 1993) and Guatemalan
(Bentley et al. 1997) infants. Furthermore, the fact that there
were highly significant delays in most of the milestones among
the LBW (small-for-gestational age) infants compared with
the infants in study 1, which is consistent with other reports
(Goldenberg et al. 1998), indicates that the method used was
able to capture biologically important differences.
It is noteworthy that crawling typically occurs just after the
4- to 6-mo interval. The mechanism by which EBF during this
interval might affect motor development is unknown. Certain
constituents of breast milk (e.g., docosohexaenoic acid) are
known to be associated with infant mental development (Ko-
letzko and Rodriguez-Palmero 1999, Uauy et al. 1995), but
there is little evidence that they affect motor development.
On the other hand, Vestergaard et al. (1999) reported that
achievement of two motor skills (crawling and pincer grip) was
linked to the duration of breastfeeding in a large sample of
Danish infants, even after adjustment for potentially con-
founding variables. It is thus possible that greater consumption
of breast milk in the EBF groups accounts for our findings,
although the difference in breast milk intake between inter-
vention groups was only 67–110 mL/d. Breastfeeding fre-
quency was similar between intervention groups after 6 mo,
but the volume of breast milk consumed may have continued
to differ for several months after the intervention period.
Other possible mechanisms include lower absorption of mi-
cronutrients by partially breastfed than exclusively breastfed
infants (Bell et al. 1987, Oski and Landaw 1980) or differences
in maternal caregiving or infant motivation to explore the
environment or be upright (Biringen et al. 1995), all of which
could be altered by the amount of time spent nursing. What-
ever the mechanisms for these findings, the differences in
motor development observed may be predictive of later func-
tional outcomes. Although motor development in infancy is
not correlated with later cognitive development in well-nour-
ished populations, Pollitt and Gorman (1990) reported that
motor test scores (although not mental scores) of Guatemalan
infants at 15 mo were significantly associated with several
indices of cognitive performance in adolescence and specu-
lated that this may also be the case in other nutritionally at
risk populations.
In summary, these results indicate that EBF from 4 to 6 mo
postpartum leads to 1) a small but significant difference in
maternal weight loss, 2) little additional maternal nutritional
DEWEY ET AL.266
burden compared with the nutrient demand of breastfeeding
plus complementary feeding, 3) a longer duration of postpar-
tum amenorrhea and 4) earlier development of certain motor
milestones by the infant. The public health implications of
these findings depend on the context; for example, greater
maternal weight loss may be beneficial in affluent populations
but detrimental in malnourished populations. The differences
in motor development may be even larger in situations where
the complementary foods are of poor nutritional and microbi-
ological quality (which was not the case in these two studies).
Taken together with previously reported results (Brown et al.
1998), these results support the conclusion that in most pop-
ulations, the advantages of EBF during this interval are likely
to outweigh any potential disadvantages.
LITERATURE CITED
Adair, L. S. & Popkin, B. M. (1992) Prolonged lactation contributes to depletion
of maternal energy reserves in Filipino women. J. Nutr. 122: 1643–1655.
American Academy of Pediatrics Work Group on Breastfeeding (1997) Breast-
feeding and the use of human milk. Pediatrics 100: 1035–1039.
Bell, J. G., Keen, C. L., & Lonnerdal, B. (1987) Effect of infant cereals on zinc
and copper absorption during weaning. Am. J. Dis. Child. 141: 1128 –1132.
Bentley, M. E., Caulfield, L. E., Ram, M., Santizo, M. C., Hurtado, E., Rivera, J. A.,
Ruel, M. T., & Brown, K. H. (1997) Zinc supplementation affects the activity
patterns of rural Guatemalan infants. J. Nutr. 127: 1333–1338.
Biringen, Z., Emde, R. N., Campos, J. J. & Appelbaum, M. I. (1995) Affective
reorganization in the infant, the mother, and the dyad: The role of upright
locomotion and its timing. Child Dev. 66: 499 –514.
Brown, K. H., Dewey, K. G. & Allen, L. H. (1998) Complementary Feeding of
Young Children in Developing Countries: A Review of Current Scientific
Knowledge. WHO/NUT/ 98 1: World Health Organization, Geneva.
Cohen, R. J., Brown, K. H., Canahuati, J., Landa Rivera, L. & Dewey, K. G. (1994)
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 344: 288 –293.
Cohen, R. J., Brown, K. H., Canahuati, J., Landa Rivera, L. & Dewey, K. G.
(1995a) Determinants of growth from birth to 12 months among breast-fed
Honduran infants in relation to age of introduction of complementary foods.
Pediatrics 96: 504–510.
Cohen, R. J., Brown, K. H., Landa Rivera, L., & Dewey, K. G. (1999) Promoting
exclusive breastfeeding for 4– 6 months in Honduras: Attitudes of mothers
and barriers to compliance. J. Hum. Lact. 15: 9 –18.
Cohen, R. J., Haddix, K., Hurtado, E. & Dewey, K. G. (1995c) Maternal activity
budgets: Feasibility of exclusive breastfeeding for six months among urban
women in Honduras. Soc. Sci. Med. 41: 527–536.
Cohen, R. J., Landa Rivera, L., Canahuati, J., Brown, K. H. & Dewey, K. G.
(1995b) Delaying the introduction of complementary food until 6 months
does not affect appetite or mother’s report of food acceptance of breastfed
infants from 6 to 12 months in a low-income, Honduran population. J. Nutr.
125: 2787–2792.
Dewey, K. G., Cohen, R .J., Brown, K. H., & Landa Rivera, L. (1999) Age of
introduction of complementary food and growth of term, low birth weight
breastfed infants: A randomized intervention study in Honduras. Am. J. Clin.
Nutr. 69: 679– 686.
Dewey, K. G., Cohen, R. J., Landa Rivera, L. & Brown, K. H. (1998a) Effects of
age of introduction of complementary foods on iron status of breastfed infants
in Honduras. Am. J. Clin. Nutr. 67: 878 – 884.
Dewey, K. G., Cohen, R. J., Landa Rivera, L., & Brown, K. H. (1998b) Effects
of age of introduction of complementary foods on micronutrient status of term
low-birthweight, breastfed infants in Honduras. FASEB J. 12: A648.
Dewey, K. G., Cohen, R. J., Landa Rivera, L., Canahuati, J. & Brown, K. H. (1997)
Effects of age of introduction of complementary foods to breastfed infants on
duration of lactational amenorrhea in Honduran women. Am. J. Clin. Nutr. 65:
1403–1409.
Frankenburg, W. K. & Dodds, J. B. (1967) The Denver Developmental Screen-
ing Test. J. Pediatr. 71: 181–191.
Frongillo, E. A., & Habicht, J. P. (1997) Investigating the weanling’s dilemma:
Lessons from Honduras. Nutr. Rev. 55: 390 –395.
Goldenberg, R. L., Hoffman, H. J., & Cliver, S. P. (1998) Neurodevelopmental
outcome of small-for-gestational-age infants. Eur. J. Clin. Nutr. 52: S54–S58.
INACG (1981) Iron Deficiency in Women. Nutrition Foundation, Washington,
D.C.
Institute of Medicine (1991) Nutrition During Lactation. National Academy
Press, Washington, D.C.
Koletzko, B., & Rodriguez-Palmero, M. (1999) Polyunsaturated fatty acids in
human milk and their role in early infant development. J. Mamm. Gland Bio.
Neoplasia 4: 269–284.
McDade, T. W. & Worthman, C. M. (1998) The weanling’s dilemma reconsid-
ered: A biocultural analysis of breastfeeding ecology. J. Dev. Behav. Pediatr.
19: 286 –299.
Oski, F. A., & Landaw, S. A. (1980) Inhibition of iron absorption from human
milk by baby food. Am. J. Dis. Child. 134: 459 – 460.
Perez-Escamilla, R., Cohen, R. J., Brown, K. H., Landa Rivera, L., Canahuati, J. &
Dewey, K. G. (1995) Maternal anthropometric status and lactation perfor-
mance in a low-income Honduran population: Evidence for the role of the
infant. Am. J. Clin. Nutr. 61: 528 –534.
Pollitt, E. & Gorman, K. (1990) Long term developmental implications of motor
maturation and physical activity in infancy in a nutritionally at risk population.
In: Activity, Energy Expenditure and Energy Requirements of Infants and
Children (Schurch, B. & Scrimshaw, N.S., eds.), pp. 279 –296, Nestle Foun-
dation, Lausanne, Switzerland.
Pollitt, E., Husaini, M.A., Harahap, H., Halati, S., Nugraheni, A. & Sherlock, A. O.
(1994) Stunting and delayed motor development in rural west Java. Am. J.
Hum. Biol. 6: 627– 635.
Pollock, M. L., Laughridge, E. E., Coleman, B., Linnerud, A. C. & Jackson, A.
(1975) Prediction of body density in young and middle-aged women.
J. Appl. Physiol. 38: 745–749.
SAS Institute Inc. (1987) SAS-PC Software, version 6.03. SAS Institute, Cary,
NC.
The Psychological Corporation (1969) The Bayley Scales of Infant Develop-
ment, New York.
Uauy, R. & De Andraca, I. (1995) Human milk and breast feeding for optimal
mental development. J. Nutr. 125: S2278 –S2280.
UNICEF (1999) Breastfeeding: Foundation for a Healthy Future. UNICEF, New
York.
Vestergaard, M., Obel, C., Henriksen, T. B., Sorensen, H. T., Skajaa, E., &
Ostergaard, J. (1999) Duration of breastfeeding and developmental mile-
stones during the latter half of infancy. Acta Paediatr. 88: 1327–1332.
Winkvist, A., & Rasmussen, K. M. (1999) Impact of lactation on maternal body
weight and body composition. J. Mamm. Gland Bio. Neoplasia 4: 309 –318.
World Health Organization (WHO) (1995b) The World Health Organization’s
infant-feeding recommendation. WHO Weekly Epidemiological Record 17:
117–220.
Yaqoob, M., Ferngren, H., Jalil, F., Nazir, R. & Karlberg, J. (1993) Early child
health in Lahore, Pakistan: XII. Milestones. Acta Paediatr. 390(suppl):151–
157.
EFFECTS OF EXCLUSIVE BREASTFEEDING FOR 4 OR 6 MO 267
