![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Departamento de Nutrição, Universidade Federal de Pelotas and * Departamento de Medicina Social, Universidade Federal de Pelotas, Pelotas, Brasil
1To whom correspondence should be addressed. E-mail: denise{at}ufpel.tche.br.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2500 g) underwent measurements for weight, height, waist,
hip and arm circumferences, triceps and subscapular skinfolds. The
following indices were calculated in 312 women: body mass index,
waist/hip ratio, arm fat area, the percentage of body fat assessed
through skinfolds, and weight and body mass index change since before
conception. The percentage of body fat was also measured through
bioimpedance for half of the sample. After adjustment for confounding,
all outcomes generally showed a similar pattern, i.e., mothers who
breast-fed for 6–11.9 mo had lower measurements than those with
shorter or longer durations. However, only the association with
bioimpedance was significant (P < 0.03), and that
for arm fat area tended to be significant (P = 0.06). Exclusive or predominant breastfeeding at 4 mo was associated
with lower waist circumference (P = 0.05) and the
percentage of body fat measured through skinfolds (P
= 0.04). This study suggests that the relationship between
breast-feeding and long-term changes in maternal weight is
complex and, in this population, not particularly strong.
KEY WORDS: • breast-feeding • lactation • body composition • anthropometry • mother nutrition
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In Brazil, the percentage of women with a body mass
index (BMI) > 30 kg/m2 increased from 8.2
in 1974 to 13.3% in 1989, an increase of 60% (2)
. The
reasons behind this secular trend include socioeconomic improvement,
but may also be affected by a number of other social, environmental and
cultural factors including reproductive performance. Changes in parity,
birth intervals and breast-feeding duration may have played a role.
It is important to differentiate studies that have addressed weight
gain during the whole reproductive cycle (net change in comparison with
prepregnancy weight) and postpartum weight retention studies (change in
weight after delivery). The literature on the effects of
breast-feeding on maternal anthropometry is controversial. Of 14
previous studies on this subject, nine assessed postpartum weight
retention and five weight gain during the reproductive cycle. Four
showed that breast-feeding significantly reduces weight
(3
4
5
6)
and seven others did not find an association
(7
8
9
10
11
12
13)
. Paradoxically, two studies reported weight loss
restricted to nonlactating women (14
,15)
and one showed an
increase in the weight of women who breast-fed for >2 mo
(16)
. Part of these differences may be due to
inconsistencies in breast-feeding definition, as well as to the
variable duration of follow-up. Seven of these studies followed up
mothers for >6 mo, and only one did so for 24 mo. Studies of short
duration generally did not show a greater effect of breast-feeding
on weight loss than those with a longer follow-up. No published
studies are available in which women were followed for >2 y.
This study was designed to investigate the effect of breast-feeding on maternal weight and on several other anthropometric indicators of fat tissue, 5 y after delivery. Its goal is to answer the most important question from a Public Health standpoint, namely, does breast-feeding offer protection against obesity in the long term and not just for a few months after delivery?
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Hospital interviews were carried out with all mothers. There were only 16 refusals (0.3%). Mothers provided information on prepregnancy weight and were weighed just before delivery and on d 1 postpartum.
A systematic sample of 1460 children, comprising all low-birth-weight
infants plus a 20% sample of the remainder, was selected for
follow-up at 6 and 12 mo of age. At the latter home visit, 1363
children (93.4%) were examined and their mothers interviewed. In 1997,
an attempt was made to locate these 1363 children and 1273 (93.4%
again) were traced, with a cumulative rate of losses to follow- up
since birth of 12.8%. After excluding twins and children who did not
live with their biological mothers, 1236 mothers were available for
interview. The last stage of the research consisted of an additional
home visit to mothers who satisfied the study’s inclusion criteria,
i.e., having delivered a baby with a birth weight 2500 g (358
exclusions), not smoking (193), and not having been pregnant since the
delivery of the index child (308). Of 377 eligible women, 363 were
located from June to October 1998 (96.3%), and a further 51 were
excluded due to a new pregnancy or for having started smoking,
resulting in the final sample of 312 women. For some of the variables,
the total numbers available for analysis were smaller because some
women refused to participate in specific measurements (particularly
subscapular skinfold for which there were 12 refusals).
Data collection was carried out by four University-trained
nutritionists. After training in anthropometric methods according to
the procedures recommended by Lohman et al. (17)
,
standardization sessions were carried out on 10 women. Intra- and
interobserver technical errors of measurements (TEM) were calculated
(18)
. Two months after the initial training,
standardization was repeated. All measurements in the field were
carried out separately by two anthropometrists. When differences
exceeded 2.8 times the mean interobserver TEM, measurements were
repeated; if the difference persisted, a third and final set of
measurements was carried out. The mean of the values obtained by the
two anthropometrists was taken as the final value.
The following anthropometric measures were obtained: weight (using a UNISCALE digital electronic scale with capacity of 150 kg and precision of 0.1 kg; UNICEF, Copenhagen, Denmark); height (using a locally developed portable aluminum height meter, with a precision of 1 mm); arm, waist and hip circumferences (using a nonextensible tape 6 mm in width and 2 m in length; CMS, London, UK); triceps and subscapular skinfolds (Holtain skinfold meter; CMS).
Interobserver technical errors of measurement were as follows: 20 g for weight; 0.18 cm for height; 0.23 cm for arm; 0.56 cm for waist and 0.33 cm for hip circumferences; 0.58 mm for triceps and 0.71 for subscapular skinfolds.
The anthropometric indices used in this analyses were the waist-hip
ratio, waist circumference, arm fat index and BMI
(kg/m2). The percentage of body fat was measured
through bioimpedance (Tanita Bodyfat Analyzer model TBF-305; Tanita,
Tokyo, Japan) and based on Siri’s equations derived from skinfold
thickness, as adapted by 19
.
Change in nutritional status was assessed by comparing the current weight of the women with their reported prepregnancy weight, and calculating changes in weight and BMI. For the BMI, the height measure was obtained in the 1998 survey.
Information on breast-feeding duration were collected at 6 mo (for
children weaned before this age) and 12 mo (for the remaining
children). Breast-feeding patterns were classified according to
Labbok and Krasovec (20)
.
Several confounding factors were measured. In the hospital
questionnaire, information was collected on family income (in minimum
wages per month; ordinal variable with five categories); education (in
years of schooling; ordinal variable with four categories); age (5-y
groups); skin color (white or nonwhite); marital status (single or
married); parity (number of children had before the index
pregnancy); weight gain during pregnancy based on the difference
between recalled prepregnancy weight and that measured in the hospital
upon admission (in four groups); prepregnancy BMI (in four groups);
prepregnancy weight (<49 kg). In the 1998 interview, information was
collected on ownership of household items (a score built from ownership
of radio, television or refrigerator, for example; ordinal variable
with five categories); employment (not working; paid work at home;
working outside the home); physical exercise in the last year (yes or
no); number of hours of sleep per night (in quartiles); use of oral
contraceptives (yes or no); number of daily meals (discrete); use of
alcohol in the last week (yes or no); use of maté tea in the last
week (yes or no); special diets (none, weight-reducing, for weight
gain); and divorce since the child was born (yes or no). The
questionnaire developed by Block et al. (21)
was used to
estimate dietary intake of fats (in five categories) and fibers (in
three categories).
In all phases of data collection, 5% of the interviews were repeated by a supervisor for quality control. Data were double entered using the Epi-Info software (Centers for Disease Control, Atlanta, GA) and checked for range and consistency.
The bivariate analyses included the comparison of the mean values
of anthropometric indicators according to breast-feeding duration
and pattern, using ANOVA. Multivariate analyses included confounding
factors that presented some degree of association (P < 0.20) (22)
with both the anthropometric outcomes and the
breast-feeding variables. The Stata (College Station, TX)
package was used for carrying out the backward elimination
method for multiple linear regression.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In 1998, very low BMI (<18.5 kg/m2) was
observed in <1% of the women, but 17.7% were obese (BMI 30
kg/m2). All variables included in Table 1
, except
marital status, were significantly associated with BMI.
|
. The women gained 
5 kg during this period, and the BMI increased by
2 kg/m2.
|
. For all but one index, the mean values were higher for women who
breast-fed for <1 mo and tended to decrease with longer
breast-feeding duration, but increased again in women who
breast-fed for >12 mo. The only exception was arm fat index, which
did not increase in the last group; this variable showed an association
that tended to be significant (test for linear trend: P
= 0.09) with breast-feeding duration.
|
were or tended to be
significant. Values were lower for mothers who breast-fed
exclusively or predominantly for 4 mo.
The associations between confounding variables, the two
breast-feeding measures (pattern and duration) and the eight
anthropometric indices are shown in Table 4
. For each of the 16 combinations of the two breast-feeding
variables (exposure) and anthropometric (outcome) variables, all
potential confounders associated in a crude analysis with both
variables (P < 0.2) were selected; these variables are
marked with "X" in Table 4
. They were then included in a backward
elimination multiple linear regression for the anthropometric outcome,
and again only those with P < 0.2 were retained (X*).
Potential confounders that were not associated in the crude analyses
with both exposure and outcome were not listed in Table 4
(i.e.,
marital status; use of oral contraceptives; prepregnancy BMI; number of
daily meals; special diets; use of maté tea; use of alcohol).
|
6
7
8
. The regression coefficients in the body of the tables indicate the
change in the anthropometric outcomes associated with
breast-feeding duration, relative to mothers who breast-fed for
12 mo, and with breast-feeding pattern, relative to mothers who
breast-fed exclusively or predominantly at 4 mo.
|
|
|
|
). The association with BMI was not significant, but the test for linear
trend with the percentage of fat mass was significant (P
= 0.03). For breast-feeding pattern, there was a borderline
linear trend (P = 0.05) in the adjusted analysis.
Associations with the other anthropometric outcomes are shown in
Tables 6
7
8
. In the adjusted analyses, there were no significant associations with
breast-feeding duration. For breast-feeding pattern, most crude
associations were of borderline significance (0.05 < P
< 0.1) but became weaker after adjustment. For the percentage of
body mass assessed through skinfolds (Table 8)
the crude association
was significant; after adjustment, however it became borderline
(P = 0.06). For all outcomes, mothers who
breast-fed for <1 mo or >12 mo tended to have the highest values;
those breast-feeding for 6–11.9 mo had the lowest. Regarding
breast-feeding pattern, those who breast-fed exclusively or
predominantly had consistently lower values.
Because the confounding variable weight gain during pregnancy may be considered as a component of the variables total weight gain and BMI gain, the analyses of these two outcomes were repeated without adjusting for pregnancy weight gain; however, the results remained unchanged.
When breast-feeding pattern at 4 mo was recoded as a dichotomous variable (exclusive or predominant vs. partial or weaned) the following adjusted regression coefficients and P-levels were observed: weight gain (ß = 1.05 kg; P = 0.30), percentage of body fat through impedance (ß = 4.04 percentage points; P = 0.059), waist circumference (ß = 3.30 cm; P = 0.049), waist-hip ratio (ß = 0.01; P = 0.28), BMI (ß = 1.18 kg/m2; P = 0.082), BMI gain (ß = 0.53 kg/m2; P = 0.20), percentage of body fat through skinfold thickness (ß = 1.45 percentage points; P = 0.043) and arm fat index (ß = 1.19 percentage points; P = 0.31). Therefore, for all variables studied, mothers who breast-fed exclusively or predominantly were thinner than those who breast-fed partially or not at all, but only two differences were significant and two others were borderline (P = 0.059 and 0.082).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The study was specifically designed to address the issue of breast-feeding and maternal anthropometry. Therefore, mothers who became pregnant again, who smoked, or whose children had a low birthweight were excluded because these three factors are strongly related to both maternal nutrition and lactation. Other factors less strongly associated to both were treated as confounding variables in the analyses. Despite the fact that weight retention (relative to prepregnancy weight) was studied, adjustment for weight gain during pregnancy means that the study can also be interpreted as a postpartum weight change study. It is reassuring that weight gain during pregnancy was not an important confounding factor because, as noted above, such adjustment did not affect the results for BMI or weight gain.
A possible limitation of this study is the use of reported prepregnancy
weight. However, several authors have shown a high correlation between
reported and measured weight (23
,24)
, including in
Brazilian samples (25
,26)
. Also, the average weight gain
over time, 1 kg/y, is consistent with other studies
(27
,28)
. The mean BMI based on reported weight was 23.9
kg/m2 in 1993, whereas in a population-based
survey of Pelotas in 1994 (29)
that included standardized
measurement of women, the projected value for the same mean age was
24.5 kg/m2.
The mean BMI in 1993 (23.9 kg/m2) was close to
the cut-off of 25.0 kg/m2 that indicates an
increased risk for chronic diseases (1)
. Similarly, the
mean waist circumference (82.1 cm) and waist/hip ratio (0.81) were
close to the corresponding cut-offs (1)
. The
percentage of body fat estimates according to impedance (39%) or to
arm fat area (40%) were very similar, but that measured through
skinfolds was 7% lower. Relative to NHANES-II, except for
triceps skinfold and arm fat area, Pelotas women had larger indices, by
up to 20% compared with women with the same mean age (obtained through
a regression approach from the original publication (19)
.
Therefore, the study sample does not present evidence of malnutrition;
on the contrary, it is closer to overweight.
These findings did not show a linear association between
breast-feeding duration and body size and composition 5 y
after delivery. For most anthropometric indicators, there was a
U-shaped curve in which mothers who breast-fed for 6–11.9 mo
had the lowest body size, and those who breast-fed for <1 mo or
for 12 mo the largest. Most of the adjusted analyses were not
significant. The two exceptions were a significant (P = 0.03) linear trend for the percentage of body fat measured through
impedance and a borderline (P = 0.06) linear trend for
the arm fat index, both of which decreased with longer
breast-feeding duration, but still showed a slight increase after
12 mo.
The initial analyses of breast-feeding pattern at 4 mo showed a significant linear trend (P = 0.05) only with impedance. For most variables, women who breast-fed exclusively or predominantly 4 mo after delivery were thinner than those who breast-fed partially or not at all. A regrouping of these categories showed significant results for two variables and borderline results (P = 0.06 and 0.08) for two others. These analyses have to be interpreted with caution because the recoding of breast-feeding pattern took place after examining the data.
In summary, although breast-feeding may help reduce weight retention 5 y after delivery in this group, the results are not clear cut, and the most beneficial duration of breast-feeding appears to be 6–11.9 mo.
The possibility of reverse causality cannot be ruled out. Mothers who
lost a substantial amount of weight by breast-feeding 6–11.9 mo
may have stopped then, whereas those who were still fat continued for
>12 mo. Also, prolonged breast-feeding may be associated with a
lesser degree of concern about body image and therefore with less
effort to reduce weight after delivery. Studies from the United Kingdom
(30)
showed that women who were concerned about their body
shape were less likely to breast-feed. Ethnographic studies are
required to investigate this possibility.
During y 1 of breast-feeding, the extra energy expenditure
associated with lactation is compensated at least in part by increased
food intake, as shown by several studies
(5
,6
,12
,14
,31
,32)
. We were unable to find any published
studies on energy intake for lactating and nonlactating women during y
2 after delivery. If energy intake remains higher and the amount of
breast-milk produced decreases, as is normally the case during y 2
of lactation, then breast-feeding for >12 mo could lead to weight
gain. This would be compatible with the findings of the present study
in which breast-feeding for 6–11.9 mo was associated with the
lowest weight retention.
This study suggests that the relationship between breast-feeding and long-term weight retention is complex and, in this population, not particularly strong.
Manuscript received May 25, 2000. Initial review completed July 23, 2000. Revision accepted September 28, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1. World Health Organization Obesity preventing and managing the global epidemic 1997 Report of a WHO Consultation on Obesity WHO, Geneva, Switzerland.
2. Monteiro C. A., Mondini L., Souza A.L.M., Popkin B. M. Da desnutrição para a obesidade: a transição nutricional no Brasil. Monteiro C. A. eds. Velhos e Novos Males de Saúde no Brasil 1995 Editora Hucitec NUPENS/USP São Paulo, Brasil.
3.
Dewey K. G., Heinig M. J., Nommsen L. A. Maternal weight-loss patterns during prolonged lactation. Am. J. Clin. Nutr. 1993;58:162-166
4.
Janney C. A., Zhang D., Sowers M. F. Lactation and weight retention. Am. J. Clin. Nutr. 1997;66:1116-1124
5. Kramer F. M., Stunkard A. J., Marshall K. A., McKinney S., Liebschutz J. Breast-feeding reduces maternal lower-body fat. J. Am. Diet. Asoc. 1993;93:429-433[Medline]
6. Motil K. J., Sheng H. P., Kertz B. L., Montandon C. M., Ellis K. J. Lean body mass of well-nourished women is preserved during lactation. Am. J. Clin. Nutr. 1998;67:292-300[Abstract]
7.
Brewer M. M., Bates M. R., Vannoy L. P. Postpartum changes in maternal weight and body fat depots in lactating vs nonlactating women. Am. J. Clin. Nutr. 1989;49:259-265
8. Butte N. F., Hopkinson J. M., Ellis K. J., Wong W. W., Smith E. O. Changes in fat-free mass and fat mass in postpartum women: a comparison of body composition models. Int. J. Obes. 1997;21:874-880
9. Dugdale A. E., Eaton-Evans J. The effect of lactation and other factors on post-partum changes in body-weight and triceps skinfold thickness. Br. J. Nutr. 1989;61:149-153[Medline]
10. Öhlin A., Rössner S. Maternal body weight development after pregnancy. Int. J. Obes. 1990;14:159-173
11. Schauberger C. W., Rooney B. L., Brimer L. M. Factors that influence weight loss in the puerperium. Obstet. Gynecol. 1992;79:424-429[Medline]
12.
Van Raaij J.M.A., Schonk C. M., Vermaat-Miedema S. H., Peek M.E.M., Hautvast J.G.A.G. Energy cost of lactation and energy balances of well-nourished Dutch lactating women reappraisal of the extra energy requirements of lactation. Am. J. Clin. Nutr. 1991;53:612-619
13. Walker L. O., Freenland-Graves J. Lifestyle factors related to postpartum weight gain and body image in bottle- and breastfeeding women. J. Obstet. Gynecol. Neonatal Nurs. 1998;27:151-160[Medline]
14. Chou T., Chan G. M., Moyer-Mileur L. Postpartum body composition changes in lactating and non-lactating primiparas. Nutrition 1999;15:481-484[Medline]
15. Potter S., Hannum S., McFarlim B., Essex-Sorlie D., Campbell E., Trupin S. Does infant feeding method influence maternal postpartum weight loss?. J. Am. Diet. Assoc. 1991;91:441-446[Medline]
16. Rookus M. A., Rokebrand P., Burema J., Deurenberg P. The effect of pregnancy on the body mass index 9 months post partum in 49 women. Int. J. Obes. 1987;11:609-618[Medline]
17. Lohman T. G., Roche A. F., Martorell R. Anthropometric Standardization Reference Manual 1988 Human Kinetics Books Champaign, IL.
18. Mueller W. H., Malina R. M. Relative reliability of circumferences and skinfolds as measures of body fat distribution. Am. J. Phys. Anthropol. 1987;72:437-439[Medline]
19. Frisancho A. R. Anthropometric Standards for the Assessment of Growth and Nutritional Status 1990 The University of Michigan Press Ann Arbor, MI.
20. Labbok M., Krasovec K. Toward consistency in breastfeeding definitions. Stud. Fam. Plann. 1990;21:226-230[Medline]
21. Block G., Clifford C., Naughton M. D., Henderson M., Mcadams M. A brief dietary screen for high fat intake. J. Nutr. Educ. 1989;21:199-207
22.
Maldonado G., Greenland S. Simulation study of confounder-selection strategies. Am. J. Epidemiol. 1993;138:923-936
23. Stevens-Simon C., Mcanarney E., Coulter M. How accurately do pregnant adolescents estimate their weight prior to pregnancy?. J. Adolesc. Health Care 1986;7:250-254[Medline]
24.
Stewart A., Jackson R., Ford M., Beaglehole R. Underestimation of relative weight by use of self-reported height and weight. Am. J. Epidemiol. 1987;125:122-126
25. Chor D., Coutinho E.S.F., Laurenti R. Reliability of self-reported weight and height among state bank employees. Rev. Saúde Pública 1999;33:16-23
26. Schmidt M., Duncan B., Tavares M., Polanczyk C., Pellanda L., Zimmer P. Validity of self-reported weight—a study of urban Brazilian adults. Rev. Saúde Pública 1993;27:271-276
27. Coitinho D. C. A influência da história reprodutiva no índice de massa corporal de mulheres brasileiras 1998 Doctoral thesis Faculdade de Saúde Pública Universidade de São Paulo, São Paulo, Brasil.
28.
Smith D. E., Lewis C. E., Caveny J. L., Perkins L. L., Burke G. L., Bild D. E. Longitudinal changes in adiposity associated with pregnancy. J. Am. Med. Assoc. 1994;271:1747-1751
29. Gigante D. P., Barros F. C., Post C.L.A., Olinto M.T.A. Prevalência de obesidade em adultos e seus fatores de risco. Rev. Saúde Pública 1997;3:236-246
30. Foster S. F., Slade P., Wilson K. Body image maternal fetal attachment. and breast feeding. J. Psychosom. Res. 1996;41:181-184[Medline]
31.
Goldberg G. R., Prentice A. M., Coward W. A., Davies H. L., Murgatroyd P. R., Sawyer M. B., Ashford J., Black A. E. Longitudinal assessment of the components of energy balance in well-nourished lactating women. Am. J. Clin. Nutr. 1991;54:788-798
32. Bradshaw M. K., Pfeiffer S. Feeding mode and anthropometric changes in primiparas. Hum. Biol. 1988;60:251-261[Medline]
This article has been cited by other articles:
|
|
 |
|
  E. P. Gunderson Breastfeeding After Gestational Diabetes Pregnancy: Subsequent obesity and type 2 diabetes in women and their offspring Diabetes Care, July 1, 2007; 30(Supplement_2): S161 - S168. [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Stuebe, J. W. Rich-Edwards, W. C. Willett, J. E. Manson, and K. B. Michels Duration of Lactation and Incidence of Type 2 Diabetes JAMA, November 23, 2005; 294(20): 2601 - 2610. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Kac, M. H. Benicio, G. Velasquez-Melendez, J. G Valente, and C. J Struchiner Breastfeeding and postpartum weight retention in a cohort of Brazilian women Am. J. Clinical Nutrition, March 1, 2004; 79(3): 487 - 493. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
