| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Division of Nutritional Sciences, Cornell University, Ithaca, NY (JLB and KMR); the Department of Human Nutrition and KVL Centre for Food Research, The Royal Veterinary and Agricultural University, Frederiksberg, Denmark (KFM); the Danish Epidemiology Science Centre, Institute of Preventive Medicine, Copenhagen University Hospital, Copenhagen (TIAS)
2 Data collection supported by the Danish National Research Foundation, Danish Pharmaceutical Association, Ministry of Health, National Board of Health, Statens Serum Institut, BIOMED, March of Dimes Birth Defects Foundation, Danish Heart Association, Danish Medical Research Council, and Sygekassernes Helsefond. Data analysis supported by Hatch grant NYC399405 (to KMR) and by a grant from the Einaudi Center at Cornell University (to JLB). JLB was supported by NIH training grant HD07331 (to KMR).
3 Address reprint requests to TIA Sørensen, Institute of Preventive Medicine, Øster Søgade 20, 1, DK-1399 Copenhagen K, Denmark. E-mail: tias{at}ipm.hosp.dk.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: We examined how maternal prepregnant body mass index (BMI; in kg/m2) and infant feeding pattern are associated with infant weight gain.
Design: In this prospective, observational study, we used multiple regression analyses adjusted for potential confounding factors to examine these associations among 3768 mother-infant dyads from the Danish National Birth Cohort.
Results: In multiple regression analyses, increasing maternal prepregnant BMI, decreasing durations of breastfeeding, and earlier complementary food introduction were associated with increased infant weight gain. An interaction was identified for short durations of breastfeeding (<20 wk). Earlier complementary food introduction (<16 wk) was associated with greater infant weight gain; however, the timing of complementary food introduction did not increase infant weight gain at longer durations of breastfeeding (
20 wk). In this sample, prepregnant obesity (BMI 30.0), short durations of breastfeeding, and earlier introduction of complementary food were associated with 0.7 kg of additional weight gain during infancy.
Conclusions: Infant weight gain is associated with maternal prepregnant BMI and with an interaction between the duration of breastfeeding and the timing of complementary food introduction. Future investigations of the effects of breastfeeding on infant weight gain should account for all of these factors.
Key Words: Breastfeeding • maternal nutrition • body mass index • pregnancy • infant nutrition • infant growth • Danish National Birth Cohort
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
It is well-established that human milk is the optimal form of infant nutrition. Breastfeeding confers immunologic, psychological, and developmental benefits to the infant (8, 9). Breastfed infants grow differently during the first year of life than do those fed infant formula. Longer durations of breastfeeding are consistently associated with infants who are lighter and shorter at 1 y of age than are infants who have shorter durations of breastfeeding or were not breastfed at all (10). Furthermore, a growing body of evidence shows that breastfeeding is associated with long-term benefits for the infant, such as a reduced risk of developing overweight and obesity during childhood (11–13). After the early cessation of breastfeeding, an infant is usually fed infant formula. Because formula-fed infants are often introduced to complementary foods earlier than are their breastfed counterparts, these infants are exposed to a different feeding pattern than are infants who continue to breastfeed (14, 15). This combination of short periods of breastfeeding and early introduction of complementary food may contribute to the altered growth patterns seen in nonbreastfed infants.
Despite the association between excess maternal prepregnant weight or high body mass index (BMI; in kg/m2) and a shortened duration of breastfeeding, few researchers have accounted for this association in their analyses of the effects of breastfeeding on infant growth. Investigators found that women who were obese before pregnancy had heavier infants at 1 y than did women who were not obese (16). However, the authors of this study did not examine how the infants were fed during the first year of life (16). Conversely, in other investigations of associations between breastfeeding duration and infant weight, researchers have not accounted for the effects of maternal prepregnant weight (11, 17). Given the known association between high maternal prepregnant BMI and unsuccessful initiation and shortened durations of breastfeeding, it is likely that maternal prepregnant BMI is also an independent determinant of infant weight gain at 1 y, but this cannot be determined from these studies (5–7, 11, 17).
Given the negative association between maternal prepregnant BMI and breastfeeding, and thus its association with alternate infant feeding methods, the objective of this analysis was to determine how maternal prepregnant BMI, duration of any breastfeeding, and infant age at the time of complementary food introduction affect infant weight gain from birth to 1 y. This objective was addressed in a contemporary national cohort of Danish women and their infants.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
12 wk and 26 wk of gestation and at 6 and 18 mo postpartum. Data about the infant’s birth was gathered via linkage to the Danish National Patient Registry with the use of the subjects’ Civil Person’s Registry numbers. This study was approved by both the University Committee on Human Subjects at Cornell University and the Danish Data Protection Agency (18).
Selection criteria
Because we used a sample from the early data collection phase of the Danish National Birth Cohort, 5845 subjects were available for our analyses. Of these subjects, we limited our analyses to breastfeeding mother-infant dyads without conditions that are known to affect growth during the first year of life. We excluded subjects whose infants were born weighing <2500 g (n = 139), were born at <37 wk gestation (n = 144), or had illnesses or conditions (such as congenital heart defects) expected to negatively affect growth (as determined by KFM) (n = 99). Subjects were also excluded if they were aged <18 y (n = 2), never breastfed (n = 65), used an alternative form of breastfeeding (such as donor milk; n = 2), or had any form of diabetes (n = 64). Because <1% of the women had never breastfed their infants, we excluded them from the analyses because they likely represented women with unusual circumstances in this population (ie, had undergone breast surgery or had psychological disturbances). Thus, 5330 subjects were eligible for our study. Infants with missing measures of birth weight (n = 53); with no information available about the 1-y exam (n = 986); with missing weight, length, or age measurements at 1 y (n = 180); or not between 43 and 61 wk of age (52 ± 9 wk) at the time of the 1-y measurements were excluded (n = 124). Dyads in which the mother’s height or weight were unavailable (n = 66) which made the calculation of BMI impossible, were excluded. Furthermore, women for whom data on the duration of breastfeeding (n = 77) or on the age at which complementary food was introduced were missing (n = 76) were also excluded. This resulted in a final sample of 3768 dyads available for these analyses.
Variable descriptions
Maternal variables
Maternal characteristics such as age, level of basic education, occupation, parity, presence of a spouse or partner during early pregnancy, and self-report of prepregnant weight and height were obtained at the 12-wk interview. Maternal education was classified into 2 categories: <10 and 10 y (compulsory level). Maternal occupation was coded by DNBC staff members into categories of jobs requiring a high level of education, management level jobs, skilled labor, unskilled labor, student, or unemployed. The mother’s BMI was calculated by using her reported prepregnant weight and height. For some analyses, BMI values were classified according to the World Health Organization’s criteria (2): underweight (<18.5), normal weight (18.5–24.9), overweight (25.0–29.9), and obese (30.0).
Information about the development of maternal health conditions (such as gestational diabetes), alcohol use, and smoking habits during pregnancy were obtained at the 12- and 26-wk gestation and 6-mo postpartum interviews. For these analyses, women were classified as nondrinkers if they consumed an average of <1 alcoholic beverage/d (in accordance with Danish Government guidelines on alcohol consumption during pregnancy at the time) and as drinkers if they consumed an average of 1 alcoholic beverage/d during pregnancy. Women were categorized as nonsmokers if they never smoked or quit during early pregnancy and as smokers if they smoked throughout pregnancy. Nonsmokers (n = 2798) and women who quit in early pregnancy (n = 226) were combined because so few women had quit and the characteristics of these groups of women and infants were not statistically different with regard to maternal prepregnant BMI, infant birth weight, timing of complementary food introduction, and duration of any breastfeeding. Gestational weight gain was obtained by self-report at the 6-mo postpartum interview. Data on family income were obtained from the 18-mo postpartum interview. Income data were expressed as a continuous variable and were classified by increments of 25000 Danish Kroner (US$3100) according to criteria set by the Danish government (19).
Infant feeding variables
At the 6-mo postpartum interview, the women were asked about their current infant feeding practices, including the duration of breastfeeding, use of formula, and introduction of complementary food. At the 18-mo postpartum interview, the women were asked about the duration of breastfeeding and use of formula. In the version of the 6-mo questionnaire used in the present analyses, the women were asked whether they ever fed infant formula to their infants and not when they first began to use formula; thus, we could not determine when full breastfeeding (the period during which the infants were fed neither formula nor complementary food) ended.
Therefore, "any breastfeeding" was used in these analyses and was defined as the total duration of breastfeeding, which was reported by mothers in months and weeks, weeks, or days and then converted to weeks. The duration of any breastfeeding was categorized into quartiles (<20, 20–31.9, 32–40, and >40 wk). Complementary foods were defined as mush or porridge in the questionnaire; infant formula was not classified as a complementary food. The age at introduction of complementary food was defined as the age at which complementary foods were first fed, which was reported by mothers in months and weeks or weeks and then converted to weeks. This variable was categorized as complementary food introduced at <16 and 16 wk. A cutoff of 16 wk was used because it was the lower limit of the Danish government’s recommendation that complementary foods should be introduced between 4 and 6 mo (20). Whether the continuous or categorical classification was used in each analysis was indicated. Formula use was classified into 2 categories: never or ever fed formula by 6 mo of age.
Infant variables
Information about the duration of gestation was available from 2 sources. At the 12-wk interview, women reported the date of their last menstrual period (LMP). Gestation duration was calculated as the number of weeks between birth and the LMP. Information was also available from the Danish National Patient Registry. In this registry, however, gestation duration was estimated by LMP or by ultrasound, but the method used for each subject was not recorded. Given the 2 sources of information, we used a method that has been applied to similar Danish data sets to validate gestation duration estimates (21). Comparisons were made between individual estimates from the LMP and from the registry, and it was shown that >96% of cases were within 14 d of agreement. The remaining 4% of cases were examined for data errors, and all but 24 cases were successfully assigned a value. All analyses were performed with the use of gestation duration estimates from the LMP, the registry, and the combined group; the results remained unchanged (data not shown). Therefore, for consistency, the LMP measure of gestation duration was used for >96% of cases; the remaining 4% of cases contained a mixture of LMP and registry estimates.
Infant birth weight and length were obtained from the Danish National Patient Registry. Information about infant weight and length at 1 y was obtained by maternal report from the "green book" at the 18-mo postpartum interview. The green book is a health record that is filled out by the infant’s physician and kept by the parents. Infant weight gain during the first year of life was calculated as the infant’s weight at the 1-y doctor visit minus birth weight.
Statistical methods
To test whether characteristics of included and excluded women were similar, Student’s t test was used (22). Student’s t test was also used to test whether infant birth weight and weight gain were similar between groups of 1) primiparous and parous women, 2) women who smoked during pregnancy and nonsmokers, and 3) male and female infants. To test whether characteristics of women categorized by the World Health Organization classification of their prepregnant BMI were similar, ANOVA was used (22). To make further comparisons among these categories, Dunnett’s t test was used to compare the characteristics of the underweight, overweight, and obese groups with the category of normal-weight women (those with prepregnant BMI values of 18.5–24.9). To test whether the timing of complementary food introduction (treated as a continuous variable) differed by quartiles of any breastfeeding duration, ANOVA was used. Dunnett’s t test was used to compare the timing of complementary food introduction among the categories of breastfeeding duration with those who breastfed for the longest duration (>40 wk). To test whether the timing of complementary food introduction was similar between women who never fed formula to their infants by 6 mo of age and those who did, Student’s t test was used. To test whether there was an interaction between the categories of any breastfeeding duration and the category of infant formula use by 6 mo of age on the timing of complementary food introduction, linear regression was used.
We proposed that maternal prepregnant BMI was positively associated with infant birth weight and with infant weight gain at 1 y. Furthermore, we proposed that maternal prepregnant BMI was negatively associated with the duration of any breastfeeding and with the age at complementary food introduction. We proposed that the duration of any breastfeeding was positively associated with age at complementary food introduction. Finally, we proposed that shorter durations of any breastfeeding and earlier introduction of complementary foods were positively associated with infant weight gain, and that these variables interact.
We evaluated these hypotheses using linear regression analyses in a stepwise manner (22). First, we examined the unadjusted associations between maternal prepregnant BMI and birth weight, weight gain, duration of any breastfeeding, and age at complementary food introduction. Second, we examined the association between duration of any breastfeeding and infant weight gain. Third, we examined the association between age at complementary food introduction and infant weight gain. Finally, we evaluated the interaction between duration of any breastfeeding and age at complementary food introduction on infant weight gain.
After examining the unadjusted associations of the variables of interest, we examined potentially confounding factors. We adjusted the regression model for the following continuous variables: maternal age, gestational weight gain, duration of gestation, birth weight, infant age at the time of the 1-y anthropometric measurements, and infant length at 1 y. The following categorical variables also were included: maternal level of education, maternal occupation in early pregnancy, maternal smoking during pregnancy, family income, presence of a spouse or partner during early pregnancy, primiparity, smoking during pregnancy, alcohol consumption during pregnancy, mode of delivery (vaginal or cesarean delivery), and infant sex. Models for infant weight gain were analyzed including and excluding infant length at 1 y. Infant length was included in some models because adjustment for this variable allowed the equation to predict the weight gain of the infant that went beyond the amount required to grow in length; this was the best way to assess high weight gain among these infants. Significance was set at 
= 0.05. All data were analyzed with the use of the Statistical Analysis System (version 8.2; SAS Institute Inc, Cary, NC) on a personal computer.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
10 y of education, were less likely to have a partner, and were more likely to smoke during pregnancy (Table 1
). Gestational weight gain, the duration of gestation, infant birth weight, and the duration of full and any breastfeeding did not differ significantly between the groups.
|
). Compared with normal-weight women, those who were overweight or obese before pregnancy had lower gestational weight gains, had higher rates of cesarean deliveries, and delivered heavier infants. Furthermore, compared with normal-weight women, obese women were older, less likely to have completed the compulsory education, and to have longer gestation durations (Table 2).
|
An examination of the association between categories of breastfeeding duration and the continuous variable of timing of complementary food introduction showed that women who breastfed for long periods (>40 wk, highest quartile of any breastfeeding duration) introduced complementary foods the latest. Compared with the category of women who breastfed for >40 wk, those who breastfed for <20 wk introduced complementary foods 2.4 wk sooner, those who breastfed for 20–31.9 wk introduced them 1.7 wk sooner, and those who breastfed for 32–40 wk introduced them 0.8 wk sooner (P < 0.0001, ANOVA with Dunnett’s t test). By 6 mo, 30% of infants in this study had never been fed formula. Compared with women who never fed formula by 6 mo of age, those who fed formula introduced complementary foods 1.4 wk sooner (P < 0.0001, Student’s t test). Furthermore, there was an interaction between never or ever feeding formula by 6 mo of age and category of breastfeeding duration and the age at complementary food introduction (P = 0.0489) (Figure 1). An examination of differences within quartiles of any breastfeeding duration showed that women who never fed formula and breastfed for the longest durations (the highest 2 quartiles) introduced complementary foods significantly later than did women who ever fed formula and breastfed for the same durations (Figure 1). The difference, however, was <1 wk.
|
). Infants of primiparous women weighed less at birth and gained more weight during the first year than did infants of parous women (Table 3). As expected, infants of smokers weighed less at birth and gained more than did the infants of nonsmokers during the first year of life (Table 3). Furthermore, male infants weighed more at birth and gained more weight than did female infants during the first year of life (Table 3).
|
). Infants of obese women were introduced to complementary foods an average of 0.5 wk sooner than were infants of normal-weight women. Furthermore, infants of obese women were breastfed an average of 4 wk less than were infants of normal-weight women.
Maternal prepregnant BMI and infant weight
Maternal prepregnant BMI was strongly associated with infant birth weight and infant weight gain from birth to 1 y. Compared with normal-weight women, those who were overweight or obese before pregnancy had infants with higher birth weights (Table 2). Also, compared with normal-weight women, those who were underweight before pregnancy had infants with lower birth weights (Table 2). Maternal prepregnant BMI was also positively associated with infant weight gain from birth to 1 y; in adjusted regression analyses without infant feeding information, as maternal prepregnant BMI increased by 1 unit, the infants gained 18.6 g more by 1 y of age (Table 4). Similar associations existed for weight gain when infant length was excluded from the model (data not shown).
|
). Additionally, the effects of the categories of any breastfeeding decreased by 3.8 to 17.0 g (3–5%), with the largest decease occurring in the category of <20 wk duration of any breastfeeding (Table 4). This indicates that maternal prepregnant BMI and the category of any breastfeeding duration explain some of the same variance in infant weight gain at 1 y.
When the categorical variable of age at complementary food introduction (<16 wk or 16 wk) was added into the model, the parameter estimate for maternal prepregnant BMI decreased by 0.9 g (5%); for each 1-unit increase in prepregnant BMI, the infant gained 17.7 g more by 1 y of age (Table 4). The effect of complementary food introduction decreased by 13.1 g (6%) when maternal prepregnant BMI was introduced into the model. This indicates that maternal prepregnant BMI and the category of complementary food introduction explain some of the same variance in infant weight gain at 1 y. Again, this suggests that maternal prepregnant BMI should be included in the analyses of infant feeding methods and infant weight gain during the first year of life.
Infant feeding and infant weight
Infant feeding methods were associated with infant weight gain from birth to 1 y. Compared with infants who were fed complementary foods at 16 wk, those who were fed these foods before 16 wk gained 224.2 g more from birth to 1 y (Table 4). Similar associations existed when infant length was excluded from the model (data not shown). To investigate whether reverse causality was underlying this association (ie, bigger infants "needed" more nutrition and were fed complementary foods sooner), we examined the association between birth weight and age at complementary food introduction (treated as a continuous variable) with regression analyses. This association was not statistically significant (P = 0.5).
Additionally, as the duration of any breastfeeding increased, the infant gained less weight from birth to 1 y. Compared with those who were breastfed for the longest duration (>40 wk), those who were breastfed for the shortest duration (<20 wk) gained 317.4 g more from birth to 1 y (Table 4). Similar associations existed when infant length was excluded from the model (data not shown). Again, to investigate whether reverse causality was underlying the observed association, we examined the association between infant birth weight and the duration of breastfeeding with regression analyses. In an unadjusted model, we found a small, but positive, association between birth weight and breastfeeding duration (treated as a continuous variable; adjusted R2 = 0.003, P < 0.001). This regression showed that for each additional 500 g of birth weight, the infant was breastfed for 1 wk longer. This association did not change appreciably when the model was adjusted for other factors associated with the duration of breastfeeding, including infant sex, primiparity, cesarean delivery, and maternal prepregnant BMI (adjusted R2 = 0.03, P < 0.0001).
To determine whether the categories of any breastfeeding duration and age at complementary food introduction explained the same part of the variance in infant weight gain at 1 y, we introduced these variables simultaneously into the model (Table 4). Both factors remained significant, but the parameter estimates for each decreased, which indicated that they did explain some of the same variance in infant weight gain at 1 y. When maternal prepregnant BMI was introduced into this model containing both categorical variables of duration of any breastfeeding and age at complementary food introduction, the parameter estimate for maternal prepregnant BMI decreased by 3.1 g (17%); for each 1-unit increase in maternal prepregnant BMI, the infants gained 15.5 g more by 1 y of age (Table 4). Additionally, the parameter estimates for any breastfeeding decreased by 3.9 to 34.1 g (3–11%), and the parameter estimates for complementary foods decreased by 57.5 g (26%). As all factors remained significant, and the effect sizes remained large enough to be physiologically significant, this confirmed that these factors need to be accounted for in analyses of infant feeding and weight gain from birth to 1 y.
Infant weight gain from birth to 1 y
Infant weight gain from birth to 1 y was predicted by both maternal and infant factors. In unadjusted regression analyses, maternal factors that were positively associated with infant weight gain included maternal prepregnant BMI, primiparity, maternal smoking during pregnancy, gestational weight gain, and cesarean delivery. Maternal factors that were negatively associated with infant weight gain included gestation duration, duration of any breastfeeding, and later introduction of complementary food. The maternal factors that had no significant association with infant weight gain included maternal age, alcohol consumption during pregnancy, level of education, occupation, family income, or the presence of a spouse or partner during early pregnancy (P > 0.1) (data not shown). The infant factors that had positive associations with infant weight gain included male sex, age at the time of the anthropometric measurements, and length at 1y. The birth weight of the infants was negatively associated with weight gain.
In the adjusted model, the interaction between the categorical variable of any breastfeeding duration (<20, 20–31.9, 32–40, or >40 wk) and the categorical variable of infant age at complementary food introduction (<16 wk or 16 wk) remained significant (overall test for interaction: P = 0.0021; Table 5). This interaction was also significant when the variables were treated as continuous (not shown). At the lowest quartile of any breastfeeding duration (<20 wk), the early introduction of complementary food (<16 wk) was associated with an increased weight gain from birth to 1 y (Figure 2). At longer durations of any breastfeeding, the timing of complementary food introduction was not associated with weight gain.
|
|
). Compared with the model that included length as a covariate, in the model without length, all variables, with the exception of gestational weight gain, remained significant. The direction of the associations remained the same in both models; however, the magnitude of the effects changed. Notably, the parameter estimate for birth weight decreased by one-fourth, the parameter estimate for infant age when the anthropometric measurements were taken nearly tripled, and the parameter estimates for infant sex and primiparity nearly doubled.
Because birth weight is positively associated with infant length at 1 y and negatively associated with infant weight gain, once length is removed from the model the effect of birth weight on infant weight gain at 1 y decreases. Because sex is strongly associated with length (boys are longer than girls), once length is removed from the model the parameter estimate for sex increases and explains the same variability in weight gain that length did. In this population, infants of primiparous women were significantly longer at 1 y (Table 3). Once length was dropped from the model, the parameter estimate for parity increased because parity and length explain the same part of variation in weight gain from birth to 1 y.
Results from the analyses of infant weight gain without length showed that much of the infant weight gain from birth to 1 y is attributable to infant length. Nonetheless, maternal prepregnant BMI, categories of any breastfeeding duration, and the category of timing of complementary food introduction remained significant in this model. These findings indicate that these factors have an association with infant weight gain that goes beyond the amount required to grow in length.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although the association between the duration of breastfeeding and later infant or childhood weight has been examined in several studies, researchers have not adjusted carefully for the effect of prepregnant BMI (11, 17, 23). The failure to do this, especially in contemporary cohorts that contain a high proportion of overweight and obese women, may lead to the overestimation of the effect of breastfeeding. This is because prepregnant BMI explains some of the same variance in infant weight gain from birth to 1 y that is also explained by the duration of any breastfeeding.
The present investigation is unusual in examining complementary food introduction among women who breastfeed for long durations. In several older studies, the early introduction of complementary foods was associated with higher infant weight at 1 y, but these cohorts consisted of infants who were primarily formula-fed (24, 25). Furthermore, the compositions of these formulas were quite different from those that are available today (26). Another unique aspect of this study was the short time (<6 mo) between the time of introduction of complementary foods and the time of the interview in which information about it was obtained, which reduced the chance of recall bias.
Maternal prepregnant BMI was an important determinant of infant growth in this population. The mean BMI of these Danish women was relatively low (23.4) compared with the BMIs of US women (54.3% of the women aged 20–39 y had a BMI 25.0; 1); yet, the range was large enough to see a potentially harmful effect of overweight and obesity on infant weight gain. There was a direct effect of maternal prepregnant BMI on the weight gain of these infants, even after the effects of birth weight and infant feeding methods were accounted for. It is notable that even in this population of women with very high rates of initiation and long durations of breastfeeding, overweight women breastfed for 2 wk less and obese women breastfed for 4 wk less than did normal-weight women. Furthermore, these overweight and obese women introduced complementary foods to their infants sooner than did normal-weight women.
Consistent with other studies, infants who were breastfed gained less weight at 1 y than did formula- or mixed-fed infants (27). Based on the current understanding that entering pregnancy at a normal-weight and following recommended infant feeding practices is associated with the best infant health outcomes, infant weight gain values were predicted. The infant of a reference woman who began her pregnancy with a mean BMI of 23.4, introduced complementary foods at 16 wk, and continued to breastfeed for >40 wk is predicted to gain 6489 g between birth and 1 y of age (Figure 2). Compared with this infant, the infant of a woman who breastfed for <20 wk and introduced complementary foods at <16 wk (a more typical pattern in the US; 28) would gain 624 g (9.6%) more (Figure 2). This represents a substantial difference in growth that could be achieved through variation in infant feeding practices alone.
Inasmuch as prepregnant BMI also contributes to infant growth, further calculations were performed in which both BMI values and infant feeding patterns were varied. Compared with the infant of the reference woman, the infant of an obese woman (prepregnant BMI = 30.0; 93rd percentile) who breastfed for >40 wk and introduced complementary foods at 16 wk would gain 102 g (1.6%) more. If this obese woman decreased any breastfeeding to <20 wk and introduced complementary foods at <16 wk, this infant would gain 726 g (11.2%) more than would the infant of the reference woman. As we and others observed previously (5–7, 29), women with higher prepregnant BMI values are likely to breastfeed for the shortest periods; thus, there is cause for concern about the growth of their infants.
There are potential limitations to our study. Our analyses were conducted in a select population of educated Danish mothers who participated in the DNBC; thus, the external validity of these results is restricted to mother-infant dyads with similar characteristics. There were significant differences between the included and excluded subjects (Table 1). These groups may differ in a meaningful way because of the larger number of smokers in the excluded group. We expect infants of these women to have lower birth weights (30), to be breastfed for shorter periods (31), and to gain more weight during the first year of life (due to "catch-up" growth associated with maternal smoking) (32); this is suggested by the results (Table 1). Given the characteristics of these excluded subjects, it is possible that our analyses underestimated the effects of smoking in the entire cohort.
Maternal prepregnant weight and height were obtained by self-report. We expect that the reports contained imprecision, but we had large numbers of subjects to compensate for this. Even if the heavier women underreported their weight more than did the lighter women, this would have caused us to underestimate the associations that we observed between prepregnant BMI and the outcomes of interest.
We had information about whether formula had ever been fed to the infant, but not the age when it was introduced. Danish women use different strategies for infant feeding than do women in other countries, such as the United States, who typically introduce formula before complementary foods (33). Comparisons of the overall feeding patterns between women who never and ever fed formula showed that women who never fed formula and breastfed for the longest times introduced complementary foods, on average, <1 wk later than did women who ever fed formula (Figure 1). Although we cannot exclude that the effect of complementary foods captures some effect of formula feeding, it is likely to be minimal. Additionally, results from a different subset of the DNBC, which contains information about the infant’s age at the time of formula introduction, support that the effect of formula is likely to be minimal. In this subset, by 6 mo of age, 37% of infants never received formula, 9% received complementary foods before formula, an additional 19% received these in the same week, and 35% received formula before complementary foods (JL Baker, unpublished observations, 2004).
Although it is difficult to evaluate the extent to which the differences in weight gain we observed among infants of women with different prepregnant BMI values and who used different feeding patterns are excessive, it is clear that infants of heavier women who were breastfed for shorter durations and received complementary foods sooner gained more weight by 1 y than did infants of normal-weight women. In this sample, there was a 500-g difference between the 75th and 85th percentiles of the weight gain distribution. The additional weight gain associated with high maternal prepregnant BMI, short durations of breastfeeding, and early complementary food introduction are sufficient to move an infant from the 75th to the 85th percentile, which is a commonly used cutoff for an infant being considered "at risk" of overweight.
The more rapid weight gain of the infants of heavier mothers could have many causes. It is possible that they received more formula or complementary foods or both, and thus more energy, than did infants of normal-weight women, possibly because of more vigorous sucking behavior (34). It is possible that their energy expenditure could have been lower, which was previously observed in a subset of infants born to heavier women (35). Additionally, it is possible that these infants had genetic characteristics different from those of the infants of normal-weight women (36) or that their genetic characteristics could have been modified in utero (37). However, this newly described phenomenon has not yet been shown to exist in human beings. Because we did not have information about the timing of formula introduction or quantity given, complementary food composition or quantity given, genetic characteristics, or energy expenditure, these data do not support one possible explanation over another.
This study provides further evidence that women should begin pregnancy at an appropriate weight to optimize breastfeeding and that women should introduce complementary foods at the appropriate time to provide their infants with the best start in life.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Mok, C. Multon, L. Piguel, E. Barroso, V. Goua, P. Christin, M.-J. Perez, and R. Hankard Decreased Full Breastfeeding, Altered Practices, Perceptions, and Infant Weight Change of Prepregnant Obese Women: A Need for Extra Support Pediatrics, May 1, 2008; 121(5): e1319 - e1324. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Scholtens, U. Gehring, B. Brunekreef, H. A. Smit, J. C. de Jongste, M. Kerkhof, J. Gerritsen, and A. H. Wijga Breastfeeding, Weight Gain in Infancy, and Overweight at Seven Years of Age: The Prevention and Incidence of Asthma and Mite Allergy Birth Cohort Study Am. J. Epidemiol., April 15, 2007; 165(8): 919 - 926. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Savage, J. O. Fisher, and L. L. Birch Parental Influence on Eating Behavior: Conception to Adolescence J. Law Med. Ethics, March 1, 2007; 35(1): 22 - 34. [PDF]
|
|
|
|
 |
|
 M. Giovannini, G. Radaelli, E. Riva, and G. Banderali Low Prepregnant Body Mass Index and Breastfeeding Practices J Hum Lact, February 1, 2007; 23(1): 44 - 51. [Abstract] [PDF]
|
|
|
|
 |
|
 L. Patterson, P. Jarvis, A. Verma, R. Harrison, and I. Buchan Measuring children and monitoring obesity: surveys of English Primary Care Trusts 2004-06 J. Public Health Med., December 1, 2006; 28(4): 330 - 336. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. J. Mayer-Davis, S. L. Rifas-Shiman, L. Zhou, F. B. Hu, G. A. Colditz, and M. W. Gillman Breast-Feeding and Risk for Childhood Obesity: Does maternal diabetes or obesity status matter? Diabetes Care, October 1, 2006; 29(10): 2231 - 2237. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. L Burdette, R. C Whitaker, W. C Hall, and S. R Daniels Breastfeeding, introduction of complementary foods, and adiposity at 5 y of age Am. J. Clinical Nutrition, March 1, 2006; 83(3): 550 - 558. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. K. Ong, P. M. Emmett, S. Noble, A. Ness, D. B. Dunger, and and the ALSPAC Study Team Dietary Energy Intake at the Age of 4 Months Predicts Postnatal Weight Gain and Childhood Body Mass Index Pediatrics, March 1, 2006; 117(3): e503 - e508. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. Ekelund, K. Ong, Y. Linne, M. Neovius, S. Brage, D. B Dunger, N. J Wareham, and S. Rossner Upward weight percentile crossing in infancy and early childhood independently predicts fat mass in young adults: the Stockholm Weight Development Study (SWEDES) Am. J. Clinical Nutrition, February 1, 2006; 83(2): 324 - 330. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. E Buchan, R. F Heller, P. Clayton, P. E Bundred, and T. J Cole Early life risk factors for obesity in childhood: Early feeding is crucial target for preventing obesity in children BMJ, August 20, 2005; 331(7514): 453 - 454. [Full Text]
|
|
|
|
 |
|
 N. Stettler, V. A. Stallings, A. B. Troxel, J. Zhao, R. Schinnar, S. E. Nelson, E. E. Ziegler, and B. L. Strom Weight Gain in the First Week of Life and Overweight in Adulthood: A Cohort Study of European American Subjects Fed Infant Formula Circulation, April 19, 2005; 111(15): 1897 - 1903. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||
