Maternal and Child Health Journal

, Volume 18, Issue 5, pp 1224–1232

Early Childhood Healthy and Obese Weight Status: Potentially Protective Benefits of Breastfeeding and Delaying Solid Foods

Authors

    • Department of SociologyWayne State University
  • William H. Yeaton
    • Institute for Social ResearchUniversity of Michigan
Article

DOI: 10.1007/s10995-013-1357-z

Cite this article as:
Moss, B.G. & Yeaton, W.H. Matern Child Health J (2014) 18: 1224. doi:10.1007/s10995-013-1357-z

Abstract

The aim of this study was to assess the relationship between breastfeeding and postponing introduction to solid food (SF) on children’s obesity and healthy weight status (WS), at 2 and 4 years. Drawing upon a nationally representative sample of children from the Early Childhood Longitudinal Study-Birth Cohort, we estimated the magnitude of the relationship between children’s WS and early feeding practices. Contingency tables and multinomial logistic regression were used to analyze obese and healthy WS for breastfed and never breastfed children and examine three timing categories for SF introduction. With both percentages and odds, breastfeeding and delaying introduction to SF until 4 months were associated with lower obesity rates and higher, healthy WS rates (typically 5–10 %). Analyses of feeding practice combinations revealed that when children were not breastfed, obesity odds decreased when SF introduction was postponed until 4 months. Obesity odds were further reduced when SF delay was combined with breastfeeding. Consistent increases in healthy WS were also observed. Benefits were stable across both follow-up periods. Breastfeeding and delaying complementary foods yielded consistently and substantially lower likelihood of obesity and greater probability of healthy WS. Health policies targeting early feeding practices represent promising interventions to decrease preschool obesity and promote healthy WS.

Keywords

Body weightBreastfeedingComplementary foodInfant foodObesity

Abbreviations

ECLS-B

Early childhood longitudinal study-birth cohort

WS

Weight status

SF

Solid food

Introduction

Pediatric experts have asserted that early feeding practices are critical for the development of healthy weight status (WS) in young children as they have been shown to protect against obesity and to promote healthy WS later in childhood. [14] Monitoring young children’s WS is important since excess weight has been associated with undesirable childhood psychological (e.g., low self-esteem, behavioral problems, and psychiatric disorders) and physiological outcomes (e.g., asthma, childhood diabetes, and elevated childhood cardiovascular risk). Similarly, there are long-term, adult, adverse health conditions such as persistent obesity, premature mortality, and cardiovascular risk. [5] Being sensitive to these risks and benefits, the American Academy of Pediatrics (AAP) advises parents to exclusively breastfeed infants until 6 months old. Moreover, prolonged breastfeeding is endorsed by AAP until at least the first birthday, after nutritious SF has been introduced at about 6 months. [6].

Research Evidence: Impact of Breastfeeding and SF Timing on Early Childhood WS

In a systematic review of 28 studies, consistent evidence emerged that, when compared to formula feeding, breastfeeding reduces the likelihood of later obesity. [4] Yet, despite a significant body of research that validates the benefits of breastfeeding on young children’s WS, several contemporary scholars have recently characterized this relationship as a “myth.” [7].

Additional evidence connecting the positive benefits of breastfeeding to the development of healthy WS has often been quite compelling. A second set of meta-analytic results indicated that breastfeeding was associated with lower levels of childhood obesity and increased likelihood of desirable WS, for children between 5 and 18 years. [2] These findings and those from the above meta-analysis by Owen et al. [4] suggest that breastfeeding serves as a foundation for promoting healthy WS development well beyond early childhood. Moreover, a number of studies have reported greater chances of healthy WS between 3 and 6 years when comparing breastfed and formula fed children [8]. Still other studies have extended the period of benefit to late childhood and adolescence. [9].

Research has also linked the timing of solid food (SF) introduction and future adiposity. In a recent study of early feeding practices, Huh et al. [3] found an increased risk for obesity when SF was introduced before 4 months. However, some research has been less conclusive about the timing of complementary food [10] and its relationship with both obesity and overweight in toddlers and preschoolers [11, 12].

To foster healthy WS, leading medical organizations (AAP, IMO, WHO) recommend that parents postpone introduction to nutritious SF until children reach 4 months and, if possible, initiate SFs at 6 months [6, 13, 14]. Brophy et al. [15] observed that infants introduced to SF prior to 3 months had higher levels of obesity at 5 years. Huh et al. [3] found that formula fed infants introduced to complementary foods before 4 months were more prone to overweight status at 3 years compared to children introduced to SF at 4–5 months (though results were non-significant due to very small sample size). In line with this evidence, delaying introduction to SF has been related to reduced overweight/obese odds compared to an age-cohort of children not experiencing delay. [16, 17] In contrast, the majority of studies included in Moorcraft et al.’s [10] systematic review reported no consistent relationship between timing of SF introduction and later, unhealthy WS.

While there was general confluence among the above findings, the weight of their conclusions should be tempered against several inferential weaknesses. Within the reviews, few studies explored children’s WS during the first years of life. When preschool WS was assessed, most studies used a single measurement period rather than multiple follow-ups. Most studies did not use prospective, nationally representative data, thus compromising generalizability. Relatively weak research designs such as case–control were common. Furthermore, many studies included in the above meta-analyses utilized data from children born more than 20 years ago. Thus, there are a considerable number of dimensions upon which methodological improvement might occur.

Evidence for the Combined Impact of both Feeding Practices, Within a Single Study

Relatively few studies have examined the combined impact of both breastfeeding and delaying SF introduction on young children’s WS, within the context of a single study. Huh et al. [3] recently reported that when SF was introduced before 4 months to non-breastfed children, the odds of obesity at 3 years increased six-fold. Yet, if children were breastfed, the timing of SF introduction had no impact on WS at age three. Unfortunately, since Huh et al.’s sample lacked racial diversity, contained predominantly high income, college educated mothers, and utilized only one follow-up period, the generalizability of their findings is an important avenue of further inquiry.

While the large magnitude of obesity risk was compelling for those non breastfed children who were introduced to SF prior to 4 months in Huh et al.’s research, the very small sample of obese children (a total of 75, with several cells ranging from three to eight) led to unreliable benefit estimates. As Ionnidis [18] has noted in a provocative paper titled “Why most published research findings are false,” there are substantial inferential dangers in such undersized research studies.

In summary, given the promise but relative uncertainty reflected in the above body of research, the primary purpose of our longitudinal study was to investigate the extent to which breastfeeding and timing of SF introduction were independently and concurrently related to WS in a national probability sample of young children.

Methods and Procedures

Data Source

Data were taken from the 9 month, 2, and 4 year waves of the Early Childhood Longitudinal Study-Birth Cohort (ECLS-B), a nationally representative sample of US children born in 2001. Periodic assessments of children’s physical growth and parental reports of feeding experiences were collected from birth through kindergarten.

ECLS-B sampled 14,200 children from 2001 birth certificates. From this initial sample, study protocol excluded approximately 3,500 children born to young mothers (< 16y), were adopted, died prior to their first birthday, or had parents who refused to participate. With later data collection waves, ECLS-B omitted children whose parents subsequently refused to participate and children who moved outside of the US or died. Also, we eliminated about 1,150 very low birth weight (<1,500 g) children from the analyses, because children with extremely low birth weight may follow very different height and weight growth trajectories than normal birth weight children [19].

These restrictions reduced the sample to 7,200 at 2 years and 6,950 at 4 years. All unweighted sample sizes were rounded to the nearest 50 in accordance with National Center for Education Statistics data security policy [20]. Data collection methods were approved by the National Center for Health Statistics and State Institutional Review Boards.

Measuring Weight Status

At 2 and 4 year waves, two height (cm) and weight (kg) measurements were taken for each child. Trained research staff measured children’s stature using a statiometer and children’s weight using a digital scale. Within-wave reliability for both height (r = .99, 2 years; r = .99, 4 years) and weight (r = .98, 2 years; r = .99, 4 years) measurements were high.

The CDC’s 2000 reference growth charts were used to classify children’s WS into one of three commonly used categories [1, 21, 22]. Children’s sex, height, weight, and age at the time of measurement were linked to determine each child’s body mass percentile at 2 and 4 years. Children with percentiles less than 85 were classified as having a healthy WS (including about 600 children who were classified as underweight with percentiles less than 5). Children whose body mass percentiles fell between 85 and 94 were considered overweight, whereas, children with percentiles at or above 95 were categorized as obese.

Breastfeeding and Solid Food Introduction

During 9 month and 2 year data collection periods, mothers were asked, “Did you ever breastfeed {child}?” or “Are you still breastfeeding {child} now?” Based on mother’s responses to these questions, we created two classifications of breastfeeding experience: children who were breastfed (69.7 %) or not (30.4 %) (total percentage was over 100, due to rounding). Children whose mothers reported that they had ever breastfed or were breastfeeding at the time of data collection were considered breastfed.

At 9 months, mothers reported the month when children were introduced to SF, which included pureed foods or cereal but not finger foods. Based on maternal report, we established three groups of children, in accordance with the AAP recommendation to postpone SF until after 4 months and, if possible, to wait until infants reached 6 months. Thus, the first group was composed of children introduced SF <4 months (23.6 %), the second group was introduced to complementary food between 4 and 5 months (49.8 %), and the third group was introduced SF ≥6 months (26.6 %).

Analytic Approach

Due to ECLS-B’s complex sampling design, sample weighting techniques were used to produce results generalizable to the US population of children born in 2001. Sample weights were applied to adjust for survey nonresponse, overall selection probabilities, and undercoverage at each data collection wave. Jackknife repeated replication variance estimation, along with base and replicate weights, were used to calculate prevalence point estimates and to conduct statistical tests with Stata v11 [23].

Rao-Scott design-adjusted F tests compared the three WS classifications for either breastfeeding categories or the three timing groups, at both 2 and 4 years. We also constructed contingency tables to assess the relationship between combination of children’s breastfeeding status and SF timing categories with children’s WS, at both follow-up periods. To evaluate the separate and conditional benefits of breastfeeding and delayed SF introduction on children’s WS we used tests of significance for differences in proportions. Furthermore, we used multinomial logistic regression to gauge both the unique and the statistically adjusted, conditional risks associated with both feeding practices and WS. Our statistical models controlled for the effects of several socioeconomic (maternal education and family poverty status) and demographic covariates (maternal age, maternal race, children’s gender, and birth weight) that previous studies had suggested might confound feeding practices or growth trajectories and WS [3, 11, 12]. Since these covariates were not reported for each study participant, there was a slight reduction in original sample size (~1.5 %) when they were included. We also considered including maternal body mass index as a covariate. However, due to a substantial degree of nonresponse, inclusion of this variable would have resulted in a large decline in sample size (~11 %), reduced statistical power, and led to far less reliable odds ratio estimates.

Results

From a public health perspective, the most desired WS outcome occurs when children develop and maintain a healthy WS. In contrast, the development or maintenance of an obese WS would be least desirable. Therefore, our results centered upon the relationship between feeding practices and healthy or obese WS, at both 2 and 4 years, and excluded any systematic discussion of overweight WS.

To better understand the ways in which feeding practices relate to WS, we divided our results into three sections. First, we reported how breastfeeding (using row percentages from Table 1a, b) or timing (using cells within Table 2) related to WS percentage differences. Second, we revisited cell percentage differences found within and between Table 1a, b to address how both feeding practices were associated with healthy and obese WS. Third, using odds ratios, we reported the overall impact of each factor (Table 3) and, finally, the conditional impact of each feeding practice across levels of the other on the odds of children’s healthy, overweight, or obese WS (Table 4).
Table 1

Weighted percentage of children’s weight status by feeding practice

Weight Status

Breastfed

Not breastfed

Solid food introduced, months

Solid food introduced, months

<4

4–5

≥6

Row total

<4

4–5

≥6

Row total

(a) 2 years

 Healthy

61.9

69.4

70.3

68.2

55.8

61.8

62.4

59.9

[57.7,65.9]

[66.4,72.2]

[66.8,73.6]

[65.9,70.3]

[50.1,61.3]

[56.9,66.4]

[57.1,67.4]

[56.1,63.5]

 Overweight

13.1

14.5

13.1

13.8

12.5

13.8

12.9

13.1

[10.9,15.7]

[12.8,16.5]

[11.1,15.3]

[12.6,15.2]

[9.7,15.9]

[11.3,16.6]

[9.6,17.1]

[11.4,15.1]

 Obese

25.0

16.1

16.6

18.0

31.7

24.5

24.7

27.0

[21.8,28.5]

[14.0,18.5]

[14.0,19.7]

[16.3,19.9]

[26.4,37.6]

[20.1,29.6]

[20.0,30.0]

[23.3,31.0]

Weight status

Breastfed×

Not Breastfed§

Solid food introduced, months

Solid food introduced, months

<4

4–5

≥6

Row total

<4

4–5

≥6

Row total

(b) 4 years

 Healthy

63.6

71.4

69.9

69.5

59.4

67.5

70.1

65.4

[59.1,67.9]

[68.9,73.8]

[66.6,72.9]

[67.7,71.2]

[53.8,64.8]

[63.5,71.]

[64.1,75.5]

[62.2,68.4]

 Overweight

19.0

17.7

17.5

17.9

18.9

16.4

12.4

16.3

[15.7,22.8]

[16.0,19.4]

[14.5,21.1]

[16.4,19.4]

[14.2,24.6]

[13.6,19.7]

[8.7,17.4]

[14.00,19.0]

 Obese

17.4

11.0

12.6

12.7

21.7

16.0

17.5

18.3

[14.0,21.3]

[9.4,12.7]

[10.3,15.3]

[11.4,14.2]

[17.7,26.4]

[12.6,20.3]

[13.9,21.9]

[15.8,21.1]

Row total and column percentages may not equal 100 due to rounding. 95 % confidence intervals are in brackets

Rao-Scott designed-adjusted F (3.68, 350) = 6.87, P < .001; ‡ Rao-Scott designed-adjusted F (3.74, 350) = 2.19, P < .10; ×Rao-Scott designed-adjusted F (3.48, 300) = 3.69, P < .01; § Rao-Scott designed-adjusted F (3.62, 300) = 2.74, P < .05

Table 2

Two and 4 year weighted percentage of children’s weight status by timing

Weight Status

2 years

4 years

Solid food introduced, months

Solid food introduced, months

<4

4–5

≥6

Row total

<4

4–5

≥6

Row total

 Healthy

59.5

67.4

68.5

65.8

62.0

70.5

70.2

68.4

[55.8,63.0]

[64.6,70.1]

[65.4,71.3]

[63.7,67.8]

[58.6,65.3]

[68.5,72.3]

[67.5,72.8]

[66.9,69.8]

 Overweight

12.9

14.2

13.0

13.6

18.9

17.2

16.1

17.3

[11.1,15.0]

[12.6,16.1]

[11.3,14.9]

[12.4,14.8]

[15.8,22.4]

[15.8,18.8]

[13.6,18.9]

[16.0,18.8]

 Obese

27.6

18.4

18.6

20.7

19.1

12.3

13.8

14.3

[24.3,31.2]

[16.1,20.9]

[16.3,21.2]

[18.7,22.8]

[16.5,22.1]

[10.8,14.0]

[11.8,16.0]

[13.1,15.6]

Row total and column percentages may not equal 100 due to rounding. 95 % confidence intervals are in brackets

Rao-Scott designed-adjusted F (3.76, 350) = 10.80, P < .001; ‡ Rao-Scott designed-adjusted F (3.61, 300) = 6.98, P < .001

Table 3

Two and 4 year obese and healthy odds ratios by breastfeeding and timing

Variable

2 years

4 years

Healthy

Overweight

Obese

Healthy

Overweight

Obese

OR

95 % CI

OR

95 % CI

OR

95 % CI

OR

95 % CI

Breastfed

 No

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

 Yes

1.08

[0.89, 1.31]

1 [Ref]

0.64***

[0.51, 0.80]

0.94

[0.75, 1.18]

1 [Ref]

0.67**

[0.53, 0.85]

Solid food timing

 <4 m

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

 4–5 m

0.99

[0.78, 1.26]

1 [Ref]

0.62***

[0.50, 0.78]

1.21^

[0.96, 1.51]

1 [Ref]

0.77^

[0.57, 1.03]

 ≥6 m

1.10

[0.86, 1.41]

1 [Ref]

0.70*

[0.53, 0.92]

1.28^

[0.97, 1.70]

1 [Ref]

0.90

[0.63, 1.29]

Odds ratios adjusted for maternal education, maternal age, maternal race, family poverty, children’s birth weight and gender

CI confidence interval, OR odds ratio

P < .10; * P < .05; ** P < .01; *** P < .001

Table 4

Two and 4 year obese and overweight odds ratios by combination of breastfeeding and timing

Combination

2 years

4 years

Healthy

Overweight

Obese

Healthy

Overweight

Obese

 

OR

95 % CI

OR

95 % CI

 

OR

95 % CI

OR

95 % CI

Breastfed

SFT

 

 No

<4

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

1 [Ref]

 Yes

<4

1 [Ref]

0.89

[0.62, 1.29]

0.64**

[0.47, 0.86]

1 [Ref]

0.91

[0.62, 1.34]

0.72^

[0.49, 1.06]

 No

4–5

1 [Ref]

0.97

[0.69, 1.36]

0.67**

[0.50, 0.89]

1 [Ref]

0.78

[0.53, 1.14]

0.66*

[0.45, 0.98]

 Yes

4–5

1 [Ref]

0.92

[0.67, 1.28]

0.38***

[0.29, 0.51]

1 [Ref]

0.80

[0.55, 1.16]

0.45***

[0.33, 0.61]

 No

≥6

1 [Ref]

0.91

[0.58, 1.44]

0.69*

[0.48, 0.99]

1 [Ref]

0.57*

[0.32, 1.00]

0.68*

[0.48, 0.96]

 Yes

≥6

1 [Ref]

0.82

[0.58, 1.15]

0.38***

[0.27, 0.53]

1 [Ref]

0.81

[0.54, 1.22]

0.51***

[0.36, 0.72]

Odds ratios adjusted for maternal education, maternal age, maternal race, family poverty, children’s birth weight and gender

CI confidence interval, OR odds ratio, SFT solid food timing (months)

P < .10; * P < .05; ** P < .01; *** P < .001

Impact of Breastfeeding

Using row totals in Table 1a, b, based on tests of significance for the difference in proportions, we found that breastfed children were more likely than never breastfed children to have healthy WS at both 2 (68.2 vs. 59.9 %, P < .001) and 4 years (69.5 vs. 65.4 %, P < .05). Breastfed children also had a lower percentage of obese WS than children not breastfed, at both 2 (18.0 vs. 27.0 %, P < .001) and 4 years (12.7 vs. 18.3 %, P < .001).

Both breastfed and never breastfed infants displayed decreased obesity percentages from 2 to 4 years. Similarly, healthy WS was higher for both breastfed and not breastfed children at 4 versus 2 years. These results indicate WS advantages between breastfed and never breastfed children at 2 years were maintained at 4 years.

Impact of Timing of Solid Food Introduction

Inspection of individual cells in Table 2 suggests that timing was related to obese WS at both time periods. Percentage of obesity at 2 years was highest among children introduced to SF <4 months (27.6 %) when compared to those children provided SFs at 4–5 months (18.4 %, P < .001) or ≥6 months (18.6 %, P < .001). Similarly, at 4 years, obesity was more common for children introduced to SF <4 months (19.1 %) compared to children who began SF at 4–5 months (12.3 %, P < .001) or ≥6 months (13.8 %, P < .01).

The specific time period when SF was introduced also had implications for development of healthy WS. At 2 years, children given SF <4 months were less likely to have healthy WS (59.5 %) versus those introduced later (4–5 months = 67.4 %, P < .001; ≥6 months = 68.5 %, P < .001). A similar trend of increased healthy WS percentages by postponed SF introduction occurred at 4 years. Healthy WS occurred less often among infants who were provided SF before 4 months (62.0 %) than children introduced SF between 4 and 5 months (70.5 %, P < .001) or ≥6 months (70.2, P < .001).

As with breastfeeding, we also evaluated whether timing of SF led to increased or decreased levels of obese and healthy WS by comparing 2–4 years results. All three SF introduction timing categories led to decreased obesity percentages at 4 years when contrasted to percentages at 2 years. For each timing period, healthy WS percentages were higher at 4 years versus 2. Again, the evidence was consistent that the advantage of both factors was maintained at 4 years.

Conditional Impact of Breastfeeding, Solid Food Timing, and Weight Status

Table 1a, b also permitted us to determine how healthy and obese WS were related to both feeding practices. At 2 years, using tests of significance for differences between proportions, we found the percentage of healthy WS increased with the delay of SF introduction, for both breastfed (from 61.9 % for <4 months to 70.3 % for ≥6 months, P < .01) and never breastfed children (from 55.8 % for <4 months to 62.4 % for ≥6 months, P = .07). Similarly, at 4 years, breastfed and never breastfed children introduced to SF <4 months maintained the lowest healthy WS percentage, which, for both groups, significantly increased as complementary food was delayed (for breastfed children from 63.6 % for <4 months to 69.9 % for ≥6 months (P < .05) and for never breastfed children from 59.4 % for <4 months to 70.1 % for ≥6 months, P < .01). When compared to never breastfed children, breastfed children introduced to SF <4 months (63.6 vs. 59.4 %, P > .05) or between 4 and 5 months (71.4 vs. 67.5 %, P > .05) possessed a non-significant, higher percentage of healthy WS.

When children were introduced to SFs <4 months, compared to the other timing periods, the highest percentage of obese children was present, regardless of whether they were breastfed or not, at 2 years (breastfed: 25 % (<4 months) vs. 16.1 % (4–5 months), P < .001 or 16.6 % (≥6 months), P < .001; repeating the above timing introduction pattern, we found not breastfed: 31.7 %, P < .01 vs. 24.5 or 24.7 %, P < .05) and 4 years (breastfed: 17.4 vs. 11.0 %, P < .01 or 12.6 %, P = .05; never breastfed: 21.7 vs. 16.0 %, P < .05 or 17.5 %, P > .05). Contrasted with breastfed children, the obesity percentage for children introduced to SF <4 months was higher at 2 years for never breastfed children (25.0 % vs. 31.7 %, P < .05); however, at 4 years the analogous percentage difference of obesity by breastfeeding experience was not significant (17.4 vs. 21.7 %, P > .05).

Impact of Breastfeeding and Early Introduction to Solid Food on Obesity: Odds Ratios

In Tables 1 and 2, we utilized percentages and were able to analyze the relationship between both feeding practices as they related to healthy, overweight, and obese WS. In Table 3, we utilized odds ratios from a multinomial logistic regression model that allowed us to adjust for the possible influence of potentially important covariates. With this analytic strategy, we were able to estimate the impact of breastfeeding and timing of SF introduction as individual variables. Breastfed children had significantly lower odds of obesity at both 2 years (odds ratio (OR): 0.64) and 4 years (OR: 0.67) than non-breastfed children. While statistical control for covariates and for timing of SF introduction led to a reduced risk of obesity, no statistically significant, beneficial relationship was found for healthy WS at either time period.

A similar pattern of lower odds of obesity continued at 2 and 4 years when SF introduction was delayed until 4–5 months (2 year OR: 0.62; 4 year OR: 0.77) or ≥6 months (2 year OR: 0.70; 4 year OR: 0.90). All four estimates indicated reduced odds of obesity, with two being significant and one with P = .08 At 4 years, we observed that waiting to introduce SF until after 4 months (i.e., between 4 and 5 or ≥6 months) led to healthy WS (P = .10 and P = .08, respectively for both results).

To provide a more detailed analysis of obesity and overweight WS risk for each factor, we assembled children into one of six feeding practice groups and conducted a second, multinomial logistic regression (Table 4). This approach allowed us to determine the impact of the level of each factor, within the levels of the other, while adjusting for the influence of the covariates.

When compared to the reference group of children who were not breastfed and who were also introduced to SF before 4 months, each of the other feeding practice groups had significantly lower odds of obesity at 2 years. The odds of obesity were also significantly lower at 4 years for all feeding practice groups, with the sole exception of those children who were introduced SF <4 months and who were breastfed (P = .10).

We further noted that, for the bottom five rows of Table 4, being breastfed resulted in a substantial, incremental benefit for reduced obesity odds versus not being breastfed. Regardless of timing of SF, breastfed children had lower odds of obesity when compared to children who were not breastfed (2y OR: 0.64 vs. 1.00 (reference group) for <4 months; 0.38 vs. 0.67 for 4–5 months; OR: 0.38 vs. 0.69 for ≥6 months; and 4 year OR: 4 year OR: 0.72 vs. 1.00 (reference group) for <4 months; 0.45 vs. 0.66 for 4–5 months; OR: 0.51 vs. 0.68 for ≥6 months).

Discussion

Using a nationally representative cohort, we addressed questions of both overall and conditional impacts of breastfeeding and timing of SF introduction on early childhood WS. Our data strongly indicate that, both independently and in concert, each factor played an important role in desirable child development, beginning in the first few years of life. Whether reported as differences in prevalence rates or as odds ratios, breastfeeding and delaying SF introduction have large, consistent, and positive impacts on favorable WS, at both 2 and 4 years. Breastfeeding’s beneficial impact was consistent, both when analyzed across and within timing categories, with differences typically 5–10 percentage points. When odds ratios were used as a basis of comparison, results were consistently significant at both 2 and 4 years, indicating that breastfeeding reduced obesity risk.

For both breastfeeding and timing, 4 year data showed greater WS benefit than those at 2 years (Table 3). The odds of obesity were substantially and consistently diminished by waiting to introduce children to SF until 4 months, and the odds for healthy WS were consistently higher within both time periods. In Table 4, we were able to gauge the conditional impact of both factors using odds ratios since an individual child’s data could only appear in a single combination of breastfeeding (yes or no) and timing. At both follow-up periods, the most desirable combination, breastfeeding and delaying complementary foods, was associated with the largest WS benefit. While SF delay reduced the risk of obesity, breastfeeding further and substantially reduced those risks by nearly one half. Thus, factor benefits were dynamic and catalytic, and the joint impact was close to additive.

Possible Feeding Practice Mechanisms

One possible explanation for breast milk’s benefit might be that breastfed children learn self-regulated appetite control that persists when they begin to self-feed solids. Being breastfed may acclimatize children to control food intake and develop the ability to identify internal cues signifying hunger gratification [24, 25]. In contrast, bottle-fed children may overeat since they rely more heavily on caregivers to signal when feeding should conclude (e.g., encourage infants to empty an entire bottle though the child’s hunger had been satisfied), leading to a conditioned tendency to overeat persisting beyond infancy.

In addition, termination of breastfeeding prior to 6 months has been linked to children being introduced to SF early, and this early introduction was more likely to be associated with foods containing high levels of fat and sugar [26]. These unhealthy SF eating habits established before children are 2 years old have been linked to habits of older children who eat food with low nutritional value and high calories [27, 28]. Therefore, the early introduction of unhealthy SF may establish feeding routines that exacerbate children’s undesirable WS.

Literature: Fit, Clarification, and Extension

These findings not only substantiate but also clarify and extend those of previous studies addressing these two feeding practices. Our estimates of breastfeeding benefit (reduced odds of obesity at 2 years = 0.64 and 4 years = 0.67) were similar to estimates by Arenz et al. [2] and Owen et al. [4] Our findings also parallel other research linking early complementary foods to significant increases in childhood obesity risk [15, 17].

Unlike previous studies, our research simultaneously addressed overall and conditional impact (based on one of six, unique combinations of each individual child’s feeding experience) of breastfeeding and timing of SF introduction. As noted above, Huh et al. [3] found a similar pattern of increased obesity risk at 3 years for children who were not breastfed but introduced to complementary food prior to 4 months. In contrast to our findings, these authors observed that when children were breastfed for more than 4 months, timing was not related to obesity. Their conclusions may have partially resulted from sample differences; our study included families across all SES levels, whereas theirs was limited to higher SES families who could afford and have access to more expensive and nutritious SF [29, 30]. Furthermore, the Huh study’s conclusions were based on unstable and imprecise estimates, due primarily to the very low, within-cell sample sizes.

Strengths and Limitations

Our data were derived from a well-designed, large, nationally representative probability sample, with repeated and direct, highly reliable height and weight measurements. This methodically sound evidence adds to the increasing body of findings for the benefits of breastfeeding and delaying SF introduction. Using two different analytic approaches, we provide estimates that suggest substantial benefits of early feeding practices on preschool WS and that these benefits occur at both 2 and 4 years, instead of the single time point used in other studies [3]. Finally, while our analysis of the joint impact of early feeding practices showed a large reduction in obesity risk, our findings did not reveal an increased rate of healthy WS since this group had been used as our reference category.

To minimize the extent of missing data (some mothers reported breastfeeding but did not provide duration), we reported breastfeeding as a dichotomy (mothers either breastfed or not). This operationalization would underestimate the benefits of breastfeeding on WS, since a woman who breastfed for only a brief period would have had a smaller impact than if she had breastfed for a considerable duration. Despite this potential underestimation, we found the benefits of breastfeeding to be substantial and consistent.

Policy Implications

Our results provide high quality, supportive evidence for health care professionals to inform parents regarding early childhood WS risks. Pediatricians and other physicians are particularly well positioned to prompt and to encourage these feeding practices during regularly scheduled, child care checkups. Professional advice might best be two-pronged to include both the benefits for the development and maintenance of healthy WS as well as the health risks associated with childhood obesity [5]. Though past attempts to translate risk factor information into effective behavioral strategies have not always been successful [31], our data offer relatively unambiguous evidence that parents seriously consider both breastfeeding and delaying introduction to SF since they operate together to curb obesity and to promote healthy WS.

Our findings strongly support existing policy statements from leading health organizations regarding the positive impact of breastfeeding and delaying solid food on children’s health [6, 13, 14]. Sustained endorsement of these feeding guidelines is potentially useful both as a prevention strategy and as a method to reduce the prevalence of childhood obesity. Consequently, specific public health policies that eliminate barriers to breastfeeding (e.g., options for storing breast milk and flexible break times at work) and increase parental awareness by aggressively publicizing the benefits of delaying introduction to complementary foods deserve serious scrutiny.

Acknowledgments

Both Drs. Moss and Yeaton developed the study concept and design, participated with the interpretation of the data, as well as contributed to the manuscript’s content. Dr. Moss had access to the restricted data and takes responsibility for the integrity of the data analyses. Neither author claims a conflict of interest and approval was granted by the National Center for Health Statistics and State Institutional Review Boards. The authors wish to express appreciation to Drs. Janet Hankin and Riva Tukachinsky for feedback on earlier drafts of this manuscript.

Conflict of interest

The authors have no competing interest.

Copyright information

© Springer Science+Business Media New York 2013