OBJECTIVES:

Breast milk has higher cholesterol than formula. Infants who are breastfed have different cholesterol synthesis and metabolism in infancy than infants who are formula fed. Little is known as to whether breastfeeding is associated with subsequent lipid profile, independent of adiposity. We assessed the association of breastfeeding in early infancy with lipid profile and adiposity at ∼17.5 years in a setting where exclusive breastfeeding is not associated with higher socioeconomic position.

METHODS:

We used multivariable linear regression with multiple imputation and inverse probability weighting to examine the associations of contemporaneously reported feeding in the first 3 months of life (exclusive breastfeeding [7.5%], mixed feeding [40%], or always formula feeding [52%]) with lipids and adiposity at ∼17.5 years in 3261 participants in the Hong Kong Chinese birth cohort Children of 1997, adjusting for sex, birth weight, gestational weeks, parity, pregnancy characteristics, parents’ highest education, mother’s place of birth, and age at follow-up.

RESULTS:

Exclusive breastfeeding, but not mixed feeding at 0 to 3 months, compared with formula feeding was associated with lower total cholesterol and low-density lipoprotein cholesterol but not with high-density lipoprotein cholesterol at ∼17.5 years. BMI and fat percentage measured by bioimpedance did not differ by type of infant feeding.

CONCLUSIONS:

Exclusive breastfeeding in early infancy may promote a healthier lipid profile in late adolescence through mechanisms unrelated to adiposity, implicating its potential long-term benefits for cardiovascular health.

What’s Known on This Subject:

Breastfed infants have higher plasma cholesterol and less endogenous cholesterol synthesis. However, whether and how an exposure to high cholesterol in breast milk programs cholesterol homeostasis and affects the trajectory of the lipid profile in adulthood is poorly understood.

What This Study Adds:

In a setting where exclusive breastfeeding is not associated with higher socioeconomic position or adiposity, exclusively breastfeeding in early infancy was associated with a healthier lipid profile in late adolescence, potentially through mechanisms unrelated to adiposity.

Cardiovascular disease is 1 of the leading causes of death globally.1 Low-density lipoprotein cholesterol (LDL-C) is a key contributor to cardiovascular disease.2,3 Whether breastfeeding has long-term benefits on lipid metabolism in adulthood has important implications for combating the epidemic because infant feeding is highly modifiable. Breast milk contains more cholesterol than infant formula,4 and infants who are breastfed have higher blood cholesterol.5,6 Higher neonatal dietary cholesterol is associated with different cholesterol metabolism and less endogenous cholesterol synthesis in infants who are breastfed.6,7 It has been hypothesized that such early changes in cholesterol metabolism may program subsequent cholesterol homeostasis and lipid profile in adulthood.8 

A randomized controlled trial (RCT) in preterm infants in the 1980s revealed a healthier lipid profile at 13 to 16 years (ie, lower ratios of both low-density lipoprotein (LDL)/high-density lipoprotein and apolipoprotein B/apolipoprotein A-1) among preterm infants fed banked breast milk compared with those fed preterm formula.9 However the RCT had a low follow-up rate (23%), hence potential selection bias despite the use of an intention-to-treat analysis. An RCT of breastfeeding promotion, the Promotion of Breastfeeding Intervention Trial, revealed no effect on apolipoprotein A-1 at 11.5 years.10 However, the impact of breastfeeding on lipids may only become evident in adulthood.5 A meta-analysis of 17 observational studies from Western settings reveled a modest association of breastfeeding (particularly exclusive breastfeeding) with lower total cholesterol in adulthood,11 although the studies were somewhat heterogeneous (P = .02), and the included studies are open to confounding because breastfeeding in Western settings is usually associated with higher socioeconomic position,12 which is also linked to better health.

Hong Kong is a developed non-Western setting with a standard of living and social infrastructure similar to Western Europe and North America; however, for historic and cultural reasons, in the 1990s, breastfeeding was not clearly associated with socioeconomic position.13 As such, the Hong Kong Children of 1997 birth cohort14 provides an opportunity to unravel the effects of breastfeeding on health.15 Notably, studies of breastfeeding in Hong Kong have replicated findings from the Promotion of Breastfeeding Intervention Trial, such as a lack of association of exclusive breastfeeding with BMI,16,17 blood pressure,18 incidence of asthma,19 and behavioral outcomes20 in childhood or adolescence. Here we assessed the association of infant feeding with lipid profile and adiposity at ∼17.5 years. We also assessed whether associations varied by sex because some studies suggested a stronger effect in women.21,23 

The Hong Kong Children of 1997 birth cohort is a population-representative Chinese birth cohort (N = 8327) that covered 88% of all births from April 1, 1997, to May 31, 1997. The study was initially established to investigate the impact of second-hand smoke exposure on infant health.14 Families were recruited at their first postnatal visit to the 49 maternal and child health centers, which parents of all newborns are encouraged to attend for free postnatal care, developmental checks, and vaccinations until the age of 5 years. Baseline characteristics were obtained at recruitment by using a self-administered questionnaire in Chinese that included information on birth characteristics (birth weight, birth order, sex, and gestational age), parents’ education, infant feeding, and second-hand smoke exposure. Passive follow-up via record linkage was instituted in 2005 to obtain the following: (1) weight and height from 0 to 5 years from the maternal and child health centers; (2) annual measurements of weight and height (grade 1 onward) and biannual assessments of pubertal status (grade 1 onward), physiologic well-being (grade 2 onward), and blood pressure (grade 5 onward) from the Student Health Service, Department of Health, which provides free annual check-ups for all school students; (3) hospital discharge records from the Hospital Authority, which manages all public hospitals; and (4) death records from the death registry. Active follow-up via direct contact was instituted in 2007, with 3 postal or telephone surveys conducted in 2008–2012.

A biobank clinical follow-up was conducted in 2013–2016. Body weight and body fat percentage were measured in the fasting state by trained staff using the bioimpedance machine (body composition monitor BD545; Tanita Corporation, Japan), and standing height was measured by using a stadiometer. At the same visit, participants’ overnight fasting blood samples were taken and assayed for lipids. High-density lipoprotein cholesterol (HDL-C) was measured by using a homogeneous enzymatic colorimetric test with PEG coupled with cholesterol esterase and cholesterol oxidase, whereas triglycerides were quantitated by using the enzymatic colorimetric method with glycerol and peroxidase (Roche cobas C8000 system; Germany). LDL-C was estimated by the Friedewald equation.

As a token of appreciation, we offered supermarket coupons worth 200 Hong Kong dollars (24 US dollars) per participant. This follow-up study was reviewed by and received approval from The University of Hong Kong-Hospital Authority Hong Kong West Cluster Joint Institutional Review Board and the Joint Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee. Informed written consent from the participants (or from the parents and assent from the participants <18 years old at the time of follow-up) was obtained.

Infant feeding in the first 3 months was categorized, on the basis of the most contemporaneous report, as always formula fed, (ie, never breastfed), mixed fed (ie, fed both breast milk and formula milk), or exclusively breastfed, consistent with our previous studies on breastfeeding in this birth cohort.16,18,20 As previously described, mixed feeding during the first 3 months of life, (ie, partially breastfed) included those partially breastfed for any length of time or exclusively breastfed for <3 months. Information on infant feeding was collected by using standardized questionnaires administered shortly after birth and at 3, 9, and 18 months. At recruitment, shortly after birth, the respondent, mainly the mother, was asked, “How is the infant currently fed?” to which the respondent could answer with 1 of the following: “exclusively breastfed,” “partially breastfed,” and “only formula fed.” At each of the follow-up visits at 3, 9, and 18 months, the respondent was asked, “How has the infant been fed from birth until now?” to which the respondent could answer with 1 of the 3 original options or with an additional option that applied to the initial “exclusively breastfed” option, “initially breastfed, but now formula fed,” and provide the age in months when breastfeeding was terminated.

The main outcomes were total cholesterol, LDL-C, HDL-C, and triglycerides, considered as continuous variables. We used BMI and body fat percentage as markers of adiposity. Age (in days) and sex-specific z scores for BMI relative to the 2007 World Health Organization (WHO) growth reference were calculated by interpolating the WHO references onto a daily scale by using the akima package in R (R Development Core Team, Vienna, Austria). We also defined overweight as a BMI for age and sex corresponding to an adult BMI of ≥25 using the International Obesity Task Force cutoffs,24 which were interpolated to obtain cutoffs for overweight at each age in days by using the akima package in R.

Baseline characteristics of the participants included and excluded were compared first by using χ2 tests (for categorical variables) and t tests (for continuous variables) and second by using Cohen’s w effect size (for categorical variables) and Cohen’s d effect size (for continuous variables), where <0.1 and <0.2, respectively, indicate small difference between groups. We used multivariable linear regression to assess the adjusted association of infant feeding with lipids and adiposity at ∼17.5 years, adjusting for age at follow-up and potential confounders. Confounders, selected as likely common factors influencing choices of infant feeding and subsequent lipids and/or adiposity, included sex, birth weight, parity, gestational age, pregnancy characteristics (gestational diabetes, preeclampsia, maternal age, and maternal smoking during pregnancy), mother’s place of birth, and parents’ highest education. We tested whether the associations varied by sex from the significance of interaction terms after including interactions with all potential confounders. We obtained the P for trend for the association of breastfeeding exclusivity with lipids and adiposity by treating the 3 infant feeding categories (always formula fed, mixed fed, or exclusively breastfed) as a continuous variable.

To minimize potential biases due to missing exposure and/or covariates and differences between respondents and nonrespondents in the biobank clinical follow-up, we combined multiple imputation and inverse probability weighting.25 We used multiple imputation to predict missing confounders on the basis of a flexible additive regression model, with predictive mean matching used to incorporate data on infant feeding, potential confounders, interactions of interest (ie, infant feeding by sex), and the outcomes (ie, lipids and adiposity).26 Confounders were imputed for <5%. We calculated the probability of an observation being present using logistic regression to identify predictors of being included in the follow-up (including sex, birth size, birth order, gestational age, maternal smoking during pregnancy, mother’s place of birth, maternal age, parents’ highest education, household income at birth, and BMI at 12 years), from which we obtained individual inverse probability weights of response. We summarized the weighted results from 50 imputed data sets into single estimates, with confidence intervals (CIs) adjusted for uncertainty due to missing data. Statistical analyses were performed by using R version 3.3.2.

Of the original 8327 cohort participants, as of December 2017, 8298 remained in the Children of 1997 birth cohort, and 29 had permanently withdrawn. For the clinical follow-up, 6850 participants were contactable, of whom 3460 attended the clinical follow-up at 16 to 18 years old; 3261 participants who provided fasting blood samples were included in the analyses. There were differences in the distributions of some baseline characteristics by inclusion in the study. However, the magnitude of the differences was small (Cohen’s d <0.2 for continuous variables; Cohen’s w <0.1 for categorical variables), and thus the included participants were comparable to the excluded participants in terms of mode of infant feeding, birth characteristics, pregnancy conditions, and markers of socioeconomic positions (Table 1). As in the entire cohort, included participants who were exclusively breastfed for at least 3 months were more likely to be girls but less likely to be exposed to preeclampsia and maternal smoking (Table 2). Their mothers were more likely to be born in the rest of China and have lower education. Exclusive breastfeeding for 3 months was not associated with parents’ higher education.

TABLE 1

Baseline Characteristics Between Included and Excluded Participants

Included (n = 3261)Excluded (n = 5037)PaEffect Sizeb
Boys, % 51 54 .03 0.02 
Birth wt z score, meanc −0.24 −0.26 .12 0.04 
Parity, %     
 First 47 48 .76 <0.01 
 Second 42 41 — — 
 Greater than or equal to third 11 11 — — 
Gestational age, wk, %     
 ≤36 4.6 5.7 .38 0.03 
 37–38 28 30 — — 
 39–40 53 52 — — 
 ≥41 13 13 — — 
Gestational diabetes, % 7.3 5.2 .002 0.04 
Preeclampsia, % 4.0 3.4 .24 0.01 
Maternal smoking during pregnancy, % 3.9 6.0 <.001 0.04 
Maternal age at delivery, y, %     
 ≤24 10 14 <.001 0.07 
 25–29 31 31 — — 
 30–34 39 37 — — 
 ≥35 20 17 — — 
Infant feeding during 0–3 mo, %     
 Always formula fed 53 60 <.001 0.07 
 Mixed 40 35 — — 
 Exclusively breastfed 7.5 5.4 — — 
Parents’ highest education, %     
 Ninth grade or below 29 31 .04 0.03 
 10th–11th grade 43 42 — — 
 ≥12th grade 28 26 — — 
Household income per head in quintiles, %     
 First 19 21 .07 0.03 
 Second 20 21 — — 
 Third 20 19 — — 
 Fourth 20 19 — — 
 Fifth 21 19 — — 
Hong Kong–born mother, % 58 62 <.001 0.04 
Included (n = 3261)Excluded (n = 5037)PaEffect Sizeb
Boys, % 51 54 .03 0.02 
Birth wt z score, meanc −0.24 −0.26 .12 0.04 
Parity, %     
 First 47 48 .76 <0.01 
 Second 42 41 — — 
 Greater than or equal to third 11 11 — — 
Gestational age, wk, %     
 ≤36 4.6 5.7 .38 0.03 
 37–38 28 30 — — 
 39–40 53 52 — — 
 ≥41 13 13 — — 
Gestational diabetes, % 7.3 5.2 .002 0.04 
Preeclampsia, % 4.0 3.4 .24 0.01 
Maternal smoking during pregnancy, % 3.9 6.0 <.001 0.04 
Maternal age at delivery, y, %     
 ≤24 10 14 <.001 0.07 
 25–29 31 31 — — 
 30–34 39 37 — — 
 ≥35 20 17 — — 
Infant feeding during 0–3 mo, %     
 Always formula fed 53 60 <.001 0.07 
 Mixed 40 35 — — 
 Exclusively breastfed 7.5 5.4 — — 
Parents’ highest education, %     
 Ninth grade or below 29 31 .04 0.03 
 10th–11th grade 43 42 — — 
 ≥12th grade 28 26 — — 
Household income per head in quintiles, %     
 First 19 21 .07 0.03 
 Second 20 21 — — 
 Third 20 19 — — 
 Fourth 20 19 — — 
 Fifth 21 19 — — 
Hong Kong–born mother, % 58 62 <.001 0.04 

—, not applicable.

a

P values were calculated by using the χ2 test for categorical variables and the t test for continuous variables.

b

Effect size was calculated by using Cohen’s w effect size, the Φ coefficient or Cramer’s V (for categorical variables), and Cohen’s d effect size (for continuous variables), in which <0.1 and <0.2, respectively, indicate that the difference between groups is small.

c

Referenced to the 2006 WHO growth standard.

TABLE 2

Baseline Characteristics by Infant-Feeding Type in the First 3 Months of Life in 3261 Included Participants

CharacteristicsAlways Formula Fed (n = 1724)Mixed (n = 1292)Exclusively Breastfed (n = 245)
Boys, %a 51 53 41 
Birth wt z score, meanb −0.23 −0.25 −0.23 
Parity, %a    
 First 42 54 41 
 Second 45 37 43 
 Greater than or equal to third 13 8.6 16 
Gestational age, wk, %a    
 ≤36 4.1 5.5 3.4 
 37–38 31 25 27 
 39–40 51 55 56 
 ≥41 14 14 14 
Gestational diabetes, % 7.6 6.7 7.9 
Preeclampsia, % 4.7 3.4 1.9 
Maternal smoking during pregnancy, %a 5.4 2.5 1.1 
Maternal age at delivery, %, ya    
 ≤24 9.7 10 11 
 25–29 32 29 36 
 30–34 37 42 36 
 ≥35 21 19 17 
Parents’ highest education, %a    
 Ninth grade or below 31 23 41 
 10th–11th grade 47 40 36 
 ≥12th grade 22 37 23 
Household income per head in first quintile, %a 19 16 35 
Hong Kong born–mother, %a 60 60 27 
CharacteristicsAlways Formula Fed (n = 1724)Mixed (n = 1292)Exclusively Breastfed (n = 245)
Boys, %a 51 53 41 
Birth wt z score, meanb −0.23 −0.25 −0.23 
Parity, %a    
 First 42 54 41 
 Second 45 37 43 
 Greater than or equal to third 13 8.6 16 
Gestational age, wk, %a    
 ≤36 4.1 5.5 3.4 
 37–38 31 25 27 
 39–40 51 55 56 
 ≥41 14 14 14 
Gestational diabetes, % 7.6 6.7 7.9 
Preeclampsia, % 4.7 3.4 1.9 
Maternal smoking during pregnancy, %a 5.4 2.5 1.1 
Maternal age at delivery, %, ya    
 ≤24 9.7 10 11 
 25–29 32 29 36 
 30–34 37 42 36 
 ≥35 21 19 17 
Parents’ highest education, %a    
 Ninth grade or below 31 23 41 
 10th–11th grade 47 40 36 
 ≥12th grade 22 37 23 
Household income per head in first quintile, %a 19 16 35 
Hong Kong born–mother, %a 60 60 27 
a

Significant difference at P < .05 by mode of infant feeding was assessed by using χ2 tests for categorical parameters and analysis of variance for continuous parameters.

b

Referenced to the 2006 WHO growth standard.

Few included participants (mean age of 17.4 ± 0.5 years) had elevated LDL-C (<1% with ≥159 mg/dL) levels, elevated triglyceride levels (3.9% boys and 1.8% girls with ≥150 mg/dL) or low HDL-C (2.9% boys with ≤39 mg/dL and 16% girls with ≤50 mg/dL) levels. Participants with overweight (including obesity; (16% boys and 9% girls) had higher total cholesterol, LDL-C, and triglycerides but lower HDL-C (data not shown).

We observed a graded relation of breastfeeding exclusivity in the first 3 months of life with lower total cholesterol and LDL-C (P for trend = .02 and .05, respectively), although none of the associations of infant feeding with lipids varied by sex (Table 3). Compared with always formula feeding (never breastfed), exclusive breastfeeding for the first 3 months was associated with lower total and LDL-C and possibly lower triglycerides but similar HDL-C at ∼17.5 years, adjusting for sex, birth weight (kg), parity, gestational weeks, pregnancy characteristics (including gestational diabetes, preeclampsia, maternal age, and maternal smoking), parents’ education, mother’s place of birth, and mother’s age (years) at measurement (Table 3). There was no difference in lipid profiles between mixed feeding and formula feeding, except that HDL-C was marginally lower among those who were mixed fed. These associations of early infant feeding with the lipid profile at ∼17.5 years remained unchanged when further adjusted for BMI or fat percentage (data not shown). BMI and body fat percentage at ∼17.5 years did not differ by type of infant feeding, with no difference by sex (Table 3).

TABLE 3

Adjusted Differences (95% CI) in Adiposity and Lipid Profile at ∼17.5 Years by Infant Feeding in the First 3 Months of Life

Mean (SD)Always Formula FedMixed, Adjusted Difference (95% CI)Exclusively Breastfed, Adjusted Difference (95% CI)P for TrendP for Sex Interaction
Adiposity       
 BMI 20.9 (3.5) Reference 0.1 (−0.1 to 0.4) 0.1 (−0.4 to 0.6) .28 .42 
 Body fat percentage 21.5 (8.8) Reference 0.0 (−0.5 to 0.5) −0.2 (−1.4 to 0.7) .80 .65 
Lipid profile       
 Total cholesterol, mg/dL 69.6 (28.2) Reference −1.7 (−3.8 to 0.3) −5.5 (−9.5 to −1.4) .01 .61 
 LDL-C, mg/dL 82.6 (25.1) Reference −0.6 (−2.5 to 1.2) −4.5 (−8.1 to −0.8) .05 .64 
 HDL-C, mg/dL 60.1 (19.2) Reference −1.4 (−2.8 to 0.0) −0.8 (−3.6 to 2.0) .11 .50 
 Triglycerides, mg/dL 71.0 (35.0) Reference −0.9 (−3.5 to 1.7) −4.3 (−9.4 to 0.8) .14 .49 
Mean (SD)Always Formula FedMixed, Adjusted Difference (95% CI)Exclusively Breastfed, Adjusted Difference (95% CI)P for TrendP for Sex Interaction
Adiposity       
 BMI 20.9 (3.5) Reference 0.1 (−0.1 to 0.4) 0.1 (−0.4 to 0.6) .28 .42 
 Body fat percentage 21.5 (8.8) Reference 0.0 (−0.5 to 0.5) −0.2 (−1.4 to 0.7) .80 .65 
Lipid profile       
 Total cholesterol, mg/dL 69.6 (28.2) Reference −1.7 (−3.8 to 0.3) −5.5 (−9.5 to −1.4) .01 .61 
 LDL-C, mg/dL 82.6 (25.1) Reference −0.6 (−2.5 to 1.2) −4.5 (−8.1 to −0.8) .05 .64 
 HDL-C, mg/dL 60.1 (19.2) Reference −1.4 (−2.8 to 0.0) −0.8 (−3.6 to 2.0) .11 .50 
 Triglycerides, mg/dL 71.0 (35.0) Reference −0.9 (−3.5 to 1.7) −4.3 (−9.4 to 0.8) .14 .49 

Adjusted for sex, birth weight (kg), parity, gestational weeks, pregnancy characteristics (including gestational diabetes, preeclampsia, maternal age, and maternal smoking), parents’ education, mother’s place of birth, and mother’s age (y) at measurement.

Exclusive breastfeeding, but not mixed feeding (ie, partial breastfeeding), in the first 3 months of life was associated with a less atherogenic lipid profile, characterized by lower total cholesterol and LDL-C at ∼17.5 years. Individuals who were exclusively breastfed, however, did not have a lower BMI or fat percentage, suggesting the potential long-term impact of exclusive breastfeeding on the lipid profile could possibly be nonattributable to adiposity. In our cohort, exclusive breastfeeding was not related to parents’ higher education, making confounding by socioeconomic position unlikely. As such, our findings are consistent with the hypothesis that exclusive breastfeeding in early infancy may protect against cardiovascular disease risk.

In our study, we used a population-representative birth cohort with contemporaneously reported information on infant feeding. Exclusive breastfeeding at the time when the birth cohort was established was not associated with higher socioeconomic position.16 Differential recall bias in retrospective studies with long-term recall of breastfeeding,27 and confounding by socioeconomic position in studies in high-income countries,12 may bias benefits of exclusive breastfeeding. Such confounding is unlikely in our study population, which adds strength to our study. The associations of exclusive breastfeeding, but not mixed breastfeeding, with healthier lipid profiles is consistent with a meta-analysis in which the authors reported lower total cholesterol among adults who had been exclusively breastfed but not among those who were ever breastfed.11 Our study suggested a potential association between mixed feeding and lower HDL-C. Although this requires further studies to confirm, the lack of association between mixed feeding and a better lipid profile is consistent with the null associations with any breastfeeding reported from meta-analyses.28,29 

Our study also reveals that the impact of breastfeeding on lipids is independent of adiposity because being exclusively breastfed was not associated with a lower BMI or fat percentage. Such a null association was not attributed to inaccurate BMI or fat percentage measures because they had expected associations with lipid profile. This null association between exclusive breastfeeding and lower adiposity was consistently observed in settings with little confounding by socioeconomic position, including our cohort,16,17 other observational studies,30 and an RCT.31 Exclusive breastfeeding was associated with lower total cholesterol and LDL-C but not with lower adiposity, suggesting that the long-term impact of exclusive breastfeeding in early infancy could be independent of mediating pathways related to adiposity, as shown in a previous study.32 

It has been well established that, compared with infants who are formula fed, infants who are breastfed have higher plasma cholesterol and less endogenous cholesterol synthesis. However, whether and how an exposure to high cholesterol in breast milk programs cholesterol homeostasis and lipid profile in adulthood, as hypothesized 40 years ago on the basis of limited evidence from rat models,8 is poorly understood. Potential mechanisms could involve changes in the expression of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase and LDL receptors, as shown in experiments in pigs33 and baboons.34 Authors of 1 study in humans reported upregulation of the fractional synthesis rate of cholesterol in infants who were formula fed at 18 months,7 but no study so far clarified the molecular mechanism for the long-term impact of breast milk on cholesterol metabolism. The genetic variant related to expression of HMG-CoA reductase (rs12916 in HMGCR) is associated with LDL-C but not with HDL-C,35 whereas genetic variants related to LDL receptors (rs11613352 [LRP1], rs3136441 [LRP4], and rs11206510 [PCSK4]) are associated with changes in both LDL-C and HDL-C. Thus, our observation that lower LDL-C, but not lower HDL-C, in individuals who are exclusively breastfed appears to be more consistent with the role of HMG-CoA reductase in the programing effect of early cholesterol exposure, if any. Such a programing effect may explain the change in lipid profile but not in markers of adiposity by type of infant feeding observed here. However, evidence is still lacking as to whether such a change in synthesis or metabolism of cholesterol in the neonatal period persists beyond weaning and into adulthood.7 Whether other differences between breast and formula milk (eg, a higher phytosterol, protein, or galactose level in formula milk) could program lipid metabolism, requires further research to elucidate.

There are some limitations to our study. First, attending the follow-up was voluntary, and we cannot rule out the possibility of selection bias. There were inevitable differences in the baseline characteristics between those included and excluded. However, the magnitude of all these differences, including the differences in the percentage of those exclusively breastfed for 3 months, were small. Infant feeding was not related to parents’ higher education, suggesting little bias from socioeconomic confounding for any observed benefits of breastfeeding. Multiple imputation with inverse probability weighting was used to minimize the impact from potential selection bias. Second, we have little information on diet after early infancy and on whether types of solid food given after weaning differed by mode of infant feeding. A higher proportion of participants who were exclusively breastfed had later solid food introduction, so we cannot rule out the possibility that the association of early infant feeding with lipids is due to subsequent dietary factors. However, in this cohort, later solid food introduction was not associated with either a higher childhood or adolescent BMI36 or a better lipid profile (data not shown). Third, BMI cannot distinguish body fat from lean mass and thus may overestimate adiposity of individuals with higher muscle mass. However, we obtained a similarly null association of exclusive breastfeeding with BMI and body fat percentage, and thus underestimation of the association of breastfeeding and adiposity was unlikely. Finally, we do not know exclusiveness of breastfeeding and cannot assess the impact of predominant breastfeeding on the adult lipid profile.

Our study suggests that exclusive breastfeeding in the first 3 months of life is associated with a less atherogenic lipid profile. Despite the modest impact on LDL-C, breastfeeding is potentially important in cardiovascular disease prevention by shifting the whole population distribution of cholesterol toward a lower level, as per the Rose37 prevention paradox. However, we cannot elucidate whether prolonged breastfeeding (ie, a longer exposure to a high-cholesterol diet during infancy) is harmful to cardiovascular health38,39 through other mechanisms because the breastfeeding duration was commonly short in the 1990s in Hong Kong. Further studies are also required to assess whether predominant breastfeeding also provides benefits. Finally, although the lipid profile could be attributed to the high cholesterol content in breast milk, this study does not provide direct evidence to support a long-term beneficial impact on the lipid profile from an increase in cholesterol in formula milk.

Exclusive breastfeeding in early infancy may be associated with a better lipid profile in late adolescence, suggesting its potential long-term benefits for cardiovascular health, which should be supported. Further studies used to unravel the biological mechanism through which breastfeeding programs lipid metabolism and the health impact of prolonged breastfeeding are warranted.

Drs Hui and Schooling conceptualized and designed the study, drafted the initial manuscript, and reviewed and revised the manuscript; Drs Kwok and Nelson, Ms Lee, and Dr Leung critically reviewed the manuscript and contributed substantially to the interpretation of data; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: This work is a substudy of the Children of 1997 birth cohort (which was initially supported by the Health Care and Promotion Fund, Health and Welfare Bureau, Government of the Hong Kong special administrative region [Health Care and Promotion Fund grant 216106] and reestablished in 2005 with support from the Health and Health Services Research Fund, Government of the Hong Kong special administrative region [Health and Health Services Research Fund grant 03040771]) and the University Research Committee Strategic Research Theme of Public Health, The University of Hong Kong. This substudy is used to build on information added to the birth cohort by RFCID grant 04050172 and Health and Health Services Research Fund grant 08090761 and was funded by the Health and Medical Research Fund, Government of the Hong Kong special administrative region (Health and Medical Research Fund grant 10111491) and the WYNG Foundation.

COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2019-0447.

We thank colleagues at the Student Health Service and Family Health Service of the Department of Health for their assistance and collaboration. We also thank the late Dr Connie O for coordinating the project and all the field work for the initial study in 1997–1998.

     
  • CI

    confidence interval

  •  
  • HDL-C

    high-density lipoprotein cholesterol

  •  
  • LDL

    low-density lipoprotein

  •  
  • LDL-C

    low-density lipoprotein cholesterol

  •  
  • RCT

    randomized controlled trial

  •  
  • WHO

    World Health Organization

1
GBD 2013 Mortality and Causes of Death Collaborators
.
Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013.
Lancet
.
2015
;
385
(
9963
):
117
171
[PubMed]
2
Shah
S
,
Casas
JP
,
Drenos
F
, et al
.
Causal relevance of blood lipid fractions in the development of carotid atherosclerosis: Mendelian randomization analysis.
Circ Cardiovasc Genet
.
2013
;
6
(
1
):
63
72
[PubMed]
3
Voight
BF
,
Peloso
GM
,
Orho-Melander
M
, et al
.
Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study.
Lancet
.
2012
;
380
(
9841
):
572
580
[PubMed]
4
Friedman
G
,
Goldberg
SJ
.
Concurrent and subsequent serum cholesterol of breast- and formula-fed infants.
Am J Clin Nutr
.
1975
;
28
(
1
):
42
45
[PubMed]
5
Owen
CG
,
Whincup
PH
,
Odoki
K
,
Gilg
JA
,
Cook
DG
.
Infant feeding and blood cholesterol: a study in adolescents and a systematic review.
Pediatrics
.
2002
;
110
(
3
):
597
608
[PubMed]
6
Wong
WW
,
Hachey
DL
,
Insull
W
,
Opekun
AR
,
Klein
PD
.
Effect of dietary cholesterol on cholesterol synthesis in breast-fed and formula-fed infants.
J Lipid Res
.
1993
;
34
(
8
):
1403
1411
[PubMed]
7
Demmers
TA
,
Jones
PJ
,
Wang
Y
,
Krug
S
,
Creutzinger
V
,
Heubi
JE
.
Effects of early cholesterol intake on cholesterol biosynthesis and plasma lipids among infants until 18 months of age.
Pediatrics
.
2005
;
115
(
6
):
1594
1601
[PubMed]
8
Reiser
R
,
Sidelman
Z
.
Control of serum cholesterol homeostasis by cholesterol in the milk of the suckling rat.
J Nutr
.
1972
;
102
(
8
):
1009
1016
[PubMed]
9
Singhal
A
,
Cole
TJ
,
Fewtrell
M
,
Lucas
A
.
Breastmilk feeding and lipoprotein profile in adolescents born preterm: follow-up of a prospective randomised study.
Lancet
.
2004
;
363
(
9421
):
1571
1578
[PubMed]
10
Martin
RM
,
Patel
R
,
Kramer
MS
, et al
.
Effects of promoting longer-term and exclusive breastfeeding on cardiometabolic risk factors at age 11.5 years: a cluster-randomized, controlled trial.
Circulation
.
2014
;
129
(
3
):
321
329
[PubMed]
11
Owen
CG
,
Whincup
PH
,
Kaye
SJ
, et al
.
Does initial breastfeeding lead to lower blood cholesterol in adult life? A quantitative review of the evidence.
Am J Clin Nutr
.
2008
;
88
(
2
):
305
314
[PubMed]
12
Brion
MJ
.
Commentary: assessing the impact of breastfeeding on child health: where conventional methods alone fall short for reliably establishing causal inference.
Int J Epidemiol
.
2010
;
39
(
1
):
306
307
[PubMed]
13
Lee
WT
,
Wong
E
,
Lui
SS
,
Chan
V
,
Lau
J
.
Decision to breastfeed and early cessation of breastfeeding in infants below 6 months old–a population-based study of 3,204 infants in Hong Kong.
Asia Pac J Clin Nutr
.
2007
;
16
(
1
):
163
171
[PubMed]
14
Schooling
CM
,
Hui
LL
,
Ho
LM
,
Lam
TH
,
Leung
GM
.
Cohort profile: ‘children of 1997’: a Hong Kong Chinese birth cohort.
Int J Epidemiol
.
2012
;
41
(
3
):
611
620
[PubMed]
15
Schooling
CM
,
Leung
GM
.
A socio-biological explanation for social disparities in non-communicable chronic diseases: the product of history?
J Epidemiol Community Health
.
2010
;
64
(
11
):
941
949
[PubMed]
16
Kwok
MK
,
Schooling
CM
,
Lam
TH
,
Leung
GM
.
Does breastfeeding protect against childhood overweight? Hong Kong’s ‘Children of 1997’ birth cohort.
Int J Epidemiol
.
2010
;
39
(
1
):
297
305
[PubMed]
17
Cheng
TS
,
Kwok
MK
,
Leung
GM
,
Schooling
CM
.
The associations of breast feeding with infant growth and body mass index to 16 years: ‘Children of 1997’.
Paediatr Perinat Epidemiol
.
2018
;
32
(
2
):
200
209
[PubMed]
18
Kwok
MK
,
Leung
GM
,
Schooling
CM
.
Breastfeeding and adolescent blood pressure: evidence from Hong Kong’s “Children of 1997” birth cohort.
Am J Epidemiol
.
2013
;
178
(
6
):
928
936
[PubMed]
19
Leung
JY
,
Kwok
MK
,
Leung
GM
,
Schooling
CM
.
Breastfeeding and childhood hospitalizations for asthma and other wheezing disorders.
Ann Epidemiol
.
2016
;
26
(
1
):
21
27.e3
20
Kwok
MK
,
Leung
GM
,
Schooling
CM
.
Breast feeding and early adolescent behaviour, self-esteem and depression: Hong Kong’s ‘Children of 1997’ birth cohort.
Arch Dis Child
.
2013
;
98
(
11
):
887
894
[PubMed]
21
Marmot
MG
,
Page
CM
,
Atkins
E
,
Douglas
JW
.
Effect of breast-feeding on plasma cholesterol and weight in young adults.
J Epidemiol Community Health
.
1980
;
34
(
3
):
164
167
[PubMed]
22
Kolacek
S
,
Kapetanović
T
,
Zimolo
A
,
Luzar
V
.
Early determinants of cardiovascular risk factors in adults. A. Plasma lipids.
Acta Paediatr
.
1993
;
82
(
8
):
699
704
[PubMed]
23
Williams
MJ
,
Williams
SM
,
Poulton
R
.
Breast feeding is related to C reactive protein concentration in adult women.
J Epidemiol Community Health
.
2006
;
60
(
2
):
146
148
[PubMed]
24
Cole
TJ
,
Bellizzi
MC
,
Flegal
KM
,
Dietz
WH
.
Establishing a standard definition for child overweight and obesity worldwide: international survey.
BMJ
.
2000
;
320
(
7244
):
1240
1243
[PubMed]
25
Seaman
SR
,
White
IR
,
Copas
AJ
,
Li
L
.
Combining multiple imputation and inverse-probability weighting.
Biometrics
.
2012
;
68
(
1
):
129
137
[PubMed]
26
Moons
KG
,
Donders
RA
,
Stijnen
T
,
Harrell
FE
 Jr
.
Using the outcome for imputation of missing predictor values was preferred.
J Clin Epidemiol
.
2006
;
59
(
10
):
1092
1101
[PubMed]
27
Huttly
SR
,
Barros
FC
,
Victora
CG
,
Beria
JU
,
Vaughan
JP
.
Do mothers overestimate breast feeding duration? An example of recall bias from a study in southern Brazil.
Am J Epidemiol
.
1990
;
132
(
3
):
572
575
[PubMed]
28
Horta
BL
,
Loret de Mola
C
,
Victora
CG
.
Long-term consequences of breastfeeding on cholesterol, obesity, systolic blood pressure and type 2 diabetes: a systematic review and meta-analysis.
Acta Paediatr
.
2015
;
104
(
467
):
30
37
[PubMed]
29
Horta
BL
,
Victora
CG
.
Long-Term Effects of Breastfeeding: A Systematic Review
.
Geneva, Switzerland
:
World Health Organization
;
2013
30
Brion
MJ
,
Lawlor
DA
,
Matijasevich
A
, et al
.
What are the causal effects of breastfeeding on IQ, obesity and blood pressure? Evidence from comparing high-income with middle-income cohorts.
Int J Epidemiol
.
2011
;
40
(
3
):
670
680
31
Martin
RM
,
Patel
R
,
Kramer
MS
, et al
.
Effects of promoting longer-term and exclusive breastfeeding on adiposity and insulin-like growth factor-I at age 11.5 years: a randomized trial.
JAMA
.
2013
;
309
(
10
):
1005
1013
[PubMed]
32
Parikh
NI
,
Hwang
SJ
,
Ingelsson
E
, et al
.
Breastfeeding in infancy and adult cardiovascular disease risk factors.
Am J Med
.
2009
;
122
(
7
):
656
663.e1
33
Jones
PJ
,
Hrboticky
N
,
Hahn
P
,
Innis
SM
.
Comparison of breast-feeding and formula feeding on intestinal and hepatic cholesterol metabolism in neonatal pigs.
Am J Clin Nutr
.
1990
;
51
(
6
):
979
984
[PubMed]
34
Mott
GE
,
Jackson
EM
,
DeLallo
L
,
Lewis
DS
,
McMahan
CA
.
Differences in cholesterol metabolism in juvenile baboons are programmed by breast- versus formula-feeding.
J Lipid Res
.
1995
;
36
(
2
):
299
307
[PubMed]
35
Teslovich
TM
,
Musunuru
K
,
Smith
AV
, et al
.
Biological, clinical and population relevance of 95 loci for blood lipids.
Nature
.
2010
;
466
(
7307
):
707
713
[PubMed]
36
Lin
SL
,
Leung
GM
,
Lam
TH
,
Schooling
CM
.
Timing of solid food introduction and obesity: Hong Kong’s “children of 1997” birth cohort.
Pediatrics
.
2013
;
131
(
5
). Available at: www.pediatrics.org/cgi/content/full/131/5/e1459
[PubMed]
37
Rose
G
.
Strategy of prevention: lessons from cardiovascular disease.
Br Med J (Clin Res Ed)
.
1981
;
282
(
6279
):
1847
1851
[PubMed]
38
Fall
CH
,
Barker
DJ
,
Osmond
C
,
Winter
PD
,
Clark
PM
,
Hales
CN
.
Relation of infant feeding to adult serum cholesterol concentration and death from ischaemic heart disease.
BMJ
.
1992
;
304
(
6830
):
801
805
[PubMed]
39
Leeson
CP
,
Kattenhorn
M
,
Deanfield
JE
,
Lucas
A
.
Duration of breast feeding and arterial distensibility in early adult life: population based study.
BMJ
.
2001
;
322
(
7287
):
643
647
[PubMed]

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.