Secondary Logo

Journal Logo

Original Articles: Nutrition

Differences in Postprandial Lipid Response to Breast- or Formula-feeding in 8-Week-Old Infants

Teller, Inga C.*; Schoen, Stefanie*; van de Heijning, Bert*; van der Beek, Eline M.*; Sauer, Pieter J.J.

Author Information
Journal of Pediatric Gastroenterology and Nutrition: April 2017 - Volume 64 - Issue 4 - p 616-623
doi: 10.1097/MPG.0000000000001341


What Is Known

  • Breast-feeding is optimal for infant nutrition and associated with short- and long-term health.
  • Distinct features of human milk lipids (ie, droplet size and architecture, structural composition of triglycerides, and gradual fat delivery during a breast-feeding session) differ from formulae and may affect digestion.

What Is New

  • Our study suggests that human milk and formula lipids may be handled differently by neonates.
  • The triglyceride response of formula-fed subjects started at 60 minutes, and thus earlier than that of breast-fed subjects.
  • Chylomicrons in breast-fed subjects carry a higher fat load, compensated by a higher chylomicron number in formula-fed subjects.

Dietary lipids, preferably from human milk (HM), are a key energy source to meet the high energy requirements of the fast-growing infant. In addition to their nutritional value, dietary lipids including fat-soluble vitamins play important roles as signaling molecules involved in infant growth and metabolism, gene expression regulation, as well as in the development and modulation of cardiovascular, immune, and nervous system functions (1,2).

The lipid components in HM vary highly over the time of day, lactational stage, maternal nutrition status, and dietary intake (1,3,4). The HM lipid moiety is distinctly different from infant milk formulae (IF) as well with regard to fatty acid composition, the presence of cholesterol, lipid droplet size and architecture (5–7), structural composition of triglycerides (TG) (8,9). There is a gradual release of lipids during a breast-feeding (BF) session (foremilk/hindmilk) in contrast to constant concentrations during formula-feeding (FF) (10). These features may directly affect lipid digestion, absorption, and bioavailability as shown previously in rodents, (preterm) infants, and adults (7,8,11–16). This may potentially have long-term consequences. Many of these physiological and metabolic aspects regarding lipid delivery have not been studied in healthy term infants.

By measuring lipid-related plasma concentrations of TG, total Cholesterol, HDL-cholesterol (HDL), LDL-cholesterol (LDL), apolipoproteina a1 (ApoA1), ApoB48, and total ApoB (ie, ApoB48 plus ApoB100) we estimated postmeal responses of 2-month-old healthy term infants to dietary lipids and compared BF with FF.


Study Design, Recruitment, and Procedure

An exploratory, observational single-center study with invasive measures (heel prick) was conducted between April 2010 and January 2012 in Groningen, the Netherlands. The study was named EAGLE 1 after a fictional space craft exploring unknown territory. Study protocol and documents were approved by the medical ethical committee of the University Medical Centre Groningen (UMCG/METc 2009.369).

Participation was voluntary and written informed consent obtained from both parents. To strictly avoid interfering with parents’ feeding choice, FF subjects were only included when the decision to exclusively feed formula was made study independently and before inclusion (Fig. 1). FF subjects received free formula once informed consent was given. For these subjects, inclusion was possible up to the age of 4 weeks. For the exclusively BF subjects, inclusion was possible until the study day at 8 weeks of age. Recruitment was the responsibility of the investigator who made the study known in hospitals and “well-baby clinics.”

Study design and procedure at study day. Exclusively breast-fed infants (dark gray) were recruited until shortly before the study day whereas exclusively formula-fed infants (light gray) were included until 4 weeks of age so that they received the same formula for at least 4 weeks before testing. Study day: during a home visit of the study nurse before the second daily feeding, subject's anthropometric measures and a capillary baseline sample (t0) were taken. After the feeding, subjects were weighed again and randomized to the timing of the second blood sampling at 1 of the 6 sample times (t1) (W = anthropometrics and test weighing; blood drop symbol = heel prick).

Healthy term subjects with normal birth and body weight at testing (>10th percentile of locally applicable growth charts for sex and age) were included. Exclusion criteria were congenital or metabolic diseases or compromised maternal health (hepatitis B, HIV, high blood pressure, hyperlipidemia, diabetes, overweight [body mass index ≥27] before pregnancy), and nonexclusive/mixed feeding. Subjects that needed other than standard cow's milk-based formula or that had received medical treatment within 4 days before study day were excluded. At the study day, all subjects received a small cuddle toy. Parents had the right to resign from the study at any time. A sequential identification number and subject code (no initials) were assigned to each subject at study entry, stored at the site, and accessible only to study staff ensuring subject anonymity.

Study Day

The study amounted to 1 test day (Fig. 1): before the age of 2 months (55 ± 5 days), parents contacted the study nurse after finishing the first feeding after 06:00 AM and arranged a home visit before the next feeding. Informed consent was confirmed, a final eligibility check performed, and data collected. Before the second feeding of the day, a capillary baseline blood sample (t0, 1 mL) was obtained, collected in heparin-coated tubes (BD Microtainer Contact-Activated Lancet, BD Diagnostics, Franklin Lakes, NJ), and stored on ice. Subjects were then fed in their usual manner either per bottle (FF) or by nursing (BF); feeding start and end were recorded. HM composition was not assessed during the present study. Subjects were randomly assigned to 1 of 6 postprandial blood collection times using an allocation list generated by a DOS-based software RANDOM provided by Nutricia Research, The Netherlands. The permuted block randomization was stratified for the type of feeding (BF, FF). The second blood sample (t1) was collected either at 30, 60, 90, 120, 180, or 240 minutes after the first sample and immediately stored on ice. All subjects were weighed before and after the meal, the difference being a proxy for volume intake (test weighing); bottle weights were recorded before and after feeding for the FF group for more precise volume intake measures in this group. Blood samples were delivered within 1 hour after collecting the second blood sample to the laboratory, centrifuged, and plasma and pellets stored separately at −80°C until analysis.

Infant Formula

After inclusion in the study, the formula-fed infants received Bambix 1 Zuigelingenmelk (Nutricia B.V., The Netherlands), a generic commercial cow's milk-based formula for infants’ age range 0 to 6 months compliant to EU directives for infant feeding (Directive 2006/141/EC). The prepared formula contained 66 kcal per 100 mL, and comprised protein 1.3 g, carbohydrates 7.4 g (of which 0.6g per 100 ml > 0.8g per 100 kcal oligosaccharides), fat 3.4 g (of which 1.5 g saturated fatty acids, 1.3 g mono-unsaturated, and 0.6 g poly-unsaturated fatty acids (of which linoleic acid 447 mg, alpha-linolenic acid 83 mg, 11 mg arachidonic, 6.4 mg docosahexanoic acid, and traces of cholesterol (<0.0003 g). Bambix 1 formula lipids originated from a vegetable oil blend, representative of the most common lipid origins in infants formulae without addition of specific fat components (such as milk fat globule membrane fractions, structured lipids, etc) that would change the generic character of the formula as reviewed in more detail, recently (17). A full description of all components of the formula is given in Supplemental Digital Content 1, Table,

Plasma Parameters

Primary and secondary parameters were total plasma TG concentrations at t0 and t1 and differences between groups at each postmeal time point, respectively. In addition to TG, total cholesterol, HDL, LDL, ApoA1, and ApoB48, total ApoB (ApoB48 plus ApoB100) concentrations were measured. Ratios of HDL:LDL, ApoA1:total ApoB, and TG:ApoB48 were calculated. From the available plasma sample pairs with sufficient volume, composite postmeal concentration profiles were generated combining data from multiple subjects. For this reason, the terms “postprandial” or “kinetics” were avoided in favor of “postmeal” and “lipid handling” as chronological blood collection from the same individual would be implied otherwise. Data are presented as (1) baseline, (2) postmeal concentrations not corrected for baseline (*PM), or (3) postmeal concentrations corrected to each subjects’ own baseline (delta/Δ).

Plasma Analyses

Plasma analyses were performed at UMCG. Briefly, TG, total Cholesterol, HDL, and LDL were measured together applying combined enzymatic and colorimetric assays using a Roche Modular P Chemistry Analyser (Roche, Mannheim, Germany); Apolipoprotein A1 and total ApoB concentrations were assessed with specific antisera in the Siemens BN*II ProSpec system (Siemens Healthcare Diagnostics Products GmbH, Marburg, Germany). ApoB48 concentrations were measured with the Shibayagi sandwich enzyme-linked immunosorbent assay kit (Shibayagi Co, Ltd, Shibukawa, Gunma, Japan via BioVendor GmbH, Heidelberg, Germany) according to manufacturer's specifications.

Approximated Fat Intake/Calculations

Approximated fat intake per body weight was calculated by relating feeding volume estimated by test weighing (see above) corrected for milk and formula density (HM 1.03 g/mL and IF 1.0 g/mL) (18) to the average body weight at testing (Table 1) assuming that the mature term HM contained 3.5 g/dL fat (19).

Subject characteristics of the per protocol group are shown as mean and median for the BF and FF groups


Little to no data were available from literature for sample size calculations. Considering the burden of collecting 2 blood samples from infants, we estimated that the minimal sample size to generate informative results would be 5 subjects/ sample time totaling 30 subjects per feeding group. τ test comparison or Wilcoxon Mann-Whitney test were used as indicated. P values <0.05 were significant, between 0.05 and <0.1 were considered worth mentioning because of the small sample size. SAS/STAT version 9.2 software (SAS Institute, Cary, NC) was used.


Study Population

Seventy six subjects were enrolled, 4 were withdrawn by parents before testing, 2 did not meet study criteria leaving 70 subjects in the intention-to-treat population; 3 subjects were excluded because of nonresolvable errors in blood collection, leaving 67 subjects in the per-protocol population (BF, n = 33) for which subject characteristics are shown (Table 1). In this population, 59 plasma sample pairs with sufficient volume were available for analysis (baseline t0/postmeal sample time t1) (BF, n = 30); mean subject age was 55.3 ± 2.8 days. Subject characteristics (gestational age, weight, and length at birth and at test day) as well as sex distribution between feeding regimens were not statistically different between groups. All subjects followed their respective feeding regimen exclusively. More details on the study population are shown in Table 1; no (serious) adverse event occurred throughout the study.

Food Intake/Feeding Duration/Feeding Number

Significant differences between groups were observed (P < 0.001) for number of meals in the 24 hours before testing with a higher frequency of feeding moments in the BF group (BF 6.55 ± 0.83, FF 5.53 ± 0.56) (Table 1). Subsequently the feeding interval before the study meal was shorter for this group (BF, 182.91 ± 22.86 minutes; FF, 214.09 ± 30.77 minutes) (Table 1). Duration of the study meal was 5 minutes longer in the BF group (BF, 20.27 ± 7.70 minutes vs FF, 14.82 ± 3.57 minutes).

Feeding volume was, however, not statistically different as indicated by body weight (BF, 174.85 ± 43.49 g; FF, 166.62 ± 25.78 g) and bottle weight differences (FF, 164.07 ± 22.87 g) before and after feeding. Feeding duration did not correlate with weight change (data not shown).

Approximated Fat Intake

We found no statistical difference between approximated fat intake per body weight (BF, 1.65 ± 0.42 g/kg body weight; FF, 1.55 ± 0.27 g/kg body weight) when calculated as described above.

Postprandial Plasma Lipid Concentrations

Triglycerides, ApoB48, and TG:ApoB48 Ratio

TG baselines were not statistically different between groups (Fig. 2A). Postmeal TG concentrations seemed higher in BF than FF subjects at 30 minutes and higher in FF than BF subjects at 60 minutes (*PM; Fig. 3A). Once values were corrected for baseline (Δ), the response of FF over BF subjects at 60 minutes was confirmed, but the TG median of the BF group increased above baseline only at 90 minutes and did not statistically differ from FF subjects at this time. ΔTG concentrations of all FF subjects remained at or above baseline within 2 hours after feeding. In contrast, ΔTG concentrations for several (but not all) BF subjects substantially fell below baseline in the 2 hours after feeding. In both groups, the postmeal response took place between 30 and 180 minutes, after which TG concentrations (Δ) returned to baseline.

Baseline concentration scatter plots of exclusively breast-fed (black circles) and exclusively formula-fed (open triangles) healthy term infants before the next meal. Each symbol represents 1 subject. Median and interquartile ranges (25/75%) are shown. Statistical significances by Wilcoxon-Mann-Whitney test are shown where applicable (n = 30 BF/n = 29 FF). BF = breast-feeding; FF = formula-feeding.
Postmeal responses. For each of the selected lipid-related parameters, the upper panel shows measured postmeal (*PM) concentrations; the lower panel data that are corrected for each subject's own baseline (Delta, Δ) of exclusively breast-fed (dark circles) and exclusively formula fed (open triangles) subjects after their respective feeding. Each symbol represents 1 subject. The median is shown as gray bar. Observations with P < 0.1, established by Wilcoxon-Mann-Whitney test, were considered relevant findings due to the small sample size per assessment time.

Baseline ApoB48 concentrations were significantly higher in FF than BF subjects (Fig. 2B) and postmeal concentrations (*PM) increased between 120 and 180 minutes in FF subjects only, returning to baseline after 4 hours (Supplemental Digital Content 2, Table In contrast, postmeal concentrations in BF subjects (*PM) did not vary over time resulting in biggest differences between groups at 120 and 180 minutes (P = 0.03 and 0.074, respectively). Once corrected for baseline, postmeal ApoB48 concentrations (Δ) were not statistically different between groups and stayed close to baseline.

The ratio of TG:ApoB48 was higher in BF than FF subjects at baseline and consistently elevated over the whole postmeal period (*PM), significantly different at 180 minutes (Figs. 2C and 3B). Postmeal TG:ApoB48 concentrations (Δ) were, however, not different between groups once corrected for baseline.

Total Cholesterol and Related Parameters

Baseline total cholesterol and total cholesterol postmeal concentrations (*PM) were higher in BF than FF subjects at all times (Fig. 2D and Fig. 3C). In BF subjects, total cholesterol concentrations (Δ) increased in response to the meal, whereas in FF subjects, total cholesterol concentrations (Δ) remained close to baseline at all times (Fig. 3C).

The other cholesterol-related parameters (HDL, LDL, ApoA1, total ApoB, and ratios of HDL:LDL and ApoA1:total ApoB), also showed some pre- and postmeal differences: baseline concentrations of HDL, its lipoprotein ApoA1, LDL, total ApoB all were higher in BF subjects. Corrected for baseline, total ApoB concentrations (Δ) showed a postmeal response similar in timing but not magnitude as total cholesterol concentrations in BF subjects only (90 minutes P = 0.027, 120 minutes P = 0.050). A complete overview of all parameter outcomes is shown in detail in Supplemental Digital Content 2, Table for baseline and postmeal concentrations (*PM and delta [Δ]).


Many studies have shown beneficial effects of BF on later-life health (20–22). If and how particular components of HM such as lipids or physiological aspects of BF such as the later release of lipids during a BF session may contribute to these effects has not yet been elucidated: few studies investigated postprandial plasma parameters in (hospitalized) term (15) or preterm infants (16,23,24); others assessed fasting concentrations (25–28) or explored specific interventions over time (29–32). Postmeal responses of healthy infants after receiving either HM or IF with focus on lipids have, however, not been studied previously.

In this small-sized study, we investigated postmeal responses of infants to dietary lipids and found indications of differences between breast- and formula-fed infants. BF subjects in our study were fed more frequently than formula-fed infants, as found also in other studies (33–35). Subsequently, this group had shorter intermeal intervals. In addition, BF lasted longer than formula-feeding. The shorter feeding interval and longer feeding duration of BF subjects in our study was not associated with different volume intakes as determined by test weighing. We do not believe that these factors confounded our observations since 80% to 90% of a BF volume is consumed in the first 4 minutes of feeding (36).

The gradual release of fat toward the end of a BF session in contrast to bottle feeding, which provides lipids steadily over time, could be a reason for the differences we observed. The drop of some individual's ΔTG concentration in the BF group below baseline shortly after the meal could indicate that the late fat release of hindmilk had not been digested fully in all BF infants at the time of baseline blood collection.

Despite a potential delay in lipid intake in BF subjects, baselines were not different between groups. Postmeal BF TG concentrations seemed to increase earlier than those in the FF group (*PM, median BF > FF, 30 minutes) indicating that HM lipids may be more quickly digested, absorbed, and reached the blood stream faster than those in formula as suggested by others (24). The effect was, however, lost with baseline correction, and the BF ΔTG concentrations stayed at baseline until a later increase above baseline at 90 minutes. This observation suggests an effect of the more lipid-rich hindmilk on plasma TG concentrations.

Our observations seem to contradict with previously published reports that TG from HM disappear earlier from the stomach of preterm infants than those from formulae (24). There was, however, no hindmilk effect in that study because the children received their feeding by nasogastric tube, and therefore with a more constant fat intake.

Recent in vitro studies indicate that formula composition has less impact on digestion and absorption than physical properties (37). Differences in lipid handling in our study could thus be related to physical properties of HM and IF lipids such as milk fat droplet size and architecture, which are larger and more complexly coated in HM (7,11–13,38), structural properties of TG such as positioning of specific fatty acids along the glycerol backbone in HM (8) or differing viscoelasticity (39,40). These factors have shown to affect gastric emptying, gastric lipolysis, intestinal digestion, and absorption of fatty acids, TG, and other lipid-related parameters.

We observed faster changes in postmeal lipid concentrations compared with studies in adults challenged with a fat load (12). Fasted adults that had ingested 0.68 g fat/kg body weight showed a ΔTG net peak height of 1.0 mmol/L after 3 to 4 hours and a return to baseline after 8 hours (12). In contrast, infants in our study showed an earlier and much lower postmeal ΔTG response (0.42 mmol/L between 60 and 90 minutes) and a return to baseline within 180 minutes, despite an estimated 2.3 times higher fat intake relative to body weight (∼1.6 g fat/kg body weight).

Differences of postmeal TG concentrations between infants and adults could be due to faster uptake from the gastrointestinal tract or a more efficient clearance of circulating lipids in newborn infants, both of which would be in line with the approximately 6 times higher energy demands in infants than adults and the dependency of the human neonate on fatty acids and ketone bodies as energy source (1). Although a higher clearance efficiency may be reflected by the lower plasma peak heights despite a higher fat load that we observed, this could also be related to differences in fasting state between infants and adults. Because fasting for >4 hours is not ethical in infants, yet common for digestion studies in adults and rodents, other studies are required for a comparison of lipid uptake and plasma clearance between neonates and adults to confirm the indications given by our data.

A single ApoB48 lipoprotein is incorporated into an individual chylomicron and can therefore serve as proxy for chylomicron number (41). By relating TG to ApoB48 concentrations, the TG load carried per chylomicron can be estimated (42,43). The significant difference in baseline TG:ApoB48 ratio may suggest that chylomicrons in BF subjects carry a higher TG load than FF subjects. This observation is supported by postmeal ApoB48 concentrations in the BF group being consistently higher than in the FF group. Because TG baselines do not indicate a difference in circulating TG concentrations between groups, the lower TG:ApoB48 ratio in the FF group could therefore be compensated for by a higher number of chylomicrons (as indicated by ApoB48) in this group.

Baseline total cholesterol concentrations were higher in the BF than FF group in line with previous studies (25,30–32,44–46). Postmeal concentrations were also consistently higher in the BF group and—once corrected for baseline—increased in response to the meal whereas concentrations in the FF group remained unchanged. This observation is most likely explained by the virtual absence of cholesterol in this studies infant formula. The presence of cholesterol in HM has been proposed to exert an inverse but permanent (long-term programming) effect resulting in low circulating cholesterol concentrations in adulthood (25) thus contributing to the protective effect proposed for BF on later cardiovascular diseases risk (47–49), yet needs further investigation.

Our study has a number of important limitations. First, the number of infants for each measuring time was very small. For ethical reasons and to limit the burden, we collected 2 blood samples from each healthy infant within hours after feeding. This allowed us to obtain consent from both the parents and the ethical committee. Second, this limitation of only collecting 1 baseline and 1 postmeal sample made it impossible to construct true postprandial kinetics or area-under-the-curve graphs of individual infants that are usually generated in food intake or fat challenge studies. Thus, our data could be regarded as a first indication of differences of plasma lipid concentration patterns in infants receiving either BF or FF. Based on our results, future studies on dietary lipid handling in healthy term infants could focus on the first 2 hours after a feeding. In addition, our data could support future power calculations for the parameters used in the present study. Third, because of the ethical reasons, infants cannot be fasted as for rodent or adult fat load experiments, which may have affected the degree of comparability, as discussed in detail above. And last, we studied 1 formula with a vegetable oil-based fat bend, only. More studies are needed to investigate postmeal responses to formulae of different composition including fat blend components and compare responses between infants and adults directly.

In summary, we found that postmeal plasma concentrations of several lipid parameters showed small differences after either BF or FF in healthy infants. Differences were indicated for TG, BF over FF subjects at 30 minutes (*PM) yet for FF over BF subjects at 60 minutes when corrected for baseline. The differences in TG:ApoB48 ratio together with baseline TG and ApoB48 concentrations suggest that chylomicrons in BF subjects may carry a higher fat load compensated by a higher chylomicron number in FF subjects. Cholesterol and lipoprotein concentrations were higher in BF subjects and seemed to increase after the meal in the BF group only.

Research Perspectives

Although knowledge about lipid requirements and physiological impact of lipids is accumulating, more research is needed. Only few studies over the past 4 decades have investigated postprandial responses of (healthy) infants to their respective feeds. Future studies should aim to

  • further increase the understanding of direct physiological effects of HM components and the process of BF on lipid handling and the infant's metabolism;
  • investigate effects and functional mechanisms of milk characteristics such as fat droplet size and architecture, composition of TG, as well as active components such as lactoferrin, lactalbumin, and others; and
  • identify impact of dietary lipids on the human epigenome and their functional role as signaling molecules.


We thank the study nurses for their role in recruiting subjects, conducting home visits, and managing study documentation. We thank the personnel at the Clinical Chemistry laboratory at the UMCG for analysis of the blood samples and Dr Hanneke Lankheet (Team Leader Data Management) and Ruurd Bennink, Statistician at the Biometrics department, Nutricia Research, for their support and contribution to the study.


1. Koletzko B, Agostoni C, Bergmann R, et al. Physiological aspects of human milk lipids and implications for infant feeding: a workshop report. Acta Paediatr 2011; 100:1405–1415.
2. Lindquist S, Hernell O. Lipid digestion and absorption in early life: an update. Curr Opin Clin Nutr Metab Care 2010; 13:314–320.
3. Daly SE, Hartmann PE. Infant demand and milk supply. Part 2: the short-term control of milk synthesis in lactating women. J Hum Lact 1995; 11:27–37.
4. Stam J, Sauer PJ, Boehm G. Can we define an infant's need from the composition of human milk? Am J Clin Nutr 2013; 98:521S–528S.
5. Michalski MC, Briard V, Michel F, et al. Size distribution of fat globules in human colostrum, breast milk, and infant formula. J Dairy Scit 2005; 88:1927–1940.
6. Oosting A, Van Vlies N, Kegler D, et al. Effect of dietary lipid structure in early postnatal life on mouse adipose tissue development and function in adulthood. Br J Nutr 2014; 111:215–226.
7. Vors C, Pineau G, Gabert L, et al. Modulating absorption and postprandial handling of dietary fatty acids by structuring fat in the meal: a randomized crossover clinical trial. Am J Clin Nutr 2013; 97:23–36.
8. Nelson CM, Innis SM. Plasma lipoprotein fatty acids are altered by the positional distribution of fatty acids in infant formula triacylglycerols and human milk. Am J Clin Nutr 1999; 70:62–69.
9. Innis SM. Dietary triacylglycerol structure and its role in infant nutrition. Adv Nutr 2011; 2:275–283.
10. Hall B. Uniformity of human milk. Am J Clin Nutr 1979; 32:304–312.
11. Armand M, Borel P, Pasquier B, et al. Physicochemical characteristics of emulsions during fat digestion in human stomach and duodenum. Am J Physiol 1996; 271 (1 pt 1):G172–G183.
12. Armand M, Pasquier B, Andre M, et al. Digestion and absorption of 2 fat emulsions with different droplet sizes in the human digestive tract. Am J Clin Nutr 1999; 70:1096–1106.
13. Michalski MC, Briard V, Desage M, et al. The dispersion state of milk fat influences triglyceride metabolism in the rat—a 13CO2 breath test study. Eur J Nutr 2005; 44:436–444.
14. Borel P, Armand M, Pasquier B, et al. Digestion and absorption of tube-feeding emulsions with different droplet sizes and compositions in the rat. JPEN J Parenter Enteral Nutr 1994; 18:534–543.
15. Lucas A, Boyes S, Bloom SR, et al. Metabolic and endocrine responses to a milk feed in six-day-old term infants: differences between breast and cow's milk formula feeding. Acta Paediatr Scand 1981; 70:195–200.
16. Moro G, Minoli I, Boehm G, et al. Postprandial plasma amino acids in preterm infants: influence of the protein source. Acta Paediatr 1999; 88:885–889.
17. Delplanque B, Gibson R, Koletzko B, et al. Lipid quality in infant nutrition: current knowledge and future opportunities. J Pediatr Gastroenterol Nutr 2015; 61:8–17.
18. Jensen RG. Handbook of Milk Composition. San Diego, CA:Academic Press; 1995.
19. Stam J, Sauer PJ, Boehm G. Can we define an infant's need from the composition of human milk? Am J Clin Nutr 2013; 98:521S–528S.
20. Dewey KG. Is breastfeeding protective against child obesity? J Hum Lact 2003; 19:9–18.
21. Harder T, Bergmann R, Kallischnigg G, et al. Duration of breastfeeding and risk of overweight: a meta-analysis. Am J Epidemiol 2005; 162:397–403.
22. Owen CG, Martin RM, Whincup PH, et al. Does breastfeeding influence risk of type 2 diabetes in later life? A quantitative analysis of published evidence. Am J Clin Nutr 2006; 84:1043–1054.
23. Rochow N, Moller S, Fusch G, et al. Levels of lipids in preterm infants fed breast milk. Clin Nutr 2010; 29:94–99.
24. Armand M, Hamosh M, Mehta NR, et al. Effect of human milk or formula on gastric function and fat digestion in the premature infant. Pediatr Res 1996; 40:429–437.
25. Wagner V, von Stockhausen HB. The effect of feeding human milk and adapted milk formulae on serum lipid and lipoprotein levels in young infants. Eur J Pediatr 1988; 147:292–295.
26. Van Biervliet JP, Rosseneu M, Bury J, et al. Apolipoprotein and lipid composition of plasma lipoproteins in neonates during the first month of life. Pediatr Res 1986; 20:324–328.
27. Van Biervliet JP, Rosseneu M, Caster H. Influence of dietary factors on the plasma lipoprotein composition and content in neonates. Eur J Pediatr 1986; 144:489–493.
28. Decsi T, Thiel I, Koletzko B. Essential fatty acids in full term infants fed breast milk or formula. Arch Dis Child Fetal Neonatal Ed 1995; 72:F23–F28.
29. Rings EH, Minich DM, Vonk RJ, et al. Functional development of fat absorption in term and preterm neonates strongly correlates with ability to absorb long-chain Fatty acids from intestinal lumen. Pediatr Res 2002; 51:57–63.
30. Akeson PM, Axelsson IE, Raiha NC. Plasma lipids and apolipoproteins in breastfed and formula-fed Swedish infants. Acta Paediatr 1999; 88:1–6.
31. Kallio MJ, Salmenpera L, Siimes MA, et al. Exclusive breast-feeding and weaning: effect on serum cholesterol and lipoprotein concentrations in infants during the first year of life. Pediatrics 1992; 89 (4 pt 1):663–666.
32. Isomura H, Takimoto H, Miura F, et al. Type of milk feeding affects hematological parameters and serum lipid profile in Japanese infants. Pediatr Int 2011; 53:807–813.
33. Dewey KG, Heinig MJ, Nommsen LA, et al. Maternal versus infant factors related to breast milk intake and residual milk volume: the DARLING study. Pediatrics 1991; 87:829–837.
34. Hornell A, Aarts C, Kylberg E, et al. Breastfeeding patterns in exclusively breastfed infants: a longitudinal prospective study in Uppsala, Sweden. Acta Paediatr 1999; 88:203–211.
35. Kent JC, Mitoulas LR, Cregan MD, et al. Volume and frequency of breastfeedings and fat content of breast milk throughout the day. Pediatrics 2006; 117:e387–e395.
36. Lucas A, Lucas PJ, Baum JD. Pattern of milk flow in breast-fed infants. Lancet 1979; 314:57–58.
37. Bourlieu C, Menard O, De La Chevasnerie A, et al. The structure of infant formulas impacts their lipolysis, proteolysis and disintegration during in vitro gastric digestion. Food Chem 2015; 182:224–235.
38. Michalski MC, Soares AF, Lopez C, et al. The supramolecular structure of milk fat influences plasma triacylglycerols and fatty acid profile in the rat. Eur J Nutr 2006; 45:215–224.
39. Michalski MC, Cariou R, Michel F, et al. Native vs. damaged milk fat globules: membrane properties affect the viscoelasticity of milk gels. J Dairy Sci 2002; 85:2451–2461.
40. Bourlieu C, Michalski MC. Structure-function relationship of the milk fat globule. Curr Opin Clin Nutr Metab Care 2015; 18:118–127.
41. Sakai N, Uchida Y, Ohashi K, et al. Measurement of fasting serum apoB-48 levels in normolipidemic and hyperlipidemic subjects by ELISA. J Lipid Res 2003; 44:1256–1262.
42. Pal S, Ellis V, Ho S. Acute effects of whey protein isolate on cardiovascular risk factors in overweight, post-menopausal women. Atherosclerosis 2010; 212:339–344.
43. Vine DF, Takechi R, Russell JC, et al. Impaired postprandial apolipoprotein-B48 metabolism in the obese, insulin-resistant JCR: LA-cp rat: increased atherogenicity for the metabolic syndrome. Atherosclerosis 2007; 190:282–290.
44. Bayley TM, Alasmi M, Thorkelson T, et al. Influence of formula versus breast milk on cholesterol synthesis rates in four-month-old infants. Pediatr Res 1998; 44:60–67.
45. Harit D, Faridi MM, Aggarwal A, et al. Lipid profile of term infants on exclusive breastfeeding and mixed feeding: a comparative study. Eur J Clin Nutr 2008; 62:203–209.
46. Innis SM, Hamilton JJ. Effects of developmental changes and early nutrition on cholesterol metabolism in infancy: a review. J Am Coll Nutr 1992; 11 (11 suppl):63S–68S.
47. Khuc K, Blanco E, Burrows R, et al. Adolescent metabolic syndrome risk is increased with higher infancy weight gain and decreased with longer breast feeding. Int J Pediatr 2012; 2012:1–6.
48. Martin RM, Ebrahim S, Griffin M, et al. Breastfeeding and atherosclerosis: intima-media thickness and plaques at 65-year follow-up of the Boyd Orr cohort. Arterioscler Thromb Vasc Biol 2005; 25:1482–1488.
49. Martin RM, Gunnell D, Smith GD. Breastfeeding in infancy and blood pressure in later life: systematic review and meta-analysis. Am J Epidemiol 2005; 161:15–26.

apolipoprotein B48; bottle feeding; dietary fat; nursing; triglycerides

Supplemental Digital Content

© 2017 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,