Objective. To examine the influence of infant feeding method on serum total cholesterol (TC) and low-density lipoprotein (LDL) cholesterol.
Methods. A cross-sectional study of 13- to 16-year-olds and a systematic review of studies (all observational) on the effects of infant feeding on cholesterol in infancy (<1 year), childhood or adolescence (1–16 years), and adulthood (≥17 years) were conducted using random effects models. Differences are presented as breastfed-bottle-fed. A total of 1532 individuals (92% white; 55% male; mean age: 15.1 years) in 10 British towns were studied, and 37 studies with 52 observations on TC (26 in infancy, 17 in childhood or adolescence, and 9 in adulthood; corresponding figures for LDL were 7, 4, and 6) were reviewed.
Results. Mean TC in childhood or adolescence (including the new study) was not related to infant feeding pattern (mean TC difference = 0.00; 95% confidence interval [CI]: −0.07 to 0.07 mmol/L). However, in infancy, mean TC was higher among those breastfed (mean TC difference = 0.64; 95% CI: 0.50–0.79 mmol/L), whereas in adults, mean TC was lower among those breastfed (mean TC difference = −0.18; 95% CI: −0.30 to −0.06 mmol/L). Patterns for LDL were similar to those for TC throughout.
Conclusions. Breastfeeding is associated with increased mean TC and LDL levels in infancy but lower levels in adulthood/adult life. These results suggest that breastfeeding may have long-term benefits for cardiovascular health and may have implications for the content of formula feed milks.
Circulating concentrations of total cholesterol (TC) and low-density lipoprotein (LDL) cholesterol in adult life are strong and reversible risk factors for coronary heart disease (CHD).1,2 However, there is growing evidence that CHD risk begins to emerge before adult life. Pathologic studies have shown that coronary atherosclerosis is already well established by early adulthood in populations with high CHD rates.3,4 Although adult TC and LDL are influenced by adult diet and adiposity,5 levels of these factors “track” with increasing strength from early childhood.6,7 Moreover, reports from historical cohort studies suggest that adult TC and LDL levels may be influenced by factors that operate in early life. Low birth weight,8 bottle-feeding,8,9 and prolonged breastfeeding8 have been related to higher levels of adult TC and LDL.
The possibility that infant feeding has long-term effects on blood cholesterol levels is of considerable interest, particularly as increasing evidence has suggested that nutritional programming may influence the development of cardiovascular risk in humans.10,11 However, studying the effects of infant feeding on blood cholesterol level is complicated by several factors. First, it is possible that the effects of infant feeding may differ with age,12,13 although most studies have examined only short-term outcomes. Second, infant feeding practice, including the content of bottle feeds, changed markedly during the 20th century,14 as did other aspects of the early life environment.10,15 Thus, the relevance of infant feeding to blood cholesterol levels may be different for historical and contemporary children.
We examined whether there are overall mean differences in TC and LDL levels between breast- and bottle-fed children in a large cross-sectional survey of contemporary young people/adolescents aged 13 to 16 years. To place our findings in context, we conducted a systematic review of all studies reporting on TC and LDL levels in breast- and bottle-fed subjects at all ages.
Study of Contemporary Adolescents
The study was the third phase of the School Heart Health Study, a survey of cardiovascular risk factors in childhood in 10 British towns.16,17 Approval was obtained from all relevant local research ethics committees. In 1998 to 1999, 69 secondary schools, corresponding to the random sample of 100 primary schools included in earlier phases, were recruited. Pupils who had previously taken part were invited (average 45 pupils per school), together with a random sample of new participants (average 13 pupils per school). Preliminary visits were made to all schools by the research team, both to provide details about the study to school staff and to encourage participation among pupils. Invitations were sent jointly to parents and pupils; written informed consent was sought for all participants, using reminders when necessary.
Measurements were made by a trained field team (4 observers) who visited each town in turn, alternating between towns with high and low cardiovascular mortality. Children were examined in light clothing without shoes. Height was measured with a portable stadiometer (CMS Ltd, Camden, United Kingdom), and weight was measured with a digital electronic weighing scale (Soehnle Ltd, Murrhardt, Germany). Pubertal status was ascertained using a confidential self-assessment questionnaire based on Tanner pubic hair, penile, and breast development scales.18 Ethnicity was defined on appearance (European, south Asian, other). A venous blood sample was requested from 2451 subjects and collected after an overnight fast. Blood samples were frozen (−20°C) within 6 hours of collection and transferred to a central laboratory for analysis within 2 weeks of collection. Total serum cholesterol (mmol/L) and high-density lipoproteins were measured using a Hitachi 747 automated analyzer (Roche Diagnostics Corp, Indianapolis, IN) at the Royal Free Hospital.19,20 LDL was calculated using the Friedewald equation.21 Triglycerides (maximum 4.7 mmol/L) were well within the limit for determining LDL (10 mmol/L) using the Friedewald equation.21 A parental questionnaire sent immediately after examination provided information on infant feeding practices in the first 3 months of life (breastfed only, bottle-fed only, and mixed feeding) and on the duration of breastfeeding when applicable. Parental occupation was used to define social class in accordance with the Registrar General’s (ONS) 1990 coding manual22; head of household’s social class was defined as paternal social class or (when not available) maternal social class.
Systematic Review Process
A systematic review of all published papers, letters, abstracts, and review articles on infant feeding and blood lipids was conducted using Embase, Medline, and Web of Science databases. All references to the effects of infant feeding on cholesterol were identified using a combined text word and MESH heading (for Medline only) search strategy of infant feeding (breast feed*, infant feed*, bottle feed*, or formula feed*) and cholesterol (cholesterol, lipoprotein, or lipid*). The review was restricted to studies written in English and conducted on human subjects and from which estimates of a mean difference and its standard error between breastfed and bottle-fed infants could be derived.
Because it was possible that the influence of infant feeding would vary with age,12,13 an a priori decision was taken to examine the data in 3 age strata: in infancy (<1 year of life), in childhood or adolescence (1–16 years), and in adults (≥17 years). Because it has been suggested that the effects of infant feeding on blood lipids might differ by gender,23,24 data on male and female subjects were presented separately when possible. The electronic search (completed in May 2001) yielded 523 references; abstract review suggested that 78 were potentially relevant, and an additional 5 references were identified from other sources. Papers were considered relevant when the abstract indicated (or stated) that cholesterol had been measured in different infant feeding groups. Of these 83 papers, 37 papers contained data on TC and/or LDL levels, which could be used in the meta-analysis (Table 1). Of the 46 studies excluded, 30 studies did not examine TC levels in breast- and bottle-fed groups, 6 studies provided insufficient data for inclusion,25–30 and 10 studies duplicated data presented elsewhere.31–40 After correspondence with the authors of 2 studies in adults, the provision of unpublished data made it possible to include 2 additional studies in the meta-analysis.8,24,41 In most studies, breast- and bottle-feeding groups were separate, although in 7 studies the breastfeeding group also included mixed feeders,12,42–47 and in 2 studies bottle feeders were mixed fed.9,48 In 7 additional studies, the exclusiveness of infant feeding could not be verified (Table 1).9,44,48–52 Bottle-fed infants on standard formulas were included in preference to specially supplemented formulas, which were systematically excluded. In longitudinal studies examining blood lipids in the first year of life, cholesterol measurement at the latest point of exclusive breast or formula feeding was recorded when possible.
Statistical analysis was performed using Intercooled Stata 6.0 for Windows software (Stata Corp, Union Station, TX).
Cross-Sectional Study in Children
Both outcome variables (TC and LDL) were normally distributed. Body mass index (BMI) was used as a measure of childhood weight-for-height because of its relative independence of height within this narrow age range in this population, unlike other indices of body stature (eg, ponderal index). Infant feeding was treated as a categorical variable with 3 levels (breast only, bottle only, mixed); breastfeeding duration was grouped (3- or 6-month intervals) to minimize the effects of imprecise recall. Associations between TC, LDL, and infant feeding practices were studied using multivariate linear regression models and taking account of possible gender differences in the relations between age, height, BMI, and cholesterol (TC and LDL). All regression models were standardized for town (10 levels) and ethnicity (3 levels: white, south Asian, other).
A meta-analysis of the effects of infant feeding (breast or bottle) on TC and LDL was conducted separately among infants (<1 year), children or adolescents (1–16 years), and adults (≥17 years). The mean difference (breastfed-bottle-fed) and standard error of the difference in TC and LDL were analyzed. In 1 study, the standard deviation quoted seemed to be a standard error and was treated as such.42 A test of heterogeneity of the differences in TC levels between infant feeding groups was conducted (using Woolf’s χ2 test) within each of the 3 age groups studied using the META command within STATA. Because heterogeneity was marked in studies in infancy and in childhood or adolescence, pooled estimates of the difference in TC and LDL between feeding groups were produced using random effects models53 throughout (META command within STATA); in adults (when there was little heterogeneity), random and fixed effects models gave very similar point estimates. To assess whether there was evidence of publication bias and whether the studies included in the meta-analysis were typical, we produced funnel plots54 for both TC and LDL for each age group. In addition, age-specific Begg55 and Egger56 tests were performed using the METABIAS command within STATA. Differences in cholesterol between age groups were examined using the METAREG command. The effect of excluding studies in which the breastfeeding group included subjects on mixed feeding or when exclusive breastfeeding status could not be ascertained was explored.
In all, 2451 subjects were invited for measurement: 1635 (67%) participated and 1532 (63%) provided blood samples; 842 (55%) were male; no subjects were excluded on the basis of their blood lipid values. Main characteristics of survey participants who provided blood samples are shown in Table 2. Almost all samples (90%) were collected during fasting; nonfasting TC and LDL levels were similar to fasting levels, and adjustment made no difference to the results presented. TC fell with increasing age in boys (−0.16 mmol/L/y; 95% confidence interval [CI]: −0.24 to −0.07; P < .001) but was unrelated to age in girls (0.02; 95% CI: −0.08 to 0.11; P = .760); a test for gender interaction was significant (P = .015). TC fell with increasing height (−0.01 mmol/L/cm; 95% CI: −0.02 to −0.01; P < .001) and rose with increasing BMI (0.02 mmol/L per kg/m2; 95% CI: 0.01–0.03; P < .001) in both genders; pubertal status had a negative relationship with TC once age and height were taken into account (−0.07 mmol/L/grade; 95% CI: −0.12 to −0.01; P = .011). Town, social class, and ethnicity showed no consistent relation with TC level. LDL showed similar patterns to TC.
Both mean TC and LDL levels were very similar among children who received exclusive breast or bottle-feeding in the first 3 months of life (Table 3). Although mean TC and LDL levels were slightly higher among mixed feeders than among the other groups, the differences were not statistically significant. Additional adjustment for body build, maturation, and social class made little difference to the results. The duration of breastfeeding (with bottle feeders included as a 0-duration group as they were not exposed to breast milk) did not have a discernible effect on TC or LDL.
Systematic Review of the Effect of Infant Feeding on TC and LDL
From the 37 studies that provided data on the relationship between infant feeding (breast or formula feeding) and cholesterol levels, 52 estimates of TC differences were derived, of which 34 included both genders and 18 were gender specific. Twenty-six observations were in infants, 17 were in children or adolescents, and 9 were in adults; these are shown in Figs 1 to 3⇓⇓⇓, respectively. There was strong evidence of a systematic difference among the 3 age groups (P < .001), justifying separate consideration of the 3 groups in turn.
Infants (<1 Year)
Mean TC was higher in breastfed infants compared with bottle-fed infants in 25 of the 26 observations. The 1 study that reported higher TC among formula-fed infants measured TC levels only 4 days after birth.42 There was evidence of marked heterogeneity between studies (χ225 = 89.5; P < .001), although there was no consistent relationship between TC difference and age or gender. In a random effects model, mean TC was higher among breastfed infants (mean difference: 0.64 mmol/L; 95% CI: 0.49–0.79 mmol/L). Mean LDL was higher among breastfed infants in 645,57–61 of 7 observations.62 There was no statistically significant heterogeneity between estimates (χ26 = 6.8; P = .34). In a random effects model, mean LDL was higher among breastfed infants (mean difference: 0.57 mmol/L; 95% CI: 0.40–0.75; Fig 1).
Children and Adolescents (1–16 Years)
Mean TC in childhood and adolescence showed no consistent difference between breast and bottle-feeding, with the CIs of 16 of 17 observations, including 1 large study (N = 4023) in a multiracial sample of children in the United States,52 crossing the line of no difference (Fig 2). There was statistically significant heterogeneity between studies (χ216 = 31.6, P = .011), although again there was no consistent age or gender effect. Maturing children (aged 12–15 years) seemed to show a similar pattern to children of younger age. A random effects model showed that breastfed infants had similar mean childhood TC levels to those who were bottle-fed (mean difference: 0.00 mmol/L; 95% CI: −0.07 to 0.07 mmol/L). Only 4 observations on the relationship between infant feeding and LDL were available47,60; 2 were from the current study. There was no statistically significant heterogeneity among these observations (χ23 = 2.5, P = .48). A random effects model showed that mean LDL levels were similar among breast- and bottle-fed subjects (mean difference: 0.01 mmol/L; 95% CI: −0.07 to 0.08 mmol/L).
Adults (≥17 Years)
In adults, mean TC was lower in those who were breastfed in 7 of the 9 observations. Despite the wide age range (with mean ages from 17 years44 to 64 years8,24) and the wide range of birth dates (from 1920 to 1975), there was no statistically significant heterogeneity between estimates (χ28 = 3.7, P = .887). A random effects model showed that breastfed subjects had mean TC levels 0.18 mmol/L lower (95% CI: 0.06–0.30 mmol/L) than those who were bottle-fed. In 4 observations, mean LDL levels all were lower among breastfed adults.8,9,24,63 There was no statistically significant heterogeneity between estimates (χ25 = 1.7, P = .887), and a random effects model suggested that mean LDL levels were 0.20 mmol/L (95% CI: 0.08–0.32) lower than in those who were formula fed (Fig 3).41
Funnel plots examining the relationship between sample size and mean difference in cholesterol (TC and LDL) provided no evidence of publication or inclusion bias in any age group. That is, there was no evidence that small studies had been published only when they reported large mean differences in cholesterol between feeding groups. Age-specific Begg and Egger tests for publication bias were not statistically significant. The exclusion of 7 studies (11 observations) in which the breastfeeding group included subjects on mixed feeding and 7 studies (7 observations) in which exclusive breastfeeding could not be ascertained had no effects on the results.
In our study of contemporary adolescents, TC and LDL levels at 13 to 16 years of age did not differ between breast and bottle feeders, even after adjustment for body build, maturation, and social class. In addition, the duration of breastfeeding (with bottle feeders included as a 0-duration group) did not have a discernible effect on cholesterol. These results were entirely consistent with our systematic review and meta-analysis of studies of infant feeding and blood cholesterol in childhood and adolescence. However, in infancy, TC and LDL levels tended to be higher among breastfed subjects than in bottle feeders, whereas in adulthood, TC and LDL levels tended to be lower among breastfed subjects.
The present study of contemporary adolescents was among the largest to examine the relation between infant feeding and blood cholesterol in any age group, and it had considerable statistical power, reliably excluding even modest mean differences in TC or LDL of the order of 0.15 mmol/L in either direction. The estimates obtained were not materially affected by adjustment for potential confounding factors. The absence of an association between breastfeeding and blood cholesterol is not likely to be attributable to poor maternal recall of infant feeding history, which has been shown to be valid up to 20 years after birth.64 The study population was drawn from subjects from widely differing geographic locations and social circumstances. Although the response rate (63%) was only moderate, the prevalence of breastfeeding in the study population (63%) was close to an estimate (64%) from a nationwide survey of infant feeding practices carried out at the same time (1985).65 Moreover, the absence of any relationship between blood lipids and social class in this study population suggests that selection bias is unlikely to be affecting the infant feeding-blood lipid relationship. Mean TC in both boys and girls were comparable to a contemporary group of British children aged 11 to 14 years.66
The systematic review of the relationship between infant feeding and blood cholesterol levels was based on all reports identified by a systematic search of Medline, Embase, and Web of Science databases, supplemented by retrospective citations from recent reviews. Although it was not possible to identify unpublished studies systematically, there was no consistent evidence of publication bias. The sources of information on infant feeding practice varied, but in most studies at all ages the information was recorded in infancy (ie, at the time of infant feeding), either directly from child health records or from the mother; few studies relied on long-term maternal recall. In a few studies (n = 7), subjects on mixed feeding were included in the breastfeeding group. It is therefore possible that the extent of differences between breast- and bottle-fed infants are being underestimated, although exclusion of these studies (along with studies in which exclusive breastfeeding status could not be obtained) did not affect the estimates obtained. In almost all studies, bottle-feeding is likely to represent the use of formula feeds67; the exceptions are 2 of the studies in adults8,9,24 in which cow milk and condensed milk preparations are likely to have been prominent. All of the comparisons are based on observational data, because the random allocation of infants to breast and bottle-feeding has generally been regarded as unfeasible and/or unethical. Thus, confounding, particularly by social factors, body build, and later diet, needs to be considered.
The results of the systematic review suggest that breastfeeding is associated with different effects on TC and LDL at different stages of the life course. During the first year of life, breastfeeding seems to be associated with higher levels of TC and LDL. This is a remarkably consistent finding in all studies in which outcome is measured at least 2 weeks after birth. This association does not seem to be confounded by marked differences in social class, which has no consistent relation with blood cholesterol level in infancy.66 In most of the studies of infants examined, breast or bottle-feeding was still being provided at the time of measurement, so confounding by later diet is unlikely. Although no relation between infant feeding pattern and TC or LDL in childhood and adolescence was observed, both TC and LDL levels were lower in adults who had been breastfed rather than bottle-fed. Although the difference is modest, it is reasonably consistent between studies including subjects of different ages (from 17 to 64 years) and years of birth (from 1920 to 1975). This suggests that the difference between patterns in infancy and in adults is not simply the result of a birth cohort effect and that it may have continuing relevance. It also makes the possibility of confounding by social circumstances unlikely, because the relation between social class and infant feeding changed during the 20th century. The absence of confounding by social class was also demonstrated within individual studies.8 Although recent studies have suggested that bottle-feeding may be related to later obesity,68,69 this association was not apparent in these adult studies and adjustment for adult body build did not affect the results.23 The possibility that dietary patterns in adult life are different among breast and bottle feeders is difficult to exclude, but, again, the marked social changes associated with breastfeeding make it unlikely that a consistent pattern of adult dietary differences between breast and bottle feeders would have persisted across the 20th century.
The duration of breastfeeding required for reduction of adult cholesterol levels is uncertain. Of the studies examining the relation between duration of breastfeeding and adult TC, 1 suggested that breastfeeding for >3 months was associated with lower TC levels,41 although Fall’s study in Hertfordshire men suggested that breastfeeding for >1 year was associated with higher TC level.8
The association between breastfeeding and higher TC and LDL in infancy is likely to be a direct consequence of nutritional differences between breast and bottle milk; this is supported by the rapid emergence and reversibility of the effect. The high cholesterol content of breast milk (its most consistent long-term difference from formula feed) may well be responsible; in infancy, dietary cholesterol intake seems to be the main determinant of TC level61 and the use of cholesterol-supplemented formula feeds in infancy leads to TC levels in infancy closer to those seen in breastfed infants.31,62 Although the causality of the association between breastfeeding and lower TC and LDL in adults has still to be established, this could represent a form of nutritional programming11 whereby a stimulus (eg, enteral feeding in early life) acting at a critical period of growth (ie, during infancy) has lasting or lifelong significance.11 Studies in rats have shown that early breastfeeding is associated with favorable lipid levels in adulthood.70 However, results in different species are not consistent, with an opposite effect being found in monkeys fed an atherogenic diet later in life.71 Although the precise nutritional stimulus is uncertain, the markedly higher cholesterol content of breast milk compared with most bottle-feeding preparations is an important possibility. This difference will certainly have applied in those post-World War II populations that used formula feeds.23,41,44,63 In the prewar or wartime populations,8,9,24 cow milk (which has a similar cholesterol and fat content to breast milk) is likely to have been an important source of bottle feed. However, in practice, this was often diluted with water (sometimes with added sugar), reducing its fat and cholesterol concentration.9,72 Thus, it is likely that the difference in cholesterol concentration between breast and bottle milks would apply to all of the studies cited here. Other differences in the composition of breast milk from formula feeds, in hormones (particularly leptin and tri-iodothyronine), in immunoglobulins, and in nucleotides, might be important.44,63 Whatever the precise programming stimulus, long-term changes in cholesterol metabolism (either by regulation of enzyme activity or by LDL-receptor activity) may be occurring. High cholesterol intake in infancy reduces endogenous synthesis of cholesterol, probably by downregulation of hepatic hydroxymethyl glutaryl coenzyme A reductase.61 However, whether such changes persist into adult life remains to be established.
Although the difference in adult TC level between breast and bottle feeders of 0.2 mmol/L (approximately 0.2 standard deviation) is modest, a reduction in mean TC of this magnitude in adult life would be associated with a reduction in CHD incidence of approximately 10%, based on observational data.1 In Britain, where the prevalence of bottle-feeding is at least one third, this could contribute an appreciable population attributable risk fraction for CHD (approximately 3%). At this stage, it would be premature to recommend breastfeeding on these grounds; the case for breastfeeding is based on evidence of protection against infection and allergic disorders,73,74 and obesity68,69 coupled with the possibility of improved neurodevelopment.32,75 However, it is possible that these benefits are accompanied by a long-term reduction in CHD risk, although other possible cardiovascular consequences of breastfeeding (eg, on arterial compliance) must also be taken into account.63 The results also may have implications for the content of formula feeds. In recent years, it has been assumed that a low intake of cholesterol in infant feed is good for long-term cardiovascular prevention, an assumption that has received support from some animal studies.70 However, our results suggest that high cholesterol intake in infancy may (paradoxically) protect against dietary fat intake later in life and that there may be a strong case for the fat content of formula feeds to match that of breast milk as closely as possible.
Examination of existing large prospective birth cohort studies may help to resolve the uncertainty about the long-term consequences of breastfeeding on cholesterol levels in later life. If the relationship between breastfeeding and lower circulating cholesterol levels in adult life receives additional support from such analyses and is independent of differences in adiposity and current diet, then it suggests that infant feeding may have long-term benefits for cardiovascular health.
The Ten Towns Heart Health Study was supported by a project grant from the Wellcome Trust (reference 051187/Z/97/A).
We are grateful to the members of the study team (Michelle Bithell, Gill Emmett, Carina Hammarstrom, Rosemary MacShane, and Cara Oliver) and to all of the participating schools, pupils, and parents. Measurements of TC and LDL cholesterol were made in the Department of Clinical Biochemistry, Royal Free Hospital (Professor Tony F. Winder and Dr Michael Thomas). Dr Caroline Fall (MRC Environmental Epidemiology Unit, University of Southampton, UK) and Professor Sanja Kolacek (Children’s Hospital Zagreb, Croatia) kindly provided unpublished data for inclusion in the review.
- ↵Law MR, Wald NJ, Thompson SG. By how much and how quickly does reduction in serum cholesterol concentration lower risk of ischaemic heart disease? BMJ.1994;308 :367– 372
- ↵Tang JL, Armitage JM, Lancaster T, Silagy CA, Fowler GH, Neil HA. Systematic review of dietary intervention trials to lower blood total cholesterol in free-living subjects. BMJ.1998;316 :1213– 1220
- ↵Webber LS, Srinivasan SR, Wattigney WA, Berenson GS. Tracking of serum lipids and lipoproteins from childhood to adulthood. The Bogalusa Heart Study. Am J Epidemiol.1991;133 :884– 899
- ↵Wilsgaard T, Jacobsen BK, Schirmer H, et al. Tracking of cardiovascular risk factors: the Tromso study, 1979–1995. Am J Epidemiol.2001;154 :418– 426
- ↵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 :801– 805
- ↵Ravelli AC, van der Meulen JH, Osmond C, Barker DJ, Bleker OP. Infant feeding and adult glucose tolerance, lipid profile, blood pressure, and obesity. Arch Dis Child.2000;82 :248– 252
- ↵Barker DJ. Mothers, Babies and Health in Later Life. London, United Kingdom: Churchill Livingstone; 1998
- ↵Lucas A. Programming by early nutrition in man. In: Bock GR, Whelan J, eds. The Childhood Environment and Adult Disease. Chichester, United Kingdom: John Wiley & Sons; 1991:38–55
- ↵Oppe TE. Present Day Practice in Infant Feeding: Third Report of a Working Party of the Panel on Child Nutrition, Committee on Medical Aspects of Food Policy. London, United Kingdom: Her Majesty’s Stationery Office; 1988
- ↵Whincup PH, Cook DG, Papacosta O, Walker M. Childhood blood pressure, body build, and birthweight: geographical associations with cardiovascular mortality. J Epidemiol Community Health.1992;46 :396– 402
- ↵Whincup PH, Cook DG, Adshead F, et al. Cardiovascular risk factors in British children from towns with widely differing adult cardiovascular mortality. BMJ.1996;313 :79– 84
- ↵Siedel J, Hagele EO, Ziegenhorn J, Wahlefeld AW. Reagent for the enzymatic determination of serum total cholesterol with improved lipolytic efficiency. Clin Chem.1983;29 :1075– 1080
- ↵Sugiuchi H, Uji Y, Okabe H, et al. Direct measurement of high-density lipoprotein cholesterol in serum with polyethylene glycol-modified enzymes and sulfated alpha-cyclodextrin. Clin Chem.1995;41 :717– 723
- ↵Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem.1972;18 :499– 502
- ↵Office for National Statistics. Standard Occupational Classification (SOC 1990). Vol 1. London, United Kingdom: The Stationery Office; 2000
- ↵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 :164– 167
- ↵Fall CH, Osmond C, Barker DJ, et al. Fetal and infant growth and cardiovascular risk factors in women. BMJ.1995;310 :428– 432
- Baranowski T, Carroll R, Harrison J, Henske J. Cholesterol fractions among 3–4 year old children from 3 ethnicities: relationships to infant-feeding practices. Circulation.1990;81 :726
- Farris RP, Frank GC, Webber LS, Srinivasan SR, Berenson GS. Influence of milk source on serum lipids and lipoproteins during the first year of life. Bogalusa Heart Study. Am J Clin Nutr.1982;35 :42– 49
- Lane DM, Pineda J, McClanahan J, McConathy WJ. Serum-lipids and apolipoproteins up to 4 weeks of age: human-milk versus SMA formula feedings. J Am Coll Nutr.1988;7 :409
- Lapinleimu H, Viikari J, Ronnemaa T, et al. Apolipoprotein E polymorphism and serum lipids in a randomized, prospective trial of an infant diet with reduced saturated fat and cholesterol. Pediatrics.1996;98 :757– 762
- McConathy WJ, Lane DM. Serum-lipids and apolipoproteins during the 1st 4 weeks of life: breast feeding versus formula feeding. Clin Res.1983;31 :A893
- Rassin DK, Gaull GE, Jarvenpaa AL, Raiha NC. Feeding the low-birth-weight infant: II. Effects of taurine and cholesterol supplementation on amino acids and cholesterol. Pediatrics.1983;71 :179– 186
- ↵Uauy R, Mize CE, Castillo-Duran C. Fat intake during childhood: metabolic responses and effects on growth. Am J Clin Nutr.2000;72 :1354S– 1360S
- ↵Friedman G, Goldberg SJ. Concurrent and subsequent serum cholesterol of breast- and formula-fed infants. Am J Clin Nutr.1975;28 :42– 45
- ↵Darmady JM, Fosbrooke AS, Lloyd JK. Prospective study of serum cholesterol levels during first year of life. BMJ.1972;2 :685– 688
- ↵Ward SD, Melin JR, Lloyd FP, Norton JA Jr, Christian JC. Determinants of plasma cholesterol in children: a family study. Am J Clin Nutr.1980;33 :63– 70
- ↵Light RJ, Pillemar DB. Summing Up: The Science of Reviewing Research. Cambridge, MA: Harvard University Press; 1984
- ↵Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ.1997;315 :629– 634
- ↵Katoku Y, Yamada M, Yonekubo A, Kuwata T, Kobayashi A, Sawa A. Effect of the cholesterol content of a formula on the lipid compositions of plasma lipoproteins and red blood cell membranes in early infancy. Am J Clin Nutr.1996;64 :871– 877
- ↵Agostoni C, Riva E, Scaglioni S, Marangoni F, Radaelli G, Giovannini M. Dietary fats and cholesterol in Italian infants and children. Am J Clin Nutr.2000;72 :1384S– 1391S
- ↵Mize CE, Uauy R, Kramer R, Benser M, Allen S, Grundy SM. Lipoprotein-cholesterol responses in healthy infants fed defined diets from ages 1 to 12 months: comparison of diets predominant in oleic acid versus linoleic acid, with parallel observations in infants fed a human milk-based diet. J Lipid Res.1995;36 :1178– 1187
- ↵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 :1403– 1411
- ↵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 :643– 647
- ↵Kark JD, Troya G, Friedlander Y, Slater PE, Stein Y. Validity of maternal reporting of breast feeding history and the association with blood lipids in 17 year olds in Jerusalem. J Epidemiol Community Health.1984;38 :218– 225
- ↵Martin J, White A. Infant Feeding 1985: A Survey of Infant Feeding Carried Out by Social Survey Division of the OPCS on Behalf of the Department of Health and Social Security and Scotish Home and Health Department. London, United Kingdom: Her Majesty’s Stationery Office; 1988
- ↵Gregory J, Lowe S. National Diet and Nutrition Survey: Young People Aged 4 to 18 Years. Vol 1. London, United Kingdom: The Stationery Office; 2000
- ↵White A, Freeth S, O’Brien M. Infant Feeding 1990: A Survey Carried Out by Social Survey Division of OPCS on Behalf of the Department of Health, the Scottish Home and Health Department, the Welsh Office and the Department of Health and Social Services in Northern Ireland. London, United Kingdom: The Stationery Office; 1992
- ↵Reiser R, Sidelman Z. Control of serum cholesterol homeostasis by cholesterol in the milk of the suckling rat. J Nutr.1972;102 :1009– 1016
- ↵Mott GE, Jackson EM, McMahan CA, McGill HC Jr. Cholesterol metabolism in adult baboons is influenced by infant diet. J Nutr.1990;120 :243– 251
- ↵Oppe TE. Present-Day Practice in Infant Feeding: Report of a Working Party of the Panel on Child Nutrition, Committee on Medical Aspects of Food Policy. London, United Kingdom: Her Majesty’s Stationery Office; 1974
- ↵Howie PW, Forsyth JS, Ogston SA, Clark A, Florey CD. Protective effect of breast feeding against infection. BMJ.1990;300 :11– 16
- ↵Rassin DK, Raiha NC, Minoli I, Moro G. Taurine and cholesterol supplementation in the term infant: responses of growth and metabolism. JPEN J Parenter Enteral Nutr.1990;14 :392– 397
- ↵Jarvenpaa AL, Raiha NC, Rassin DK, Gaull GE. Milk protein quantity and quality in the term infant. I. Metabolic responses and effects on growth. Pediatrics.1982;70 :214– 220
- ↵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
- Copyright © 2002 by the American Academy of Pediatrics