Breastfeeding and Atherosclerosis
http://www.100md.com
动脉硬化血栓血管生物学 2005年第7期
From the Department of Social Medicine (R.M.M., S.E., G.D.S., S.W., S.F., D.G.), University of Bristol, Bristol, UK; The Vascular Noninvasive Screening and Diagnostic Centre (M.G.), London, UK; Department of Neurovascular Sciences (A.N.N., N.G.), The Cyprus Institute of Neurology and Genetics, Nicosia; and Department of Surgery (J.M.P.H.), University of Bristol, Bristol Royal Infirmary, UK.
Correspondence to Dr Martin, University of Bristol, Canynge Hall, Whiteladies Rd, Bristol, Avon BS8 2PR UK. E-mail richard.martin@bristol.ac.uk
Abstract
Objectives— The impact of breastfeeding in infancy on cardiovascular disease risk is uncertain. We related breastfeeding in infancy to atherosclerosis in adulthood.
Methods and Results— A historic cohort study based on a 65-year follow-up of the Carnegie (Boyd Orr) survey of diet and health in prewar Britain, 1937 to 1939. A total of 732 eligible cohort members living in or around Aberdeen, Bristol, Dundee, Wisbech, and London were invited for follow-up examinations in 2002, and 405 (55%) participated. In models controlling for age and sex, breastfeeding was inversely associated with common carotid intima-media thickness (IMT; difference –0.03 mm; 95% CI, –0.07 to 0.01), bifurcation IMT (difference –0.19 mm; 95% CI, –0.37 to –0.01), carotid plaque (odds ratio [OR], 0.52; 95% CI, 0.29 to 0.92), and femoral plaque (OR, 0.54; 95% CI, 0.26 to 1.12), compared with bottle-feeding. Controlling for socioeconomic variables in childhood and adulthood, smoking and alcohol made little difference to effect estimates. Controlling for factors potentially on the causal pathway (blood pressure, adiposity, cholesterol, insulin resistance, and C-reactive protein) made little difference to observed associations.
Conclusions— Breastfeeding may be associated with a reduced risk of atherosclerosis in later life. Measurement error and power considerations limit the extent to which conclusions about the mechanisms underlying this relationship can be made.
We investigated the association between breastfeeding in infancy and later atherosclerosis among 405 participants in a 65-year follow-up study. Breastfeeding was inversely associated with intima-media thickness and carotid and femoral plaque prevalence, even after controlling for socioeconomic and behavioral factors. Breastfeeding may lower atherosclerosis risk in later life.
Key Words: infant nutrition ? breastfeeding ? cardiovascular disease risk factors ? intima-media thickness ? atherosclerosis ? historical cohort
Introduction
Breastfeeding in infancy is a possible determinant of later coronary heart disease and its risk factors.1–7 A postmortem study found fewer coronary plaques among breastfed versus bottle-fed young accident victims,8 and breastfeeding was inversely associated with coronary heart disease mortality, except when prolonged.6 Several mechanisms could underpin these observations. In meta-analyses, breastfeeding was associated with a 0.18 mmol/L reduction in total cholesterol in adults1 and a 1.10 mm Hg reduction in systolic blood pressure.2 Bottle-feeding, in contrast, is positively associated with blood pressure9 and insulin resistance.3,4
Others show no relationship of breastfeeding with coronary heart disease,10–12 and prolonged breastfeeding may adversely affect arterial distensibility (a suggested predictor of coronary heart disease).5 In yellow baboons, prolonged breastfeeding followed by a high-fat diet was positively associated with atherosclerosis,13 although generalizing to humans is problematic.
Common carotid and bifurcation intima-media thickness (IMT), and the presence of coronary and femoral plaques, are established measures of preclinical atherosclerosis and predict incident stroke and ischemic heart disease.14–16 We investigated the association between breastfeeding and atherosclerosis measured by arterial ultrasound in 63- to 82-year-old participants in the Boyd Orr cohort.
Methods
The Boyd Orr cohort comprises 4999 participants from 1343 families in 16 centers in England and Scotland who participated in a 1-week survey of diet and health when aged 0 to 19 years between 1937 and 1939.17,18 The National Health Service Central Register (NHSCR) was used to trace 4379 (88%) individuals. Between 1997 and 1998, all 3182 traced survivors were sent questionnaires. Of the 1648 responses, 1378 (84%) consented to further follow-up. In February 2002, 2563 of the original cohort were alive and living in Britain, and 1295 (51%) participants who had consented to further follow-up were known to be still alive and contactable. We contacted all 732 (29%) participants living near clinics in Bristol, London, Wisbech, Aberdeen, and Dundee, and 85% (n=619) responded, of whom 405 (16% of total; 55% of those contacted) underwent clinical examination; and 339 (13% of total; 46% of those contacted) returned for arterial ultrasound scans (Figure I, available online at http://atvb.ahajournals.org). Ethical approval was obtained from Multi-centre Research Ethics Committee Scotland. All participants gave informed consent.
Method of infant feeding (any breastfeeding and its duration, or exclusively bottle-fed) was obtained by direct questioning of mothers at the time of the original survey. For full details of other variables measured at baseline in childhood and at the follow-up clinic in adulthood, please see the online supplement, available at http://atvb.ahajournals.org.
Arterial Ultrasound Scan
The right and left carotid and common femoral arterial bifurcations were studied with an Advanced Technology Laboratories HDI (high-definition imaging) 3000 triplex system using a high-resolution broadband-width linear array transducer 7-4 MHz (Phillips Medical Systems). Measurements were made of IMT and plaques, where present, by 1 vascular technologist, blind to infant feeding mode.14 (For full details please see http://atvb.ahajournals.org.) The common carotid IMT was measured at its thickest point, 1.5- to 2-cm proximal to the flow divider, on the distal wall of the common carotid artery. Bifurcation IMT was defined as described previously.14 In the presence of a plaque, its maximum thickness was measured, and this was taken as the bifurcation IMT. In the absence of a plaque, the IMT measured at the bulb origin was defined as the bifurcation IMT. Plaques were defined at the time of ultrasound measurements as described previously.14
Statistical Analysis
Associations of breastfeeding with continuously distributed variables were investigated with random-effects linear regression modeling since clustering effects (shared genetic influences on atherosclerosis and propensity to being breastfed) may have arisen because several cohort members belonged to the same families (the 339 subjects were from 261 families). Associations between breastfeeding and the prevalence of plaques were investigated using logistic regression and robust SEs computed to account for clustering. (Please see http://atvb.ahajournals.org for a detailed modeling strategy.) To assess the sensitivity of our conclusions to possible selection bias, we repeated the analyses using inverse probability weighting19 (for details, please see http://atvb.ahajournals.org).
Results
Overall, 182 (45%) men and 223 (55%) women were followed up in clinic, and 155 (46%) men and 184 (54%) women were scanned (Figure I). Their mean age was 71 years (range 63 to 82) with no sex difference (P=0.5). Method of infant feeding was available for 362 participants, of whom 272 (75%) were breastfed with no sex difference (P=0.7). The median duration of breastfeeding was 9 months (interquartile range [IQR], 5 to 9) in both sexes (P=0.7). This is similar to the prevalence (70%) and median duration (9 months; IQR, 4 to 9) of breastfeeding in the full cohort.20 Breastfed subjects were 284 g (95% CI, 65 to 503) heavier at birth, but there was little difference in infant feeding mode by age, year born, sex, childhood social class, food expenditure, nutrient intake, adult social class, smoking, or alcohol use (Table I, available online at http://atvb.ahajournals.org).
Representativeness
Compared with the remaining surviving survey members (n=2563), clinic participants were 10 months younger at baseline (95% CI, 4 to 14 months), taller (difference in height z score, 0.19; 95% CI, 0.07 to 0.32), more likely to have been breastfed (75% versus 69%), and when they were children, the family per-capita weekly food expenditure was >5 shillings (ie, 25 pence, equivalent to ￡12.16 at current prices) among 55% of participants versus 41% of nonparticipants. Birth year, sex, birth weight, father’s social class, and childhood body mass index (BMI) were similar whether subjects were followed up or not.
Cardiovascular Disease Risk Factors
In general, there was little evidence of differences in risk factors (adiposity, blood pressure, lipids, or insulin resistance) between breastfed and bottle-fed participants (Table 1). There was some evidence that breastfeeding was associated with lower average glycemia measured by hemoglobin A1c (HbA1c) in those without diabetes (difference –0.07%; 95% CI, –0.17 to 0.02). In models controlling for age, sex, socioeconomic, and behavioral factors and BMI, the difference in HbA1c between breastfed and bottle-fed subjects was –0.12% (95% CI, –0.26 to 0.02; P=0.1) in all subjects and –0.10% (95% CI, –0.19 to 0.00; P=0.05) in subjects without diabetes. There was no evidence of an association of breastfeeding with type 2 diabetes (odds ratio [OR], 0.97; 95% CI, 0.41 to 2.30; P=0.9). There was some evidence of a reduction in odds of being on an antihypertensive drug associated with breastfeeding (OR, 0.67; 0.40 to 1.12; P=0.1).
TABLE 1. Distribution of Cardiovascular Disease Risk Factors by Infant Feeding Mode
Atherosclerosis
In line with other population-based studies,14 the common carotid IMT was normally distributed with means (SD) of 0.79 (0.18) and 0.72 (0.13) mm for men and women, respectively; the mean (SD) bifurcation IMT was 1.82 (0.78) and 1.63 (0.69) mm for men and women, respectively. In age- and sex-adjusted models, breastfeeding was associated with reductions in bifurcation IMT (difference, –0.19; 95% CI, –0.37 to –0.01) and odds of carotid plaque (OR, 0.52; 95% CI, 0.29 to 0.92; Table II, available online at http://atvb.ahajournals.org).
In models controlling for age, sex, and socioeconomic and behavioral factors, breastfeeding was associated with reductions in common carotid (difference –0.03 mm; 95% CI, –0.07 to 0.01) and bifurcation (–0.23 mm; 95% CI, –0.40 to –0.06) IMT compared with bottle-feeding (Table 2). Breastfeeding was also associated with reductions in odds of carotid (OR, 0.45; 0.24 to 0.86) and femoral (0.46; 95% CI, 0.21 to 1.01) plaques (Table 3). Further adjustment for cardiovascular disease risk factors hardly altered the effect estimates, except HbA1c, which attenuated the association between breastfeeding and bifurcation IMT by 13%.
TABLE 2. Association of Breastfeeding With Carotid and Bifurcation IMT Controlling for Potential Confounding Variables and Risk Factors for Coronary Heart Disease
TABLE 3. Association of Breastfeeding With Carotid and Femoral Plaques Controlling for Potential Confounding Variables and Risk Factors for Coronary Heart Disease
Neither birth weight (a marker for fetal growth), childhood leg length (a marker for childhood growth and adverse exposures during growth),21 nor specific nutrient intakes in childhood confounded the breastfeeding–atherosclerosis associations. There was little evidence of interaction by sex, age at examination, year of birth category, childhood BMI, energy, fat, or saturated fat intake. There was little evidence of a duration-response relationship between breastfeeding and atherosclerosis (Table III, available online at http://atvb.ahajournals.org). The mean difference in bifurcation IMT was 0.12 mm (95% CI, –0.04 to 0.28; P=0.16) and the OR for carotid plaques was 2.05 (95% CI, 1.03 to 4.08; P=0.04) per category of increasing breastfeeding duration (reference < 6 months of breastfeeding). There was little evidence of an association of breastfeeding with self-reported ischemic heart disease (OR, 0.88; 95% CI, 0.49 to 1.56; P=0.7). Breastfeeding-IMT (P interaction >0.5) and -plaque (P interaction >0.1) associations did not differ among those with and without clinical evidence of ischemia.
Sensitivity Analyses
When analyses were reweighted to assess the impact of missing data, effect sizes were little altered. In weighted models based on all 2563 surviving nonrespondents in England, Wales and Scotland, inverse associations of breastfeeding with common carotid (difference –0.03; 95% CI: –0.07 to 0.01) and bifurcation (difference –0.17; 95% CI, –0.4 to 0.03) IMT, carotid (OR, 0.49; 95% CI, 0.25 to 0.99) and femoral plaques (OR, 0.62; 95% CI, 0.31 to 1.24) remained. Results were similar when analyses were repeated using study area nonrespondents.
Discussion
Breastfeeding was inversely associated with atherosclerosis, measured by IMT and plaque prevalence. We also observed a 0.12% reduction in HbA1c in breastfed versus bottle-fed subjects. Although of borderline statistical significance, the findings are of interest for at least 2 reasons. First, the differences in IMT and plaque prevalence associated with breastfeeding were of a similar magnitude to differences seen in smokers versus never-smokers and those with and without evidence of coronary heart disease.14 The potential public health importance of the reduction in HbA1c associated with breastfeeding is suggested by the observation that lowering the nondiabetic population mean HbA1c by just 0.1% has been predicted to reduce total mortality by 5%.22 Second, the decision to breastfeed in the pre–World War II era was less socially patterned than it is now,23 providing some control for socioeconomic confounding at the design stage of this study. In contrast, breastfeeding mothers of children born during the last 30 to 40 years are more educated and more health-conscious than mothers who bottle-feed,24 and the influence of possible confounding factors in recent studies of the long-term effects of breastfeeding is probably impossible to completely control for.25 The breastfeeding–atherosclerosis associations were independent of other early life factors, such as birth weight, nutrition, and socioeconomic conditions in childhood, of socioeconomic environment in adulthood, and of factors (smoking and alcohol) operating in later life that may be related to healthy upbringing among children of mothers who breastfed.
We were unable to establish a mechanism by which breastfeeding may influence atherosclerosis. We had hypothesized that any association may operate via blood pressure,2,9 cholesterol levels,1 glycemia, or insulin resistance,3,4 but effect estimates were only altered a little after controlling for glycosylated hemoglobin. However, measurement error may account for this apparent lack of any substantial effect of controlling for these factors. Furthermore, we cannot exclude the possibility that lifelong exposure (as opposed to concurrent risk factor levels) to increased blood pressure, cholesterol levels, or insulin resistance may be underlying mechanisms. Breastfeeding has been associated with a reduced prevalence of arterial plaques in children,8 and cardiovascular disease risk factors measured in childhood are prospectively associated with IMT in adulthood, independently of contemporaneous risk factors.26 Breastfeeding may be more strongly associated with blood pressure, cholesterol levels, glycemia, or insulin resistance much earlier in life, perhaps during the infant feeding period,1,27,28 protecting against early arterial damage.
Acute and chronic viral/bacterial infections have been associated with atherosclerosis, although the evidence is inconclusive.29 We could not investigate whether breastfeeding protects against atherosclerosis by reducing exposure to persistent infections in infancy.23,30
Limitations
First, selection bias is possible because the study was based on a proportion of the original cohort, but it requires that breastfeeding–atherosclerosis associations differ among clinic participants versus nonparticipants. This seems unlikely for the following reasons. Effect sizes were little altered when analyses were reweighted to account for missing values, although models assume data were missing at random. Inverse breastfeeding–ischemic heart disease associations observed during clinic follow-up (OR, 0.88) were also observed among 418 nonclinic participants who returned questionnaires (OR, 0.84). Breastfeeding–atherosclerosis associations were similar in those with and without ischemic heart disease, suggesting selection by disease status does not explain the results. Survival bias was an unlikely explanation for the absence of associations with blood pressure, adiposity, or cholesterol levels because there was little evidence of interaction by age. Second, method of infant feeding was obtained when the children were a mean age of 6.3 years, suggesting the possibility of misclassification. However, long-term recall appears to be a valid method of obtaining infant-feeding information up to 20 years earlier.31 Recall bias seems implausible because atherosclerosis was measured prospectively. Third, in line with other historic cohorts, breastfed babies were heavier than bottle babies at birth, suggesting a size-based influence on choice to breastfeed.32 Birth weight has been inversely related to the degree of subclinical atherosclerosis in adulthood,33 pointing to the potential importance of genetic or in utero factors. Although we found little evidence that controlling for birth weight attenuated associations between breastfeeding and atherosclerosis, birth weight was self-reported in a subsample (59%) of the cohort, and residual confounding is possible. The finding that rapid weight gain in the first 2 weeks postnatally is positively associated with insulin resistance,34 and endothelial dysfunction,35 independent of birth weight, supports the suggestion that factors promoting early postnatal growth (eg, formula feeding compared with breastfeeding) might adversely affect later cardiovascular health.7 Finally, the findings may have arisen by chance because we conducted a number of statistical tests. However, an inverse breastfeeding–atherosclerosis relationship was a prespecified, end point–specific hypothesis.
Generalizability
Baboon studies suggest that infant feeding method may interact with a diet high in saturated fat in childhood to influence the development of atherosclerosis.13 Given subspecies differences in the apolipoprotein A genetic variants and levels of lipoprotein A, and a lack of data from other primates, it is possible that these findings are specific for this particular subspecies of baboon and thus have little relevance to humans.36 However, an influence of breastfeeding in humans may depend on later dietary patterns, which are now very different from those in the early 20th century.37 We found no interactions by childhood BMI, energy, fat, or saturated fat intake on breastfeeding–atherosclerosis associations, although power to detect these was limited.
Artificial feeds in the 1920s to 1930s were largely based on unmodified cow’s milk.23 Unlike formula milks of today (low in cholesterol, saturated fatty acid, and sodium), unmodified cow’s milk (unless it was diluted) had a high sodium concentration (low in breastmilk) but more closely resembled the cholesterol and saturated fatty acid content of mature breastmilk.38 Distinct associations with particular components of artificial feeds (such as differences in salt content) may produce different results in contemporary versus historic cohorts. Other differences between the composition of breast milk and cow’s milk and modern formula feeds in hormones (eg, leptin and thyroxine), immunoglobulins, and nucleotides might be important.1 Altered hormonal responses to breast and formula feeds, for example, different insulin27 and growth factor,39 effects may explain variations in outcomes in later life. We had no information on breastfeeding exclusivity and do not know whether results differ among infants who were exclusively or partially breastfed.
Comparison With Other Studies
In line with our findings, a postmortem study in young adults found lower rates of coronary atheroma among those breastfed (25%) compared with those artificially fed (60%).8 In Hertfordshire, men who had been partially or exclusively breastfed <1 year had lower standardized mortality ratios (SMRs; 73 and 79, respectively) compared with those who had been exclusively breastfed for >1 year (SMR, 97) or exclusively bottle-fed (SMR, 95).6 The results are in line with data from a recently published study on 87 252 participants in the Nurses Health Study, born between 1921 and 1946. Ever having been breastfed was associated with an 8% to 10% reduction in risk of coronary heart disease and stroke; the reduction in risk of coronary heart disease was 16% for women breastfed >9 months.40 However, other studies have found no association between breastfeeding and coronary artery plaques among young accident victims at postmortem,11 nonfatal myocardial infarction,12 and cardiovascular or coronary heart disease mortality.10 Small sample size11,12 and selection bias10 are a concern with these studies. We have shown recently no association of breastfeeding with ischemic heart disease mortality (hazard ratio, 1.02) in the Boyd Orr cohort.20 Although breastfeeding may influence subclinical atherosclerosis, other factors may be important for survival among those with disease.
The inverse relationship between breastfeeding and HbA1c, a measure of average glycemia, concurs with 2 studies showing lower levels of impaired glucose tolerance3 and type 2 diabetes4 in adult life among breastfed subjects. We observed differences in systolic and diastolic blood pressure of –1.62 mm Hg and –0.74 mm Hg, respectively, between breastfed and bottle-fed subjects, in line with a recent meta-analysis.2 However, the current study was only powered to detect differences of 6.5 mm Hg systolic and 3.2 mm Hg diastolic blood pressure. Despite recent interest in the relationship between breastfeeding and obesity, our findings indicate no evidence of any such association.
Breastfeeding is associated with a reduction in atherosclerosis, but the mechanism is unclear. Prospective investigations of the association between breastfeeding and ischemic heart disease are lacking. Furthermore, such studies would have to be prohibitively large and long term to detect small but (on a population level) important reductions in ischemic heart disease. Approximately 40% of infants are never breastfed in the United Kingdom.24 In the absence of prospective evidence, this study suggests the possibility that the promotion of breastfeeding could be a potential component of the public health strategy to reduce future levels of ischemic cardiovascular disease. However, further studies in large adult populations are needed to confirm these findings. In particular, the hypothesis that breastfeeding influences later cardiovascular disease risk factors could ethically and feasibly be tested on an intention-to-treat basis in large controlled trials of successful breastfeeding promotion interventions with long-term follow-up.41
Acknowledgments
The Wellcome Trust funded R.M.M. to undertake the clinical follow-up as part of a research training fellowship in clinical epidemiology (grant GR063779FR). The arterial ultrasound scans were performed and analyzed with funding by the British Heart Foundation (BHF; project grant PG/02/125). We are very grateful to the cohort members who participated so willingly in the follow-up study. We also acknowledge all the research workers in the original survey in 1937-9. Susie Potts is thanked for all her hard work in providing secretarial and administrative support to the study. The hypothesis was developed by D.G., G.D.S., R.M.M., S.F., and S.E. The field work was conducted by R.M.M., N.G., M.G., and S.W. under the direction of a steering group (S.E., G.D.S., S.F., D.G., and J.H.). A.N. provided expert advice on the conduct and analysis of the arterial scans. N.G., M.G., and A.N. analyzed the ultrasound scan data. R.M.M. did the statistical analysis, wrote the first draft and coordinated completion of the article. R.M.M. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors critically commented on and edited earlier drafts and approved the final version of this article.
References
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: 597–608.
Owen CG, Whincup PH, Gilg JA, Cook DG. Effect of breast feeding in infancy on blood pressure in later life: systematic review and meta-analysis. BMJ. 2003; 327: 1189–1195.
Ravelli ACJ, van der Meulen JH, Osmond C, Barker DJP, Bleker OP. Infant feeding and adult glucose tolerance, lipid profile, blood pressure, and obesity. Arch Dis Child. 2000; 82: 248–252.
Pettitt DJ, Forman MR, Hanson RL, Knowler WC, Bennett PH. Breastfeeding and incidence of non-insulin-dependent diabetes mellitus in Pima Indians. Lancet. 1997; 350: 166–168.
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.
Fall CHD, Barker DJP, Osmond C, Winter PD, Clark PMS, Hales CN. Relation of infant feeding to adult serum cholesterol concentration and death from ischaemic heart disease. BMJ. 1992; 304: 801–805.
Singhal A, Lucas A. Early origins of cardiovascular disease: is there a unifying hypothesis? Lancet. 2004; 363: 1642–1645.
Osborn GR. Stages in development of coronary disease observed from 1500 young subjects: relationship of hypotension and infant feeding to aetiology. Colloq Int Cent Natl Rech Sci. 1967; 169: 93–139.
Martin RM, McCarthy A, Davies DP, Davey Smith G, Ben-Shlomo Y. Association between infant nutrition and blood pressure in early adulthood: the Barry Caerphilly Growth cohort study. Am J Clin Nutr. 2003; 77: 1489–1497.
Wingard DL, Criqui MH, Edelstein SL, Tucker J, Tomlinson-Keasey C, Schwartz JE, Friedman HS. Is breast-feeding associated with adult longevity? Am J Public Health. 1994; 84: 1458–1462.
Burr ML, Beasley WH, Fisher CB. Breast feeding, maternal smoking and early atheroma. Eur Heart J. 1984; 5: 588–591.
Cowen DD. Myocardial infarction and infant feeding. Practitioner. 1973; 210: 661–663.
Lewis DS, Mott GE, McMahan CA, Masoro EJ, Carey KD, McGill HC Jr. Deferred effects of preweaning diet on atherosclerosis in adolescent baboons. Arteriosclerosis. 1988; 8: 274–280.
Ebrahim S, Papacosta O, Whincup P, Wannamethee G, Walker M, Nicolaides AN, Dhanjil S, Griffin M, Belcaro G, Rumley A, Lowe GD. Carotid plaque, intima media thickness, cardiovascular risk factors, and prevalent cardiovascular disease in men and women: the British Regional Heart Study. Stroke. 1999; 30: 841–850.
Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE. Common carotid intima-media thickness and risk of stroke and myocardial infarction: the Rotterdam Study. Circulation. 1997; 96: 1432–1437.
Leng GC, Papacosta O, Whincup P, Wannamethee G, Walker M, Ebrahim S, Nicolaides AN, Dhanjil S, Griffin M, Belcaro G, Rumley A, Lowe GD. Femoral atherosclerosis in an older British population: prevalence and risk factors. Atherosclerosis. 2000; 152: 167–174.
Gunnell DJ, Frankel S, Nanchahal K, Braddon FE, Smith GD. Lifecourse exposure and later disease: a follow-up study based on a survey of family diet and health in pre-war Britain (1937–1939). Public Health. 1996; 110: 85–94.
Rowett Research Institute. Family Diet and Health in Pre-War Britain. Dunfermline: Carnegie United Kingdom Trust; 1955.
Ziegler A, Kastner C, Chang-Claude J. Analysis of pregnancy and other factors on detection of human papilloma virus (HPV) infection using weighted estimating equations for follow-up data. Stat Med. 2003; 22: 2217–2233.
Martin RM, Davey Smith G, Tilling K, Frankel S, Gunnell D. Breastfeeding and cardiovascular mortality: the Boyd Orr cohort and a systematic review with meta-analysis. Eur Heart J. 2004; 25: 778–786.
Gunnell D. Can adult anthropometry be used as a biomarker for prenatal and childhood exposures? Int J Epidemiol. 2002; 31: 390–394.
Khaw KT, Wareham N, Luben R, Bingham S, Oakes S, Welch A, Day N. Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of European prospective investigation of cancer and nutrition (EPIC-Norfolk). BMJ. 2001; 322: 15–18.
Fildes V. Infant feeding practices and infant mortality in England, 1900–1919. Contin Chang. 1998; 13: 251–280.
Foster K, Lader D, Cheesborough S. Infant Feeding 1995: A Survey of Infant Feeding Practices in the United Kingdom Carried Out by the Social Survey Division of ONS on Behalf of the Department of Health, the Scottish Office Department of Health, the Welsh Office and the Department of Health and Social Services in Northern Ireland. London, UK: The Stationery Office; 1997.
Simmons D. NIDDM and breastfeeding. Lancet. 1997; 350: 157–158.
Raitakari OT, Juonala M, Kahonen M, Taittonen L, Laitinen T, Maki-Torkko N, Jarvisalo MJ, Uhari M, Jokinen E, Ronnemaa T, Akerblom HK, Viikari JSA. Cardiovascular risk factors in childhood and carotid artery intima-media thickness in adulthood: the Cardiovascular Risk in Young Finns Study. J Am Med Assoc. 2003; 290: 2277–2283.
Lucas A, Sarson DL, Blackburn AM, Adrian TE, Aynsley-Green A, Bloom SR. Breast vs bottle: endocrine responses are different with formula feeding. Lancet. 1980; 1: 1267–1269.
Hofman A, Hazebroek A, Valkenburg HA. A randomized trial of sodium intake and blood pressure in newborn infants. J Am Med Assoc. 1983; 250: 370–373.
Danesh J. Coronary heart disease, Helicobacter pylori, dental disease, Chlamydia pneumoniae, and cytomegalovirus: meta-analyses of prospective studies. Am Heart J;. 138: S434–S437.
Fall CHD, Goggin PM, Hawtin P, Fine D, Duggleby S. Growth in infancy, infant feeding, childhood living conditions, and Helicobacter pylori infection at age 70. Arch Dis Child. 1997; 77: 310–314.
Kark JD, Troya G, Friedlander Y, Slater PE, Stein Y. Validity of maternal reporting history and the association with blood lipids in 17 year olds in Jerusalem. J Epidemiol Community Health. 1984; 38: 218–225.
Baxter-Jones ADG, Cardy AH, Helms PJ, Phillips DO, Smith WC. Influence of socioeconomic conditions on growth in infancy: the 1921 Aberdeen birth cohort. Arch Dis Child. 1999; 81: 5–9.
Martyn CN, Gale CR, Jespersen S, Sherriff SB. Impaired fetal growth and atherosclerosis of carotid and peripheral arteries. Lancet. 1998; 352: 173–178.
Singhal A, Fewtrell M, Cole TJ, Lucas A. Low nutrient intake and early growth for later insulin resistance in adolescents born preterm. Lancet. 2003; 361: 1089–1097.
Singhal A, Cole TJ, Fewtrell M, Deanfield J, Lucas A. Is Slower Early Growth Beneficial for Long-Term Cardiovascular Health? Circulation. 2004; 109: 1108–1113.
Williams-Blangero S, Rainwater DL. Variation in Lp(a) levels and apo(a) isoform frequencies in five baboon subspecies. Hum Biol. 1991; 63: 65–76.
Maynard M, Gunnell D, Emmett P, Frankel S, Davey Smith G. Fruit, vegetables, and antioxidants in childhood and risk of adult cancer: the Boyd Orr cohort. J Epidemiol Community Health. 2003; 57: 218–225.
Emmett PM, Rogers IS. Properties of human milk and their relationship with maternal nutrition. Early Hum Dev. 1997; 49 (suppl): S7–S28.
Diaz-Gomez NM, Domenech E, Barroso F. Breast feeding and growth factors in preterm newborn infants. J Pediatr Gastroenterol Nutr. 1997; 24: 322–327.
Rich-Edwards JW, Stampfer MJ, Manson JE, Rosner B, Hu FB, Michels KB, Willett W. Breastfeeding during infancy and the risk of cardiovascular disease in adulthood. Epidemiology. 2004; 15: 550–556.
Kramer MS, Chalmers B, Hodnett ED, Sevkovskaya Z, Dzikovich I, Shapiro S, Collet JP, Vanilovich I, Mezen I, Ducruet T, Shishko G, Zubovich V, Mknuik D, Gluchanina E, Dombrovskiy V, Ustinovitch A, Kot T, Bogdanovich N, Ovchinikova L, Helsing E; PROBIT Study Group (Promotion of Breastfeeding Intervention Trial). Promotion of Breastfeeding Intervention Trial (PROBIT): a randomized trial in the Republic of Belarus. J Am Med Assoc. 2001; 285: 413–420.(Richard M. Martin; Shah E)
Correspondence to Dr Martin, University of Bristol, Canynge Hall, Whiteladies Rd, Bristol, Avon BS8 2PR UK. E-mail richard.martin@bristol.ac.uk
Abstract
Objectives— The impact of breastfeeding in infancy on cardiovascular disease risk is uncertain. We related breastfeeding in infancy to atherosclerosis in adulthood.
Methods and Results— A historic cohort study based on a 65-year follow-up of the Carnegie (Boyd Orr) survey of diet and health in prewar Britain, 1937 to 1939. A total of 732 eligible cohort members living in or around Aberdeen, Bristol, Dundee, Wisbech, and London were invited for follow-up examinations in 2002, and 405 (55%) participated. In models controlling for age and sex, breastfeeding was inversely associated with common carotid intima-media thickness (IMT; difference –0.03 mm; 95% CI, –0.07 to 0.01), bifurcation IMT (difference –0.19 mm; 95% CI, –0.37 to –0.01), carotid plaque (odds ratio [OR], 0.52; 95% CI, 0.29 to 0.92), and femoral plaque (OR, 0.54; 95% CI, 0.26 to 1.12), compared with bottle-feeding. Controlling for socioeconomic variables in childhood and adulthood, smoking and alcohol made little difference to effect estimates. Controlling for factors potentially on the causal pathway (blood pressure, adiposity, cholesterol, insulin resistance, and C-reactive protein) made little difference to observed associations.
Conclusions— Breastfeeding may be associated with a reduced risk of atherosclerosis in later life. Measurement error and power considerations limit the extent to which conclusions about the mechanisms underlying this relationship can be made.
We investigated the association between breastfeeding in infancy and later atherosclerosis among 405 participants in a 65-year follow-up study. Breastfeeding was inversely associated with intima-media thickness and carotid and femoral plaque prevalence, even after controlling for socioeconomic and behavioral factors. Breastfeeding may lower atherosclerosis risk in later life.
Key Words: infant nutrition ? breastfeeding ? cardiovascular disease risk factors ? intima-media thickness ? atherosclerosis ? historical cohort
Introduction
Breastfeeding in infancy is a possible determinant of later coronary heart disease and its risk factors.1–7 A postmortem study found fewer coronary plaques among breastfed versus bottle-fed young accident victims,8 and breastfeeding was inversely associated with coronary heart disease mortality, except when prolonged.6 Several mechanisms could underpin these observations. In meta-analyses, breastfeeding was associated with a 0.18 mmol/L reduction in total cholesterol in adults1 and a 1.10 mm Hg reduction in systolic blood pressure.2 Bottle-feeding, in contrast, is positively associated with blood pressure9 and insulin resistance.3,4
Others show no relationship of breastfeeding with coronary heart disease,10–12 and prolonged breastfeeding may adversely affect arterial distensibility (a suggested predictor of coronary heart disease).5 In yellow baboons, prolonged breastfeeding followed by a high-fat diet was positively associated with atherosclerosis,13 although generalizing to humans is problematic.
Common carotid and bifurcation intima-media thickness (IMT), and the presence of coronary and femoral plaques, are established measures of preclinical atherosclerosis and predict incident stroke and ischemic heart disease.14–16 We investigated the association between breastfeeding and atherosclerosis measured by arterial ultrasound in 63- to 82-year-old participants in the Boyd Orr cohort.
Methods
The Boyd Orr cohort comprises 4999 participants from 1343 families in 16 centers in England and Scotland who participated in a 1-week survey of diet and health when aged 0 to 19 years between 1937 and 1939.17,18 The National Health Service Central Register (NHSCR) was used to trace 4379 (88%) individuals. Between 1997 and 1998, all 3182 traced survivors were sent questionnaires. Of the 1648 responses, 1378 (84%) consented to further follow-up. In February 2002, 2563 of the original cohort were alive and living in Britain, and 1295 (51%) participants who had consented to further follow-up were known to be still alive and contactable. We contacted all 732 (29%) participants living near clinics in Bristol, London, Wisbech, Aberdeen, and Dundee, and 85% (n=619) responded, of whom 405 (16% of total; 55% of those contacted) underwent clinical examination; and 339 (13% of total; 46% of those contacted) returned for arterial ultrasound scans (Figure I, available online at http://atvb.ahajournals.org). Ethical approval was obtained from Multi-centre Research Ethics Committee Scotland. All participants gave informed consent.
Method of infant feeding (any breastfeeding and its duration, or exclusively bottle-fed) was obtained by direct questioning of mothers at the time of the original survey. For full details of other variables measured at baseline in childhood and at the follow-up clinic in adulthood, please see the online supplement, available at http://atvb.ahajournals.org.
Arterial Ultrasound Scan
The right and left carotid and common femoral arterial bifurcations were studied with an Advanced Technology Laboratories HDI (high-definition imaging) 3000 triplex system using a high-resolution broadband-width linear array transducer 7-4 MHz (Phillips Medical Systems). Measurements were made of IMT and plaques, where present, by 1 vascular technologist, blind to infant feeding mode.14 (For full details please see http://atvb.ahajournals.org.) The common carotid IMT was measured at its thickest point, 1.5- to 2-cm proximal to the flow divider, on the distal wall of the common carotid artery. Bifurcation IMT was defined as described previously.14 In the presence of a plaque, its maximum thickness was measured, and this was taken as the bifurcation IMT. In the absence of a plaque, the IMT measured at the bulb origin was defined as the bifurcation IMT. Plaques were defined at the time of ultrasound measurements as described previously.14
Statistical Analysis
Associations of breastfeeding with continuously distributed variables were investigated with random-effects linear regression modeling since clustering effects (shared genetic influences on atherosclerosis and propensity to being breastfed) may have arisen because several cohort members belonged to the same families (the 339 subjects were from 261 families). Associations between breastfeeding and the prevalence of plaques were investigated using logistic regression and robust SEs computed to account for clustering. (Please see http://atvb.ahajournals.org for a detailed modeling strategy.) To assess the sensitivity of our conclusions to possible selection bias, we repeated the analyses using inverse probability weighting19 (for details, please see http://atvb.ahajournals.org).
Results
Overall, 182 (45%) men and 223 (55%) women were followed up in clinic, and 155 (46%) men and 184 (54%) women were scanned (Figure I). Their mean age was 71 years (range 63 to 82) with no sex difference (P=0.5). Method of infant feeding was available for 362 participants, of whom 272 (75%) were breastfed with no sex difference (P=0.7). The median duration of breastfeeding was 9 months (interquartile range [IQR], 5 to 9) in both sexes (P=0.7). This is similar to the prevalence (70%) and median duration (9 months; IQR, 4 to 9) of breastfeeding in the full cohort.20 Breastfed subjects were 284 g (95% CI, 65 to 503) heavier at birth, but there was little difference in infant feeding mode by age, year born, sex, childhood social class, food expenditure, nutrient intake, adult social class, smoking, or alcohol use (Table I, available online at http://atvb.ahajournals.org).
Representativeness
Compared with the remaining surviving survey members (n=2563), clinic participants were 10 months younger at baseline (95% CI, 4 to 14 months), taller (difference in height z score, 0.19; 95% CI, 0.07 to 0.32), more likely to have been breastfed (75% versus 69%), and when they were children, the family per-capita weekly food expenditure was >5 shillings (ie, 25 pence, equivalent to ￡12.16 at current prices) among 55% of participants versus 41% of nonparticipants. Birth year, sex, birth weight, father’s social class, and childhood body mass index (BMI) were similar whether subjects were followed up or not.
Cardiovascular Disease Risk Factors
In general, there was little evidence of differences in risk factors (adiposity, blood pressure, lipids, or insulin resistance) between breastfed and bottle-fed participants (Table 1). There was some evidence that breastfeeding was associated with lower average glycemia measured by hemoglobin A1c (HbA1c) in those without diabetes (difference –0.07%; 95% CI, –0.17 to 0.02). In models controlling for age, sex, socioeconomic, and behavioral factors and BMI, the difference in HbA1c between breastfed and bottle-fed subjects was –0.12% (95% CI, –0.26 to 0.02; P=0.1) in all subjects and –0.10% (95% CI, –0.19 to 0.00; P=0.05) in subjects without diabetes. There was no evidence of an association of breastfeeding with type 2 diabetes (odds ratio [OR], 0.97; 95% CI, 0.41 to 2.30; P=0.9). There was some evidence of a reduction in odds of being on an antihypertensive drug associated with breastfeeding (OR, 0.67; 0.40 to 1.12; P=0.1).
TABLE 1. Distribution of Cardiovascular Disease Risk Factors by Infant Feeding Mode
Atherosclerosis
In line with other population-based studies,14 the common carotid IMT was normally distributed with means (SD) of 0.79 (0.18) and 0.72 (0.13) mm for men and women, respectively; the mean (SD) bifurcation IMT was 1.82 (0.78) and 1.63 (0.69) mm for men and women, respectively. In age- and sex-adjusted models, breastfeeding was associated with reductions in bifurcation IMT (difference, –0.19; 95% CI, –0.37 to –0.01) and odds of carotid plaque (OR, 0.52; 95% CI, 0.29 to 0.92; Table II, available online at http://atvb.ahajournals.org).
In models controlling for age, sex, and socioeconomic and behavioral factors, breastfeeding was associated with reductions in common carotid (difference –0.03 mm; 95% CI, –0.07 to 0.01) and bifurcation (–0.23 mm; 95% CI, –0.40 to –0.06) IMT compared with bottle-feeding (Table 2). Breastfeeding was also associated with reductions in odds of carotid (OR, 0.45; 0.24 to 0.86) and femoral (0.46; 95% CI, 0.21 to 1.01) plaques (Table 3). Further adjustment for cardiovascular disease risk factors hardly altered the effect estimates, except HbA1c, which attenuated the association between breastfeeding and bifurcation IMT by 13%.
TABLE 2. Association of Breastfeeding With Carotid and Bifurcation IMT Controlling for Potential Confounding Variables and Risk Factors for Coronary Heart Disease
TABLE 3. Association of Breastfeeding With Carotid and Femoral Plaques Controlling for Potential Confounding Variables and Risk Factors for Coronary Heart Disease
Neither birth weight (a marker for fetal growth), childhood leg length (a marker for childhood growth and adverse exposures during growth),21 nor specific nutrient intakes in childhood confounded the breastfeeding–atherosclerosis associations. There was little evidence of interaction by sex, age at examination, year of birth category, childhood BMI, energy, fat, or saturated fat intake. There was little evidence of a duration-response relationship between breastfeeding and atherosclerosis (Table III, available online at http://atvb.ahajournals.org). The mean difference in bifurcation IMT was 0.12 mm (95% CI, –0.04 to 0.28; P=0.16) and the OR for carotid plaques was 2.05 (95% CI, 1.03 to 4.08; P=0.04) per category of increasing breastfeeding duration (reference < 6 months of breastfeeding). There was little evidence of an association of breastfeeding with self-reported ischemic heart disease (OR, 0.88; 95% CI, 0.49 to 1.56; P=0.7). Breastfeeding-IMT (P interaction >0.5) and -plaque (P interaction >0.1) associations did not differ among those with and without clinical evidence of ischemia.
Sensitivity Analyses
When analyses were reweighted to assess the impact of missing data, effect sizes were little altered. In weighted models based on all 2563 surviving nonrespondents in England, Wales and Scotland, inverse associations of breastfeeding with common carotid (difference –0.03; 95% CI: –0.07 to 0.01) and bifurcation (difference –0.17; 95% CI, –0.4 to 0.03) IMT, carotid (OR, 0.49; 95% CI, 0.25 to 0.99) and femoral plaques (OR, 0.62; 95% CI, 0.31 to 1.24) remained. Results were similar when analyses were repeated using study area nonrespondents.
Discussion
Breastfeeding was inversely associated with atherosclerosis, measured by IMT and plaque prevalence. We also observed a 0.12% reduction in HbA1c in breastfed versus bottle-fed subjects. Although of borderline statistical significance, the findings are of interest for at least 2 reasons. First, the differences in IMT and plaque prevalence associated with breastfeeding were of a similar magnitude to differences seen in smokers versus never-smokers and those with and without evidence of coronary heart disease.14 The potential public health importance of the reduction in HbA1c associated with breastfeeding is suggested by the observation that lowering the nondiabetic population mean HbA1c by just 0.1% has been predicted to reduce total mortality by 5%.22 Second, the decision to breastfeed in the pre–World War II era was less socially patterned than it is now,23 providing some control for socioeconomic confounding at the design stage of this study. In contrast, breastfeeding mothers of children born during the last 30 to 40 years are more educated and more health-conscious than mothers who bottle-feed,24 and the influence of possible confounding factors in recent studies of the long-term effects of breastfeeding is probably impossible to completely control for.25 The breastfeeding–atherosclerosis associations were independent of other early life factors, such as birth weight, nutrition, and socioeconomic conditions in childhood, of socioeconomic environment in adulthood, and of factors (smoking and alcohol) operating in later life that may be related to healthy upbringing among children of mothers who breastfed.
We were unable to establish a mechanism by which breastfeeding may influence atherosclerosis. We had hypothesized that any association may operate via blood pressure,2,9 cholesterol levels,1 glycemia, or insulin resistance,3,4 but effect estimates were only altered a little after controlling for glycosylated hemoglobin. However, measurement error may account for this apparent lack of any substantial effect of controlling for these factors. Furthermore, we cannot exclude the possibility that lifelong exposure (as opposed to concurrent risk factor levels) to increased blood pressure, cholesterol levels, or insulin resistance may be underlying mechanisms. Breastfeeding has been associated with a reduced prevalence of arterial plaques in children,8 and cardiovascular disease risk factors measured in childhood are prospectively associated with IMT in adulthood, independently of contemporaneous risk factors.26 Breastfeeding may be more strongly associated with blood pressure, cholesterol levels, glycemia, or insulin resistance much earlier in life, perhaps during the infant feeding period,1,27,28 protecting against early arterial damage.
Acute and chronic viral/bacterial infections have been associated with atherosclerosis, although the evidence is inconclusive.29 We could not investigate whether breastfeeding protects against atherosclerosis by reducing exposure to persistent infections in infancy.23,30
Limitations
First, selection bias is possible because the study was based on a proportion of the original cohort, but it requires that breastfeeding–atherosclerosis associations differ among clinic participants versus nonparticipants. This seems unlikely for the following reasons. Effect sizes were little altered when analyses were reweighted to account for missing values, although models assume data were missing at random. Inverse breastfeeding–ischemic heart disease associations observed during clinic follow-up (OR, 0.88) were also observed among 418 nonclinic participants who returned questionnaires (OR, 0.84). Breastfeeding–atherosclerosis associations were similar in those with and without ischemic heart disease, suggesting selection by disease status does not explain the results. Survival bias was an unlikely explanation for the absence of associations with blood pressure, adiposity, or cholesterol levels because there was little evidence of interaction by age. Second, method of infant feeding was obtained when the children were a mean age of 6.3 years, suggesting the possibility of misclassification. However, long-term recall appears to be a valid method of obtaining infant-feeding information up to 20 years earlier.31 Recall bias seems implausible because atherosclerosis was measured prospectively. Third, in line with other historic cohorts, breastfed babies were heavier than bottle babies at birth, suggesting a size-based influence on choice to breastfeed.32 Birth weight has been inversely related to the degree of subclinical atherosclerosis in adulthood,33 pointing to the potential importance of genetic or in utero factors. Although we found little evidence that controlling for birth weight attenuated associations between breastfeeding and atherosclerosis, birth weight was self-reported in a subsample (59%) of the cohort, and residual confounding is possible. The finding that rapid weight gain in the first 2 weeks postnatally is positively associated with insulin resistance,34 and endothelial dysfunction,35 independent of birth weight, supports the suggestion that factors promoting early postnatal growth (eg, formula feeding compared with breastfeeding) might adversely affect later cardiovascular health.7 Finally, the findings may have arisen by chance because we conducted a number of statistical tests. However, an inverse breastfeeding–atherosclerosis relationship was a prespecified, end point–specific hypothesis.
Generalizability
Baboon studies suggest that infant feeding method may interact with a diet high in saturated fat in childhood to influence the development of atherosclerosis.13 Given subspecies differences in the apolipoprotein A genetic variants and levels of lipoprotein A, and a lack of data from other primates, it is possible that these findings are specific for this particular subspecies of baboon and thus have little relevance to humans.36 However, an influence of breastfeeding in humans may depend on later dietary patterns, which are now very different from those in the early 20th century.37 We found no interactions by childhood BMI, energy, fat, or saturated fat intake on breastfeeding–atherosclerosis associations, although power to detect these was limited.
Artificial feeds in the 1920s to 1930s were largely based on unmodified cow’s milk.23 Unlike formula milks of today (low in cholesterol, saturated fatty acid, and sodium), unmodified cow’s milk (unless it was diluted) had a high sodium concentration (low in breastmilk) but more closely resembled the cholesterol and saturated fatty acid content of mature breastmilk.38 Distinct associations with particular components of artificial feeds (such as differences in salt content) may produce different results in contemporary versus historic cohorts. Other differences between the composition of breast milk and cow’s milk and modern formula feeds in hormones (eg, leptin and thyroxine), immunoglobulins, and nucleotides might be important.1 Altered hormonal responses to breast and formula feeds, for example, different insulin27 and growth factor,39 effects may explain variations in outcomes in later life. We had no information on breastfeeding exclusivity and do not know whether results differ among infants who were exclusively or partially breastfed.
Comparison With Other Studies
In line with our findings, a postmortem study in young adults found lower rates of coronary atheroma among those breastfed (25%) compared with those artificially fed (60%).8 In Hertfordshire, men who had been partially or exclusively breastfed <1 year had lower standardized mortality ratios (SMRs; 73 and 79, respectively) compared with those who had been exclusively breastfed for >1 year (SMR, 97) or exclusively bottle-fed (SMR, 95).6 The results are in line with data from a recently published study on 87 252 participants in the Nurses Health Study, born between 1921 and 1946. Ever having been breastfed was associated with an 8% to 10% reduction in risk of coronary heart disease and stroke; the reduction in risk of coronary heart disease was 16% for women breastfed >9 months.40 However, other studies have found no association between breastfeeding and coronary artery plaques among young accident victims at postmortem,11 nonfatal myocardial infarction,12 and cardiovascular or coronary heart disease mortality.10 Small sample size11,12 and selection bias10 are a concern with these studies. We have shown recently no association of breastfeeding with ischemic heart disease mortality (hazard ratio, 1.02) in the Boyd Orr cohort.20 Although breastfeeding may influence subclinical atherosclerosis, other factors may be important for survival among those with disease.
The inverse relationship between breastfeeding and HbA1c, a measure of average glycemia, concurs with 2 studies showing lower levels of impaired glucose tolerance3 and type 2 diabetes4 in adult life among breastfed subjects. We observed differences in systolic and diastolic blood pressure of –1.62 mm Hg and –0.74 mm Hg, respectively, between breastfed and bottle-fed subjects, in line with a recent meta-analysis.2 However, the current study was only powered to detect differences of 6.5 mm Hg systolic and 3.2 mm Hg diastolic blood pressure. Despite recent interest in the relationship between breastfeeding and obesity, our findings indicate no evidence of any such association.
Breastfeeding is associated with a reduction in atherosclerosis, but the mechanism is unclear. Prospective investigations of the association between breastfeeding and ischemic heart disease are lacking. Furthermore, such studies would have to be prohibitively large and long term to detect small but (on a population level) important reductions in ischemic heart disease. Approximately 40% of infants are never breastfed in the United Kingdom.24 In the absence of prospective evidence, this study suggests the possibility that the promotion of breastfeeding could be a potential component of the public health strategy to reduce future levels of ischemic cardiovascular disease. However, further studies in large adult populations are needed to confirm these findings. In particular, the hypothesis that breastfeeding influences later cardiovascular disease risk factors could ethically and feasibly be tested on an intention-to-treat basis in large controlled trials of successful breastfeeding promotion interventions with long-term follow-up.41
Acknowledgments
The Wellcome Trust funded R.M.M. to undertake the clinical follow-up as part of a research training fellowship in clinical epidemiology (grant GR063779FR). The arterial ultrasound scans were performed and analyzed with funding by the British Heart Foundation (BHF; project grant PG/02/125). We are very grateful to the cohort members who participated so willingly in the follow-up study. We also acknowledge all the research workers in the original survey in 1937-9. Susie Potts is thanked for all her hard work in providing secretarial and administrative support to the study. The hypothesis was developed by D.G., G.D.S., R.M.M., S.F., and S.E. The field work was conducted by R.M.M., N.G., M.G., and S.W. under the direction of a steering group (S.E., G.D.S., S.F., D.G., and J.H.). A.N. provided expert advice on the conduct and analysis of the arterial scans. N.G., M.G., and A.N. analyzed the ultrasound scan data. R.M.M. did the statistical analysis, wrote the first draft and coordinated completion of the article. R.M.M. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors critically commented on and edited earlier drafts and approved the final version of this article.
References
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: 597–608.
Owen CG, Whincup PH, Gilg JA, Cook DG. Effect of breast feeding in infancy on blood pressure in later life: systematic review and meta-analysis. BMJ. 2003; 327: 1189–1195.
Ravelli ACJ, van der Meulen JH, Osmond C, Barker DJP, Bleker OP. Infant feeding and adult glucose tolerance, lipid profile, blood pressure, and obesity. Arch Dis Child. 2000; 82: 248–252.
Pettitt DJ, Forman MR, Hanson RL, Knowler WC, Bennett PH. Breastfeeding and incidence of non-insulin-dependent diabetes mellitus in Pima Indians. Lancet. 1997; 350: 166–168.
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.
Fall CHD, Barker DJP, Osmond C, Winter PD, Clark PMS, Hales CN. Relation of infant feeding to adult serum cholesterol concentration and death from ischaemic heart disease. BMJ. 1992; 304: 801–805.
Singhal A, Lucas A. Early origins of cardiovascular disease: is there a unifying hypothesis? Lancet. 2004; 363: 1642–1645.
Osborn GR. Stages in development of coronary disease observed from 1500 young subjects: relationship of hypotension and infant feeding to aetiology. Colloq Int Cent Natl Rech Sci. 1967; 169: 93–139.
Martin RM, McCarthy A, Davies DP, Davey Smith G, Ben-Shlomo Y. Association between infant nutrition and blood pressure in early adulthood: the Barry Caerphilly Growth cohort study. Am J Clin Nutr. 2003; 77: 1489–1497.
Wingard DL, Criqui MH, Edelstein SL, Tucker J, Tomlinson-Keasey C, Schwartz JE, Friedman HS. Is breast-feeding associated with adult longevity? Am J Public Health. 1994; 84: 1458–1462.
Burr ML, Beasley WH, Fisher CB. Breast feeding, maternal smoking and early atheroma. Eur Heart J. 1984; 5: 588–591.
Cowen DD. Myocardial infarction and infant feeding. Practitioner. 1973; 210: 661–663.
Lewis DS, Mott GE, McMahan CA, Masoro EJ, Carey KD, McGill HC Jr. Deferred effects of preweaning diet on atherosclerosis in adolescent baboons. Arteriosclerosis. 1988; 8: 274–280.
Ebrahim S, Papacosta O, Whincup P, Wannamethee G, Walker M, Nicolaides AN, Dhanjil S, Griffin M, Belcaro G, Rumley A, Lowe GD. Carotid plaque, intima media thickness, cardiovascular risk factors, and prevalent cardiovascular disease in men and women: the British Regional Heart Study. Stroke. 1999; 30: 841–850.
Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE. Common carotid intima-media thickness and risk of stroke and myocardial infarction: the Rotterdam Study. Circulation. 1997; 96: 1432–1437.
Leng GC, Papacosta O, Whincup P, Wannamethee G, Walker M, Ebrahim S, Nicolaides AN, Dhanjil S, Griffin M, Belcaro G, Rumley A, Lowe GD. Femoral atherosclerosis in an older British population: prevalence and risk factors. Atherosclerosis. 2000; 152: 167–174.
Gunnell DJ, Frankel S, Nanchahal K, Braddon FE, Smith GD. Lifecourse exposure and later disease: a follow-up study based on a survey of family diet and health in pre-war Britain (1937–1939). Public Health. 1996; 110: 85–94.
Rowett Research Institute. Family Diet and Health in Pre-War Britain. Dunfermline: Carnegie United Kingdom Trust; 1955.
Ziegler A, Kastner C, Chang-Claude J. Analysis of pregnancy and other factors on detection of human papilloma virus (HPV) infection using weighted estimating equations for follow-up data. Stat Med. 2003; 22: 2217–2233.
Martin RM, Davey Smith G, Tilling K, Frankel S, Gunnell D. Breastfeeding and cardiovascular mortality: the Boyd Orr cohort and a systematic review with meta-analysis. Eur Heart J. 2004; 25: 778–786.
Gunnell D. Can adult anthropometry be used as a biomarker for prenatal and childhood exposures? Int J Epidemiol. 2002; 31: 390–394.
Khaw KT, Wareham N, Luben R, Bingham S, Oakes S, Welch A, Day N. Glycated haemoglobin, diabetes, and mortality in men in Norfolk cohort of European prospective investigation of cancer and nutrition (EPIC-Norfolk). BMJ. 2001; 322: 15–18.
Fildes V. Infant feeding practices and infant mortality in England, 1900–1919. Contin Chang. 1998; 13: 251–280.
Foster K, Lader D, Cheesborough S. Infant Feeding 1995: A Survey of Infant Feeding Practices in the United Kingdom Carried Out by the Social Survey Division of ONS on Behalf of the Department of Health, the Scottish Office Department of Health, the Welsh Office and the Department of Health and Social Services in Northern Ireland. London, UK: The Stationery Office; 1997.
Simmons D. NIDDM and breastfeeding. Lancet. 1997; 350: 157–158.
Raitakari OT, Juonala M, Kahonen M, Taittonen L, Laitinen T, Maki-Torkko N, Jarvisalo MJ, Uhari M, Jokinen E, Ronnemaa T, Akerblom HK, Viikari JSA. Cardiovascular risk factors in childhood and carotid artery intima-media thickness in adulthood: the Cardiovascular Risk in Young Finns Study. J Am Med Assoc. 2003; 290: 2277–2283.
Lucas A, Sarson DL, Blackburn AM, Adrian TE, Aynsley-Green A, Bloom SR. Breast vs bottle: endocrine responses are different with formula feeding. Lancet. 1980; 1: 1267–1269.
Hofman A, Hazebroek A, Valkenburg HA. A randomized trial of sodium intake and blood pressure in newborn infants. J Am Med Assoc. 1983; 250: 370–373.
Danesh J. Coronary heart disease, Helicobacter pylori, dental disease, Chlamydia pneumoniae, and cytomegalovirus: meta-analyses of prospective studies. Am Heart J;. 138: S434–S437.
Fall CHD, Goggin PM, Hawtin P, Fine D, Duggleby S. Growth in infancy, infant feeding, childhood living conditions, and Helicobacter pylori infection at age 70. Arch Dis Child. 1997; 77: 310–314.
Kark JD, Troya G, Friedlander Y, Slater PE, Stein Y. Validity of maternal reporting history and the association with blood lipids in 17 year olds in Jerusalem. J Epidemiol Community Health. 1984; 38: 218–225.
Baxter-Jones ADG, Cardy AH, Helms PJ, Phillips DO, Smith WC. Influence of socioeconomic conditions on growth in infancy: the 1921 Aberdeen birth cohort. Arch Dis Child. 1999; 81: 5–9.
Martyn CN, Gale CR, Jespersen S, Sherriff SB. Impaired fetal growth and atherosclerosis of carotid and peripheral arteries. Lancet. 1998; 352: 173–178.
Singhal A, Fewtrell M, Cole TJ, Lucas A. Low nutrient intake and early growth for later insulin resistance in adolescents born preterm. Lancet. 2003; 361: 1089–1097.
Singhal A, Cole TJ, Fewtrell M, Deanfield J, Lucas A. Is Slower Early Growth Beneficial for Long-Term Cardiovascular Health? Circulation. 2004; 109: 1108–1113.
Williams-Blangero S, Rainwater DL. Variation in Lp(a) levels and apo(a) isoform frequencies in five baboon subspecies. Hum Biol. 1991; 63: 65–76.
Maynard M, Gunnell D, Emmett P, Frankel S, Davey Smith G. Fruit, vegetables, and antioxidants in childhood and risk of adult cancer: the Boyd Orr cohort. J Epidemiol Community Health. 2003; 57: 218–225.
Emmett PM, Rogers IS. Properties of human milk and their relationship with maternal nutrition. Early Hum Dev. 1997; 49 (suppl): S7–S28.
Diaz-Gomez NM, Domenech E, Barroso F. Breast feeding and growth factors in preterm newborn infants. J Pediatr Gastroenterol Nutr. 1997; 24: 322–327.
Rich-Edwards JW, Stampfer MJ, Manson JE, Rosner B, Hu FB, Michels KB, Willett W. Breastfeeding during infancy and the risk of cardiovascular disease in adulthood. Epidemiology. 2004; 15: 550–556.
Kramer MS, Chalmers B, Hodnett ED, Sevkovskaya Z, Dzikovich I, Shapiro S, Collet JP, Vanilovich I, Mezen I, Ducruet T, Shishko G, Zubovich V, Mknuik D, Gluchanina E, Dombrovskiy V, Ustinovitch A, Kot T, Bogdanovich N, Ovchinikova L, Helsing E; PROBIT Study Group (Promotion of Breastfeeding Intervention Trial). Promotion of Breastfeeding Intervention Trial (PROBIT): a randomized trial in the Republic of Belarus. J Am Med Assoc. 2001; 285: 413–420.(Richard M. Martin; Shah E)