Early nutrition in preterm infants and later blood pressure: twocohorts after randomised trials

7/31/2019 Early nutrition in preterm infants and later blood pressure: twocohorts after randomised trials

1/7

For personal use only. Reproduce with permission from The Lancet Publishing Group.

ARTICLES

Summary

Background Despite data relating body size in early life to

later cardiovascular outcomes, the hypothesis that nutrition

affects such outcomes has not been established.

Breastfeeding has been associated with lower blood

pressure in later life, but previous studies have not

controlled for possible confounding factors by using a

randomised design with prospective follow-up. We undertook

such a study to test the hypothesis that early dietprogrammes blood pressure in later life in children randomly

assigned different diets at birth.

Methods Blood pressure was measured at age 1316 years

in 216 (23%) of a cohort of 926 children who were born

prematurely and had participated at birth in two parallel

randomised trials in five neonatal units in the UK. Dietary

interventions were: donated banked breastmilk versus

preterm formula and standard term formula versus preterm

formula.

Findings Children followed up at age 1316 years were

similar to those not followed up in terms of social class and

anthropometry at birth. Mean arterial blood pressure at

age 1316 years was lower in the 66 childrenassigned banked breastmilk (alone or in addition to

mothers milk) than in the 64 assigned preterm formula

(mean 819 [SD 78] vs 861 [65] mm Hg; 95% CI for

difference 66 to 16; p=0001). In non-randomised

analyses, the proportion of enteral intake as human milk in

the neonatal period was inversely related to later mean

arterial pressure (=03 mm Hg per 10% increase [95% CI

05 to 01]; p=0006). No differences were found in the

term formula (n=44) versus preterm formula (n=42)

comparison.

Interpretation Breastmilk consumption was associated with

lower later blood pressure in children born prematurely. Our

data provide experimental evidence of programming of a

cardiovascular risk factor by early diet and further support thelong-term beneficial effects of breastmilk.

Lancet 2001; 357: 41319

See Commentary page 406

Medical Research Council Childhood Nutrition Research Centre

(A Singhal MRCP, Prof A Lucas FRCP) and Department of Paediatric

Epidemiology and Biostatistics (Prof T J Cole PhD), Institute of

Child Health, London WC1N 1EH, UK

Correspondence to: Dr Atul Singhal

(e-mail: [email protected])

IntroductionThe potential effect of early nutrition on cardiovasculardisease is an issue of major public-health importance.Associations between low weight1 and thinness2 at birthand hypertension and coronary-artery disease in later lifeare thought to be consistent with the hypothesis thatundernutrition in early life programmes latercardiovascular outcomes such as blood pressure.38 Agreater risk of death from coronary heart disease in men

who were light for age at 1 year

9

also suggests that thecritical period of sensitivity for nutrition in terms of itsinfluence on cardiovascular disease extends into the firstyear of postnatal life.

Despite these strong epidemiological data relating sizein early life to later outcome, an effect of fetal or infantnutrition on later risk factors for ischaemic heart disease istheoretical and has not been adequately establishedbecause of the lack of a randomised design in previousstudies.

Breastfeeding has been associated with lower bloodpressure in later life,8,10 but observational data cannotextricate any possible effects of diet from environmentalinfluences associated with the maternal choice tobreastfeed. An experimental study would be needed, but

randomisation to breastfeeding or formula feeding wouldbe difficult. We had an opportunity, however, to test thehypothesis that feeding of breastmilk to premature infantsprogrammes later blood pressure in a formal experimentalintervention trial. In the early 1980s, human-milk bankswere in common use to provide preterm infants withbreastmilk from unrelated donors. Random assignment ofpreterm infants to donated human milk or formula wasethical because at that time the optimum diet for babiesborn preterm was uncertain and the long-term outcomesof early feeding regimens unknown.11

Our study started in 1982. It consisted of two parallelrandomised trials: one compared banked breastmilk withnutrient-enriched preterm formula; the other compared astandard term formula with the nutrient-enriched preterm

formula. Our planned follow-up allowed us to test thegeneric hypothesis that early nutrition influences laterblood pressure and the a-priori specific hypothesis thatconsumption of human milk in infancy leads to lowerblood pressure in later life.

MethodsParticipants

The study participants were recruited from a cohort of926 children born preterm who had participated inrandomised controlled trials initially investigating theeffects of early diet on later cognitive function.12 Plannedlater follow-up of this cohort was designed to test thehypothesis that early diet influences risk factors forcardiovascular disease.13

Early nutrition in preterm infants and later blood pressure: two

cohorts after randomised trials

Atul Singhal, Tim J Cole, Alan Lucas

Articles

THE LANCET Vol 357 February 10, 2001 413


2/7


Between 1982 and 1985, babies with no majorcongenital anomalies and of birthweight less than1850 g were recruited in five UK centres (Norwich,Cambridge, Sheffield, Ipswich, and Kings Lynn) andrandomly assigned, in two parallel trials, breastmilkdonated by unrelated lactating women, or nutrient-enriched preterm formula (Farleys Osterprem, CrookesHealth Care, Kendal, Cumbria, UK) or standard termformula (Farleys Ostermilk) versus nutrient-enrichedpreterm formula.11 Within each trial the diets wererandomly assigned in two strata: the trial diets alone (A)and, in mothers who elected to express their own milk,the trial diets were assigned as supplements to mothersmilk (B; figure 1). To compare the randomised diets as

originally planned, strata A and B have been combinedwithin each trial as a balanced addition, therebypreserving randomisation. Diet assignment was madewithin 48 h of birth by means of sealed envelopes.11

Ethical approval for the trial was obtained from eachcentre, and informed consent was obtained from eachparent (no parent refused consent).

The compositions of the assigned diets, given until theinfant weighed 2000 g or was discharged home, havebeen described fully elsewhere.1113 Preterm formula,compared with standard formula, was enriched in

protein and fat (20 g protein and 49 g fat per 100 mLcompared with 15 g protein and 38 g fat per 100 mL)but not in carbohydrate (70 g per 100 mL in both).12

For infants fed maternal milk, intakes of protein andenergy were estimated from 4935 complete 24 hcollections of milk (about 15 g protein, 30 g fat, and70 g carbohydrate per 100 mL), and nutrient values fordonated banked breastmilk were measured in more than600 samples of pooled banked breastmilk and theaverages calculated (about 11 g protein, 20 g fat, and70 g carbohydrate per 100 mL). The sodiumconcentration was 196 mmol/L in preterm formula,110 mmol/L in expressed breastmilk, 83 mmol/L interm formula, and 72 mmol/L in banked breastmilk.13

Preterm formula was also enriched in vitamins, zinc,and copper12 and was designed overall to meet thespecial nutrient needs of the preterm infant.

Extensive demographic, social, anthropometric,biochemical, and clinical data were collected throughoutthe hospital admission.1113 Infants were weighed daily bytrained staff, and weight was expressed as an absolutevalue and as an SD from expected weight (z score)based on birth centiles for preterm infants.14 The z scorefor weight at discharge (or on discontinuation of theassigned diet) was also derived and represented both

ARTICLES

414 THE LANCET Vol 357 February 10, 2001

502 infants

randomised

253 assigned

banked breastmilk

A

B

83

170

232 survived

A

B

75

157

66 followed up

A

B

13

53

249 assigned

preterm formula

A

B

76

173

224 survived

A

B

67

157

64 followed up

A

B

17

47

Trial 1: Banked breastmilk versus preterm formula Trial 2: Term formula versus preterm formula

424 infants

randomised

211 assigned

term formula

A

B

79

132

188 survived

A

B

71

117

44 followed up

A

B

14

30

213 assigned

preterm formula

A

B

81

132

188 survived

A

B

70

118

42 followed up

A

B

17

25

Figure 1: Derivation of sample followed up at age 1316 yearsA=infants receiving assigned milk as sole diet. B=infants receiving assigned milk as supplement to mothers expressed milk.

Variable Banked breastmilk vs preterm formula Term formula vs preterm formula

Number of children Mean (SD) value p Number of children Mean (SD) value p

Followed up Not Followed up Not Followed up Not Followed up Notfollowed up followed up followed up followed up

Growth

Birthweight (kg) 130 372 14 (03) 14 (03) 085 86 338 14 (03) 14 (03) 071

Duration of gestation (weeks) 130 372 311 (26) 307 (29) 013 85 338 307 (28) 308 (29) 089

z score birthweight 130 372 10 (12) 07 (13) 010 85 338 08 (11) 07 (13) 079

z score discharge weight 130 367 21 (10) 20 (11) 053 85 338 21 (10) 21 (10) 067

Weight gain (g/kg daily) 109 310 149 (35) 144 (44) 031 71 250 144 (35) 144 (38) 097

Length gain (mm daily) 72 216 13 (04) 13 (07) 075 47 174 15 (10) 14 (06) 027

Demographic and clinical

Social class 130 360 34 (15) 36 (19) 030 84 324 35 (16) 38 (18) 011

Apgar score at 5 min 124 353 83 (17) 80 (19) 012 80 323 78 (18) 80 (20) 052

Days of ventilation* 130 370 0 (04) 1 (05) 012 86 338 1 (04) 1 (06) 055

Infections* 130 370 1 (01) 1 (01) 075 86 338 1 (01) 1 (01) 080

*Median (IQR), analysed by Mann-Whitney test.

Based on number of courses of antibiotics.

Table 1: Characteristics of children followed up and not followed up at age 1316 years


3/7


intrauterine and extrauterine growth. Social class wasbased on the occupation of the parent providing themain financial support for the family (or if both parentsworked, the fathers occupation) and expressed as acode according to the Registrar Generals Classification.

Follow-up

The follow-up at age 1316 years involved measurementof four key cardiovascular outcomes: blood pressure,endothelial dysfunction15 (an indicator of early

atheroma), plasma cholesterol concentration, andconcentration of 3233 split proinsulin (as a measure ofinsulin resistance). This paper reports the follow-upmeasurement of blood pressure. Sample size wasestimated to exclude a difference of half an SD in theseoutcomes between randomised dietary groups in each ofthe trials. We needed a maximum sample of around 250participants to detect this difference (125 in each trial)at 80% power and 5% significance, and a minimumsample of around 200 participants for 70% power and5% significance.16 This sample size was sufficient todetect a difference of 5 mm Hg (half 1 SD) in systolicblood pressure between the dietary groups, an estimateddifference similar to the difference in blood pressurepreviously reported in adolescents breastfed for more

than 3 months and those not breastfed45 mm Hg ingirls and 65 mm Hg in boys.8 In all, 216 children agreedto participate. Ethical approval for the follow-up studywas obtained from national multicentre and localresearch ethics committees, and written consent wasobtained from all children and their guardians.

The original dietary assignments were concealed fromall researchers and participants. Blood pressure wasmeasured in the left arm after the participant had lainsupine for 10 min, with an automated device(Accutorrsat, Datascope Corporation, Montvale, NJ,USA) and cuff of appropriate size.17 The measurementwas repeated after a non-invasive ultrasonographicassessment of the right brachial artery,15 with theparticipant resting supine throughout the procedure.Systolic, diastolic, and mean arterial blood pressureswere recorded on each occasion, and the mean value forthe two measurements was calculated. The meanarterial blood pressure was measured automatically asthe moving average of the area under the curve of bloodpressure versus time for each cardiac cycle.

Height was measured with a portable stadiometeraccurate to 1 mm (Holtain Instruments Ltd,Crymmych, Dyfed, UK) and weight with electronicscales accurate to 01 kg (Seca, CMS WeightEquipment Ltd, London, UK). Most measurements(99%) were made by one of two trained observers.Equipment was calibrated before each field visit, and theobservers measuring techniques were monitoredthroughout. Socioeconomic status was obtained from

the number of people per room in the house (excludingkitchen and bathrooms) and from social class at birth.Tanner staging was done by the participant in privatewith standard Tanner stage photographs.

Analysis

Comparisons of normally distributed variables betweenrandomised groups were made with a two-sidedStudents ttest. Simultaneous multiple linear regressionanalyses were used for epidemiological (non-randomised) analyses of data from trials 1 and 2combined, with adjustment for baseline differences.

ResultsObservations in cohorts after randomised trials

There were no significant differences in birthweight,duration of gestation, social class, SD scores forbirthweight and discharge weight, and clinical variablesbetween children who were or were not followed up atage 1316 years (table 1).

As expected, early weight gain was significantlygreater in infants assigned nutrient-enriched pretermformula than in those assigned banked breastmilk(table 2). Among the children followed up at age 1316years, there were no significant differences in systolic ordiastolic blood pressure between the randomised groupsin trials 1 and 2 at age 7580 years (data notpresented).

At age 1316 years, mean arterial pressure wassignificantly lower in children assigned banked

breastmilk than in those assigned preterm formula(mean difference 41 mm Hg [95% CI 66 to 16],p=0001; table 3). Diastolic blood pressure was alsosignificantly lower in infants assigned banked breastmilkthan in those assigned preterm formula, but thedifferences for systolic blood pressure were not

ARTICLES


Variable Mean (SD) blood pressure (mm Hg) Mean (95% CI) difference p Mean (SD) blood pressure (mm Hg) Mean (95% CI) difference p

Banked breastmilk Preterm formula Term formula Preterm formula

(n=66) (n=64) (n=44) (n=42)

Diastolic 619 (81) 650 (67) 32 (58 to 06) 0016 641 (77) 643 (68) 02 (34 to 29) 088

Systolic 1136 (90) 1163 (80) 27 (57 to 03) 0075 1179 (80) 1164 (78) 10 (19 to 49) 037

Mean arterial 819 (78) 861 (65) 41 (66 to 16) 0001 855 (73) 845 (64) 15 (20 to 39) 051

Table 3: Blood pressure in relation to early diet

Variable Banked Preterm Term Preterm

breastmilk formula formula formula

(n=66) (n=64) (n=44) (n=42)

At follow-up

Number male 32 (49%) 32 (50%) 20 (45%) 13 (31%)

Age (years) 152 (09) 151 (10) 148 (08) 148 (08)

Weight (kg) 539 (99) 550 (122) 586 (97) 549 (101)

Height (cm) 1613 (102) 1608 (94) 1625 (89) 1619 (73)

Body-mass index (kg/m2) 208 (39) 211 (39) 222 (35) 209 (32)

People per room 07 (02) 07 (02) 07 (02) 08 (03)

Tanner stage

Pubic hair 43 (07) 42 (07) 41 (07) 44 (05)Genitalia/breast 42 (08) 41 (08) 40 (08) 41 (07)

Neonatal

Birthweight (kg) 14 (03) 13 (03) 13 (03) 14 (04)

Gestation (weeks) 311 (25) 312 (26) 306 (29) 309 (28)

z score birthweight 08 (12) 11 (12) 08 (11) 07 (11)

z score discharge weight 21 (10) 20 (10) 23 (08) 18 (09)*

Weight gain (g/kg daily) 140 (29) 158 (37)* 137 (23) 149 (43)

Length gain (mm daily) 12 (04) 14 (05) 15 (11) 15 (10)

Social class 34 (17) 35 (13) 36 (16) 34 (17)

Number of non-manual 28 (42%) 25 (39%) 17 (39%) 18 (43%)

social class

Apgar score at 5 min 82 (16) 84 (19) 75 (19) 82 (15)

Days of ventilation 0 (03) 0 (04) 1 (04) 1 (04)

Infections 1 (01) 1 (01) 1 (02) 1 (01)

Number who received 53 47 30 25

assigned diet as

supplement

Data are mean (SD) unless otherwise stated. *p


4/7


significant (table 3). Blood pressure remained lower ininfants fed banked breastmilk than in those fed pretermformula after adjustment for sex, mean daily enteralsodium intake in infancy, and current body-mass index(for mean blood pressure adjusted mean 821 vs 860mm Hg, p=0005; for diastolic blood pressure adjustedmean 619 vs 650 mm Hg, p=003).

Both mean and diastolic blood pressure weresignificantly lower in children assigned bankedbreastmilk than in those assigned preterm formula whenthe analysis was confined to children whose mothers hadno evidence of pregnancy-induced hypertension or tochildren whose diastolic blood pressure was below 1 SDabove the mean for the whole population, a cut-offchosen to reduce the influence of children with higherblood pressures on analyses (data not presented).

Diastolic, systolic, and mean arterial blood pressuresdid not, however, significantly differ between childrenassigned term versus preterm formula (table 3); thesedifferences remained non-significant after adjustmentfor sex, mean daily sodium intake in infancy, andcurrent body-mass index (data not presented).

Mean arterial pressure was lower in both boys andgirls assigned banked breastmilk than in those assigned

preterm formula (boys mean 832 [SD 66] vs 867 [59]mm Hg; 95% CI for difference 67 to 04; p=0028;girls 808 [87] vs 854 [70] mm Hg; 95% CI fordifference 85 to 08; p=002). The interaction of sexwith diet was not significant (p=066).

Among children whose mothers chose to providebreastmilk in addition to the assigned diet, diastolic,systolic, and mean arterial blood pressures were lower inthose assigned banked breastmilk than in those assignedpreterm formula (diastolic 625 [78] vs 654 [66]mm Hg; 95% CI for difference 58 to 0; p=005;systolic 1139 [87] vs 1178 [79] mm Hg; 95% CI fordifference 72 to 06; p=002; mean arterial 823[76] vs 871 [62] mm Hg; 95% CI for difference76 to 20; p=0001). Diastolic and mean arterial

blood pressures were also lower in children who receivedonly banked breastmilk than in those who received onlypreterm formula but these differences were notsignificant (data not presented).

Mean arterial pressure was just significantly greater inchildren assigned term formula alone than in thoseassigned preterm formula (p=005), but the differencewas not significant after adjustment for body-mass indexand sex (p=022). There were no other significantdifferences in any measure of blood pressure in the trialof term formula versus preterm formula in childrenreceiving the assigned diet alone or as a supplement tomothers milk (data not presented).

Non-randomised observations

The proportion of enteral intake (by volume) consumedas human milk was significantly related to mean arterialpressure (=031 mm Hg per 10% increase in human-milk intake [95% CI 052 to 010], p=0006) anddiastolic blood pressure (028 mm Hg per 10%increase [050 to 005], p=0015) but not to systolicblood pressure (024 mm Hg per 10% increase[049 to 002], p=007). Diastolic blood pressure inthe thirds of proportion human-milk intake is shown infigure 2.

After adjustment for possible confounding factors(age, sex, body-mass index, duration of gestation, andsocial class), blood pressure remained significantlyassociated with proportion human-milk intake (meanarterial pressure =030 mm Hg per 10% increase

[050 to 009], p=0006; diastolic 025 mm Hg per10% increase [047 to 003], p=0027; systolic025 mm Hg per 10% increase [049 to 001],p=0042).

The proportion of enteral intake consumed asexpressed breastmilk was significantly related to meanarterial pressure (025 mm Hg per 10% increase[050 to 0], p=005) and diastolic blood pressure(026 mm Hg per 10% increase [052 to 0],p=0048), but not to systolic blood pressure(016 mm Hg per 10% increase [046 to 014],p=029). After adjustment for confounding factors asabove, the association between intake of expressed

breastmilk and blood pressure was significant for meanarterial pressure (029 mm Hg per 10% increase[053 to 005], p=0019) and diastolic (026mm Hg per 10% increase [051 to 001], p=0041)but not systolic blood pressure (023 mm Hg per 10%increase [051 to 005], p=010).

Mean arterial pressure was not associated withbirthweight z score without (022 mm Hg per SDbirthweight; p=060) or with adjustment for currentweight z score (016 mm Hg per SD birthweight;p=069). There was no relation between blood pressureand discharge-weight z score (an indicator of antenataland postnatal growth) or this score after adjustment forbirthweight z score or current-weight z score (data notpresented).

Neither energy intake (total or that from enteralfeeds) nor protein intake was related to mean arterialpressure (data not presented).

DiscussionWe tested the hypothesis that diet in infancy influencesor programmes a key cardiovascular risk factor, namelyblood pressure, in later life. Our study in preterm infantstook two approaches: two parallel randomised trials thatcompared banked donated breastmilk with an infantformula or two formulas differing in nutrient content;and an observational study comparing human milk withother diets.

Our major finding related to the studies includinguse of breastmilk. We found lower blood pressure at age

ARTICLES


68

66

64

62

60

58

0

Diastolicblood

pressure(mmH

g)

Lowest Middle Highest

Third of human-milk-intake distribution

Figure 2: Diastolic blood pressure in thirds of proportional

intake of human milk

Error bars=95% CI.


5/7


1316 years in children who had been fed human milkboth in the randomised trial and in the observationalstudy. The effect on blood pressure in the randomisedtrial was significant in both male and female subgroups.Our findings are consistent with the hypothesis that highblood pressure in later life, at least partly, has earlynutritional origins and provide supportive evidence for apotential long-term beneficial effect of breast-feeding.Many observational studies have explored the relationbetween early size and later blood pressure, butpreviously there has been no evidence that earlynutrition could explain the observed effects. Our study,however, initiated nearly 20 years ago, providesprospective, experimental evidence for early nutritionalprogramming of a cardiovascular outcome.

Of relevance to our main finding was that systolicblood pressure at a mean age of 73 years was lower in aprevious study of children who had received breastmilk10

than in those who had been entirely bottle fed. Also, inadolescents, breastfeeding for longer than 3 months wasassociated with lower systolic blood pressure by up to65 mm Hg.8 However, in both of those epidemiologicalstudies, confounding from the differences insocioeconomic backgrounds of breastfed and formula-

fed babies could have accounted for the benefits ofbreastfeeding.In this study, we planned to control for such

confounding factors by random allocation of infantsborn preterm to human or formula milk. Diastolic bloodpressure was 32 mm Hg lower in infants assignedbanked breastmilk than in those assigned a pretermformula, a benefit similar to the 23 mm Hg differencein diastolic and 35 mm Hg difference in systolic bloodpressure observed previously in the non-randomisedstudy10 of children born at term. The association ofproportion of enteral intake consumed as human milk(or expressed breastmilk) with lower blood pressure inadolescence, independent of anthropometry andsocioeconomic status, was consistent with a dose-

response relation, further supporting a causalexplanation (figure 2). Collectively, current data nowsupport the hypothesis that human milk has a long-termprogramming effect on blood pressure, although thepossibility that formula milk increases later bloodpressure cannot be excluded.

Follow-up of this preterm cohort at age 75 to 80years did not show any effect of diet on later bloodpressure,13 even in the subpopulation followed up at1316 years. However, amplification of dietaryinfluences on blood pressure with increasing age1,6,7

could explain the lack of effect of breastmilk intake onblood pressure in our earlier follow-up and in previousstudies by others in younger children.4,18 Nutritionalprogramming of blood pressure might not manifest until

puberty or beyond, analogous to the deferred effects ofearly diet on obesity19 and atheroma20 observed only afteradolescence in baboons. In a longitudinal study, therewas stronger correlation between birthweight andsystolic blood pressure after rather than before puberty,again consistent with this hypothesis.8

The only other area tested in randomised trials on theissue of early nutrition and later cardiovascular riskfactors is the effect of early salt intake on later bloodpressure. In one randomised study, a low sodium intakein infancy was associated with lower systolic bloodpressure.21 This effect, however, emerged only afterunplanned, post-hoc adjustment for several variablesthat effectively removed the formal experimentalelement of the study.21 Moreover, the importance of

early sodium intake in the programming of later bloodpressure is not established, and two other prospectivestudies showed no effect of early salt intake on laterblood pressure.22,23 In our study that compared the twoformulas (standard versus nutrient-enriched pretermformula), we found no effect on later blood pressuredespite the major difference in formula sodium content;this finding suggests that further work is needed toestablish a connection between early sodium intake andlater hypertension.

The possible mechanisms for the early dietaryprogramming of later blood pressure require furtherinvestigation. In our study of adolescents, bloodpressure was not related to birthweight for gestationalage, which is consistent with reports suggesting that thisrelation, seen at other ages, may be disturbed by thepubertal growth spurt.1,24,25 Adjustment for the z scorefor current weight did not affect this association, whichsuggests that neither fetal growth nor the change inweight between birth and adolescence was related tolater blood pressure.26 The possibility that a relationbetween early size and later outcome may not occur ininfants born preterm (in whom the cause of growthfailure differs from that in infants born at term) is

unlikely, because prematurity does not affectassociations between birthweight for gestation and latercardiovascular outcomes,2,27 including blood pressure atother ages.3,5,7

We have noted possible statistical flaws in theinterpretation of data relating birthweight to lateroutcomes, including blood pressure.26 We suggestedthat much of what was claimed to be fetal in originmay in fact relate to postnatal nutrition and growth.26

The findings of this study are consistent with thisproposal.

In the UK, donor breastmilk (provided by unrelateddonors to a milk bank) was generally the low-fatforemilk that dripped from the contralateral breastduring breastfeeding of the donors own baby. We first

considered the possibility that the lower fat andtherefore energy content of such donor breastmilk,rather than any unique property of human milk itself,accounted for our findings. We now think thispossibility is unlikely. First, we found anepidemiological association between intake of theinfants own mothers expressed breastmilk (which has ahigher nutrient content than donated breastmilk) andlower later blood pressure. Second, the publishedepidemiological data8,10 on breastfeeding and subsequentlow blood pressure were derived from normallynourished individuals born at full term. Our findingsnarrow the range of dietary constituents that mightbring about the programming of later blood pressure.The benefits of breastmilk were likely to be independent

of enteral sodium intake; indeed blood pressure wassimilar in children receiving term or preterm formuladespite differences in contents of protein, energy,minerals, and micronutrients. These observationssuggest either a role for non-nutritive factors inbreastmilk or for dietary factors in breastmilk but not inthe formulas used.

A decrease of about 3 mm Hg in diastolic bloodpressure, estimated from our own and previousepidemiological studies, is likely to have substantialpublic-health implications. For instance, analysis ofrandomised trials and the Framingham study suggeststhat a lowering of population-wide diastolic bloodpressure in adults by only 2 mm Hg would reduce theprevalence of hypertension by 17%, the risk of coronary

ARTICLES



6/7


heart disease by 6%, and the risk of stroke and transientischaemic attacks by 15%.28 Addition of such anintervention to existing hypertension treatment couldtherefore prevent an estimated 67 000 (6%) events ofcoronary heart disease and 34 000 (13%) of stroke ortransient ischaemic attack, each year, among the 3564age-group, in the USA alone.28 Furthermore, a meanreduction in blood pressure of 3 mm Hg underestimatesthe degree of lowering and benefit to children at theextremes of the distribution for blood-pressure changeand, because blood pressure tracks from childhood intoadult life with amplification of early differences,1,6,7 theblood-pressure difference between breastfed and non-breastfed populations is likely to be substantially greater inadulthood. Finally, in a non-hypertensive population, thepotential benefit of breastfeeding is greater than thebenefit, estimated in a meta-analysis, from all other non-pharmacological means of lowering blood pressure, suchas weight loss, salt restriction, or exercise.29

We studied only a subsample of our original studypopulation. However, this sample was representative ofthe children recruited at birth, and characteristics ofchildren who were reviewed at age 1316 years did notdiffer between the randomised groups at birth or at follow-

up. Moreover, there is no prior reason why theepidemiological association between proportion of humanmilk consumed and lower later blood pressure shoulddiffer between adolescents reviewed and those notfollowed up. As discussed previously,13 our population wassimilar to the general population in terms of social class,an important factor in studies of the early influence onlater cardiovascular disease.

The applicability of our data to populations born atterm is important. A randomised controlled trial ofbreastfeeding versus formula feeding would be difficult ininfants born at term. However, our finding that the effectof human milk on blood pressure was independent ofgestation and the similarity of the degree of benefit in ourstudy and that observed in the two most robust

prospective epidemiological studies in infants born atterm, strongly suggest that the benefits of breastfeeding onblood pressure are not confined to preterm infants.Moreover, even if our study applied only to prematurebabies (which now seems unlikely), it would still havesubstantial clinical importance to the 6% of the populationborn preterm.

Consumption of human milk has been shown to havemany benefits for infantsboth preterm and full termincluding a reduced risk of necrotising enterocolitis, atopy,and infection and improved later cognitive development.The hypothesis that consumption of human milk also maylower blood pressure, up to 16 years later, is supported byour randomised trial (involving dietary manipulation foran average of only 1 month) and three observational

studies (ours, reported here, and two previous studies onhealthy children born at term8,10). Our study raises thepossibility that non-nutrient factors, or at least factors notfound in infant formula,30 could be implicated in theeffects observed. Breastmilk contains a wide range of suchfactors, including trophic substances and hormones.Further investigation would be of potential importance inelucidating mechanisms for nutritional programming andfor the beneficial effects of human milk.

ContributorsAtul Singhal initiated the study and wrote the first draft of the paper.

Tim Cole provided statistical expertise. Alan Lucas was responsible for the

design of long-term studies of early nutrition. All three investigators

contributed to study design, analysis of data, and preparation of the

report.

AcknowledgmentsWe thank the staff at the special-care baby units at Cambridge, Ipswich,

Kings Lynn, Sheffield, and Norwich for their assistance and cooperation;

M F Bamford, J F B Dossetor, P Crowle, A Boon, and R Pearse for help

and collaboration; Farley Health Products for the manufacture of the

preterm formula and technical assistance; Mia Kattenhorn, Eve Smith,

Emma Sutton, Kathy Kennedy, and Ian Merryweather for technnical

assistance; and the children and their families who participated in this

project.

References

1 Law CM, Shiell AW. Is blood pressure inversely related to birth

weight? The strength of evidence from a systematic review of the

literature. J Hyper tens 1996; 14: 93541.

2 Barker DJP, Osmond C, Simmonds SJ, Wield GA. The relation of

small head circumference and thinness at birth to death from

cardiovascular disease in adult life. BMJ1993; 306: 42226.

3 Barker DJP, Osmond C, Golding J, Kuh D, Wadsworth MEJ.

Growth in utero, blood pressure in childhood and adult life,

and mortality from cardiovascular disease. BMJ1989; 298:

56467.

4 Whincup PH, Cook DG, Shaper AG. Early influences on blood

pressure: a study of children aged 57 years. BMJ1989; 299:

58791.

5 Whincup PH, Cook DG, Papacosta O. Do maternal and

intrauterine factors influence blood pressure in childhood? Arch Dis

Child1992; 67: 142329.

6 Law CM, De Swiet M, Osmond C, et al. Initiation of hypertension

in utero and its amplification throughout life. BMJ1993; 306:

2427.

7 Whincup P, Cook D, Papacosta O, Walker M. Birth weight and

blood pressure: cross sectional and longitudinal relations in

childhood. BMJ1995; 311: 77376.

8 Taittonen L, Nuutinen M, Turtinen J, Uhari M. Prenatal and

postnatal factors in predicting later blood pressure among children:

cardiovascular risk in young Finns. Pediatr Res 1996; 40: 62732.

9 Barker DJP, Winter PD, Osmond C, Margetts B, Simmonds SJ.

Weight in infancy and death from ischaemic heart disease. Lancet

1989; 2: 57780.

10 Wilson AC, Forsyth JS, Greene SA, Irvine L, Hau C, Howie PW.

Relation of infant diet to childhood health: seven year follow up of

cohort in Dundee infant feeding study. BMJ1998; 316: 2125.

11 Lucas A, Gore SM, Cole TJ, et al. Multicentre trial on feeding of

low birthweight infants: effects of diet on early growth. Arch Dis

Child1984; 59: 72230.

12 Lucas A, Morley R, Cole TJ. Randomised trial of early diet in

preterm babies and later intelligence quotient. BMJ1998;317:

148187.

13 Lucas A, Morley R. Does early nutrition in infants born before term

programme later blood pressure. BMJ1994; 309: 30408.

14 Freeman JV, Cole TJ, Chinn S, Jones PRM, White EM, Preece MA.

Cross-sectional stature and weight reference curves for the UK,

1990.Arch Dis Child1990; 73: 1724.

15 Leeson CPM, Whincup PH, Cook DG, et al. Flow-mediated

dilation in 9- to 11-year-old children: the influence of intrauterine

and childhood factors. Circulation 1997; 96: 223338.

16 Gore SM. Statistical considerations in infant nutrition trials. Lipids

1999; 34: 18597.

17 Anwar YA, Tendler BE, McCabe EJ, Mansoor GA, White WB.

Evaluation of the datascope Accutorr plus according to the

recommendations of the association for the advancement of medical

instrumentation. Blood Press Monit1997; 2: 10510.

18 Baranowski T, Bryan GT, Harrison JA, Rassin DK, Greaves KA,

Baranowski JH. Height, infant-feeding practices and cardiovascular

functioning among 3 or 4 year old children in three ethnic groups.J Clin Epidemiol1992; 45: 51318.

19 Lewis DS, Bertrand HA, McMahan CA, McGill HC, Carey KD,

Masoro EJ. Preweaning food intake influences the adiposity of

young adult baboons.J Clin Invest1986; 78: 899905.

20 Lewis DS, Mott GE, McMahan CA, Masoro EJ, Carey KD,

McGill HC. Deferred effects of preweaning diet on atherosclerosis

in adolescent baboons.Atherosc lerosis 1988; 8: 27480.

21 Geleijnse JM, Hofman A, Witteman JCM, Hazebroek AAJM,

Valkenburg HA, Grobbee DE. Long-term effects of neonatal

sodium restriction on blood pressure. Hypertension 1997;

29: 91317.

22 Whitten CF, Stewart RA. The effect of dietary sodium in infancy on

blood pressure and related factors.Acta Pediatr Scand1980; 279:

117.

23 Lucas A, Morley R, Hudson GJ, et al. Early sodium intake and later

blood pressure in preterm infants.Arch Dis Child1988;

63: 65657.

ARTICLES



7/7


Clinical picture: Occult ectopic ACTH syndrome

D J R Cuthbertson, S Vinjamuri, D Bowen-Jones

A 38-year-old male was diagnosed with Cushing's syndrome. Biochemical tests showedhypokalaemia and hypercortisolaemia, with failure of suppression following high dosedexamethasone, suggesting an ectopic or adrenal source. Adrenocorticotrophic hormonelevels were >100 IU/L (normal 314 IU/L). Pituitary magnetic resonance imaging wasnormal as was the chest radiograph, spiral computed tomography and bronchoscopy.

Adrenal imaging revealed hyperplastic glands with no adenoma.111

In-octreotide scan wasperformed, demonstrating a focal area of increased tracer uptake in the left hemithoraxclose to the midline (figure). The patient was treated with metyrapone and underwentbilateral adrenalectomy but died postoperatively. Post-mortem examination demonstrateda normal pituitary and bilaterally hyperplastic adrenal glands with a bronchial carcinoid inthe left hilar region corresponding to the site of increased tracer uptake. 111In-octreotidescanning may have a useful role in occult ectopic ACTH syndrome.

Department of Nuclear Medicine, Royal Liverpool University Hospital, Liverpool (D J R Cuthbertson MRCP,

S Vinjamuri MD); Department of Diabetes and Endocrinology, Arrow Park Hospital, Upton, Wirral, UK

(D Bowen-Jones MRCP)

24 Seidam DS, Laor A, Gale R, Stevenson DK, Mashiach S,

Danon YL. Birth weight, current body weight, and blood pressure in

late adolescence. BMJ1991; 302: 123537.

25 Laor A, Stevenson DK, Shemer J, Gale R, Seidman DS. Size at

birth, maternal nutritional status in pregnancy, and blood

pressure at age 17: a population based analysis. BMJ1997; 315:

44953.

26 Lucas A, Fewtrell MS, Cole TJ. Fetal origins of adult diseasethe

hypothesis revisited. BMJ1999; 319: 24549.

27 Leon DA, Lithell HO, Vagero D, et al. Reduced fetal growth rate

and increased risk of death from ischaemic heart disease: cohort

study of 15 000 Swedish men and women born 191529. BMJ

1998; 317: 24145.

28 Cook NR, Cohen J, Herbert PR, Taylor JO, Hennekens CH.

Implications of small reductions in diastolic blood pressure for

primary prevention.Arch Intern Med1995; 155: 70109.

29 Ebrahim S, Smith GD. Lowering blood pressure: a systematic

review of sustained effects of non-pharmacological interventions.

J Public Health Med1998; 20: 44148.

30 Garofalo RP, Goldman AS. Cytokines, chemokines, and colony-

stimulating factors in human milk: the 1997 update. Biol Neonate

1998; 74: 13442.

ARTICLES


Early nutrition in preterm infants and later blood pressure: twocohorts after randomised trials

Documents

Transcript of Early nutrition in preterm infants and later blood pressure: twocohorts after randomised trials